Nondestructive Evaluation and Quality Control. Metals Handbook

ASM INTERNATIONAL

The Materials Information Company

®

Publication Information and Contributors

Nondestructive Evaluation and Quality Control was published in 1989 as Volume 17 of the 9th Edition Metals Handbook. With the second printing (1992), the series title was changed to ASM Handbook. The Volume was prepared under the direction of the ASM Handbook Committee.

Authors and Reviewers • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

Rafael Menezes Nunes UFRGS D.A. Aldrich Idaho National Engineering Laboratory EG&G Idaho, Inc. Craig E. Anderson Nuclear Energy Services Gerald L. Anderson American Gas and Chemical Company Glenn Andrews Ultra Image International Bruce Apgar DuPont NDT Systems R.A. Armistead Advanced Research and Applications Corporation Ad Asead University of Michigan at Dearborn David Atherton Queen's University Yoseph Bar-Cohen Douglas Aircraft Company McDonnell Douglas Corporation R.C. Barry Lockheed Missiles & Space Company, Inc. John Bassart Iowa State University George Becker DuPont NDT Systems R.E. Beissner Southwest Research Institute Alan P. Berens University of Dayton Research Institute Harold Berger Industrial Quality, Inc. Henry Bertoni Polytechnic University of New York R.A. Betz Lockheed Missiles & Space Company, Inc. Craig C. Biddle United Technologies Research Center Kelvin Bishop Tennessee Valley Authority Carl Bixby Zygo Corporation Dave Blackham Consultant Gilbert Blake Wiss, Janney, Elstner Associates James Bolen Northrop Aircraft Division Jim Borges Intec Corporation J.S. Borucki Ardox Inc. Richard Bossi Boeing Aerospace Division The Boeing Company Byron Brendan Battelle Pacific Northwest Laboratories G.L. Burkhardt Southwest Research Institute Paul Burstein Skiametrics, Inc. Willard L. Castner National Aeronautics and Space Administration Lyndon B. Johnson Space Center V.S. Cecco Atomic Energy of Canada, Ltd. Chalk River Nuclear Laboratories Francis Chang General Dynamics Corporation Tsong-how Chang University of Wisconsin, Milwaukee F.P. Chiang Laboratory for Experimental Mechanics Research State University of New York at Stony Brook D.E. Chimenti Wright Research & Development Center Wright-Patterson Air Force Base P. Cielo National Research Council of Canada Industrial Materials Research Institute T.N. Claytor Los Alamos National Laboratory J.M. Coffey CEGB Scientific Services J.F. Cook Idaho National Engineering Laboratory EG&G Idaho, Inc. Thomas D. Cooper Wright Research & Development Center Wright-Patterson Air Force Base William D. Cowie United States Air Force Aeronautical Systems Division

• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

L.D. Cox General Dynamics Corporation Robert Cribbs Folsom Research Inc. J.P. Crosson Lucius Pitkin, Inc. Darrell Cutforth Argonne National Laboratory William Dance LTV Missiles & Electronics Group Steven Danyluk University of Illinois Oliver Darling Spectrum Marketing, Inc. E.A. Davidson Wright Research & Development Center Wright-Patterson Air Force Base Vance Deason EG&G Idaho, Inc. John DeLong Philadelphia Electric Company Michael J. Dennis NDE Systems & Services General Electric Company Richard DeVor University of Illinois at Urbana-Champaign Robert L. Ditz GE Aircraft Engines General Electric Company Kevin Dooley University of Minnesota Thomas D. Dudderar AT&T Bell Laboratories Charles D. Ehrlich National Institute of Standards & Technology Ralph Ekstrom University of Nebraska Lincoln Robert Erf United Technologies Research Center K. Erland United Technologies Corporation Pratt & Whitney Group J.L. Fisher Southwest Research Institute Colleen Fitzpatrick Spectron Development Laboratory William H. Folland United Technologies Corporation Pratt & Whitney Group Joseph Foster Texas A&M University Kenneth Fowler Panametrics, Inc. E.M. Franklin Argonne National Laboratory Argonne--West Larry A. Gaylor Dexter Water Management Systems David H. Genest Brown & Sharpe Manufacturing Company Dennis German Ford Motor Company Ron Gerow Consultant Scott Giacobbe GPU Nuclear Robert S. Gilmore General Electric Research and Development Center J.N. Gray Center for NDE Iowa State University T.A. Gray Center for NDE Iowa State University Robert E. Green, Jr. The Johns Hopkins University Arnold Greene Micro/Radiographs Inc. Robert Grills Ultra Image International Donald Hagemaier Douglas Aircraft Company McDonnell Douglas Corporation John E. Halkias General Dynamics Corporation Grover L. Hardy Wright Research & Development Center Wright-Patterson Air Force Base Patrick G. Heasler Battelle Pacific Northwest Laboratories Charles J. Hellier Hellier Associates, Inc. Edmond G. Henneke Virginia Polytechnic Institute and State University B.P. Hildebrand Failure Analysis Associates, Inc. Howard E. Housermann ZETEC, Inc. I.C.H. Hughes BCIRA International Centre Phil Hutton Battelle Pacific Northwest Laboratories Frank Iddings Southwest Research Institute Bruce G. Isaacson Bio-Imaging Research, Inc. W.B. James Hoeganaes Corporation D.C. Jiles Iowa State University Turner Johnson Brown & Sharpe Manufacturing Company John Johnston Krautkramer Branson William D. Jolly Southwest Research Institute M.H. Jones Los Alamos National Laboratory

• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

Gail Jordan Howmet Corporation William T. Kaarlela General Dynamics Corporation Robert Kalan Naval Air Engineering Center Paul Kearney Welch Allyn Inc. William Kennedy Canadian Welding Bureau Lawrence W. Kessler Sonoscan, Inc. Thomas G. Kincaid Boston University Stan Klima NASA Lewis Research Center Kensi Krzywosz Electric Power Research Institute Nondestructive Evaluation Center David Kupperman Argonne National Laboratory H. Kwun Southwest Research Institute J.W. Lincoln Wright Research & Development Center Wright-Patterson Air Force Base Art Lindgren Magnaflux Corporation D. Lineback Measurements Group, Inc. Charles Little Sandia National Laboratories William Lord Iowa State University D.E. Lorenzi Magnaflux Corporation Charles Loux GE Aircraft Engines General Electric Company A. Lucero ZETEC, Inc. Theodore F. Luga Consultant William McCroskey Innovative Imaging Systems, Inc. Ralph E. McCullough Texas Instruments, Inc. William E.J. McKinney DuPont NDT Systems Brian MacCracken United Technologies Corporation Pratt & Whitney Group Ajit K. Mal University of California, Los Angeles A.R. Marder Energy Research Center Lehigh University Samuel Marinov Western Atlas International, Inc. George A. Matzkanin Texas Research Institute John D. Meyer Tech Tran Consultants, Inc. Morey Melden Spectrum Marketing, Inc. Merlin Michael Rockwell International Carol Miller Wright Research & Development Center Wright-Patterson Air Force Base Ron Miller MQS Inspection, Inc. Richard H. Moore CMX Systems, Inc. Thomas J. Moran Consultant John J. Munro III RTS Technology Inc. N. Nakagawa Center for NDE Iowa State University John Neuman Laser Technology, Inc. H.I. Newton Babcock & Wilcox G.B. Nightingale General Electric Company Mehrdad Nikoonahad Bio-Imaging Research, Inc. R.C. O'Brien Hoeganaes Corporation Kanji Ono University of California, Los Angeles Vicki Panhuise Allied-Signal Aerospace Company Garrett Engine Division James Pellicer Staveley NDT Technologies, Inc. Robert W. Pepper Textron Specialty Materials C.C. Perry Consultant John Petru Kelly Air Force Base Richard Peugeot Peugeot Technologies, Inc. William Plumstead Bechtel Corporation Adrian Pollock Physical Acoustic Corporation George R. Quinn Hellier Associates, Inc. Jay Raja Michigan Technological University Jack D. Reynolds General Dynamics Corporation

• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

William L. Rollwitz Southwest Research Institute A.D. Romig, Jr. Sandia National Laboratories Ward D. Rummel Martin Marietta Astronautics Group Charles L. Salkowski National Aeronautics and Space Administration Lyndon B. Johnson Space Center Thomas Schmidt Consultant Gerald Scott Martin Marietta Manned Space Systems D.H. Shaffer Westinghouse Electric Corporation Research and Development Center Charles N. Sherlock Chicago Bridge & Iron Company Thomas A. Siewert National Institute of Standards and Technology Peter Sigmund Lindhult & Jones, Inc. Lawrence W. Smiley Reliable Castings Corporation James J. Snyder Westinghouse Electric Company Oceanic Division Doug Steele GE Aircraft Engines General Electric Company John M. St. John Caterpillar, Inc. Bobby Stone Jr. Kelly Air Force Base George Surma Sundstrand Aviation Operations Lyndon J. Swartzendruber National Institute of Standards and Technology Richard W. Thams X-Ray Industries, Inc. Graham H. Thomas Sandia National Laboratories R.B. Thompson Center for NDE Iowa State University Virginia Torrey Welch Allyn Inc. James Trolinger Metro Laser Michael C. Tsao Ultra Image International Glen Wade University of California, Santa Barbara James W. Wagner The Johns Hopkins University Henry J. Weltman General Dynamics Corporation Samuel Wenk Consultant Robert D. Whealy Boeing Commercial Airplane Company David Willis Allison Gas Turbine Division General Motors Corporation Charles R. Wojciechowski NDE Systems and Services General Electric Company J.M. Wolla U.S. Naval Research Laboratory John D. Wood Lehigh University Nello Zuech Vision Systems International

Foreword Volume 17 of Metals Handbook is a testament to the growing importance and increased sophistication of methods used to nondestructively test and analyze engineered products and assemblies. For only through a thorough understanding of modern techniques for nondestructive evaluation and statistical analysis can product reliability and quality control be achieved and maintained. As with its 8th Edition predecessor, the aim of this Volume is to provide detailed technical information that will enable readers to select, use, and interpret nondestructive methods. Coverage, however, has been significantly expanded to encompass advances in established techniques as well as introduce the most recent developments in computed tomography, digital image enhancement, acoustic microscopy, and electromagnetic techniques used for stress analysis. In addition, material on quantitative analysis and statistical methods for design and quality control (subjects covered only briefly in the 8th Edition) has been substantially enlarged to reflect the increasing utility of these disciplines. Publication of Volume 17 also represents a significant milestone in the history of ASM International. This Volume completes the 9th Edition of Metals Handbook, the largest single source of information on the technology of metals that has ever been compiled. The magnitude, respect, and success of this unprecedented reference set calls for a special tribute to its many supporters. Over the past 13 years, the ASM Handbook Committee has been tireless in its efforts, ASM members have been unflagging in their support, and the editorial staff devoted and resourceful. Their efforts, combined with the considerable knowledge and technical expertise of literally thousands of authors, contributors, and reviewers,

have resulted in reference books which are comprehensive in coverage and which set the highest standards for quality. To all these men and women, we extend our most sincere appreciation and gratitude. Richard K. Pitler President, ASM International Edward L. Langer Managing Director, ASM International

Preface The subject of nondestructive examination and analysis of materials and manufactured parts and assemblies is not new to Metals Handbook. In 1976, Volume 11 of the 8th Edition--Nondestructive Inspection and Quality Control--provided what was at that time one of the most thorough overviews of this technology ever published. Yet in the relatively short time span since then, tremendous advances and improvements have occurred in the field--so much so that even the terminology has evolved. For example, in the mid-1970s the examination of an object or material that did not render it unfit for use was termed either nondestructive testing (NDT) or nondestructive inspection (NDI). Both are similar in that they involve looking at (or through) an object to determine either a specific characteristic or whether the object contains discontinuities, or flaws. The refinement of existing methods, the introduction of new methods, and the development of quantitative analysis have led to the emergence of a third term over the past decade, a term representing a more powerful tool. With nondestructive evaluation (NDE), a discontinuity can be classified by its size, shape, type, and location, allowing the investigator to determine whether or not the flaw is acceptable. The title of the present 9th Edition volume was modified to reflect this new technology. Volume 17 is divided into five major sections. The first contains four articles that describe equipment and techniques used for qualitative part inspection. Methods for both defect recognition (visual inspection and machine vision systems) and dimensional measurements (laser inspection and coordinate measuring machines) are described. In the second section, 24 articles describe the principles of a wide variety of nondestructive techniques and their application to quality evaluation of metallic, composite, and electronic components. In addition to detailed coverage of more commonly used methods (such as magnetic particle inspection, radiographic inspection, and ultrasonic inspection), newly developed methods (such as computed tomography, acoustic microscopy, and speckle metrology) are introduced. The latest developments in digital image enhancement are also reviewed. Finally, a special six-page color section illustrates the utility of color-enhanced images. The third section discusses the application of nondestructive methods to specific product types, such as one-piece products (castings, forgings, and powder metallurgy parts) and assemblies that have been welded, soldered, or joined with adhesives. Of particular interest is a series of reference radiographs presented in the article "Weldments, Brazed Assemblies, and Soldered Joints" that show a wide variety of weld discontinuities and how they appear as radiographic images. The reliability of discontinuity detection by nondestructive methods, referred to as quantitative NDE, is the subject of the fourth section. Following an introduction to this rapidly maturing discipline, four articles present specific guidelines to help the investigator determine the critical discontinuity size that will cause failure, how long a structure containing a discontinuity can be operated safely in service, how a structure can be designed to prevent catastrophic failure, and what inspections must be performed in order to prevent failure. The final section provides an extensive review of the statistical methods being used increasingly for design and quality control of manufactured products. The concepts of statistical process control, control charts, and design of experiments are presented in sufficient detail to enable the reader to appreciate the importance of statistical analysis and to organize and put into operation a system for ensuring that quality objectives are met on a consistent basis. This Volume represents the collective efforts of nearly 200 experts who served as authors, contributors of case histories, or reviewers. To all we extend our heartfelt thanks. We would also like to acknowledge the special efforts of Thomas D.

Cooper (Wright Research & Development Center, Wright-Patterson Air Force Base) and Vicki E. Panhuise (AlliedSignal Aerospace Company, Garrett Engine Division). Mr. Cooper, a former Chairman of the ASM Handbook Committee, was instrumental in the decision to significantly expand the material on quantitative analysis. Dr. Panhuise organized the content and recruited all authors for the section "Quantitative Nondestructive Evaluation." Such foresight and commitment from Handbook contributors over the years has helped make the 9th Edition of Metals Handbook--all 17 volumes and 15,000 pages--the most authoritative reference work on metals ever published. The Editors

General Information Officers and Trustees of ASM International Officers

• • • •

Richard K. Pitler President and Trustee Allegheny Ludlum Corporation (retired) Klaus M. Zwilsky Vice President and Trustee National Materials Advisory Board Academy of Sciences William G. Wood Immediate Past President and Trustee Kolene Corporation Robert D. Halverstadt Treasurer AIMe Associates

Trustees

• • • • • • • • • •

John V. Andrews Teledyne Allvac Edward R. Burrell Inco Alloys International, Inc. Stephen M. Copley University of Southern California H. Joseph Klein Haynes International, Inc. Gunvant N. Maniar Carpenter Technology Corporation Larry A. Morris Falconbridge Limited William E. Quist Boeing Commercial Airplanes Charles Yaker Howmet Corporation Daniel S. Zamborsky Consultant Edward L. Langer Managing Director ASM International

Members of the ASM Handbook Committee (1988-1989) • • • • • • • • • • • • • • • • •

Dennis D. Huffman (Chairman 1986-; Member 1983-) The Timken Company Roger J. Austin (1984-) ABARIS Roy G. Baggerly (1987-) Kenworth Truck Company Robert J. Barnhurst (1988-) Noranda Research Centre Peter Beardmore (1986-1989) Ford Motor Company Hans Borstell (1988-) Grumman Aircraft Systems Gordon Bourland (1988-) LTV Aerospace and Defense Company Robert D. Caligiuri (1986-1989) Failure Analysis Associates Richard S. Cremisio (1986-1989) Rescorp International, Inc. Gerald P. Fritzke (1988-) Metallurgical Associates J. Ernesto Indacochea (1987-) University of Illinois at Chicago John B. Lambert (1988-) Fansteel Inc. James C. Leslie (1988-) Advanced Composites Products and Technology Eli Levy (1987-) The De Havilland Aircraft Company of Canada Arnold R. Marder (1987-) Lehigh University John E. Masters (1988-) American Cyanamid Company L.E. Roy Meade (1986-1989) Lockheed-Georgia Company

National

• • • • • • • •

Merrill L. Minges (1986-1989) Air Force Wright Aeronautical Laboratories David V. Neff (1986-) Metaullics Systems Dean E. Orr (1988-) Orr Metallurgical Consulting Service, Inc. Ned W. Polan (1987-1989) Olin Corporation Paul E. Rempes (1986-1989) Williams International E. Scala (1986-1989) Cortland Cable Company, Inc. David A. Thomas (1986-1989) Lehigh University Kenneth P. Young (1988-) AMAX Research & Development

Previous Chairmen of the ASM Handbook Committee • • • • • • • • • • • • • • • • • • • • • •

R.S. Archer (1940-1942) (Member, 1937-1942) L.B. Case (1931-1933) (Member, 1927-1933) T.D. Cooper (1984-1986) (Member, 1981-1986) E.O. Dixon (1952-1954) (Member, 1947-1955) R.L. Dowdell (1938-1939) (Member, 1935-1939) J.P. Gill (1937) (Member, 1934-1937) J.D. Graham (1966-1968) (Member, 1961-1970) J.F. Harper (1923-1926) (Member, 1923-1926) C.H. Herty, Jr. (1934-1936) (Member, 1930-1936) J.B. Johnson (1948-1951) (Member, 1944-1951) L.J. Korb (1983) (Member, 1978-1983) R.W.E. Leiter (1962-1963) (Member, 1955-1958, 1960-1964) G.V. Luerssen (1943-1947) (Member, 1942-1947) G.N. Maniar (1979-1980) (Member, 1974-1980) J.L. McCall (1982) (Member, 1977-1982) W.J. Merten (1927-1930) (Member, 1923-1933) N.E. Promisel (1955-1961) (Member, 1954-1963) G.J. Shubat (1973-1975) (Member, 1966-1975) W.A. Stadtler (1969-1972) (Member, 1962-1972) R. Ward (1976-1978) (Member, 1972-1978) M.G.H. Wells (1981) (Member, 1976-1981) D.J. Wright (1964-1965) (Member, 1959-1967)

Staff ASM International staff who contributed to the development of the Volume included Joseph R. Davis, Manager of Handbook Development; Kathleen M. Mills, Manager of Book Production; Steven R. Lampman, Technical Editor; Theodore B. Zorc, Technical Editor; Heather J. Frissell, Editorial Supervisor; George M. Crankovic, Editorial Coordinator; Alice W. Ronke, Assistant Editor; Jeanne Patitsas, Word Processing Specialist; and Karen Lynn O'Keefe, Word Processing Specialist. Editorial assistance was provided by Lois A. Abel, Wendy L. Jackson, Robert T. Kiepura, Penelope Thomas, and Nikki D. Wheaton. The Volume was prepared under the direction of Robert L. Stedfeld, Director of Reference Publications. Conversion to Electronic Files ASM Handbook, Volume 17, Nondestructive Evaluation and Quality Control was converted to electronic files in 1998. The conversion was based on the fifth printing (1997). No substantive changes were made to the content of the Volume, but some minor corrections and clarifications were made as needed. ASM International staff who contributed to the conversion of the Volume included Sally Fahrenholz-Mann, Bonnie Sanders, Marlene Seuffert, Gayle Kalman, Scott Henry, Robert Braddock, Alexandra Hoskins, and Erika Baxter. The electronic version was prepared under the direction of William W. Scott, Jr., Technical Director, and Michael J. DeHaemer, Managing Director.

Copyright Information (for Print Volume) Copyright © 1989 ASM International. All rights reserved No part of this book may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the written permission of the copyright owner. First printing, September 1989 Second printing, May 1992 Third printing, May 1994 Fourth printing, January 1996 Fifth printing, December 1997 ASM Handbook is a collective effort involving thousands of technical specialists. It brings together in one book a wealth of information from world-wide sources to help scientists, engineers, and technicians solve current and long-range problems. Great care is taken in the compilation and production of this Volume, but it should be made clear that no warranties, express or implied, are given in connection with the accuracy or completeness of this publication, and no responsibility can be taken for any claims that may arise. Nothing contained in the ASM Handbook shall be construed as a grant of any right of manufacture, sale, use, or reproduction, in connection with any method, process, apparatus, product, composition, or system, whether or not covered by letters patent, copyright, or trademark, and nothing contained in the ASM Handbook shall be construed as a defense against any alleged infringement of letters patent, copyright, or trademark, or as a defense against any liability for such infringement. Comments, criticisms, and suggestions are invited, and should be forwarded to ASM International. Library of Congress Cataloging-in-Publication Data (for Print Volume) Metals handbook. Includes bibliographies and indexes. Contents: v. 1. Properties and selection--v. 2. Properties and selection--nonferrous alloysand pure metals--[etc.]--v. 17. Nondestructiveevaluation and quality control. 1. Metals--Handbooks, manuals, etc. I. ASM Handbook Committee. II. ASM International. Handbook Committee. TA459.M43 1978 669 78-14934 ISBN 0-87170-007-7 (v. 1) SAN 204-7586

Visual Inspection

Introduction VISUAL INSPECTION is a nondestructive testing technique that provides a means of detecting and examining a variety of surface flaws, such as corrosion, contamination, surface finish, and surface discontinuities on joints (for example, welds, seals, solder connections, and adhesive bonds). Visual inspection is also the most widely used method for detecting and examining surface cracks, which are particularly important because of their relationship to structural failure mechanisms. Even when other nondestructive techniques are used to detect surface cracks, visual inspection often provides a useful supplement. For example, when the eddy current examination of process tubing is performed, visual inspection is often performed to verify and more closely examine the surface disturbance. Given the wide variety of surface flaws that may be detectable by visual examination, the use of visual inspection may encompass different techniques, depending on the product and the type of surface flaw being monitored. This article focuses on some equipment used to aid the process of visual inspection. The techniques and applicability of visual inspection for some products are considered in the Selected References in this article and in the Section "Nondestructive Inspection of Specific Products" in this Volume. The methods of visual inspection involve a wide variety of equipment, ranging from examination with the naked eye to the use of interference microscopes for measuring the depth of scratches in the finish of finely polished or lapped surfaces. Some of the equipment used to aid visual inspection includes: • • • •

Flexible or rigid borescopes for illuminating and observing internal, closed or otherwise inaccessible areas Image sensors for remote sensing or for the development of permanent visual records in the form of photographs, videotapes, or computer-enhanced images Magnifying systems for evaluating surface finish, surface shapes (profile and contour gaging), and surface microstructures Dye and fluorescent penetrants and magnetic particles for enhancing the observation of surface cracks (and sometimes near-surface conditions in the case of magnetic particle inspection)

This article will review the use of the equipment listed above in visual inspection, except for dye penetrants and magnetic particles, which are discussed in the articles "Liquid Penetrant Inspection" and "Magnetic Particle Inspection," respectively, in this Volume.

Acknowledgements ASM International would like to thank Oliver Darling and Morley Melden of Spectrum Marketing, Inc., for their assistance in preparing the section on borescopes. They provided a draft of a textbook being developed for Olympus Corporation. Thanks are also extended to Virginia Torrey of Welch Allyn, Inc., for the information on videoscopes and to Peter Sigmund of Lindhult and Jones, Inc., for the information on instruments from Lenox, Inc. Visual Inspection

Borescopes A borescope (Fig. 1) is a long, tubular optical device that illuminates and allows the inspection of surfaces inside narrow tubes or difficult-to-reach chambers. The tube, which can be rigid or flexible with a wide variety of lengths and diameters, provides the necessary optical connection between the viewing end and an objective lens at the distant, or distal, tip of the borescope. This optical connection can be achieved in one of three different ways:

• • •

By using a rigid tube with a series of relay lenses By using a tube (normally flexible but also rigid) with a bundle of optical fibers By using a tube (normally flexible) with wiring that carries the image signal from a charge-coupled device (CCD) imaging sensor at the distal tip

These three basic tube designs can have either fixed or adjustable focusing of the objective lens at the distal tip. The distal tip also has prisms and mirrors that define the direction and field of view (see Fig. 2). These views vary according to the type and application of borescope. The design of illumination system also varies with the type of borescope. Generally, a fiber optic light guide and a lamp producing white light is used in the illumination system, although ultraviolet light can be used to inspect surfaces treated with liquid fluorescent penetrants. Light-emitting diodes at the distal tip are sometimes used for illumination in videoscopes with working lengths greater than 15 m (50 ft).

Fig. 1 Three typical designs of borescopes. (a) A rigid borescope with a lamp at the distal end. (b) A flexible fiberscope with a light source. (c) A rigid borescope with a light guide bundle in the shaft

Rigid Borescopes Rigid borescopes are generally limited to applications with a straight-line path between the observer and the area to be observed. The sizes range in lengths from 0.15 to 30 m (0.5 to 100 ft) and in diameters from 0.9 to 70 mm (0.035 to 2.75 in.). Magnification is usually 3 to 4×, but powers up to 50× are available. The illumination system is either an incandescent lamp located at the distal tip end (Fig. 1a) or a light guide bundle made from optical fibers (Fig. 1c) that conduct light from an external source.

Fig. 2 Typical directions and field of view with rigid borescopes

The choice of viewing heads for rigid borescopes (Fig. 2) varies according to the application, as described in the section "Selection" in this article. Rigid borescopes generally have a 55° field of view, although the fields of view can range from 10 to 90°. Typically, the distal tips are not interchangeable, but some models (such as the extendable borescopes) may have interchangeable viewing heads.

Some rigid borescopes have orbital scan (Fig. 1c), which involves the rotation of the optical shaft for scanning purposes. Depending on the borescope model, the amount of rotation can vary from 120 to 370°. Some rigid borescopes also have movable prisms at the tip for scanning. Rigid borescopes are available in a variety of models having significant variations in the design of the shaft, the distal tip, and the illumination system. Some of these design variations are described below. Basic Design. The rigid borescope typically has a series of achromatic relay lenses in the optical tube. These lenses

preserve the resolution of the image as it travels from the objective lens to the eyepiece. The tube diameter of these borescopes ranges from 4 to 70 mm (0.16 to 2.75 in.). The illumination system can be either a distal lamp or a light guide bundle, and the various features may include orbital scan, various viewing heads, and adjustable focusing of the objective lens. Miniborescopes. Instead of the conventional relay lenses, miniborescopes have a single image-relaying rod or quartz

fiber in the optical tube. The lengths of miniborescopes are 110 and 170 mm (4.3 and 6.7 in.), and the diameters range from 0.9 to 2.7 mm (0.035 to 0.105 in.). High magnification (up to 30×) can be reached at minimal focal lengths, and an adjustable focus is not required, because the scope has an infinite depth of field. The larger sizes have forward, side view, and forward-oblique views. The 0.9 mm (0.035 in.) diam size has only a forward view. Miniborescopes have an integral light guide bundle. Hybrid borescopes utilize rod lenses combined with convex lenses to relay the image. The rod lenses have fewer

glass-air boundaries; this reduces scattering and allows for a more compact optical guide. Consequently, a larger light guide bundle can be employed with an increase in illumination and an image with a higher degree of contrast. Hybrid borescopes have lengths up to 990 mm (39 in.), with diameters ranging from 5.5 to 12 mm (0.216 to 0.47 in.). All hybrid borescopes have adjustable focusing of the objective lens and a 370° rotation for orbital scan. The various viewing directions are forward, side, retrospective, and forward-oblique. Extendable borescopes allow the user to construct a longer borescopic tube by joining extension tubes. Extendable

borescopes are available with either a fiber-optic light guide or an incandescent lamp at the distal end. Extendable borescopes with an integral lamp have a maximum length of about 30 m (100 ft). Scopes with a light guide bundle have a shorter maximum length (about 8 m, or 26 ft), but do allow smaller tube diameters (as small as 8 mm, or 0.3 in.). Interchangeable viewing heads are also available. Extendable borescopes do not have adjustable focusing of the objective lens.

Rigid chamberscopes allow more rapid inspection of larger chambers. Chamberscopes (Fig. 3) have variable

magnification (zoom), a lamp at the distal tip, and a scanning mirror that allows the user to observe in different directions. The higher illumination and greater magnification of chamberscopes allow the inspection of surfaces as much as 910 mm (36 in.) away from the distal tip of the scope. Mirror

sheaths can convert a direct-viewing borescope into a side-viewing scope. A mirror sheath is designed to fit over the tip of the scope and thus reflect an image from the side of the scope. However, not all applications are suitable for this device. A side, forward-oblique, or retrospective viewing head provides better resolution and a higher degree of image contrast. A mirror sheath also produces an inverse image and may produce unwanted reflections from the shaft.

Fig. 3 Typical chamberscope. Instrument Company

Courtesy

of

Scanning. In addition to the orbital scan feature described earlier, some rigid borescopes have the ability to scan longitudinally along the axis of the shaft. A Lenox movable prism with a control at the handle accomplishes this scanning. Typically, the prism can shift the direction of view through an arc of 120°.

Flexible Borescopes Flexible borescopes are used primarily in applications that do not have a straight passageway to the point of observation. The two types of flexible borescopes are flexible fiberscopes and videoscopes with a CCD image sensor at the distal tip. Flexible Fiberscopes. A typical fiberscope (Fig. 1b) consists of a light guide bundle, an image guide bundle, an

objective lens, interchangeable viewing heads, and remote controls for articulation of the distal tip. Fiberscopes are available in diameters from 1.4 to 13 mm (0.055 to 0.5 in.) and in lengths up to 12 m (40 ft). Special quartz fiberscopes are available in lengths up to 90 m (300 ft). The fibers used in the light guide bundle are generally 30 m (0.001 in.) in diameter. The second optical bundle, called the image guide, is used to carry the image formed by the objective lens back to the eyepiece. The fibers in the image guide must be precisely aligned so that they are in an identical relative position to each other at their terminations for proper image resolution. The diameter of the fibers in the image guide is another factor in obtaining good image resolution. With smaller diameter fibers, a brighter image with better resolution can be obtained by packing more fibers in the image guide. With higher resolution, it is then possible to use an objective lens with a wider field of view and also to magnify the image at the eyepiece. This allows better viewing of objects at the periphery of the image (Fig. 4). Image guide fibers range from 6.5 to 17 m (255 to 670 in.).

Fig. 4 Two views down a combustor can with the distal tip in the same position. A fiberscope with smaller diameter fibers and 40% more fibers in the image bundle provides better resolution (a) than a fiberscope with larger fibers (b). Courtesy of Olympus Corporation

The interchangeable distal tips provide various directions and fields of view on a single fiberscope. However, because the tip can be articulated for scanning purposes, distal tips with either a forward or side viewing direction are usually sufficient. Fields of view are typically 40 to 60°, although they can range from 10 to 120°. Most fiberscopes provide adjustable focusing of the objective lens. Videoscopes with CCD probes involve the electronic transmission of color or black and white images to a video monitor. The distal end of electronic videoscopes contains a CCD chip, which consists of thousands of light-sensitive elements arrayed in a pattern of rows and columns. The objective lens focuses the image of an object on the surface of the CCD chip, where the light is converted to electrons that are stored in each picture element, or pixel, of the CCD device. The image of the object is thus stored in the form of electrons on the CCD device. At this point, a voltage proportional to the number of electrons at each pixel is determined electronically for each pixel site. This voltage is then amplified, filtered, and sent to the input of a video monitor.

Videoscopes with CCD probes produce images (Fig. 5) with spatial resolutions of the order of those described in Fig. 6. Like rigid borescopes and flexible fiberscopes, the resolution of videoscopes depends on the object-to-lens distance and the fields of view, because these two factors affect the amount of magnification (see the section "Magnification and Field of View" in this article). Generally, videoscopes produce higher resolution than fiberscopes, although fiberscopes with smaller diameter fibers (Fig. 4a) may be competitive with the resolution of videoscopes.

Fig. 5 Videoscope images (a) inside engine guide vanes (b) of an engine fuel nozzle. Courtesy of Welch Allyn, Inc.

Fig. 6 Typical resolution of CCD videoscopes with a 90° field of view (a), 60° field of view (b), 30° field of view (c). Source: Welch Allyn, Inc.

Another advantage of videoscopes is their longer working length. With a given amount of illumination at the distal tip, videoscopes can return an image over a greater length than fiberscopes. Other features of videoscopes include: • • • •

The display can help reduce eye fatigue (but does not allow the capability of direct viewing through an eyepiece) There is no honeycomb pattern or irregular picture distortion as with some fiberscopes (Fig. 7) The electronic form of the image signal allows digital image enhancement and the potential for integration with automatic inspection systems. The display allows the generation of reticles on the viewing screen for point-to-point measurements.

Fig. 7 Image from a videoscope (a) and a fiberscope (b). In some fiberscope images, voids between individual glass fibers can create a honeycomb pattern that adds graininess to the image. Courtesy of Welch Allyn, Inc.

Special Features Measuring borescopes and fiberscopes contain a movable cursor that allows measurements during viewing (Fig. 8). When the object under measurement is in focus, the movable cursor provides a reference for dimensional measurements in the optical plane of the object. This capability eliminates the need to know the object-to-lens distance when determining magnification factors.

Working channels are used in borescopes and fiberscopes to pass working

devices to the distal tip. Working channels are presently used to pass measuring instruments, retrieval devices, and hooks for aiding the insertion of thin, flexible fiberscopes. Working channels are used in flexible fiberscopes with diameters as small as 2.7 mm (0.106 in.). Working channels are also under consideration for the application and removal of dye penetrants and for the passage of wires and sensors in eddy current measurements. Selection

Fig. 8 View through a measuring fiberscope with reticles for 20° and 40° fieldof-view lenses. Courtesy of Olympus Corporation

Flexible and rigid borescopes are available in a wide variety of standard and customized designs, and several factors can influence the selection of a scope for a particular application. These factors include focusing, illumination, magnification, working length, direction of view, and environment. Focusing and Resolution. If portions of long objects are at different planes, the

scope must have sufficient focus adjustment to achieve an adequate depth of field. If the scope has a fixed focal length, the object will be in focus only at a specific lensto-object distance.

To allow the observation of surface detail at a desired size, the optical system of a borescope must also provide adequate resolution and image contrast. If resolution is adequate but contrast is lacking, detail cannot be observed. In general, the optical quality of a rigid borescope improves as the size of the lens increases; consequently, a borescope with the largest possible diameter should be used. For fiberscopes, the resolution is dependent on the accuracy of alignment and the diameter of the fibers in the image bundle. Smaller-diameter fibers provide more resolution and edge contrast (Fig. 4), when combined with good geometrical alignment of the fibers. Typical resolutions of videoscopes are given in Fig. 6. Illumination. The required intensity of the light source is determined by the reflectivity of the surface, the area of

surface to be illuminated, and the transmission losses over the length of the scope. At working lengths greater than 6 m (20 ft), rigid borescopes with a lamp at the distal end provide the greatest amount of illumination over the widest area. However, the heat generated by the light source may deform rubber or plastic materials. Fiber-optic illumination in scopes with working lengths less than 6 m (20 ft) is always brighter and is suitable for heat-sensitive applications because filters can remove infrared frequencies. Because the amount of illumination depends on the diameter of the light guide bundle, it is desirable to use the largest diameter possible. Magnification and field of view are interrelated; as magnification is increased, the field of view is reduced. The

precise relationship between magnification and field of view is specified by the manufacturer. The degree of magnification in a particular application is determined by the field of view and the distance from the objective lens to the object. Specifically, the magnification increases when either the field of view or the lens-to-object distance decreases. Working Length. In addition to the obvious need for a scope of sufficient length, the working length can sometimes

dictate the use of a particular type of scope. For example, a rigid borescope with a long working length may be limited by the need for additional supports. In general, videoscopes allow a longer working length than fiberscopes. Direction of View. The selection of a viewing direction is influenced by the location of the access port in relation to the

object to be observed. The following sections describe some criteria for choosing the direction of view shown in Fig. 2. Flexible fiberscopes or videoscopes, because of their articulating tip, are often adequate with either a side or forward viewing tip. Circumferential or panoramic heads are designed for the inspection of tubing or other cylindrical structures. A centrally located mirror permits right-angle viewing of an area just scanned by the panoramic view. The forward viewing head permits the inspection of the area directly ahead of the viewing head. It is commonly used when examining facing walls or the bottoms of blind holes and cavities.

Forward-oblique heads bend the viewing direction at an angle to the borescope axis, permitting the inspection of corners at the end of a bored hole. The retrospective viewing head bends the cone of view at a retrospective angle to the borescope axis, providing a view of the area just passed by the advancing borescope. It is especially suited to inspecting the inside neck of cylinders and bottles. Environment. Flexible and rigid borescopes can be manufactured to withstand a variety of environments. Although most scopes can operate at temperatures from -34 to 66 °C (-30 to 150 °F), especially designed scopes can be used at temperatures to 1925 °C (3500 °F). Scopes can also be manufactured for use in liquid media.

Special scopes are required for use in pressures above ambient and in atmospheres exposed to radiation. Radiation can cause the multicomponent lenses and image bundles to turn brown. When a scope is used in atmospheres exposed to radiation, quartz fiberscopes are generally used. Scopes used in a gaseous environment should be made explosionproof to minimize the potential of an accidental explosion. Applications Rigid and flexible borescopes are available in different designs suitable for a variety of applications. For example, when inspecting straight process piping for leaks rigid borescopes with a 360° radial view are capable of examining inside diameters of 3 to 600 mm (0.118 to 24 in.). Scopes are also used by building inspectors and contractors to see inside walls, ducts, large tanks, or other dark areas. The principal use of borescope is in equipment maintenance programs, in which borescopes can reduce or eliminate the need for costly teardowns. Some types of equipment, such as turbines, have access ports that are specifically designed for borescopes. Borescopes provide a means of checking in-service defects in a variety of equipment, such as turbines (Fig. 9), automotive components (Fig. 10), and process piping (Fig. 11).

Fig. 9 Turbine flaws seen through a flexible fiberscope. (a) Crack near a fuel burner nozzle. (b) Crack in an outer combustion liner. (c) Combustion chamber and high pressure nozzle guide vanes. (d) Compressor damage showing blade deformation. Courtesy of Olympus Corporation

Fig. 10 In-service defects as seen through a borescope designed for automotive servicing. (a) Carbon on valves. (b) Broken transmission gear tooth. (c) Differential gear wear. Courtesy of Lenox Instrument Company

Fig. 11 Operator viewing a weld 21 m (70 ft) inside piping with a videoscope. Courtesy of Olympus Corporation

Borescopes are also extensively used in a variety of manufacturing industries to ensure the product quality of difficult-toreach components. Manufacturers of hydraulic cylinders, for example, use borescopes to examine the interiors of bores for pitting, scoring, and tool marks. Aircraft and aerospace manufacturers also use borescopes to verify the proper placement and fit of seals, bonds, gaskets, and subassemblies in difficult-to-reach regions. Visual Inspection

Optical Sensors Visible light, which can be detected by the human eye or with optical sensors, has some advantages over inspection methods based on nuclear, microwave, or ultrasound radiation. For example, one of the advantages of visible light is the capability of tightly focusing the probing beam on the inspected surface (Ref 1). High spatial resolution can result from this sharp focusing, which is useful in gaging and profiling applications (Ref 1). Some different types of image sensors used in visual inspection include: • • •

Vidicon or plumbicon television tubes Secondary electron-coupled (SEC) vidicons Image orthicons and image isocons

• •

Charge-coupled device sensors Holographic plates (see the article "Optical Holography" in this Volume)

Television cameras with vidicon tubes are useful at higher light levels (about 0.2 lm/m2, or 10-2 ftc), while orthicons, isocons, and SEC vidicons are useful at lower light levels. The section "Television Cameras" in the article "Radiographic Inspection" in this Volume describes these cameras in more detail. Charge-coupled devices are suitable for many different information-processing applications, including image sensing in television-camera technology. Charge-coupled devices offer a clear advantage over vacuum-tube image sensors because of the reliability of their solid-state technology, their operation at low voltage and low power dissipation, extensive dynamic range, visible and near-infrared response, and geometric reproducibility of image location. Image enhancement (or visual feedback into robotic systems) typically involve the use of CCDs as the optical sensor or the use of television signals that are converted into digital form. Optical sensors are also used in inspection applications that do not involve imaging. The articles "Laser Inspection" and "Speckle Metrology" in this Volume describe the use of optical sensors when laser light is the probing tool. In some applications, however, incoherent light sources are very effective in non-imaging inspection applications utilizing optical sensors.

Example 1: Monitoring Surface Roughness on a Fast-Moving Cable. A shadow projection configuration that can be used at high extrusion speeds is shown in Fig. 12. A linear-filament lamp is imaged by two spherical lenses of focal length f1 on a large-area single detector. Two cylindrical lenses are used to project and recollimate a laminar light beam of uniform intensity, nearly 0.5 mm (0.02 in.) wide across the wire situated near their common focal plane. The portion of the light beam that is not intercepted by the wire is collected on the detector, which has an alternating current output that corresponds to the defect-related wire diameter fluctuations. The wire speed is limited only by the detector response time. With a moderate detector bandwidth of 100 kHz, wire extrusion speeds up to 50 m/s (160 ft/s) can be accepted. Moreover, the uniformity of the nearly collimated projected beam obtained with such a configuration makes the detected signal relatively independent of the random wire excursions in the plane of Fig. 12. It should be mentioned that the adoption of either a single He-Ne laser or an array of fiber-pigtailed diode lasers proved to be inadequate in this case because of speckle noise, high-frequency laser amplitude or mode-to-mode interference fluctuations, and line nonuniformity.

Fig. 12 Schematic of line projection method for monitoring the surface roughness on fast-moving cables

An industrial prototype of such a sensor was tested on the production line at extruding speeds reaching 30 m/s (100 ft/s). Figure 13 shows the location of the sensor just after the extruder die. Random noise introduced by vapor turbulence could be almost completely suppressed by high-pass filtering. Figure 14 shows two examples of signals obtained with a wire of acceptable and unacceptable surface quality. As shown, a roughness amplitude resolution of a few micrometers can be

obtained with such a device. Subcritical surface roughness levels can thus be monitored for real time control of the extrusion process.

Fig. 13 Setup used in the in-plant trials of the line projection method for monitoring the surface roughness of cables. Courtesy of P. Cielo, National Research Council of Canada

Fig. 14 Examples of signals obtained with the apparatus shown in Fig. 13. (a) Acceptable surface roughness. (b) Unacceptable surface roughness

Reference cited in this section

1. P. Cielo, Optical Techniques for Industrial Inspection, Academic Press, 1988, p 243

Note cited in this section: Example 1 in this section was adapted with permission from P. Cielo, Optical Techniques for Industrial Inspection, Academic Press, 1988. Visual Inspection

Magnifying Systems In addition to the use of microscopes in the metallographic examination of microstructures (see the article "Replication Microscopy Techniques for NDE" in this Volume), magnifying systems are also used in visual reference gaging. When tolerances are too tight to judge by eye alone, optical comparators or toolmakers' microscopes are used to achieve magnifications ranging from 5 to 500×. A toolmakers' microscope consists of a microscope mounted on a base that carries an adjustable stage, a stage transport mechanism, and supplementary lighting. Micrometer barrels are often incorporated into the stage transport mechanism to permit precisely controlled movements, and digital readouts of stage positioning are becoming increasingly available. Various objective lenses provide magnifications ranging from 10 to 200×. Optical comparators (Fig. 15) are magnifying devices that project the silhouette of small parts onto a large projection

screen. The magnified silhouette is then compared against an optical comparator chart, which is a magnified outline drawing of the workpiece being gaged. Optical comparators are available with magnifications ranging from 5 to 500×.

Fig. 15 Schematic of an optical comparator

Parts with recessed contours can also be successfully gaged on optical comparators. This is done with the use of a pantograph. One arm of the pantograph is a stylus that traces the recessed contour of the part, and the other arm carries a follower that is visible in the light path. As the stylus moves, the follower projects a contour on the screen.

Visual Inspection

Reference 1. P. Cielo, Optical Techniques for Industrial Inspection, Academic Press, 1988, p 243 Visual Inspection

Selected References • • • • • • • • •

Robert C. Anderson, Inspection of Metals: Visual Examination, Vol 1, American Society for Metals, 1983 Detecting Susceptibility to Intergranular Attack in Austenitic Stainless Steels, ASTM A 262, Annual Book of ASTM Standards, American Society for Testing and Materials Detecting Susceptibility to Intergranular Attack in Ferritic Stainless Steels, ASTM A 763, Annual Book of ASTM Standards, American Society for Testing and Materials Detecting Susceptibility to Intergranular Corrosion in Severely Sensitized Austenitic Stainless Steel, ASTM A 708, Annual Book of ASTM Standards, American Society for Testing and Materials W.R. DeVries and D.A. Dornfield, Inspection and Quality Control in Manufacturing Systems, American Society of Mechanical Engineers, 1982 C.W. Kennedy and D.E. Andrews, Inspection and Gaging, Industrial Press, 1977 Standard Practice for Evaluating and Specifying Textures and Discontinuities of Steel Castings by Visual Examination, ASTM Standard A 802, American Society for Testing and Materials Surface Discontinuities on Bolts, Screws, and Studs, ASTM F 788, Annual Book of ASTM Standards, American Society for Testing and Materials Visual Evaluation of Color Changes of Opaque Materials, ASTM D 1729, Annual Book of ASTM Standards, American Society for Testing and Materials

Laser Inspection Carl Bixby, Zygo Corporation

Introduction THE FIRST LASER was invented in 1960, and many useful applications of laser light have since been developed for metrology and industrial inspection systems. Laser-based inspection systems have proved useful because they represent a fast, accurate means of noncontact gaging, sorting, and classifying parts. Lasers have also made interferometers a more convenient tool for the accurate measurement of length, displacement, and alignment. Lasers are used in inspection and measuring systems because laser light provides a bright, undirectional, and collimated beam of light with a high degree of temporal (frequency) and spatial coherence. These properties can be useful either singly or together. For example, when lasers are used in interferometry, the brightness, coherence, and collimation of laser light are all important. However, in the scanning, sorting, and triangulation applications described in this article, lasers are used because of the brightness, unidirectionality, and collimated qualities of their light; temporal coherence is not a factor. The various types of laser-based measurement systems have applications in three main areas: •

Dimensional measurement

• •

Velocity measurement Surface inspection

The use of lasers may be desirable when these applications require high precision, accuracy, or the ability to provide rapid, noncontact gaging of soft, delicate, hot, or moving parts. Photodetectors are generally needed in all the applications, and the light variations or interruptions can be directly converted into electronic form. Laser Inspection Carl Bixby, Zygo Corporation

Dimensional Measurements Lasers can be used in several different ways to measure the dimensions and the position of parts. Some of the techniques include: • • • • • • • •

Profile gaging of stationary and moving parts with laser scanning equipment Profile gaging of stationary parts by shadow projection on photodiode arrays Profile gaging of small gaps, and small-diameter parts from diffraction patterns Gaging of surfaces that cannot be seen in profile (such as concave surfaces, gear teeth, or the inside diameters of bores) with laser triangulation sensors Measuring length, alignments, and displacements with interferometers Sorting of parts Three-dimensional gaging of surfaces with holograms Measuring length from the velocity of moving, continuous parts (see the section "Velocity Measurements" in this article)

These techniques provide high degrees of precision and accuracy as well as the capability for rapid, noncontact measurement. Scanning Laser Gage. Noncontact sensors are used in a variety of inspection techniques, such as those involving

capacitive gages, eddy-current gages, and air gages. Optical gages, however, have advantages because the distance from the sensor to the workpiece can be large and because many objects can be measured simultaneously. Moreover, the light variations are directly converted into electronic signals, with the response time being limited only by the photodetector and its electronics. Optical sensors for dimensional gaging employ various techniques, such as shadow projection, diffraction phenomena, and scanning light beams. If the workpiece is small or does not exhibit large or erratic movements, diffraction phenomena or shadow projection on a diode array sensor can work well. However, diffraction techniques become impractical if the object has a dimension of more than a few millimeters or if its movement is large. Shadow projection on a diode array sensor may also be limited if the size or movement of the part is too large. The scanning laser beam technique, on the other hand, is suited to a broad range of product sizes and movements. The concept of using a scanning light beam for noncontact dimensional gaging predates the laser; the highly directional and collimated nature of laser light, however, greatly improves the precision of this method over techniques that use ordinary light. The sensing of outside diameters of cylindrical parts is probably the most common application of a laser scanning gage. A scanning laser beam gage consists of a transmitter, a receiver, and processor electronics (Fig. 1). A thin band of scanning laser light is projected from the transmitter to the receiver. When an object is placed in a beam, it casts a timedependent shadow. Signals from the light entering the receiver are used by the microprocessor to extract the dimension represented by the time difference between the shadow edges. The gages can exhibit accuracies as high as ±0.25 m (±10

μin.) for diameters of 10 to 50 mm (0.5 to 2 in.) For larger parts (diameters of 200 to 450 mm, or 8 to 18 in.), accuracies are less.

Fig. 1 Schematic of a scanning laser gage

There are two general types of scanning laser gages: separable transmitters and receivers designed for in-process applications, and self-contained bench gages designed for off-line applications. The in-process scanning gage consists of a multitasking electronic processor that controls a number of scanners. The separable transmitters and receivers can be configured for different scanning arrangements. Two or more scanners can also be stacked for large parts, or they can be oriented along different axes for dual-axis inspection. High-speed scanners are also available for the detection of small defects, such as lumps, in moving-part applications. The bench gage is compact and can measure a variety of part sizes quickly and easily (Fig. 2). As soon as measurements are taken, the digital readout displays the gaged dimension and statistical data. It indicates the total number of measurements taken, the standard deviation, and the maximum, minimum, and mean readings of each batch tested.

Applications. A wide variety of scanning laser gages

are available to fit specific applications. Measurement capabilities fall within a range of 0.05 to 450 mm (0.002 to 18 in.), with a repeatability of 0.1 m (5 in.) for the smaller diameters. Typical applications include centerless grinding, precision machining, extrusions, razor blades, turbine blades, computer disks, wire lines, and plug gages. Photodiode

Array Imaging. Profile imaging closely duplicates a shadowgraph or contour projector where the ground-glass screen has been replaced by a solid-state diode array image sensor. The measurement system consists of a laser light source, imaging optics, a photodiode array, and signal-processing electronics (Fig. 3). The object casts a shadow, or profile image, of the part on the photodiode array. A scan of the array determines the edge image location and then the location of the part edges from which the dimension of the part can be determined. Fig. 2 Self-contained laser bench micrometer

Fig. 3 Schematic of profile imaging. The laser beam passing the edge of a cylindrical part is imaged by a lens onto a photodiode array. A scan of the array determines the edge image location and then the location of the part edges from which the dimension of the part can be determined.

Accuracies of ±0.05 m (±2 in.) have been achieved with photodiode arrays. For large-diameter parts, two arrays are used--one for each edge. When large or erratic movement of the part is involved, photodiode arrays are not suitable. However, when part rotation is involved, stroboscopic illumination can freeze the image of the part. Diffraction Pattern Technique. Diffraction pattern metrology systems can be used on-line and off-line to measure

small gaps and small diameters of thin wire, needles, and fiber optics. They can also be used to inspect such defects as burrs of hypodermic needles and threads on bolts. In a typical system, the parallel coherent light in a laser beam is diffracted by a small part, and the resultant pattern is focused by a lens on a linear diode array. One significant characteristic of the diffraction pattern is that the smaller the part is, the more accurate the measurement becomes. Diffraction is not suitable for diameters larger than a few millimeters. The distance between the alternating light and dark bands in the diffraction pattern bears a precise mathematical relationship to the wire diameter, the wavelength of the laser beam, and the focal length of the lens. Because the laser

beam wavelength and the lens focal length are known constants of the system, the diameter can be calculated directly from the diffraction pattern measurement. Laser triangulation sensors determine the standoff distance between a surface and a microprocessor-based sensor. Laser triangulation sensors can perform automatic calculations on sheet metal stampings for gap and flushness, hole diameters, and edge locations in a fraction of the time required in the past with manual or ring gage methods.

The principle of single-spot laser triangulation is illustrated in Fig. 4. In this technique, a finely focused laser spot of light is directed at the part surface. As the light strikes the surface, a lens in the sensor images this bright spot onto a solid-state, position-sensitive photodetector. As shown in Fig. 4, the location of the image spot is directly related to the standoff distance form the sensor to the object surface; a change in the standoff distance results in a lateral shift of the spot along the sensor array. The standoff distance is calculated by the sensor processor.

Fig. 4 Schematic of laser triangulation method of measurement. As light strikes the surface, a lens images the point of illumination onto a photosensor. Variations in the surface cause the image dot to move laterally along the photosensor.

Laser triangulation sensors provide quick measurement of deviations due to changes in the surface. With two sensors, the method can be used to measure part thickness or the inside diameters of bores. However, it may not be possible to probe the entire length of the bore. Laser triangulation sensors can also be used as a replacement for tough-trigger probes on coordinate measuring machines. In this application, the sensor determines surface features and surface locations by utilizing an edge-finding device. The accuracy of laser triangulation sensors varies, depending on such performance requirements as standoff and range. Typically, as range requirements increase, accuracy tends to decrease; therefore, specialized multiple-sensor units are designed to perform within various specific application tolerances.

Interferometers provide precise and accurate measurement of relative and absolute length by utilizing the wave

properties of light. They are employed in precision metal finishing, microlithography, and the precision alignment of parts. The basic operational principles of an interferometer are illustrated in Fig. 5. Monochromatic light is directed at a halfsilvered mirror acting as a beam splitter that transmits half the beam to a movable mirror and reflects the other half 90° to a fixed mirror. The reflections from the movable and fixed mirrors are recombined at the beam splitter, where wave interference occurs according to the different path lengths of the two beams.

Fig. 5 Schematic of a basic interferometer

When one of the mirrors is displaced very slowly in a direction parallel to the incident beam without changing its angular alignment, the observer will see the intensity of the recombined beams increasing and decreasing as the light waves from the two paths undergo constructive and destructive interference. The cycle of intensity change from one dark fringe to another represents a half wavelength displacement of movable mirror travel, because the path of light corresponds to two times the displacement of the movable mirror. If the wavelength of the light is known, the displacement of the movable mirror can be determined by counting fringes. Variations on the basic concepts described above produce interferometers in different forms suited to diverse applications. The two-frequency laser interferometer, for example, provides accurate measurement of displacements (see the following section in this article). Multiple-beam interferometers, such as the Fizeau interferometer, also have useful applications. The Fizeau interferometer has its greatest application in microtopography and is used with a microscope to provide high resolution in three dimensions (see the section "Interference Microscopes" in this article). Displacement and Alignment Measurements. Laser interferometers provide a high-accuracy length standard

(better than 0.5 ppm) when measuring linear positioning, straightness in two planes, pitch, and yaw. The most accurate systems consists of a two-frequency laser head, beam directing and splitting optics, measurement optics, receivers, wavelength compensators, and electronics (Fig. 6).

Fig. 6 Components of a laser interferometer. The components include a laser head, beam directing and measuring optics, receivers, electronic couplers, and a computer system.

The most important element in the system is a two-frequency laser head that produces one frequency with a P polarization and another frequency with an S polarization. The beam is projected from the laser head to a remote interferometer, where the beam is split at the polarizing beam splitter into its two separate frequencies. The frequency with the P polarization becomes the measurement beam, and the frequency with the S polarization becomes the reference beam (Fig. 7).

Fig. 7 Schematic of a two-frequency laser interferometer

The measurement beam is directed through the interferometer to reflect off a moving optical element, which may be a target mirror or retroreflector attached to the item being measured. The reference beam is reflected from a stationary optical element, which is usually a retroreflector. The measurement beam then returns to the interferometer, where it is recombined with the reference beam and directed to the receiver.

Whenever the measurement target mirror or retroreflector moves, the accompanying Doppler effect induces a frequency shift in the returning beam. Because of their orthogonal polarization, the frequencies do not interfere to form fringes until the beam reaches the receiver. Consequently, the receiver can monitor the frequency shift associated with the Doppler effect, which is compared to the reference frequency to yield precise measurement of displacement. The principal advantage of the two-frequency system is that the distance information is sensed in terms of frequency. Because a change in frequency is used as the basis for measuring displacement, a change in beam intensity cannot be interpreted as motion. This provides greater measurement stability and far less sensitivity to noise (air turbulence, electrical noise, and light noise). Because motion detection information is embedded in the frequency of the measurement signal, only one photodetector per measurement axis is required; this decreases the sensitivity of optical alignment. Another advantage of the two-frequency interferometer is that the laser head need not be mounted on the machine or instrument being tested. Typical applications include the calibration of length-measuring standards such as glass scales and the characterization of positional, angular, and straightness errors in precision equipment, such as machine tools, coordinate measuring machines, and X-Y stages. The linear resolution of a two-frequency displacement interferometer is 1 nm (0.05 μin.), the angular resolution is 0.03 arc seconds, and the straightness resolution is 40 nm (1.6 μin.). The laser interferometric micrometer uses interferometric technology and a laser beam to perform absolute length

measurements to a resolution of 0.01 μm (0.4 μin.) with an accuracy of ±0.08 mm (±0.003 in.). A contact probe interfaces with an internal interferometer that measures changes in distance (Fig. 8). The part to be measured is placed under the probe, and the interference effects are electronically analyzed and displayed in terms of the distance from the probe to the datum (Fig. 8). Based on user-entered information, the system can automatically compensate for room temperature, humidity, atmospheric pressure, the temperature of the part, and the thermal expansion of the probe. The instrument performs gage comparison, maximum and minimum surface deviation, and total indicator reading measurements. Simple statistical functions include mean and one standard deviation reporting. Actual measurement readings can be compared to user-entered tolerance limits for automatic "go, no-go" testing. Sorting. Parts can be sorted by dimension, prior to

automatic assembly, with an in-process inspection system. A laser beam sorting system can provide accept-or-reject measurements of length, height, diameter, width, thread presence, and count. Each production run of a different part requires a simple setup to accommodate the part to be measured. In operation, a collimated laser beam is optically processed, focused, and directed onto the part. A photodetector converts the light signals to electrical signals for processing. For length inspection, the laser beam is split into three beams. The center beam is Fig. 8 Schematic of a laser interferometric micrometer stationary and acts as a reference beam. Both of the other two beams are adjusted independently by micrometer dials to define the distance between an over- or undersize measurement. The parts can then be gravity fed past the laser quantification system for accept-or-reject measurement. In many applications, parts can be inspected at a rate of 100 to 700 parts per minute. A typical application for a laser-based sorting system is the in-process, accept-or-reject measurement of bolts, nuts, rivets, bearings, tubes, rollers, and stampings. Holography is an important measurement technique in the three-dimensional contouring of large spatial areas.

Holography can determine small deviations (as small as 0.1 μm, or 4 μin.) in surface shape over large areas for all types of surface microstructure. This is accomplished by illuminating both the object and the hologram of its original or desired

shape with the original reference wave (see the article "Optical Holography" in this Volume). If the object deviates from its original or desired shape, interference fringes will appear during illumination with the reference wave. The exactness of the holographic image makes it invaluable for detecting faults in such diverse items as automobile clutch plates, brake drums, gas pipelines, and high-pressure tanks. The holographic image also depicts the vibration pattern of mechanical components and structures such as turbine blades. Laser Inspection Carl Bixby, Zygo Corporation

Velocity Measurements Velocity measurements from the Doppler effect on laser light have become a useful tool for measuring gas, liquid, and solid-surface velocities. Many systems are intended for laboratory applications. However, instruments have also been introduced for industrial process control. One application is in the primary metals industries and consists of measuring length in a unidirectional flow process. The laser Doppler velocity gage is a non-contact instrument that uses laser beams and microprocessors to measure the speed and length of a moving surface. It can measure almost any type of continuously produced material without coming into contact with it, whether it is hot, cold, soft, or delicate. It outputs various types of measurements, such as current speed, average speed, current lengths, and total length. The instrument consists of a sensor, the controller, and a computer. The sensor emits two laser beams that converge on the surface of the product being measured. The light reflected from the product surface exhibits Doppler shifts because of the movement. The frequency of the beam pointing toward the source of the product is shifted up, and the beam pointing toward the destination is shifted down. The processor measures the frequency shift and uses this information to calculate the speed and length of the product. Laser Inspection Carl Bixby, Zygo Corporation

Surface Inspection Surface inspection includes the in-process detection of surface flaws and the measurement of surface defects and roughness. Lasers are used in both of these functions. One technique of in-process flaw detection is illustrated in Fig. 9. A rotating polygonal mirror scans the laser light across the moving sheet, and a stationary photo-multiplier detects the scattered light. If there is a defect on the surface, the intensity of the scattered light increases; if the defect is a crack, the intensity decreases. Similar arrangements can estimate surface roughness by analyzing the envelope of scattered light.

Fig. 9 System for the high-speed scanning of steel sheets for surface defects. In one system, the scanner acquired 104 data points per millisecond on a 3 mm2 (0.005 in.2) sheet.

Interference microscopes are used to measure the microtopography of surfaces. The interference microscope divides

the light from a single point source into two or more waves. In multiple-beam interference microscopes, this is done by placing a partially transmitting and partially reflecting reference mirror near the surface of the specimen (Fig. 10). The multiple beams illustrated in Fig. 10 are superimposed after traveling different lengths. This produces interference patterns, which are magnified by the microscope. The interference fringes having a perfectly flat surface appear as straight, parallel lines of equal width and spacing. Height variations cause the fringes to appear curved or jagged, depending on the unit used. With multiple-beam interferometers, height differences as small as λ/200 can be measured, where λ is the wavelength of the light source. Fig. 10 General principle of a multiple-beam interferometer

Lasers can provide a monochromatic light source, which is required in interference microscopes. One such system is shown in Fig. 11. This system involves the use of photodetectors with displays of isometric plots, contour plots, and up to five qualitative parameters, such as surface roughness, camber, crown, radius of curvature, cylindrical sag, and spherical sag.

Fig. 11 Laser interference microscope with displays. Fizeau and Mirau interferometers are mounted in the turret.

Laser Inspection Carl Bixby, Zygo Corporation

Selected References • •

D. Belforte, Industrial Laser Annual Handbook, Vol 629, Pennwell, 1986 R. Halmshaw, Nondestructive Testing, Edward Arnold, 1987

Coordinate Measuring Machines David H. Genest, Brown & Sharpe Manufacturing Company

Introduction THE COORDINATE MEASURING MACHINE (CMM) fulfills current demands on manufacturing facilities to provide extremely accurate as well as flexible three-dimensional inspection of both in-process and finished parts on the assembly line. Manufacturers are under tremendous financial pressure to increase production and to minimize waste. Part tolerances once quoted in fractional figures are now quoted in thousandths of a millimeter, and manufacturers are under everincreasing pressure to meet ever more demanding specifications on a regular basis. The CMM, which first appeared some 25 years ago, has developed rapidly in recent years as the state-of-the-art measuring tool available to manufacturers. The capabilities, accuracy, and versatility of the CMM, as well as the roles it will play in manufacturing, continue to increase and evolve almost daily. Coordinate measuring machines are an object of intense interest and aggressive development because they offer potentially viable solutions to a number of challenges facing manufacturers:

• • • • •

The need to integrate quality management more closely into the manufacturing process The need to improve productivity and to reduce waste by eliminating the manufacture of out-oftolerance parts faster The realization that the measurement process itself needs to be monitored and verified The objective to eliminate fixtures and fixed gages (and their inherent rebuild costs due to evolving products), thus providing increased gaging flexibility The potential to incorporate existing technology into new and more efficient hybrid systems

Historically, traditional measuring devices and CMMs have been largely used to collect inspection data on which to make the decision to accept or reject parts. Although CMMs continue to play this role, manufacturers are placing new emphasis on using CMMs to capture data from many sources and bring them together centrally where they can be used to control the manufacturing process more effectively and to prevent defective components from being produced. In addition, CMMs are also being used in entirely new applications--for example, reverse engineering and computer-aided design and manufacture (CAD/CAM) applications as well as innovative approaches to manufacturing, such as the flexible manufacturing systems, manufacturing cells, machining centers, and flexible transfer lines. Before purchasing a CMM, the user needs to understand and evaluate modern CMMs and the various roles they play in manufacturing operations today and in the future. This article will: • • • • • •

Define what a CMM is Examine various types of machines available Outline CMM capabilities Examine major CMM components and systems Examine various applications in which CMMs can be employed Provide guidelines for use in specifying and installing CMMs

Terminology germane to CMMs includes: • •

• • • • • •

Ball bar: A gage consisting of two highly spherical tooling balls of the same diameter connected by a rigid bar Gage: A mechanical artifact of high precision used either for checking a part or for checking the accuracy of a machine; a measuring device with a proportional range and some form of indicator, either analog or digital Pitch: The angular motion of a carriage, designed for linear motion, about an axis that is perpendicular to the motion direction and perpendicular to the yaw axis Pixel: The smallest element into which an image is divided, such as the dots on a television screen Plane: A surface of a part that is defined by three points Repeatability: A measure of the ability of an instrument to produce the same indication (or measured value) when sequentially sensing the same quantity under similar conditions Roll: The angular motion of a carriage, designed for linear motion, about the linear motion axis Yaw: The angular motion of a carriage, designed for linear motion, about a specified axis perpendicular to the motion direction. In the case of a carriage with horizontal motion, the specified axis should be vertical unless explicitly specified. For a carriage that does not have horizontal motion, the axis must be explicitly specified

Coordinate Measuring Machines David H. Genest, Brown & Sharpe Manufacturing Company

CMM Operating Principles Technically speaking, a CMM is a multi-axial device with two to six axes of travel or reference axes, each of which provides a measurement output of position or displacement. Coordinate measuring machines are primarily characterized by their flexibility, being able to make many measurements without adding or changing tools. As products evolve, the same CMM can generally be used (depending on size and accuracy limitations) simply by altering software instead of altering equipment mechanics or electronics. Practically speaking, CMMs consist of the machine itself and its probes and moving arms for providing measurement input, a computer for making rapid calculations and comparisons (to blueprint specifications, for example) based on the measurement input, and the computer software that controls the entire system. In addition, the CMM has some means of providing output to the user (printer, plotter CRT, and so on) and/or to other machines in a complete manufacturing system. Coordinate measuring machines linked together in an overall inspection or manufacturing system are referred to as coordinate measuring systems. When CMMs were first introduced in the late 1950s, they were called universal measuring machines. Today, they are sometimes referred to as flexible inspection systems or flexible gages. The most important feature of the CMM is that it can rapidly and accurately measure objects of widely varying size and geometric configuration--for example, a particular part and the tooling for that part. Coordinate measuring machines can also readily measure the many different features of a part, such as holes, slots, studs, and weldnuts, without needing other tools. Therefore, CMMs can replace the numerous hand tools used for measurement as well as the open-plate and surface-plate inspection tools and hard gages traditionally used for part measurement and inspection. Coordinate measuring machines do not always achieve the rates of throughput or levels of accuracy possible with fixed automation-type measuring systems. However, if any changes must be made in a fixed system for any reason--for example, a different measurement of the same part or measurement of a different part--making the change will be costly and time consuming. This is not the case with a CMM. Changes in the measurement or inspection routine of a CMM are made quickly and easily by simply editing the computer program that controls the machine. The greater or more frequent the changes required, the greater the advantage of the CMM over traditional measuring devices. This flexibility, as well as the resulting versatility, is the principal advantage of the CMM. CMM Measurement Techniques A CMM takes measurements of an object within its work envelope by moving a sensing device called a probe along the various axes of travel until the probe contacts the object. The precise position of the contact is recorded and made available as a measurement output of position or displacement (Fig. 1). The CMM is used to make numerous contacts, or hits, with the probe; using all axes of travel, until an adequate data base of the surfaces of the object has been constructed. Various features of an object require different quantities of hits to be accurately recorded. For example, a plane, surface, or circular hole can be recorded with a minimum of three hits.

Fig. 1 Elements of a CMM showing typical digital position readout. The probe is positioned by brackets slid along two arms. Coordinate distances from one point to another are measured in effect by counting electronically the lines in gratings ruled along each arm. Any point in each direction can be set to zero, and the count is made in a plus or minus direction from there.

Once repeated hits or readings have been made and stored, they can be used in a variety of ways through the computer and geometric measurement software of the CMM. The data can be used to create a master program, for example, of the precise specifications for a part; they can also be compared (via the software) to stored part specification data or used to inspect production parts for compliance with specifications. A variety of other sophisticated applications are also possible using the same captured measurement data--for example, the reverse engineering of broken parts or the development of part specifications from handmade models. Coordinate Systems. The CMM registers the various measurements (or hits) it takes of an object by a system of coordinates used to calibrate the axes of travel. There are several coordinate systems in use. The most commonly used system is Cartesian, a three-dimensional, rectangular coordinate system the same as that found on a machine tool. In this system, all axes of travel are square to one another. The system locates a position by assigning it values along the x, y, and z axes of travel.

Another system used is the polar coordinate system. This system locates a point in space by its distance or radius from a fixed origin point and the angle this radius makes with a fixed origin line. It is analogous to the coordinate system used on a radial-arm saw or radial-arm drill. Types of Measurements. As stated earlier, fundamentally, CMMs measure the size and shape of an object and its

contours by gathering raw data through sensors or probes. The data are then combined and organized through computer software programs to form a coherent mathematical representation of the object being measured, after which a variety of inspection reports can be generated. There are three general types of measurements for which CMMs are commonly used, as follows. Geometric measurement deals with the elements commonly encountered every day--points, lines, planes, circles,

cylinders, cones, and spheres. In practical terms, these two-dimensional and three-dimensional elements and their numerous combinations translate into the size and shape of various features of the part being inspected. A CMM can combine the measurements of these various elements into a coherent view of the part and can evaluate the measurements. It can, for example, gage the straightness of a line, the flatness of a plane surface, the degree of parallelism

between two lines or two planes, the concentricity of a circle, the distance separating two features on a part, and so on. Geometric measurement clearly has broad application to many parts and to a variety of industries. Contour measurement deals with artistic, irregular, or computed shapes, such as automobile fenders or aircraft wings. The measurements taken by a CMM can be easily plotted with an exaggerated display of deviation to simplify evaluation. Although contour measurements are generally not as detailed as geometric measurements, presenting as they do only the profile of an object with its vector deviation from the nominal or perfect shape, they too have broad application. Specialized surface measurement deals with particular, recurring shapes, such as those found on gear teeth or

turbine blades. In general, these shapes are highly complex, containing many contours and forms, and the part must be manufactured very precisely. Tight tolerances are absolutely critical. Because manufacturing accuracy is critical, measurement is also highly critical, and a specialty in measuring these forms has evolved. By its nature, specialized surface measurement is applied to far fewer applications than the other two types. CMM Capabilities Coordinate measuring machines have the fundamental ability to collect a variety of different types of very precise measurements and to do so quickly, with high levels of repeatability and great flexibility. In addition, they offer other important capabilities based on computational functions. Automatic Calculation of Measurement Data. The inclusion of a computer in the CMM allows the automatic calculation of such workpiece features as hole size, boss size, the distance between points, incremental distances, feature angles, and intersections. Prior to this stage of CMM development, an inspector had to write down the measurements he obtained and manually compare them to the blueprint. Not only is such a process subject to error, but it is relatively time consuming. While waiting for the results of the inspection, production decisions are delayed and parts (possibly not being produced to specifications) are being manufactured. Compensation for Misaligned Parts. Coordinate measuring machines no longer require that the parts being

measured be manually aligned to the coordinate system of the machine. The operator cannot casually place the part within the CMM work envelope. Once the location of the appropriate reference surface or line has been determined through a series of hits on the datum features of the part, the machine automatically references that position as its zero-zero starting point, creates an x, y, z part coordinate system, and makes all subsequent measurements relative to that point. In addition, the part does not have to be leveled within the work envelope. Just as the CMM will mathematically compensate if the part is rotationally misaligned, it will also compensate for any tilt in the part. Multiple Frames of Reference. The CMM can also create and store multiple frames of reference or coordinate systems; this allows features to be measured on all surfaces of an object quickly and efficiently. The CMM automatically switches to the appropriate new alignment system and zero point (origin) for each plane (face) of the part. The CMM can also provide axis and plane rotation automatically. Probe Calibration. The CMM automatically calibrates for the size and location of the probe tip (contact element) being

used. It also automatically calibrates each tip of a multiple-tip probe. Part Program and Data Storage. The CMM stores the program for a given part so that the program and the machine

are ready to perform whenever this part comes up for inspection. The CMM can also store the results of all prior inspections of a given part or parts so that a complete history of its production can be reconstructed. This same capability also provides the groundwork for all statistical process control applications. Part programs can also be easily edited, rather than completely rewritten, to account for design changes. When a dimension or a feature of a part is changed, only that portion of the program involving the workpiece revision must be edited to conform to part geometry. Interface and Output. As mentioned earlier, CMMs can be linked together in an overall system or can be integrated

with other devices in a complete manufacturing system. The CMM can provide the operator with a series of prompts that tell him what to do next and guide him through the complete measurement routine.

Output is equally flexible. The user can choose the type and format of the report to be generated. Data can be displayed in a wide variety of charts and graphs. Inspection comments can be included in the hard copy report and/or stored in memory for analysis of production runs. CMM Applications Coordinate measuring machines are most frequently used in two major roles: quality control and process control. In the area of quality control, CMMs can generally perform traditional final part inspection more accurately, more rapidly, and with greater repeatability than traditional surface-plate methods. With regard to process control, CMMs are providing new capabilities. Because of the on-line, real-time analytical capability of many CMM software packages, CMMs are increasingly used to monitor and identify evolving trends in production before scrap or out-of-spec parts are fabricated in the first place. Thus, the emphasis has shifted from inspecting parts and subsequently rejecting scrap parts at selected points along the production line to eliminating the manufacture of scrap parts altogether and producing in-tolerance parts 100% of the time. In addition to these uses, there is a trend toward integrating CMMs into systems for more complete and precise control of production. Some shop-hardened CMMs, also known as process control robots, are being increasingly used in sophisticated flexible manufacturing systems in the role of flexible gages. Coordinate measuring machines can also be used as part of a CAD/CAM system. The CMM can measure a part, for example, and feed that information to the CAD/CAM program, which can then create an electronic model of the part. Going in the other direction, the model of the desired part in the CAD/CAM system can be used to create the part program automatically. Coordinate Measuring Machines David H. Genest, Brown & Sharpe Manufacturing Company

Types of CMMs The ANSI/ASME B89 standard formally classifies CMMs into ten different types based on design. All ten types employ three axes of measurement along mutually perpendicular guideways. They differ in the arrangement of the three movable components, the direction in which they move, and which one of them carries the probe, as well as where the workpiece is attached or mounted. However, among the many different designs of CMMs, each with its own strengths, weaknesses, and applications, there are only two fundamental types: vertical and horizontal. They are classified as such by the axis on which the probe is mounted and moves. The ANSI/ASME B89 Performance Standard classifies coordinate measuring machines as:

Vertical

Horizontal

Fixed-table cantilever

Moving ram, horizontal arm

Moving-table cantilever

Moving table, horizontal arm

Moving bridge

Fixed table, horizontal arm

Fixed bridge

L-shaped bridge

Column

Gantry

In addition, the two types of machine can be characterized to some degree by the levels of accuracy they each achieve (although there is a considerable degree of overlap based on the design of an individual machine), by the size of part they can handle, and by application. The prospective buyer/user of a CMM cannot make an intelligent choice of the type of machine that will best meet his needs, let alone the specific make and model of CMM, until he thoroughly evaluates and plans both the specific intended application of the CMM and the overall manufacturing and quality context in which it will operate. Table 1 provides a general comparison of CMM types, applications, and levels of measurement accuracy a CMM user can expect. Table 1 Typical CMM specifications Application

CMM type

Bearing type

Minimum measurement

mm

in.

Laboratory quality(a)

Laboratory grade

Vertical, moving bridge

Air bearings

90°: capillary depression

Fig. 3 Schematic showing the forces involved in capillary rise: the downward force from weight of the liquid column, and the upward force from surface tension along the meniscus perimeter

The height to which the liquid rises is directly proportional to the surface tension of the liquid and to the cosine of the angle of contact, and it is inversely proportional to the density of the liquid and to the radius of the capillary tube. If, as in Fig. 4, the capillary tube is closed, a wetting liquid will still rise in the tube; however, there will be extra pressure resulting from the air and vapor compressed in the closed end of the tube, and the capillary rise will not be as great.

Fig. 4 Schematic showing that the rise and depression of liquids in closed capillary tubes are affected by the compressed air entrapped in the closed end. Compare height of liquid in tube shown in Fig. 3 with that in the tube here.

These examples of surface wetting and capillary rise illustrate the basic physical principles by which a penetrant enters fine surface discontinuities even though the practical circumstances encountered in the use of liquid penetrants are more complex than these examples may suggest. Cracks, for example, are not capillary tubes, but simulate the basic interaction between a liquid and a solid surface, which is responsible for the migration of penetrants into fine surface cracks. The viscosity of the liquid is not a factor in the basic equation of capillary rise. Viscosity is related to the rate at which a liquid will flow under some applied unbalanced stress; in itself, viscosity has a negligible effect on penetrating ability. In general, however, very viscous liquids are unsuitable as penetrants because they do not flow rapidly enough over the surface of the workpiece; consequently, they require excessively long periods of time to migrate into fine flaws. Another necessary property of a penetrant is its capability of dissolving an adequate amount of suitable fluorescent or visible-dye compounds. Finally, the penetrant liquid must be removable with a suitable solvent/remover or emulsifier without precipitating the dye. Just as it is important that a penetrant enter surface flaws, it is also important that the penetrant be retained in the flaws and emerge from the flaw after the superficial coating is removed from the surface and the developer is applied. It is apparently a paradox that the same interaction between a liquid and a surface that causes the liquid to enter a fine opening is also responsible for its emergence from that opening. The resolution of the paradox is simple: Once the surface has been freed of excess penetrant it becomes accessible to the entrapped liquid, which, under the effect of the adhesive forces between liquid and solid, spreads over the newly cleaned surface until an equilibrium distribution is attained (Fig. 5).

Fig. 5 Bead of liquid penetrant formed when, after excess penetrant has been removed from a workpiece surface, the penetrant remaining in a discontinuity emerges to the surface until an equilibrium is established.

Although in some cases the amount of penetrant in a surface bead at equilibrium is sufficient to be detected visually, sensitivity is vastly increased by the use of a developer, of which there are several types available (forms A, B, C, and D). When applied, a developer forms a fine surface film, that is, a spongelike system of very fine, random capillary paths. If the penetrant contacts the powder, the powder then competes with the freshly cleaned surface of the workpiece for the penetrant as it flows out of the defect. If the developer is properly designed, it readily adsorbs the penetrant from the flaw. The penetrant continues to migrate by capillary action, spreading through the developer until an equilibrium is reached. This migrating action is illustrated schematically in Fig. 6. The visibility of the entrapped penetrant within the flaw is greatly increased by the spreading or enlargement of the indication.

Fig. 6 Sectional view showing result of the migrating action of a liquid penetrant through a developer

Liquid Penetrant Inspection Revised by J.S. Borucki, Ardrox Inc., and Gail Jordan, Howmet Corporation

Evolution of the Process The exact origin of liquid penetrant inspection is not known, but it has been assumed that the method evolved from the observation that the rust on a crack in a steel plate in outdoor storage was somewhat heavier than the rust on the adjacent surfaces as a result of water seeping into the crack and forcing out the oxide it had helped to produce. The obvious conclusion was that a liquid purposely introduced into surface cracks and then brought out again would reveal the locations of those cracks. The only material that fulfilled the known criteria of low viscosity, good wettability, and ready availability was kerosene. It was found, however, that although wider cracks showed up easily, finer ones were sometimes missed because of the difficulty of detecting, by purely visual means, the small amounts of kerosene exuding from them. The solution was to provide a contrasting surface that would reveal smaller seepages. The properties and availability of whitewash made it the logical choice. This method, known as the kerosene-and-whiting test, was the standard for many years. The sensitivity of the kerosene-and-whiting test could be increased by hitting the object being tested with a hammer during testing. The resulting vibration brought more of the kerosene out of the cracks and onto the whitewash. Although this test was not as sensitive as those derived from it, it was quick, inexpensive, and reasonably accurate. Thus, it provided a vast improvement over ordinary visual examination. The first step leading to the methods now available was the development of the fluorescent penetrant process by R.C. Switzer. This liquid, used jointly with a powder developer, brought penetrant inspection from a relatively crude procedure to a more scientific operation. With fluorescent penetrant, minute flaws could be readily detected when exposed to ultraviolet light (commonly called black light). This development represented a major breakthrough in the detection of surface flaws. Switzer's work also included the development of the visible-color contrast method, which allowed for inspection under white light conditions. Although not as sensitive as fluorescent penetrant inspection, it is widely used in industry for noncritical inspection. Through the developments described above, liquid penetrant inspection has become a major nondestructive inspection method. Liquid Penetrant Inspection Revised by J.S. Borucki, Ardrox Inc., and Gail Jordan, Howmet Corporation

Penetrant Methods Because of the vast differences among applications for penetrant inspection, it has been necessary to refine and develop the two types of penetrants (type I, fluorescent, and type II, visible) into four basic methods to accommodate the wide variations in the following principal factors: • • • • •

Surface condition of the workpiece being inspected Characteristics of the flaws to be detected Time and place of inspection Size of the workpiece Sensitivity required

The four methods are broadly classified as:

• • • •

Water washable, method A Postemulsifiable lipophilic, method B Solvent removable, method C Postemulsifiable hydrophilic, method D

These four methods are described below. Water-washable penetrant (method A) is designed so that the penetrant is directly water washable from the

surface of the workpiece; it does not require a separate emulsification step, as does the postemulsifiable penetrant methods. It can be used to process workpieces quickly and efficiently. It is important, however, that the washing operation be carefully controlled because water-washable penetrants are susceptible to overwashing. The essential operations involved in this method are illustrated schematically in Fig. 7.

Fig. 7 Five essential operations for liquid penetrant inspection using the water-washable system

Postemulsifiable penetrants (methods B and D) are designed to ensure the detection of minute flaws in some

materials. These penetrants are not directly water washable. Because of this characteristic, the danger of overwashing the penetrant out of the flaws is reduced. The difference between the water-washable and postemulsifiable method lies in the use of an emulsifier prior to final rinsing. The emulsifier makes the residual surface penetrant soluble in water so that the excess surface penetrant can be removed by the water rinse. Therefore, the emulsification time must be carefully controlled so that the surface penetrant becomes water soluble but the penetrant in the flaws does not. The operations involved in the postemulsifiable method are illustrated schematically in Fig. 8 for the lipophilic system and in Fig. 9 for the hydrophilic system. Despite the additional processing steps involved with the postemulsifiable methods B and D, these methods are the most reliable for detecting minute flaws.

Fig. 8 Operations (in addition to precleaning) for the postemulsifiable, method B, lipophilic liquid penetrant system

Fig. 9 Operations (in addition to precleaning) for the postemulsifiable, method D, hydrophilic liquid penetrant system

Solvent-removable penetrant (method C) is available for use when it is necessary to inspect only a localized area of a workpiece or to inspect a workpiece at the site rather than on a production inspection basis. Normally, the same type of solvent is used for precleaning and for removing excess penetrant. This penetrant process is convenient and broadens the range of applications of penetrant inspections. However, because the solvent-removable method is labor intensive, it is not practical for many production applications. When properly conducted and when used in the appropriate applications, the solvent-removable method can be one of the most sensitive penetrant methods available. The operations for this process are illustrated schematically in Fig. 10.

Fig. 10 Operations (in addition to precleaning) for the solvent-removable liquid penetrant system

Whichever penetrant method is chosen, the degree and speed of excess penetrant removal depend on such processing conditions as spray nozzle characteristics, water pressure and temperature, duration of the rinse cycle, surface condition of the workpiece, and inherent removal characteristics of the penetrant employed. Liquid Penetrant Inspection Revised by J.S. Borucki, Ardrox Inc., and Gail Jordan, Howmet Corporation

Description of the Process Regardless of the type of penetrant used, that is, fluorescent (type I) or visible (type II), penetrant inspection requires at least five essential steps, as follows. Surface Preparation. All surfaces to be inspected, whether localized or the entire workpiece, must be thoroughly

cleaned and completely dried before being subjected to penetrant inspection. Flaws exposed to the surface must be free from oil, water, or other contaminants if they are to be detected. Penetration. After the workpiece has been cleaned, penetrant is applied in a suitable manner so as to form a film of the

penetrant over the surface. This film should remain on the surface long enough to allow maximum penetration of the penetrant into any surface openings that are present. Removal of Excess Penetrant. Excess penetrant must be removed from the surface. The removal method is

determined by the type of penetrant used. Some penetrants can be simply washed away with water; others require the use of emulsifiers (lipophilic or hydrophilic) or solvent/remover. Uniform removal of excess surface penetrant is necessary for effective inspection, but overremoval must be avoided. Development. Depending on the form of developing agent to be used, the workpiece is dried either before or directly

after application of the developer. The developer forms a film over the surface. It acts as a blotter to assist the natural seepage of the penetrant out of surface openings and to spread it at the edges so as to enhance the penetrant indication. Inspection. After it is sufficiently developed, the surface is visually examined for indications of penetrant bleedback from surface openings. This examination must be performed in a suitable inspection environment. Visible penetrant inspection is performed in good white light. When fluorescent penetrant is used, inspection is performed in a suitably darkened area using black (ultraviolet) light, which causes the penetrant to fluoresce brilliantly. Liquid Penetrant Inspection Revised by J.S. Borucki, Ardrox Inc., and Gail Jordan, Howmet Corporation

Materials Used in Penetrant Inspection In addition to the penetrants themselves, liquids such as emulsifiers, solvent/cleaners and removers, and developers are required for conducting liquid penetrant inspection. Penetrants There are two basic types of penetrants:

• •

Fluorescent, type I Visible, type II

Each type is available for any one of the four methods (water washable, postemulsifiable lipophilic, and postemulsifiable hydrophilic, and solvent removable) mentioned in the section "Penetrant Methods" in this article. Type I fluorescent penetrant utilizes penetrants that are usually green in color and fluoresce brilliantly under

ultraviolet light. The sensitivity of a fluorescent penetrant depends on its ability to form indications that appear as small sources of light in an otherwise dark area. Type I penetrants are available in different sensitivity levels classified as follows: • • • • •

Level : Ultralow Level 1: Low Level 2: Medium Level 3: High Level 4: Ultrahigh

Type II visible penetrant employs a penetrant that is usually red in color and produces vivid red indications in

contrast to the light background of the applied developer under visible light. The visible penetrant indications must be viewed under adequate white light. The sensitivity of visible penetrants is regarded as Level 1 and adequate for many applications. Penetrant selection and use depend on the criticality of the inspection, the condition of the workpiece surface, the

type of processing, and the desired sensitivity (see the section "Selection of Penetrant Method" in this article). Method A, water-washable penetrants are designed for the removal of excess surface penetrant by water rinsing

directly after a suitable penetration (dwell) time. The emulsifier is incorporated into the water-washable penetrant. When this type of penetrant is used, it is extremely important that the removal of excess surface penetrant be properly controlled to prevent overwashing, which can cause the penetrant to be washed out of the flaws. Methods B and D, lipophilic and hydrophilic postemulsifiable penetrants are insoluble in water and therefore not removable by water rinsing alone. They are designed to be selectively removed from the surface of the workpiece by the use of a separate emulsifier. The emulsifier, properly applied and left for a suitable emulsification time, combines with the excess surface penetrant to form a water-washable surface mixture that can be rinsed from the surface of the workpiece. The penetrant that remains within the flaw is not subject to overwashing. However, proper emulsification time must be established experimentally and maintained to ensure that overemulsification, which results in the loss of flaws, does not occur. Method C, solvent-removable penetrants are removed by wiping with clean, lint-free material until most traces of

the penetrant have been removed. The remaining traces are removed by wiping with clean, lint-free material lightly moistened with solvent. This type of penetrant is primarily used where portability is required and for the inspection of localized areas. To minimize the possibility of removing the penetrant from discontinuities, the use of excessive amounts of solvent must be avoided. Physical and Chemical Characteristics. Both fluorescent and visible penetrants, whether water washable,

postemulsifiable, or solvent removable, must have certain chemical and physical characteristics if they are to perform their intended functions. The principal requirements of penetrants are as follows: • • • •

Chemical stability and uniform physical consistency A flash point not lower than 95 °C (200 °F); penetrants that have lower flash points constitute a potential fire hazard A high degree of wettability Low viscosity to permit better coverage and minimum dragout

• • • • • • • • •

Ability to penetrate discontinuities quickly and completely Sufficient brightness and permanence of color Chemical inertness with materials being inspected and with containers Low toxicity to protect personnel Slow drying characteristics Ease of removal Inoffensive odor Low cost Resistance to ultraviolet light and heat fade

Emulsifiers Emulsifiers are liquids used to render excess penetrant on the surface of a workpiece water washable. There are two methods used in the postemulsifiable method: method B, lipophilic, and method D, hydrophilic. Both can act over a range of durations from a few seconds to several minutes, depending on the viscosity, concentration, method of application, and chemical composition of the emulsifier, as well as on the roughness of the workpiece surface. The length of time an emulsifier should remain in contact with the penetrant depends on the type of emulsifier employed and the roughness of the workpiece surface. Method B, lipophilic emulsifiers are oil based, are used as supplied, and function by diffusion (Fig. 11). The

emulsifier diffuses into the penetrant film and renders it spontaneously emulsifiable in water. The rate at which it diffuses into the penetrant establishes the emulsification time. The emulsifier is fast acting, thus making the emulsification operation very critical. The emulsifier continues to act as long as it is in contact with the workpiece; therefore, the rinse operation should take place quickly to avoid overemulsification.

Fig. 11 Elements in the functioning of lipophilic emulsifiers

Method D, hydrophilic emulsifiers are water based and are usually supplied as concentrates that are diluted in water

to concentrations of 5 to 30% for dip applications and 0.05 to 5% for spray applications. Hydrophilic emulsifiers function by displacing excess penetrant from the surface of the part by detergent action (Fig. 12). The force of the water spray or

the air agitation of dip tanks provides a scrubbing action. Hydrophilic emulsifier is slower acting than the lipophilic emulsifier; therefore, it is easier to control the cleaning action. In addition to the emulsifier application, method D also requires a prerinse. Utilizing a coarse water spray, the prerinse helps remove the excess penetrant to minimize contamination of the emulsifier. Of greater significance, only a very thin and uniform layer of penetrant will remain on the surface, thus allowing easy removal of the surface layer with minimum opportunity of removing penetrant from the flaws. This step is required because the penetrant is not miscible with the hydrophilic emulsifier.

Fig. 12 Elements in the functioning of hydrophilic emulsifiers

The penetrant manufacturer should recommend nominal emulsification times for the specific type of emulsifier in use. Actual emulsification times should be determined experimentally for the particular application. The manufacturer should also recommend the concentrations for hydrophilic emulsifiers. Solvent Cleaner/Removers Solvent cleaner/removers differ from emulsifiers in that they remove excess surface penetrant through direct solvent action. There are two basic types of solvent removers: flammable and nonflammable. Flammable cleaners are essentially free of halogens but are potential fire hazards. Nonflammable cleaners are widely used. However, they do contain halogenated solvents, which may render them unsuitable for some applications. Excess surface penetrant is removed by wiping, using lint-free cloths slightly moistened with solvent cleaner/remover. It is not recommended that excess surface penetrant be removed by flooding the surface with solvent cleaner/remover, because the solvent will dissolve the penetrant within the defect and indications will not be produced. Developers The purpose of a developer is to increase the brightness intensity of fluorescent indications and the visible contrast of visible-penetrant indications. The developer also provides a blotting action, which serves to draw penetrant from within the flaw to the surface, spreading the penetrant and enlarging the appearance of the flaw.

The developer is a critical part of the inspection process; borderline indications that might otherwise be missed can be made visible by the developer. In all applications of liquid penetrant inspection, use of a developer is desirable because it decreases inspection time by hastening the appearance of indications. Required Properties. To carry out its functions to the fullest possible extent, a developer must have the following

properties or characteristics (rarely are all these characteristics present to optimum degrees in any given material or formulation, but all must be considered in selecting a developer): • • • • • • • • • •

The developer must be adsorptive to maximize blotting It must have fine grain size and a particle shape that will disperse and expose the penetrant at aflaw to produce strong and sharply defined indications of flaws It must be capable of providing a contrast background for indications when color-contrast penetrants are used It must be easy to apply It must form a thin, uniform coating over a surface It must be easily wetted by the penetrant at the flaw (the liquid must be allowed to spread over the particle surfaces) It must be nonfluorescent if used with fluorescent penetrants It must be easy to remove after inspection It must not contain ingredients harmful to parts being inspected or to equipment used in the inspection operation It must not contain ingredients harmful or toxic to the operator

Developer Forms. There are four forms of developers in common use:

• • • •

Form A, dry powder Form B, water soluble Form C, water suspendible Form D, nonaqueous solvent suspendible

The characteristics of each form are discussed below. Dry powder developers (form A) are widely used with fluorescent penetrants, but should not be used with visible-

dye penetrants because they do not produce a satisfactory contrast coating on the surface of the workpiece. Ideally, dry powder developers should be light and fluffy to allow for ease of application and should cling to dry surfaces in a fine film. The adherence of the powder should not be excessive, as the amount of black light available to energize fluorescent indications will be reduced. For purposes of storage and handling as well as applications, powders should not be hygroscopic, and they should remain dry. If they pick up moisture when stored in areas of high humidity, they will lose their ability to flow and dust easily, and they may agglomerate, pack, or lump up in containers or in developer chambers. For reasons of safety, dry powder developers should be handled with care. Like any other dust particle, it can dry the skin and irritate the lining of the air passages, causing irritation. If an operator will be working continuously at a developer station, rubber gloves and respirators may be desirable. Modern equipment often includes an exhaust system on the developer spray booth or on the developer dust chamber that prevents dust from escaping. Powder recovery filters are included in most such installations. Water-soluble developers (form B) can be used for both fluorescent (type I) or visible (type II) postemulsifiable or

solvent-removable penetrants. Water-soluble developers are not recommended for use with water-washable penetrants, because of the potential to wash the penetrant from within the flaw if the developer is not very carefully controlled. Water-soluble developers are supplied as a dry powder concentrate, which is then dispersed in water in recommended proportions, usually from 0.12 to 0.24 kg/L (1 to 2 lb/gal.). The bath concentration is monitored for specific gravity with

the appropriate hydrometer. Necessary constituents of the developers include corrosion inhibitors and biocides. The advantages of this form of developer are as follows: • • •

The prepared bath is completely soluble and therefore does not require any agitation The developer is applied prior to drying, thus decreasing the development time The dried developer film on the workpiece is completely water soluble and is thus easily and completely removed following inspection by simple water rinsing

Water-suspendible developers (form C) can be used with either fluorescent (type I) or visible (type II) penetrants.

With fluorescent penetrant, the dried coating of developer must not fluoresce, nor may it absorb or filter out the black light used for inspection. Water-suspendible developers are supplied as a dry powder concentrate, which is then dispersed in water in recommended proportions, usually from 0.04 to 0.12 kg/L ( to 1 lb/gal.). The amount of powder in suspension must be carefully maintained. Too much or too little developer on the surface of a workpiece can seriously affect sensitivity. Specific gravity checks should be conducted routinely, using a hydrometer to check the bath concentration. Water-soluble developers contain dispersing agents to help retard settling and caking as well as inhibitors to prevent or retard corrosion of workpieces and equipment, and biocides to extend the working life of the aqueous solutions. In addition, wetting agents are present to ensure even coverage of surfaces and ease of removal after inspection. Water-suspendible developer is applied before drying; therefore, developing time can be decreased because the heat from the drier helps to bring penetrant back out of surface openings. In addition, with the developer film already in place, the developing action begins at once. Workpieces are ready for inspection in a shorter period of time. Nonaqueous solvent-suspendible developers (form D) are commonly used for both the fluorescent and the

visible penetrant process. This form of developer produces a white coating on the surface of the part. This coating yields the maximum white color contrast with the red visible penetrant indication and extremely brilliant fluorescent indication. Nonaqueous solvent-suspendible developers are supplied in the ready-to-use condition and contain particles of developer suspended in a mixture of volatile solvents. The solvents are carefully selected for their compatibility with the penetrants. Nonaqueous solvent-suspendible developers also contain surfactants in a dispersant whose functions are to coat the particles and reduce their tendency to clump or agglomerate. Nonaqueous solvent-suspendible developers are the most sensitive form of developer used with type I fluorescent penetrants because the solvent action contributes to the absorption and adsorption mechanisms. In many cases where tight, small flaws occur, the dry powder (form A), water-soluble (form B), and water-suspendible (form C) developers do not contact the entrapped penetrant. This results in the failure of the developer to create the necessary capillary action and surface tension that serve to pull the penetrant from the flaw. The nonaqueous solvent-suspendible developer enters the flaw and dissolves into the penetrant. This action increases the volume and reduces the viscosity of the penetrant. The manufacturer must carefully select and compound the solvent mixture. There are two types of solvent-base developers: nonflammable (chlorinated solvents) and flammable (nonchlorinated solvents). Both types are widely used. Selection is based on the nature of the application and the type of alloy being inspected. Liquid Penetrant Inspection Revised by J.S. Borucki, Ardrox Inc., and Gail Jordan, Howmet Corporation

Selection of Penetrant Method The size, shape, and weight of workpieces, as well as the number of similar workpieces to be inspected, often influence the selection of a penetrant method.

Sensitivity and Cost. The desired degree of sensitivity and cost are usually the most important factors in selecting the

proper penetrant method for a given application. The methods capable of the greatest sensitivity are also the most costly. Many inspection operations require the ultimate in sensitivity, but there are a significant number in which extreme sensitivity is not required and may even produce misleading results. On a practical basis, the fluorescent penetrant methods are employed in a wider variety of production inspection operations than the visible penetrant methods, which are utilized primarily for localized inspections. As stated earlier, penetrants are classified on the basis of penetrant type: • • • • • •

Type I: Fluorescent Type II: Visible Method A: Water washable Method B: Postemulsifiable-lipophilic Method C: Solvent removable Method D: Postemulsifiable-hydrophilic

Penetrants are also classified in terms of sensitivity levels: • • • • •

Level : Ultralow Level 1: Low Level 2: Medium Level 3: High Level 4: Ultrahigh

Advantages and Limitations of Penetrant Methods. Each penetrant method, whether postemulsifiable (either lipophilic or hydrophilic), solvent removable, or water washable, using fluorescent or visible-dye penetrants, has inherent advantages and limitations. The postemulsifiable fluorescent penetrant method is the most reliable and sensitive penetrant method. This

procedure will locate wide, shallow flaws as well as tight cracks and is ideal for high-production work. On the other hand, emulsification requires an additional operation, which increases cost. Also, this method requires a water supply and facilities for inspection under black light. The postemulsifiable, lipophilic fluorescent penetrant method is less sensitive and less reliable than the hydrophilic method. Its use is therefore declining. The solvent-removable fluorescent penetrant method employs a procedure similar to that used for the

postemulsifiable fluorescent method, except that excess penetrant is removed with a solvent/remover. This method is especially recommended for spot inspection or where water cannot be conveniently used. It is more sensitive than the water-washable system, but the extreme caution and additional time required for solvent removal often preclude its use. The water-washable fluorescent penetrant method is the fastest of the fluorescent procedures. It is also highly

sensitive, reliable, and reasonably economical. It can be used for both small and large workpieces and is effective on most part surfaces. However, it will not reliably reveal open, shallow flaws if overwashed and in some cases, depending on the sensitivity level of the penetrant, will not locate the very tightest cracks. There is also the danger of overwashing by applying water for an excessive period of time or with a pressure sufficient to remove the penetrant from the flaws. The postemulsifiable visible penetrant method is used whenever sensitivity required is greater than that

provided by the water-washable visible penetrant method. However, the additional step of applying emulsifier makes this system more costly than the water-washable visible penetrant dye method that requires water, but otherwise no location limitations are imposed. The solvent-removable visible penetrant method has a distinct advantage in that all the necessary ingredients

are portable; accordingly, it can be used in a practically limitless number of locations, both in the shop and in the field. Because of the problems involved in penetrant removal, however, the method is generally confined to spot inspection or to inspection under circumstances that prohibit the use of other methods because of workpiece size or location.

The water-washable visible penetrant system is the fastest and simplest of all penetrant techniques. It is,

however, the least sensitive because the penetrant is likely to be removed from wide, shallow flaws. Therefore, it is most useful in those applications where shallow and relatively wide flaws are not significant. This method is also the least sensitive for locating tight cracks. It requires a water source, but can be performed in almost any location because neither a darkened area nor electricity is required. Liquid Penetrant Inspection Revised by J.S. Borucki, Ardrox Inc., and Gail Jordan, Howmet Corporation

Equipment Requirements The equipment used in the penetrant inspection process varies from spray or aerosol cans to complex, automated, computer-driven processing systems. Some of the more generally used types of equipment are described in the following sections. Portable Equipment For occasional inspections, especially in the field, where equipment portability is necessary, minimal kits for either visible or fluorescent penetrant inspection are commercially available. (Generally, portable penetrant applications are limited to localized areas or spot inspections rather than entire part surfaces.) Such a kit for visible penetrant inspection work includes a precleaner, a penetrant, and a penetrant remover and developer, all in pressurized spray cans. Penetrant removal requires wiping with lint-free cloths or paper towels. A similar kit is available for fluorescent work; a precleaner, a penetrant, penetrant remover and developer are likewise supplied in pressurized cans. Cleaning is accomplished by wiping with lint-free cloths or paper towels. This kit includes a small, portable black light for conducting the inspection. Stationary Inspection Equipment The type of equipment most frequently used in fixed installations consists of a series of modular subunits. Each subunit performs a special task. The number of subunits in a processing line varies with the type of penetrant method used. The subunits are: • • • • • • •

Drain and/or dwell stations Penetrant and emulsifier stations Pre- and post-wash stations Drying station Developer station Inspection station Cleaning stations

The drain or dwell stations are actually roller-top benches that hold the parts during the processing cycle. The usual arrangement is to position a drain or dwell station following each of the dip tanks, the wash station, and the drying oven. The subunits are described in more detail below. Penetrant Station. The principal requirement of a penetrant station is that it provide a means for coating workpieces

with penetrant--either all over, for small workpieces, or over small areas of large workpieces when only local inspection is required. In addition, means should be provided for draining excess penetrant back into the penetrant reservoir, unless the expendable technique is being used. Draining racks usually serve the additional purpose of providing a storage place for parts during the time required for penetration (dwell time).

Small workpieces are easily coated by dipping them into a reservoir of penetrant. This may be done individually or in batches in a wire basket. The penetrant container should be equipped with an easily removable cover to reduce evaporation when not in use. A drain cock should also be provided to facilitate draining of the tank for cleaning. Containers are usually made of steel, but stainless steel containers should be used with water-base penetrants. For large workpieces, penetrant is often applied by spraying or flowing. This is done mainly for convenience but also for economy, because the volume of penetrant needed to immerse a large object may be so great as to increase unnecessarily the original cost of installation. A small reservoir of penetrant equipped with a pump, a hose, and a spray or flow nozzle is usually almost as fast a means of coating large objects as the dipping operation. For this type of operation, the penetrant station consists of a suitably ventilated booth with a rotatable grill platform on which the workpiece is set. A drain under the platform returns penetrant runoff to the sump, from which it is pumped back to the spray nozzle. The booth enclosure prevents the overspraying of penetrant on areas outside the penetrant station. In some applications, it has been found that only a small amount of penetrant is recoverable and reusable, and this has led to the adoption of the expendable technique for some very large workpieces. In this technique, penetrant is sprayed over the workpiece in a penetrant station similar to the one mentioned previously. The penetrant is stored in a separate pressure tank fitted with a hose and a spray nozzle. The spray booth is not equipped with a sump to recover excess penetrant. Instead, the booth is fitted with water spray nozzles and a drain so that it can serve the multiple purpose of draining and washing. A decision to use the expendable technique and related equipment should be based on a careful analysis and consideration of cost, time, rate of production, and handling problems. Emulsifier Station. The emulsifier liquid is contained in a tank of sufficient size and depth to permit immersion of the

workpieces, either individually or in batches. Covers are sometimes provided to reduce evaporation, and drain valves are supplied for cleanout when the bath has become contaminated. Suitable drain racks are also a part of this station and are used to permit excess emulsifier to drain back into the tank. If large workpieces must be coated with emulsifier, methods must be devised to achieve the fastest possible coverage. Multiple spraying or copious flowing of emulsifier from troughs or perforated pipes can be used on some types of automatic equipment. For the local coating of large workpieces, spraying is often satisfactory, using the expendable technique described for the application of penetrant. Pre- and Postrinse Stations. The water rinsing (washing) of small workpieces is frequently done by hand, either

individually or in batches in wire baskets. The workpieces are held in the wash tank and cleaned with a hand-held spray using water at tap pressure and temperature. The wash trough or sink should be large enough and deep enough so that workpieces can be easily turned to clean all surfaces. Splash shields should separate the rinse station from preceding (penetrant or emulsifier) and succeeding (wet developer) stations. Rinse stations are always equipped with at least one ultraviolet light so that the progress of removal of fluorescent penetrant can be easily followed. The automatic rinsing of small workpieces is satisfactorily accomplished by means of a rotating table. The basket is placed on the table, and water-spray heads are properly located so as to rinse all surfaces of the workpieces thoroughly. Specially built automatic washers for rinsing workpieces that are large and of irregular contour are often installed. Spray nozzles must be located to suit the individual application. The removal of excess penetrant by simply submerging the workpiece in water is generally not recommended. However, in some cases, simple submersion in an air-agitated water bath is satisfactory. The rinse station is subject to corrosion. All steel should be protected by rustproofing and painting. Most satisfactory, but more costly, is the use of stainless steel equipment. Drying Station. The recirculating hot-air drier is one of the most important equipment components. The drier must be

large enough to easily handle the type and number of workpieces being inspected. Heat input, air flow, and rate of movement of workpieces through the drier, as well as temperature control, are all factors that must be balanced. The drier may be of the cabinet type, or it may be designed so that the workpieces pass through on a conveyor. If conveyor operation is used, the speed must be considered with the required drying cycle.

Electric-resistance elements are frequently used as sources of heat, but gas, hot water, and steam are also used. Heat input is controlled by suitably located thermostats and is determined by workpiece size, composition, and rate of movement. Integrated equipment invariably includes the recirculating hot-air drier mentioned previously. Makeshift driers are sometimes used--often because nothing better is available. Electric or gas hot-air blowers of commercial design have been used, but because no control of temperature is possible, these are very unsatisfactory and are ordinarily used only on an emergency basis. Infrared lamps are not suitable for drying washed workpieces, because the radiant heat cannot be readily controlled. Equipment designed to handle workpieces of special size and shape requires a specially designed drier. Each drier is a separate engineering problem involving a special combination of workpiece composition, mass, surface area, speed of movement, and other considerations unique to the circumstances. Developer Station. The type and location of the developer station depend on whether dry or wet developer is to be used. For dry developer, the developer station is downstream from the drier, but for wet developer it immediately precedes the drier, following the rinse station. The dry-developer station usually consists of a simple bin containing the powder. Dried workpieces are dipped into

the powder, and the excess powder is shaken off. Larger workpieces may not be so easily immersed in the powder, so a scoop is usually provided for throwing powder over the surfaces, after which the excess is shaken off. The developer bin should be equipped with an easily removable cover to protect the developer from dust and dirt when not in use. Dust control systems are sometimes needed when dry developer is used. Control is accomplished by a suction opening across the back of the bin at the top, which draws off any developer dust that rises out of the bin. The dust-laden air is passed through filter bags, from which the developer dust can be reclaimed for further use (Fig. 13).

Fig. 13 Dry-developer bin equipped with dust control and reclaimer system

Developer powder can also be applied with air pressure. This system requires no bin, but it does require a booth or a cabinet and also makes dust collection mandatory. Equipment for the automatic application of dry developer consists of a cabinet through which the dried workpieces are passed on a conveyor. The air in the cabinet is laden with dust that is kept agitated by means of a blower. As workpieces pass through, all surfaces are brought into contact with developer powder carried by the air. Air must be exhausted from the cabinet and either recirculated or cleaned by being passed through a dust-collecting filter. Wet developer, when used, is contained in a tank similar to that used for penetrant or emulsifier. The tank should be

deep enough to permit workpieces to be submerged in the developer. There should also be a rack or conveyor on which parts can rest after dipping. This will permit excess developer to run back into the tank.

Suspendible developer baths settle out when not in use; therefore, a paddle for stirring should be provided. Continuous agitation is essential because the settling rate is rapid. Pumps are sometimes incorporated into the developer station for flowing the developer over large workpieces through a hose and nozzle and for keeping the developer agitated. In automatic units, special methods of applying developer are required. Flow-on methods are frequently used. This technique requires a nozzle arrangement that permits the workpieces to be covered thoroughly and quickly. Inspection Station. Essentially, the inspection station is simply a worktable on which workpieces can be handled under proper lighting. For fluorescent methods, the table is usually surrounded by a curtain or hood to exclude most of the white light from the area. For visible-dry penetrants, a hood is not necessary.

Generally, black (ultraviolet) lights (100 W or greater) are mounted on brackets from which they can be lifted and moved about by hand. Because of the heat given off by black lights, good air circulation is essential in black light booths. For automatic inspection, workpieces are moved through booths equipped with split curtains, either by hand, monorail, or by conveyor. In some large inspection installations, fully enclosed rooms have been built for black light inspection. Access to the room is provided by a light lock. Inspection rooms must be laid out efficiently to prevent rejected workpieces from reentering the production line. Figures 14 and 15 illustrate typical penetrant inspection components and their layout in an inspection station installation.

Fig. 14 Typical seven-station package equipment unit for inspecting workpieces by the water-washable fluorescent penetrant system

Fig. 15 Arrangement of equipment used in one foundry for the liquid penetrant inspection of a large variety of castings to rigid specifications. Many of the castings require handling by crane or roller conveyor.

Automated Inspection Equipment For many years, the penetrant inspection of production parts has been a manual operation of moving parts from station to station through the penetrant line. Properly trained and motivated operators will do an excellent job of processing and inspecting parts as well as controlling the process. There are, however, many situations in which manual processing simply cannot keep up with the production rates required or control the process properly. The use of automated inspection systems, therefore, has become a significant factor in performing penetrant inspections of high-volume production parts. Modern automated penetrant inspection systems provide precise and repeatable process control, improved inspection reliability, increased productivity, and lower inspection costs. Automated penetrant inspection systems incorporate programmable logic control (PLC) units, which are programmed to control the handling of parts through the system, to control the processing cycle precisely, and to monitor the functions at each processing station. Figures 16, 17, and 18 show typical automated penetrant systems currently in use.

Fig. 16 Automated inspection equipment setup (upper left) with close-up of operator checking PLC panel (upper right), which includes a screen display (lower right). The setup incorporates material handling devices

such as the robot shown (lower left) to transfer workpieces from station to station and to apply penetrants and other solutions needed to inspect components.

Fig. 17 Typical automated fluorescent penetrant inspection installation

Fig. 18 Automated inspection installation for the fluorescent penetrant inspection of large workpieces, such as castings. The installation incorporates a complex roller conveyor system.

Liquid Penetrant Inspection Revised by J.S. Borucki, Ardrox Inc., and Gail Jordan, Howmet Corporation

Precleaning Regardless of the penetrant chosen, adequate precleaning of workpieces prior to penetrant inspection is absolutely necessary for accurate results. Without adequate removal of surface contamination, relevant indications may be missed because:

• • •

The penetrant does not enter the flaw The penetrant loses its ability to identify the flaw because it reacts with something already in it The surface immediately surrounding the flaw retains enough penetrant to mask the true appearance of the flaw

Also, nonrelevant (false) indications may be caused by residual materials holding penetrants. Cleaning methods are generally classified as chemical, mechanical, solvent, or any combination of these. Chemical cleaning methods include alkaline or acid cleaning, pickling or chemical etching, and molten salt bath

cleaning. Mechanical cleaning methods include tumbling, wet blasting, dry abrasive blasting, wire brushing, and high-

pressure water or steam cleaning. Mechanical cleaning methods should be used with care because they often mask flaws by smearing adjacent metal over them or by filling them with abrasive material. This is more likely to happen with soft metals than with hard metals. Solvent cleaning methods include vapor degreasing, solvent spraying, solvent wiping, and ultrasonic immersion

using solvents. Probably the most common method is vapor degreasing. However, ultrasonic immersion is by far the most effective means of ensuring clean parts, but it can be a very expensive capital equipment investment. Cleaning methods and their common uses are listed in Table 1. A major factor in the selection of a cleaning method is the type of contaminant to be removed and the type of alloy being cleaned. This is usually quite evident, but costly errors can be avoided by accurate identification of the contaminant. Before the decision is made to use a specific method, it is good practice to test the method on known flaws to ensure that it will not mask the flaws. Table 1 Applications of various methods of precleaning for liquid penetrant inspection Method

Use

Mechanical methods

Abrasive tumbling

Removing light scale, burrs, welding flux, braze stopoff, rust, casting mold, and core material; should not be used on soft metals such as aluminum, magnesium, or titanium

Dry abrasive grit blasting

Removing light or heavy scale, flux, stopoff, rust, casting mold and core material, sprayed coatings, carbon deposits--in general, any friable deposit. Can be fixed or portable

Wet abrasive grit blasting

Same as dry except, where deposits are light, better surface and better control of dimensions are required

Wire brushing

Removing light deposits of scale, flux, and stopoff

High-pressure water and steam

Ordinarily used with an alkaline cleaner or detergent; removing typical machine shop soils such as cutting oils, polishing compounds, grease, chips, and deposits from electrical discharge machining; used when surface finish must be maintained; inexpensive

Ultrasonic cleaning

Ordinarily used with detergent and water or with a solvent; removing adherent shop soil from large quantities of small parts

Chemical methods

Alkaline cleaning

Removing braze stopoff, rust, scale, oils, greases, polishing material, and carbon deposits; ordinarily used on large articles where hand methods are too laborious; also used on aluminum for gross metal removal

Acid cleaning

Strong solutions for removing heavy scale; mild solutions for light scale; weak (etching) solutions for removing lightly smeared metal

Molten salt bath cleaning

Conditioning and removing heavy scale

Solvent methods

Vapor degreasing

Removing typical shop soil, oil, and grease; usually employs chlorinated solvents; not suitable for titanium

Solvent wiping

Same as for vapor degreasing except a hand operation; may employ nonchlorinated solvents; used for localized low-volume cleaning

Equally important in choosing a cleaning method is knowledge of the composition of the workpiece being cleaned. For example, abrasive tumbling can effectively remove burrs from a machined steel casting and leave a surface that is fully inspectable. This method, however, is not suitable for aluminum or magnesium, because it smears these metals and frequently hides flaws. Particular care must be taken in selecting a cleaning method for workpieces fabricated from more than one alloy (brazed assemblies are notable examples); a chemical cleaning method to remove scale, stopoff material, or flux must be chosen carefully to ensure that neither the braze nor the components of the assembly will be attacked. The surface finish of the workpiece must always be considered. When further processing is scheduled, such as

machining or final polishing, or when a surface finish of 3.20 m (125 in.) or coarser is allowed, an abrasive cleaning method is frequently a good choice. Generally, chemical cleaning methods have fewer degrading effects on surface finish than mechanical methods (unless the chemical used is strongly corrosive to the material being cleaned). Steam cleaning and solvent cleaning rarely have any effect on surface finish. Some materials are subject to delayed reactions as a result of improper cleaning. Two notable examples are high-strength steel and titanium. If it is ever necessary to chemically etch a high-strength steel workpiece, it should be baked at an appropriate temperature for a sufficient time to avoid hydrogen embrittlement. This should be done as soon after etching as possible but no later than 1 h. Titanium alloys can be subject to delayed cracking if they retain halogenated compounds and are then exposed to temperatures exceeding 480 °C (900 °F). Consequently, halogenated solvents should not be used for titanium and its alloys if their complete removal cannot be ensured. Choice of cleaning method may be dictated by Occupational Safety and Health Administration and Environmental Protection Agency health and safety regulations. Quantities of materials that will be used, toxicity, filtering, neutralization and disposal techniques, and worker protection all are crucial factors. Liquid Penetrant Inspection Revised by J.S. Borucki, Ardrox Inc., and Gail Jordan, Howmet Corporation

Penetrant Inspection Processing Parameters

It is extremely important to understand the significance of adhering to the established process parameters for a given application. Failure to control the process parameters will affect the quality of the inspection. For example, excessive overwashing or overemulsification can remove the penetrant from the flaws; minimal washing or underemulsification can result in excessive background, which could mask the flaws and render them undetectable. Processing time in each station, the equipment used, and other factors can vary widely, depending on workpiece size and shape, production quantities of similar workpieces, and required customer specifications for process parameters. Postemulsifiable Method The processing cycles for the postemulsifiable processes, method B (lipophilic) and method D (hydrophilic) are illustrated in the processing flow diagrams (Fig. 19 and 20, respectively). The major difference between the two methods, as described below, is the additional prerinse step utilized in method D.

Fig. 19 Processing flow diagram for the postemulsifiable, method B, lipophilic liquid penetrant system

Fig. 20 Processing flow diagram for the postemulsifiable, method D, hydrophilic liquid penetrant system

Application of Penetrant. Workpieces should be thoroughly and uniformly coated with penetrant by flowing,

brushing, swabbing, dipping, or spraying. Small workpieces requiring complete surface inspection are usually placed in a basket and dipped in the penetrant. Larger workpieces are usually brushed or sprayed. Electrostatic spray application is also very effective and economical. After the workpiece has been coated with a light film of penetrant, it should be positioned so that it can drain and so that excess penetrant cannot collect in pools. Workpieces should not be submerged during the entire penetration dwell time. Heating the workpiece is also not necessary or recommended, because certain disadvantages can occur, such as volatilization of the penetrant, difficulty in washing, and a decrease in fluorescence. Dwell Time. After the penetrant has been applied to the workpiece surface, it should be allowed to remain long enough for complete penetration into the flaws. Dwell time will vary, depending mainly on the size of the defects sought, cleanliness of the workpiece, and sensitivity and viscosity of the penetrant. In most cases, however, a minimum of 10 min and a maximum of 30 min is adequate for both fluorescent- and visible-penetrant types. A lengthy dwell time could cause the penetrant to begin drying on the surface, resulting in difficult removal. If drying does occur, it is necessary to reapply the penetrant to wet the surface and then begin the removal steps. Recommendations from the penetrant supplier will help establish the time, but experimentation will determine optimum dwell time. Prerinse. When using method D (hydrophilic), a coarse waterspray prerinse is needed to assist in penetrant removal and

to reduce contamination of the emulsifier. A coarse water spray is recommended, using a pressure of 275 to 345 kPa (40 to 50 psi). The prerinse water temperature should be 10 to 40 °C (50 to 100 °F). The prerinse time should be kept to a minimum (that is, 30 to 90 s) because the purpose is to remove excess penetrant so that the emulsifier does not become contaminated quickly. Emulsifier Application. It is very important that all surfaces of the workpiece be coated with the emulsifier at the same time. Small workpieces are dipped individually or in batches in baskets or on racks, whichever is the most convenient. For large workpieces, methods must be devised to achieve the fastest possible coverage; two methods often used are spraying or immersing. Localized emulsification of large workpieces can be achieved by spraying. The temperature of the emulsifier is not extremely critical, but a range of 20 to 30 °C (70 to 90 °F) is referred. Emulsification Time. The length of time the emulsifier is allowed to remain on the workpiece and in contact with the

penetrant is the emulsification time and depends mainly on the type of emulsifier employed, its concentration, and on the surface condition of the workpieces. Recommendations by the manufacturer of the emulsifier can serve as guidelines, but the optimum time for a specific workpiece must be established by experimentation. The surface finish, size, and composition of the workpiece will determine more precisely the choice of emulsifier and emulsification time. Emulsification time ranges from approximately 30 s to 3 min and is directly related to the concentration of the emulsifier. If emulsification time is excessive, penetrant will be removed from the flaws, making detection impossible. Rinsing. For all methods, removing the penetrant from the workpiece is probably the most important step in obtaining

reproducible results. If penetrant removal is performed properly, penetrant will be stripped from the surface and will remain only in the flaws. More variability in individual technique enters into this particular phase of inspection than any other step. Therefore, removal must be performed with the same sequence of operations time after time if results are to be reproducible. This is especially important when inspecting for tight or shallow flaws. Rinse time should be determined experimentally for specific workpieces; it usually varies from 10 s to 2 min. For spray rinsing, water pressure should be constant. A pressure of about 275 kPa (40 psi) is desirable; too much pressure may remove penetrants from the flaws. A coarse water spray is recommended and can be assisted with air (the combined water and air pressure should not exceed the pressure recommended for water alone). Water temperature should be maintained at a relatively constant level. Most penetrants can be removed effectively with water in a range of 10 to 40 °C (50 to 100 °F). Drying is best done in a recirculating hot-air drier that is thermostatically controlled. The temperature in the drier is normally between 65 and 95 °C (150 and 200 °F). The temperature of the workpieces should not be permitted to exceed 70 °C (160 °F). Workpieces should not remain in the drier any longer than necessary; drying is normally accomplished within a few minutes. Excessive drying at high temperatures can impair the sensitivity of the inspection. Because drying time will vary, the exact time should be determined experimentally for each type of workpiece. Developing depends on the form of developer to be used. Various types of developers are discussed below.

Dry-developer powder (form A) is applied after the workpiece has been dried and can be applied in a variety of

ways. The most common is dusting or spraying. Electrostatic spray application is also very effective. In some cases, application by immersing the workpiece into the dry powder developer is permissible. For simple applications, especially when only a portion of the surface of a large part is being inspected, applying with a soft brush is often adequate. Excess developer can be removed from the workpiece by a gentle air blast (140 kPa, or 20 psi, maximum) or by shaking or gentle tapping. Whichever means of application is chosen, it is important that the workpiece be completely and evenly covered by a fine film of developer. Water-soluble developer (form B) is applied just after the final wash and immediately prior to drying by dip, flow-

on, or spray techniques. No agitation of the developer bath is required. Removal of the developer coating from the surface of the workpiece is required and easily accomplished because the dried developer coating is water soluble and therefore completely removable by a water rinse. Water-suspendible developer (form C) is applied just after the final wash and immediately before drying. Dip,

flow-on, and spray are common methods of application. Care must be taken to agitate the developer thoroughly so that all particles are in suspension; otherwise, control of the concentration of the applied coating is impossible. Removal of the water-suspendible developer can best be achieved by water spray rinsing. If allowed to remain indefinitely on the workpiece, the developer can become difficult to remove. Solvent-suspendible nonaqueous developer (form D) is always applied after drying by spraying, either with

aerosol containers or by conventional or electrostatic methods. Proper spraying produces a thin, uniform layer that is very sensitive in producing either fluorescent or red visible indications. The volatility of the solvent makes it impractical to use in open tanks. Not only would there be solvent loss, reducing the effectiveness of the developer, but there would also be a hazardous vapor condition. Dipping, pouring, and brushing are not suitable for applying solvent-suspendible developer. Developing Time. In general, 10 min is the recommended minimum developing time regardless of the developer form

used. The developing time begins immediately after application of the developer. Excessive developing time is seldom necessary and usually results in excessive bleeding of indications, which can obscure flaw delineation. Inspections. After the prescribed development time, the inspection should begin. The inspection area should be

properly darkened for fluorescent penetrant inspection. Recommended black light intensity is 1000 to 1600 W/cm2. The intensity of the black light should be verified at regular intervals by the use of a suitable black light meter such as a digital radiometer. The intensity of the black light should be allowed to warm up prior to use--generally for about 10 min. The inspector should allow time for adapting to darkness; a 1-min period is usually adequate. White light intensity should not exceed 20 lx (2 ftc) to ensure the best inspection environment. Visible-penetrant systems provide vivid red indications that can be seen in visible light. Lighting intensity should be adequate to ensure proper inspection; 320 to 540 lx (30 to 50 ftc) is recommended. Lighting intensity should be verified at regular intervals by the use of a suitable white light meter such as a digital radiometer. Detailed information on inspection techniques is available in the sections "Inspection and Evaluation" and "Specifications and Standards" in this article. Water-Washable Method As indicated by the flow diagram in Fig. 21, the processing cycle for the water-washable method is similar to that for the postemulsifiable method. The difference lies in the penetrant removal step. As discussed in the section "Materials Used in Penetrant Inspection" in this article, the water-washable penetrants have a built-in emulsifier, thus eliminating the need for an emulsification step. One rinse operation is all that is required, and the washing operation should be carefully controlled because water-washable penetrants are susceptible to overwashing.

Fig. 21 Processing flow diagram for the water-washable liquid penetrant system

Rinse time should be determined experimentally for a specific workpiece; it usually varies from 10 s to 2 min. The best

practical way of establishing rinse time is to view the workpiece under a black light while rinsing and washing only until the fluorescent background is removed to a satisfactory degree. On some applications, such as castings, an immersion rinse followed by a final spray rinsing is desirable to remove tenacious background fluorescence. This technique, however, must be very carefully controlled to ensure that overwashing does not occur. For spray rinsing, a nominal water pressure of 140 to 275 kPa (20 to 40 psi) is recommended; too much pressure can

result in overwashing, that is, the removal of penetrant from within flaws. Hydro-air spray guns can be used. The air pressure, however, should not exceed 170 kPa (25 psi). The temperature of the water should be controlled to 10 to 40 °C (50 to 100 °F). Drying, developing, and inspection process parameters are the same as the postemulsifiable method process parameters described in the section "Postemulsifiable Method" in this article. Solvent-Removable Method The basic sequence of operations for the solvent-removable penetrant system is generally similar to that followed for the other methods. A typical sequence is shown by the flow diagram in Fig. 22. A notable difference is that with the solventremovable method the excess penetrant is removed by wiping with clean, lint-free material moistened with solvent. It is important to understand that flooding the workpiece to remove excess surface penetrant will also dissolve the penetrant from within the flaws.

Fig. 22 Processing flow diagram for the solvent-removable liquid penetrant system

The processing parameters for the use of developer are the same as those described above for the postemulsifiable method. Dry-powder developers, however, are not recommended for use with the visible solvent-removable penetrant method. Liquid Penetrant Inspection Revised by J.S. Borucki, Ardrox Inc., and Gail Jordan, Howmet Corporation

Penetrant Inspection Processing Parameters It is extremely important to understand the significance of adhering to the established process parameters for a given application. Failure to control the process parameters will affect the quality of the inspection. For example, excessive overwashing or overemulsification can remove the penetrant from the flaws; minimal washing or underemulsification can result in excessive background, which could mask the flaws and render them undetectable. Processing time in each station, the equipment used, and other factors can vary widely, depending on workpiece size and shape, production quantities of similar workpieces, and required customer specifications for process parameters. Postemulsifiable Method The processing cycles for the postemulsifiable processes, method B (lipophilic) and method D (hydrophilic) are illustrated in the processing flow diagrams (Fig. 19 and 20, respectively). The major difference between the two methods, as described below, is the additional prerinse step utilized in method D.

Fig. 19 Processing flow diagram for the postemulsifiable, method B, lipophilic liquid penetrant system

Fig. 20 Processing flow diagram for the postemulsifiable, method D, hydrophilic liquid penetrant system

Application of Penetrant. Workpieces should be thoroughly and uniformly coated with penetrant by flowing,

brushing, swabbing, dipping, or spraying. Small workpieces requiring complete surface inspection are usually placed in a basket and dipped in the penetrant. Larger workpieces are usually brushed or sprayed. Electrostatic spray application is also very effective and economical. After the workpiece has been coated with a light film of penetrant, it should be positioned so that it can drain and so that excess penetrant cannot collect in pools. Workpieces should not be submerged during the entire penetration dwell time. Heating the workpiece is also not necessary or recommended, because certain disadvantages can occur, such as volatilization of the penetrant, difficulty in washing, and a decrease in fluorescence. Dwell Time. After the penetrant has been applied to the workpiece surface, it should be allowed to remain long enough for complete penetration into the flaws. Dwell time will vary, depending mainly on the size of the defects sought, cleanliness of the workpiece, and sensitivity and viscosity of the penetrant. In most cases, however, a minimum of 10 min and a maximum of 30 min is adequate for both fluorescent- and visible-penetrant types. A lengthy dwell time could cause the penetrant to begin drying on the surface, resulting in difficult removal. If drying does occur, it is necessary to reapply the penetrant to wet the surface and then begin the removal steps. Recommendations from the penetrant supplier will help establish the time, but experimentation will determine optimum dwell time. Prerinse. When using method D (hydrophilic), a coarse waterspray prerinse is needed to assist in penetrant removal and

to reduce contamination of the emulsifier. A coarse water spray is recommended, using a pressure of 275 to 345 kPa (40 to 50 psi). The prerinse water temperature should be 10 to 40 °C (50 to 100 °F). The prerinse time should be kept to a minimum (that is, 30 to 90 s) because the purpose is to remove excess penetrant so that the emulsifier does not become contaminated quickly. Emulsifier Application. It is very important that all surfaces of the workpiece be coated with the emulsifier at the same time. Small workpieces are dipped individually or in batches in baskets or on racks, whichever is the most convenient. For large workpieces, methods must be devised to achieve the fastest possible coverage; two methods often used are spraying or immersing. Localized emulsification of large workpieces can be achieved by spraying. The temperature of the emulsifier is not extremely critical, but a range of 20 to 30 °C (70 to 90 °F) is referred. Emulsification Time. The length of time the emulsifier is allowed to remain on the workpiece and in contact with the

penetrant is the emulsification time and depends mainly on the type of emulsifier employed, its concentration, and on the surface condition of the workpieces. Recommendations by the manufacturer of the emulsifier can serve as guidelines, but the optimum time for a specific workpiece must be established by experimentation. The surface finish, size, and composition of the workpiece will determine more precisely the choice of emulsifier and emulsification time. Emulsification time ranges from approximately 30 s to 3 min and is directly related to the concentration of the emulsifier. If emulsification time is excessive, penetrant will be removed from the flaws, making detection impossible.

Rinsing. For all methods, removing the penetrant from the workpiece is probably the most important step in obtaining

reproducible results. If penetrant removal is performed properly, penetrant will be stripped from the surface and will remain only in the flaws. More variability in individual technique enters into this particular phase of inspection than any other step. Therefore, removal must be performed with the same sequence of operations time after time if results are to be reproducible. This is especially important when inspecting for tight or shallow flaws. Rinse time should be determined experimentally for specific workpieces; it usually varies from 10 s to 2 min. For spray rinsing, water pressure should be constant. A pressure of about 275 kPa (40 psi) is desirable; too much pressure may remove penetrants from the flaws. A coarse water spray is recommended and can be assisted with air (the combined water and air pressure should not exceed the pressure recommended for water alone). Water temperature should be maintained at a relatively constant level. Most penetrants can be removed effectively with water in a range of 10 to 40 °C (50 to 100 °F). Drying is best done in a recirculating hot-air drier that is thermostatically controlled. The temperature in the drier is normally between 65 and 95 °C (150 and 200 °F). The temperature of the workpieces should not be permitted to exceed 70 °C (160 °F). Workpieces should not remain in the drier any longer than necessary; drying is normally accomplished within a few minutes. Excessive drying at high temperatures can impair the sensitivity of the inspection. Because drying time will vary, the exact time should be determined experimentally for each type of workpiece. Developing depends on the form of developer to be used. Various types of developers are discussed below. Dry-developer powder (form A) is applied after the workpiece has been dried and can be applied in a variety of

ways. The most common is dusting or spraying. Electrostatic spray application is also very effective. In some cases, application by immersing the workpiece into the dry powder developer is permissible. For simple applications, especially when only a portion of the surface of a large part is being inspected, applying with a soft brush is often adequate. Excess developer can be removed from the workpiece by a gentle air blast (140 kPa, or 20 psi, maximum) or by shaking or gentle tapping. Whichever means of application is chosen, it is important that the workpiece be completely and evenly covered by a fine film of developer. Water-soluble developer (form B) is applied just after the final wash and immediately prior to drying by dip, flow-

on, or spray techniques. No agitation of the developer bath is required. Removal of the developer coating from the surface of the workpiece is required and easily accomplished because the dried developer coating is water soluble and therefore completely removable by a water rinse. Water-suspendible developer (form C) is applied just after the final wash and immediately before drying. Dip,

flow-on, and spray are common methods of application. Care must be taken to agitate the developer thoroughly so that all particles are in suspension; otherwise, control of the concentration of the applied coating is impossible. Removal of the water-suspendible developer can best be achieved by water spray rinsing. If allowed to remain indefinitely on the workpiece, the developer can become difficult to remove. Solvent-suspendible nonaqueous developer (form D) is always applied after drying by spraying, either with

aerosol containers or by conventional or electrostatic methods. Proper spraying produces a thin, uniform layer that is very sensitive in producing either fluorescent or red visible indications. The volatility of the solvent makes it impractical to use in open tanks. Not only would there be solvent loss, reducing the effectiveness of the developer, but there would also be a hazardous vapor condition. Dipping, pouring, and brushing are not suitable for applying solvent-suspendible developer. Developing Time. In general, 10 min is the recommended minimum developing time regardless of the developer form

used. The developing time begins immediately after application of the developer. Excessive developing time is seldom necessary and usually results in excessive bleeding of indications, which can obscure flaw delineation. Inspections. After the prescribed development time, the inspection should begin. The inspection area should be

properly darkened for fluorescent penetrant inspection. Recommended black light intensity is 1000 to 1600 W/cm2. The intensity of the black light should be verified at regular intervals by the use of a suitable black light meter such as a digital radiometer. The intensity of the black light should be allowed to warm up prior to use--generally for about 10 min. The inspector should allow time for adapting to darkness; a 1-min period is usually adequate. White light intensity should not exceed 20 lx (2 ftc) to ensure the best inspection environment.

Visible-penetrant systems provide vivid red indications that can be seen in visible light. Lighting intensity should be adequate to ensure proper inspection; 320 to 540 lx (30 to 50 ftc) is recommended. Lighting intensity should be verified at regular intervals by the use of a suitable white light meter such as a digital radiometer. Detailed information on inspection techniques is available in the sections "Inspection and Evaluation" and "Specifications and Standards" in this article. Water-Washable Method As indicated by the flow diagram in Fig. 21, the processing cycle for the water-washable method is similar to that for the postemulsifiable method. The difference lies in the penetrant removal step. As discussed in the section "Materials Used in Penetrant Inspection" in this article, the water-washable penetrants have a built-in emulsifier, thus eliminating the need for an emulsification step. One rinse operation is all that is required, and the washing operation should be carefully controlled because water-washable penetrants are susceptible to overwashing.

Fig. 21 Processing flow diagram for the water-washable liquid penetrant system

Rinse time should be determined experimentally for a specific workpiece; it usually varies from 10 s to 2 min. The best

practical way of establishing rinse time is to view the workpiece under a black light while rinsing and washing only until the fluorescent background is removed to a satisfactory degree. On some applications, such as castings, an immersion rinse followed by a final spray rinsing is desirable to remove tenacious background fluorescence. This technique, however, must be very carefully controlled to ensure that overwashing does not occur. For spray rinsing, a nominal water pressure of 140 to 275 kPa (20 to 40 psi) is recommended; too much pressure can

result in overwashing, that is, the removal of penetrant from within flaws. Hydro-air spray guns can be used. The air pressure, however, should not exceed 170 kPa (25 psi). The temperature of the water should be controlled to 10 to 40 °C (50 to 100 °F). Drying, developing, and inspection process parameters are the same as the postemulsifiable method process parameters described in the section "Postemulsifiable Method" in this article. Solvent-Removable Method The basic sequence of operations for the solvent-removable penetrant system is generally similar to that followed for the other methods. A typical sequence is shown by the flow diagram in Fig. 22. A notable difference is that with the solventremovable method the excess penetrant is removed by wiping with clean, lint-free material moistened with solvent. It is important to understand that flooding the workpiece to remove excess surface penetrant will also dissolve the penetrant from within the flaws.

Fig. 22 Processing flow diagram for the solvent-removable liquid penetrant system

The processing parameters for the use of developer are the same as those described above for the postemulsifiable method. Dry-powder developers, however, are not recommended for use with the visible solvent-removable penetrant method. Liquid Penetrant Inspection Revised by J.S. Borucki, Ardrox Inc., and Gail Jordan, Howmet Corporation

Postcleaning Some residue will remain on workpieces after penetrant inspection is completed. In many cases, this residue has no deleterious effects in subsequent processing or in service. There are, however, instances in which postcleaning is required. Residues can result in the formation of voids during subsequent welding or unwanted stopoff in brazing, in the contamination of surfaces (which can cause trouble in heat treating), or in unfavorable reactions in chemical processing operations. Drastic chemical or mechanical methods are seldom required for postcleaning. When justified by the volume of work, an emulsion cleaning line is effective and reasonable in cost. In special circumstances, ultrasonic cleaning may be the only satisfactory way of cleaning deep crevices or small holes. However, solvents or detergent-aided steam or water is almost always sufficient. The use of steam with detergent is probably the most effective of all methods. It has a scrubbing action that removes developers, the heat and detergent remove penetrants, it leaves a workpiece hot enough to promote rapid, even drying, and it is harmless to nearly all materials. Vapor degreasing is very effective for removing penetrants, but it is practically worthless for removing developers. It is frequently used in combination with steam cleaning. If this combination is used, the steam cleaning should always be done first because vapor degreasing bakes on developer films. Where conditions do not warrant or permit permanent cleaning installations, hand wiping with solvents is effective. Dried developer films can be brushed off, and residual penetrants can be rinsed off by solvent spraying or wiped off with a solvent-dampened cloth. Liquid Penetrant Inspection Revised by J.S. Borucki, Ardrox Inc., and Gail Jordan, Howmet Corporation

Quality Assurance of Penetrant Inspection Materials It is important to provide the controls necessary to ensure that the penetrant materials and equipment are operating at an acceptable level of performance. The frequency of the required checks should be based on a facility operating for a full, one-shift operation daily. In general, it is good practice to check the overall system performance on a daily basis. This check should be performed by processing a known defect standard through the line, using appropriate processing parameters and comparing the indications thus obtained to those obtained with fresh, unused penetrant material samples. When the performance of the in-use materials falls below that of the unused materials, the in-use material quality should be checked with the appropriate tests (as described below) and corrected prior to conducting any further penetrant inspection. Key quality assurance tests to be periodically conducted on in-use penetrants, emulsifiers, and developers are listed in Table 2. Also listed are the intervals at which the light sources and the overall system performance should be checked. Table 2 Intervals at which solutions, light sources, and system performance should be checked Minimum test frequency

Requirement

Fluorescent brightness

Quarterly

Not less than 90% of reference standard

Sensitivity

Monthly

Equal to reference standard

Removability (method A water wash only)

Monthly

Equal to reference standard

Water content (method A water wash penetrant only)

Monthly

Not to exceed 5%

Contamination

Weekly

No noticeable tracers

Removability

Weekly

Equal to reference standard

Water content (method B, lipophilic)

Monthly

Not to exceed 5%

Concentration (method D, hydrophilic)

Weekly

Not greater than 3% above initial concentration

Contamination

Weekly

No noticeable tracers

Daily

Must be fluffy, not caked

Test

Penetrants

Emulsifiers

Developers

Dry-developer form

Contamination

Daily

Not more than ten fluorescent specks observed in a 102 mm (4 in.) circle of sample

Wetting/coverage

Daily

Must be uniform/wet and must coat part

Contamination

Daily

Must not show evidence of fluorescence contaminates

Concentration

Weekly

Concentration shall be maintained as specified

Black lights

Daily

Minimum 1000

White light

Weekly

Minimum 2200 lx (200 ftc)

System performance

Daily

Must equal reference standards

Aqueous (soluble and suspended) developer

Other

W/cm2 at 381 mm (15 in.)

Military standard 6866 specifies the specific test procedure to use for the tests defined in Table 2. Penetrants applied by spray application from sealed containers are not likely to be exposed to the same working environment as with open dip tanks and are therefore not required to be tested as defined in Table 2 unless contamination is suspected. Liquid Penetrant Inspection Revised by J.S. Borucki, Ardrox Inc., and Gail Jordan, Howmet Corporation

Maintenance of Materials With constant open-tank use, penetrant materials are inherently subject to potential deterioration. Such factors as evaporation losses and contamination from various sources can contribute to deterioration. It is essential, therefore, to monitor the condition of these materials as described in Table 2. The evaporation of the volatile constituents of penetrants can alter their chemical and performance characteristics, thus resulting in changes in inherent brightness, removability, and sensitivity. Liquid penetrant materials qualified to MIL-I25135D (and subsequent revisions) have a flash point requirement of a minimum of 95 °C (200 °F) (per Pensky Martens flash point test procedure), assuring the minimization of evaporation losses. The contamination of water-washable penetrant with water is the most frequent source of difficulty. When present

beyond a critical percentage, this contamination will render the penetrant tank useless. For postemulsifiable penetrants, water contamination is not as critical a problem, because water is usually not miscible with postemulsifiable penetrants and will separate from the penetrant, which can then be subsequently removed. Water contamination can be minimized by implementing and following proper processing procedures. It is important to recognize that acid contamination (carryover from precleaning) will render fluorescent penetrants ineffective. Acid contamination changes the consistency of the penetrant and damages or destroys the fluorescent dye.

Dust, dirt and lint, and similar foreign materials get into the penetrant in the ordinary course of shop usage. These contaminants do no particular harm unless present to the extent that the bath is scummy with floating or suspended foreign material. Reasonable care should be taken to keep the penetrant clean. Workpieces containing adhering sand and dirt from the shop floor should be cleaned before being dipped into the penetrant. Contamination of the emulsifier must also be considered. Method B, lipophilic emulsifiers inherently become contaminated by penetrant through the normal processing of parts coated with penetrant being dipped into the emulsifier. It is imperative, therefore, that the lipophilic emulsifier have a high tolerance (that is, 10%) for penetrant contamination. Water contamination of the lipophilic emulsifier is always a potential problem due to the nature of the process. Generally, 5% water contamination can be tolerated.

Method D, hydrophilic emulsifiers are not normally subject to appreciable amounts of penetrant contamination, mainly because of the prerinse processing step, which removes most of the excess surface penetrant before emulsification. Because hydrophilic emulsifiers are water based, water contamination is not a problem, except for the fact that the bath concentration must be maintained at the prescribed limits. In general, emulsifiers that become severely contaminated will not properly emulsify the surface penetrant on the parts. Periodic monitoring is essential. Developer must also be maintained to ensure proper performance. Contamination of the dry-powder developer with water or moisture in the air can result in caking. Dry developers must remain fluffy and free flowing if they are to perform properly. In addition, contamination from the fluorescent penetrant must not occur. Fluorescent specks in the developer powder could be misinterpreted as an indication. Wet developer (soluble or suspendible) must not become contaminated with penetrant or any contaminant that could affect its ability to wet and evenly cover the workpiece. Liquid Penetrant Inspection Revised by J.S. Borucki, Ardrox Inc., and Gail Jordan, Howmet Corporation

Training and Certification of Personnel The apparent simplicity of the penetrant method is deceptive. Very slight variations in performing the penetrant process and the inspection can invalidate the inspection results by failing to indicate all flaws. Therefore, many companies require that penetrant inspection be conducted only by trained and certified personnel. Minimum requirements for personnel training and certification are described by various military and industry specifications (such as MIL-STD-410 and ASNT SNT-TC-1A). The following are examples of the most commonly followed training programs; however, specific customer training requirements are usually defined within the contract. Training is minimal for level I penetrant inspection operators (personnel responsible for the processing). However, the

penetrant process must be correctly performed to ensure accurate inspection. Operator training consists of the satisfactory completion of a period of on-the-job training, as determined by immediate supervision, conducted under the guidance of a certified level I inspector. Training for level II inspectors (personnel responsible for the inspection and evaluation) is more extensive than that for the level I operators. Training usually consists of 40 h of formal training, followed by several weeks of on-the-job training under the supervision of a designated trainer, usually a certified level II operator. Certification. Personnel of sufficient background and training in the principles and procedures of penetrant inspection

are usually certified by the successful completion of a practical test, which demonstrates their proficiency in penetrant techniques, and a written test, which documents their knowledge of penetrant inspection. Certified personnel are also normally required to pass a periodic eye examination, which includes a color-vision test. Certification can be obtained onsite through a certified level III inspector who may be with an outside source contracted to certify personnel or a company employee who has been certified as a level III inspector by the appropriate agency.

Liquid Penetrant Inspection Revised by J.S. Borucki, Ardrox Inc., and Gail Jordan, Howmet Corporation

Inspection and Evaluation After the penetrant process is completed, inspection and evaluation of the workpiece begin. Table 3 lists the more common types of flaws that can be found by penetrant inspection, together with their likely locations and their characteristics. Table 3 Common types, locations, and characteristics of flaws or discontinuities revealed by liquid penetrant inspection Locations

Characteristics

Shrinkage cracks

Castings (all metals)--on flat surfaces

Open

Hot tears

Castings (all metals)--at inside corners

Open

Cold shuts

Castings (all metals)--at changes in cross section

Tight, shallow

Folds

Castings (all metals)--anywhere

Tight, shallow

Inclusions

Castings, forgings, sheet, bar--anywhere

Tight, shallow, intermittent

Microshrinkage pores

Castings--anywhere

Spongy

Laps

Forgings, bar--anywhere

Tight, shallow

Forging cracks

Forgings--at inside or outside corners and at changes in cross section

Tight or open

Pipe

Forgings, bar--near geometric center

Irregular shape

Laminations

Sheet--at edges

Tight or open

Center bead cracks

Welds--at center of reinforcement

Tight or open

Cracks zone

Welds--at edge of reinforcement

Tight or open

Welds--at end of bead

Star-shaped

Type

Relevant indications

in

Crater cracks

heat-affected

Porosity

Castings, welds

Spherical

Grinding cracks

Any hard metal--ground surfaces

Tight, shallow, random

Quench cracks

Heat treated steel

Tight to open, oxidized

Stress-corrosion cracks

Any metal

Tight to corrosion

Fatigue cracks

Any metal

Tight

Weld spatter

Arc welds

Spherical or surface

Incomplete penetration

Fillet welds

Open, full weld length

Surface expulsion

Resistance welds

Raised metal at weld edge

Scuff marks

Seam welds

Surface of seam welds

Press-fit interface

Press fits

Outlines press fit

Braze runoff

Brazed parts

Edge of excess braze

Burrs

Machined parts

Bleeds heavily

Nicks, dents, scratches

All parts

Visible without penetrant aids

open;

may

show

Nonrelevant indications(a)

(a) These may be prohibited flaws, but are usually considered nonrelevant in penetrant testing.

Inspection Tools. An inspector must have tools that are capable of providing the required accuracy. These tools usually include suitable measuring devices, a flashlight, small quantities of solvent, small quantities of dry developers or aerosol cans of nonaqueous wet developers, pocket magnifiers ranging from 3 to 10×, and a suitable black light for fluorescent penetrants or sufficient white light for visible penetrants. Photographic standards or workpieces that have specific known flaws are sometimes used as inspection aids. A typical inspection begins with an overall examination to determine that the workpiece has been properly processed

and is in satisfactory condition for inspection. Inspection should not begin until the wet developers are completely dry. If developer films are too thick, if penetrant bleedout appears excessive, or if the penetrant background is excessive, the workpiece should be cleaned and reprocessed. When the inspector is satisfied that the workpiece is inspectable, it is examined according to a specified plan to be sure no areas have been missed. An experienced inspector can readily determine which indications are within acceptable limits and which ones are not. The inspector then measures all other indications. If the length or diameter of an indication exceeds allowable limits, it must be evaluated. One of the most

common and accurate ways of measuring indications is to lay a flat gage of the maximum acceptable dimension of discontinuity over the indication. If the indication is not completely covered by the gage, it is not acceptable. Evaluation. Each indication that is not acceptable should be evaluated. It may actually be unacceptable, it may be worse

than it appears, it may be false, it may be real, but nonrelevant, or it may actually be acceptable upon closer examination. One common method of evaluation includes the following steps: • • •

Wipe the area of the indication with a small brush or clean cloth that is dampened with a solvent Dust the area with a dry developer or spray it with a light coat of nonaqueous developer Remeasure under lighting appropriate for the type of penetrant used

If the discontinuity originally appeared to be of excessive length because of bleeding of penetrant along a scratch, crevice, or machining mark, this will be evident to a trained eye. Finally, to gain maximum assurance that the indication is properly interpreted, it is good practice to wipe the surface again with solvent-dampened cotton and examine the indication area with a magnifying glass and ample white light. This final evaluation may show that the indication is even larger than originally measured, but was not shown in its entirety because the ends were too tight to hold enough penetrant to reach the surface and become visible. Disposition of Unacceptable Workpieces. A travel ticket will usually accompany each workpiece or lot of

workpieces. Provision should be made on this ticket to indicate the future handling of unacceptable material, that is, scrapping, rework, repair, or review board action. There is often room on such tickets for a brief description of the indication. More often, indications are identified directly on the workpiece by circling them with some type of marking that is harmless to the material and not easily removed by accident, but removable when desired. Reworking an unacceptable flaw is often allowable to some specified limit; indications can be removed by sanding,

grinding, chipping, or machining. Repair welding is sometimes needed; in this case, the indication should be removed as in reworking before it is repair welded, or welding may move the flaw to a new location. In addition, it is imperative that all entrapped penetrant be removed prior to repair welding, because entrapped penetrant is likely to initiate a new flaw. Verification that the indication and the entrapped penetrant have been removed is required. Because reworking is usually required, it is good practice to finish it off with moderately fine sanding, followed by chemical etching to remove smeared metal. All traces of the etching fluid should be rinsed off, and the area should be thoroughly dried before reprocessing for reinspection. Reprocessing can be the same as original processing for penetrant inspection, or can be done locally by applying the materials with small brushes or swabs. False and Nonrelevant Indications. Because penetrant inspection provides only indirect indications or flaws, it

cannot always be determined at first glance whether an indication is real, false, or nonrelevant. A real indication is caused by an undesirable flaw, such as a crack. A false indication is an accumulation of penetrant not caused by a discontinuity in the workpiece, such as a drop of penetrant left on the workpiece inadvertently. A nonrelevant indication is an entrapment of penetrant caused by a feature that is acceptable even though it may exceed allowable indication lengths, such as a press-fit interface. Liquid Penetrant Inspection Revised by J.S. Borucki, Ardrox Inc., and Gail Jordan, Howmet Corporation

Specifications and Standards It has not been practical to establish any type of universal standardization, because of the wide variety of components and assemblies subjected to penetrant inspection, the differences in the types of discontinuities common to them, and the differences in the degree of integrity required. Generally, quality standards for the types of discontinuities detected by penetrant inspection are established by one or more of the following methods: •

Adoption of standards that have been successfully used for similar workpieces

• •

Evaluation of the results of penetrant inspection by destructive examination Experimental and theoretical stress analysis

Specifications. Normally, a specification is a document that delineates design or performance requirements. A

specification should include the methods of inspection and the requirements based on the inspection or test procedure. With penetrant inspection, this becomes difficult. Too often the wording in quality specifications is ambiguous and meaningless, such as "workpieces shall be free from detrimental defects" or "workpieces having questionable indications shall be held for review by the proper authorities." Specifications applicable to penetrant inspection are generally divided into two broad categories: those involving materials and equipment, and those concerning methods and standards. There are, however, several standards and specifications that are in common use; some of these are listed in Table 4. Because the equipment used for penetrant inspection covers such a broad scope, that is, ranging from small dip-tank setups to large automated installations, most emphasis in standards and specifications has been placed on the materials used in this inspection process. Table 4 Partial listing of standards and specifications for liquid penetrant inspection Number

Title or explanation of standard or specification

ASTM standards

ASTM E 165

Standard Practice for Liquid-Penetrant Inspection Method

ASTM E 270

Standard Definitions of Terms Relating to Liquid-Penetrant Inspection

ASTM E 1208

Standard Method for Fluorescent Liquid-Penetrant Examination Using the Lipophilic Post-Emulsification Process

ASTM E 1209

Standard Method for Fluorescent-Penetrant Examination Using the Water-Washable Process

ASTM E 1210

Standard Method for Fluorescent-Penetrant Examination Using the Hydrophilic Post-Emulsification Process

ASTM E 1219

Standard Method for Fluorescent-Penetrant Examination Using the Solvent-Removable Process

ASTM E 1220

Standard Method for Visible-Penetrant Examination Using the Solvent-Removable Process

ASTM E 1135

Standard Test Method for Comparing the Brightness of Fluorescent Penetrants

ASTM D 2512

Compatibility of Materials with Liquid Oxygen (Impact-Sensitivity Threshold Technique)

Test for AMS-SAE specifications

AMS 2647

Fluorescent Penetrant Inspection--Aircraft and Engine Component Maintenance

ASME specifications

ASME SEC V

ASME Boiler and Pressure Vessel Code Section V, Article 6

U.S. military and government specifications

MIL-STD-6866

Military Standard Inspection, Liquid Penetrant

MIL-STD-410

Nondestructive Testing Personnel Qualifications & Certifications

MIL-I-25135

Inspection Materials, Penetrant

MIL-I-25105

Inspection Unit, Fluorescent Penetrant, Type MA-2

MIL-I-25106

Inspection Unit, Fluorescent Penetrant, Type MA-3

MIL-STD-271 (Ships)

Nondestructive Testing Requirements for Metals

Control Systems. In conjunction with the specifications listed in Table 4, several methods and several types of

standards are used to check the effectiveness of liquid penetrants. One of the oldest and most frequently used methods involves chromium-cracked panels, which are available in sets containing fine, medium, and coarse cracks. Many other types of inspection standards have been produced--often for specific indications needed for a unique application. A comparison of indications from two water-washable penetrants of different sensitivity that were applied to a chromiumcracked panel containing fine cracks is shown in Fig. 23.

Fig. 23 Comparison of indications on chromium-cracked panels developed with water-washable liquid penetrants of low sensitivity (panel at left) and high sensitivity (panel at right)

Acceptance and rejection standards for liquid penetrant inspection are usually established for each individual item

or group of items by the designer. In most cases, acceptance and rejection standards are based on experience with similar items, the principal factor being the degree of integrity required. At one extreme, for certain noncritical items, the standard may permit some specific types of discontinuities all over the workpiece or in specified areas. Inspection is often applied only on a sampling basis for noncritical items. At the opposite extreme, items are subjected to 100% inspection, and requirements are extremely stringent to the point of defining the limitations on each specific area. Magnetic Particle Inspection Revised by Art Lindgren, Magnaflux Corporation

Introduction MAGNETIC PARTICLE INSPECTION is a method of locating surface and subsurface discontinuities in ferromagnetic materials. It depends on the fact that when the material or part under test is magnetized, magnetic discontinuities that lie in a direction generally transverse to the direction of the magnetic field will cause a leakage field to be formed at and above the surface of the part. The presence of this leakage field, and therefore the presence of the discontinuity, is detected by the use of finely divided ferromagnetic particles applied over the surface, with some of the particles being gathered and held by the leakage field. This magnetically held collection of particles forms an outline of the discontinuity and generally indicates its location, size, shape, and extent. Magnetic particles are applied over a surface as dry particles, or as wet particles in a liquid carrier such as water or oil. Ferromagnetic materials include most of the iron, nickel, and cobalt alloys. Many of the precipitation-hardening steels, such as 17-4 PH, 17-7 PH, and 15-4 PH stainless steels, are magnetic after aging. These materials lose their ferromagnetic properties above a characteristic temperature called the Curie point. Although this temperature varies for different materials, the Curie point for most ferromagnetic materials is approximately 760 °C (1400 °F). Magnetic Particle Inspection Revised by Art Lindgren, Magnaflux Corporation

Method Advantages and Limitations Nonferromagnetic materials cannot be inspected by magnetic particle inspection. Such materials include aluminum alloys, magnesium alloys, copper and copper alloys, lead, titanium and titanium alloys, and austenitic stainless steels. In addition to the conventional magnetic particle inspection methods described in this article, there are several proprietary methods that employ ferromagnetic particles on a magnetized testpiece. Three of these methods--magnetic rubber inspection, magnetic printing, and magnetic painting--are described in the Appendix to this article. Applications. The principal industrial uses of magnetic particle inspection are final inspection, receiving inspection, in-

process inspection and quality control, maintenance and overhaul in the transportation industries, plant and machinery maintenance, and inspection of large components. Although in-process magnetic particle inspection is used to detect discontinuities and imperfections in materials and parts as early as possible in the sequence of operations, final inspection is needed to ensure that rejectable discontinuities or imperfections detrimental to the use or function of the part have not developed during processing. During receiving inspection, semifinished purchased parts and raw materials are inspected to detect any initially defective material. Magnetic particle inspection is extensively used on incoming rod and bar stock, forging blanks, and rough castings. The transportation industries (truck, railroad, and aircraft) have planned overhaul schedules at which critical parts are magnetic particle inspected for cracks. Planned inspection programs are also used in keeping plant equipment in operation without breakdowns during service. Because of sudden and severe stress applications, punch-press crankshafts, frames,

and flywheels are vulnerable to fatigue failures. A safety requirement in many plants is the inspection of crane hooks; fatigue cracks develop on the work-hardened inside surfaces of crane hooks where concentrated lifting loads are applied. The blading, shaft, and case of steam turbines are examined for incipient failure at planned downtimes. Advantages. The magnetic particle method is a sensitive means of locating small and shallow surface cracks in

ferromagnetic materials. Indications may be produced at cracks that are large enough to be seen with the naked eye, but exceedingly wide cracks will not produce a particle pattern if the surface opening is too wide for the particles to bridge. Discontinuities that do not actually break through the surface are also indicated in many cases by this method, although certain limitations must be recognized and understood. If a discontinuity is fine, sharp, and close to the surface, such as a long stringer of nonmetallic inclusions, a clear indication can be produced. If the discontinuity lies deeper, the indication will be less distinct. The deeper the discontinuity lies below the surface, the larger it must be to yield a readable indication and the more difficult the discontinuity is to find by this method. Magnetic particle indications are produced directly on the surface of the part and constitute magnetic pictures of actual discontinuities. There is no electrical circuitry or electronic readout to be calibrated or kept in proper operating condition. Skilled operators can sometimes make a reasonable estimate of crack depth with suitable powders and proper technique. Occasional monitoring of field intensity in the part is needed to ensure adequate field strength. There is little or no limitation on the size or shape of the part being inspected. Ordinarily, no elaborate precleaning is necessary, and cracks filled with foreign material can be detected. Limitations. There are certain limitations to magnetic particle inspection the operator must be aware of; for example,

thin coatings of paint and other nonmagnetic coverings, such as plating, adversely affect the sensitivity of magnetic particle inspection. Other limitations are: • • • • • • •

The method can be used only on ferromagnetic materials For best results, the magnetic field must be in a direction that will intercept the principal plane of the discontinuity; this sometimes requires two or more sequential inspections with different magnetizations Demagnetization following inspection is often necessary Postcleaning to remove remnants of the magnetic particles clinging to the surface may sometimes be required after testing and demagnetization Exceedingly large currents are sometimes needed for very large parts Care is necessary to avoid local heating and burning of finished parts or surfaces at the points of electrical contact Although magnetic particle indications are easily seen, experience and skill are sometimes needed to judge their significance

Specifications and standards for magnetic particle inspection have been developed by several technical associations

and divisions of the U.S. Department of Defense. Sections III, V, and VIII of the ASME Boiler and Pressure Vessel Code contain specifications for nondestructive inspection of the vessels. Several Aerospace Material Specifications (published by the Society of Automotive Engineers) and standards from the American Society for Testing and Materials cover magnetic particle inspection. Various military standards include specifications for vendors to follow in establishing inspection procedures for military equipment and supplies. American Society for Nondestructive Testing Recommended Practice SNT-TC-1A is a guide to the employer for establishing in-house procedures for training, qualification, and certification of personnel whose jobs require appropriate knowledge of the principles underlying the nondestructive inspection they perform. Magnetic Particle Inspection Revised by Art Lindgren, Magnaflux Corporation

Description of Magnetic Fields

Magnetic fields are used in magnetic particle inspection to reveal discontinuities. Ferromagnetism is the property of some metals, chiefly iron and steel, to attract other pieces of ferromagnetic materials. A horseshoe magnet will attract magnetic materials to its ends, or poles. Magnetic lines of force, or flux, flow from the south pole through the magnet to the north pole. Magnetized Ring. When a magnetic material is placed across the poles of a horseshoe magnet having square ends, forming a closed or ringlike assembly, the lines of force flow from the north pole through the magnetic material to the south pole (Fig. 1a). (Magnetic lines of force flow preferentially through magnetic material rather than through nonmagnetic material or air.) The magnetic lines of force will be enclosed within the ringlike assembly because no external poles exist, and iron filings or magnetic particles dusted over the assembly are not attracted to the magnet even though there are lines of magnetic force flowing through it. A ringlike part magnetized in this manner is said to contain a circular magnetic field that is wholly within the part.

Fig. 1 Schematics of magnetic lines of force. (a) Horseshoe magnet with a bar of magnetic material across poles, forming a closed, ringlike assembly, which will not attract magnetic particles. (b) Ringlike magnet assembly with an air gap, to which magnetic particles are attracted

If one end of the magnet is not square and an air gap exists between that end of the magnet and the magnetic material, the poles will still attract magnetic materials. Magnetic particles will cling to the poles and bridge the gap between them, as shown in Fig. 1(b). Any radial crack in a circularly magnetized piece will create a north and a south magnetic pole at the edges of a crack. Magnetic particles will be attracted to the poles created by such a crack, forming an indication of the discontinuity in the piece. The fields set up at cracks or other physical or magnetic discontinuities in the surface are called leakage fields. The strength of a leakage field determines the number of magnetic particles that will gather to form indications; strong indications are formed at strong fields, weak indications at weak fields. The density of the magnetic field determines its strength and is partly governed by the shape, size, and material of the part being inspected. Magnetized Bar. A straight piece of magnetized material (bar magnet) has a pole at each end. Magnetic lines of force flow through the bar from the south pole to the north pole. Because the magnetic lines of force within the bar magnet run the length of the bar, it is said to be longitudinally magnetized or to contain a longitudinal field.

If a bar magnet is broken into two pieces, a leakage field with north and south poles is created between the pieces, as shown in Fig. 2(a). This field exists even if the fracture surfaces are brought together (Fig. 2b). If the magnet is cracked but not broken completely in two, a somewhat similar result occurs. A north and a south pole form at opposite edges of the crack, just as though the break were complete (Fig. 2c). This field attracts the iron particles that outline the crack. The strength of these poles will be different from that of the fully broken pieces and will be a function of the depth of the crack and the width of the air gap at the surface.

Fig. 2 Leakage fields between two pieces of a broken bar magnet. (a) Magnet pieces apart. (b) Magnet pieces together (which would simulate a flaw). (c) Leakage field at a crack in a bar magnet

The direction of the magnetic field in an electromagnetic circuit is controlled by the direction of the flow of magnetizing current through the part to be magnetized. The magnetic lines of force are always at right angles to the direction of current flow. To remember the direction taken by the magnetic lines of force around a conductor, consider that the conductor is grasped with the right hand so that the thumb points in the direction of current flow. The fingers then point in the direction taken by the magnetic lines of force in the magnetic field surrounding the conductor. This is known as the right-hand rule. Circular Magnetization. Electric current passing through any straight conductor such as a wire or bar creates a

circular magnetic field around the conductor. When the conductor of electric current is a ferromagnetic material, the passage of current induces a magnetic field in the conductor as well as in the surrounding space. A part magnetized in this manner is said to have a circular field or to be circularly magnetized, as shown in Fig. 3(a).

Fig. 3 Magnetized bars showing directions of magnetic field. (a) Circular. (b) Longitudinal

Longitudinal Magnetization. Electric current can also be used to create a longitudinal magnetic field in magnetic materials. When electric current is passed through a coil of one or more turns, a magnetic field is established lengthwise or longitudinally, within the coil, as shown in Fig. 3(b). The nature and direction of the field around the conductor that forms the turns of the coil produce longitudinal magnetization. Effect of Flux Direction. To form an indication, the magnetic field must approach a discontinuity at an angle great

enough to cause the magnetic lines of force to leave the part and return after bridging the discontinuity. For best results, an intersection approaching 90° is desirable. For this reason, the direction, size, and shape of the discontinuity are important. The direction of the magnetic field is also important for optimum results, as is the strength of the field in the area of the discontinuity. Figure 4(a) illustrates a condition in which the current is passed through the part, causing the formation of a circular field around the part. Under normal circumstances, a discontinuity such as A in Fig. 4(a) would give no indication of its presence, because it is regular in shape and lies parallel to the magnetic field. If the discontinuity has an irregular shape but is predominantly parallel to the magnetic field, such as B, there is a good possibility that a weak indication would form. Where the predominant direction of the discontinuity is at a 45° angle to the magnetic field, such as C, D, and E, the conditions are more favorable for detection regardless of the shape of the discontinuity. Discontinuities whose predominant directions, regardless of shape, are at a 90° angle to the magnetic field produce the most pronounced indications (F, G, and H, Fig. 4a).

Fig. 4 Effect of direction of magnetic field or flux flow on the detectability of discontinuities with various orientations. (a) Circular magnetization. (b) Longitudinal magnetization. See text for discussion.

A longitudinally magnetized bar is shown in Fig. 4(b). Discontinuities L, M, and N, which are at about 45° to the magnetic field, would produce detectable indications as they would with a circular field. Discontinuities J and K would display pronounced indications, and weak indications would be produced at discontinuities P, Q, and R. Magnetization Methods. In magnetic particle inspection, the magnetic particles can be applied to the part while the magnetizing current is flowing or after the current has ceased, depending largely on the retentivity of the part. The first technique is known as the continuous method; the second, the residual method.

If the magnetism remaining in the part after the current has been turned off for a period of time (residual magnetism) does not provide a leakage field strong enough to produce readable indications when magnetic particles are applied to the surface, the part must be continuously magnetized during application of the particles. Consequently, the residual method can be used only on materials having sufficient retentivity; usually the harder the material, the higher the retentivity. The continuous method can be used for most parts. Magnetic Particle Inspection Revised by Art Lindgren, Magnaflux Corporation

Magnetizing Current Both direct current (dc) and alternating current (ac) are suitable for magnetizing parts for magnetic particle inspection. The strength, direction, and distribution of magnetic fields are greatly affected by the type of current used for magnetization. The fields produced by direct and alternating current differ in many respects. The important difference with regard to magnetic particle inspection is that the fields produced by direct current generally penetrate the cross section of the part,

while the fields produced by alternating current are confined to the metal at or near the surface of the part, a phenomenon known as the skin effect. Therefore, alternating current should not be used in searching for subsurface discontinuities. Direct Current. The best source of direct current is the rectification of alternating current. Both the single-phase (Fig. 5a) and three-phase types of alternating current (Fig. 5b) are furnished commercially. By using rectifiers, the reversing alternating current can be converted into unidirectional current, and when three-phase alternating current is rectified in this manner (Fig. 5c), the delivered direct current is entirely the equivalent of straight direct current for purposes of magnetic particle inspection. The only difference between rectified three-phase alternating current and straight direct current is a slight ripple in the value of the rectified current, amounting to only about 3% of the maximum current value.

Fig. 5 Alternating current wave forms. (a) Single-phase. (b) Three-phase. (c) Three-phase rectified. (d) Halfwave rectified single-phase. (e) Full-wave rectified single-phase

When single-phase alternating current is passed through a simple rectifier, current is permitted to flow in one direction only.The reverse half of each cycle is completely blocked out (Fig. 5d). The result is unidirectional current (called halfwave current) that pulsates; that is, it rises from zero to a maximum and then drops back to zero. During the blocked-out reverse of the cycle, no current flows, then the half-cycle forward pulse is repeated, at a rate of 60 pulses per second. A rectifier for alternating current can also be connected so that the reverse half of the cycle is turned around and fed into the circuit flowing in the same direction as the first half of the cycle (Fig. 5e). This produces pulsating direct current, but with no interval between the pulses. Such current is referred to as single-phase full-wave direct current or full-wave rectified single-phase alternating current. There is a slight skin effect from the pulsations of the current, but it is not pronounced enough to have a serious impact on the penetrations of the field. The pulsation of the current is useful because it imparts some slight vibration to the magnetic particles, assisting them in arranging themselves to form indications. Half-wave current, used in magnetization with prods and dry magnetic particles, provides the highest sensitivity for discontinuities that are wholly below the surface, such as those in castings and weldments.

Magnetization employing surges of direct current can be used to increase the strength of magnetic fields; for example, a rectifier capable of continuously delivering 400-A current can put out much more than 400 A for short intervals. Therefore, it is possible, by suitable current-control and switching devices, to pass a very high current for a short period (less than a second) and then reduce the current, without interrupting it, to a much lower value. Alternating current, which must be single-phase when used directly for magnetizing purposes, is taken from

commercial power lines and usually has a frequency of 50 or 60 Hz. When used for magnetizing, the line voltage is stepped down, by means of transformers, to the low voltages required. At these low voltages, magnetizing currents of several thousand amperes are often used. One problem encountered when alternating current is used is that the resultant residual magnetism in the part may not be at a level as high as that of the magnetism generated by the peak current of the ac cycle. This is because the level of residual magnetism depends on where in the cycle the current was discontinued. Magnetic Particle Inspection Revised by Art Lindgren, Magnaflux Corporation

Power Sources Early power sources were general-purpose units designed to use either alternating or direct current for magnetization. When direct current was used, it was derived directly from a bank of storage batteries, and a carbon-pile rheostat was used to regulate current level. Subsequent advances in technology have made the storage battery obsolete as a power supply and have given rise to many innovations, especially in the area of current control. Portable, mobile, and stationary equipment is currently available, and selection among these types depends on the nature and location of testing. Portable equipment is available in light-weight (16 to 40 kg, or 35 to 90 lb) power source units that can be readily

taken to the inspection site. Generally, these portable units are designed to use 115-, 230-, or 460-V alternating current and to supply magnetizing-current outputs of 750 to 1500 A in half-wave or alternating current. Machines capable of supplying half-wave current and alternating current and having continuously variable (infinite) current control can be used for magnetic particle inspection in a wide range of applications. Primary application of this equipment is hand-held prod inspection utilizing the half-wave output in conjunction with dry powder. In general, portable equipment is designed to operate with relatively short power supply cables, and the output is very limited when it is necessary to use longer cables. The major disadvantage of portable equipment is the limited amount of current available. For the detection of deep-lying discontinuities and for coverage of a large area with one prod contact, a machine with higher-amperage output is required. Also, portable equipment cannot supply the full-wave direct current necessary for some inspections and does not have the accessories found on larger mobile equipment. Mobile units are generally mounted on wheels to facilitate transportation to various inspection sites. Mobile equipment

usually supplies half-wave or alternating magnetizing-current outputs. Inspection of parts is accomplished with flexible cables, yokes, prod contacts, contact clamps, and coils. Instruments and controls are mounted on the front of the unit. Magnetizing current is usually controlled by a remote-control switch connected to the unit by an electric cord. Quickcoupling connectors for connecting magnetizing cables are on the front of the unit. Mobile equipment is usually powered by single-phase, 60-Hz alternating current (230 or 460 V) and has an output range of 1500 to 6000 A. On some units, current control is provided by a power-tap switch, which varies the voltage applied to the primary coil of the power transformer; most of these have either 8 or 30 steps of current control. However, current control on more advanced units is provided either by solid-state phase control of the transformer or by use of a saturablecore reactor to control the transformer. Phase control of the transformer is achieved by silicon-controlled rectifiers or triacs in series with the transformer. A solid-state control circuit is used to rapidly switch the ac supply on and off for controlled fractions of each cycle. A triac provides current control in both directions, while a saturable-core reactor provides current control in one direction only. In a circuit employing a saturable-core reactor to control magnetizingcurrent output, a silicon-controlled rectifier is used in conjunction with phase control to control a saturable-core reactor that is in series with, and that controls the input to, the power transformer.

Standard instruments and controls on mobile equipment are as follows: • • • • • •

Ammeters to indicate the magnetizing current flowing through the yokes, prods, or coil as being alternating, half-wave, or direct current Switches for controlling the magnetizing or demagnetizing current Pilot light to indicate when power to the unit is on Current control, either stepped or continuously variable, for controlling the amount of magnetizing and demagnetizing current Remote-control cable receptacle that permits turning the magnetizing current on and off at some distance from the unit Receptacles to permit the connection of the magnetizing-current cables

Built-in demagnetizers as contained in mobile equipment for magnetic particle inspection are available that demagnetize by one of four methods: •

• •

•

Low-voltage high-amperage alternating current with a motor-driven power-tap switch, arranged to automatically provide periods of current-on and periods of current-off in succession, with the amount of demagnetizing current reduced with each successive step Low-voltage high-amperage alternating current provided by a continuously variable current control that affords complete control of the demagnetization current from full-on to zero Current-decay method, in which low-voltage high-amperage alternating current is caused to decay from some maximum value to zero in an automatic and controlled manner. Because the entire cycle can be completed in a few seconds, the current-decay method offers an advantage over some of the more timeconsuming methods Low-voltage high-amperage direct current, with which demagnetization is accomplished by a motordriven power-tap switch in conjunction with means for reversing the current direction from positive to negative as the current is systematically reduced in steps of current-on periods followed by current-off periods

Stationary equipment can be obtained as either general-purpose or special-purpose units. The general-purpose unit is primarily intended for use in the wet method and has a built-in tank that contains the bath pump, which continuously agitates the bath and forces the fluid through hoses onto the part being inspected. Pneumatically operated contact heads, together with a rigid-type coil, provide capabilities for both circular and longitudinal magnetization. Self-contained ac or dc power supplies are available in amperage ratings from 2500 to 10,000 A. Maximum opening between contact plates varies from 0.3 to 3.7 m (1 to 12 ft).

Optional features that are available include self-regulating current control, automatic magnetizing circuit, automatic bath applicator, steady rests for heavy parts, and demagnetizing circuitry. Other options are curtains or hoods and an ultraviolet light; these are used during inspection with fluorescent particles. Stationary power packs serve as sources of high-amperage magnetizing current to be used in conjunction with special fixtures or with cable-wrap or clamp-and-contact techniques. Rated output varies from a customary 4000 to 6000 A to as high as 20,000 A. The higher-amperage units are used for the overall magnetization of large forgings or castings that would otherwise require systematic prod inspection at much lower current levels. Multidirectional Magnetizing. Some units feature up to three output circuits that are systematically energized in

rapid sequence, either electrically or mechanically, for effectively magnetizing a part in several directions in the same time frame. This reveals discontinuities lying in any direction after only a single processing step. Special-purpose stationary units are designed for handling and inspecting large quantities of similar items.

Generally, conveyors, automatic markers, and alarm systems are included in such units to expedite the handling of parts.

Magnetic Particle Inspection Revised by Art Lindgren, Magnaflux Corporation

Methods of Generating Magnetic Fields One of the basic requirements of magnetic particle inspection is that the part undergoing inspection be properly magnetized so that the leakage fields created by discontinuities will attract the magnetic particles. Permanent magnets serve some useful purpose in this respect, but magnetization is generally produced by electromagnets or the magnetic field associated with the flow of electric current. Basically, magnetization is derived from the circular magnetic field generated when an electric current flows through a conductor. The direction of this field is dependent on the direction of current flow, which can be determined by applying the right-hand rule (see the section "Description of Magnetic Fields" in this article). General applications, advantages, and limitations of the various methods of magnetizing parts for magnetic particle inspection are summarized in Table 1. Additional information can be found in the article "Magnetic Field Testing" in this Volume. Table 1 General applications, advantages, and limitations of the various magnetizing methods used in magnetic particle inspection Advantages

Limitations

Medium-size parts whose length predominates, such as a crankshaft or camshaft

All generally longitudinal surfaces are longitudinally magnetized to locate transverse discontinuities.

Part should be centered in coil to maximize length effectively magnetized during a given shot. Length may dictate additional shots as coil is repositioned.

Large castings, forgings, or shafts

Longitudinal field easily attained by wrapping with a flexible cable.

Multiple processing may be required because of part shape.

Miscellaneous small parts

Easy and fast, especially where residual method is applicable. Noncontact with part. Relatively complex parts can usually be processed with same ease as simple cross section.

Length-to-diameter (L/D) ratio is important in determining adequacy of ampere-turns; effective ratio can be altered by utilizing pieces of similar cross-sectional area. Sensitivity diminishes at ends of part because of general leakage field pattern. Quick break of current is desirable to minimize end effect on short parts with low L/D ratios.

Inspection of large surface areas for surface discontinuities

No electrical contact. Highly portable. Can locate discontinuities in any direction, with proper yoke orientation

Time consuming. Yoke must be systematically repositioned to locate discontinuities with random orientation.

Miscellaneous parts requiring inspection of localized areas

No electrical contact. Good sensitivity to surface discontinuities. Highly portable. Wet or dry method can be used. Alternating current yoke can also serve as demagnetizer in some cases.

Yoke must be properly positioned relative to orientation of discontinuity. Relatively good contact must be established between part and poles of yoke; complex part shape may cause difficulty. Poor sensitivity to subsurface discontinuities except in isolated areas

Application

Coils (single or multiple loop)

Yokes

Central conductors

Miscellaneous short parts having holes through which a conductor can be threaded, such as bearing rings, hollow cylinders, gears, large nuts, large clevises, and pipe couplings

No electrical contact, so that possibility of burning is eliminated. Circumferentially directed magnetic field is generated in all surfaces surrounding the conductor. Ideal for parts for which the residual method is applicable. Lightweight parts can be supported by the central conductor. Multiple turns can be used to reduce the amount of current required.

Size of conductor must be ample to carry required current. Ideally, conductor should be centrally located within hole. Large-diameter parts require several setups with conductor near or against inner surface and rotation of part between setups. Where continuous method is being employed, inspection is required after each setup.

Long tubular parts such as pipe, tubing, hollow shafts

No electrical contact. Both inside (ID) and outside (OD) surfaces can be inspected. Entire length of part is circularly magnetized.

Sensitivity of outer surface to indications may be somewhat diminished relative to inner surface for large-diameter and thick-wall parts.

Large valve bodies and similar parts

Good sensitivity to inner-surface discontinuities

Same as for long tubular parts, above

Fast, easy process. Complete circular field surrounds entire current path. Good sensitivity to surface and near-surface discontinuities. Simple as well as relatively complex parts can usually be easily inspected with one or more shots.

Possibility of burning part exists if proper contact conditions are not met. Long parts should be inspected in sections to facilitate bath application without resorting to an excessively long current shot.

Large castings and forgings

Large surface areas can be inspected in a relatively short time.

High amperage requirements (8000-20,000 A) dictate use of special direct current power pack.

Long tubular parts such as tubing, pipe, and hollow shafts

Entire length can be circularly magnetized by contacting end-to-end.

Effective field is limited to outer surface so process cannot be used to inspect inner surface. Part ends must be shaped to permit electrical contact and must be able to carry required current without excessive heating.

Long solid parts such as billets, bars, and shafts

Entire length can be circularly magnetized by contacting end-to-end. Amperage requirements are independent of length. No loss of magnetism at ends

Voltage requirements increase as length increases because of greater impedance of cable and part. Ends of parts must have shape that permits electrical contact and must be capable of carrying required current without excessive heating.

Circular field can be selectively directed to weld area by prod placement. In conjunction with halfwave current and dry powder, provides excellent sensitivity to subsurface discontinuities. Prods, cables, and power packs can be brought to

Only small area can be inspected at one time. Arc burn can result from poor contact. Surface must be dry when dry powder is being used. Prod spacing must be in accordance with

Direct contact, head shot

Solid, relatively small parts (cast, forged, or machined) that can be inspected on a horizontal wet-method unit

Direct contact, clamps and cables

Prod contacts

Welds, for cracks, inclusions, open roots, or inadequate joint penetration

inspection site.

magnetizing-current level.

Entire surface area can be inspected in small increments using nominal current values. Circular magnetic field can be concentrated in specific areas likely to contain discontinuities. Prods, cables, and power packs can be brought to the inspection site.

Coverage of large surface areas requires a multiplicity of shots, which can be very time consuming. Arc burn can result from poor contact. Surface must be dry when dry powder is being used.

Ring-shaped parts, for circumferential discontinuities

No electrical contact. All surfaces of part are subjected to a toroidal magnetic field. 100% coverage is obtained in single magnetization. Can be automated

Laminated core is required through ring to enhance magnetic path. Type of magnetizing current must be a compatible with magnetic hardness or softness of metal inspected. Other conductors encircling field must be avoided.

Balls

No electrical contact. Permits 100% coverage for indications of discontinuities in any direction by use of a three-step process with reorientation of ball between steps. Can be automated

For small-diameter balls, use is limited to residual method of magnetization.

Disks and gears

No electrical contact. Good sensitivity at or near periphery or rim. Sensitivity in various areas can be varied by selection of core or pole piece. In conjunction with half-wave current and dry powder, provides excellent sensitivity to discontinuities lying just below the surface

100% coverage may require two-step process. Type of magnetizing current must be compatible with magnetic hardness or softness of metal inspected.

Large castings or forgings

Induced current

Yokes There are two basic types of yokes that are commonly used for magnetizing purposes: permanent-magnet and electromagnetic yokes. Both are hand held and therefore quite mobile. Permanent-magnet yokes are used for applications where a source of electric power is not available or where arcing is not permissible (as in an explosive atmosphere). The limitations of permanent-magnet yokes include the following:

• • • •

Large areas or masses cannot be magnetized with enough strength to produce satisfactory crack indications Flux density cannot be varied at will If the magnet is very strong, it may be difficult to separate from a part Particles may cling to the magnet, possibly obscuring indications

Electromagnetic yokes (Fig. 6) consist of a coil wound around a U-shaped core of soft iron. The legs of the yoke can

be either fixed or adjustable. Adjustable legs permit changing the contact spacing and the relative angle of contact to accommodate irregularly-shaped parts. Unlike a permanent-magnet yoke, an electromagnetic yoke can readily be switched on or off. This feature makes it convenient to apply and remove the yoke from the testpiece.

Fig. 6 Electromagnetic yoke showing position and magnetic field for the detection of discontinuities parallel to a weld bead. Discontinuities across a weld bead can be detected by placing the contact surfaces of the yoke next to and on either side of the bead (rotating yoke about 90° from position shown here).

The design of an electromagnetic yoke can be based on the use of either direct or alternating current or both. The flux density of the magnetic field produced by the direct current type can be changed by varying the amount of current in the coil. The direct current type of yoke has greater penetration while the alternating current type concentrates the magnetic field at the surface of the testpiece, providing good sensitivity for the disclosure of surface discontinuities over a relatively broad area. In general, discontinuities to be disclosed should be centrally located in the area between pole pieces and oriented perpendicular to an imaginary line connecting them (Fig. 6). Extraneous leakage fields in the immediate vicinity of the poles cause an excessive buildup of magnetic particles. In operation, the part completes the magnetic path for the flow of magnetic flux. The yoke is a source of magnetic flux, and the part becomes the preferential path completing the magnetic circuit between the poles. (In Fig. 6, only those portions of the flux lines near the poles are shown.) Yokes that use alternating current for magnetization have numerous applications and can also be used for demagnetization. Coils Single-loop and multiple-loop coils (conductors) are used for the longitudinal magnetization of components (Fig. 3b and 4b). The field within the coil has a definite direction, corresponding to the direction of the lines of force running through it. The flux density passing through the interior of the coil is proportional to the product of the current, I, in amperes, and the number of turns in the coil, N. Therefore, the magnetizing force of such a coil can be varied by changing either the current or the number of turns in the coil. For large parts, a coil can be produced by winding several turns of a flexible cable around the part, but care must be taken to ensure that no indications are concealed beneath the cable. Portable magnetizing coils are available that can be plugged into an electrical outlet. These coils can be used for the inplace inspection of shaftlike parts in railroad shops, aircraft maintenance shops, and shops for automobile, truck, and tractor repair. Transverse cracks in spindles and shafts are easily detected with such coils. Most coils used for magnetizing are short, especially those wound on fixed frames. The relationship of the length of the part being inspected to the width of the coil must be considered. For a simple part, the effective overall distance that can

be inspected is 150 to 230 mm (6 to 9 in.) on either side of the coil. Thus, a part 305 to 460 mm (12 to 18 in.) long can be inspected using a normal coil approximately 25 mm (1 in.) thick. In testing longer parts, either the part must be moved at regular intervals through the coil or the coil must be moved along the part. The ease with which a part can be longitudinally magnetized in a coil is significantly related to the length-to-diameter (L/D) ratio of the part. This is due to the demagnetizing effect of the magnetic poles set up at the ends of the part. This demagnetizing effect is considerable for L/D ratios of less than 10 to 1 and is very significant for ratios of less than 3 to 1. Where the L/D ratio is extremely unfavorable, pole pieces of similar cross-sectional area can be introduced to increase the length of the part and thus improve the L/D ratio. The magnetization of rings, disks, and other parts with low L/D ratios is discussed and illustrated in the section "Induced Current" in this article. The number of ampere-turns required to produce sufficient magnetizing force to magnetize a part adequately for inspection is given by:

NI = 45,000 (L/D)

(Eq 1)

where N is the number of turns in the coil, I is the current in amperes, and L/D is the length-to-diameter ratio of the part. When the part is magnetized at this level by placing it on the bottom of the round magnetizing coil, adjacent to the coil winding, the flux density will be about 110 lines/mm2 (70,000 lines/in.2). Experimental work has shown that a flux density of 110 lines/mm2 is more than satisfactory for most applications of coil magnetization and that 54 lines/mm2 (35,000 lines/in.2) is acceptable for all but the most critical applications. When it is desirable to magnetize the part by centering it in the coil, Eq 1 becomes:

(Eq 2)

where r is the radius of the coil in inches and eff = (6L/D) - 5. Equation 2 is applicable to parts that are centered in the coil (coincident with the coil axis) and that have cross sections constituting a low fill factor, that is, with a cross-sectional area less than 10% of the area encircled by the coil. When using a coil for magnetizing a bar-like part, strong polarity at the ends of the part could mask transverse defects. An advantageous field in this area is assured on full wave, three phase, direct current units by special circuitry known as "quick" or "fast" break. A "controlled" break feature on alternating current, half wave, and on single-phase full wave direct current units provides a similar advantageous field. Central Conductors For many tubular or ring-shaped parts, it is advantageous to use a separate conductor to carry the magnetizing current rather than the part itself. Such a conductor, commonly referred to as a central conductor, is threaded through the inside of the part (Fig. 7) and is a convenient means of circularly magnetizing a part without the need for making direct contact to the part itself. Central conductors are made of solid and tubular nonmagnetic and ferromagnetic materials that are good conductors of electricity.

Fig. 7 Use of central conductors for the circular magnetization of long, hollow cylindrical parts (a) and short, hollow cylindrical or ringlike parts (b) for the detection of discontinuities on inside and outside surfaces

The basic rules regarding magnetic fields around a circular conductor carrying direct current are as follows: • • •

The magnetic field outside a conductor of uniform cross section is uniform along the length of the conductor The magnetic field is 90° to the path of the current through the conductor The flux density outside the conductor varies inversely with the radial distance from the center of the conductor

Solid Nonmagnetic Conductor Carrying Direct Current. The distribution of the magnetic field inside a

nonmagnetic conductor, such as a copper bar, when carrying direct current is different from the distribution external to the bar. At any point inside the bar, the flux density is the result of only that portion of the current that is flowing in the metal between the point and the center of the bar. Therefore, the flux density increases linearly, from zero at the center of the bar to a maximum value at the surface. Outside the bar, the flux density decreases along a curve, as shown in Fig. 8(a). In calculating flux densities outside the bar, the current can be considered to be concentrated at the center of the bar. If the radius of the bar is R and the flux density, B, at the surface of the bar is equal to the magnetizing force, H, then the flux density at a distance 2R from the center of the bar will be H/2; at 3R, H/3; and so on.

Fig. 8 Flux density in and around solid conductors of the same diameter. (a) Nonmagnetic conductor ( = 1.0) carrying direct current. (b) Ferromagnetic conductor ( > 1.0) carrying direct current. (c) Ferromagnetic conductor ( > 1.0) carrying alternating current. See text for discussion.

Solid Ferromagnetic Conductor Carrying Direct Current. If the conductor carrying direct current is a solid bar

of steel or other ferromagnetic material, the same distribution of magnetic field exists as in a similar nonmagnetic conductor, but the flux density is much greater. Figure 8(b) shows a conductor of the same diameter as that shown in Fig. 8(a). The flux density at the center is zero, but at the surface it is H, where is the material permeability of the magnetic material. (Permeability is the ease with which a material accepts magnetism.) The actual flux density, therefore, may be many times that in a nonmagnetic bar. Just outside the surface, however, the flux density drops to exactly the same value as that for the nonmagnetic conductor, and the decrease in flux density with increasing distance follows the same curve. Solid Ferromagnetic Conductor Carrying Alternating Current. The distribution of the magnetic field in a solid

ferromagnetic conductor carrying alternating current is shown in Fig. 8(c). Outside the conductor, the flux density decreases along the same curve as if direct current produced the magnetizing force; however, while the alternating current is flowing, the field is constantly varying in strength and direction. Inside the conductor, the flux density is zero at the center and increases toward the outside surface--slowly at first, then accelerating to a high maximum at the surface. The flux density at the surface is proportional to the permeability of the conductor material. Central Conductor Enclosed Within Hollow Ferromagnetic Cylinder. When a central conductor is used to

magnetize a hollow cylindrical part made of a ferromagnetic material, the flux density is maximum at the inside surface of the part (Fig. 9). The flux density produced by the current in the central conductor is maximum at the surface of the conductor (H, in Fig. 9) and then decreases along the same curve outside the conductor, as shown in Fig. 8, through the space between the conductor and the inside surface of the part. At this surface, however, the flux density is immediately increased by the permeability factor, , of the material of the part and then decreases to the outer surface. Here the flux density again drops to the same decreasing curve it was following inside the part.

Fig. 9 Flux density in and around a hollow cylinder made of magnetic material with direct current flowing through a nonmagnetic central conductor

This method, then, produces maximum flux density at the inside surface and therefore gives strong indications of discontinuities on that surface. Sometimes these indications may even appear on the outside surface of the part. The flux density in the wall of the cylindrical part is the same whether the central conductor is of magnetic or nonmagnetic material, because it is the field external to the conductor that constitutes the magnetizing force for the part. If the axis of a central conductor is placed along the axis of a hollow cylindrical part, the magnetic field in the part will be concentric with its cylindrical wall. However, if the central conductor is placed near one point on the inside circumference of the part, the flux density of the field in the cylindrical wall will be much stronger at this point and will be weaker at the diametrically opposite point. In small hollow cylinders, it is desirable that the conductor be centrally placed so that a uniform field for the detection of discontinuities will exist at all points on the cylindrical surface. In larger-diameter tubes, rings, or pressure vessels, however, the current necessary in the centrally placed conductor to produce fields of adequate strength for proper inspection over the entire circumference becomes excessively large. An offset central conductor should then be used (Fig. 10). When the conductor passing through the inside of the part is placed against an inside wall of the part, the current levels given in the section "Magnitude of Applied Current" in this article apply except that the diameter will be considered the sum of the diameter of the central conductor and twice the wall thickness. The distance along the part circumference (interior or exterior) that is effectively magnetized will be taken as four times the diameter of the central conductor, as illustrated in Fig. 10. The entire circumference will be inspected by rotating the part on the conductor, allowing for approximately a 10% magnetic field overlap.

Fig. 10 Schematic showing that the effective region of inspection when using an offset central conductor is equal to four times the diameter of the conductor

The diameter of a central conductor is not related to the inside diameter or the wall thickness of the cylindrical part. Conductor size is usually based on its current-carrying capacity and ease of handling. In some applications, conductors larger than that required for current-carrying capacity can be used to facilitate centralizing the conductor within the part. Residual magnetization is usually employed whenever practicable because the background is minimized and contrast is therefore enhanced. Also, residual magnetization is faster and less critical than continuous magnetization. The central-conductor type of inspection is sometimes required on components having parallel multiple openings, such as engine blocks. The cylinders can be processed with a single central conductor in the normal manner. However, a multiple central-conductor fixture can be designed that enables the operator to process two or more adjacent cylinders at one time with the same degree of sensitivity as if processed individually. In fact, in the areas between the central conductors, the circular fields reinforce one another to enhance sensitivity. Direct-Contact Method For small parts having no openings through the interior, circular magnetic fields are produced by direct contact to the part. This is done by clamping the parts between contact heads (head shot), generally on a bench unit (Fig. 11) that incorporates the source of the current. A similar unit can be used to supply the magnetizing current to a central conductor (Fig. 7).

Fig. 11 Bench unit for the circular magnetization of workpieces that are clamped between contact heads (direct-contact, head-shot method). The coil on the unit can be used for longitudinal magnetization.

The contact heads must be constructed so that the surfaces of the part are not damaged--either physically by pressure or structurally by heat from arcing or from high resistance at the points of contact. Heat can be especially damaging to hardened surfaces such as bearing races. For the complete inspection of a complex part, it may be necessary to attach clamps at several points on the part or to wrap cables around the part to orient fields in the proper directions at all points on the surface. This often necessitates several magnetizations. Multiple magnetizations can be minimized by using the overall magnetization method, multidirectional magnetization, or induced-current magnetization. Prod Contacts For the inspection of large and massive parts too bulky to be put into a unit having clamping contact heads, magnetization is often done by using prod contacts (Fig. 12) to pass the current directly through the part or through a local portion of it. Such local contacts do not always produce true circular fields, but they are very convenient and practical for many purposes. Prod contacts are often used in the magnetic particle inspection of large castings and weldments.

Fig. 12 Single and double prod contacts. Discontinuities are detected by the magnetic field generated between the prods.

Advantages. Prod contacts are widely used and have many advantages. Easy portability makes them convenient to use

for the field inspection of large tanks and welded structures. Sensitivity to defects lying wholly below the surface is greater with this method of magnetization than with any other, especially when half-wave current is used in conjunction with dry powder and the continuous method of magnetization. Limitations. The use of prod contacts involves some disadvantages:

•

Suitable magnetic fields exist only between and near the prod contact points. These points are seldom

• •

more than 305 mm (12 in.) apart and usually much less; therefore, it is sometimes necessary to relocate the prods so that the entire surface of a part can be inspected Interference of the external field that exists between the prods sometimes makes observation of pertinent indications difficult; the strength of the current that can be used is limited by this effect Great care must be taken to avoid burning of the part under the contact points. Burning may be caused by dirty contacts, insufficient contact pressure, or excessive currents. The likelihood of such damage is particularly great on steel with a carbon content of 0.3 to 0.4% or more. The heat under the contact points can produce local spots of very hard material that can interfere with later operations, such as machining. Actual cracks are sometimes produced by this heating effect. Contact heating is less likely to be damaging to low-carbon steel such as that used for structural purposes

Induced Current Induced current provides a convenient method of generating circumferential magnetizing current in ring-shaped parts without making electrical contact. This is accomplished by properly orienting the ring within a magnetizing coil such that it links or encloses lines of magnetic flux (flux linkage), as shown in Fig. 13(a). As the level of magnetic flux changes (increases or decreases), a current flows around the ring in a direction opposing the change in flux level. The magnitude of this current depends on the total flux linkages, rate of flux linkage changes, and the electrical impedance associated with the current path within the ring. Increasing the flux linkages and the rate of change increases the magnitude of current induced in the ring. The circular field associated with this current takes the form of a toroidal magnetic field that encompasses all surface areas on the ring and that is conducive to the disclosure of circumferential types of discontinuities. This is shown schematically in Fig. 13(b). To enhance the total flux linkages, laminated soft iron pole pieces are usually inserted through the hole in the part as shown in Fig. 13(a).

Fig. 13 Induced-current method of magnetizing a ring-shaped part. (a) Ring being magnetized by induced current. Current direction corresponds to decreasing magnetizing current. (b) Resulting induced current and toroidal magnetic field in a ring

Direct Versus Alternating Current. The choice of magnetizing current for the induced-current method depends on the magnetic properties of the part to be inspected. In cases in which the residual method is applicable, such as for most bearing races or similar parts having high magnetic retentivity, direct current is used for magnetizing. The rapid interruption of this current, by quick-break circuitry, results in a rapid collapse of the magnetic flux and the generation of a high-amperage, circumferentially directed single pulse of current in the part. Therefore, the part is residually magnetized with a toroidal field, and the subsequent application of magnetic particles will produce indications of circumferentially oriented discontinuities.

Passing an alternating current through a conductor will set up a fluctuating magnetic field as the level of magnetic flux rapidly changes from a maximum value in one direction to an equal value in the opposite direction. This is similar to the current that would flow in a single-shorted-turn secondary of a transformer. The alternating induced current, in conjunction with the continuous method, renders the method applicable for processing magnetically soft, or less retentive, parts. Applications. The induced-current method, in addition to eliminating the possibility of damaging the part, is capable of

magnetizing in one operation parts that would otherwise require more than one head shot. Two examples of this type of

part are illustrated in Fig. 14 and 15. These parts cannot be completely processed by one head shot to disclose circumferential defects, because regions at the contact point are not properly magnetized. Therefore, a two-step inspection process would be required for full coverage, with the part rotated approximately 90° prior to the second step. On the other hand, the induced-current method provides full coverage in one processing step. The disk-shaped part shown in Fig. 15 presents an additional problem when the contact method is employed to disclose circumferential defects near the rim. Even when a two-step process is employed, as with the ring shown in Fig. 14, the primary current path through the disk may not develop a circular field of ample magnitude in the rim area. The induced current can be selectively concentrated in the rim area by proper pole piece selection to provide full coverage (rim area) in a single processing step. The pole pieces shown in Fig. 15(b) are hollow and cylindrical, with one on each side of the disk. These pole pieces direct the magnetic flux through the disk such that the rim is the only portion constituting a totally enclosing current path.

Fig. 14 Current and magnetic-field distribution in a ring being magnetized with a head shot. Because the regions at the contact points are not magnetized, two operations are required for full coverage. With the induced-current method, parts of this shape can be completely magnetized in one operation.

Fig. 15 Current paths in a rimmed disk-shaped part that has been magnetized by (a) head-shot magnetization and (b) induced-current magnetization

Pole pieces used in conjunction with this method are preferably constructed of laminated ferromagnetic material to minimize the flow of eddy currents within the pole pieces, which detract from the induced (eddy) current developed within the part being processed. Pole pieces can also be made of rods, wire-filled nonconductive tubes, or thick-wall pipe saw cut to break up the eddy-current path. In some cases, even a solid shaft protruding from one side of a gear or disk can be used as one of the pole pieces.

Inspection of Steel Balls. Direct contact is not permitted during the inspection of hardened, finished steel balls for

heat treating or grinding cracks, because of the highly polished surface finish. The discontinuities may be oriented in any direction, and 100% inspection of the balls is required. The induced-current method can provide the required inspection without damaging the surface finish. The L/D ratio of 1:1 for spheres is unfavorable for magnetization with a coil; therefore, laminated pole pieces are used on each side of the balls to provide a more favorable configuration for magnetizing. Because of the highly retentive nature of the material, residual magnetization with direct current and quickbreak circuitry is used for magnetizing the balls. The smallness of the heat-treating or grinding cracks and the high surface finish dictate that the inspection medium be a highly oil-suspendible material. Balls are inspected along the x-, y-, and z-axes in three separate operations. The operation for each axis consists of: • • •

An induced-current shot Bathing the ball with the wet-particle solution Inspection while rotating the ball 360°

Rotation and reorientation can be accomplished in a simple manually operated fixture, or the entire operation can be automated. Magnetic Particle Inspection Revised by Art Lindgren, Magnaflux Corporation

Permeability of Magnetic Materials The term permeability is used to refer to the ease with which a magnetic field or flux can be set up in a magnetic circuit. For a given material, it is not a constant value but a ratio. At any given value of magnetizing force, permeability, , is B/H, the ratio of flux density, B, to magnetizing force, H. Several permeabilities have been defined, but material permeability, maximum permeability, effective (apparent) permeability, and initial permeability are used with magnetic particle testing. Material permeability is of interest in magnetic particle inspection with circular magnetization. Material permeability

is the ratio of the flux density, B, to the magnetizing force, H, where the flux density and magnetizing force are measured when the flux path is entirely within the material. The magnetizing force and the flux density produced by that force are measured point by point for the entire magnetization curve with a fluxmeter and a prepared specimen of material. Maximum Permeability. For magnetic particle inspection, the level of magnetization is generally chosen to be just

below the knee of a normal magnetization curve for the specific material; the maximum material permeability occurs near this point. For most engineering steels, the maximum material permeability ranges from 0.06 to 0.25 T/A · m-1 (500 to 2000 G/Oe) or more. The 500 value is for 400-series stainless steels. Specific permeability values for the various engineering materials are not readily available, but even if they were, they could be misleading. To a large extent, the numerous rules of thumb consider the variations in permeability, so that knowledge of permeability values is not a prerequisite for magnetic particle inspection. Effective (apparent) permeability is the ratio of the flux density in the part to the magnetizing force, when the

magnetizing force is measured at the same point in the absence of the part. Effective permeability is not solely a property of the material, but is largely governed by the shape of the part and is of prime importance for longitudinal magnetization. Initial permeability is exhibited when both the flux density, B, and the magnetizing force, H, approach zero (Fig.

16a). With increasing magnetizing force, the magnetic field in the part increases along the virgin curve of the hysteresis loop.

Fig. 16 Representative magnetization (hysteresis) curve for a ferromagnetic material. Dashed line in each of the charts is the virgin hysteresis curve. See text for discussion.

Magnetic Particle Inspection Revised by Art Lindgren, Magnaflux Corporation

Magnetic Hysteresis Some ferromagnetic materials, when magnetized by being introduced into an external field, do not return to a completely unmagnetized state when removed from that field. In fact, these materials must be subjected to a reversed field of a certain strength to demagnetize them (discounting heating the material to a characteristic temperature, called the Curie point, above which the ferromagnetic ordering of atomic moments is thermally destroyed, or mechanically working the material to reduce the magnetization). If an external field that can be varied in a controlled manner is applied to a completely demagnetized (virgin) specimen and if instrumentation for measuring the magnetic induction within the specimen is at hand, the magnetization curve of the material can be determined. A representative magnetization (hysteresis) curve for a ferromagnetic material is shown in Fig. 16. As shown in Fig. 16(a), starting at the origin (O) with the specimen in the unmagnetized condition and increasing the magnetizing force in small increments, the flux in the material increases quite rapidly at first, then more slowly until it reaches a point beyond which any increase in the magnetizing force does not increase the flux density. This is shown by the dashed (virgin) curve Oa. In this condition, the specimen is said to be magnetically saturated. When the magnetizing force is gradually reduced to zero, curve ab results (Fig. 16b). The amount of magnetism that the steel retains at point b is called residual magnetism, Br. When the magnetizing current is reversed and gradually increased in value, the flux continues to diminish. The flux does not become zero until point c is reached, at which time the magnetizing force is represented by Oc (Fig. 16c), which

graphically designates the coercive force, Hc, in the material. Ferromagnetic materials retain a certain amount of residual magnetism after being subjected to a magnetizing force. When the magnetic domains of a ferromagnetic material have been oriented by a magnetizing force, some domains remain so oriented until an additional force in the opposite direction causes them to return to their original random orientation. This force is commonly referred to as coercive force. As the reversed field is increased beyond c, point d is reached (Fig. 16d). At this point, the specimen is again magnetically saturated. The magnetizing force is now decreased to zero, and the de line is formed and retains reversed-polarity residual magnetism, Br, in the specimen. Further increasing the magnetizing force in the original direction completes the curve efa. The cycle is now complete, and the area within the loop abcdefa is called the hysteresis curve. The definite lag throughout the cycle between the magnetization force and the flux is called hysteresis. If the hysteresis loop is slender (Fig. 16e), the indication usually means that the material has low retentivity (low residual field) and is easy to magnetize (has low reluctance). A wide loop (Fig. 16f) indicates that the material has high reluctance and is difficult to magnetize. Magnetic Particle Inspection Revised by Art Lindgren, Magnaflux Corporation

Magnetic Particles and Suspending Liquids Magnetic particles are classified according to the medium used to carry the particles to the part. The medium can be air (dry-particle method) or a liquid (wet-particle method). Magnetic particles can be made of any low-retentivity ferromagnetic material that is finely subdivided. The characteristics of this material, including magnetic properties, size, shape, density, mobility, and degree of visibility and contrast, vary over wide ranges for different applications. Magnetic Properties. The particles used for magnetic particle inspection should have high magnetic permeability so

that they can be readily magnetized by the low-level leakage fields that occur around discontinuities and can be drawn by these fields to the discontinuities themselves to form readable indications. The fields at very fine discontinuities are sometimes extremely weak. Low coercive force and low retentivity are desirable for magnetic particles. If high in coercive force, wet particles become strongly magnetized and form an objectionable background. In addition, the particles will adhere to any steel in the tank or piping of the unit and cause heavy settling-out losses. Highly retentive wet particles tend to clump together quickly in large aggregates on the test surface. Excessively large clumps of particles have low mobility, and indications are distorted or obscured by the heavy, coarse-grain backgrounds. Dry particles having coercive force and high retentivity would become magnetized during manufacture or in first use and would therefore become small, strong permanent magnets. Once magnetized, their control by the weak fields at discontinuities would be subdued by their tendency to stick magnetically to the test surface wherever they first touch. This would reduce the mobility of the powder and would form a high level of background, which would reduce contrast and make indications difficult to see. Effect of Particle Size. Large, heavy particles are not likely to be arrested and held by weak fields when such particles

are moving over a part surface, but fine particles will be held by very weak fields. However, extremely fine particles may also adhere to surface areas where there are no discontinuities (especially if the surface is rough) and form confusing backgrounds. Coarse dry particles fall too fast and are likely to bounce off the part surface without being attracted by the weak leakage fields at imperfections. Finer particles can adhere to fingerprints, rough surfaces, and soiled or damp areas, thus obscuring indications. Magnetic particles for the wet method are applied as a suspension in some liquid medium, and particles much smaller than those for the dry method can be used. When such a suspension is applied over a surface, the liquid drains away, and the film remaining on the surface becomes thinner. Coarse particles would quickly become stranded and immobilized. The stranding of finer particles as a result of the draining away of the liquid occurs much later, giving these particles mobility for a sufficient period of time to be attracted by leakage fields and to accumulate and thus form true indications.

Effect of Particle Shape. Long, slender particles develop stronger polarity than globular particles. Because of the

attraction exhibited by opposite poles, these tiny, slender particles, which have pronounced north and south poles, arrange themselves into strings more readily than globular particles. The ability of dry particles to flow freely and to form uniformly dispersed clouds of powder that will spread evenly over a surface is a necessary characteristic for rapid and effective dry-powder testing. Elongated particles tend to mat in the container and to be ejected in uneven clumps, but globular particles flow freely and smoothly under similar conditions. The greatest sensitivity for the formation of strong indications is provided by a blend of elongated and globular shapes. Wet particles, because they are suspended in a liquid, move more slowly than do dry particles to accumulate in leakage fields. Although the wet particles themselves may be of any shape, they are fine and tend to agglomerate, or clump, into unfavorable shapes. These unfavorable shapes will line up into magnetically held elongated aggregates even when the leakage field is weak. This effect contributes to the relatively high sensitivity of fine wet particles. Visibility and contrast are promoted by choosing particles with colors that are easy to see against the color of the

surface of the part being inspected. The natural color of the metallic powders used in the dry method is silver-gray, but pigments are used to color the particles. The colors of particles for the wet method are limited to the black and red of the iron oxides commonly used as the base for wet particles. For increased visibility, particles are coated with fluorescent pigment by the manufacturer. The search for indications is conducted in total or partial darkness, using ultraviolet light to activate the fluorescent dyes. Fluorescent magnetic particles are available for both the wet and dry methods, but fluorescent particles are more commonly used with the wet method. Types of Magnetic Particles. The two primary types of particles for magnetic particle inspection are dry particles

and wet particles, and each type is available in various colors and as fluorescent particles. Particle selection is principally influenced by: • • •

Location of the discontinuity, that is, on the surface or beneath the surface Size of the discontinuity, if on the surface Which type (wet or dry particles) is easier to apply

Dry particles, when used with direct current for magnetization, are superior for detecting discontinuities lying wholly below the surface. The use of alternating current with dry particles is excellent for revealing surface cracks that are not exceedingly fine, but is of little value for discontinuities even slightly beneath the surface (Fig. 17).

Fig. 17 Comparison of sensitivity of alternating current, direct current, direct current with surge, and half-wave current for locating defects at various distances below a surface by means of the dry-particle continuous magnetizing method. See text for description of test conditions.

Wet particles are better than dry particles for detecting very fine surface discontinuities regardless of which form of magnetizing current is used. Wet particles are often used with direct current to detect discontinuities that lie just beneath the surface. The surface of a part can easily be covered with a wet bath because the bath flows over and around surface contours. This is not easily accomplished with dry powders. Colored particles are used to obtain maximum contrast with the surface of the part being inspected. Black stands out against most light-colored surfaces (Fig. 18), and gray against dark-colored surfaces. Red particles are more visible than black or gray particles against silvery and polished surfaces. Fluorescent particles, viewed under ultraviolet light, provide the highest contrast and visibility.

Fig. 18 Black magnetic-powder indications of a cold shut in a casting when seen under black light. Courtesy of Magnaflux Corporation

Dry particles are available with yellow, red, black, and gray pigmented coloring and with fluorescent coatings. The magnetic properties and particle sizes are similar in all colors, making them equally efficient. Dry particles are most sensitive for use on very rough surfaces and for detecting flaws beneath the surface. They are ordinarily used with portable equipment. The reclamation and reuse of dry particles is not recommended.

Air is used to carry the particles to the surface of the part, and care must be taken to apply the particles correctly. Dry powders should be applied in such a way that they reach the magnetized surface in a uniform cloud with a minimum of motion. When this is done, the particles come under the influence of the leakage fields while suspended in air and have three-dimensional mobility. This condition can be best achieved when the magnetized surface is vertical or overhead. When particles are applied to a horizontal or sloping surface, they settle directly to the surface and do not have the same degree of mobility. Dry powders can be applied with small rubber spray bulbs or specially designed mechanical powder blowers. The air stream of such a blower is of low velocity so that a cloud of powder is applied to the test area. Mechanical blowers can also deliver a light stream of air for the gentle removal of excess powder. Powder that is forcibly applied is not free to be

attracted by leakage fields. Neither rolling a magnetized part in powder nor pouring the powder on the part is recommended. Wet particles are best suited for the detection of fine discontinuities such as fatigue cracks. Wet particles are commonly

used in stationary equipment where the bath can remain in use until contaminated. They are also used in field operations with portable equipment, but care must be taken to agitate the bath constantly. Wet particles are available in red and black colors or as fluorescent particles that fluoresce a blue-green or a bright yellow-green color. The particles are supplied in the form of a paste or other type of concentrate that is suspended in a liquid to produce the coating bath. The liquid bath may be either water or a light petroleum distillate having specific properties. Both require conditioners to maintain proper dispersion of the particles and to allow the particles the freedom of movement to form indications on the surfaces of parts. These conditioners are usually incorporated in the powder concentrates. Oil Suspending Liquid. The oil used as a suspending liquid for magnetic particles should be an odorless, well-refined light petroleum distillate of low viscosity having a low sulfur content and a high flash point. The viscosity of the oil should not exceed 3 × 10-6 m2/s (3 cSt) as tested at 40 °C (100 °F) and must not exceed 5 × 10-6 m2/s (5 cSt) when tested at the bath temperature. Above 5 × 10-6 m2/s (5 cSt), the movement of the magnetic particles in the bath is sufficiently retarded to have a definite effect in reducing buildup, and therefore the visibility, of an indication of a small discontinuity. Parts should be precleaned to remove oil and grease because oil from the surface accumulates in the bath and increases its viscosity. Water Suspending Liquid. The use of water instead of oil for magnetic particle wet-method baths reduces costs and

eliminates bath flammability. Figures 19, 20, 21, and 22 illustrate the effectiveness of water-base magnetic particle testing. Water-suspendible particle concentrates include the necessary wetting agents, dispersing agents, rust inhibitors, and antifoam agents.

Fig. 19 Spindle defects revealed with water-base magnetic particle testing. Courtesy of Circle Chemical Company

Fig. 20 Compressor vane microcracks revealed with water-base magnetic particle testing. Courtesy of Circle Chemical Company

Fig. 21 50 mm (2 in.) diam gear subjected to water-base magnetic particle testing showing cracks 0.25 mm (0.0098 in.) long by 0.1 mm (0.004 in.) wide. Courtesy of Circle Chemical Company

Fig. 22 Defects in the leading edge and fillet areas of a hydroplane propeller as revealed by water-base

magnetic particle testing. (a) Overall view. (b) Close-up of fillet area. Courtesy of Circle Chemical Company

Water baths ordinarily should not be used where the temperature is below freezing. However, ethylene glycol can be employed to protect against reasonably low temperatures. Care must be exercised, however, because a high percentage of ethylene glycol can impede particle mobility. For the inspection of large billets, a solution containing three parts water and one part anti-freeze has been successfully used. Because water is a conductor of electricity, units in which water is to be used must be designed to isolate all high-voltage circuits so as to avoid all possibility of operator shock, and the equipment must be thoroughly and positively grounded. Also, electrolysis of parts of the unit can occur if preventive measures are not taken. Units specifically designed to be used with water as a suspensoid are safe for the operator and minimize electrolytic corrosion. The strength of the bath is a major factor in determining the quality of the indications obtained. The proportion of magnetic particles in the bath must be maintained at a uniform level. If the concentration varies, the strength of the indications will also vary, and the indications may be misinterpreted. Fine indications may be missed entirely with a weak bath. Too heavy a concentration of particles gives a confusing background and excessive adherence of particles at external poles, thus interfering with clean-cut indications of extremely fine discontinuities.

The best method for ensuring optimum bath concentration for any given combination of equipment, bath application, type of part, and discontinuities sought is to test the bath using parts with known discontinuities. Bath strength can be adjusted until satisfactory indications are obtained. This bath concentration can then be adopted as standard for those conditions. The concentration of the bath can be measured with reasonable accuracy with the settling test. In this test, 100 mL (3.4 oz) of well-agitated bath is placed in a pear-shaped centrifuge tube. The volume of solid material that settles out after a predetermined interval (usually 30 min) is read on the graduated cylindrical part of the tube. Dirt in the bath will also settle and usually shows as a separate layer on top of the oxide. The layer of dirt is usually easily discernible because it is different in color from the magnetic particles. Magnetic Particle Inspection Revised by Art Lindgren, Magnaflux Corporation

Ultraviolet Light A mercury-arc lamp is a convenient source of ultraviolet light. This type of lamp emits light whose spectrum has several intensity peaks within a wide band of wave-lengths. When used for a specific purpose, emitted light is passed through a suitable filter so that only a relatively narrow band of ultraviolet wavelengths is available. For example, a band in the long-wave ultraviolet spectrum is used for fluorescent liquid penetrant or magnetic particle inspection. Fluorescence is the characteristic of an element or combination of elements to absorb the energy of light at one frequency and emit light of a different frequency. The fluorescent materials used in liquid penetrant and magnetic particle inspection are combinations of elements chosen to absorb light in the peak energy band of the mercury-arc lamp fitted with a Kopp glass filter. This peak occurs at about 365 nm (3650 Å). The ability of fluorescent materials to emit light in the greenish-yellow wavelengths of the visible spectrum depends on the intensity of ultraviolet light at the workpiece surface. In contrast to the harmful ultraviolet light of shorter wavelengths, which damages organs such as the eyes and the skin, the black light of 365 nm (3650 Å) wavelengths poses no such hazards to the operator and provides visible evidence of defects in materials, as shown in Fig. 23, 24, 25, 26, 27, and 28.

Fig. 23 Casting viewed under black light showing strong magnetic particle indication with minimal background fluorescence. Courtesy of Magnaflux Corporation

Fig. 24 Potentially dangerous cracks in a lawn mower blade revealed when fluorescent magnetic particles are exposed to black light. Courtesy of Magnaflux Corporation

Fig. 25 Fluorescent magnetic particle indications of quenched-and-drawn drill casing tubing that shows rejectable, cracked spiral seams in the tubing when exposed to black light. Courtesy of Magnaflux Corporation

Fig. 26 Black light used to reveal magnetic particle indication of a crack in the centering stud of a ball-yoketype universal-joint section of a drive-shaft assembly. Courtesy of Magnaflux Corporation

Fig. 27 Black light used to reveal magnetic particle indication of a lap that could lead to potential failure of a forged connecting rod. Courtesy of Magnaflux Corporation

Fig. 28 Black light used to reveal magnetic particle indication of a dangerous crack near a wheel stud. The crack was detected during manufacture. Courtesy of Magnaflux Corporation

Early specifications required 970 lx (90 ftc) of illumination at the workpiece surface measured with a photographic-type light meter, which responds to white light as well as ultraviolet light. Because the ultraviolet output of a mercury-arc lamp decreases with age and with hours of service, the required intensity of the desired wavelength is often absent, although the light meter indicates otherwise. Meters have been developed that measure the overall intensity of long-wave ultraviolet light only in a band between 300 to 400 nm (3000 to 4000 Å) and that are most sensitive near the peak energy band of the mercury-arc lamp used for fluorescent inspection. These meters read the intensity level in microwatts per square centimeter (μW/cm2). For aircraft-quality fluorescent inspection, the minimum intensity level of ultraviolet illumination is 1000 μW/cm2. High-intensity 125-W ultraviolet bulbs are available that provide up to 5000 W/cm2 at 380 mm (15 in.). Magnetic Particle Inspection Revised by Art Lindgren, Magnaflux Corporation

Nomenclature Used in Magnetic Particle Inspection An indication is an accumulation of magnetic particles on the surface of the part that forms during inspection. Relevant indications are the result of errors made during or after metal processing. They may or may not be

considered defects. A nonrelevant indication is one that is caused by flux leakage. This type of indication is usually weak and has no relation to a discontinuity that is considered to be a defect. Examples are magnetic writing, change in section due to part design, or a heat affected zone line in welding. False indications are those in which the particle patterns are held by gravity or surface roughness. No magnetic

attraction is involved. A discontinuity is any interruption in the normal physical configuration or composition of a part. It may not be a defect. A defect is any discontinuity that interferes with the utility or service of a part. Interpretation consists of determining the probable cause of an indication, and assigning it a discontinuity name or

label.

Evaluation involves determining whether an indication will be detrimental to the service of a part. It is a judgement

based on a well-defined accept/reject standard that may be either written or verbal. Magnetic Particle Inspection Revised by Art Lindgren, Magnaflux Corporation

Detectable Discontinuities The usefulness of magnetic particle inspection in the search for discontinuities or imperfections depends on the types of discontinuities the method is capable of finding. Of importance are the size, shape, orientation, and location of the discontinuity with respect to its ability to produce leakage fields. Surface Discontinuities. The largest and most important category of discontinuity consists of those that are exposed

to the surface. Surface cracks or discontinuities are effectively located with magnetic particles. Surface cracks are also more detrimental to the service life of a component than are subsurface discontinuities, and as a result they are more frequently the object of inspection. Magnetic particle inspection is capable of locating seams, laps, quenching and grinding cracks, and surface ruptures in castings, forgings, and weldments. The method will also detect surface fatigue cracks developed during service. Magnetizing and particle application methods may be critical in certain cases, but in most applications the requirements are relatively easily met because leakage fields are usually strong and highly localized. For the successful detection of a discontinuity, there must be a field of sufficient strength oriented in a generally favorable direction to produce strong leakage fields. For maximum detectability, the field set up in the part should be at right angles to the length of a suspected discontinuity (Fig. 3 and 4). This is especially true if the discontinuity is small and fine. The characteristics of a discontinuity that enhance its detection are: • • • •

Its depth is at right angles to the surface Its width at the surface is small, so that the air gap it creates is small Its length at the surface is large with respect to its width It is comparatively deep in proportion to the width of its surface opening

Many incipient fatigue cracks and fine grinding cracks are less than 0.025 mm (0.001 in.) deep and have surface openings of perhaps one-tenth that or less. Such cracks are readily located using wet-method magnetic particle inspection. The depth of the crack has a pronounced effect on its detectability; the deeper the crack, the stronger the indication for a given level of magnetization. This is because the stronger leakage flux causes greater distortion of the field in the part. However, this effect is not particularly noticeable beyond perhaps 6.4 mm ( in.) in depth. If the crack is not close-lipped but wide open at the surface, the reluctance (opposition to the establishment of magnetic flux in a magnetic circuit) of the resulting longer air gap reduces the strength of the leakage field. This, combined with the inability of the particles to bridge the gap, usually results in a weaker indication. Detectability generally involves a relationship between surface opening and depth. A surface scratch, which may be as wide at the surface as it is deep, usually does not produce a magnetic particle pattern, although it may do so at high levels of magnetization. Because of many variables, it is not possible to establish any exact values for this relationship, but in general a surface discontinuity whose depth is at least five times its opening at the surface will be detectable. There are also limitations at the other extreme. For example, if the faces of a crack are tightly forced together by compressive stresses, the almost complete absence of an air gap may produce so little leakage field that no particle indication is formed. Shallow cracks produced in grinding or heat treating and subsequently subjected to strong compression by thermal or other stresses usually produce no magnetic particle indications. Sometimes, with careful, maximum-sensitivity techniques, faint indications of such cracks can be produced.

One other type of discontinuity that sometimes approaches the lower limit of detectability is a forging or rolling lap that, although open to the surface, emerges at an acute angle. In this case, the leakage field produced may be quite weak because of the small angle of emergence and the resultant relatively high reluctance of the actual air gap; consequently, very little leakage flux takes the path out through the surface lip of the lap to cross this high reluctance gap. When laps are being sought (usually when newly forged parts are being inspected), high-sensitivity, such as combining dc magnetizing with the wet fluorescent method, is desirable. Figure 29 shows two indications of forging laps in a 1045 steel crane hook.

Fig. 29 1045 steel crane hook showing indications of forging laps of the type revealed by magnetic particle inspection. Dimensions given in inches

A seam that was found during the magnetic particle inspection of a forged crane hook is shown in Fig. 30. The seam was present in the material before the hook was forged. The cold shut shown in Fig. 31 was found in the flange of a cast drum after machining. A faint indication was noted in the rough casting, but the size of the cold shut was not known until after machining of the drum.

Fig. 30 Magnetic particle indications of a seam (at arrow) in the shank of a forged crane hook

Fig. 31 Cold shut (at arrow) in the flange of a machined cast drum. Magnetic particle inspection revealed faint indications of the cold shut in the rough casting.

The magnetic particle inspection of a 460 mm (18 in.) diam internally splined coupling revealed the indications shown in Fig. 32, one of which was along the fusion zone of a repair weld. Routine magnetic particle inspection of a 1.2 m (4 ft) diam weldment revealed cracks in the weld between the rim and web, as shown in Fig. 33.

Fig. 32 Magnetic particle indications of cracks in a large cast splined coupling. Indication (at arrow) in photo at right is along the fusion zone of a repair weld.

Fig. 33 Indications of cracks (at arrows) in the weld between the web and rim of a 1.2 m (4 ft) diam weldment

Internal Discontinuities. The magnetic particle method is capable of indicating the presence of many discontinuities that do not break the surface. Although radiography and ultrasonic methods are inherently better for locating internal

discontinuities, sometimes the shape of the part, the location of the discontinuity, or the cost or availability of the equipment needed makes the magnetic particle method more suitable. The internal discontinuities that can be detected by magnetic particle inspection can be divided into two groups: • •

Subsurface discontinuities (those lying just beneath the surface of the part) Deep-lying discontinuities

Subsurface discontinuities comprise those voids or nonmetallic inclusions that lie just beneath the surface.

Nonmetallic inclusions are present in all steel products to some degree. They occur as scattered individual inclusions, or they may be aligned in long stringers. These discontinuities are usually very small and cannot be detected unless they lie very close to the surface, because they produce highly localized but rather weak fields. Deep-lying discontinuities in weldments may be caused by inadequate joint penetration, subsurface incomplete

fusion, or cracks in weld beads beneath the last weld bead applied. In castings, they result from internal shrinkage cavities, slag inclusions, or gas pockets. The depth to which magnetic particle testing can reach in locating internal discontinuities cannot be established in millimeters, because the size and shape of the discontinuity itself in relation to the size of the part in which it occurs is a controlling factor. Therefore, the deeper the discontinuity lies within a section, the larger it must be to be detected by magnetic particle inspection. In considering the detectability of a discontinuity lying below the surface, of primary concern are the projected area presented as an obstruction to the lines of force and the sharpness of the distortion of the field produced. It is helpful to think of the magnetic field as flowing through the specimen like a stream of water. A coin, on edge below the surface and at right angles to the surface and flow, would cause a sharp disturbance in the movement of the water, but a straight round stick placed at a similar depth and parallel to the direction of flow would have very little effect. A ball or marble of the same size as the coin would present the same projected area, but would be much less likely to be detected than the coin because the flow lines would be streamlines around the sphere and the disturbance created in the field would be much less sharp. The orientation of the discontinuity is another factor in detection. The coin-shaped obstruction discussed above was considered to be 90° to the direction of the flux. If the same discontinuity were inclined, either vertically or horizontally, at an angle of only 60 or 70°, there would be a noticeable difference in the amount of leakage field and therefore in the strength of the indication. This difference would result not only because the projected area would be reduced but also because of a streamlining effect, as with the sphere. The strength and direction of the magnetic field are also important in the detection of deep-lying discontinuities. Direct current yokes are effective if the discontinuity is close to the surface. Prod magnetization using direct current or half-wave current is more effective for discontinuities deeper in the part. Magnetic Particle Inspection Revised by Art Lindgren, Magnaflux Corporation

Nonrelevant Indications Nonrelevant indications are true patterns caused by leakage fields that do not result from the presence of flaws. Nonrelevant indications have several possible causes and therefore require evaluation, but they should not be interpreted as flaws. Sources of Nonrelevant Indications Particle patterns that yield nonrelevant indications can be the result of design, fabrication, or other causes and do not imply a condition that reduces the strength or utility of the part. Because nonrelevant indications are true particle buildups, they are difficult to distinguish from buildups caused by flaws. Therefore, the investigator must be aware of design and fabrication conditions that would contribute to or cause nonrelevant indications.

Particle Adherence Due to Excessive Magnetizing Force. One type of nonrelevant indication is that caused by

particle adherence at leakage fields around sharp corners, ridges, or other surface irregularities when magnetized longitudinally with too strong a magnetizing force. The use of too strong a current with circular magnetization can produce indications of the flux lines of the external field. Both of the above phenomena (excessive magnetizing force or excessive current) are clearly recognized by experienced operators and can be eliminated by a reduction in the applied magnetizing force. Mill Scale. Tightly adhering mill scale will cause particle buildup, not only because of mechanical adherence but also

because of the difference in magnetic permeability between the steel and the scale. In most cases, this can be detected visually, and additional cleaning followed by retesting will confirm the absence of a true discontinuity. Configurations that result in a restriction of the magnetic field are a cause of nonrelevant indications. Typical

restrictive configurations are internal notches such as splines, threads, grooves for indexing, or keyways. Abrupt changes in magnetic properties, such as those between weld metal and base metal or between dissimilar

base metals, result in nonrelevant indications. Depending on the degree of change in the magnetic property, the particle pattern may consist of loosely adhering particles or may be strong and well defined. Again, it is necessary for the investigator to be aware of such conditions. Magnetized writing is another form of nonrelevant indication. Magnetic writing is usually associated with parts

displaying good residual characteristics in the magnetized state. If such a part is contacted with a sharp edge of another (preferably magnetically soft) part, the residual field is locally reoriented, giving rise to a leakage field and consequently a magnetic particle indication. For example, the point of a common nail can be used to write on a part susceptible to magnetic writing. Magnetic writing is not always easy to interpret, because the particles are loosely held and are fuzzy or intermittent in appearance. If magnetic writing is suspected, it is only necessary to demagnetize the parts and retest. If the indication was magnetic writing, it will not reappear. Additional Sources. Some other conditions that cause nonrelevant indications are brazed joints, voids in fitted parts,

and large grains. Distinguishing Relevant From Nonrelevant Indications There are several techniques for differentiating between relevant and nonrelevant indications: • • • • •

Where mill scale or surface roughness is the probable cause, close visual inspection of the surface in the area of the discontinuity and use of magnification up to ten diameters Study of a sketch or drawing of the part being tested to assist in locating welds, changes in section, or shape constrictions Demagnetization and retesting Careful analysis of the particle pattern. The particle pattern typical of nonrelevant indications is usually wide, loose, and lightly adhering and is easily removable even during continuous magnetization Use of another method of nondestructive inspection, such as ultrasonic testing or radiography, to verify the presence of a subsurface defect

The following two examples illustrate how nonrelevant indications are used in nondestructive testing to verify product quality. Example 1: Nonrelevant Indications in Electric Motor Rotors. An instance where nonrelevant indications are used to advantage is in the inspection of rotors for squirrel cage electric motors. These rotors are usually fabricated from laminations made of magnetic material. The conductor-bar holes are aligned during assembly of the laminations. The end rings and conductor bars are cast from an aluminum alloy in a single operation. The integrity of the internal cast aluminum alloy conductor bars must be checked to ensure that each is capable of carrying the required electrical current; voids and internal porosity would impair their electrical properties.

These rotors are tested by clamping them between the heads of a horizontal unit and processing them by the continuous method using wet magnetic particles. The end rings distribute the magnetizing current through each of the conductor bars, which constitute parallel paths. All conductor bars that are sound and continuous produce broad, pronounced, subsurfacetype indications on the outside surface of the rotor. The absence of such indications is evidence that the conductor bar is discontinuous or that its current-carrying capacity is greatly impaired. This is a direct departure from the customary inspection logic in that negative results, or the absence of indications, are indicative of defects. Example 2: Nonrelevant Indications Present in Welding A537 Grade 2 Carbon Steel With E8018-C1 Weld Wire. Linear magnetic particle indications have frequently been observed in the heat-affected zone of A537 grade 2 (quenchedand-tempered) materials joined with E8018-C1 (2 % Ni) weld wire. However, when liquid penetrant examinations of these areas are performed, no indications are apparent. Four types of welds were used in the investigation: • • • •

Weld A: a T-joint with fillet welds on both sides (Fig. 34) Weld B: a butt weld (Fig. 35) Weld C: a pickup simulating a repair where a fit-up device had been torn off (Fig. 36) Weld D: a T-joint with fillet welds on both sides, one of which contained a longitudinal crack (Fig. 37)

The fourth sample was made in such a manner that it would contain a linear discontinuity in or near the heat-affected zone. Examination with magnetic particle and liquid penetrant inspection methods yielded the following results: •

•

•

•

Weld A: Magnetic particle inspection showed loosely held, slightly fuzzy linear indications at the toes of the fillet, with the bottom edge showing the strongest pattern due to gravity and the configuration. The liquid penetrant examination produced no relevant indications. The etched cross section revealed no discontinuities Weld B: Magnetic particle inspection showed loosely held, slightly fuzzy linear indications along both edges of the butt weld. The liquid penetrant examination produced no relevant indications. The etched cross section revealed no discontinuities Weld C: Magnetic particle inspection showed a very loosely held, fuzzy pattern over the entire pickup weld. The liquid penetrant examination produced no relevant indications. The etched cross section revealed only a minor slag inclusion, which had no bearing on the magnetic particle pattern produced Weld D: Magnetic particle inspection showed a tightly held, sharply defined linear indication along one edge of the fillet with connecting small, transverse linear indications at various locations. The liquid penetrant examination in this case produced the same pattern of indications. The etched cross section revealed a crack completely through the toe of one leg of the fillet

Fig. 34 T-joint weld of A537 grade 2 (quenched and tempered) material joined with E8018-C1 (2.5% Ni) weld

wire showing linear magnetic particle indications. Weld A, shown in cross section (a), was examined along its length using both the (b) liquid penetrant inspection method and the (c) magnetic particle inspection method. See text for discussion. Courtesy of Chicago Bridge & Iron Company

Fig. 35 Butt weld of A537 grade 2 (quenched and tempered) material joined with E8018-C1 (2.5% Ni) weld wire showing linear magnetic particle indications. Weld B, shown in cross section (a), was checked using both the (b) liquid penetrant inspection method and the (c) magnetic particle inspection method. See text for discussion. Courtesy of Chicago Bridge & Iron Company

Fig. 36 A pickup, simulating a repair of A537 grade 2 (quenched and tempered) material joined with E8018-C1 (2.5% Ni) weld wire where a fit-up device had been torn off, showing linear magnetic particle indications. Weld C, shown in cross section (a), was checked using both the (b) liquid penetrant inspection method and the (c) magnetic particle inspection method. See text for discussion. Courtesy of Chicago Bridge & Iron Company

Fig. 37 T-joint weld of A537 grade 2 (quenched and tempered) material joined with E8018-C1 (2.5% Ni) weld wire showing a longitudinal crack in one of the fillet welds (arrow in cross section) made visible by linear magnetic particle indications. Weld D, shown in cross section (a), was checked using both the (b) liquid penetrant inspection method and the (c) magnetic particle inspection method. See text for discussion. Courtesy

of Chicago Bridge & Iron Company

The difference in magnetic properties between this parent material and the weld metal creates magnetic leakage fields in the heat-affected zone or metallic interface of these materials during a magnetic particle inspection. This can result in magnetic particle indications in the heat-affected zone along the toe of the fillets, along the edges of butt welds, or over entire shallow pickup welds. The magnetic particle indications produced by metallic interface leakage fields are readily distinguished from indications caused by real discontinuities by their characteristic of being loosely held and slightly fuzzy in appearance and by their location in or at the edge of the heat-affected zone. Magnetic Particle Inspection Revised by Art Lindgren, Magnaflux Corporation

General Procedures for Magnetic Particle Inspection In magnetic particle inspection, there are many variations in procedure that critically affect the results obtained. These variations are necessary because of the many types of discontinuities that are sought and the many types of ferromagnetic materials in which these discontinuities must be detected. Establishing a set of procedures for the magnetic particle inspection of a specific part requires that the part be carefully analyzed to determine how its size and shape will affect the test results. The magnetic characteristics of the material and the size, shape, location, and direction of the expected discontinuity also affect the possible variations in the procedure. The items that must be considered in establishing a set of procedures for the magnetic particle inspection of a specific part include: • • • • • •

Type of current Type of magnetic particles Method of magnetization Direction of magnetization Magnitude of applied current Equipment

Type of Current The electric current used can be either alternating current or some form of direct current. This choice depends on whether the discontinuities are surface or subsurface and, if subsurface, on the distance below the surface. Alternating Current. The skin effect of alternating current at 50 or 60 Hz limits its use to the detection of discontinuities that are open to the surface or that are only a few thousandths of an inch below the surface. With alternating current at lower frequencies, the skin effect is less pronounced, resulting in deeper penetration of the lines of force.

The rapid reversal of the magnetic field set up by alternating current imparts mobility to dry particles. Agitation of the powder helps it move to the area of leakage fields and to form stronger indications. The strength of magnetization, which is determined by the value of the peak current at the top of the sine wave of the cycle, is 1.41 times that of the current indicated on the meter. Alternating current meters indicate more nearly the average current for the cycle than the peak value. Obtaining an equivalent magnetizing effect from straight direct current requires more power and heavier equipment. Direct current, on the other hand, magnetizes the entire cross section more or less uniformly in a longitudinal direction,

and with a straight-line gradient of strength from a maximum at the surface to zero at the center of a bar in the case of circular magnetization. This effect is demonstrated in Fig. 8(a) and 8(b).

Alternating Current Versus Direct Current. In an experiment designed to compare the effectiveness of 60-Hz

alternating current and three types of direct current, 12 holes representing artificial defects were drilled in a 127 mm (5 in.) OD by 32 mm (1

in.) ID by 22 mm (

in.) thick ring made of unhardened O1 tool steel (0.40% C). The 12 holes,

1.8 mm (0.07 in.) in diameter and spaced 19 mm ( in.) apart, were drilled through the ring parallel to the cylindrical surface at increasing distances from that surface. The centerline distances ranged from 1.8 to 21.3 mm (0.07 to 0.84 in.), in increments of 1.8 mm (0.07 in.). A central conductor, dry magnetic particles, and continuous magnetization were used for this test. The three types of direct current were straight direct current from batteries, three-phase rectified alternating current with surge, and half-wave rectified single-phase 60-Hz alternating current. The threshold values of current necessary to give readable indications of the holes in the ring are plotted in Fig. 17. Current levels as read on the usual meters were varied from the minimum needed to indicate hole 1 (1.8 mm, or 0.07 in., below the surface) for each type of current, up to a maximum of over 1000 A. To produce an indication at hole 1 using alternating current, about 475 A was required, and at hole 2 (3.56 mm, or 0.14 in., below the surface), over 1000 A. Hole 3 (5.33 mm, or 0.21 in., below the surface) could not be revealed with alternating current at any current level available. Indications at hole 2 were produced using 450-A straight direct current, 320-A direct current preceded by a surge of twice that amperage, and 250-A half-wave current. Indications were produced at hole 12 (21.3 mm, or 0.84 in., below the surface) using 750-A half-wave current, while 975-A straight direct current was required for hole 10 (17.8 mm, or 0.70 in., below the surface). The current levels needed to produce indications using wet particles were somewhat higher. For example, an indication for hole 1 using direct current and wet particles required approximately 440 A, and for hole 3, approximately 910 A. Over 625 A was required to detect hole 1 using alternating current and wet particles. The hardness of the testpiece also had an effect on the current level needed to produce indications. At a hardness of

63 HRC, to produce an indication at hole 1, approximately 200 A of half-wave current, 300 A of direct current with surge, and 450 A of direct current were needed. For hole 3, the current levels needed for the three types of current were approximately 1300, 1875, and 2700 A, respectively. Tests similar to the one described above have been performed on ring specimens made of 1020 and 4130 steels (Ref 1). For the inspection of finished parts such as machined and ground shafts, cams, and gears of precision machinery, direct current is frequently used. Alternating current is used for detecting fine cracks that actually break the surface, but direct current is better for locating very fine nonmetallic stringers lying just beneath the surface. Method of Magnetization The method of magnetization refers to whether residual magnetism in the part provides a leakage field strong enough to produce readable indications when particles are applied or if the part must be continuously magnetized while the particles are applied. Residual Magnetism. The procedure for magnetic particle inspection with residual magnetism, using either wet or dry

particles, basically consists of two steps: establishing a magnetic field in the part and subsequently applying the magnetic particles. The method can be used only on parts made of metals having sufficient retentivity. The residual magnetic field must be strong enough to produce leakage fields at discontinuities that in turn produce readable indications. This method is reliable only for detecting surface discontinuities. Either the dry or the wet method of applying particles can be used with residual magnetization. With the wet method, the magnetized parts can either be immersed in a gently agitated bath of suspended metallic particles or flooded by a curtain spray. The time of immersion of the part in the bath can affect the strength of the indications. By leaving the magnetized part in the bath or under the spray for a considerable time, the leakage fields, even at fine discontinuities, can have time to attract and hold the maximum number of particles. The location of the discontinuity on the part as it is immersed has an effect on particle buildup. Buildup will be greatest on horizontal upper surfaces and will be less on vertical surfaces and horizontal lower surfaces. Parts should be removed from the bath slowly because rapid removal can wash off indications held by weak leakage fields. Continuous Magnetism. In the continuous method, parts are continuously magnetized while magnetic particles are

applied to the surfaces being inspected. In the dry-particle continuous method, care must be taken not to blow away

indications held by weak leakage fields. For this reason, the magnetizing current is left on during the removal of excess particles. In the wet-particle continuous method, the liquid suspension containing the magnetic particles is applied to the part, and the magnetizing current is applied simultaneously with completion of particle application. This prevents washing away of indications held by weak leakage fields. For reliability of results, the wet continuous method requires more attention to timing and greater alertness on the part of the operator than the wet residual method. The continuous method can be used on any metal that can be magnetized because in this method residual magnetism and retentivity are not as important in producing a leakage field at a discontinuity. This method is mandatory for inspection of low-carbon steels or iron having little or no retentivity. It is frequently used with alternating current on such metals because the ac field produces excellent mobility of dry magnetic particles. Maximum sensitivity for the detection of very tine discontinuities is achieved by immersing the part in a wet-particle bath, passing the magnetizing current through the part for a short time during immersion, and leaving the current on as the part is removed and while the bath drains from the surface. Direction of Magnetization The shape and orientation of the suspected discontinuity in relation to the shape and principal axis of the part have a bearing on whether the part should be magnetized in a circular or a longitudinal direction or in both directions. The rule of thumb is that the current must be passed in a direction parallel to the discontinuity. If the principal direction of the discontinuities is unknown, to detect all discontinuities, both circular and longitudinal magnetization must be used; with the prod and yoke methods, the prods or yoke must be repositioned at 90° from the first magnetizing position. The magnetic particle background held on a part by extraneous leakage fields is minimized when using the circular method of magnetizing because the field generally is self-contained within the part. Magnitude of Applied Current The amount of magnetized current or the number of ampere-turns needed for optimum results is governed by the types and minimum dimensions of the discontinuities that must be located or by the types and sizes of discontinuities that can be tolerated. The amount of current for longitudinal magnetization with a coil is initially determined by Eq 1 and 2. For circular magnetization, when magnetizing by passing current directly through a part, the current should range from 12 to 31 A/mm (300 to 800 A/in.) of the diameter of the part. The diameter is defined as the largest distance between any two points on the outside circumference of the part. Normally, the current used should be 20 A/mm (500 A/in.) or lower, with the higher current values of up to 31 A/mm (800 A/in.) used to inspect for inclusions or to inspect alloys such as precipitation-hardened steels. The prod method of magnetization usually requires 4 to 4.92 A/mm (100 to 125 A/in.) of prod spacing. Prod spacing should not be less than 50 mm (2 in.) nor more than 203 mm (8 in.). Equipment Selection of equipment for magnetic particle inspection depends on the size, shape, number, and variety of parts to be tested. Bench Units. For the production inspection of numerous parts that are relatively small but not necessarily identical in shape, a bench unit with contact heads for circular magnetization, as well as a built-in coil for longitudinal magnetization, is commonly used (Fig. 11). Portable units using prods, yokes, or hand-wrapped coils may be most convenient for large parts. Half-wave current

and dry particles are often used with portable equipment. Wet particles can be used with portable equipment, but the bath is usually not recovered. Mass Production Machinery. For large lots of identical or closely similar parts, single-purpose magnetization-and-

inspection units or fixtures on multiple-purpose units can be used.

Reference cited in this section

1. Mater. Eval., Vol 30 (No. 10), Oct 1972, p 219-228 Magnetic Particle Inspection Revised by Art Lindgren, Magnaflux Corporation

Inspection of Hollow Cylindrical Parts Some hollow cylindrical parts requiring magnetic particle inspection present difficulties in processing because of configuration, extraneous leakage field interference, requirements for noncontact of magnetizing units, the overall time required, or low L/D ratio. Techniques for inspecting long pieces of seamless tubing (oil well tubing), butt-welded and longitudinally welded carbon steel pipe or tubing, and a cylinder with a closed end are described in the following sections. Oil well tubing is made of high-strength steel using hot finishing operations and has upset ends for special threading.

The discontinuities most expected are longitudinally oriented on the main body of the tube and transversely oriented on the upset ends. For these reasons, the entire length of the tube is circularly magnetized and inspected for longitudinal-type discontinuities. Also, the upset ends are longitudinally magnetized and inspected for transverse-type discontinuities. Tube sections are usually more than 6 m (20 ft) long. An insulated central conductor is used to introduce circular magnetism instead of passing the current through the material by head contacts. The central conductor facilitates inspection of the inside surface of the upset ends; the direct-contact method would not provide the required field. The central-conductor technique used for circumferential magnetization of the tube body is illustrated in Fig. 38(a). The magnetic particles are applied to the outside surface of the tube. The residual-magnetism technique is used. The current density for this test is usually 31 to 39 A/mm (800 to 1000 A/in.) of tube diameter.

Fig. 38 Magnetic particle inspection of oil well tubing for longitudinal (a) and circumferential (b) discontinuities

The encircling-coil technique used to magnetize the upset ends in the longitudinal direction is shown in Fig. 38(b). The residual-magnetism technique is used. Both the inside and outside surfaces are inspected for discontinuities.

Welds in Carbon Steel Pipe. Magnetic particle inspection using the prod technique is a reliable method of detecting

discontinuities in consumable-insert root welds and final welds in carbon steel pipe up to 75 mm (3 in.) in nominal diameter. (For larger-diameter pipe, less time-consuming magnetic particle techniques can be used.) The types of discontinuities found in root welds are shown in Fig. 39(a), and those found in final welds are shown in Fig. 39(b).

Fig. 39 Magnetic particle inspection for detection of discontinuities in consumable-insert root welds and final welds in carbon steel pipe. Dimensions given in inches

Placement of the prods is important to ensure reliable inspection of the welds. Circular magnetization, used to check for longitudinal discontinuities, is accomplished by placing the prods at 90° intervals (four prod placements) around the pipe, as shown in Fig. 39(c). For pipes larger than 25 mm (1 in.) in nominal diameter, prods should be spaced around the pipe at approximately 50 mm (2 in.) intervals, as shown in Fig. 39(e). Circumferentially oriented discontinuities are revealed by placing the prods as shown in Fig. 39(d). The prods are placed adjacent to and on opposite sides of the weld bead to ensure flux flow across the weld metal. If the circumferential distance between the prods is greater than 75 mm (3 in.) when positioned as shown in Fig. 39(d), the prods should be positioned as shown in Fig. 39(f). To ensure proper magnetization, the areas inspected should overlap approximately 25 mm (1 in.). Magnetic particles are applied to the weld area while the current is on because of the low retentivity of carbon steel. A buildup of magnetic particles in the fusion-line crevice of the weld is indicative of either a subsurface discontinuity or a nonrelevant indication because of the abrupt change in material thickness and/or the crevicelike depression between the weld metal and the base metal. However, a true indication, as from incomplete fusion between the weld metal and the base metal, would be a sharply defined particle pattern. This pattern would be difficult if not impossible to remove by blowing with a hand-held powder blower while the magnetizing current is being applied. If the indication at the fusion zone can be blown away with a hand-held powder blower, the indication is nonrelevant. The current used is approximately 3.9 A/mm (100 A/in.) of prod spacing. Hollow cylinders closed at one end, such as drawn shells or forged fluid-power cylinders, can be magnetized circumferentially for the inspection of longitudinal discontinuities using a head shot end-to-end. However, this technique does not provide sensitivity for discontinuities on the inside surface.

As shown in Fig. 40, a central conductor can be used in such a manner that the closed end of the cylinder completes the current path. Also, the open end of the cylinder is accessible for the application of a wet-particle bath to the inside surface, which can then be inspected directly. For thin-wall cylinders, discontinuities on the inside surface produce subsurface-type indications on the outside surface. The central-conductor method for magnetization is advantageous when the inside diameter is too small to permit direct internal viewing.

Fig. 40 Use of a central conductor for the magnetic particle testing of a cylinder with one end closed

Magnetic Particle Inspection Revised by Art Lindgren, Magnaflux Corporation

Inspection of Castings and Forgings Castings and forgings may be difficult to inspect because of their size and shape. External surfaces can usually be inspected with prods; however, on large parts this can be time consuming, and inspection of interior surfaces may not be adequate. High-amperage power supplies, in conjunction with flexible cable used with clamps (as contact heads), central conductors, or wrapping, can effectively reduce inspection time because relatively large areas can be inspected with each processing cycle. Figure 41 illustrates the direct-contact, cable-wrap, and central-conductor techniques that can be used to inspect two relatively large parts. The three circuits for each part can be applied on a single-shot basis or, if high-output multidirectional equipment is available, can be combined into a single inspection cycle.

Fig. 41 Methods of using cable for applying magnetizing circuits to large forgings and castings. For the forging in (a), circuits 1 and 3 are head shots and circuit 2 is a cable wrap. For the casting in (b), circuits 1 and 3 are central conductors and circuit 2 is a cable wrap.

Generally, the higher-amperage power supplies are of the dc type. Alternating current and half-wave current supplies are limited to outputs of approximately 5000 to 6000 A because of the reactive impedance component associated with these types of current. Wet particles are generally preferred as the inspection medium in the presence of strong dc fields because wet particles exhibit much more mobility on the surface of a part than dry particles. Wet particles also readily permit full coverage of large surface areas and are easy to apply to internal surfaces. Crane Hooks. The inspection of crane hooks, as required by the Occupational Safety and Health Act, has focused attention on cracks and other discontinuities in these components. Crane hooks are generally magnetic particle inspected with electromagnetic yokes having flexible legs. Power supplies are 115-V, 60-Hz alternating current and half-wave current. Stress areas in a crane hook are:

• • •

The bight (in tension) on both sides and in the throat (area A, Fig. 42) The area below the shank (in compression and tension) on four sides (area B, Fig. 42) The shank (in tension), mainly in threads and fillet (area C, Fig. 42)

Fig. 42 Forged crane hook showing stress areas subject to inspection

The steps involved in the magnetic particle inspection of crane hooks are as follows:

1. Remove dirt and oil from hook 2. Magnetize and apply particles to areas A and B in Fig. 42, using a yoke and an ac field parallel to the axis of the hook 3. For hooks out of surface, inspect the shank (area C, Fig. 42), using a yoke and an ac field parallel to the axis of the hook 4. For hooks in service, inspect the shank ultrasonically 5. Repeat steps 2 and 3 using a dc field for subsurface indications

The 50 kN (6 tonf) crane hook shown in Fig. 43 was removed from a jib crane after an indication was found during magnetic particle inspection. The discontinuity was found to be a deep forging lap. Sectioning the hook through the lap revealed a discontinuity about 19 mm ( in.) deep in a 50 mm (2 in.) square section (section A-A, in Fig. 43). All inspections of the hook were made with the magnetic field parallel to the hook axis, thus inspecting for transverse cracks and other discontinuities. Because of the depth of the lap, the defect was detected even though the field was parallel to its major dimension. A sufficient flux-leakage field occurred at the surface to attract the magnetic particles.

Fig. 43 50 kN (6 tonf) crane hook showing magnetic particle indication of a forging lap and section through hook showing depth of lap

Inspection of a 90 kN (10 tonf) crane hook revealed a forging lap in the area below the shank. The lap was transverse and was located in the fillet below the keeper hole. A fatigue crack had initiated in the lap. Drive-Pinion Shaft. An annual preventive maintenance inspection was performed on the large forged drive-pinion shaft shown in Fig. 44(a), and detection of a large crack (arrows, Fig. 44b) in the fillet between the shaft and coupling flange prevented a costly breakdown. Three areas were magnetic particle inspected for cracks:

• • •

Along the shaft The fillet between the shaft and the coupling flange (cracked area, Fig. 44b) At each fillet in the wobbler coupling half (arrows, Fig. 44a)

Fig. 44 Forged drive-pinion shaft and coupling in which the detection of a crack during preventive maintenance magnetic particle inspection prevented a costly breakdown. (a) Drive-pinion shaft and coupler assembly; arrows show locations of fillets on wobbler coupling half that were inspected. (b) Fillet between shaft and coupling flange showing crack (at arrows) found during inspection

Inspection was conducted using a portable power source capable of up to 1500-A output in alternating current or halfwave direct current. Double prod contacts and a 4/0 cable were used to introduce the magnetic fields in the shaft. The fillets in the wobbler coupling half were inspected for discontinuities with double prod contacts. The steps involved in inspecting the shaft were as follows:

1. Clean all areas to be tested 2. Wrap cable around shaft and apply current providing 2900 ampere-turns. Inspect for transverse cracks in the shaft portion 3. Place prods across fillet at coupling flange (Fig. 44b) at spacing of 152 to 203 mm (6 to 8 in.); apply 500-A current, producing a circular field perpendicular to fillet. Inspect for discontinuities parallel to fillet 4. Place prods across fillet at flange on wobbler coupling half and across fillets in shaft portion (arrows, Fig. 44a) at spacing of 152 to 203 mm (6 to 8 in.); apply 500-A current, producing a circular field perpendicular to the fillets. Inspect for discontinuities parallel to the fillets

Disk or Gear on Shaft. A forged or cast disk or gear on a heavy through shaft can be wrapped with a cable to form

two opposing coils, as shown for the disk in Fig. 45. The two opposing coils produce radial magnetic fields on each side of the disk. This type of field reveals circumferential discontinuities on the sides of the disk and transverse discontinuities in the shaft. Also, the shaft can be used as a central conductor for the purpose of locating radial discontinuities in the disk and longitudinal discontinuities in the shaft.

Fig. 45 Disk on a through shaft in which the shaft was cable wrapped to produce a longitudinal magnetic field in the shaft and a radial field in the disk. Using the shaft as a central conductor produced a circular magnetic field in both the shaft and the disk.

In parts where the shaft extends from one side only, a pole piece can be used to simulate a through shaft. The pole piece should have approximately the same diameter as the shaft. Y-shaped parts may not be processed on a horizontal wet-particle unit in such a manner that only one shot will be

required for complete inspection. This is true even though steps are taken to ensure that the current is divided equally through the two upper legs of the Y. With equally divided current, an area of no magnetization will exist at the junction of the two legs. A Y-shaped part can be processed in two steps using a conventional horizontal wet-particle unit, or it can be processed in one operation using a modified unit having a double headstock and special current assurance magnetizing circuit. Magnetic Particle Inspection Revised by Art Lindgren, Magnaflux Corporation

Inspection of Weldments Many weld defects are open to the surface and are readily detectable by magnetic particle inspection with prods or yokes. For the detection of subsurface discontinuities, such as slag inclusions, voids, and inadequate joint penetration at the root of the weld, prod magnetization is best, using half-wave current and dry powder. Yokes, using alternating current, direct current, or half-wave current, are suitable for detecting surface discontinuities in weldments. The positioning of a yoke with respect to the direction of the discontinuity sought is different from the corresponding positioning of prods. Because the field traverses a path between the poles of the yoke, the poles must be placed on opposite sides of the weld bead to locate cracks parallel to the bead (Fig. 6), and adjacent to the bead to locate transverse cracks. Prods are spaced adjacent to the weld bead for parallel cracks and on opposite sides of the bead for transverse cracks. For applications in which holding the prod contacts by hand is difficult or tiring, prods incorporating magnetic clamps, or leeches, that magnetically hold the prods to the work are available. The prods carrying the magnetizing current are held firmly to the work by an electromagnet. Both prods can be attached by the magnets, or one of the prods can be held magnetically and the other by hand.

There is one type of weld on which the penetrating power of half-wave current results in nonrelevant indications: a Tjoint welded from one or both sides for which complete joint penetration is not specified and in which an open root is permissible and almost always present. When half-wave current is used with prods, this open root will probably be indicated on the weld surface. This nonrelevent indication can be eliminated by using alternating current instead of halfwave current. A case in which nonrelevant magnetic particle indications were found occurred at the welded T-joints between six tubes and the end plates of a complicated assembly (Fig. 46). The welds were made from the outside of the tubes only. Liquid penetrant and radiographic inspections and metallographic examination disclosed that the integrity of the welds was good. Investigations revealed that the depth of penetration of the field produced by dc magnetization was sufficient to reveal the joint along the inside wall of the cylinder. Inspection by ac magnetization, which had less depth of penetration of the field, eliminated these indications.

Fig. 46 Assembly in which T-joints between tubes and end plates that were welded (as specified) with partial penetration produced nonrelevant magnetic particle indications when inspected using dc magnetization

The detectability of subsurface discontinuities in butt welds between relatively thin plates can often be improved by positioning a dc yoke on the side opposite the weld bead, as shown in Fig. 47. Magnetic particles are applied along the weld bead. Improvement is achieved because of the absence of extraneous leakage flux that normally emanates from the

yoke pole pieces. Techniques developed and used for the magnetic particle inspection of weldments are described in the two examples that follow.

Fig. 47 Method of using a yoke for detecting subsurface discontinuities in butt welds joining thin plates

Example 3: Magnetic Particle Inspection of Shielded Metal-Arc Welds in the Outer Hull of a Deep-Submergence-Vessel Float Structure. The outer hull of the float structure of a deep-submergence vessel was made of 3.2 mm ( in.) thick steel plate and was butt welded using the shielded metal-arc process. For magnetic particle inspection, the weld areas were magnetized with prods, using a half-wave current of 3.9 A/mm (100 A/in.) of prod spacing. Dry, red magnetic particles provided adequate sensitivity and color contrast for inspection. The crowns on the surfaces of the welds were removed with a flat chisel. Neither grinding of the welds nor the use of pneumatic needle guns was permitted. The surfaces were cleaned of paint, scale, and slag by sand or vapor blasting. The area to be inspected was shielded from air currents to prevent disturbing the magnetic particle indications. Magnetic particles were applied with a powder dispenser hand-held 305 to 381 mm (12 to 15 in.) from the weld at an angle of 30 to 45° to the surface of the plate along the axis of the weld. The powder dispenser was held stationary once it was positioned for each prod placement. The powder was allowed to float to the weld surface so that a very light dustlike coating of particles settled on the surface being tested. Excessive application of magnetic particles required removal of all particles and reapplication. Longitudinal discontinuities were detected by placing the prods at a 152 mm (6 in.) spacing parallel to the longitudinal axis of the weld and on the surface of the weld. Successive prod spacings overlapped approximately 25 mm (1 in.) to ensure complete coverage. For the detection of transverse discontinuities, prods were positioned at right angles to the weld axis, 75 to 152 mm (3 to 6 in.) apart on each side of the weld, and at intervals not exceeding one-fourth the prod spacing. Magnetic particle inspection was used instead of radiography. Following are the maximum radiographic acceptance standards from which the magnetic particle procedures were developed: •

A linear or linearly disposed indication 1.6 mm ( times greater than its width

•

A single slag (subsurface) indication longer than 0.8 mm ( together than 4.8 mm ( length

•

in.) or more in length, with its length at least four in.), and multiple slag indications closer

in.) or totaling more than 3.2 mm (

in.) in any 152 mm (6 in.) of weld

A cluster of four or more indications of any size spaced within 3.2 mm (

in.) of each other

Example 4: Magnetic Particle Inspection of a Girder Weldment for an Overhead Crane. The girder weldment for an overhead crane (Fig. 48a) was magnetic particle inspected in the presence of the customer's representative. Fabrication of the weldment was complete, and the weldment had been sandblasted in preparation for painting.

Fig. 48 Magnetic particle inspection of a girder weldment for an overhead crane. (a) Girder weldment, with section showing welds and detail showing prod locations. (b) Magnetograph of magnetic field used to check for discontinuities in the welds. (c) Magnetograph of field used to check for discontinuities in top plate. Note different prod placement in (b) and (c). Dimensions given in inches

The dry-powder, continuous-magnetization method was used with half-wave current. The current used was 3.9 to 5.9 A/mm (100 to 150 A/in.) of prod contact spacing, which was 152 mm (6 in.). Inspected were the fillet welds joining the top and bottom plates to the web plate (section A-A in Fig. 48), and all butt seam welds of the top, bottom, and web plates. The edges of the top and bottom plates were also inspected for discontinuities. All prod-contact points on the girder were ground to ensure good contact; prod locations are shown in detail B in Fig. 48.

During magnetic particle testing of the welds, the prods were alternated on either side of the welds. A magnetograph of the magnetic flux during inspection of a weld joining the top plate to the web plate is shown in Fig. 48(b). Inspection of the edges of the top and bottom plates was performed by placing the prods parallel to the weld with contacts on the edge of the plate; a magnetograph of this test is shown in Fig. 48(c). Longitudinal indications were observed in both the plate and the weld. Nonrelevant indications that the process was capable of detecting were established prior to inspecting the girder. All the indications detected were verified as either false or nonrelevant, signifying an absence of defects. Magnetic Particle Inspection Revised by Art Lindgren, Magnaflux Corporation

Inspection of Billets A billet is the last semifinished intermediate step between the ingot and the finished shape. Steel billets are rectangular or square and range from 2600 to 32,000 mm2 (4 to 49 in.2) in cross-sectional area. The magnetic particle inspection of billets requires a large unit equipped to handle billets 50 to 184 mm (2 to 7 in.) square and 2.4 to 12 m (8 to 40 ft) long. The amperage setting on the testing unit should be 1200 to 4000 A. The discontinuities shown in Fig. 49 would appear as bright fluorescent indications under ultraviolet light.

Fig. 49 Discontinuities on the surfaces of steel billets that can be detected by magnetic particle inspection. (a) Arrowhead cracks. (b) Longitudinal cracks. (c) Normal seams. (d) Brush seams. (e) Laps. (f) Scabs. See text for discussion.

Cracks in billets appear as deep vertical breaks or separations in the surface of the steel. Arrowhead cracks (Fig. 49a)

occur early in processing, usually as the result of primary mill elongation of an ingot containing a transverse crack. Longitudinal cracks (Fig. 49b) appear as relatively straight lines in the direction of rolling. They are 0.3 m (1 ft) or more in length and usually occur singly or in small numbers.

Seams are longitudinal discontinuities that appear as light lines in the surface of the steel. Normal seams (Fig. 49c) are similar to longitudinal cracks, but produce lighter indications. Seams are normally closed tight enough that no actual opening can be visually detected without magnetic particle inspection. Seams have a large number of possible origins, some mechanical and some metallurgical. Brush seams (Fig. 49d) are clusters of short ( 10. For magabsorption measurements on wire, the filling factor, F, which is the ratio of the volume of the wire sample to the volume inside of the RF coil, is less than 0.01. This value of the filling factor keeps the loaded Q of the coil also greater than 10. The theoretical derivation has shown that ΔR ωΔL. Taking into account the above assumptions and because Ri 10 Lo/RoC and ΔR is less than 0.1 Ro, the voltage change, ΔV, can be approximated within 1% to be directly proportional to the ΔR from magabsorption. The change in resonant frequency can also be shown to be directly related to ΔL. The Magabsorption Bridge Detector. The voltage across the resonant RF circuit in Fig. 12 is Vc + ΔV. Although the

actual value is approximately 0.1 Ei, there may be problems in amplifying Vc + ΔV because of the dynamic range of many amplifiers. Therefore, the bridge circuit in Fig. 16 has been used to eliminate Vc and to give an output of only ΔV. The left resonant circuit contains the material sample, while the right resonant circuit contains no sample. The rectified voltages across each resonant circuit are detected, subtracted, filtered, and supplied to the output. With no magabsorption signal, the output is zero. With a magabsorption signal, the output of the bridge in Fig. 16 is:

(Eq 12)

The magabsorption bridge is used when the variation rate for R is very low (1 MeV) x-rays for the radiographic field inspection of thicker sections. Certain types of flaws are difficult to detect by radiography. Cracks cannot be detected unless they are essentially parallel to the radiation beam. Tight cracks in thick sections may not be detected at all, even when properly oriented. Minute discontinuities such as inclusions in wrought material, flakes, microporosity, and microfissures may not be detected unless they are sufficiently segregated to yield a detectable gross effect. Laminations are nearly impossible to detect with radiography; because of their unfavorable orientation, delaminations do not yield differences in absorption that enable laminated areas to be distinguished from delaminated areas. It is well known that large doses of x-rays or -rays can kill human cells, and in massive doses can cause severe disability or death. Protection of personnel--not only those engaged in radiographic work but also those in the vicinity of radiographic inspection--is of major importance. Safety requirements impose both economic and operational constraints on the use of radiography for inspection.

Radiographic Inspection Revised by the ASM Committee on Radiographic Inspection*

Principles of Radiography Three basic elements of radiography include a radiation source, the testpiece or object being evaluated, and a sensing material. These elements are shown schematically in Fig. 1. The testpiece in Fig. 1 is a plate of uniform thickness containing an internal flaw that has absorption characteristics different from those of the surrounding material. Radiation from the source is absorbed by the testpiece as the radiation passes through it; the flaw and surrounding material absorb different amounts of radiation. Thus, the intensity of radiation that impinges on the sensing material in the area beneath the flaw is different from the amount that impinges on adjacent areas. This produces an image, or shadow, of the flaw on the sensing material. This section briefly reviews the general characteristics and safety principles associated with radiography.

Fig. 1 Schematic of the basic elements of a radiographic system showing the method of sensing the image of an internal flaw in a plate of uniform thickness

Radiation Sources Two types of electromagnetic radiation are used in radiographic inspection: x-rays and -rays. X-rays and -rays differ from other types of electromagnetic radiation (such as visible light, microwaves, and radio waves) only in their wavelengths, although there is not always a distinct transition from one type of electromagnetic radiation to another (Fig. 2). Only x-rays and -rays, because of their relatively short wavelengths (high energies), have the capability of penetrating opaque materials to reveal internal flaws.

Fig. 2 Schematic of the portion of the electromagnetic spectrum that includes x-rays, -rays, ultraviolet and visible light, and infrared radiation showing their relationship with wavelength and photon energy

X-rays and -rays are physically indistinguishable; they differ only in the manner in which they are produced. X-rays result from the interaction between a rapidly moving stream of electrons and atoms in a solid target material, while rays are emitted during the radioactive decay of unstable atomic nuclei. The amount of exposure from x-rays or -rays is measured in roentgens (R), where 1 R is the amount of radiation exposure that produces one electrostatic unit (3.33564 × 10-10 C) of charge from 1.293 mg (45.61 × 10-6 oz) of air. The intensity of an x-ray or -ray radiation is measured in roentgens per unit time. Although the intensity of x-ray or -ray radiation is measured in the same units, the strengths of x-ray and -ray sources are usually given in different units. The strength of an x-ray source is typically given in roentgens per minute at one meter (RMM) from the source or in some other suitable combination of time or distance units (such as roentgens per hour at one meter, or RHM). The strength of a -ray source is usually given in terms of the radioactive decay rate, which has the traditional unit of a curie (1 Ci = 37 × 109 disintegrations per second). The corresponding unit in the Système International d'Unités (SI) system is a gigabecquerel (1 GBq = 1 × 109 disintegrations per second). The spectrum of radiation is often expressed in terms of photon energy rather than as a wavelength. Photon energy is measured in electron volts, with 1 eV being the energy imparted to an electron by an accelerating potential of 1 V. Figure 2 shows the radiation spectrum in terms of both wavelength and photon energy. Production of X-Rays. When x-rays are produced from the collision of fast-moving electrons with a target material, two types of x-rays are generated. The first type of x-ray is generated when the electrons are rapidly decelerated during collisions with atoms in the target material. These x-rays have a broad spectrum of many wavelengths (or energies) and are referred to as continuous x-rays or by the German word bremsstrahlung, which means braking radiation. The second type of x-ray occurs when the collision of an electron with an atom of the target material causes a transition of an orbital electron in the atom, thus leaving the atom in an excited state. When the orbital electrons in the excited atom rearrange themselves, x-rays are emitted that have specific wavelengths (or energies) characteristic of the particular electron rearrangements taking place. These characteristic x-rays usually have much higher intensities than the background of bremsstrahlung having the same wavelengths. Production of

-Rays. Gamma rays are generated during the radioactive decay of both naturally occurring and

artificially produced unstable isotopes. In all respects other than their origin, -rays and x-rays are identical. Unlike the broad-spectrum radiation produced by x-ray sources, -ray sources emit one or more discrete wavelengths of radiation, each having its own characteristic photon energy (or wavelength).

Many of the elements in the periodic table have either naturally occurring radioactive isotopes or isotopes that can be made radioactive by irradiation with a stream of neutrons in the core of the nuclear reactor. However, only certain isotopes are extensively used for radiography (see the section "Gamma-Ray Sources" in this article). Image Conversion The most important process in radiography is the conversion of radiation into a form suitable for observation or further signal processing. This conversion is accomplished with either a recording medium (usually film) or a real-time imaging medium (such as fluorescent screens or scintillation crystals). The imaging process can also be assisted with the use of intensifying or filtration screens, which intensify the conversion process or filter out scattered radiation. Recording media provide a permanent image that is related to the variations in the intensity of the unabsorbed

radiation and the time of exposure. With a recording medium such as film, for example, an invisible latent image is formed in the areas exposed to radiation. These exposed areas become dark when the film is processed (that is, developed, rinsed, fixed, washed, and dried), with the degree of darkening (the photographic density) depending on the amount of exposure to radiation. The film is then placed on an illuminated screen so that the image formed by the variations in photographic density can be examined and interpreted. Real-time imaging media provide an immediate indication of the intensity of radiation passing through a testpiece.

With fluorescent screens, for example, visible light is emitted with a brightness that is proportional to the intensity of the incident x-ray or -ray radiation. This emitted light can be observed directly, amplified, and/or converted into a video signal for presentation on a television monitor and subsequent recording. The various types of imaging systems are described in the section "Real-Time Radiography" in this article. Intensifying and filtration screens are used to improve image contrast, particularly when the radiation intensity is

low or when the radiation energy is high. The screens are useful at higher energies because the sensitivity of films and fluorescent screens decreases as the energy of the penetrating radiation increases. The various types of screens are discussed in the section "Image Conversion Media" in this article. Image Quality and Radiographic Sensitivity The quality of radiographs is affected by many variables, and image quality is measured with image-quality indicators known as penetrameters. These devices are thin specimens made of the same material as the testpiece; they are described in more detail in the section "Identification Markers and Penetrameters (Image-Quality Indicators)" in this article. When placed on the testpiece during radiographic inspection, the penetrameters measure image contrast and, to a limited extent, resolution. Detail resolution is not directly measured with penetrameters, because flaw detection depends on the nature of the flaw, its shape, and orientation to the radiation beam. Image quality is governed by image contrast and resolution, which are also sometimes referred to as radiographic contrast and radiographic definition. These two factors are interrelated in a complex way and are affected by several factors described in the sections "Radiographic contrast" and "Radiographic definition" in this article. In real-time systems, image contrast and resolution are also described in terms of the detective quantum efficiency (DQE) (also known as the quantum detection efficiency, or QDE) and the modulation transfer function (MTF), respectively. These terms, which are also used in computed tomography, are defined in the text and in Appendix 2 of the article "Industrial Computed Tomography" in this Volume. Radiographic sensitivity, which should be distinguished from image quality, generally refers to the size of the smallest detail that can be seen on a radiograph or to the ease with which the images of small details can be detected. Although radiographic sensitivity is often synonymous with image quality in applications requiring the detection of small details, a distinction should be made between radiographic sensitivity and radiographic quality. Radiographic sensitivity refers more to detail resolvability, which should be distinguished from spatial resolution and contrast resolution. For example, if the density of an object is very different from the density of the surrounding region, the flaw might be resolved because of the large contrast, even if the flaw is smaller than the spatial resolution of the system. On the other hand, when the contrast is small, the area must be large to achieve resolvability. Radiographic contrast refers to the amount of contrast observed on a radiograph, and it is affected by subject contrast and the contrast sensitivity of the image-detecting system. Radiographic contrast can also be affected by the unsharpness of the detected image. Figure 3, for example, shows how contrast is affected when the sharp edge of a step is blurred

because of unsharpness. If the unsharpness is much smaller than d in Fig. 3, then the contrast is not reduced and the edges in the image are easily defined. However, if d is much smaller than the unsharpness, then the contrast is reduced. If the unsharpness is too large, the image cannot be resolved.

Fig. 3 Effect of geometric unsharpness on image contrast. (a) Flaw size, d, is larger than the unsharpness, then full contrast occurs. (b) Flaw size, d, is smaller than the unsharpness, then contrast is reduced.

Subject contrast is the ratio of radiation intensities transmitted by various portions of a testpiece. It depends on the

thickness, shape, and composition of the testpiece; the intensity of the scattered radiation; and the spectrum of the incident radiation. It does not depend on the detector, the source strength, or the source-to-detector distance. However, it may depend on the testpiece-to-detector distance because the intensity of scattered radiation (from within the testpiece) impinging upon a detector is almost eliminated when projective enlargement approaches a magnification factor of three (see the section "Enlargement" in this article). In general, a radiograph should be made with the lowest-energy radiation that will transmit adequate radiation intensities to the detector, because long wavelengths tend to improve contrast. However, radiation energies that are too low produce excessive amounts of scattered radiation that washes out fine detail. On the other hand, energies that are too high, although they reduce scattered radiation, may produce images having contrast that is too low to resolve small flaws. For each situation, there is an optimum radiation energy that produces the best combination of radiographic contrast and definition, or the greatest sensitivity.

The amount of subject contrast can be modeled as follows. If there is a change in thickness, Δx, of a testpiece with a uniform attenuation coefficient, μ, then the subject contrast (without scattering) is equal to μΔx. Similarly, if there is a change in the attenuation coefficient, Δμ, over a distance X, then the subject contrast (without scatter) is equal to XΔμ. If there is scattering, the subject contrast is reduced such that:

(Eq 1)

where IS is the intensity of the scattered radiation and ID is the intensity of the direct (or primary) radiation passing through the testpiece. The scattering may occur from within the testpiece (Fig. 4a) or from the surroundings (Fig. 4b and c). The various mechanisms of scattering and attenuation are described in the section "Attenuation of Electromagnetic Radiation" in this article.

Fig. 4 Schematic of the three types of scattered radiation encountered when using x-rays and

-rays

Contrast sensitivity refers to the ability of responding to and displaying small variations in subject contrast. Contrast sensitivity depends on the characteristics of the image detector and on the level of radiation being detected (or on the amount of exposure for films). The relationship between the contrast sensitivity and the level of radiation intensity (or film exposure) can be illustrated by considering two extremes. At low levels of radiation intensity, the contrast sensitivity of the detectors is reduced by a smaller signal-to-noise ratio, while at high levels, the detectors become saturated. Consequently, contrast sensitivity is a function of dynamic range.

In film radiography, the contrast sensitivity is:

(Eq 2)

where D is the smallest change in photographic density that can be observed when the film is placed on an illuminated screen. The factor GD is called the film gradient or film contrast. The film gradient is the inherent ability of a film to record a difference in the intensity of the transmitted radiation as a difference in photographic density. It depends on film type, development procedure, and film density. For all practical purposes, it is independent of the quality and distribution of the transmitted radiation (see the section "Film Radiography" in this article). Contrast sensitivity in real-time systems is determined by the number of bits (if the image is digitized) and the signal-tonoise ratio (which is affected by the intensity of the radiation and the efficiency of the detector). The factors affecting the contrast sensitivity and signal-to-noise ratio of real-time systems are discussed in the section "Modern Image Intensifiers" in this article. The best contrast sensitivity of digitized images from fluorescent screens is about 8 bits (or 256 gray levels). Another way of specifying the contrast sensitivity of fluorescent screens is with a gamma factor, which is defined by the fractional unit change in screen brightness, ΔB/B, for a given fractional change in the radiation intensity, ΔI/I. Most fluorescent screens have a gamma factor of about one, which is not a limiting factor. At low levels of intensity, however,

the contrast is reduced because of quantum mottle (see the section "Screen Unsharpness" in this article) just as unsharpness reduced the contrast in the example illustrated in Fig. 3. Dynamic range, or latitude, describes the ability of the imaging system to produce a suitable signal over a range of

radiation intensities. The dynamic range is given as the ratio of the largest signal that can be measured to the precision of the smallest signal that can be discriminated. A large dynamic range allows the system to maintain contrast sensitivity over a wide range of radiation intensities or testpiece thicknesses. Film radiography has a dynamic range of up to 1000:1, while digital radiography with discrete detectors can achieve 100,000:1. The latitude, or dynamic range, of film techniques is discussed in the section "Exposure Factors" in this article. Radiographic definition refers to the sharpness of a radiograph. In Fig. 5(a), for example, the radiograph exhibits poor radiographic definition even when the contrast is high. Better radiographic definition is achieved in Fig. 5(b), even though the contrast is lower.

Fig. 5 Radiographs with poor and improved radiographic definition. (a) Advantage of a higher radiographic contrast is offset by poor definition. (b) Despite lower contrast, better detail is obtained with improved definition. Source: Ref 1

The degree of radiographic definition is usually specified in line pairs per millimeter or by the minimum distance by which two features can be physically distinguished. The factors affecting radiographic definition are geometric unsharpness and the unsharpness (or spatial resolution) of the imaging system. The net effect of these two unsharpnesses is not directly additive, because the combination of the two unsharpnesses is complex. For most practical radiography, the total unsharpness, UT, is:

=

+

(Eq 3)

where Ug is the geometric unsharpness and Ud is the unsharpness or spatial resolution of the image-detecting system. Another important type of unsharpness occurs when in-motion radiography is performed. This type of unsharpness, termed motion unsharpness, is discussed in the section "In-Motion Radiography" in this article. Geometric unsharpness is usually the largest contributor to maximum unsharpness in film radiography, while the unsharpness of fluorescent screens in real-time radiography often overshadows the effect of geometric unsharpness. Distortion and Geometric Unsharpness. Because a radiograph is a two-dimensional representation of a three-

dimensional object, the radiographic images of most testpieces are somewhat distorted in size and shape compared to actual testpiece size and shape. The mechanisms of geometric unsharpness and distortion are described in the section "Principles of Shadow Formation" in this article. The severity of unsharpness and distortion depends primarily on the

source size (focal-spot size for x-ray sources), source-to-object and source-to-detector distances, and position and orientation of the testpiece with respect to source and detector. Film unsharpness specifies the spatial resolution of the film and must not be confused with film graininess. Film unsharpness and film graininess must be treated as two separate parameters.

Film unsharpness depends not only on the film grain size and the development process but also on the energy of the radiation and the thickness and material of the screens. When a quantum of radiation sensitizes a silver halide crystal in a film, adjacent crystals may also be sensitized if the quantum has sufficient energy to release electrons during the absorption by the first crystal. Consequently, a small volume of crystals would be sensitized, thus producing a small disk instead of a point image. Figure 6 illustrates an experimental curve of film unsharpness versus radiation energy. The screens used will also affect the unsharpness with the production of electrons.

Fig. 6 Experimental curve of film unsharpness, Uf, against x-ray energy (filtered sources). Source: Ref 2

Screen Unsharpness. In addition to the spatial resolution established by the video or image-processing equipment, the

spatial resolution of real-time systems is also affected by screen unsharpness. Screen unsharpness depends not only on the grain size and thickness of the screen but also on the detection efficiency and the energy and intensity of the radiation. Statistical fluctuations (quantum mottle) of screen brightness becomes more significant at low levels of screen brightness, and this can occur when either the radiation intensity or the detection efficiency is lowered. Nevertheless, screen unsharpness due to statistical fluctuation can be eliminated with the image-processing operation of frame summation, which makes an image representing the amount of radiation exposure during the frame-summing interval.

Projective Magnification With Microfocus X-Ray Sources. Even though the imaging systems have limits in

their spatial resolution, detail of the testpiece can be enlarged by projective magnification (see the section "Enlargement" in this article). This process of image magnification minimizes the limits of graininess and unsharpness caused by the imaging system. With microfocus x-ray sources, a resolution greater than 20 line pairs per millimeter (or a spatial resolution of 0.002 in.) can be achieved. Detail Perceptibility of Images. Studies have shown that the perceptibility of image details depends principally on

the product: MTF2 · DQE; where MTF is the modulation transfer function and DQE is the detective quantum efficiency of the imaging system. These terms, as previously mentioned in this section, are described in the article "Industrial Computed Tomography" in this Volume. Radiation Safety

The principles of radiation safety have evolved from experience in health physics, which is the study of the biological effects of ionizing radiation. There are two main aspects of safety: monitoring radiation dosage and protecting personnel. This section summarizes the major factors involved in both. More detailed information on radiation safety is available in the Selected References at the end of this article. Radiation Units. Ordinarily, x-ray and -ray radiation is measured in terms of its ionizing effect on a given quantity of atoms. The roentgen, as described in the section "Radiation Sources" in this article, is a unit derived on this basis for xrays and -rays. However, even though the amount of x-ray and -ray exposure can be quantified in the roentgen unit, the effects of radiation on the human body also depend on the amount of energy absorbed and the type of radiation. Consequently, the roentgen equivalent man (rem) unit and the SI-based sievert (Sv) unit have been developed for the purposes of radiation safety. As described below, these two units include safety factors assigned to the absorbed energy dose of a particular type of radiation (where the absorbed energy dose is measured either in the radiation absorbed dose, or rad, unit or the SI-based gray unit). Rem and Rad Units. The specification of radiation in rems is equal to the product of the rad and a factor known as the

relative biological effectiveness (rbe). The rbe (as established by the U.S. Government) is 1.0 for x-rays, -rays, or particles; 5.0 for thermal neutrons; 10 for fast neutrons; and 20 for particles. One rad represents the absorption of 100 ergs of energy per gram of irradiated material; it applies to any form of penetrating radiation. Sievert and Gray Units. A sievert is the SI-based version of the rem unit. The same rbe factors are used, but the sievert unit is derived from the absorbed dose specified in the SI-based gray (Gy) unit. Because 1 Gy represents the absorption of one joule of energy per kilogram of material, then 1 Gy = 100 rads and 1 Sv = 100 rem. Units for X-Rays and

-Rays. Because 1 R (which applies only to the ionizing effect of x-rays or -rays) corresponds to the absorption of only about 83 ergs per gram of air, the unit of 1 R can be considered equal to 0.01 Gy (1 rad) for radiation-safety measurements. With this approach, then, the measurement of x-ray and -ray dose in roentgens can be considered equivalent to direct measurement of dose in rems because the rbe is equal to one. This considerably simplifies the process of monitoring dose levels. Maximum Permissible Dose. Current practice in health physics specifies a threshold value of accumulated dose

above which an individual should not be exposed. The value of this threshold, or maximum permissible dose, increases at the rate of 0.05 Sv/yr (5 rem/yr) for all persons over the age of 18. This so-called banking concept is based on experience and assumes that no person will be exposed to penetrating radiation before the age of 18. The total amount of radiation absorbed by any individual should never exceed his continually increasing accumulated permissible dose; furthermore, no individual should be exposed to more than 0.12 Sv (12 rem) in any given year, regardless of other factors. For administrative purposes, maximum permissible dose rate is evaluated for each individual on a quarterly or weekly basis. The applicable dose rates are 0.0125 Sv (1 week.

rems) per calendar quarter, or about 1 millisievert (1 mSv, or 100 mrem) per

Radiation Protection. The two main factors in radiation protection are controlling the level of radiation exposure and

licensing the facility. The main criterion in controlling radiation exposure is the philosophy outlined in Regulatory Guide 8.10 of the United States Nuclear Regulatory Commission, which specifies the objective of reducing radiation exposures to a level as low as reasonably achievable. The licensing process involves the issuance of permission for operating a radiographic facility; permission may be required from federal, state, and local governments. The federal licensing program is concerned with those companies that use radioactive isotopes as sources. However, in some localities, state and local agencies exercise similar regulatory prerogatives. To become licensed under any of these programs, a facility or operator must show that certain minimum requirements have been met for the protection of both operating personnel and the general public from excessive levels of radiation. Although local regulations may vary in the degree and type of protection afforded, certain general principles apply to all. First, the amount of radiation that is allowed to escape from the area over which the licensee has direct and exclusive control is limited to an amount that is safe for continuous exposure. In most cases, total dose values of 0.02 mSv (2 mrem) in 1 h, 1 mSv (100 mrem) in 7 consecutive days, and 5 mSv (500 mrem) in a calendar year can be considered safe. These values correspond to those adopted by the federal government for radioactive sources and are also applicable under most local regulations. For purposes of control, unrestricted areas within a given facility are often afforded the same status as

areas outside the facility. An unrestricted area is usually defined as an area where employees are not required to wear personal radiation-monitoring devices and where there is unrestricted access by either employees or the public. Second, main work areas are shielded in accordance with applicable requirements. Some facilities must be inspected by an expert in radiation safety in order to qualify for licensing; it is usually recommended that the owner consult an expert when planning a new facility or changes in an older facility to ensure that all local requirements are met. Appropriate shielding usually consists of lead or thick concrete on all sides of the main work area. The room in which the actual exposures are made is the most heavily shielded and may be lined on all sides with steel or lead. Particular attention is given to potential paths of leakage such as access doors and passthroughs and to seemingly unimportant paths of leakage such as the nails and screws that attach lead sheets to walls and doors. Third, portable radiation sources are used in strict accordance with all regulations. In most cases, safe operation is ensured by a combination of: • • •

Movable shielding (usually lead) Restrictions on the intensity and direction of radiation emitted from the source during exposure Exclusion of all personnel from the immediate area during exposure

The best protection is afforded by distance because radiation intensity decreases in proportion to the square of the distance from the source. As long as personnel stay far enough away from the source while an exposure is being made, portable sources can be used with adequate safety. Finally, access to radiographic work areas, including field sites and radioactive-source storage areas, must at all times be under the control of competent and properly trained radiographers. Radiographers must be responsible for admitting only approved personnel into restricted work areas and must ensure that each individual admitted to a restricted work area carries appropriate devices for monitoring absorbed radiation doses. In addition to keeping records of absorbed dose for all monitored personnel, radiographers must maintain accurate and complete records of radiation levels in both restricted and adjacent unrestricted areas. Radiation Monitoring. Every safety program must be controlled to ensure that both the facility itself and all personnel

subject to radiation exposure are monitored. Facility monitoring is generally accomplished by periodically taking readings of radiation leakage during the operation of each source under maximum-exposure conditions. Calibrated instruments can be used to measure radiation dose rates at various points within the restricted area and at various points around the perimeter of the restricted area. This information, in conjunction with knowledge of normal occupancy, is used to evaluate the effectiveness of shielding and to determine maximum duty cycles for x-ray equipment. To guard against the inadvertent leakage of large amounts of radiation from a shielded work area, interlocks and alarms are often required. Basically, an interlock disconnects power to the x-ray source if an access door is opened. Alarms are connected to a separate power source, and they activate visible and/or audible signals whenever the radiation intensity exceeds a preset value, usually 0.02 mSv/h (2 mrem/h). All personnel within the restricted area must be monitored to ensure that no one absorbs excessive amounts of radiation. Devices such as pocket dosimeters and film badges are the usual means of monitoring, and often both are worn. Pocket dosimeters should be direct reading. One disadvantage of the remote-reading type is that it must be brought to a charger unit to be read. Both types are sensitive to mechanical shock, which could reduce the internal charge, thus implying that excessive radiation has been absorbed. If this should happen, the film badge should be developed immediately and the absorbed dose evaluated. Under normal conditions, pocket dosimeters are read daily, and the readings are noted in a permanent record. At less frequent intervals, but at least monthly, film badges are developed and are evaluated by comparison with a set of reference films. Because of various factors, the values of absorbed radiation indicated by dosimeters and film badges may differ. This is particularly true at low dose rates and with high radiation energies (1 MeV or more). However, the accuracy of both dosimeters and film badges increases with increasing dose, and at dose rates near the maximum allowable dose of 0.0125 Sv (1

rem) per calendar quarter, the readings usually check within ±20%.

Access Control. Permanent facilities are usually separated from unrestricted areas by shielded walls. Sometimes,

particularly in on-site radiographic inspection, access barriers may consist only of ropes. In such cases, the entire perimeter around the work area should be under continual visual surveillance by radiographic personnel. Signs that carry the international symbol for radiation must be posted around any high-radiation area. This helps to inform bystanders of the potential hazard, but should never be assumed to prevent unauthorized entry into the danger zone. In fact, no interlock, no radiation alarm, and no other safety device should be considered a substitute for constant vigilance on the part of radiographic personnel.

References cited in this section

1. Radiography & Radiation Testing, Vol 3, 2nd ed., Nondestructive Testing Handbook, American Society for Nondestructive Testing, 1985 2. R. Halmshaw, Nondestructive Testing, Edward Arnold Publishers, 1987 Radiographic Inspection Revised by the ASM Committee on Radiographic Inspection*

X-Ray Tubes X-ray tubes are electronic devices that convert electrical energy into x-rays. Typically, an x-ray tube consists of a cathode structure containing a filament and an anode structure containing a target--all within an evacuated chamber or envelope (Fig. 7). A low-voltage power supply, usually controlled by a rheostat, generates the electric current that heats the filament to incandescence. This incandescence of the filament produces an electron cloud, which is directed to the anode by a focusing system and accelerated to the anode by the high voltage applied between the cathode and the anode. Depending on the size of the focal spot achieved, x-ray tubes are sometimes classified into three groups: • • •

Conventional x-ray tubes with focal-spot sizes between 2 by 2 mm (0.08 by 0.08 in.) and 5 by 5 mm (0.2 by 0.2 in.) Minifocus tubes with focal-spot sizes in the range of 0.2 mm (0.008 in.) and 0.8 mm (0.03 in.) Microfocus tubes with focal-spot sizes in the range of 0.005 mm (0.0002 in.) and 0.05 mm (0.002 in.)

The design of conventional and microfocus x-ray tubes is discussed in the following section.

Fig. 7 Schematic of the principal components of an x-ray unit

When the accelerated electrons impinge on the target immediately beneath the focal spot, the electrons are slowed and absorbed, and both bremsstrahlung and characteristic x-rays are produced. Most of the energy in the impinging electron beam is transformed into heat, which must be dissipated. Severe restrictions are imposed on the design and selection of materials for the anode and target to ensure that structural damage from overheating does not prematurely destroy the target. Anode heating also limits the size of the focal spot. Because smaller focal spots produce sharper radiographic images, the design of the anode and target represents a compromise between maximum radiographic definition and maximum target life. In many x-ray tubes, a long, narrow, actual focal spot is projected as a roughly square effective focal spot by inclining the anode face at a small angle (usually about 20°) to the centerline of the x-ray beam, as shown in Fig. 8.

Fig. 8 Schematic of the actual and effective focal spots of an anode that is inclined at 20° to the centerline of the x-ray beam

Tube Design and Materials The cathode structure in a conventional x-ray tube incorporates a filament and a focusing cup, which surrounds the filament. The focusing cup, usually made of pure iron or pure nickel, functions as an electrostatic lens whose purpose is to direct the electron beam toward the anode. The filament, usually a coil of tungsten wire, is heated to incandescence by an electric current produced by a relatively low voltage, similar to the operation of an ordinary incandescent light bulb. At incandescence, the filament emits electrons, which are accelerated across the evacuated space between the cathode and the anode. The driving force for acceleration is a high electrical potential (voltage) between anode and cathode, which is applied during exposure. The anode usually consists of a button of the target material embedded in a mass of copper that absorbs much of the heat generated by electron collisions with the target. Tungsten is the preferred material for traditional x-ray tubes used in radiography because its high atomic number makes it an efficient emitter of x-rays and because its high melting point enables it to withstand the high temperatures of operation. Gold and platinum are also used in x-ray tubes for radiography, but targets made of these metals must be more effectively cooled than targets made of tungsten. Other materials are used, particularly at low energies, to take advantage of their characteristic radiation. Most high-intensity x-ray tubes have forced liquid cooling to dissipate the large amounts of anode heat generated during operation. Tube envelopes are constructed of glass, ceramic materials or metals, or combinations of these materials. Tube envelopes must have good structural strength at high temperatures to withstand the combined effect of forces imposed by atmospheric pressure on the evacuated chamber and radiated heat from the anode. The shape of the envelope varies with

the cathode-anode arrangement and with the maximum rated voltage of the tube. Electrical connections for the anode and cathode are fused into the walls of the envelope. Generally, these are made of metals or alloys having thermal-expansion properties that match those of the envelope material. X-ray tubes are inserted into metallic housings that contain an insulating medium such as transformer oil or an insulating gas. The main purpose of the insulated housing is to provide protection from high-voltage electrical shock. Housings usually contain quick disconnects for electrical cables from the high-voltage power supply or transformer. On selfcontained units, most of which are portable, both the x-ray tube and the high-voltage transformer are contained in a single housing, and no high-voltage cables are used. Microfocus X-Ray Tubes. Developments in vacuum technology and manufacturing processes have led to the design and manufacture of microfocus x-ray systems. Some of these systems incorporate designs that allow the opening of the tube head and the replacement of component parts, such as targets and filaments (Fig. 9). A vacuum is generated by the use of a roughing pump and turbomolecular pumps that rapidly evacuate and maintain the system to levels as low as 0.1 × 10-3 Pa (10-6 torr). Electrostatic or magnetic focusing and x-y deflection are used to provide very small focal spots and to guide the beam to various locations on the target. Focal-spot sizes are adjustable from 0.005 to 0.2 mm (0.0002 to 0.008 in.). If the target becomes pitted in one area, the beam can be deflected to a new area, extending the target life. When the target, filament, or other interior component is no longer useful in producing the desired focal-spot size or x-ray output, the tube can be opened and the component replaced at minimal cost. The tube can then be evacuated to operating levels in a few minutes or hours, depending on the length of time the tube is open to the atmosphere and the amount of contamination present.

Fig. 9 Schematic of a microfocus x-ray tube

These systems are available with voltages varying from 10 to 360 kV at beam currents from 0.01 to 2 mA. To avoid excessive pitting of the target, the beam current is varied according to the desired focal-spot size and/or kilovolt level (Fig. 10).

Fig. 10 Maximum ratings for no burn-in with 200-kV cathode/120-kV anode for electron beam widths of m (0.4 mil)

10

Microfocus x-ray systems having focal spots that approach a point source are useful in obtaining very high resolution images. A radiographic definition of 20 line pairs per millimeter (or a spatial resolution of 0.002 in.), using real-time radiography has been achieved with microfocus x-ray sources. This high degree of radiographic definition is accomplished by image enlargement, which allows the imaging of small details (see the section "Enlargement" in this article). Microfocus x-ray systems have found considerable use in the inspection of integrated circuits and other miniature electronic components. Microfocus x-ray systems with specially designed anodes as small as 13 mm (0.5 in.) in diameter and several inches long also enable an x-ray source to be placed inside otherwise inaccessible areas, such as aircraft structures and piping. The imaging medium is placed on the exterior, and this allows for the single-wall inspection of otherwise uninspectable critical components. Because of the small focal spot, the source can be close to the test area with minimal geometric unsharpness (see the section "Principles of Shadow Formation" in this article for the factors that influence geometric unsharpness). X-Ray Tube Operating Characteristics There are three important electrical characteristics of x-ray tubes: • • •

The filament current, which controls the filament temperature and in turn the quantity of electrons that are emitted The tube voltage, or anode-to-cathode potential, which controls the energy of impinging electrons and therefore the energy or penetrating power, of the x-ray beam The tube current, which is directly related to filament temperature and is usually referred to as the milliamperage of the tube

The strength, or radiation output, of the beam is approximately proportional to milliamperage, which is used as one of the variables in exposure calculations. This radiation output, or R-output, is usually expressed in roentgens per minute (or hour) at 1 m (as mentioned in the section "Radiation Sources" in this article). X-Ray Spectrum. The output of a radiographic x-ray tube is not a single-wavelength beam, but rather a spectrum of

wavelengths somewhat analogous to white light. The lower limit of wavelengths, min, in manometers, at which there is an abrupt ending of the spectrum, is inversely proportional to tube voltage, V. This corresponds to an upper limit on photon energy, Emax, which is proportional to the tube voltage, V:

Emax = aV

(Eq 4)

where a = 11 eV/volt. Figure 11 illustrates the effect of variations in tube voltage and tube current on photon energy and the intensity (number of photons). As shown in Fig. 11(a), increasing the tube voltage increases the intensity of radiation and adds higherenergy photons to the spectrum (crosshatched area, Fig. 11a). On the other hand, as shown in Fig. 11(b), increasing the tube current increases the intensity of radiation but does not affect the energy distribution.

Fig. 11 Effect of (a) tube voltage and (b) tube current on the variation of intensity with wavelength for the

bremsstrahlung spectrum of an x-ray tube. See text for discussion.

The energy of the x-rays is important because higher-energy radiation has greater penetrating capability; that is, it can penetrate through greater thickness of a given material or can penetrate denser materials than can lower-energy radiation. This effect is shown in Fig. 12, which relates the penetrating capability to tube voltage. An applied voltage of about 200 kV can penetrate steel up to 25 mm (1 in.) thick, while almost 2000 kV is needed when the thickness to be penetrated is 100 mm (4 in.).

Fig. 12 Effect of tube voltage on the penetrating capability of the resulting x-ray beam

As previously stated, most of the energy in the electron stream is converted into heat rather than into x-rays. The efficiency of conversion, expressed in terms of the percentage of electron energy that is converted into x-rays, varies with electron energy (or tube voltage), as shown in Fig. 13. At low electron energies (tube voltages of about 100 to 200 kV), conversion is only about 1%; about 99% of the energy must be dissipated from the anode as heat. However, at high electron energies (above 1 MeV), the process is much more efficient, varying from about 7% conversion efficiency at 1 MeV to 37% at 5 MeV. Therefore, in addition to producing radiation that is more penetrating, high-energy sources produce greater intensity of radiation for a given amount of electrical energy consumed.

Fig. 13 Effect of tube voltage or electron energy on the efficiency of energy conversion in the target of an x-ray source

The R-output of an x-ray tube varies with tube voltage (accelerating potential), tube current (number of electrons impinging on the target per unit time), and physical features of the individual equipment. Because of the last factor, the R-

output of an individual source also varies with position in the radiation beam, position usually being expressed as the angle relative to the central axis of the beam. Effect of Tube Voltage. Both the mean photon energy and the R-output of an x-ray source are altered by changes in

tube voltage. The effect of tube voltage on the variation of intensity (R-output) is shown in Fig. 11(a). The overall Routput varies approximately as the square root of tube voltage. The combined effect of greater photon energy and increased R-output produces, for film radiography, a decrease in exposure time of about 50% for a 10% increase in tube voltage. The effect is similar with other permanent-image recording media, as in paper radiography and xeroradiography. Effect of Tube Current. The spectral distribution of wavelengths is not altered by changes in tube current; only the R-

output (intensity) varies (Fig. 11b). Because tube current is a direct measure of the number of electrons impinging on the target per unit of time, and therefore the number of photons emitted per unit of time at each value of photon energy, Routput varies directly with tube current. This leads to the so-called reciprocity law for radiographic exposure, expressed as:

it = constant

(Eq 5)

where i is the tube current (usually expressed in milliamperes) and t is the exposure time. The reciprocity law (Eq 5) is valid for any recording medium whose response depends solely on the amount of

radiation impinging on the testpiece, regardless of the rate of impingement (radiation intensity). For example, the reciprocity law is valid for most film and paper radiography and most xeroradiography but not for fluoroscopic screens or radiometric detectors, both of which respond to radiation intensity rather than to total amount of radiation. However, radiographic films, when used with fluorescent screens, exhibit reciprocity law failure because the response of film emulsions varies with the rate of impingement of photons of visible light (screen brightness) and with total exposure. If tube current is decreased and exposure time is increased according to the reciprocity law, a fluorescent screen will emit the same amount of light (same number of photons) but at a lower level of brightness over a longer period of time. The lower brightness level will result in lower film density compared to an equivalent exposure made with a higher tube current. Deviations from reciprocity law are usually small for minor changes in tube current, causing little difficulty in resolving testpiece features or interpreting radiographs made on screen-type film. However, when the tube current is changed by a factor of four or more, there may be a 20% or more deviation from reciprocity law, and it will be necessary to compensate for the effect of screen brightness on film density. In most cases, when tube voltage is maintained constant, exposures made in accordance with the reciprocity law should produce identical film densities. This assumes that tube voltage does not vary with tube current. Heavy-duty equipment designed for stationary installations contains electrical circuitry (current stabilizers and voltage compensators) that tend to maintain R-output in accordance with the reciprocity law. However, equipment intended for on-site use is often designed to minimize weight, size, and cost and does not contain such complex circuitry. Consequently, the R-output and x-ray spectrum of many portable or transportable machines vary with tube current. These types of machines may exhibit apparent reciprocity law failure at both ends of the useful range of tube currents. For example, at currents approaching the maximum rated current, the tube voltage tends to be depressed because of the heavy electrical load on the transformer. This produces lower values of both R-output and mean photon energy than would normally be expected. Conversely, at very low tube currents, the tube voltage may actually exceed the calibrated value because the transformer is more efficient. This produces higher values of both R-output and mean photon energy than are normally expected. These deviations from expected values are not always indicated on the monitoring instruments attached to x-ray machines. Because of deviations from reciprocity law caused by equipment characteristics, it is often desirable to prepare exposure charts (usually graphs) for each x-ray unit. Such charts or graphs should be based on film exposure data at several values of tube voltage within the useful kilovoltage range of the unit. Use of these charts for determining exposure times, especially at abnormally high or low tube currents, will help to avoid unsatisfactory image quality. Effect of Electrical Waveform. In all x-ray tubes that are powered by transformer circuits, the waveform of the tube

voltage affects R-output. X-ray tubes are direct current (dc) devices, yet almost all power supplies are driven by alternating current (ac). If ac power is supplied directly to the x-ray tube, the tube itself will provide half-wave rectification. Many low-power x-ray machines use self-rectifying x-ray tubes. As illustrated in Fig. 14(a), half-wave rectification results in a sinusoidal variation of instantaneous tube voltage with time, except that no tube current flows

(and there is no effective tube voltage) during the portion of each cycle when the anode is electrically negative, and consequently there is no emission of x-rays during that portion of the cycle.

Fig. 14 Waveforms for accelerating potential (tube voltage) and tube current for four generally used types of xray-tube circuits. (a) Half-wave rectified circuit and (b) full-wave rectified circuit, in which tube voltage is equal to the peak voltage of the electrical transformer. (c) Villard-type circuit and (d) constant-potential circuit, in which tube voltage is twice the peak voltage of the transformer

In a full-wave-rectified circuit (Fig. 14b), the instantaneous tube voltage varies sinusoidally from zero to peak voltage (kVp) and back to zero, then repeats the sinusoidal pattern during the second half of each cycle. This effectively doubles the time during which x-rays are emitted on each cycle, thus doubling the R-output compared to half-wave-rectified equipment. In Villard-circuit equipment, the electrical input to the x-ray tube varies sinusoidally, but as Fig. 14(c) shows, the zero potential occurs at the minimum instantaneous tube voltage rather than at midrange. Consequently, the accelerating potential across the tube has a peak value (kVp) that is twice the peak voltage of the transformer; for example, 2-MeV electrons can be produced with transformer rated at 1 MVp. X-rays produced by Villard-circuit equipment are harder (that is, of higher mean photon energy) than x-rays produced by rectified equipment having a transformer that operates at the same peak voltage. Villard circuits exhibit a variation in instantaneous tube current that parallels the variation in instantaneous tube voltage. A modified Villard circuit, also known as a Greinacher circuit or a constant-potential circuit, produces an accelerating potential and instantaneous tube current that are nearly constant, varying only slightly with time, as shown in Fig. 14(d). The chief value of constant-potential equipment lies in the fact that its R-output is about 15% higher than for a full-waverectified unit of equivalent tube voltage. This means that a nominal 15% reduction in exposure times can be achieved by using constant-potential instead of rectified equipment. Heel Effect. X-ray tubes exhibit a detrimental feature known as the heel effect. When the direction in which x-rays are emitted from the target approaches the anode heel plane, the intensity of radiation at a given distance from the focal spot is less than the intensity of the central beam because of self-absorption by the target. Figure 15(a) shows the heel effect schematically, and Fig. 15(b) is a graph of the variation of intensity as a function of angle of emergence relative to the anode heel plane. The direction of the central beam is at 20°--for this particular tube design, an angle equal to the anode heel angle.

Fig. 15 Heel effect in a typical conventional x-ray tube having an anode face inclined 70° to the tube axis (20° anode heel angle). (a) Diagram showing relation of useful beam (crosshatched region) to tube components. (b) Graph of beam intensity (expressed as a percentage of the intensity of the central beam) as a function of angle of emergence relative to the anode heel plane

Radiographs of large-area testpieces that are made at relatively short source-to-detector distances will exhibit less photographic density (film) or less brightness (real-time) in the region where the incident radiation is less intense because of the heel effect. This can lead to errors in interpretation unless the heel effect is recognized. The consequences of heel effect can be minimized by using a source-to-detector distance that ensures that the desired area of coverage is entirely within the useful beam emanating from the x-ray tube or by making multiple exposures with a reduced area of coverage for each exposure. The heel effect also influences image sharpness when there is some geometric enlargement. The projected size of the focal spot is reduced near the "heel," which reduces the geometric unsharpness of the radiographic image nearest the heel. Conversely, the radiographic image is less sharp on the cathode side of the radiation beam. Stem Radiation. In an x-ray tube, electrons that produce the useful x-rays are known as primary electrons. The kinetic

energy of the primary electrons is converted partly to radiation and partly to heat when these electrons collide with the target. However, some of the primary electrons collide with other components inside the tube envelope, giving rise to slower-moving electrons known as secondary electrons or stray electrons. Some of these stray electrons bounce around inside the tube and eventually strike the anode. This produces x-rays, called stem radiation, that are of relatively long wavelength because of the low kinetic energy of the stray electrons that produced them. Stem radiation, which is about as intense as the useful beam, is produced all over the anode surface, not only at the focal spot. Because stem radiation is projected from a comparatively large area instead of emanating only from the focal spot, it invariably causes image unsharpness.

Some x-ray tubes are designed to eliminate this phenomenon by the use of hollow-end anodes called hooded anodes (Fig. 16). Hooded anodes reduce stem radiation both by shielding the target from some of the stray electrons and by absorbing much of the radiation produced by stray electrons that enter through the open end of the anode cavity.

Fig. 16 Schematic of an x-ray tube having a hooded anode to minimize stem radiation

Tube Rating. X-ray tubes produce a great amount of heat. At 100-kV tube voltage, only about 1% of the electrical

energy is converted to x-rays; the remaining 99% is converted to heat. Heat removal constitutes the most serious limitation on x-ray tube design. The size of the focal spot and the design of the anode are the main factors that determine the rating of a particular x-ray tube. A tube rating is a limiting or maximum allowable combination of tube voltage (kilovoltage) and tube current (milliamperage). Most industrial tubes are rated for continuous service at the maximum tube current that will give reasonable target life at maximum tube voltage. Therefore, if a tube has a continuous rating of 10 mA at 150-kV constant potential, the rating is 1500 W. Any product of kilovoltage and milliamperage equal to 1500 will not exceed the heat limit on the anode. The tube should be able to operate continuously at 20 mA and 75 kV, 30 mA and 50 kV, and so on. However, another limitation on operating characteristics is that, to sustain the same tube current, the filament must be heated to a higher temperature at low tube voltages than at higher tube voltages. Therefore, the tube current at low voltage must be limited to achieve satisfactory service life. Figure 17 shows a tube-rating chart for continuous operation in which the total length of time for each exposure is not a factor in determining operating conditions. On the left side of the chart, at low tube voltages, operating conditions are determined by filament life; on the right side of the chart, at high tube voltages, anode heating (focal-spot loading) limits operating conditions. Any combination of tube voltage and tube current below the curve can be used without impairing the service life of either anode or filament. In practice, unless there are special considerations, values approaching the continuous-duty rating are ordinarily used to minimize exposure time.

Fig. 17 Tube-rating chart for a typical beryllium-window tube with a 0.3 mm (0.012 in.) focal spot. Tube rating is for continuous operation with a full-wave-rectified single-phase power supply. Below 22.5 kVp, tube current is limited by desired filament life; above 34.5 kVp, tube current is limited by focal-spot loading. Filament voltage is 1.7 to 2.6 V, and filament current is 3.2 to 4.0 A.

Some tube manufacturers use an alternative form of rating (intermittent rating), in which the time involved in making a particular film (or paper) exposure enters into the rating. As illustrated in Fig. 18(a), for each combination of tube current and tube voltage there is a maximum exposure time that must not be exceeded. Longer exposure times can cause the target to melt. Normally, there is a minimum waiting time between exposures for tubes rated for intermittent operation. For the tube rating shown in Fig. 18(a), the waiting time is specified as 5 s when maximum exposure times are used. Some manufacturers have developed rating systems in which the minimum waiting time can be calculated.

Fig. 18 Tube-rating chart (a) and cooling chart (b) for a typical industrial x-ray tube having a 1.5 mm (0.06 in.) focal spot. Tube rating is for intermittent operation with a full-wave-rectified single-phase power supply; a minimum 5-s waiting time between exposures is specified. Cooling chart is for an oil-filled housing without an air circulator. When housing is fitted with an optional air circulator (fan), cooling times are reduced to one-half

the values determined from the chart.

Tubes that are rated for intermittent operation can accumulate only a certain amount of heat, then they must be allowed to cool before additional exposures are made. Figure 18(b) is the cooling chart used in conjunction with the tube-rating chart in Fig. 18(a). According to the cooling chart, the tube can store 500,000 units of heat (one heat unit equals 1 mA · 1 kVp · 1 s). Therefore, if a series of exposures is to be made at 70 kVp, 30 mA, and 5 s, each exposure will expend energy equal to 10,500 heat units. The number of successive film exposures that can be made, with a 5-s wait between exposures, is 47 (500,000 heat units divided by 10,500 heat units per exposure). After the 47th exposure, the tube must be allowed to cool according to the cooling curve in Fig. 18(b). The tube can be allowed to cool for any length of time, and then another exposure can be made as long as the additional exposure does not increase the stored heat above 500,000 heat units. After the tube has reached its capacity for heat storage, waiting times between successive exposures must be in accordance with the cooling curve. In this case, as shown in Fig. 18(b), 2 h would be required (or 1 h with an optional air circulator) for the tube to cool completely. Inherent Filtration. In the radiography (film or real-time) of thin or lightweight materials, which requires low-energy

radiation, filtration by the glass walls of the x-ray tube becomes a problem. Ninety-five percent of a 30-kV x-ray beam is absorbed by the glass walls of an ordinary x-ray tube. Consequently, in a tube used to radiograph thin or lightweight materials, a beryllium window is fused into the glass wall in the path of the x-ray beam. Beryllium is one of the lightest of metals and is more transparent to x-rays than any other metal. The beryllium-window tube has a minimum of inherent filtration and allows most of the very low energy x-rays to escape from the tube, as shown by the comparison with an ordinary oil-insulated glass tube in Fig. 19. The results are quite noticeable with both film and real-time radiography, particularly in contrast improvement. This effect is sometimes noticeable even beyond 150 kV. Very costly 250- or 300kV tubes with beryllium windows are sometimes used when thin, light materials as well as thick, dense materials must be inspected with the same apparatus. In beryllium-window tubes designed for very low voltage work, the window can be as thin as 0.25 mm (0.010 in.). Consequently, to avoid rupture due to external pressure on the evacuated tube, thin beryllium windows are also of small diameter; this restricts the size of the x-ray beam and the corresponding field diameter.

Fig. 19 Comparison of the R-output of a beryllium-window tube with that of an ordinary oil-insulated glass xray tube

In general, the higher the tube voltage, the larger the tube must be, and consequently the thicker the glass walls must be to support the internal elements and to withstand external atmospheric pressure. Furthermore, as the tube voltage increases, there is more likely to be high-voltage arcing from the tube to various parts of the housing. High-voltage tubes are generally surrounded by oil, which not only insulates the tube from the housing but also is often the primary means of extracting heat from the anode. In some tube assemblies, the oil surrounding the tube may be 50 mm (2 in.) thick or more. Both the oil surrounding the tube and the window in the housing that allows the useful beam of radiation to escape act as filters, particularly affecting the long-wavelength portion of the emitted radiation. Above about 200 kV, inherent filtration by components of the tube assembly is less important, primarily because inherent filtration by the testpiece would not allow the long-wavelength portion of the spectrum to reach the film anyway.

Radiographic Inspection Revised by the ASM Committee on Radiographic Inspection*

High-Energy X-Ray Sources Above about 400 kV, the conventional design of an x-ray tube and its high-voltage iron-core transformer becomes cumber-some and large. Although x-ray machines with iron-core transformers have been built for 600 kV (maximum), there are no commercial versions operating above 500 kV. For higher-energy x-rays, other designs are used. Some of the machine designs for the production of high-energy x-rays include: • • • •

Linear accelerators Betatrons Van de Graaff generators X-ray tubes with a resonant transformer

Linear accelerators, which produce high-velocity electrons by means of radio-frequency energy coupled to a waveguide, have extended industrial radiography to about 25-MeV photon energy. Betatrons, which accelerate electrons by magnetic induction, are used to produce 20- to 30-MeV x-rays. Portable linear accelerators and betatrons are also used in field inspection. The electron energy of portable units is of the order of 1.5 MeV for portable linear accelerators and 2 to 6 MeV for portable betatrons. Van de Graaff generators and x-ray tubes with resonant transformers are less widely used in industrial radiography. The Van de Graaff generator is an electrostatic device that operates from 500 kV to about 6 MV. X-ray tubes with resonant transformers were developed in the 1940s, and some units are still in operation. The output of these units is limited to about 4000 keV (4 MeV) in maximum photon energy that can be produced. Above these energies, the efficiency of resonant transformers is somewhat impaired. The radiation spectrum is broad, containing a large amount of radiation produced at much lower energies than the maximum. In addition, the accelerating potential fluctuates, which makes focusing of the electron beam difficult and creates a focal spot that is much larger than desired for optimum radiographic definition. In terms of penetrating capability, expressed as the range of steel thickness that can be satisfactorily inspected, Table 2 compares high-energy sources with conventional x-ray tubes. The maximum values in Table 2 represent the thicknesses of steel that can be routinely inspected using exposures of several minutes' duration and with medium-speed film. Thicker sections can be inspected at each value of potential by using faster films and long exposure times, but for routine work the use of higher-energy x-rays is more practical. Sections thinner than the minimum thicknesses given in Table 2 can be penetrated, but radiographic contrast is not optimum. Table 2 Penetration ranges of x-ray tubes and high-energy sources Maximum accelerating potential

Penetration range in steel

mm

in.

X-ray tubes

150 kV

Up to 15 Up to

250 kV

Up to 40 Up to 1

400 kV

Up to 65 Up to 2

1000 kV (1 MV)

5-90 -3

High-energy sources

2.0 MeV

5-250 -10

4.5 MeV

25-300

7.5 MeV

60-460

1-12

2

20.0 MeV

75-610

-18

3-24

R-Output of High-Energy Sources. The R-output of a pulsed x-ray generator such as a linear accelerator or betatron depends on several factors, including pulse length and frequency of pulse repetition, although the output of x-ray generators of similar type and design varies approximately as V2.7. For high-energy sources, electron beam current generally is not a satisfactory measure of R-output, and comparisons of beam current with the tube current of conventional x-ray tubes have little meaning. In fact, even the measurement of R-output is complicated and difficult, especially for energies exceeding 3 MeV. It is often more realistic to estimate the output of high-energy sources from film exposure data.

In terms of effective output, linear accelerators have the highest beam strength; machines with estimated outputs as high as 25,000 RMM have been built. Betatrons, although capable of producing higher-energy radiation than linear accelerators, are limited in beam strength to about 500 RMM. By contrast, 200 RMM is about the maximum output of conventional rectified, Villard-circuit or constant-potential equipment, of resonant-transformer-based machines, or of machines powered by an electrostatic Van de Graaff generator. Portable betatrons and linear accelerators have an Routput in the range of 1 to 3 RMM. With most high-energy sources (in which the x-ray beam is emitted from the back side of the target in approximately the same direction of travel as the impinging stream of electrons), the combined effects of self-absorption and electron scattering within the target itself produce a variation in intensity with angle relative to the axis of the central beam that parallels the heel effect in conventional x-ray tubes. The magnitude of the effect varies with electron energy and is extremely pronounced at energies exceeding 2 MeV. The variation of R-output with angle of emergence, shown in Fig. 20 for three values of electron energy, is of considerable practical importance because this effect can severely limit the area of coverage, especially for short source-to-detector distances. The calculated angles between the central beam and the focus of points of half intensity (that is, where the intensity is one-half that of the central beam), as well as the corresponding field diameters at 1 m (3.3 ft), are given in Table 3 for several values of electron energy, assuming normal incidence of the electrons on the target. Table 3 Half intensity and field diameter at various electron energy values Electron energy, MeV

Angle to half intensity, degrees

Field diameter at 1 m (3.3 ft)

mm

in.

5

11

380

15.0

10

7

244

9.6

18

5

174

6.9

31

3.4

119

4.7

50

1.9

66

2.6

Fig. 20 Variation of R-output with angle of emergence relative to the central beam of a high-energy source. Curves were calculated for x-rays emitted at three levels of electron energy from back side of tungsten target, assuming electron impingement normal to target surface.

As Table 3 indicates, the useful field produced by a parallel stream of electrons with energy above 10 MeV is inconveniently small for most applications. Fortunately, many high-energy sources employ some form of device to focus the electron beam on the target so that a converging beam rather than a parallel beam of electrons strikes the target. Originally intended to achieve a reduced focal-spot size, beam convergence has the secondary effect of increasing the angle to half intensity, thus increasing the effective field at any focal distance. For most high-energy sources, the effect of electron beam convergence is to at least double the field diameter. In addition to electron beam convergence, almost all high-energy sources are equipped with a field compensator, which further increases the effective field size. A field compensator is a metallic filter that is thicker in the center than it is at the edges. By progressively absorbing less radiation with increasing angle of emergence relative to the central beam, a field compensator can almost double the effective field size. The intensity of the entire beam of radiation is reduced by a field compensator, but even a reduction in central beam intensity of about 35% does not seriously increase film exposure times for most applications. In fact, the increase in effective field size usually more than compensates for the loss of intensity.

There is one type of application in which the inherent variation in intensity with angle of emergence is beneficial. A solid sphere or solid cylinder varies continuously in thickness across a diametral plane. When incident radiation from a uniform-intensity source penetrates a solid sphere or cylinder parallel to a diametral plane, there is a progressive variation in transmitted intensity, because of thickness variations, which leads to underexposure of the central area of the radiograph and overexposure along the edges, often resulting in unsatisfactory definition. With a nonuniform radiation field, however, inherent variations in penetrated thickness are at least partly counter-balanced by variations in incident intensity across the radiation field, which results in a much better resolution of testpiece detail. Radiographic Inspection Revised by the ASM Committee on Radiographic Inspection*

Gamma-Ray Sources Gamma rays are high-energy electromagnetic waves of relatively short wavelength that are emitted during the radioactive decay of both naturally occurring and artificially produced unstable isotopes. In all respects other than their origin, γ-rays and x-rays are identical. Unlike the broad-spectrum radiation produced by an x-ray tube, γ-ray sources emit one or more discrete wavelengths of radiation, each having its own characteristic photon energy. The two most common radioactive isotopes used in radiography are iridium-192 and cobalt-60. Ytterbium-169 has also gained a measure of acceptability in the radiography of thin materials and small tubes, especially boiler tubes in power plants. Cesium-137, although not used in radiography, has been useful in radiometry and computed tomography. Radiation Characteristics of

-Ray Sources. Because gamma radiation is produced by the radioactive decay of unstable atomic nuclei, there is a continuous reduction in the intensity of emitted radiation with time as more and more unstable nuclei transform to stable nuclei. This reduction follows a logarithmic law, and each radioactive isotope has a characteristic half-life, or amount of time that it takes for the intensity of emitted radiation to be reduced by one-half. The term half-life should not be misinterpreted as meaning that the intensity of emitted radiation will be zero at the end of the second half-life. Rather, the intensity remaining at the end of a second half-life will be one-half the intensity at the end of the first half-life, or one-fourth of the initial intensity. The intensity at the end of the third half-life will be one-eighth of the initial intensity, and so on.

As described in the section "Radiation Sources" in this article, the strength of γ-ray sources is the source activity given in the units of gigabecquerel or curies. If the source activity is known at any given time, a table of radioactive-decay factors for the specific radioactive isotope (or the corresponding logarithmic decay formula) can be used to calculate source activity at any subsequent time. Then, radiation intensity, which is usually measured in roentgens per hour, can be found by multiplying source activity by the specific -ray constant given in roentgens per hour per gigabecquerel (or curie). Although radiation output is constant for a given isotope, large-size sources emit slightly less radiation per gigabecquerel (curie) than small-size sources; that is, large-size sources have somewhat lower effective output. The difference between calculated radiation output and effective radiation output is the result of self-absorption within the radioactive mass, in which -rays resulting from decay of atoms in the center of the mass are partly absorbed by the mass itself before they can escape. Most source manufacturers measure source activity in units of effective gigabecquerel (or curies), which is the net of self-absorption. Specific activity, commonly expressed as gigabecquerels (curies) per gram or gigabecquerels (curies) per cubic centimeter, is a characteristic of radioactive sources that expresses their degree of concentration. A source that has a high specific activity will be smaller than another source of the same isotope and activity that has a lower specific activity. Generally, sources of high specific activity are more desirable because they have lower self-absorption and provide less geometric unsharpness in the radiographs they produce than sources of low specific activity. The more important characteristics of γ-ray sources are summarized in Table 4. Source activity and specific activity are not listed, because these characteristics vary according to the physical size of the source, the material and design of its encapsulation, and the degree of concentration at the time the source was originally produced.

Table 4 Characteristics of

-ray sources used in industrial radiography

γ-ray source

Half-life

Photon energy, MeV

Radiation output, RHM/Ci(a)

Thulium-170

128 days

0.054 and 0.084(b)

0.003

Penetrating power, mm (in.) of steel

13 (

)

Iridium-192

74 days

12 rays from 0.21-0.61

0.48

75 (3)

Cesium-137

33 years

0.66

0.32

75 (3)

Cobalt-60

5.3 years

1.17 and 1.33

1.3

230 (9)

(a) Output for typical unshielded, encapsulated sources; RHM/Ci, roentgens per hour at 1 m per curie.

(b) Against strong background of higher-MeV radiation.

-Ray Sources. Radioactive sources, which are usually metallic but may be salts or gases adsorbed on a block of charcoal, are encapsulated in a protective covering, usually a thin stainless steel sheath or a thicker sheath of aluminum. Encapsulation prevents abrasion, spillage, or leakage of the radioactive material; reduces the likelihood of loss or accidental mishandling; and provides a convenient means by which wires, rods, or other handling devices can be attached to the source.

Physical Characteristics of

Gamma-ray sources are housed in protective containers made of lead, depleted uranium, or other dense materials that absorb the -rays, thus providing protection from exposure to the radiation. Two basic types of containers are generally used. One type incorporates an inner rotating cylinder that contains the isotope and is rotated toward a conical opening in the outer cylinder to expose the source and allow the escape of radiation. This type is often referred to as a radioisotope camera. The second type incorporates a remote-controlled mechanical or pneumatic positioner that moves the encapsulated radioactive source out of the container and into a predetermined position where it remains until exposure is completed; the source is then returned to the protective container, again by remote control. This type is more versatile than the first type and is much more extensively used. Remote control of the positioner allows the operator to remain at a safe distance from the source while manipulating the capsule out of and into the protective container. Radiographic Inspection Revised by the ASM Committee on Radiographic Inspection*

Attenuation of Electromagnetic Radiation X-rays and -rays interact with any substance, even gases such as air, as the rays pass through the substance. It is this interaction that enables parts to be inspected by differential attenuation of radiation and that enables differences in the intensity of radiation to be detected and recorded. Both these effects are essential to the radiographic process. The attenuation characteristics of materials vary with the type, intensity, and energy of the radiation and with the density and atomic structure of the material.

The attenuation of electromagnetic radiation is a complex process. Because of their electromagnetic properties, x-rays and -rays are affected by the electric fields surrounding atoms and their nuclei. It is chiefly interaction with these electric fields that causes the intensity of electromagnetic radiation to be reduced as it passes through any material. The intensity of radiation varies exponentially with the thickness of homogeneous material through which it passes. This behavior is expressed as:

I = I0 exp (- t)

(Eq 6)

where I is the intensity of the emergent radiation, I0 is the initial intensity, t is the thickness of homogeneous material, and is a characteristic of the material known as the linear absorption coefficient. The coefficient μ is constant for a given situation, but varies with the material and with the photon energy of the radiation. The units of μ are reciprocal length (for example, cm-1). The absorption coefficient of a material is sometimes expressed as a mass absorption coefficient (μ/ρ), where ρ is the density of the material. Alternatively, the absorption coefficient can be expressed as an effective absorbing area of a single atom; this property, called the atomic absorption coefficient (μa) or cross section, is equal to the linear absorption coefficient divided by the number of atoms per unit volume. The cross section, usually expressed in barns (1 barn = 10-24 cm2), indicates the probability that a photon of radiation will interact with a given atom. Atomic Attenuation Processes Theoretically, there are four possible interactions between a photon (quantum) of electromagnetic radiation and material. Also, there are three possible results that an interaction can have on the photon. Thus, there are 12 possible combinations of interaction and result. However, only four of these have a high enough probability of occurrence to be important in the attenuation of x-rays or γ-rays. These four combinations are photoelectric effect, Rayleigh scattering, Compton scattering, and pair production. The photoelectric effect is an interaction with orbital electrons in which a photon of electromagnetic radiation is

consumed in breaking the bond between an orbital electron and its atom. Energy in excess of the bond strength imparts kinetic energy to the electron. The photoelectric effect generally decreases with increasing photon energy, E, as E-3.5, except that, at energies corresponding to the binding energies of electrons to the various orbital shells in the atom, there are abrupt increases in absorption. These abrupt increases are called absorption edges and are given letter designations corresponding to the electron shells with which they are associated. At photon energies exceeding an absorption edge, the photoelectric effect again diminishes with increasing energy. For elements of low atomic number, the photoelectric effect is negligible at photon energies exceeding about 100 keV. However, the photoelectric effect varies with the fourth to fifth power of atomic number; thus, for elements of high atomic number, the effect accounts for an appreciable portion of total absorption at photon energies up to about 2 MeV. Rayleigh scattering, also known as coherent scattering, is a form of direct interaction between an incident photon and

orbital electrons of an atom in which the photon is deflected without any change in either the kinetic energy of the photon or the internal energy of the atom. In addition, no electrons are released from the atom. The angle between the path of the deflected photon and that of the incident radiation varies inversely with photon energy, being high for low-energy photons and low for high-energy photons. There is a characteristic photon energy, which varies with atomic number, above which Rayleigh scattering is entirely in the forward direction and no attenuation of the incident beam can be detected. Rayleigh scattering is most important for elements of high atomic number and low photon energies. However, Rayleigh scattering never accounts for more than about 20% of the total attenuation. Compton scattering is a form of direct interaction between an incident photon and an orbital electron in which the electron is ejected from the atom and only a portion of the kinetic energy of the photon is consumed. The photon is scattered incoherently, emerging in a direction that is different from the direction of incident radiation and emerging with reduced energy and a correspondingly lower wavelength. The relationship of the intensity of the scattered beam to the intensity of the incident beam, scattering angle, and photon energy in the incident beam is complex, yet is amenable to

theoretical evaluation. Compton scattering varies directly with the atomic number of the scattering element and varies approximately inversely with the photon energy in the energy range of major interest. Pair production is an absorption process that creates two 0.5-MeV photons of scattered radiation for each photon of high-energy incident radiation consumed; a small amount of scattered radiation of lower energy also accompanies pair production. Pair production is more important for heavier elements; the effect varies with atomic number, Z, approximately as Z(Z + 1). The effect also varies approximately logarithmically with photon energy.

In pair production, a photon of incident electromagnetic radiation is consumed in creating an electron-positron pair that is then ejected from an atom. This effect is possible only at photon energies exceeding 1.02 MeV because, according to the theory of relativity, 0.51 MeV is consumed in the creation of the mass of each particle, electron, or positron. Any energy of the incident photon exceeding 1.02 MeV imparts kinetic energy to the pair of particles. The positron created by pair production is destroyed by interaction with another electron after a very short life. This destruction produces electromagnetic radiation, mainly in the form of two photons that travel in opposite directions, each photon having an energy of about 0.5 MeV. Most of the electrons created by pair production are absorbed by the material, producing bremsstrahlung of energy below 0.5 MeV. Total absorption is the sum of the absorption or scattering effects of the four processes. The atomic absorption

coefficient can be expressed as: a

=

pe

+

R

+

C

+

(Eq 7)

pr

where pe is the attenuation due to the photoelectric effect, Compton scattering, and pr is that due to pair production.

R

is that due to Rayleigh scattering,

C

is that due to

From calculated atomic absorption coefficients, the mass absorption coefficient can be determined by multiplying Avogadro's number, N, and dividing by the atomic weight of the element, A:

a

by

(Eq 8)

Figure 21 shows the variation of the mass absorption coefficient for uranium as a function of photon energy and indicates the amount of total absorption that is attributable to the various atomic processes over the entire range of 10 keV to 100 MeV. Tables of cross sections and absorption coefficients for 40 of the more usually encountered elements at various photon energies from 0.01 to 30 MeV are presented in Ref 1. Absorption coefficients can be evaluated for alloys, compounds, and mixtures by calculating a weighted-average mass absorption coefficient in proportion to the relative abundance of the constituent elements by weight.

Fig. 21 Calculated mass absorption coefficient for uranium as a function of photon energy (solid line) and contributions of various atomic processes (dashed lines). Rayleigh scattering is the difference between total scattering and Compton scattering.

Effective Absorption of X-Rays The preceding discussion of mass absorption and atomic absorption characteristics is based on the assumption of socalled narrow-beam geometry, in which any photon that is scattered, no matter how small the angle, is considered absorbed. Experimentally, narrow-beam geometry is approximated by interposing a strong absorber such as lead between the material being evaluated and the radiation detector, then allowing the beam of attenuated radiation to pass through a small hole in the absorber to impinge on the detector. Only total absorption can be determined experimentally; however, it has been shown that the sum of calculated absorption coefficients for the various atomic processes correlates closely with experimentally determined narrow-beam absorption coefficients for a wide variety of materials and photon energies. Broad-Beam Absorption. The evaluation of absorption coefficients under narrow-beam geometry is mainly a

convenience that simplifies theoretical calculations. Whenever there is a measurable width to the radiation beam, narrowbeam conditions no longer exist, and absorption must be evaluated under so-called broad-beam geometry. Under broadbeam geometry, photons that are scattered forward at small angles are not lost, but increase the measured intensity of the attenuated beam. Therefore, the broad-beam absorption coefficient of any material is less than the narrow-beam absorption coefficient. There are many degrees of broad-beam geometry, depending on the portion of the attenuated-beam intensity that is comprised of scattered radiation. However, broad-beam absorption characteristics are usually determined in such a manner that variations in the areas of either the radiation beam or the testpiece have no effect on the intensity of the transmitted radiation. Therefore, broad-beam absorption coefficients are useful because the ordinary radiographic arrangement (in which the film or screen is placed in close proximity to the testpiece) effectively duplicates this condition. Effect of Radiation Spectrum. The theoretical assessment of absorption characteristics is based on an assumption of

monoenergetic radiation (radiation having a single wavelength). As Fig. 11 shows, the actual output of an x-ray tube is not monoenergetic; for example, 200-kV x-rays have an upper limit of 200 keV of photon energy and are comprised of a broad spectrum of x-rays having lower photon energies. Even if an incident beam were monoenergetic, it would soon contain x-rays of lower photon energies because of Compton scattering. Therefore, in addition to the effect of broad-beam geometry, the effective absorption coefficients of materials are further modified by the multienergetic character of impinging radiation. The correction of calculated absorption coefficients for the known wavelength distribution of a beam of x-rays is time consuming but not impossible.

Empirical values of absorption coefficients based on broad-beam experiments are the values that are most useful in exposure calculations, because these values are corrected for both broad-beam geometry and the variation of intensity with photon energy in the incident radiation. Published values of calculated absorption coefficient are based on narrowbeam geometry. If these values are used to estimate radiographic exposures, the estimates may be inaccurate, especially for low-to-medium energy radiation from an x-ray tube and for relatively thick testpieces. Scattering Factor. The ratio of the intensity of scattered radiation, IS, to the intensity of direct radiation, ID, impinging on the detector is known as the scattering factor and is important in determining image quality. Scattered radiation does not reveal characteristics of the testpiece; it only reduces radiographic contrast, obscuring detail and reducing radiographic sensitivity. The scattering factor is a function of photon energy and the thickness, density, and material type (μ/ρ) of the testpiece. Figure 22 illustrates an example of experimental values for the build-up factor (1 + IS/ID) with steel.

Fig. 22 Experimental values of build-up factor for different thicknesses of steel and different x-ray energies. Source: Ref 2

Because radiation of low photon energy is scattered more than radiation of high photon energy, it is desirable to use relatively high-energy radiation for optimum radiographic subject contrast. This can be accomplished by either raising the voltage of the x-ray tube, which moves the entire bremsstrahlung spectrum to higher photon energies, or by interposing a filter between the source and the testpiece, which effectively raises photon energy by selectively absorbing or scattering the long wavelength portion of the spectrum. The optimum choice of filter material and thickness varies with different combinations of testpiece material and thickness, film or detector type, and bremsstrahlung spectrum as determined by the type of x-ray tube and tube voltage used. Also, the optimum combinations of filter material and thickness, and tube current and exposure time, must be evaluated on the basis of a compromise between image quality and costs. Generally, filters made of metals having high atomic numbers (lead, for example) produce a cleaner image because of a reduction in the amount of scattered radiation. However, such filters also cause a large reduction in the intensity of the radiation beam, which leads to greater expense because of the need for longer exposure times with films or more sensitive detector systems in real-time radiography. Also, because long wavelength radiation is absorbed preferentially, there is a reduction in the ability to image very small flaws. Raising the voltage of the x-ray tube, which is sometimes restricted by equipment limitations, does not necessarily involve a compromise between image quality and costs. In fact, high-voltage x-ray equipment is frequently capable of producing higher-quality radiographs at reduced cost. Image quality is improved primarily by reducing the proportion of the image background that is directly attributable to scattered radiation. For example, 80% or more of the photographic density of a given film may be attributed to non-image-forming scattered radiation when low-voltage (soft) x-rays are

employed. On the other hand, for a 100 mm (4 in.) thick steel testpiece, only about 40% of the photographic density is caused by scattered radiation when the incident beam is 1.3-MeV radiation and only about 15% with 20-MeV x-rays. Radiographic Equivalence. The absorption of x-rays and

-rays by various materials becomes less dependent on composition as radiation energy increases. For example, at 150 keV, 25 mm (1 in.) of lead is equivalent to 350 mm (14 in.) of steel, but at 1000 keV, 25 mm (1 in.) of lead is equivalent to only 125 mm (5 in.) of steel. Approximate radiographic absorption equivalence factors for several metals are given in Table 5. When film exposure charts are available only for certain common materials (such as steel or aluminum), exposure times for other materials can be estimated by determining the film exposure time for equal thickness of a common material from the chart, then multiplying by the radiographic equivalence factor for the actual testpiece material (see the section "Exposure Factors" in this article for detailed information on radiographic exposure). Table 5 Approximate radiographic absorption equivalence for various metals Materials

X-rays, keV

X-rays, MeV

γ-rays

50

100

150

220

450

1

2

4 to 25

Ir-192

Cs-137

Co-60

Ra

Magnesium

0.6

0.6

0.05

0.08

...

...

...

...

...

...

...

...

Aluminum

1.0

1.0

0.12

0.18

...

...

...

...

0.35

0.35

0.35

0.40

Aluminum alloy 2024

2.2

1.6

0.16

0.22

...

...

...

...

0.35

0.35

0.35

...

Titanium

...

...

0.45

0.35

...

...

...

...

...

...

...

...

Steel

...

12.0

1.0

1.0

1.0

1.0

1.0

1.0

1.0

1.0

1.0

1.0

18-8 stainless steel

...

12.0

1.0

1.0

1.0

1.0

1.0

1.0

1.0

1.0

1.0

1.0

Copper

...

18.0

1.6

1.4

1.4

...

...

1.3

1.1

1.1

1.1

1.1

Zinc

...

...

1.4

1.3

1.3

...

...

1.2

1.1

1.0

1.0

1.0

Brass(a)

...

...

1.4

1.3

1.3

1.2

1.2

1.2

1.1

1.1

1.1

1.1

Inconel alloys

...

16.0

1.4

1.3

1.3

1.3

1.3

1.3

1.3

1.3

1.3

1.3

Zirconium

...

...

2.3

2.0

...

1.0

...

...

...

...

...

...

Lead

...

...

14.0

12.0

...

5.0

2.5

3.0

4.0

3.2

2.3

2.0

Uranium

...

...

...

25.0

...

...

...

3.9

12.6

5.6

3.4

...

(a) Containing no tin or lead; absorption equivalence is greater than these values when either element is present.

References cited in this section

1. Radiography & Radiation Testing, Vol 3, 2nd ed., Nondestructive Testing Handbook, American Society for Nondestructive Testing, 1985 2. R. Halmshaw, Nondestructive Testing, Edward Arnold Publishers, 1987 Radiographic Inspection Revised by the ASM Committee on Radiographic Inspection*

Principles of Shadow Formation The image formed on a radiograph is similar to the shadow cast on a screen by an opaque object placed in a beam of light. Although the radiation used in radiography penetrates an opaque object whereas light does not, the geometric laws of shadow formation are basically the same. X-rays, -rays, and light all travel in straight lines. Straight-line propagation is the chief characteristic of radiation that enables the formation of a sharply discernible shadow. The geometric relationships among source, object, and screen determine the three main characteristics of the shadow--the degrees of enlargement, distortion, and unsharpness (Fig. 23). This theoretical unsharpness is added to the scattering from the atomic attenuation processes (photoelectric effect, Rayleigh and Compton scattering, and pair production, as mentioned in the previous section) to obtain the total unsharpness and blurring of the image.

Fig. 23 Schematic of the effect of geometric relationships on image formation with point sources and a radiation source of finite size. (a) Image enlargement, (b) image distortion, and (c) image overlap for point sources of radiation. (d) Degree of image unsharpness from a radiation source of finite size. Lo, source-toobject distance; Li, source-to-image distance; So, size of object; Si, size of image; Ug, geometric unsharpness; F, size of focal spot; t, object-to-image distance

Enlargement. The shadow of the object (testpiece) is always farther from the source than the object itself. Therefore, as

illustrated for a point source in Fig. 23(a), the dimensions of the shadow are always greater than the corresponding dimensions of the object. Mathematically, the size of the image or degree of enlargement can be calculated from the relationship:

(Eq 9)

where M is the degree of enlargement (magnification), Si is the size of the image, So is the size of the object, Li is the source-to-film or source-to-detector distance, and Lo is the source-to-object distance. Variations in the position of a given object relative to the source and recording surface affect image size. For example, when the source-to-image distance, Li, is decreased without changing the distance of the object from the image, the size of the image, Si, is increased. Alternatively, when the source-to-object distance, Lo, is increased without changing the distance of the recording surface from the source (Li constant), the size of the image is decreased. Film Radiography. The effect of enlargement is normally of little consequence in film and paper radiography, mainly

because the recording medium is placed close behind the testpiece to minimize geometric unsharpness. Even with this arrangement, however, images of portions of the testpiece farthest from the recording plane will be larger than corresponding portions of the testpiece itself. This effect is greatest for short source-to-image distances. Microradiography and Real-Time Radiography. There are certain circumstances when enlargement is

advantageous. In real-time radiography, for example, the significance of screen unsharpness can be minimized by positioning a testpiece closer to the source. Details that are otherwise invisible in a radiograph can also be enlarged enough to become visible. This technique has been used for microradiography, in which considerable enlargement renders minute details visible. Because of geometric unsharpness, however, only sources with extremely small focal spots can be used to produce enlarged radiographs. For example, focal spots approximately 5 m (0.2 mil) in diameter are used for microradiography (Fig. 24). One shortcoming of this technique may be a low level of radiation intensity, which could limit applications to the inspection of low-density materials. Nevertheless, real-time microradiography has found application in the inspection of high-quality castings and high-integrity weldments.

Fig. 24 Geometric magnification of radiographic images with a microfocus x-ray tube. (a) Greater geometric magnification and unsharpness occur when the testpiece is moved away from the detector and toward the radiation source. With a microfocus x-ray tube, however, the amount of geometric unsharpness (Eq 11) is

minimized, as shown by an integrated circuit magnified at 1.25× (b), 20× (c), and 40× (d). Courtesy of Rockwell International

Reduction of Scattering. Another advantage of placing the testpiece closer to the source is a reduction of secondary (scattered) radiation. When the magnification factor, M, in Eq 9 is greater than 3, the build-up factor, 1 + IS/ID, approaches 1 (Ref 2). Distortion. As long as the plane of a two-dimensional object and the plane of the imaging surface are parallel to each

other, the image will be undistorted regardless of the angle at which the beam of radiation impinges on the object. In addition, the degree of enlargement at different points in a given image is constant because the ratio Li/Lo is invariant. However, as shown in Fig. 23(b), if the plane of the object and the plane of the recording surface are not parallel, the image will be distorted because the degree of enlargement is different between rays passing through any two points on the surface of the object. This is the main reason radiographic images almost always present a distorted picture of testpiece features; the shape of most testpieces is such that one or more of its features is not parallel to the imaging surface. The degree of distortion is directly related to the degree of nonparallelism; small amounts of nonparallelism produce small degrees of distortion, and large amounts produce large degrees of distortion. Although there is no distortion of images of features that are parallel to the imaging surface, spatial relationships among testpiece features are distorted, with the amount of distortion being directly related to the cosine of the angle of the beam to the imaging surface. For example, two circular features that are parallel to the imaging surface but at different objectto-detector distances may produce two separate circular images, or one noncircular image because two images overlap, as shown in Fig. 23(c), depending on the direction of the radiation beam. In most actual radiographs, regions where images of different features overlap will appear as regions of greater or lesser contrast than adjacent regions where the images do not overlap, depending on whether the features absorb less radiation or more radiation than portions of the testpiece surrounding the features. For example, overlapping images of voids or cavities will appear darker on film or brighter on fluorescent screens in regions of overlap because less radiation is absorbed along paths intersecting overlapping voids than along paths that intersect only one void. Geometric Unsharpness. In reality, any radiation source is too large to be approximated by a point. Conventional x-

ray tubes have focal spots between 2 by 2 mm (0.08 by 0.08 in.) and 5 by 5 mm (0.2 by 0.2 in.) in size, while microfocus x-ray systems have focal-spot sizes as small as 5 m (0.2 mil). Even high-energy sources have focal spots of appreciable size, although seldom exceeding 2 mm (0.08 in.) in diameter. Gamma-ray sources vary widely in size, depending on source strength and specific activity, but are seldom less than about 2.5 mm (0.1 in.) in diameter. Radiographic definition varies according to the geometric relationships among source size, source-to-object distance, and object-to-image distance. When radiation from a source of any finite size produces a shadow, that portion of the image that is in shadow for radiation emanating from all points on the surface of the source is a region of complete shadow known as the umbra. Portions of the image that are in shadow for radiation emanating from any portion of the source, but not in shadow for radiation from some portion, are regions of partial shadow known as the penumbra. The degree of geometric unsharpness is equal to the width of the penumbra. Mathematically, the geometric unsharpness, Ug, is determined from the laws of similar triangles, as illustrated in Fig. 23(d), and can be expressed as:

(Eq 10)

where F is the size of the focal spot -ray source, t is the object-to-image distance, and Lo is the source-to-object distance. In most cases, t is considered to be the difference between the source-to-image distance, Li, and the source-toobject distance, Lo; therefore, Eq 10 can be alternatively expressed as:

(Eq 11) The size of the penumbra, or the geometric unsharpness, can be reduced by lengthening the source-to-object distance, by reducing the size of the focal spot, or by reducing the object-to-image distance. In practice, the source size is determined by the characteristics of a given x-ray tube or by the physical dimensions of a radioactive pellet, and the object-to-image

distance is minimized by placing the surface of the imaging medium as close as possible to the testpiece. Therefore, the only variable over which the radiographer has any appreciable amount of control is source-to-object distance. However, as discussed in the following section in this article, this distance is a significant variable affecting the intensity of the penetrating radiation. Consequently, the radiographer often must compromise between maximum definition (minimum unsharpness) and the cost associated with the lower radiation intensity that occurs when the source-to-object distance is increased. Higher costs in film radiography occur with a lower intensity of radiation because exposure time must be increased. In real-time radiography, costs may be less of a factor, but increases in detector efficiency or image-processing time might be required. Geometric unsharpness is one of several unsharpness factors and is usually the largest contributor to unsharpness in film radiography. When the distance between the actual surface of the imaging medium and the adjacent (facing) surface of the testpiece is neglected (which is an appropriate assumption in film radiography because this distance is usually quite small in relation to testpiece thickness), the geometric unsharpness can be calculated for any source size and can be expressed as a series of straightline plots relating geometric unsharpness, Ug, to testpiece thickness, t, for various values of the source-to-object distance, Lo. A typical series is shown in Fig. 25 for a 5 mm (0.2 in.) diam source. It is often helpful to prepare graphs such as the one in Fig. 25 for each source size used.

Fig. 25 Relation of geometric unsharpness to testpiece thickness for various source-to-object distances when the source is 5 mm (0.2 in.) in diameter

In applications where the maximum unsharpness must be kept below a specific known value to resolve certain types and sizes of flaws, the radiographer can determine the minimum source-to-object distance for a given part from Eq 10. At this value of Lo, the types and sizes of flaws to which the specified value of Ug applies will be resolved if the flaws are in the plane of the testpiece farthest from the imaging surface. Flaws in planes closer to the imaging surface will be even more clearly resolved. Shadow Intensity and the Inverse-Square Law. The intensity of radiation that penetrates a testpiece and

produces a satisfactory image is governed by the energy and spectral quality of the incident radiation, the strength of the radiation, and the source-to-detector distance. In practice, the energy of the incident radiation, which depends mainly on the tube voltage of an x-ray machine or on the radioactive isotope in a -ray source, is chosen to be sufficiently penetrating for the type of material and thickness to be inspected. With this factor fixed, the remaining interrelated factors--the source strength (determined by tube current in milliamperes for x-ray sources or by source activity in becquerels, or curies, for -ray sources) and the source-to-detector distance--determine the intensity of radiation impinging on the detector. The intensity of the shadow image determines the amount of exposure time for a given (fast or slow) film or the detector efficiency required in real-time radiography. Similar to visible light, x-rays or -rays diverge when emitted from the

radiation source and cover an increasingly larger area, with lessened intensity, as the distance from the source increases. The intensity, or amount of radiation falling on a unit area per unit time, varies inversely with the square of the distance from the source. This can be expressed mathematically as:

IL2 = constant

(Eq 12)

where I is the intensity of radiation at a given distance, L, from a source of given strength. More often, the inverse-square law is expressed as a ratio:

(Eq 13)

where the subscripts 1 and 2 refer to different points along a line radiating from the source. Because of this inherent characteristic of radiation, if the radiation has a certain intensity at 1 m (3.3 ft) from the source, it will have four times that intensity at 0.5 m (1.65 ft) but only one-fourth that intensity at 2 m (6.6 ft) and only one-ninth that intensity at 3 m (9.9 ft). Scattered Radiation. When a beam of x-rays or -rays strikes a testpiece, secondary radiation is scattered in all directions. This causes a haze over all or part of the image, thus reducing contrast and visibility of detail. Methods for reducing scattered radiation are discussed in the section "Control of Scattered Radiation" in this article. Internal scatter occurs within the testpiece being radiographed (Fig. 4a). Internal scatter is fairly uniform throughout a

testpiece of uniform thickness, but adversely affects definition by diffusing or obscuring the outline of the image. Side scatter consists of secondary radiation generated by walls, internal holes in the testpiece (as shown in Fig. 4b), or

objects near the testpiece. This radiation may pass through the testpiece and may obscure the image of the testpiece. Back scatter consists of rays that are generated from the floor or table (Fig. 4c) or from any other object that is located

on the opposite side of the imaging surface from the testpiece and source. Back scatter increases the background noise and therefore reduces contrast and definition in the recorded image.

Reference cited in this section

2. R. Halmshaw, Nondestructive Testing, Edward Arnold Publishers, 1987 Radiographic Inspection Revised by the ASM Committee on Radiographic Inspection*

Image Conversion Media After penetrating the testpiece, the x-rays or γ-rays pass through a medium on the imaging surface. This medium may be a recording medium (such as film) or a medium that responds to the intensity of the radiation (such as fluorescent screen or a scintillation crystal in a discrete detector). These two types of media provide the images for subsequent observation. Radiography can also utilize radiographic screens, which are pressed into intimate contact with the imaging medium. Radiographic screens include: • •

Metallic screens placed over film, paper, or the screens used in real-time systems Fluorescent screens placed over film or photographic paper

•

Fluorometallic screens placed over film

These screens are used to improve radiographic contrast by intensifying the conversion of radiation and by filtering the lower-energy radiation produced by scattering. Recording Media By definition, a radiographic recording medium provides a permanent visible image of the shadow produced by the penetrating radiation. Permanent images are recorded on x-ray film, radiographic paper, or electrostatically sensitive paper such as that used in the xeroradiographic process. X-ray film is constructed of a thin, transparent plastic support called a film base, which is usually coated on both sides (but occasionally on one side only) with an emulsion consisting mainly of grains of silver salts dispersed in gelatin (Fig. 26). These salts are very sensitive to electromagnetic radiation, especially x-rays, -rays, and visible light. The film base is usually tinted blue and is about 0.18 mm (0.007 in.) thick. An adhesive undercoat fastens the emulsion to the film base. A very thin but tough coating of gelatin called a protective overcoat covers the emulsion to protect it against minor abrasion. The total thickness of the x-ray film is approximately 0.23 mm (0.009 in.), including film base, two emulsions, two adhesive undercoats, and two protective overcoats.

Fig. 26 Schematic cross section of a typical x-ray film. Dimensions given in inches

The presence of emulsions on both sides of the base effectively increases the speed of the x-ray film because penetrating radiation affects each emulsion almost equally. In addition, thin emulsions allow development, fixing, and washing of the exposed film to be accomplished effectively and in a reasonably short time. Where two emulsions are used, two images are produced--one in the front emulsion and one in the back emulsion. Viewing these images with the unaided eye causes no problems, because the images are separated by only a few hundredths of a millimeter of film base. However, if magnification is used, parallax can cause the two images to be seen slightly separated, which sometimes hampers interpretation. When appropriate, x-ray film with only one emulsion (single-coated x-ray film) can be used to avoid parallax. When electromagnetic radiation and certain forms of particulate radiation (namely, electrons) react with the sensitive emulsion of x-ray films, they produce a latent image in the emulsion, which cannot be detected visually or by ordinary physical measurements. To produce a visible image, the exposed film must be chemically processed. When the film is treated with a developer, selective chemical reaction converts the exposed silver halide grains in the latent image into black metallic silver. The metallic silver remains suspended in the gelatin and is responsible for blackening or shades of gray in the developed (visible) image. After development, the film is treated with a chemical called a fixer, which converts unexposed silver halide grains into a water-soluble compound. The reaction products of development and fixing

are washed off, and the film is dried. Detailed information on the processing of radiographic film is available in the Appendix to this article. Although x-ray film is sensitive to light, the characteristics of its emulsion are different from those of emulsions used in photography. Industrial x-ray film is of two general types: that used for direct exposure (called direct-exposure, no-screen, or nonscreen film) and that used with fluorescent screens (often called screen-type film). Most industrial x-ray film is of the direct-exposure type and is available in various combinations of film speed, gradient, and graininess. The choice of film for a given application depends on the type of radiography to be performed. Film types and selection are discussed in the section "Film Radiography" in this article. Screen-type films, which are marketed primarily for medical radiography, are only occasionally used for industrial applications--for example, when a low-power x-ray machine is used and exposure time is excessive with direct-exposure x-ray film. Screen-type films are more sensitive to light than to x-radiation and are particularly sensitive to the wavelengths emitted by the fluorescent screen with which they are used. Although blue-sensitive emulsion types are the ones most often used for industrial radiography, other emulsions sensitive to ultraviolet and to green screen light are available. Radiographic Paper. Ordinary photographic paper can be used to record x-ray images, although its characteristics are

not always satisfactory. Photographic paper has a low speed, and the resulting image is low in contrast. However, photographic paper used with fluorescent screens can be effective for some applications. Specially designed radiographic paper (also known as paper-base film or opaque-base film) is essentially opaque and has only a single emulsion. The emulsion is made of gelatin and silver salts, as in x-ray film, but it also contains developer chemicals. After exposure, the developing action is triggered by an alkaline solution called an activator. Instead of being fixed after development of the radiographic paper, chemicals remaining in the emulsion are made chemically inactive by stabilization processing, which is performed very rapidly in a small automatic processor. The images are similar to those on film radiographs and will not deteriorate for 2 to 3 months. If it is desired to store the radiographs for a longer period, they can be fixed in ordinary x-ray fixer, washed, and dried. Although paper radiographs can be exposed directly to x-radiation, it is preferred to make the exposure with fluorescent screens that produce ultraviolet light. Fluorescent-screen exposures are preferred because exposure time is short and radiographic contrast is high. With direct exposure, the radiographic contrast is low and the corresponding exposure time is long. The image is viewed in reflected light like a photograph; however, the greatest amount of detail can be seen when the image is viewed in specular light from a mirrorlike reflector (like a spotlight) directed at about 30° to the surface and to one side of the radiograph. The density (blackening) of paper radiographs, which is known as reflection density, is quite different from the transmission density of film, but the differences do not cause problems in interpretation. Radiographic paper can exhibit adequate sensitivity, which in many respects matches or exceeds that of fast directexposure x-ray films. Radiographic paper does not match the sensitivities of slow x-ray films, but because of their speed, convenience, and low cost, radiographic papers are being used both for the radiography of materials that do not require critical examination and for in-process control. A new (3M) dry silver film has recently been introduced that is developed with heat in about 12 s. It also is exposed with a fluorescent screen and achieves resolutions up to 7 line pairs per millimeter (or a resolution of 0.0055 in.). Viewing is by reflected or transmitted light. Xeroradiography (dry radiography) is a form of imaging that uses electrostatic principles for the formation of a

radiographic image. In film radiography, a latent image is formed in the emulsion of a film; in xeroradiography, the latent image is formed on a plate coated with selenium. Before use, the plate is given an even charge of static electricity over the entire surface. As soon as the plate is charged, it becomes sensitive to light as well as to x-radiation and must be protected from light by a rigid holder similar to a film cassette. In practice, the holder is used for radiography as though it contained film. X-radiation will differentially discharge the plate according to the amount of radiation received by different areas. This forms an electrostatic latent image of the testpiece on the plate.

The exposed plate is developed by subjecting the plate, in the absence of light, to a cloud of very fine plastic powder, each particle in the cloud being charged opposite to the electrostatic charges remaining on the plate. The charged plastic powder is attracted to the residual charges on the plate. The visible radiographic image at this state is not permanent, because the powder is simply held in place by electrostatic charges. However, the image can be made permanent by placing a piece of specially treated paper over the plate and transferring the powder to the paper, which is then heated to fix the powder in place. Selenium-coated plates can be easily damaged by fingerprints, dirt, and abrasion. For this reason, automated equipment is used for charging and for development and image transfer to paper. The image produced by xeroradiography can be either positive or negative, depending on the color of the plastic powder (toner) and the color of the paper upon which the image is deposited. If a negative xeroradiographic image is desired, dark-colored paper with light-colored toner can be used; a positive image would require light-colored paper and darkcolored toner. Because of an electrostatic edge effect, xeroradiographic images exhibit excellent detail; even small indications in the image are sharply outlined. Optimum results from xeroradiography are more likely when the user develops special techniques rather than employing the radiographic techniques normally used with film. Lead Screens It is the combination of filtration and intensification that makes lead screens the most widely used in industrial radiography. Lead absorbs radiation to a greater extent than most other materials, with the amount of absorption depending largely on the penetrating quality of the radiation (photon energy or wavelength). High-energy (shortwavelength) radiation passes through lead much more readily than low-energy radiation; in other words, low-energy radiation is more readily absorbed by a lead screen than high-energy radiation. Because scattered radiation from a testpiece is always of a lower energy than the incident beam passing through a testpiece, a lead screen will absorb a relatively high percentage of unwanted scattered radiation, but will absorb a somewhat lower percentage of the imageforming radiation. This effect is known as filtration, and lead screens are sometimes referred to as lead-filter screens. Filtration of Secondary Radiation. As discussed in the earlier section "Principles of Shadow Formation" in this article, secondary radiation can arise by internal scatter from within the testpiece and by back scatter from objects behind the film, detector, or screen. Because of the need to filter out both internal scatter and back scatter, two screens are normally used in film radiography. The screen that faces the top of the film is referred to as the front screen, and the screen behind the film toward the table or floor is referred to as the back screen. Both screens absorb scattered radiation.

In practice, the front screen is sometimes the thinner of the two because the image-forming radiation must always pass through this screen. The usual thickness of the front lead screen is 0.13 or 0.25 mm (0.005 to 0.010 in.), but may be greater or less depending on the type of material being inspected and the photon energy of the incident beam. The main function of the beam screen is to absorb unwanted back scatter; back screens can be of any thickness that performs this function adequately, although usually they are of the same thickness as the front screen. In the radiography of thin-gage or low-density materials, in which low photon energies are used, care must be exercised to ensure that the front screen does not excessively filter the image-forming radiation. Excessive filtration affects subject contrast and tends to reduce radiographic sensitivity. In such cases, lead screens less than 0.13 mm (0.005 in.) thick should be used; lead screens 0.05 or 0.025 mm (0.002 or 0.001 in.) thick or less are available. For thicker testpieces or denser material, where higher photon energies are used, the front screen can be thicker than 0.13 mm (0.005 in.), ranging as high as 1 mm (0.040 in.) when betatrons or high-MeV linear accelerators are used. The back screen can be 0.5 to 10 mm (0.020 to 0.400 in.) thick in these cases. Intensification. When lead is excited by x-ray or

-ray radiation, it produces electrons. The number of electrons emitted is directly proportional to the photon energy of the radiation that passes through the testpiece and reaches the screens. In film radiography, the emitted electrons expose additional silver halide crystals in the film emulsion. After development, film densities are greater than they would have been without the intensifying action of emitted electrons. Intensification not only increases overall photographic density, thus requiring shorter exposure times for producing a given density, but also enhances radiographic contrast, thus improving the ability to resolve small flaws.

In real-time radiography with a fluorescent-screen system, a similar effect occurs. The phosphor material used in fluorescent screens is generally more sensitive to electrons than to primary x-rays. With high-energy x-rays, the electrons generated from a suitable lead screen (or other heavy metal, such as tantalum or tungsten) can be used to enhance the imaging process. These screens are often useful in real-time radiography with MeV radiation. Because lead screens have properties of both filtration and intensification, there is a combination of thickness of the subject material and photon energy of the incident radiation at which intensification just balances filtration and there is no net advantage. With steel testpieces in film radiography, this null point is generally considered to occur at a combination of 6 mm ( in.) testpiece thickness and 140-keV x-rays when a 0.13 mm (0.005 in.) thick front screen and a 0.25 mm (0.010 in.) thick back screen are used. At lower tube voltages or with thinner testpieces, the filtration effect is dominant, resulting in longer exposure times. At high voltages or with thicker testpieces, the intensification factor becomes dominant, and exposure times can be reduced to as much as one-third of the time for direct exposure (no screens), at a tube voltage of 200 to 300 kV. With cobalt-60 radiation and steel testpieces, the exposure time using lead screens is about one-third that for direct exposure. With light metals such as aluminum, the null point occurs at a greater thickness than for steel. With metals of greater atomic number than steel, it occurs in thinner sections. For both lighter and heavier metals, the null point will occur at a radiation energy different from 140-keV x-rays. Although electrons are produced throughout the volume of lead on both screens when excited by x-rays or -rays, electrons produced by radiation having photon energies below 1 MeV are largely low-energy electrons. Such electrons are readily absorbed by the volume of lead in the screen. Only those electrons produced at the surface adjacent to the film escape to intensify the latent image on the film. Therefore, the closer the film emulsion is to the surface of the lead, the more effectively the electrons interact with the emulsion. This is why lead screens should always be in intimate contact with the film. Low-energy electrons have little penetrating capability. They will affect the emulsion closest to the screen, but will not penetrate the film base to affect the emulsion on the other side. Although electrons have little penetrating capability, they will penetrate interleaving paper, so this should always be removed from the film to avoid a paper-pattern image on the radiograph. Similarly, dust, dirt, lint, and other foreign material between the film and screen must be avoided to prevent extraneous images (artifacts) on the radiograph. Precautions. Lead sheet of usual screen thicknesses is easily bent. For this reason, lead screens are often backed by

cardboard or other material to facilitate handling. Even so, the care and handling of lead screens is important. Deep scratches or dents must be avoided because they can appear as artifacts on the radiographic image. Wrinkles or folds in the screens can also be detected on the radiograph. Chemical spills, dust, and dirt must be carefully removed from the screen before use. Oxidation of the screen, which occurs with age and appears as a gray coating, does not seem to affect the utility of the screen. Some screens are coated with a special material to prevent oxidation. Although coated screens exhibit a reduced capability for intensification, they are easier to keep clean and free of tiny scratches. The use of spray lacquers or acrylics to protect the surface of the screens should be avoided; such coatings often have a highly detrimental effect on the radiographic image. The composition of the lead used for screens is important. Pure lead is soft and may rub off on the film to produce lead smudge on the radiograph. Lead screens made from 94Pb-6Sb alloy are most commonly used because they are harder and more resistant to scratching. However, care must be taken that there is no segregation of antimony (which appears as shiny or different-colored streaks on the screen), because antimony segregation produces low-density streaks on the radiograph. Screens must be flat and free of roll marks or chatter. Variations in screen thickness result in areas of poor contact with the film, which can produce fuzzy areas in the radiograph. Prolonged contact with lead screens can produce an effect called lead-screen fog. Consequently, films should never be left in contact with lead screens longer than is reasonably necessary. This is particularly important under conditions of high temperature and humidity, or within 24 h after cleaning screens with very fine steel wool or other abrasive. As a general rule, lead screens should be used whenever they can improve the resolution of detail, even though the exposure time may be longer. A single back screen can be used for intensification purposes, but in the absence of a front screen, forward scatter may be present in the radiograph. Sometimes, the single back screen technique is useful in the radiography of very thin materials at low photon energies. If scatter is a problem, other means of control (such as a copper filter located at the tube) are quite effective. Lead screens should always be used in radiography using high photon

energies (above 300-keV x-rays or with most irradiation of the film holder.

-ray sources) to avoid extraneous paper patterns or other effects due to

Certain types of film packaging incorporate thin lead foil (0.025 or 0.050 mm, or 0.001 or 0.002 in. in thickness) on both sides of the film, with the film-foil composite contained in a lighttight envelope. These prepackaged composites are convenient because the envelopes containing screens and film can be used as a film holder or cassette. In addition, loading in the darkroom is eliminated, and there is less likelihood of foreign particles being present to create artifacts on the radiograph. Lead Oxide Screens A variation of lead screens is lead oxide screens, which are made by evenly coating a paper base with lead oxide. The result is an extremely flexible screen, equivalent to about 0.013 mm (0.0005 in.) of lead foil, that is lightweight and free of antimony segregation. Lead oxide screens are available only in lighttight envelopes containing screens on both sides of the film. The principle of lead oxide screens is essentially the same as for lead screens. The main differences are a lesser degree of filtration than lead screens and a greater intensification below 140 keV as well as slightly lesser intensification at 300 keV than with 0.13 mm (0.005 in.) thick lead screens. Although lead oxide screens are particularly advantageous for the radiography of light alloys or thin material, they can also be used in the radiography of heavier material. Screens of Metals Other Than Lead Foils of many of the heavier metals can be as effective as lead for radiographic screens, but usually are not as practical, either because the foils are not as flexible or because of cost. Gold screens perform as well or better than lead screens, but gold costs considerably more and tends to work harden

when bent, which eventually cracks the screen. Tantalum screens exhibit slightly lower filtration but higher intensification than lead. However, tantalum foil is stiff

and springy, which does not allow it to be shaped to fit around a testpiece. Tantalum foil can be used in solid cassettes, can be polished, and is quite resistant to minor abrasion, but is expensive to use. Depleted-uranium screens exhibit greater filtration than lead screens. However, uranium is brittle, is somewhat

difficult to obtain in thin sheets, and is also expensive to use. Copper screens have been used, especially with cobalt-60 radiography. Copper has a lower degree of filtration than

lead and a lower intensification factor, but copper provides greater radiographic sensitivity. Composite screens of lead, copper, and aluminum (and sometimes other metal foils) to control the radiation reaching

the film have occasionally been used with supervoltage x-ray machines for the inspection of thick steel testpieces. Fluorescent Intensifying Screens The efficiency of film and paper radiography can be improved by fluorescent intensifying screens, which emit radiation in the ultraviolet, blue, or green portion of the electromagnetic spectrum. These screens are called fluorescent intensifying screens because they fluoresce, or produce light, when excited by x-rays or -rays. Certain compounds, such as calcium tungstate or barium lead sulfate, often containing trace elements of some other chemical or phosphor, have the property of emitting light immediately upon excitation by short-wavelength radiation. Crystals of these chemicals are finely powdered, mixed with a binder, and coated on some mildly flexible support such as cardboard or plastic to make a fluorescent screen. A thin, tough, transparent overcoat is applied to the sensitized surface of the screen to prevent damage to the crystals during use. Screen Speeds. Fluorescent intensifying screens, widely used in medical radiography, are available in a variety of

speeds. Perhaps the most common of these are blue-emitting screens, which can be characterized in speed as very slow, slow, medium, medium high, high, and super. In industrial radiography, the screen most often used (with the appropriate blue-sensitive screen-type film) is the medium-speed fluorescent screen.

By the addition of proprietary rare-earth phosphors to the fluorescent crystals, green-emitting screens are produced that, when used with the appropriate green-sensitive film, have relatively fast screen speeds characterized as high speed, superspeed, and beyond superspeed. With the addition of barium lead sulfate and a proprietary trace element, the crystals can be made to fluoresce in the ultraviolet range. Used with the appropriate ultraviolet-sensitive film or radiographic paper, these screens have speeds that are equivalent to the slow and medium-high-speed blue-emitting screens. For the greatest radiographic effect, fluorescent screens should be used with a film that is sensitive to the particular wavelengths of light emitted by the screen. In general, though not always true, the slower the screen-film combination, the better the radiographic definition. However, even slow-speed combinations can reduce exposure times by 20 to 98% compared to the fastest direct-exposure film, depending on photon energy, type of screen, type of film, and testpiece material. Disadvantages. The main reason for using fluorescent intensifying screens in industrial radiography is to reduce

exposure times. However, in comparison with lead screens, fluorescent screens have the following disadvantages. First, the screen light reaches the film as cones of light from each fluorescing crystal. Because this tends to blur the image, radiographic definition using lead screens is superior. As in real-time radiography, inherent unsharpness in a fluorescent screen often overshadows any effect of geometric unsharpness. Second, the purely statistical variations in the number of x-ray quanta (the total amount of x-radiation) absorbed from one small area to another on the screen result in uneven brightness, which in turn is recorded on the film as screen mottle (also known as quantum mottle). Screen mottle is particularly apparent at photon energies in the range 150 to 300 keV. There is essentially no screen mottle associated with lead screens. Finally, lead screens filter out scattered radiation in addition to intensifying the radiation beam, but fluorescent screens have little or no filtering capability and intensify scattered radiation in the same proportion as transmitted radiation. However, they can be placed between lead screens, using the lead to protect against secondary radiation and back scatter. Precautions in Use. The combination of a screen and the film that is sensitive to the wavelength of the screen light

results in minimum exposures. However, fluorescent screens can be used with the direct-exposure films ordinarily used in industrial radiography. With direct-exposure film, the intensification factor may range from 2 to 25 times, but screen mottle and noticeable light spreading will be less because of the lower speed of the screen-film combination. Fluorescent intensifying screens are often permanently mounted inside rigid film holders (cassettes). Cassettes of this type are designed to provide even pressure between screen and film to maintain intimate contact during exposure. Lack of contact results in unsharpness in the image and should be carefully avoided. Fluorescent screens may not be permanently mounted and can be used in regular film holders. Under these conditions, however, maintenance of good contact between fluorescent screen and film is more difficult. The fluorescent crystals on the surface of the screens can be easily damaged by handling, and the screen cannot be repaired. Because fluorescent screens are expensive, great care must be taken to avoid scratches, abrasion, or chemical spills, which can damage the screen and affect the radiographic image. It is even more important to avoid fingerprints, stains, grease, dirt, and dust on the surface of the screens. Because these all absorb screen light, their image will be transmitted to the radiograph. If the screen becomes dirty it may be cleaned, but the manufacturer's cleaning recommendations should be followed to avoid damage to the protective overcoat. Because of the high radiation intensity used in industrial radiography, great care must be exercised not to expose any portion of a fluorescent screen to the full intensity of the primary beam. When exposed to high-intensity radiation, a screen may continue to fluoresce after being removed from the beam; this semipermanent effect, known as screen lag, will produce false images on subsequent radiographs. Fluorescent intensifying screens are most often used for the radiography of thick testpieces when the x-ray machine has limited penetrating capability. With medium-speed, blue-sensitive fluorescent intensifying screens (with appropriate screen-type film), it is possible to radiograph 75 mm (3 in.) of steel with 250-keV x-rays in a reasonable exposure time. Fluorescent screens are rarely used with -rays or with photon energies exceeding 1 MeV because of screen mottle and low intensification factors. In all cases where fluorescent screens are used, reciprocity law failure is likely to occur;

reciprocity law failure (discussed in the section "X-Ray Tubes" in this article) is not a problem in direct or lead-screen exposures. Fluorometallic Screens Fluorometallic screens consist basically of lead screens placed on either side of a fluorescent-screen/film combination. The theory behind combining lead and fluorescent screens is that the lead preferentially absorbs scattered radiation (which fluorescent screens will not do) and that the fluorescent screen next to the film intensifies the radiation beam to shorten the exposure. Commercial screens incorporating lead and fluorescent materials laminated together require significantly shorter exposure times than lead screens, and the image quality exceeds that of fluorescent screens. When direct-exposure film is used with fluorometallic screens, the exposure time can be reduced to as little as one-seventh of that for lead screens with almost the same radiographic sensitivity. When screen-type film is used, exposure time can be reduced to as little as one-ninth of that for other methods. The speed factor varies with type of film, type of fluorometallic screen, and photon energy of radiation. With high-voltage x-rays, and with cobalt-60 -rays, the speed factor is about two to seven. Radiographic Inspection Revised by the ASM Committee on Radiographic Inspection*

Real-Time Radiography Various types of image conversion techniques allow the viewing of radiographic images while the testpiece is being irradiated and moved with respect to the radiation source and the radiation detector. These radioscopic techniques can be classified as real-time radiography (also known as real-time radioscopy) and near real-time radiography (or near real-time radioscopy). The distinction between these two classifications is that the formation of near real-time images occurs after a time delay, and this requires limitation of the test object motion. An example of near real-time radiographic imaging involves the use of discrete detectors (primarily linear arrays) that scan the area being irradiated. The outputs are then processed digitally to form images in near real time. This technique is often referred to as digital radiography. Background. The predecessor of the modern methods of real-time radiography is fluoroscopy. This technique, which is

now largely obsolete, involves the projection of radiographic images on a fluorescent screen (Fig. 27). The screen consists of fluorescent crystals, which emit light in proportion to the intensity of the impinging radiation. The radiographic image can then be viewed, with appropriate measures taken to protect the viewer from radiation.

Fig. 27 Diagram of the components and principles of operation of a fluoroscope

The main problem with fluoroscopes is the low level of light output from the fluorescent screen. This requires the suppression of background light and about 30 min for the viewer's eyes to become acclimated. Moreover, radiation safety dictates viewing through leaded glass or indirectly by mirrors. Because of these limitations, other methods have been developed to improve safety and to amplify the images from fluorescent screens. One of the early systems (1950s) involved the development of image-intensifier tubes. Image-intensifier tubes (Fig. 28) are glass-enclosed vacuum devices that convert a low-intensity x-ray image or a low-brightness fluorescent-screen image into a high-brightness visible-light image. Image intensification is achieved by a combination of electronic amplification and image minification. The image brightness at the output window of an image-intensifier tube is about 0.3 × 103 cd/m2 (10-1 lambert) as compared to about 0.3 cd/m2 (10-4 lambert) for conventional fluoroscopic screen. These early image intensifiers were originally developed for medical purposes and were limited to applications with lowenergy radiation (because of low detection efficiencies at high energies). Consequently, industrial radiography with these devices was restricted to aluminum, plastics, or thin sections of steel. By the mid-1970s, other technological developments led to further improvements in real-time radiography. These advances included high-energy x-ray sensitivity for image intensifiers, improved screen materials, digital video processing for image enhancement, and high-definition imaging with microfocus x-ray generators. Modern Image Intensifiers. Although the early image

intensifiers were suitable for medical applications and the inspection of light materials and thin sections of steel, the image quality was not sufficient for general use in radiography. Therefore, image intensifiers had to be redesigned for industrial material testing (Ref 3). The modern image intensifier is a very practical imaging device Fig. 28 Schematic of an image-intensifier tube for radiographic inspection with radiation energies up to 10 MeV. With the image intensifier, a 2% difference in absorption can be routinely achieved in production inspection applications. The typical dynamic range of an image intensifier before image processing is about 2000:1. A schematic of a modern x-ray image intensifier is shown in Fig. 29, along with a graph indicating the level of signal strength as the radiation passes through a sample and impinges on the entrance screen of the image-intensifier tube. These screens are usually made of a scintillating material, such as cesium iodide (CsI), and fixed to a photocathode. The energy of the incident radiation quanta generates electrons that produce light in the CsI entrance screen (approximately 150 photons per absorbed gamma quantum). To avoid degradation of the image quality by lateral dispersion of the light in the conversion screen, the CsI scintillating material, with a cubic crystalline structure, is grown under controlled conditions, resulting in small, needle-shaped elements. This structure causes the scintillating screen to act as a fiber-optical faceplate. Light generated in one crystal needle does not spread laterally, but is confined to the needle in a direction parallel to the incident radiation. Therefore, the thickness of the conversion screen does not cause appreciable deterioration in the spatial resolution of the system.

Fig. 29 Schematic of a typical radioscopic system using an x-ray image intensifier

The light from the scintillating screen then impinges on a photocathode in contact with the entrance screen. The photocathode emits photoelectrons. The electron image produced at the cathode is reduced by a factor of ten and is intensified by means of an accelerating voltage. The final phosphor screen presents a relatively bright image (approximately 5 million photons per second per square millimeter), caused by the impinging electrons. The image then passes through an optical system, which directs the image to a television camera tube, such as a vidicon or plumbicon tube. Vidicons have a dynamic range of 70:1; plumbicons, 200:1. Contrast Sensitivity. Figure 29 shows the importance of detection efficiency with regard to the contrast sensitivity of

the radioscopic system. The example is for a 370 GBq (10 Ci) cobalt-60 source, but a similar analysis can be performed for an x-ray source. A cobalt-60 source with an activity of 3700 GBq (100 Ci) generates 60,000 photons per second per square millimeter at 1 m with energies of 1.1 and 1.3 MeV. After passing through a 100 mm (4 in.) thick steel specimen, the radiation is attenuated in intensity by a factor of 15. Therefore, only 4000 photons per second per square millimeter are available for imaging. The detection efficiency of the CsI entrance screen in a typical modern image-intensifier tube is approximately 2%. This reduces the photon intensity to 80 per second per square millimeter. With this intensity, the laws of statistics (see the section "Quantum Noise (Mottle) and the Detective Quantum Efficiency" in the article "Industrial Computed Tomography" in this Volume) predict a noise-to-signal ratio of approximately 11%, which limits the attainable contrast resolution.

From Fig. 29, it is apparent that the critical factor for determining the contrast resolution is the relatively low detection efficiency of the entrance screen. To overcome this limitation, one must either increase the detection efficiency of the conversion screen or collect the photons over a longer period by summing television frames to improve the photon statistics in each picture element. The detection efficiency of the entrance screen can be improved by reducing the energy of the radiation or by making the entrance screen thicker. The minimum energy of the radiation is generally dictated by the need to have sufficient energy to penetrate the subject and to provide sufficient intensity for imaging at the conversion screen. Therefore, it is generally impossible to reduce the energy to improve conversion efficiency. Although a thicker entrance screen would improve detection efficiency, it would reduce the attainable spatial resolution. In general, improving the detection efficiency and improving the spatial resolution are competitive processes. The most conventional means of enhancing contrast resolution is through the summation of television frames with a digital image-processing system. The spatial resolution of a real-time system is principally defined by the size and thickness of the detectors (or grain

size of fluorescent crystals), the raster scan of the television system, and, for quantum statistical reasons, the intensity of the radiation. The dependencies are plotted in Fig. 30 for a typical image intensifier. Curve A shows a spatial resolution of about four line pairs per millimeter (or a resolution of 0.01 in.) at 10 R/h without television monitoring. Curve C shows the typical resolution of an image intensifier with television monitoring.

Fig. 30 Spatial resolution dependence of an image-intensifier system as a function of radiation intensity at the entrance screen. A, image-converter resolution; B, 625 TV lines limit; C, combined resolution of image converter and TV line

Real-Time Radiography With Fluorescent Screens. Because of the development of low-level television camera

tubes and low-noise video circuitry, the dim images on a fluorescent screen can be monitored with video systems. The contrast sensitivity and the spatial resolution of fluorescent-screen systems are comparable to those of image intensifiers, but the use of fluorescent screens is limited to lower radiation energies (below about 320 keV without intensifying screens and about 1 MeV with intensifying screens). Nevertheless, fluorescent screens can provide an unlimited field of view, while image intensifiers have a field of view limited to about 300 mm (12 in.). Example 1 in this article describes an application in which a fluorescent screen (with isocon camera) was preferred over an image intensifier. The dynamic

range of systems with fluorescent screens can vary from 20:1 for raw images to 1000:1 with digital processing and a large number of frames averaged. Television Cameras. The types of low-level television cameras available include image orthicons, image isocons, and secondary electron-coupled vidicons. Radiographic inspection is also performed with x-ray sensitive television cameras, which require no extra conversion or optics. X-ray sensitive television cameras are limited in image size and sensitivity and are primarily used in the inspection of low-density materials and electronic components.

Image orthicons and image isocons are return beam tubes with internal electron multipliers. The orthicon is useful to light levels of about 1.076 × 10-3 lm/m2 (10-4 ftc), and the isocon is useful to 1.076 × 10-4 lm/m2 (10-5 ftc). The isocon provides the best noise performance and resolution of all tubes for low-light levels and static scenes, but degrades for moving scenes and is highly complex. Both orthicons and isocons exhibit little target overloading damage. Both the image orthicon and isocon camera tubes must be carefully adjusted for optimum performance; the isocon is the most difficult to adjust and demands a skilled technician. Both are temperature sensitive and should be operated under stabilized conditions. The dynamic range for the isocon is about 1000:1. Secondary electron-coupled vidicons employ an internal initial stage of target amplification, using a secondary target. They are similar in performance to isocons, are less complex, but differ in lag and require special protection circuitry to prevent target destruction from overloading. Their primary use is in low-light level, low-contrast applications down to 1.076 × 10-4 lm/m2 (10-5 ftc). Fluorescent screens used in real-time radiography are similar in construction and properties to fluorescent

intensifying screens for film radiography. The commonly used fluorescent screens consist of a plastic substrate coated with a powdered luminescent material and an epoxy binder. The spatial resolution of the screen depends on the grain size of the luminescent material and the range and distribution of the conversion electrons and light photons. The x-ray absorption properties of the screen depend on the atomic weight and density of the luminescent material. The light conversion efficiency, the color of the emitted light, and the speed of the screen are additional properties that are dependent on the characteristics of the luminescent material. A wide variety of fluorescent screens are available. Screens using zinc sulfide as the fluorescent material have a high light output and a short afterglow, but the contrast and sharpness of this type of screen are poor. Zinc sulfide screens also use cadmium, which has a higher atomic number and a higher density than zinc. Other phosphor materials used include calcium tungstate and phosphors with rare-earth materials. Until the 1970s, many applications utilized the popular calcium tungstate screens. With the availability of rare-earth materials in high states of purity and at reasonable costs, however, one can design new phosphors with improved x-ray absorptions, greater density, and higher x-ray to light conversion efficiencies. Screens have been made of such materials as gadolinium oxysulfide doped with terbium, LaOBr:Tm, BaFCl:Eu, or YTaO2:Nb. These screens are discussed more fully in Ref 4. It should be noted that many rareearth screens are not as effective as calcium tungstate in absorbing very high energy x-rays. Because the intensity of the image on a fluorescent screen is rather low, very sensitive TV cameras are required for remote viewing. Different types of low light level cameras are available that have one- or two-stage light amplifiers or extremely sensitive camera tubes. When designing a fluoroscopic system, it is important to recognize that high light amplification from a double-stage amplifier generally yields poor spatial resolution and contributes additional noise to the image. In many cases, depending on the size of the fluorescent screen, the camera can also be the limiting factor in image quality. Therefore, both the entrance screen and the camera must be optimized for the particular inspection application. Digital Radiography. A third method of radiographic imaging involves the formation of an image by scanning a linear

array of discrete detectors along the object being irradiated. This method directly digitizes the radiometric output of the detectors and generates images in near real time. Direct digitization (as opposed to digitizing the output of a TV camera or image intensifier) enhances the signal-to-noise ratio and can result in a dynamic range up to 100,000:1. The large dynamic range of digital radiography allows the inspection of parts having a wide range of thicknesses and densities. Discrete detector arrangements also allow the reduction of secondary radiation from scattering by using a fan-beam detector arrangement like that of computed tomography (CT) systems. In fact, industrial CT systems are used to obtain digital radiographs (see the article "Industrial Computed Tomography" in this Volume). The detectors used in digital radiography include scintillator photodetectors, phosphor photodetectors, photomultiplier tubes, and gas ionization detectors. Scintillator and phosphor photodetectors are compact and rugged, and they are used in flying-spot and fan-beam detector arrangements. Photomultiplier tubes are fragile and bulky, but do provide the capability

of photon counting when signal levels are low. Gas ionization detectors have low detection efficiencies but better longterm stability than scintillator- and phosphor-photodetector arrays. A typical phosphor-photodetector array for the radiographic inspection of welds consists of 1024 pixel elements with 25 m (1 mil) spacing, covering 25 mm (1 in.) in length perpendicular weld seam. The linear photodiode array is covered with a fiber-optic faceplate and can be cooled in order to reduce noise. For the conversion of the x-rays to visible light, fluorescent screens are coupled to the array by means of the fiber optics. A linear collimator parallel to the array is arranged in front of the screen. The resolution perpendicular to the array is defined by the width of this slit and the speed of the manipulator. A second, single-element detector is provided to detect instabilities of the x-ray beam. Using 100-kV radiation, a spatial resolution of 0.1 mm (0.004 in.) can be achieved with a scanning speed of 1 to 10 mm/s (0.04 to 0.4 in./s). The data from the detector system are digitized and then stored in a fast dual-ported memory. This permits quasisimultaneous access to the data during acquisition. Before the image is stored in the frame buffer and displayed on the monitor, simple preprocessing can be done, such as intensity correction of the x-ray tube by the data of the second detector and correction of the sensitivity for different array elements. If further image processing or automatic defect evaluation is required, the system can be equipped with fast image-processing hardware. All standard devices for digital storage can be utilized. Image Processing. Because real-time systems generally do not provide the same level of image quality and contrast as radiographic film, image processing is often used to enhance the images from image intensifiers, fluorescent screens, and detector arrays. With image processing, the video images from real-time systems can compete with the image quality of film radiography. Moreover, image processing also increases the dynamic range of real-time systems beyond that of film (which typically has a dynamic range of about 1000:1).

Images can be processed in two ways: as an analog video signal and as a digitized signal. An example of analog processing is to shade the image after the signal leaves the camera. Shading compensates for irregularities in brightness across the video image due to thickness or density variations in the testpiece. This increases the dynamic range (or latitude) of the system, which allows the inspection of parts having larger variations in thickness and density. After analog image processing, the images can be digitized for further image enhancement. This digitization of the signal may involve some detector requirements. In all real-time radiological applications, the images have to be obtained at low dose rates (around 20 R/s). In digital x-ray imaging, however, the most often used dose is around 1 mR, in order to reduce the signal fluctuations that would result from a weak x-ray flux. This means that only a short exposure time (of the order of a few milliseconds) and a frequency of several images can be used if kinetic (motion) blurring is to be avoided. The resulting requirements for the x-ray detector are therefore: • • •

The capability of operating properly in a pulsed mode, which calls for a fast temporal response An excellent linearity, to allow the use of the simplest and most efficient form of signal processing A wide operating dynamic range in terms of dose output (around 2000:1)

Once the image from the video camera has been digitized in the image processor, a variety of processing techniques can be implemented. The image-processing techniques may range from the relatively simple operation of frame integration to more complex operations such as automatic defect evaluation. The following discussions briefly describe three imageprocessing operations. More detailed information on image processing is available in the article "Digital Image Enhancement" in this Volume. Frame integration (or summing) is a simple technique that can improve the signal-to-noise ratio and decrease (or

eliminate) the screen mottle associated with low levels of radiation intensity. This technique adds the digital images from a number of frames and displays the result on a video monitor. Frame integration can be done in real time, that is, at standard television frequency of 30 frames per second. The number of television frames that can be integrated depends on the number of frame buffers available for this procedure. Typical summation times range from a few seconds for applications using x-rays with high intensities up to 1 min or more for applications using radiation from low-activity radionuclide sources. However, these times are small compared to the time required for exposing and developing x-ray film.

Image subtraction can be used to improve sensitivity through the subtraction of one integrated image from a second

integrated image. The advantage of the subtraction technique is that all areas of the scene having the same brightness cancel each other to produce an extremely flat or uniformly gray image. With such an image available, the contrast and brightness can then be electronically enhanced to reveal details not visible in either a real-time or a simple integrated image. Example 1 in this article describes one method called pairs subtraction. Automated defect evaluation systems have been developed for the radiographic inspection of welds (Ref 5). The

automatic detection of defects with such systems is based on a mathematical algorithm that evaluates a set of parameters for each indication and compares this set to a catalog of known defects. For any particular problem, the inspection parameters must be selected to provide optimum and reproducible results. To compare different radiographs and to extract features such as the depth of a defect, it is necessary to calibrate the gray levels against material thickness and to calibrate the linear dimensions. Once established, the optimized conditions must be consistently used for all inspections. Figure 31 shows a flow chart of an automatic image evaluation procedure, and Fig. 32 shows the images from different steps in the procedure. The first step is to remove irrelevant details from the digitized image. This can be done by subtracting a background image. This background image is generated by creating an unsharp image with a low-pass filter. Although radiographs usually have a center brightness that differs from that at the edges, this gradient is eliminated by subtracting the background image from the original image. At this state (Fig. 32b), it is possible to set a threshold and to identify as possible defects all pixels whose gray levels exceed this threshold.

Fig. 31 Flow chart of an automatic image evaluation procedure

Fig. 32 Images developed by an image-processing procedure that identifies and evaluates the images of weld defects in real-time radiographs. (a) Original image. (b) Binary image. (c) Binary image after noise reduction. (d) Original image with defects identified. Courtesy of RTS Technology, Inc

In the next step, these potential defects are examined. First, all very small indications are examined, such as single pixels or groups of only a few pixels, which result from image noise. Larger details, which are obviously separated by one or two missing pixels, are connected (Fig. 32c). This algorithm prevents the computer from spending time analyzing noise that exceeds the threshold. For each potential defect, a number of parameters are evaluated. These parameters include position, orientation, and a set of descriptors for size and shape. The values of these parameters are compared with the mean value and standard deviation for each parameter in the catalog data already stored. Where some consistency is found between measured and stored data, the program identifies the defects (Fig. 32d) and calculates the probability that the indication is a particular type of defect. Where there is insufficient consistency between the measured and stored data, the program calls for the assistance of an interpreter and enters a learn mode. Once the interpreter has evaluated the defect, the program stores the values for each parameter of the defect, along with the evaluation in the catalog. This information is now considered in future automated evaluations. Relevant defects cause the rejection of the workpiece. Applications. Real-time radiography is often applied to objects on assembly lines for rapid inspection. Digital radiography with discrete detectors, in particular, is being more widely used because it provides an unlimited field of view and the largest dynamic range of all the radiographic methods. Because a larger dynamic range allows the inspection of parts having a wider range of thicknesses, digital radiography has found application in the inspection of castings (see the article "Castings" in this Volume).

Radiography is also an accepted method of detecting internal discontinuities in welds. Image intensifiers are the most widely used system for the real-time radiographic inspection of welds, although digital radiography with discrete detectors is also used. Projective magnification with microfocus x-ray sources is also useful in weld and diffusion bond inspection. Advantages and Limitations. Real-time radiography provides immediate information without the delay and expense

of film development. Real-time systems also allow the manipulation of the testpiece during inspection and the improvement of resolution and dynamic range by digital image enhancement. Real-time radiography with discrete detectors (that is, digital radiography) also provides good scatter rejection. Nevertheless, film radiography is a simpler technique that represents less of a capital investment. The sensitivity and resolution of real-time systems are usually not as good as those of film, although real-time radiography can approach the

sensitivity and resolution of film radiography when good inspection techniques are used. Good technique is often more important than the details of the imaging system itself.

Example 1: Design of a Rapid, High-Sensitivity Real-Time Radiographic System. A system was needed for the real-time radiographic inspection of jet-engine turbine blades. These were critical nickel alloy castings that required inspection to at least a 2-1T (1.4%) penetrameter sensitivity level through thicknesses that ranged from about 0.5 to more than 13 mm (0.02 to 0.5 in.). The volume of parts produced necessitated inspection at a rate of 60 parts per hour or more. For consistency, all aspects of the inspection were to be computer controlled. In addition, it was necessary to visually compare the actual parts with their radiographic images to separate castings with excess metal due to mold flaws that might be acceptable from castings with high-density inclusions that might be rejectable. Image Processing. A preliminary evaluation of the parts indicated that the 2-1T penetrameter sensitivity requirement could be met with suitable image-processing techniques. Both analog and digital processing were used for these complex parts.

The analog processing technique used was an electronic adjustment called shading. Shading is performed in the camera control unit after the signal leaves the camera and before it is sent to the image processor for digitizing and further processing. Shading compensates for variations in brightness across the video image due to thickness variations in the part. In effect, proper shading adjustment increases the latitude (dynamic range) of the image. In the case of the airfoils to be inspected with the real-time radiographic system, shading permitted blade thicknesses ranging from 0.5 to 4 mm (0.02 to 0.16 in.) to be adequately imaged in a single view. Once the image from the video camera had been digitized in the image processor, a great many processing techniques were implemented. The simplest was to integrate the image. This technique added the digital images from a number of frames and displayed the result on the video monitor. Integration improved the signal-to-noise ratio over that of a single frame from the video camera. For these castings, 128 frames of integration produced good results. To achieve the required sensitivity, an image subtraction technique was also used. This was an extremely useful processing technique in which one integrated image was subtracted from a second integrated image. There are a number of variations of this technique. It was experimentally determined that the best technique for this application was the one termed "pairs subtraction." To produce a "pairs subtraction," the part was displaced slightly between the two integrated images that were subsequently subtracted from each other. X-Ray Source. Because of the range of thicknesses to be inspected, two x-ray sources were selected for the system--one

rated at 160 kV for the thinner sections and the other at 320 kV for the thicker sections. The use of 0.4 mm (0.015 in.) focal spots for the 160-kV unit and 0.8 mm (0.03 in.) focal spots for the 320-kV unit provided sufficient resolution to detect 0.25 mm (0.01 in.) defects even at the relatively short source-to-screen distances available in the final real-time radiographic system design. Because it was required that the blade be exposed to only one source of radiation at a time and that the x-ray tubes be run continuously, two computer-controlled, high-speed lead shutters were designed to block the unwanted radiation. Imaging System. The imaging camera, which was housed in its own shielded enclosure, included a 100 × 100 mm (4

× 4 in.) rare-earth fluorescent screen for converting x-rays to visible light. Light from the screen was directed to the objective lens of an image isocon camera by a 45° front surface mirror (Fig. 33). An isocon camera was selected over an image intensifier for two reasons. First, an isocon camera is much less susceptible to blooming due to bypass radiation. This permitted the use of relatively loose masking around the part. Second, computer-selectable shading channels were available with the isocon camera; therefore, shading could be preadjusted for different views. The computer program could then set the required shading channel during an inspection. The camera system was designed to include remote controls to facilitate focusing of the camera and adjustment of the field of view from the full-screen size down to a 25-mm (1-in.) square to increase the optical magnification if necessary.

Fig. 33 Schematic of the equipment described in Example 1

Part Handling. To minimize part-handling time, a very accurate, high-speed, commercially available robot with four

axes of movement (X, Y, Z, and rotation) was selected. The robot was equipped with a computer and sophisticated operating programs. After the robot was taught a few fixed points, it determined its own path from one position to another. This was especially convenient because a large number of parts were to be placed into the real-time radiographic system in a tray. The robot only needed to be told the number of rows and columns in the tray and taught three corner pickup locations. From this information, it could calculate the exact pickup points for all other locations in the tray. The part containers required considerable development before consistent robot pickup was achieved. In the final design, parts were dropped into molded plastic holders with no special care, and they were positioned well enough for the robot to pick them up with a high degree of reliability. System Layout and Operating Console. To permit installation in a factory, the x-ray sources, imaging camera, and

robot were located in an exempt, radiation-shielded enclosure. Trays of parts to be inspected were inserted through a sliding access door from an adjacent environmental enclosure. An environmental enclosure was included for the operator's station to provide a relatively clean environment in the manufacturing area. The inspection operator's station was designed to include a desktop console with a video monitor on which either real-time or processed images could be viewed. The console also included a computer terminal and keyboard; a joystick control was used for performing various image processor functions and for manual control of the robot. The x-ray controls, the robot controller, the isocon camera controls, and the computer and image processor were located near the console. A console for reviewing images while examining parts visually was also provided in the environmental enclosure. In addition to the video monitor, image processor, and computer terminal, a voice recognition system was added so that commands and information could be input to the computer without using the keyboard.

System Software. Three separate programs were developed for the real-time radiographic system: Inspect, Review,

and Supervisor. The computer selected included a multiuser operating system to allow the Inspect and Review programs to be run concurrently at different terminals. All the programs were menu driven; the operator simply entered information or exercised options when prompted to do so by the menu on the computer terminal. The Inspect program was designed for day to day inspection operation. In this program, a preestablished scan plan controlled all the parameters required for an inspection. As long as the operator continued accepting views, the program moved through the plan for each part in turn until all parts in the tray had been completed. In running the scan plan, the computer interfaced with the x-ray machines to control the kilovolt and milliamp settings, the shutters, and x-ray on and off. It also interfaced with the isocon camera control unit to turn the camera beam on and off and to call the required shading channel, and it interfaced with the robot controller to direct the part movements. The software was designed so that the scan plan could be interrupted at any time. One option in the interrupt mode gave the operator complete manual control of all system functions through the terminal keyboard. The ability to call an image quality indicator (IQI) to check for proper radiographic sensitivity was also designed into the interrupt mode. When the IQI was called for, the part was returned to its tray position, and the IQI was picked up and moved into the x-ray beam. An image of the IQI was generated using the same x-ray parameters and image processing as employed for the part view being inspected at the time. The Review program was designed to allow stored images to be recalled when the parts were available for visual review. In addition to accessing images, this program allowed various enhancement techniques to be performed on the image. All program functions could be exercised using the voice recognition capability so that the operator's hands were free to handle parts. Once a final disposition for a part was entered, the views for that part were automatically deleted from memory. The Supervisor program was written to include a number of program functions that were to be protected from normal operator access. These included changing the lists of authorized users and their passwords, access to an operations log, and creation or modification of scan plans. System Performance. The sensitivity of the system was demonstrated with plaque penetrameters, a wire

penetrameter, and an actual part with small holes. The plaque penetrameter sensitivity of the system is shown in Fig. 34. Figure 34(a) shows a pairs-subtracted image of a 0.025 mm (0.001 in.) thick penetrameter through 1.25 mm (0.05 in.) of material, and Fig. 34(b) shows a 0.125 mm (0.005 in.) penetrameter through 6.4 mm (0.25 in.) of material. All three holes can be seen near the centers of the circles as adjacent white and black areas that result from the pairs subtraction technique.

Fig. 34 Examples of plaque penetrameter indications. (a) 0.025 mm (0.001 in.) thick penetrameter on a testpiece with a thickness of 1.3 mm (0.050 in.). (b) 0.125 mm (0.005 in.) penetrameter on a testpiece with a thickness of 6.4 mm (0.25 in.)

Figure 35 shows a pairs-subtracted image of an actual part in which a series of holes was drilled using the electrical discharge machining process. The holes have a nominal diameter of 0.25 mm (0.01 in.) and a depth that is 2% of the material thickness at the location where they were drilled. The material thickness in this airfoil view ranged from 1 to approximately 4 mm (0.04 to 0.16 in.).

Fig. 35 Radiographic image of holes with a nominal diameter of 0.25 mm (0.01 in.) and a depth of 2% of wall thickness. Holes are indicated with arrows. Courtesy of Lockheed Missiles & Space Company, Inc.

References cited in this section

3. R. Link, W. Nuding, and K. Sauerwein, Br. J. Non-Destr. Test., Vol 26, 1984, p 291-295 4. L.H. Brixner, Mater. Chem. Phys., Vol 16, 1987, p 253-281 5. W. Daum, P. Rose, H. Heidt, and J.H. Builtjes, Br. J. Non-Destr. Test., Vol 29, 1987, p 79-82 Radiographic Inspection Revised by the ASM Committee on Radiographic Inspection*

Film Radiography Film selection and exposure time are two major factors in film radiography. The selection of radiographic film for a particular application is generally a compromise between the desired quality of the radiograph and the cost of exposure time. The quality of the radiograph depends mainly on film density, gradient, graininess, and fog, which are functions of film type and development procedure. Exposure time depends mainly on film speed and on radiation intensity at the film surface. The factors affecting film selection and exposure time are discussed in more detail in this section. Characteristics of X-Ray Film Three general characteristics of film--speed, gradient, and graininess--are primarily responsible for the performance of the film during exposure and processing and for the quality of the resulting image. Film speed, gradient, and graininess are interrelated; that is, the faster the film, the larger the graininess and the lower the gradient--and vice versa. Film speed and gradient are derived from the characteristic curve for a film emulsion, which is a plot of film density versus the exposure required for producing that density in the processed film. Graininess is an inherent property of the emulsion, but can be influenced somewhat by the conditions of exposure and development.

There are other characteristics of x-ray film that can be used to produce special effects. However, the following discussion is mainly confined to film speed, gradient, graininess, and the factors that affect them. Density. The quantitative measure of the blackening of a photographic emulsion is called density. Density is usually

measured directly with an instrument called a densitometer. There are two kinds of density: • •

Transmission density, which is associated with transparent-base radiographic film Reflection density, which is associated with opaque-base imaging material such as radiographic paper

In the following discussion, the unqualified term density will refer to transmission density. Transmission density is defined by:

(Eq 14)

where D is the density, I0 is the intensity of light incident on the film, and It is the intensity of light transmitted through the film. Reflection density is defined by:

(Eq 15)

where Dr is the reflection density, I0 is the light intensity incident on the radiographic image, and Ir is the light intensity reflected from the image. Although Eq 14 and 15 are similar, the densities are quite different. The ratio of incident intensity, I0, to transmitted intensity, It, in Eq 14 is called opacity. The inverse of this ratio, It/I0, is called transmittance and indicates the fraction of incident light transmitted through the film. The relationship of density to opacity and transmittance is as tabulated below:

Density, log (I0/It)

Opacity, I0/It

Transmittance, It/I0

0

1

1.00

0.3

2

0.50

0.6

4

0.25

1.0

10

0.10

2.0

100

0.01

3.0

1,000

0.001

4.0

10,000

0.0001

If the light transmitted through the film is half of the incident light (transmittance = 0.5), the density is only 0.3, and for a density of 1.0, only one-tenth of the incident light intensity is transmitted. The fact that only of the incident light is transmitted at a density of 2.0 is the reason that it is necessary to use high-intensity illuminators for viewing industrial radiographs at densities exceeding 2.0. Exposure is the intensity of radiation multiplied by the time during which it acts, that is, the amount of energy that

reaches a particular area of the film and that is responsible for producing a particular density on the developed film. Exposure can be specified either in absolute units (such as ergs per square centimeter or roentgens) or in relative units (where one particular exposure is used as a reference and all others are specified in terms of that reference). Unless equipment and time are available for making the precise radiation measurements that are required for defining exposure in absolute units, relative exposure is much more convenient than absolute and is equally as useful. In discussing film properties, only absolute exposure or relative exposure as defined above is appropriate. Relative exposure is used in the present discussion. Characteristic Curves. The relation between the exposure applied to a given type of radiographic film and the

resulting density is expressed in a curve known as the characteristic curve of that particular type of film. Other names for this curve are the H and D curve, D log E curve, or sensitometric curve. Such curves are determined by applying a series of known exposures to the film and, after processing the film according to a standard procedure, reading the resulting densities. The curve is generated by plotting density against the logarithm of relative exposure. Figure 36(a) shows the characteristic curves of three commercial films exposed to x-radiation between lead screens.

Fig. 36 Characteristic curves for x-ray film that determine type of film and film gradient, speed, and density. (a) Typical curves for three industrial x-ray films exposed to x-radiation between lead screens. (b) Evaluation of gradients at two points on the curve for film A in (a). (c) Density differences ( D) corresponding to a 20% difference in relative exposure ( log E = 0.08) determined for the two values of gradients evaluated in (b). (d) Average gradients for film A determined over two density ranges

Relative exposure is used partly because there are no convenient units suitable to all kilovoltages and scattering conditions and partly because it is easy to determine the logarithm of relative exposure. Using the logarithm of relative exposure instead of only relative exposure has several other advantages; for example, the otherwise long scale is compressed, and ratios of intensities or exposures (which are determined by simply subtracting logarithms) are usually more significant in radiography than actual exposures or intensities. Characteristic curves are very useful in determining the speed and gradient of the film as well as indicating the type of film. For example, in Fig. 36(a), characteristic curves for films A and B are J-shaped, which is typical of industrial x-ray film of types 1, 2, and 3. The curve for film C begins to flatten at a density between 3.0 and 3.5, giving it the S-shape typical of type 4 x-ray films used with fluorescent screens. Film speed is inversely related to the time required for a given intensity of radiation to produce a particular density on

the film; the shorter the exposure, the faster the film. In absolute units, film speed is inversely proportional to the total energy (roentgens) of a particular radiation spectrum (wavelength distribution at a given kilovoltage) that produces a given density on the film. For most practical applications, it is convenient and effective to deal with relative speeds. In using relative speeds, film speeds are expressed in terms of the speed of one particular film whose relative speed is arbitrarily assigned a value, for example, 100. Thus, if film A requires half the exposure of film B, the slower film (film B) is chosen as the standard and assigned an arbitrary speed of 100. Film A, which is twice as fast, would have a relative speed of 200.

To avoid making absolute measurements to determine film speed, it is convenient to refer to a group of film curves such as those in Fig. 36(a). Curves positioned to the left of the chart require less exposure for a given density; those to the right, more exposure. Thus, for a density of 2.0, film C in Fig. 36(a) is faster than film A, and both are faster than film B. Their relative speed is calculated by determining the differences in log relative exposure and converting to the antilog. Film B is the slowest film and requires a log relative exposure of 2.5 for a density of 2.0. Film B is chosen as the standard and is assigned a speed of 100. Film A requires a log relative exposure of 1.9 for the same density. Subtracting 1.9 from 2.5 gives a difference of 0.6, for which the antilog is about 4, so the relative speed of film A is four times that of film B, or has a relative speed of 400. Similarly, at a density of 2.0, film C requires a log relative exposure of 1.6. Subtracting 1.6 from 2.5 gives a log difference of 0.9, for which the antilog is about 8. Thus, film C is eight times faster than film B, or has a relative speed of 800 at a density of 2.0. Another advantage of using groups of characteristic curves is that the visual assessment of relative speeds can be easily made. For example, although films A and B have similarly shaped curves running almost parallel, the curve for film C is radically different. As calculated above, film C has a relative speed of 800 at a density of 2.0; however, at greater densities, the relative speed of film C is lower than 800, and at lesser densities (down to a density of 1.0), the relative speed of film C is greater than 800. Therefore, whenever relative speeds are used, the density at which they were determined must be given. Film gradient, also called film contrast, is a measure of the slope of the characteristic curve. If the difference is great, the gradient (contrast) is said to be high. If the difference is slight, the gradient is said to be low. The contrast seen on a radiograph is known as radiographic contrast and is composed of two factors:

• •

Subject contrast, which is the result of variations in the amount of radiation absorbed by the testpiece and which causes variations in radiation intensity impinging on the film Film gradient (or film contrast), which is a measure of the response of the emulsion to the intensity of impinging radiation and is a characteristic of a given film

Film gradient is determined from the characteristic curve by finding the slope of the curve at a given density. The slope of the characteristic curve changes continuously over its entire length, as indicated in Fig. 36(a). The steeper the slope of the curve over a range of relative exposures, the greater the difference in density and therefore the greater the resolution of detail; thus, high gradient is important for good radiography. In Fig. 36(b) tangents to a characteristic curve for a typical x-ray film have been drawn at two points, and the two corresponding gradients (a/b and c/d) have been evaluated. Note that the gradient varies from 0.8 in the shallow portion of the curve (called the toe) to 5 in the steeper portion of the curve. Two regions of a testpiece that are slightly different in thicknesses will transmit slightly different intensities of radiation to the film. On the characteristic curve, this would represent a small difference in the log relative exposure. For example, assume that at a given kilovoltage the thinner section transmits 20% more radiation than the thicker section. The difference in log relative exposure is 0.08 and is independent of tube current, exposure time, or source-to-film distance. In the toe of the characteristic curve where, the gradient is only 0.8, the density difference is only 0.06, as indicated in Fig. 36(c). However, in the steeper portion of the curve where the gradient is 5, the density difference is 0.40. This effect of film gradient is the main reason why it is best to use high exposures, obtaining the highest density that can be viewed on a given illuminator for the greatest resolution of minor differences in transmitted radiation. It is often more convenient to express gradient as an average over a given useful range of densities than for a single density. This is a simple calculation in which the difference between the two densities is divided by the difference between the log relative exposures for these densities; this ratio, known as the average gradient, is shown in Fig. 36(d). Graininess. The silver halide grains that are contained in the emulsion of x-ray film are minute and can be seen only

with a high-power microscope, such as an electron microscope. Even though the emulsion on each side of the film is only about 0.013 mm (0.0005 in.) thick, the grains are piled on top of each other in countless numbers. When the exposed-andprocessed radiograph is viewed, these small individual silver grains appear grouped together in relatively large masses. This clumping, which is visible to the unaided eye or at low magnification, produces the visual impression called graininess.

All films exhibit some degree of graininess. In general, slower films have less graininess than faster films. This general relation is reflected in Table 6, which is an old classification scheme specified by the American Society for Testing and Materials (ASTM) in the earlier 1984 edition of ASTM E 94. Although this classification scheme is somewhat arbitrary, it is still sometimes specified in codes and specifications. Type 1 films have the least graininess, type 3 films exhibit the most graininess, and type 2 films are intermediate. Table 6 General characteristics of the four types of radiographic film specified in the earlier (1984) edition of ASTM E 94 These groupings are given only for qualitative comparisons. For a more detailed discussion on film classification, see the section "Film Types" in this article. Film type

Film characteristic

Speed

Gradient

Graininess

1

Low

Very high

Very fine

2

Medium

High

Fine

3

High

Medium

Coarse

4(a)

Very high(b)

Very high(b)

(c)

(a) Normally used with fluorescent screens.

(b) When used with fluorescent screens.

(c) Graininess is mainly a characteristic of the fluorescent screens.

(d) When used for direct exposure or with lead screens

The energy of the radiation also affects the graininess of films. As the penetrating capability of radiation increases (photon energy increases and wavelength decreases), the graininess of all films increases, but the rate of increase may be different for different films. As a result, it is possible that of two film types, one will show more graininess at long wavelengths while the other shows more graininess at short wavelengths. The graininess of the images that are produced at high photon energies makes the inherently fine-grain films such as types 1 and 2 particularly useful at photon energies of 1 MeV or more. Another source of graininess occurs in radiography using fluorescent intensifying screens. The graininess seen in fluorescent-screen radiographs is primarily caused by the graininess that is inherent in the screen. This effect overshadows the effect of film graininess, especially with films that are particularly sensitive to screen fluorescence. The graininess of fluorescent-screen radiographs increases significantly with increased kilovoltage. Lead screens have little effect on the graininess of the image; only slight increases in graininess at equal densities have been noted when lead-screen radiographs are compared with direct radiographs made at the same photon energy.

Spectral Sensitivity. The shape of the characteristic curve of a given x-ray film is for all practical purposes unaffected

by the wavelength distribution in the x-ray or -ray beam used for the exposure. However, the sensitivity of the film in terms of roentgens required to produce a given density is strongly affected by radiation energy (beam spectrum of a given kilovoltage or given -ray source). Figure 37 shows the exposure required for producing a density of 1.0 on radiographic film developed in an x-ray developer for 5 min at 20 °C (68 °F). The exposures were made directly, without screens. The spectral-sensitivity curves for all x-ray films have approximately the same general features as the curves shown in Fig. 37. The details, however, differ from film to film; for example, the ratio of maximum-to-minimum sensitivity of the film to roentgens of exposure varies. The details of the curves also differ with the degree that long wavelengths have been filtered from the beam (Fig. 37).

Fig. 37 Spectral-sensitivity curves for a radiographic film showing exposure required to produce a density of 1.0

The spectral sensitivity of a given film, or the difference of spectral sensitivity between films, is usually considered in the preparation of exposure charts and tables of relative speeds. For example, in tables of relative speeds, the speed of a particular film usually varies with tube voltage. Also, in the preparation of exposure calculators for -rays, where roentgens necessary to expose a given film to produce a given density are extensively used, spectral sensitivity is important. As a result, manufacturers of radiographic film often furnish tables of spectral sensitivity for the films they manufacture, indicating the roentgens required to produce several net densities at various x-ray tube voltages and -ray photon energies. Effect of Development on Film Characteristics. Although the shape of the characteristic curve is not affected by variations in photon energy, it is affected by variations in the degree of development. The degree of development depends on the type of developer used (including its degree of activity), temperature, and developing time. These factors are discussed in the Appendix to this article. Within limits, an increased degree of development increases the speed and gradient of a given film. If the limits are exceeded (that is, the film is overdeveloped), film speed based on density (density above fog) no longer increases and may even decrease. Also, the fog (gray cloudiness) resulting from excessive action of developer chemicals on unexposed grains will increase, and the gradient may decrease.

Figure 38 shows the characteristic curve of a radiographic film as affected by various times of manual development at 20 °C (68 °F) in an x-ray developer mixed from powder. As the developing time increases, the characteristic curve progressively becomes steeper (gradient increases) and positioned more to the left on the chart (speed increases). It is from such curves that values of relative speed and average gradient can be determined and developing recommendations can be derived.

Fig. 38 Effect of various developing times on the characteristic curve of a radiographic film developed at 20 °C (68 °F) in a manual process

Figure 39 shows the values of relative speed, average gradient, and fog plotted against developing time for the same film and processing conditions detailed in Fig. 38. These curves indicate that at an 8-min developing time the average gradient has reached its peak and further development would decrease film gradient. The speed, gradient, and fog curves of Fig. 39 are characteristic of this type of film, and similar curves are different for other types of films.

Fig. 39 Effect of developing time on (a) relative speed, (b) average gradient, and (c) fog for a radiographic film. These graphs were derived from the same data as Fig. 38.

The degree of development also affects graininess. For example, if development time is increased to produce higher film speed, the graininess is also increased. On the other hand, a developer or developing technique that gives an appreciable reduction in graininess will also result in an appreciable loss of film speed. Adjustments in developing techniques made to compensate for changes in temperature or in activity of a developer have little effect on graininess. The variation of film gradient, speed, fog, and graininess with development time and temperature for a given radiographic film and developer solution makes it imperative that standardized procedures for development of radiographic film be meticulously followed. Unless this is done, it is impossible to obtain consistent resolution of variations in radiation intensity. Inconsistency in the resolution of variations in radiation intensity can seriously impair the ability to detect flaws in metals. Film Types and Selection The selection of radiographic film for a particular application is generally a compromise between the desired quality of the radiograph and the cost of exposure time. This compromise occurs because slower films generally provide a higher film gradient and a lower level of graininess and fog. Film Types. The classification of radiographic film is complicated, as evidenced by changes in ASTM standard practice

E 94. The 1988 edition of ASTM E 94 references a new document (ASTM E 746-87), which describes a standard test method for determining the relative image-quality response of industrial radiographic film. Careful study of this document is required to arrive at a conclusive classification index suitable for the given radiographic film requirements of a facility. Earlier editions (1984 and prior) of ASTM E 94 contained a table listing the characteristics of industrial films grouped into four types. The general characteristics of these four types are summarized in Table 6. This relatively simple classification method is referenced by many codes and specifications, which may state only that a type 1 or 2 film can be used for their specification requirements. However, because of this relatively arbitrary method of classification, many film manufacturers may be reluctant to assign type numbers to a given film. Moreover, the characteristics of radiographic films can vary within a type classification of Table 6 because of inherent variations among films produced by different manufacturers under different brand names and because of variations in film processing that affect both film speed and radiographic density. These variations make it essential that film processing be standardized and that characteristic curves for each brand of film be obtained from the film manufacturer for use in developing exposure charts. Because the variables that govern the classification of film are no longer detailed in ASTM E 94-88, it is largely the responsibility of the film manufacturer to determine the particular type numbers associated with his brand names. Some manufacturers indicate the type number together with the brand name on the film package. If there is doubt regarding the type number of a given brand, it is advisable to consult the manufacturer. Most manufacturers offer a brand of film characterized as very low speed, ultrahigh gradient, and extremely fine grain. Film selection for radiography is a compromise between the economics of exposure (film speed and latitude) and the

quality desired in the radiograph. In general, fine-grain, high-gradient films produce the highest-quality radiographs. However, because of the low speed typically associated with these films, high-intensity radiation or long exposure times are needed. Other factors affecting radiographic quality and film selection are the type and thickness of the testpiece and the photon energy of the incident radiation. Although the classification of film is more complex than the types given in Table 6, a general guide is that better radiographic quality will be promoted by the lowest type number in Table 6 that economic and technical considerations will allow. In this regard, Table 7 suggests a general comparison of film characteristics for achieving a reasonable level of radiographic quality for various metals and radiation-source energies. It should be noted, however, that the film types are only a qualitative ranking of the general film characteristics given in Table 6. Many radiographic films, particularly those designed for automatic processing, cannot be adequately classified according to the system in Table 6. This compounds the problem of selecting film for a particular application.

Table 7 Guide to the selection of radiographic film for steel, aluminum, bronze, and magnesium in various thicknesses Type of film(a) for use with these x-ray tube voltages, or radioactive isotopes:

Thickness

mm

(in.)

50-80 kV

80-120 kV

120-150 kV

150-250 kV

Ir-192

250-400 kV

1 MeV

Co-60

2 MeV

Ra

6-31 MeV

3

3

2

1

...

...

...

...

...

...

...

4

3

2

2

...

1

...

...

...

...

...

...

4

3

2

2

2

1

...

1

2

...

Steel

0-6 (0-

)

6-13 (

-

)

(

-1)

13-25

25-50

(1-2)

...

...

...

3

2

2

1

2

1

2

1

50-100

(2-4)

...

...

...

4

3

4

2

2

2

3

1

100-200

(4-8)

...

...

...

...

...

4

3

3

2

3

2

>200

(>8)

...

...

...

...

...

...

...

...

3

...

2

1

1

...

...

...

...

...

...

...

...

...

2

1

1

1

...

...

...

...

...

...

...

2

1

1

1

...

1

...

...

...

...

...

Aluminum

0-6 (0-

)

6-13 (

-

(

-1)

13-25

)

25-50

(1-2)

3

2

2

1

1

1

...

...

...

...

...

50-100

(2-4)

4

3

2

2

1

2

...

...

...

...

...

100-200

(4-8)

...

4

3

3

2

3

...

...

...

...

...

>200

(>8)

...

...

...

...

4

...

...

...

...

...

...

Bronze

0-6 (0-

4

3

2

1

1

1

1

...

...

...

...

...

3

2

2

2

1

1

...

1

...

...

...

4

4

3

2

2

1

2

1

2

...

)

6-13 (

-

(

-1)

)

13-25

25-50

(1-2)

...

...

4

4

3

3

1

2

1

2

1

50-100

(2-4)

...

...

...

...

3

4

2

3

2

3

1

100-200

(4-8)

...

...

...

...

...

...

3

3

2

...

2

>200

(>8)

...

...

...

...

...

...

...

...

3

...

2

1

1

...

...

...

...

...

...

...

...

...

1

1

1

...

...

...

...

...

...

...

...

2

1

1

...

1

...

...

...

...

...

...

Magnesium

0-6 (0-

)

6-13 (

-

(

-1)

13-25

)

25-50

(1-2)

2

1

1

1

1

...

...

...

...

...

...

50-100

(2-4)

3

2

2

1

2

...

...

...

...

...

...

100-200

(4-8)

...

3

2

2

3

...

...

...

...

...

...

>200

(>8)

...

...

...

4

...

...

...

...

...

...

...

(a) These recommendations represent a usually acceptable level of radiographic quality and are based on the qualitative classification of films defined in Table 6. Optimum radiographic quality will be promoted by use of the lowest-number film type that economic and technical considerations will allow. The recommendations for type 4 film are based on the use of fluorescent screens.

Film latitude, which is the range of testpiece thickness that can be recorded with a single exposure, also influences film

selection (see the following section in this article for a discussion of latitude). High-gradient films generally have narrow latitude, that is, only a narrow range of testpiece thickness can be imaged with optimum density for interpretation. If the

testpiece is of nonuniform thickness, more than one exposure may have to be made (using different x-ray spectra or different exposure times) for complete inspection of the piece. The number of exposures, as well as the exposure times, can often be reduced by using a faster film of lower gradient but wider latitude, although there is usually an accompanying reduction in ability to image small flaws. Exposure Factors The exposure time in film radiography depends mainly on film speed, the intensity of radiation at the film surface, the characteristics of any screens used, and the desired level of photographic density. In practice, the energy of the radiation is first chosen to be sufficiently penetrating for the type of material and thickness to be inspected. The film type and the desired photographic density are then selected according to the sensitivity requirements (Eq 2) for recording the expected variations in the intensity of the transmitted radiation. Once these factors are fixed, then the source strength, the source-tofilm distance, and the characteristics of any screens used determine the exposure time. With a given type of film and screen, the exposure time to produce the desired photographic density obeys the reciprocity law for equivalent exposures (Eq 5). The reciprocity law can be modified to include the inverse-square law (Eq 13), which models the relation between radiation intensity at the film and the source-to-film distance. Therefore, because the intensity is inversely proportional to the square of distance from the source according to Eq 13, the reciprocity law for equivalent exposures with an x-ray tube (Eq 5) can be rewritten as:

(Eq 16) where i is the tube current, t is the exposure time, L is the source-to-film distance, and the subscripts refer to two different combinations that produce images with the desired photographic density. The parallel expression that applies to exposures made with a -ray source is:

(Eq 17) where a is the source strength in gigabecquerel (curies). For practical applications using the same quality of x-rays (same kilovoltage) or the same radioactive source, Eq 16 or 17 is applied in the following manner. If a satisfactory radiograph of a given testpiece can be obtained at a 1 m (3.3 ft) source-to-film distance, 20-mA tube current, and 10-s exposure time, an equivalent radiograph can be obtained in 6.4 s with 20-mA tube current or in 8 s with 16-mA tube current when the distance is reduced to 0.8 m (2.6 ft). However, if the source-to-film distance is increased to 2 m (6.6 ft), it will take 40 s at 20 mA or 32 s at 25 mA or 16 s at 50 mA to produce equivalent film density. Similarly, if a satisfactory radiograph can be obtained with a 440 GBq (12 Ci) cobalt-60 source at 1 m (3.3 ft) source-tofilm distance in 20 min, a radiograph of equivalent density will require a 45-min exposure time with the same 440 GBq (12 Ci) source if the distance is increased to 1.5 m (4.9 ft). Exposure Charts for X-Ray Radiography. A starting point must be determined for the calculations described in the discussions above. The starting point is ordinarily derived from exposure charts of the type shown in Fig. 40. Equipment manufacturers usually publish exposure charts for each type of x-ray generator that they manufacture. These published charts, however, are only approximations; each particular unit and each installation is unique. Radiographic density is affected by such factors as radiation spectrum, film processing, setup technique, amount and type of filtration, screens, and scattered radiation.

Fig. 40 Typical radiographic exposure charts for (a) aluminum and (b) steel for a film density of 2.0 without screens that relate exposure to thickness of testpieces for several values of tube voltage. Charts for aluminum and steel were prepared specifically for an Andrex 160-kV directional x-ray machine, using a source-to-film distance of 910 mm (36 in.) and Industrex AA film (Eastman Kodak) developed in a manual process for 7 min in PIX developer (Picker).

Although published exposure charts are acceptable guides for equipment selection, more accurate charts that are prepared under normal operating conditions are recommended for each x-ray machine. A simple method for preparing accurate exposure charts is as follows:

1. Make a series of radiographs of a calibrated multiple-thickness step wedge, using several different values of exposure at each of several different tube voltage settings 2. Process the exposed films together under conditions identical to those that will be used for routine application 3. From the several densities corresponding to different thicknesses, determine which density (and thickness) corresponds exactly with the density desired for routine application. This step must be done with a densitometer because no other method is accurate. If the desired density does not appear on the radiograph, the thickness corresponding to the desired density can be found by interpolation 4. Using the thickness determined in step 3 and the tube voltage (kilovoltage) and exposure (milliampsecond or milliampmin) corresponding to that piece of film, plot the relation of thickness to exposure on semilogarithmic paper with exposure on the logarithmic scale 5. Draw lines of constant tube voltage through the corresponding points on the graph

The resulting chart will be similar to those in Fig. 40, but will be accurate for the x-ray machine, type of film, and film development technique that were used. Different charts should be prepared for different film types, both with and without screens.

Charts prepared as described above will be strictly accurate only for testpieces of uniform thickness. Some adjustment in exposure or source-to-film distance will have to be made for the more usual circumstances involving testpieces of nonuniform thickness. Latitude charts can sometimes be used to help determine the correct exposure when the thickness of a testpiece varies within the area of coverage. Latitude is a range of metal thickness that produces a specific range of density in the processed film for a given combination of radiation spectrum, source-to-film distance, and exposure. The bands in Fig. 41 are latitude curves for the inspection of steel testpieces using x-rays from a 250-kV x-ray machine. As shown in Fig. 41(a), increasing the tube voltage not only decreases the exposure needed to radiograph a section of given nominal thickness but also increases the range of testpiece thickness that is satisfactorily recorded on the radiograph. For example, at a nominal steel thickness of 15 mm ( in.), about 2200 mA-s exposure is needed to produce a satisfactory image on type 2 film with 150-kV x-rays; the image will have satisfactory density for a range of thickness from about 14 to 18 mm (0.57 to 0.70 in.). On the other hand, when 250-kV x-rays are used, an equivalent image can be produced on type 2 film with only about 220 mA-s exposure, and thicknesses from about 13 to 20 mm (0.50 to 0.78 in.) can be recorded with satisfactory density.

Fig. 41 Latitude curves for the radiographic inspection of steel at film densities ranging from 1.5 to 3.0. (a) Effect of tube voltage on latitude of a type 2 radiographic film. (b) Effect of film type on latitude for radiography using 250-kV x-rays. Curves were prepared from data obtained using a 250-kV x-ray machine, 910-mm (36-

in.) source-to-film distance, and 0.25 mm (0.010 in.) thick lead screens. Films were processed using a standardized technique.

Alternatively, latitude charts for different types of film can be used to determine the ranges of thickness covered by each type at a given exposure (Fig. 41b). The difference in latitude for different types of film is one basis for the use of multiple-film techniques. For example, if a steel testpiece is radiographed using 250-kV x-rays and a film holder loaded with one sheet each of types 1, 2, and 3 film, then a range of testpiece thickness of 13 to 40 mm (0.5 to 1.6 in.) can be recorded with a single exposure of 600 mA-s. Each of the films is viewed and interpreted separately for the range of thickness corresponding to its optimum density range. For thicknesses corresponding to gaps between optimum density ranges for the different films, enough contrast and detail usually exist in one or both of the films on either side of the gap to yield a satisfactory image. When filters are used or when other variations in technique are introduced, further adjustments in exposure or source-tofilm distance will have to be made. In fact, if large numbers of a given part are to be inspected, it is usually worthwhile to make a series of radiographs of a representative testpiece using different tube voltages and different exposures. From the results of these experiments a standard setup and exposure can be established. When there is only one opportunity to inspect a part for which there is no established standard technique or when circumstances will not permit experimentation, one of several alternative techniques can be used. For example, replicate radiographs can be made using different values of one or more of the following: tube voltage, tube current, exposure time, source-to-film distance, or film speed. Alternatively, film holders can be loaded with two or more sheets of film, either of the same type or of different types; the resulting radiographs can be viewed both as double-film and single-film images to obtain wider latitude. The latter technique, using two or more sheets of film, is especially useful when there are large variations in testpiece thickness. Sometimes, it can be advantageous to adopt the technique for routine applications. Another method involves making duplicate radiographs with and without a filter or with and without lead screens. In such cases, exposure time and tube current will have to be varied to compensate for either beam attenuation with a filter or the combined effects of filtration and intensification with lead screens. Exposure charts apply only to the material of which the step wedge was composed. Most often, this is a standard material such as aluminum or steel. However, exposure charts for a standard material can be used to determine exposure factors for other materials by applying radiographic equivalence factors such as those listed in Table 5. First, the exposure is derived from the exposure chart as if the part were actually made out of the standard material. The exposure so derived is then multiplied by the radiographic equivalence factor from Table 5. For example, if a 13 mm (

in.) thick part made of

titanium is to be radiographed using 150-kV x-rays without screens, the exposure for a 13 mm ( in.) thick part made of steel can be determined from the exposure chart in Fig. 40(b). That exposure would be 4000 mA-s for a source-to-film distance of 910 mm (36 in.). From Table 5, the radiographic equivalence of titanium for 150-kV x-rays is 0.45. Thus, for 150-kV x-rays and a source-to-film distance of 910 mm (36 in.), the exposure for the titanium part would be 0.45 times 4000 mA-s, or 1800 mA-s. A reasonable exposure for this part would be 2 min (120 s) at 15-mA tube current. In addition to exposure charts, nomograms or specially constructed slide rules are often used for calculating radiographic exposures. These devices can be constructed using the same type of information as that used in constructing exposure charts. The main advantage of these devices is speed in making a calculation, which can reduce setup time and produce economic benefits. Gamma-ray exposure charts are constructed in a manner similar to the exposure charts used in determining x-ray

exposures. Instead of expressing the exposure in milliampere-seconds, -ray exposures are expressed in curie-hours or curie-minutes. To use a -ray exposure chart, the source strength must be known. Source strength decreases exponentially with time, and each radioactive isotope had a characteristic half-life. This behavior can be used to determine the strength of a radioactive source at any one time, provided its strength at one time is known. (Normally, source manufacturers provide the source strength as of a given date along with each new source). A graph of source strength versus time is constructed as described in the following paragraph. On the logarithmic scale of semilogarithmic graph paper, plot the known source strength. (For convenience, the date when the determination of source strength was made should be noted.) Divide the linear scale into convenient units of time (with the known source strength corresponding to zero time), extending the scale at least one half-life. At the time corresponding to one half-life, plot a point corresponding to half of the known source strength that was plotted at zero

time. Draw a straight line between the two points, extending the line as far beyond the second point as desired. This line represents the source strength at any time; if desired, the linear scale may now be renumbered using dates, so that at any time the source strength corresponding to a given date can be seen at a glance. Figure 42 shows two types of -ray exposure charts for cobalt-60 radiation and testpieces of steel, gray iron, or ductile iron. In Fig. 42(a), exposure in curie-minutes to produce an average photographic density of 2.0 is read directly for each combination of testpiece thickness and source-to-film distance. Exposure time is determined by dividing exposure determined from the chart by the source strength determined as outlined in the preceding paragraph. In Fig. 42(b), an exposure factor is read directly for each combination of testpiece thickness and desired photographic density. The exposure time is calculated from:

(Eq 18) where t is the exposure time in minutes, E is the exposure factor, Li is the source-to-film distance (in inches), and S is the source strength (in curies). For example, if it is desired to radiograph a 3 in. thick steel testpiece on Industrex AA film with a photographic density of 3.0 and if a 15-Ci cobalt-60 source is to be used at a source-to-film distance of 12 in., then the corresponding exposure factor is 0.58 (dashed line, Fig. 42), and from Eq 18, the exposure time is t = 0.58 × 122/15 = 5.6 min.

Fig. 42 Two types of exposure chart for computing -ray exposures that apply to Co-60 radioisotopes and steel, gray iron, or ductile iron testpieces. (a) Exposure in curie-minutes to produce a photographic density of

2.0 on NDT 75 film (DuPont) as a function of testpiece thickness for various source-to-film distances. (b) Exposure factor, E, for Industrex AA film (Eastman Kodak) as a function of testpiece thickness; four lines of constant photographic density are shown.

As with x-ray exposures, latitude charts such as the one shown in Fig. 43 can be used to determine the range of thickness that can be recorded with a given exposure. By comparing Fig. 43 with Fig. 41, it can be seen that the latitude of type 2 film with cobalt-60 radiation is almost four times the latitude of type 2 film with 250-kV x-rays. Also, as with x-ray exposures, nomograms or slide rules can be constructed to simplify calculation of -ray exposures.

Fig. 43 Latitude curve for radiographic inspection of steel using Co-60 -rays to produce a film density range of 1.5 to 3.0. Curve is for type 2 film exposed at a source-to-film distance of 810 mm (32 in.) and with 0.25 mm (0.010 in.) thick lead screens both front and back.

Selection of View The view selected for the radiography of a testpiece is a major factor that controls the ability to detect certain types of flaws. In some circumstances, although flaws are detected, the selected view presents an unsatisfactory or distorted picture of the relationship of the flaws to testpiece shape. For example, a crack in the fillet of a T-shape section of a casting or in a weld in a T-shape weldment is most likely to be revealed when the radiation is directed along the bisector of the angle between the legs of the section, because most cracks in such a location run perpendicular to the surface midway between the legs of the T. When the radiation is directed at an equal angle to both legs, it is most likely to be parallel to any crack that is present, which is the most favorable orientation for revealing cracks. With this orientation, however, there is no distinct line of demarcation between the sections of different thickness, and the portions of the testpiece that have the greatest object-to-detector distance will be recorded with a considerable degree of unsharpness and distortion. Radiographic Shadows. As discussed in the section on "Principles of Shadow Formation" in this article, radiographic images can be compared to the shadows formed when a beam of light is interrupted by an opaque object. The difference between shadows and radiographic images is that in the umbra of a shadow there is a complete absence of radiation from the light source, while in the umbral region of a radiograph the amount of radiation that falls on the image conversion medium depends mainly on the absorption characteristics of the testpiece. Ordinarily, only the umbral region of a radiograph is of interest because it is only in this region that internal features of the testpiece are revealed.

When a section of the testpiece contains a discontinuity that is essentially planar, such as a crack or cold shut, the planar discontinuity will be revealed only when the radiation is parallel or nearly parallel to that discontinuity (Fig. 44a). Such a discontinuity can be revealed as a difference in absorption compared to the surrounding material only when the radiation impinges on the edge of the discontinuity, as shown at left in the actual setup in Fig. 44(a). In all other orientations, such

as that shown at right in the actual setup in Fig. 44(a), the effective thickness of the testpiece is the same, whether in the region of the planar discontinuity or not, and there is no difference in absorption.

Fig. 44 Effect of direction of radiation beam on the appearance of (a) planar discontinuities (such as cracks or cold shuts) and (b) globular voids. See text for discussion.

When a section of the testpiece contains a flaw that is spherical in shape, such as a globular void or inclusion, the flaw will be revealed regardless of viewing direction. However, as shown in Fig. 44(b), the shape of a void will be undistorted only when it is aligned with the portion of the radiation beam that is perpendicular to the plane of the recording medium. Distortion is less of a problem for long source-to-detector distances and small testpieces than for short source-to-detector distances or large testpieces. The size of almost all spherical flaws is difficult to assess. The thickness of a spherical flaw is greatest along a line through the center of the flaw and decreases to nil at the outer edge. Correspondingly, the thickness of the surrounding material is at least at the center of the spherical flaw. Thus, for example, the amount of radiation passing through the testpiece is greatest in the shadow of the center of a void. Progressively outward from the center, the image of the void will grow gradually less dense as the effective thickness of the void grows smaller, and finally the image of the void will fade into the image of the surrounding material. Images of spherical flaws are more distinct and true in size when the flaws account for a substantial portion of the thickness of the section or when the features have absorption characteristics that are markedly different from those of the matrix. Also, the image will be larger than the flaw itself, less distinct, and more likely to be distorted the farther the flaw is above the plane of the image conversion medium. Figure 45 shows the effect of viewing direction on the image of a solid block of material. When the center of the block is aligned with the central beam and one side of the block rests on the image conversion medium (in this case, film as shown in setup 1 in Fig. 45), the central region of the image presents an undistorted view of the lower surface of the block. Surrounding the undistorted region is a region representing an enlarged view of the upper portion of the block, and surrounding this region is the penumbra. When the block is tilted onto one edge but still aligned with the central beam (setup 2, in Fig. 45), the entire image is distorted and is surrounded by a distorted penumbra. When the block rests on the film but the radiation beam is inclined to both the block and the film (setup 3, Fig. 45), only the image of the lower surface of the block remains undistorted. The remainder of the image is enlarged or otherwise distorted; the greater the object-to-detector distance, the greater the enlargement or distortion. The penumbra is enlarged or distorted similarly to the way the image is affected.

Fig. 45 Effect of viewing direction on the radiographic image of a solid block of material. See text for discussion.

It is important to understand that almost all radiographic images are somewhat distorted in size or shape or both, as compared to the actual features they represent. When viewing or interpreting radiographs, the individual features in the image should be recognized as projected shadows of the actual features in the testpiece. Inspection of Simple Shapes It is usually best to direct the radiation at right angles to a surface, along a path that presents minimum thickness to the radiation. This not only increases the radiation intensity reaching the detector but, more important, also ensures that any internal feature (except for a planar feature that is not parallel to the radiation beam) will generate the greatest subject contrast. Obviously, for a flaw of a given size, the difference between the amount of radiation transmitted through the material in the area of the flaw and the amount transmitted through an adjacent area will be greater if the flaw occupies a greater portion of the testpiece thickness. When the presence of planar discontinuities (cracks) is suspected, radiation must be directed essentially parallel to the expected crack plane, regardless of testpiece thickness in that direction. In film radiography, any increased inspection cost resulting from increased exposure time is well justified because only a view parallel to the crack plane will reveal a crack. Flat plates have the simplest shape and are usually inspected by radiography as an integral part of a more complex

assembly or component. The most favorable direction for viewing a flat plate is one in which the radiation impinges perpendicular to the surface of the plate and penetrates the shortest dimension. The image is produced with minimum distortion and with maximum sensitivity and resolution of internal features. When inspecting a difficult-to-reach region of a complex-shape part, it may be necessary to avoid superimposing the image of one section on another, by selecting a viewing direction other than 90° for a flat-plate section, as shown in Fig. 46 with film as the conversion medium. Nevertheless, except when there is a good reason to do otherwise, it is always desirable to use a viewing direction as close to 90° as possible. Except in the most unusual circumstances, views parallel

or nearly parallel to the plate surface should be avoided. These views give severely reduced resolution of detail, a high degree of image distortion, and an excessively low level of radiation intensity transmitted through the testpiece.

Fig. 46 Use of offset radiation beam to avoid superimposing image of one section of a complex-shape part on another when inspecting difficult-to-reach regions

When large areas are being radiographed, they should be inspected as a series of radiographs, each one overlapping the area of coverage of all adjacent radiographs. Use of a relatively short source-to-detector distance and multiple, overlapping radiographs is frequently more satisfactory than a single large radiograph. The long source-to-detector distance necessary to avoid the type of distortion illustrated in Fig. 44 and 45 reduces the intensity of transmitted radiation and may require frame summing in real-time radiography or longer exposures in film radiography. In film radiography, the longer exposures of the single-exposure technique may make the single-exposure technique uneconomical as compared to a multiple-exposure technique. Curved plates are most satisfactorily inspected using views similar to those for flat plates. For optimum resolution of detail, the image conversion plane should be shaped to conform with that of the back surface of the curved plate. If the curved plate has its convex side toward the incident radiation, it is usually advantageous to minimize distortion by making multiple radiographs with reduced individual areas of coverage. If the curved plate has its concave side toward the incident radiation, a distortion-free image can be achieved by placing the source at the center of the radius of curvature. In this latter case, there is no angular limit to the area that can be inspected in a single radiograph, except for any limitations imposed by the directionality of the radiation field. Frequently, this technique is used for the on-site radiography of largediameter pipes. The source is placed on the centerline of the pipe, and simultaneous exposures are made on one or more films wrapped around the outside of the pipe. The entire circumference is inspected with one shot because the source emits equal amounts of radiation in all directions.

This panoramic technique can also be used for the simultaneous inspection or exposure of several small parts. In a panoramic exposure, testpieces are placed in a circle around a source that emits equal radiation intensity in all directions. Separate films are placed behind each testpiece, and all are exposed at one time. Solid cylinders can be inspected either by a longitudinal view, which is generally satisfactory only for short large-

diameter cylinders, or by a transverse view, which is most satisfactory for relatively small-diameter cylinders. If the cylinder is neither short nor small in diameter, either view poses problems to the radiographer. However, in such cases, the transverse view is usually satisfactory. The thickness of a cylindrical testpiece varies across a diametral plane in a similar manner as for a sphere, being thickest in the center and progressively thinning out to nil thickness at the edges. Edge definition is relatively good for light-metal cylinders less than about 50 mm (2 in.) in diameter and for heavy-metal cylinders less than about 25 mm (1 in.) in diameter. When the diameter of the cylinder exceeds these values, multiple radiographs using different tube voltages are usually needed to increase the latitude (or dynamic range) of the inspection. In real-time radiography, image processing can also increase the latitude for inspecting the inner and outer portions of the cylinder. In film radiography, double-film

techniques can also correct for the inevitable overexposure of the outer portions of the image. An alternative technique is to make a radiograph at the proper exposure for all but the outer portions of the image, then rotate the cylinder 90° and make a second radiograph at the same exposure. Sometimes, section-equalizing techniques are helpful in film (or paper) radiography. In a section-equalizing technique, the outer edges of the cylinder, where the section is thinnest, are built up to present a greater radiographic density to the xrays. Close-fitting solid cradles, liquid absorbers, or many layers of shim stock--all having radiographic absorption characteristics equivalent to those of the cylinder--are alternative means of equalizing radiographic density (Fig. 47). Regardless of the form of the absorber, it is placed in contact with the cylinder around one-half of the circumference, as shown in Fig. 47. The effect of a section-equalizing technique is to reduce the subject contrast by about one-half. This is usually sufficient to ensure that the entire image of the cylinder will have a photographic density within the range that can be interpreted correctly.

Fig. 47 Three alternative methods of equalizing radiographic density to avoid overexposure of the image of thinner portions of a solid cylinder. The solid cradle (a), liquid absorber (b), and shim stock (c) have radiographic absorption characteristics equivalent to those of the cylinder.

Because the probable orientation of planar discontinuities is difficult to define for cylinders, it is good practice to make more than one radiograph, rotating the cylinder equal amounts between them. Two radiographs, 90° apart, would be considered a minimum number. In general, hollow cylinders (or tubular sections) present fewer problems to the radiographer than solid cylinders. For example, in hollow cylinders or tubular sections, there is much less subject contrast because of the nonabsorbing cavity in the center. The inspection of hollow cylinders is discussed in the section "Inspection of Tubular Sections" in this article. Inspection of Complex Shapes The inspection of complex shapes most often requires multiple exposures, usually with different viewing directions. The selection of view for each exposure depends primarily on the shape of the section of the testpiece to be inspected with that exposure and the probable orientation of suspected flaws. In principle, the view should be chosen with the objectives of minimizing geometric unsharpness and of reducing the image to a simple shape for easy interpretation. A small variation in testpiece thickness is desirable because this will highlight subject contrast due to flaws. However, the use of highenergy sources, multiple radiographs with different radiation energy levels, or double-film techniques with a single view can circumvent most problems arising from excessive material-thickness variations over a given area of coverage. The following example illustrates the use of three different views to obtain complete coverage of a cast stainless steel impeller. Example 2: Three-View, Nine-Shot Inspection of a Large Cast CF-8M Stainless Steel Impeller. A large impeller cast from CF-8M stainless steel (Fig. 48) was specified to be radiographically inspected in accordance with ASTM E 94, E 1030, and E 142, using the standard radiographs of ASTM E 186 and E 446 as a reference for

acceptance. The impeller was to be inspected in the as-cast, center-bored condition; specified image quality with plaque penetrameters was 2-2T.

Fig. 48 Cast CF-8M stainless steel impeller that required three different viewing directions and nine separate shots to provide complete coverage in radiographic inspection. Dimensions given in inches

Three different views were chosen to provide complete coverage of the impeller. These are shown in Fig. 48 as view A, selected to inspect the hub (a hollow cylinder) with a single shot; view B, using two shots with each providing coverage of half of the 18 volutes in the impeller; and view C, using six shots with each providing coverage of the thicker section of three volutes in a direction that is perpendicular to view B. View A presented certain problems, namely, the high absorption of the 115 mm (4 in.) thick hub, which required highintensity radiation of relatively high energy for adequate penetration, and the bore area along the central axis, which allowed incident radiation to impinge directly on the film and favored undercut because of side scattering along the cylindrical wall of the bore. The hub could not be radiographed in a radial plane because of excessive interference by the surrounding volutes, so the bore was filled with lead shot to absorb the direct radiation in this region and to eliminate undercut. All radiographs were made with a source-to-film distance of 1.5 in (5 ft), and 1-MV x-rays were used from a source having an 8 mm (0.3 in.) focal-spot size. Shot 1 (view A) was made with 3.0 mA tube current and 4.7-min exposure time on a 180 × 430 mm (7 × 17 in.) double film of NDT 75 (DuPont). Shots 2 and 3 (view B) were made with 1.0-mA tube current and 2.2-min exposure time on three 350 × 430 mm (14 × 17 in.) films of different types--one sheet each of NDT 45 (DuPont), Industrex M, and Industrex AA (Eastman Kodak). Shots 4 to 9 (view C) were made with 1.0-mA tube current and 2.5-min exposure time on 125 × 175 mm (5 × 7 in.) double films of Industrex M (Eastman Kodak). Lead screens, 0.13 mm (0.005 in.) thick, were used both front and back.

For control of image quality, standard ASTM penetrameters were used. The penetrameter for view A was 2 mm (0.08 in.) thick and was placed on a 110 mm (4

in.) thick block adjacent to the impeller. Four penetrameters were used for view

B: a 0.25 mm (0.010 in.) thick penetrameter on a 13 mm ( a 20 mm (

in.) thick block, a 0.38 mm (0.015 in.) thick penetrameter on

in.) thick block, a 0.5 mm (0.020 in.) thick penetrameter on a 25 mm (1 in.) thick block, and a 1.15 mm

(0.045 in.) thick penetrameter on a 57 mm (2 in.) thick block. For view C, a 0.38 mm (0.015 in.) thick penetrameter was placed directly on the part. All penetrameters and blocks were made of austenitic stainless steel. This procedure gave both the required complete coverage and the specified 2-2T image quality. Castings were routinely inspected to a minimum quality level corresponding to severity level 1 (in ASTM E 186) in the hub area and to severity level 2 to 3 (in ASTM E 446) in the shroud area. Inspection of Weldments The radiographic inspection of weldments is usually intended to cover only welds and heat-affected zones. The quality of the base metal beyond the heat-affected zones is seldom of interest at this stage of inspection. Therefore, the viewing direction is usually selected to reveal clearly the features of welds and heat-affected zones, regardless of the overall shape of the weldment. See the article "Weldments, Brazed Assemblies, and Soldered Joints" in this Volume for further information on the NDE of welds. Arc Welds. Generally, arc-welded joints are radiographed using a viewing direction normal to the surface of the weld.

For butt joints, this view directs the central beam perpendicular to the surfaces of the adjacent plates, but for other types of joints such as T-joints and lap joints, the central beam is usually inclined to both adjacent legs. The maximum area of coverage for a single radiograph is normally recommended to be no more than would result in a 6 to 10% difference in penetrated thickness between the center of the area and the extremities. Mathematically, this can be expressed for flat plates of uniform thickness as:

(Eq 19) where x is the field size, L is the source-to detector distance, and the values 0.35 and 0.46 are derived from the laws of similar triangles for 6 and 10% increase in penetrated thickness, respectively. The numerical range in Eq 19 is equivalent to a radiation-beam cone angle of about 39 to 49°. This range is not intended to limit the thickness range inspected in a single radiograph but rather to ensure the detection of transverse cracklike discontinuities by keeping the image relatively sharp and undistorted.

Views normal to the surface of the weld are not always satisfactory. Flaws in a weld such as dense inclusions due to the deposition of metal from a tungsten electrode, excessive segregation in light alloys, porosity, and shrinkage can be revealed regardless of view. However, cracks or nonfusion along the edge of the fusion zone can be best revealed when the viewing direction is parallel to the face of the joint. The detection of centerline cracks or incomplete penetration requires a normal view. Only square-groove butt joints, U-groove joints, and J-groove joints can be satisfactorily inspected with a normal view; other types of joints, including V-groove joints, bevel joints, flare joints, and fillets, require two views, one parallel to each original prepared edge, for correct inspection for cracks or regions of incomplete fusion. Certain flaws cannot be detected at all because they lie along planes that are more or less parallel to the weld surface. The shape of most weldments prohibits selection of a viewing direction parallel to the weld surface, which is the most favorable view for detecting underbead cracking or delamination. (Ultrasonic inspection is a much more suitable method for detecting either underbead cracking or delamination.) The most favorable view for fillet welds is a direction that roughly bisects the angle between the legs of the section. Preferred arrangements for several types of fillet welds are shown in Fig. 49. Single-fillet T-joints can be inspected with the simple arrangement shown in Fig. 49(a) or with an arrangement incorporating an equalizing wedge (Fig. 49b) to reduce the inherent subject contrast caused by variations in penetrated thickness across the weld zone. The arrangement in Fig. 49(b) is more relevant for film radiography than for real-time radiography because image processing can be used in real-time radiography. In film radiography, the high subject contrast can result in overexposure of the thin section on one side of the weld or in underexposure of both the weld and thick sections. Either underexposure or overexposure reduces

sensitivity and may not resolve images of flaws having low subject contrast. An equalizing wedge is most effective when the size of the fillet weld exceeds the thickness of the thin section. The technique is mainly used on straight seams; special-shape wedges would be required for other seam shapes.

Fig. 49 Preferred viewing directions for the radiographic inspection of several types of fillet-welded joints. (a) Single-fillet T-joint. (b) Single-fillet T-joint with equalizing wedge. (c) Two adjacent single-fillet joints radiographed simultaneously. (d) Double-fillet T-joint. (e) Corner joint with film positioned at the inside surface. (f) Corner joint with film positioned at the outside surface. (g) and (h) Alternative views for double-welded lap joint

Occasionally, the shape of the weldment is such that two or more seams can be inspected with a single exposure in film radiography (Fig. 49c). The savings in time and material often justify selection of a single view whose area of coverage encompasses two or more seams. However, if the shape of the seam is such that a single view would give severely reduced sensitivity, multiple views should be specified on the radiographic standard shooting sketch. T-joints that are welded on both sides usually cannot be inspected satisfactorily with a single radiographic view or exposure. Two radiographs, one for each fillet (Fig. 49d), are needed. Corner welds can be inspected from either side, depending on which side provides access (Fig. 49e and f). Lap joints can be inspected using multiple views or by a single view, as shown in Fig. 49(g) and 49(h). The main factor that determines which technique is most appropriate is the amount of separation between welds in a double-welded joint. When multiple views are used, the central beam of radiation should be aligned with the weld for each view, as in Fig. 49(g). When a single view is used, the central beam should be aligned with a spot midway between the welds, as in Fig. 49(h). Only a single radiograph is needed for each weld segment in a single-welded lap joint. The techniques discussed above, particularly those for the inspection of fillet welds, are also appropriate for the inspection of junction areas of castings, forgings, and formed parts. Circumferential welds, which are one of the more usual types of arc welds inspected by radiography, can be inspected with techniques discussed in the section "Inspection of Tubular Sections" in this article. Resistance welds, mainly spot welds and seam welds, are used to join relatively thin sheets. Most often, a resistance-

welded joint can be considered a type of lap joint, with the weld positioned inboard of the overlapping edge. Usually, the only outward sign of weld position is a slight depression on the surface of the workpiece resulting from electrode pressure applied during welding. In the radiography of spot welds and seam welds, the central beam is ordinarily directed normal to the surface of one of the sheets and is centered on the depression that indicates the location of the weld. Radiography can detect discontinuities in both weld metal and base metal, including cracks, inclusions and porosity in the weld, cracking and deformation in the

base metal, segregation in the fusion zone of resistance welds in light alloys, and expulsion of molten metal between the faying surfaces. Normally, deficiencies such as underbead cracking and incomplete fusion are not revealed by radiography, although incomplete fusion can sometimes be inferred from the presence of other discontinuities. The area of coverage should be limited in a manner similar to that suggested for arc welds so that maximum sensitivity to cracklike discontinuities can be maintained. A view other than the normal view is sometimes dictated by the shape of the weldment, especially when portions of the weldment may cast shadows that would detract from the clarity of the weld-zone image. Even in such cases, it is best to select a view as near to normal as possible so as to minimize distortion. In contrast to arc welds, in which an oblique view adversely affects both the intensity of the transmitted radiation and the sensitivity, only sensitivity is affected to any great degree by using an oblique view for the inspection of resistance welds. Most resistance-welded assemblies are thin enough that the section can be penetrated by relatively low-energy radiation even with oblique views. The sensitivity, expressed as minimum detectable discontinuity size, varies with the cosecant of the angle between the viewing direction and the sheet surface. This relationship indicates that the smallest value of minimum detectable discontinuity size occurs with a 90° viewing direction, and viewing directions between 90° and 65° give no more than about a 10% increase in this value. The main disadvantage of oblique views is their inability to resolve the images of transverse cracks (throughthickness cracks) in either the weld or adjacent heat-affected zones. Inspection of Tubular Sections Although film radiography is not well suited to the inspection of continuously produced tubular products such as tubing and pipe, the method is suited to the inspection of tubular sections in a wide variety of products and assemblies. Real-time radiography, on the other hand, is more suited to the inspection of continuously produced tubular products, but not tubular sections in large, complex assemblies. Radiography is the method most often specified for the inspection of welded joints between tubes or pipes or between a tube and a pressure vessel when the weldment is manufactured to ASME Boiler and Pressure Vessel Code specifications. Other hollow sections such as cylindrical bosses on castings and forgings, brazed tube-and-socket joints, large-diameter transmission pipe (line pipe), and longitudinally welded square or circular structural tubing, are frequently inspected by radiography. There are three major inspection techniques for tubular sections: • • •

The double-wall, double-image technique The double-wall, single-image technique The single-wall, single-image technique

The double-wall, double-image technique is mainly applicable to sections with an outside diameter of no more

than 90 mm (3 in.). This technique produces a radiograph in which the images of both walls of a tubular section are superimposed on one another. The beam of radiation is directed toward one side of the section, and the radiation conversion medium surface is placed on the opposide side, usually tangent to the section. As shown in Fig. 50(a), two radiographs 90° apart are required to provide complete coverage when the ratio of outside diameter to inside diameter is 1.4 or less. In exposure 1 in Fig. 50(a), the area between the phantom lines is recorded as a through-thickness image, and the area outside the lines is recorded essentially as an image in the plane of the tube wall that exhibits too much subject contrast for meaningful interpretation. In exposure 2, the area outside the lines in exposure 1 is recorded as a through-thickness image, and there is a certain amount of overlap between the through-thickness images of exposures 1 and 2, which is shown as darker shading in exposure 2 in Fig. 50(a).

Fig. 50 Diagrams illustrating the relation of radiation beam, testpiece, and film for the double-wall, doubleimage technique of inspecting hollow cylinders or welds in tubular sections. Dimensions given in inches

When the ratio of outside diameter to inside diameter is greater than 1.4--that is, when radiographing a thick-wall tube-the number of views required to provide complete coverage can be determined by multiplying that ratio by 1.7 and rounding off to the next higher integer. For example, to examine a 50 mm (2 in.) OD cylinder with a 25 mm (1 in.) diam axial hole, a total of 1.7 × 2 = 3.4, or 4 views, will provide complete coverage. The circumferential displacement between each view is found by dividing 180° by the number of shots. In this case, 180°/4, or 45°, rotation of the cylinder with respect to the viewing direction is required between exposures, as shown in Fig. 50(b) The area within the phantom lines for each exposure in Fig. 50(b) indicates the region of a 50 mm (2 in.) OD cylinder with a 25 mm (1 in.) diam axial hole that is recorded as a through-thickness image; heavy shading indicates regions recorded as a through-thickness image on preceding views. If only three shots are used for a cylinder with a 2:1 ratio of outside to inside diameters, there is only a minimal amount of overlap between shots, inadequate to ensure that all detectable discontinuities in the regions of overlap will be recorded. When an odd number of views are required for complete coverage, the angular spacing between shots can be determined by dividing 360° by the number of views as an alternative to dividing 180° by the number. This alternative cannot be used when the number of views is even, because half of the resulting radiographs would be mirror images of the remaining radiographs and sections of the outside circumference would not receive adequate coverage. The double-wall, double-image technique can be applied to the inspection of diagonally opposite corner welds in a section of rectangular structural tubing constructed by welding four pieces of metal strip together to form a box section. Only two views are required for the inspection of all four longitudinal welds, as shown in Fig. 50(c). As with circular sections, hollow box sections can be inspected by the double-wall, double-image technique only when they are small; that is, when section height in direction or selected view is 100 mm (4 in.) or less. A variation of the double-wall, double-image technique, sometimes called the corona or offset technique, is often used for the inspection of circumferential butt welds in small-diameter tubing and pipe. In the corona technique, the central beam is directed at an acute angle to the run of the tube (Fig. 50d) so that the weld is projected on the film as an ellipse rather than a straight band. The offset angle of the radiation beam shown in Fig. 50(d) must be large enough that the image of the upper section of the weld zone does not overlap the image of the lower portion, but not so large as to introduce an unnecessary degree of distortion. Also, the larger the offset angle, the greater the probability that the technique will fail to detect incomplete fusion at the root of the plain butt weld. Incomplete fusion at the root can have the radiographic appearance of a root void, but if it exists in a tightly butted joint or if the root gap is filled by capillary action of the

molten weld metal, incomplete fusion has the appearance of a cracklike discontinuity lying in the plane of the joint. The incident radiation must be parallel to the plane of the joint so as to detect incomplete root fusion in a tight joint. The correct number of views and the circumferential location of corresponding views can be determined for the corona technique in the same manner as for the basic double-wall, double-image technique. Sometimes, as when a weld must be inspected for both radial cracks and internal voids, it is beneficial to use both the basic technique and the corona technique--the former to detect cracks and the latter to locate voids. In all variations of the double-wall, double-image technique, the image of the section of the cylinder wall that is closest to the radiation source will exhibit the greatest amount of enlargement, the greatest degree of shape distortion, and the greatest degree of unsharpness. Selection of proper source-to-detector distance should be based on obtaining the specified image quality for that section of the cylinder. Also, penetrameters should be located where they will evaluate the image of the section closest to the source. The double-wall, single-image technique is mainly applicable to hollow cylinders and tubular sections exceeding

90 mm (3 in.) in outside diameter. This technique produces a radiographic image of only the section of the wall that is closest to the imaging plane, although the radiation penetrates both walls. The source is positioned relatively close to the section, so that blurring caused by geometric unsharpness in the image of the cylinder wall closest to the source makes the image completely indistinguishable. Only the image of the wall section closest to the detector is sharply defined. In film (or paper) radiography, exposures are calculated on the basis of double the wall thickness of the hollow section, as they are for the double-wall, double-image technique. The area of coverage is limited by geometric unsharpness and distortion at the extremities of the resolved image for hollow cylinders that are less than about 380 mm (15 in.) in outside diameter. For larger cylinders, film size is the usual limiting factor. Usually, at least five separate exposures equally spaced around the circumference are required for complete coverage. There must be enough overlap between adjacent exposures to ensure that all of the outside circumference is clearly recorded. The radiation is usually directed normal to the surface of the cylinder. However, in some cases, it may be advantageous to use viewing direction similar to that of a corona exposure. For example, if a circumferential weld has a high and irregular crown, it may be desirable to have the radiation that impinges on the area of the weld being inspected pass through the tube wall adjacent to the weld so that the incident radiation is attenuated by a more uniform and thinner section, as shown in Fig. 51.

Fig. 51 Schematic of the double-wall, single-image inspection technique applied to a circumferential butt weld in a large-diameter pipe showing relation of radiation beam, weld, and film

The single-wall, single-image technique is an alternative to either of the double-wall techniques and can be used

only when the interior of a section is accessible. With this technique, the radiation source can be placed outside the cylinder and the radiation detector inside the cavity (Fig. 52a), or the detector can be placed outside the cylinder and the radiation source inside (Fig. 52b). In both setups, only a single wall is radiographed. A significant advantage of the singlewall technique is that it is more sensitive than double-wall techniques because any flaw that may be present occupies a greater percentage of the penetrated thickness.

Fig. 52 Schematic of techniques employing different arrangements of radiation source, cylinder, and film for the single-wall, single-image radiography of cylindrical sections using external and internal radiation sources. See text for discussion.

When using the technique shown in Fig. 52(a), the cylinder is rotated to generate a series of radiographs. In film radiography, the amount of rotation must be controlled so that successive radiographs contain regions of overlap. The correct amount of overlap can be proved by placing lead view markers on the surface that is remote from the film; if the image of a view marker appears in both adjacent radiographs, complete coverage is ensured. In the technique shown in Fig. 52(b), the radiation source is placed inside the cylinder, and one continuous piece of film is wrapped around the outside of the cylinder. Alternatively, several overlapping pieces can be used, as shown in Fig. 52(c). The radiation source, which is usually either a rod-anode type of x-ray tube having 360° radiation emission or a -ray source, is placed on the central axis of the hollow cylinder. The entire circumference is inspected with one exposure, regardless of whether a single strip of film or several overlapping pieces are used. Special adaptations of these techniques can be used. For example, when maximum sensitivity is required for the technique in Fig. 52(b), the radiation source can be moved to a position close to the inside surface of the cylinder diametrically opposite a single piece of film, as illustrated in Fig. 52(d). This adaptation can almost double the source-to-film distance, thus reducing geometric unsharpness by almost 50%. Although there is a gain in sensitivity, there is a loss in efficiency because the entire circumference can no longer be inspected by a single exposure. Several exposures must be made, with the radiation source and film repositioned to several points equally spaced around the circumference to obtain complete coverage. Another special adaptation is shown in Fig. 52(e). In this arrangement, a continuous piece of film is placed in contact with the inside surface of the cylinder. The film is held in place by a close-fitting, lead-covered mandrel. Incident radiation is directed through a slit in a lead diaphragm to impinge on a narrow section of the circumference of the cylinder. The cylinder, film, and mandrel are rotated slowly past the slit one or more times, so that the total time that a given point on the cylinder is in front of the slit equals the required exposure time. Radiographs produced in this manner are remarkably clear and distinct compared to other single-wall or double-wall techniques. Because the radiation beam impinges on only a narrow region of the cylinder and because the film is in contact with the inner surface, both geometric unsharpness and variations in transmitted intensity caused by variations in penetrated thickness are minimized. As a result, the subject contrast and definition of internal flaws are maximum. The technique shown in Fig. 52(e) is a form of in-motion radiography. In-Motion Radiography

In-motion radiography is useful when large areas must be inspected. Generally, real-time radiography is more suitable for in-motion radiography because a detector can be used to scan the test area during motion. Nevertheless, in-motion film radiography can be used to inspect longitudinal welds in very large pipe, such as line pipe. In one method, a portable xray unit is attached to a traveling carriage inside the pipe and is positioned so that a narrowly collimated x-ray beam impinges on the weld, as shown in Fig. 53(a). The carriage moves slowly along the pipe, successively exposing a series of films placed along the outside surface over the weld. The correct rate of travel is equal to the width of the beam in the direction of relative motion divided by the required exposure time.

Fig. 53 Arrangements for the in-motion single-wall, single-image film radiography of longitudinal welds in line pipe, using (a) a conventional portable x-ray unit and (b) a rod-anode x-ray tube

An alternative setup for in-motion film radiography, in which a rod-anode x-ray tube is substituted for the portable x-ray unit, is shown in Fig. 53(b). The principle of operation is the same; only the x-ray equipment and method of travel are different. This setup is particularly advantageous when two or more welded joints are to be radiographed because the joints can be radiographed simultaneously. With a rod-anode tube, radiation is emitted 360° around the axis of the tube. As long as the center of the target is maintained on the axis of the pipe, the disk of radiation that emerges between the two lead disks that act as a collimator will produce equal nominal film densities on all radiographs. Even with the portable xray unit illustrated in Fig. 53(a), the source-to-film distance must be kept constant because if the distance is permitted to vary there will be a gradual change in nominal film density from radiograph to radiograph as the radiation source moves along the pipe. The principal advantage of in-motion radiography is that less time is needed for setup and operation. The inspection of long welds by conventional methods requires frequent repositioning of the testpiece with respect to the source and film; consequently, there is a considerable amount of time spent in setting up for successive exposures. With in-motion radiography, a much smaller proportion of total inspection time is required for setup. Therefore, in-motion radiography is seldom, if ever, accomplished with a radioactive source. Techniques similar to those described above for radiographing longitudinal welds in large-diameter pipe can also be adapted for other product forms when radiographic inspection is to be applied to long testpieces. Also, it may be desirable to adapt in-motion radiography for the inspection of relatively small parts, especially when the diverging beam used in normal radiography would cause unacceptable distortion at the edges of the radiograph. In general, a real-time radiographic technique is preferred for in-motion radiography. However, if only film techniques are available, then in-motion radiography would be considered whenever more than four to six shots are required for complete coverage of a long straight or circumferential weld in a relatively thin section. It is essential that the source move at constant speed in film radiography because any variations in speed will produce parallel bands of overexposure and underexposure in the radiograph, oriented at right angles to the direction of relative motion. Motion Unsharpness. The real-time and film techniques used for in-motion radiography cause a specific type of unsharpness called motion unsharpness. Any amount of relative motion of source or testpiece with respect to the image conversion plane results in blurring of the image; the greater the amount of motion, the greater the blurring. In conventional still radiography, it is considered good practice to eliminate all sources of motion, even vibrations, so that motion unsharpness does not detract from radiographic quality. The amount of unsharpness caused by relative motion of the source with respect to the testpiece and detector in in-motion radiography can be evaluated from characteristics of the setup, much as geometric unsharpness is evaluated for conventional techniques. Motion unsharpness, Um, is described by:

(Eq 20) where w is the width of the radiation beam in the direction of motion, t is the penetrated thickness, and Lo is the source-toobject distance. Motion unsharpness is generally greater for in-motion radiography than geometric unsharpness for an equivalent conventional technique. This limits the use of in-motion techniques to thin testpieces or to situations in which a long source-to-detector distance can be used. It is essential that there be no relative movement between testpiece and detector, so that only the movement of the source contributes to motion unsharpness. Radiographic Inspection Revised by the ASM Committee on Radiographic Inspection*

Identification Markers and Penetrameters (Image-Quality Indicators) It is important to relate a particular radiographic image to the direction of the radiation beam used to generate that image. Furthermore, it is important that any permanent record of inspection be traceable to the particular testpiece that was inspected or to the production lot represented by that testpiece. Identification markers are used for these purposes.

In addition to identification, the image should contain some means of evaluating the radiographic procedure in terms of its sensitivity to test conditions. This sensitivity, known as image quality, bears directly on the ability of the radiographic process to record images of small flaws and is usually determined by the use of penetrameters, which are also called image-quality indicators. Identification Markers Identification markers are made of lead or lead alloy and are usually in the form of a coded series of letters and numbers. The markers are placed on the testpiece or on the film adjacent to the testpiece during setup. When the testpiece is radiographed, a distinct, clear image of the radiographically dense identification markers is produced at the same time. Identification markers must be located so that their projected shadows do not coincide with the shadows of any regions being inspected in the testpiece. Because markers are radiographically dense, their images will obscure any coinciding image of the testpiece. Both view identification and testpiece identification almost always appear in coded form. View identification is usually a simple code (such as A, B, C, or 1, 2, 3) that relates some inherent feature of the testpiece or some specific location on the testpiece to the view used; the term sequence numbering, which is used in some specifications, refers to view identification. Often, the location of view markers is handwritten in chalk or crayon directly on the testpiece so that the radiographic image can be correlated with the testpiece itself during interpretation and evaluation of the radiograph. Sometimes, the locations of markers are steel-stamped on the surface of the part being radiographed and become a permanent reference. It is advisable always to mark the location of view markers because it may be necessary to recheck the setup procedure against specifications. Although the specific location of view markers is largely a matter of personal preference, several suggested arrangements are discussed in this section. It is required only that the markers be placed so that their images on the radiograph are legible and do not obscure any area being inspected. View markers need not be different if identical viewing directions are used for radiographing a series of testpieces of the same type and size. Actually, it is much simpler to begin at the letter "A" or the number "1" for view identification on each individual testpiece. Most radiographers adopt standard methods of identifying views and indicate the type of view markers to be used on the specification sheet for the radiographic procedure, which is commonly called a radiographic standard shooting sketch. The code for identification of the testpiece is usually more complex than the code used for identification of viewing direction. Each radiographic laboratory must adopt a system of identification that is suited to its specific requirements. As a minimum requirement, the identification code must enable each radiograph to be traced to a particular testpiece or section of a testpiece. Identification codes can be based on part number, lot number, inspection date, customer code, or manufacturing code; or they can be merely a series of consecutive multiple-digit numbers, with the pertinent data concerning testpiece identification recorded in a logbook opposite the corresponding testpiece number. Penetrameters Penetrameters, or image-quality indicators, are of known size and shape and have the same attenuation characteristics as the material in the testpiece. They are placed on the testpiece or on a block of identical material during setup and are radiographed at the same time as the testpiece. Penetrameters are preferably located in regions of maximum testpiece thickness and greatest testpiece-to-detector distance and near the outer edge of the central beam of radiation. Because of this location, the degree to which features of the penetrameter are visible in the developed image is a measure of the quality of that image. The image of the penetrameter that appears on the finished radiograph is evaluated during interpretation to ensure that the desired sensitivity, definition, and contrast have been achieved in the developed image. Penetrameters of different designs have been developed by various standards-making organizations; several of the standard designs most widely used are shown in Fig. 54. Regardless of the design, all penetrameters have the following in common: •

•

Material used for penetrameters is the same as that of the testpiece or has the same absorption characteristics. Suitable penetrameter materials for various metallic testpieces are grouped in ASTM E 1025 and E 142 as shown in Table 8 In use, the penetrameter is normally placed directly on the surface of the testpiece that faces the source. Alternatively, when the testpiece is small or when its shape is unfavorable, the penetrameter is placed on

• •

• •

a block or shim of the same nominal composition and thickness as the testpiece, with the upper surface of the block at the same distance from the recording plane as the upper surface of the testpiece. In pipe radiography, when permitted by specification, the penetrameter can be placed on the surface of the pipe that faces the film. This film-side placement of the penetrameter is usually read to a tighter sensitivity requirement (for example, 2-1T instead of 2-2T) The location of the penetrameter should be such that its projected image does not coincide with any area of interest in the image of the testpiece The image of the penetrameter is viewed at the same time as the image of the testpiece; the two images appear simultaneously on the recording medium. The image of the penetrameter is evaluated separately from the image of the testpiece and is used solely as a direct check on the quality of the recorded image Image-quality levels are usually expressed as the size of the smallest penetrameter feature (such as hole size or wire size) that is clearly visible in the processed image A penetrameter is never used as a size standard against which flaw sizes are compared

Table 8 ASTM penetrameter material grades for the material groups in ASTM E 1025 and E 142 Material groups of testpieces

Suitable penetrameter grade and material

Group Mg: magnesium

Penetrameter grade 000: made of all magnesium, or magnesium being the predominant constituent

Group Al: aluminum

Penetrameter grade 00: made of all aluminum, or aluminum being the predominant constituent

Group Ti: titanium

Penetrameter grade 0: made of all titanium, or titanium being the predominant constituent

Group 1: all carbon steels, all low-alloy steels, all stainless steels, manganese-nickelaluminum bronze

Penetrameter grade 1: carbon steel or type 304 stainless steel

Group 2: all aluminum bronzes, all nickel-aluminum bronzes, Haynes alloy IN-100

Penetrameter grade 2: may be aluminum bronze (Alloy No. 623 of ASTM Specification B 150) or equivalent, or nickel-aluminum bronze (Alloy No. 630 of Specification B 150) or equivalent

Group 3: nickel-chromium-iron alloys and 18% Ni maraging steel. Some specific alloys include: Haynes alloy No. 713C, Hastelloy alloy D, G.E. alloy SEL, Haynes Stellite alloy No. 21, GMR-235 alloy, Haynes alloy No. 93, Inconel X, Inconel 718, Haynes Stellite alloy No. 6 S 816

Penetrameter grade 3: nickel-chromium-iron alloy UNS N06600 or equivalent

Group 4: nickel, copper, all the nickel-copper series or copper-nickel series of alloys, all the brasses (copper-zinc alloys) exclusive of leaded brasses. There is no restriction on using grade 4 penetrameters for the leaded brasses, because they are more attenuating of radiation, the degree depending on the lead content. This would be equivalent to using a lower grade penetrameter, which is permissible.

Penetrameter grade 4: may be 70Ni-30Cu alloy (Monel) (Class A or B of ASTM Specification B 164) or equivalent, or 70Cu30Ni alloy (Alloy G of ASTM Specification B 161) or equivalent

Group 5: tin bronzes including gun metal and valve bronze, but excluding any leaded tin bronzes of higher lead content than valve bronze. There is no restriction on using grade 5 penetrameters for bronzes of higher lead content,

Penetrameter grade 5: tin bronze (Alloy D of ASTM Specification B 139) or equivalent

because they become more attenuating of radiation with increasing lead content. This would be equivalent to using a lower grade penetrameter, which is permissible.

Fig. 54 Designs of several widely used penetrameters (image-quality indicators). (a) Rectangular plaque-type penetrameter (ASTM-ASME standard) for plaque thicknesses of 0.13 to 1.3 mm (0.005 to 0.050 in.). (b) Circular plaque-type penetrameter (ASTM-ASME standard) for plaque thicknesses of 4.6 mm (0.180 in.) or more. (c) Typical wire-type penetrameter (Deutsche Industrie Norm standard DIN 54109). (d) Square-step step wedge penetrameter used by British Welding Research Association (BWRA standard). (e) Hexagonal and (f) linear triangular-step step wedge penetrameters used by the French Navy (AFNOR standard). Dimensions given in inches, except where otherwise indicated

Applicable codes, specifications, or purchase agreements usually determine the type of penetrameter to be used. Even when the specification does not require the use of a penetrameter, it is advisable to use a penetrameter to ensure that appropriate image quality has been achieved. Some of the standard penetrameters, including those illustrated in Fig. 54, are described below. Plaque-type penetrameters consist of strips of materials of uniform thickness with holes drilled through them. There are two general types of plaque penetrameters specified by ASTM and the American Society of Mechanical Engineers (ASME): rectangular plaque penetrameters (Fig. 54a) and circular plaque penetrameters (Fig. 54b).

The rectangular plaque design shown in Fig. 54(a) is specified by ASTM and ASME for plaque thicknesses of 0.13 to 1.3 mm (0.005 to 0.050 in.). The holes in rectangular plaque penetrameters are T, 2T, and 4T in diameter, where T is the thickness of the plaque. A notch system (Fig. 55) is also used to identify the ASTM grade of the penetrameter. For example, the penetrameter shown in Fig. 54(a) would be a grade 1 penetrameter for testpiece materials in group 1 of Table 8.

Fig. 55 Identification system of ASTM penetrameter material composition grades

The circular plaque design is larger than the rectangular plaque design and is specified for plaque thicknesses of 1.5 to 4 mm (0.060 to 0.160 in.). Figure 54(b) shows the circular design specified by ASTM and ASME for plaque-type penetrameters with thicknesses of 4.6 mm (0.180 in.) or more. Various degrees of image quality can be measured by using plaque-type penetrameters of different thicknesses. Sensitivity is usually expressed in terms of penetrameter thickness (as a percentage of testpiece), and resolution is determined by the smallest hole size visible in the radiograph. For example, an image-quality level of 2-2T indicates that the thickness of the penetrameter equals 2% of section thickness and the 2T hole is visible. If image quality of 1-1T were required, a radiograph would be acceptable if the outline of a 1% penetrameter were distinguishable. Alternatively, image quality can be expressed as a percentage only. In the ASTM or ASME systems, the equivalent sensitivity in percent is based on visibility of the 2T hole. Table 9 lists equivalent sensitivities for various standard image-quality levels. Table 9 Equivalent sensitivities of various standard ASTM or ASME sensitivity levels Equivalent sensitivity is a percentage equivalent for penetrameter thickness in which 2T is the smallest distinguishable hole size. For example, 1-1T is equivalent to 0.7-2T. Image-quality level

Penetrameter thickness, % of testpiece thickness

Smallest visible hole size

Equivalent sensitivity, %

1-1T

1

1T

0.7

1-2T

1

2T

1.0

2-1T

2

1T

1.4

2-2T

2

2T

2.0

2-4T

2

4T

2.8

4-2T

4

2T

4.0

Wire-type penetrameters are widely used in Europe, and a standard design is used in the United Kingdom, Germany, the Netherlands, and Scandinavia and by the International Organization for Standardization (ISO) and the International Institute of Welding (IIW). In the United States, the penetrameter design specified in ASTM E 747 is widely used. The ISO design of wire-type penetrameters has a group (typically seven) of straight 30 mm (1.2 in.) wires made of the same material as the testpiece. The diameters of the seven wires are sized in a geometric progression from a range of 21 wire sizes with a numbered geometric progression of diameters ranging from wire number 1 (0.032 mm, or 0.00126 in., in diameter) to wire number 21 (3.200 mm, or 0.126 in., in diameter). This ISO standard is similar to the standard of Deutsche Industrie Norm (DIN 54109), which consists of sixteen wire sizes of three metals--steel, aluminum, and copper. However, the wire numbers in the DIN standard are the reverse of the ISO standard; in the DIN standard, wire diameters decrease in geometric progression from wire number 1 (which has a 3.20 mm, or 0.126 in., diameter) to wire number 16 (which has a 0.10 mm, or 0.004 in., diameter).

The wire sizes on a wire penetrameter consist of different groupings. In the DIN system, for example, a wire penetrameter may have one of three groupings of seven wire sizes:

• • •

One group contains wire numbers 1 through 7, which correspond to wire diameters of 3.20 through 0.80 mm (0.126 through 0.031 in.) The second group (Fig. 54c) contains wire numbers 6 through 12, which correspond to wire diameters of 1.00 through 0.25 mm (0.039 through 0.010 in.) The third group contains wire numbers 10 through 16, which correspond to wire diameters of 0.40 through 0.10 mm (0.016 through 0.004 in.)

Regardless of whether the wire type is designed to ISO, DIN, or ASTM specifications, image quality is denoted by the wire number of the thinnest wire distinguishable on the radiograph. In contrast to the plaque system, however, the wire system does not provide constant sensitivity, because the sensitivity varies with testpiece thickness. Therefore, the equivalent sensitivities of wire penetrameter indications are defined for a range of testpiece thickness (as indicated in Table 10 for the DIN system). Table 11 lists the wire sizes equivalent to a 2-2T sensitivity for a variety of testpiece thicknesses. Table 10 DIN specification for minimum image quality and equivalent-sensitivity range for each range of testpiece thickness Minimum image quality is expressed as wire number (BZ) of thinnest wire distinguishable in radiograph. Testpiece thickness

High-sensitivity level (category 1)

Normal-sensitivity level (category 2)

mm

in.

Wire No., BZ

Equivalent sensitivity, %

Wire No., BZ

Equivalent sensitivity, %

0-6

>0-0.25

16

1.7 min

14

2.7 min

6-8

0.25-0.30

15

2.0-1.6

13

3.3-2.5

8-10

0.30-0.40

14

2.0-1.6

12

3.1-2.5

10-16

0.40-0.60

13

2.0-1.3

11

3.2-2.0

16-25

0.60-1.00

12

1.6-1.0

10

2.5-1.6

25-32

1.00-1.25

11

1.3-1.0

9

2.0-1.6

32-40

1.25-1.60

10

1.3-1.0

8

2.0-1.6

40-50

1.60-2.00

9

1.3-1.0

7

2.0-1.6

50-80

2.00-3.15

8

1.3-0.8

6

2.0-1.3

80-200

3.15-8.00

7

1.0-0.4

...

...

80-150

3.15-6.00

...

...

5

1.6-0.8

150-170

6.00-6.70

...

...

4

1.1-0.9

170-180

6.70-7.00

...

...

3

1.2-1.1

180-190

7.00-7.50

...

...

2

1.4-1.3

190-200

7.50-8.00

...

...

1

1.7-1.6

Table 11 Wire sizes equivalent to 2-2T hole-type levels Minimum specimen thickness

Wire diameter

mm

in.

mm

in.

6.35

0.25

0.1(a)

0.004(a)

9.5

0.375

0.13(a)

0.005(a)

13

0.500

0.16

0.0063

16

0.625

0.2

0.008

19

0.750

0.25

0.010

22

0.875

0.33

0.013

25

1.00

0.4

0.016

32

1.25

0.5

0.020

38

1.5

0.64

0.025

44

1.75

0.81

0.032

50

2

1.0

0.040

65

2.5

1.3

0.050

75

3

1.6

0.063

90

3.5

2.0

0.080

100

4

2.5

0.100

125

5

3.2

0.126

150

6

4.0

0.160

Source: ASTM E 747 (a) Wire diameters for use with specimens less than 13 mm ( relationship as the hole type.

in.) in thickness do not represent the true 2-2T level. They follow the same

Step wedge penetrameters usually have either an arithmetic or a geometric progression of step thicknesses. A plain step wedge penetrameter is useful only for determining the ability of a radiograph to resolve variations in testpiece thickness; it cannot be used to evaluate the effect of imaging unsharpness, which is often the chief factor that determines image quality. However, if a plain step wedge is modified by drilling holes in each step, it becomes sensitive to imaging unsharpness. This type of design is used by the British Welding Research Association (BWRA) and the French Navy (AFNOR). In the BWRA design (Fig. 54d), the holes in a given step wedge are all of the same size--0.635 mm (0.025 in.) diameter for the step wedge ranging from 0.13 to 1.0 mm (0.005 to 0.040 in.) in step thickness and 1.3 mm (0.050 in.) diameter for the step wedge ranging from 1.0 to 2.0 mm (0.040 to 0.080 in.) in step thickness. AFNOR step wedges (Fig. 54e and f) have holes that are the same in diameter as the step thickness. The BWRA standard specifies a uniform increment between step thicknesses; the AFNOR standard specifies a constant ratio between successive thicknesses (similar to that of successive wire diameters in the DIN standard). The BWRA standard incorporates only two penetrameters, both of which are shown in Fig. 54(d). The AFNOR standard specifies penetrameters having four different series of step thicknesses, as follows:

Series

Step thicknesses, mm

1

0.125, 0.16, 0.20, 0.25, 0.32, 0.40

2

0.32, 0.40, 0.50, 0.63, 0.80, 1.00

3

0.80, 1.00, 1.25, 1.60, 2.00, 2.50

4

2.00, 2.50, 3.20, 4.00, 5.00, 6.30

The four series cover the entire range of testpiece thicknesses from about 4 to 300 mm (1.6 to 12 in.). Series 3 is illustrated as a hexagonal penetrameter in Fig. 54(e), and series 2 is illustrated as a linear penetrameter in Fig. 54(f). Both the BWRA and AFNOR penetrameters are sensitive to image definition and to contrast. Definition is judged on the visibility of holes in the specified step. Sometimes, the image quality defined by a step wedge penetrameter is somewhat ambiguous because the image can be evaluated both on the visibility of steps and on the visibility of holes. When contrast is good and definition is poor, more steps than holes can be seen on the radiograph. However, when the image quality is judged as intended--that is, on the visibility of individual holes for AFNOR penetrameters and on the visibility of the symbol (not necessarily the visibility of individual holes) for BWRA penetrameters--the step wedge penetrameters are quite sensitive to variations in radiographic technique. One minor limitation of AFNOR penetrameters is that the hole size in the thinnest steps are comparable to the size of graininess visible in the radiograph. Sometimes it is not easy to be certain that a hole is visible, but the use of two holes in the thinnest steps partly overcomes this limitation.

Penetrameters for electronic components use spherical particles and wire sizes, typical of the wire used in such

devices, for determinations of resolution or sensitivity. The wires are arranged in a three-dimensional grid with closetoleranced spacing for the purpose of providing a measure of distortion. The image of the grid is measured on the radiograph, and the distortion is calculated by:

(Eq 21)

where D is the percent distortion, Sm is the wire spacing as measured on the radiograph, and SA is the actual wire spacing. The ASTM standard E 801 describes a set of eight penetrameters having different cover thicknesses and wire sizes. The cover densities and wire sizes are typical of the case materials and internal connecting wires of electronic components. Two of these penetrameters are typically used for each exposure--usually the number having the density closest to that of the component being radiographed and the next-higher-number penetrameter. Because electronic components are typically exposed in groups, the penetrameters are placed at opposing corners of the group edges. This ensures that the worst-case parameter values will be indicated. Radiographic Inspection Revised by the ASM Committee on Radiographic Inspection*

Placement of Identification Markers and Penetrameters The location of identification markers with respect to the testpiece is important only to the extent that shadows cast by the identification markers should not obscure shadows cast by the testpiece itself. This is accomplished most easily by attaching the lead letters or numbers to the film holder in a region outside the area being inspected, usually along the edges of the holder. When it is important to ensure that identification markers do not obscure the image of some welldefined region of the testpiece, such as a weld, it may be desirable to attach the identification markers to the testpiece adjacent to that region. When several views of the same testpiece are to be shot, it is good practice to attach an identification (view) marker to the testpiece at each end of the area to be inspected in each view. These markers should be left in place until after the adjacent exposures have been shot. Each view marker should be visible in two adjacent radiographs; if it is not, incomplete coverage has been obtained. In some codes and specifications, this practice (known as sequence numbering) is required. It is desirable to mark the testpiece with chalk, crayon, or a metal stamp to indicate the exact location of identification markers. This can avoid possible difficulties either in identifying defective testpieces or in correlating radiographs with testpieces. The placement of penetrameters is important because incorrect placement with respect to the testpiece can result in an incorrect assessment of image quality. On simple shapes, especially flat plates and similar shapes of uniform thickness, it is seldom necessary to be concerned about factors other than placing the penetrameter where it will properly represent maximum unsharpness, will not obscure any region being inspected, and will be located in the outer cone of the radiation beam. When the shape of a testpiece is complex or when there is a large variation in the thickness of the testpiece, placement of penetrameters can be critical. Several suggested means of achieving proper placement of penetrameters are shown in Fig. 56 for welds between plates of different thickness and for circumferential welds in pipe. When no level testpiece surface is available for placement of the penetrameter, penetrameter blocks placed beside the testpiece are the only reasonable alternative. It is sometimes advantageous to use a stepped wedge as a penetrameter block, with the penetrameter on each step. For example, in the technique used for three-view inspection of the large cast stainless steel impeller discussed in Example 2 in this article, four penetrameters ranging in thickness from 0.25 to 1.15 mm (0.010 to 0.045 in.) were used for six of the exposures. The four penetrameters, each placed on a different block between 13 and 57 mm (

and 2

in.)

thick, were needed for assurance that the specified level of image quality was achieved over the entire range of impeller thicknesses.

Fig. 56 Correct placement of view markers, location markers, and penetrameters for radiographic inspection. Dimensions given in inches

Even though the following discussion and Fig. 56 illustrate the placement of markers and penetrameters on weldments, similar locations for markers and penetrameters can be used on testpieces that do not contain welds. In all arrangements, penetrameters should be placed in the outer cone of the radiation beam. Radiography of Plates. Figure 56(a) illustrates three alternative arrangements of penetrameters and identification markers for the radiography of a weld joining one plate to another plate of different thickness. In all three arrangements, the identification markers and penetrameters are placed parallel to the weld. View markers and penetrameters are usually

placed 3 to 20 mm ( to in.) from the edge of the weld zone, but no more than 40 mm (1 in.). Testpiece identification markers, however, can be placed farther away if necessary to ensure that their image is outside the image of the weld zone in the processed radiograph. Identification markers are usually placed on the film but view markers should be placed on the surface of the testpiece closest to the radiation source so that correct overlap between adjacent exposures can be verified. (If the view markers were located on the film side, a portion of the testpiece directly above the view markers could be missed even though the images of the markers appeared in adjacent radiographs.) In Fig. 56(a), the preferred setup (setup 1) has two penetrameters located on the thinner plate. In the alternative setups (setups 2 and 3), two penetrameters are located on the thicker plate (setup 2) or one penetrameter on each plate (setup 3). Shims made of an alloy that has the same absorption characteristics as the weld metal are used under the penetrameters in each instance to compensate for any difference between the thickness of the weld zone, including reinforcement, and the thickness of the plate on which the penetrameter is located. Any shim used should be larger than the penetrameter placed on it, so that the image of the penetrameter can be clearly seen within the umbral image of the shim. Also, the direction of radiation with respect to shim and penetrameter location should be considered, especially with thick shims or penetrameter blocks, to ensure that the shim properly represents the effective penetrated thickness of the testpiece. Some codes and specifications require that the image of a penetrameter be used to evaluate the quality of only that portion of the radiographic image of the testpiece that has similar photographic density. Strict limits can be placed on the allowable density difference between penetrameter image and testpiece image. For this reason, it may be necessary to use two or more penetrameters to evaluate image quality in different regions on the radiograph. When plaque-type penetrameters are used, plaques of different thickness are used for different regions, depending on testpiece thickness in each region. Radiography of Cylinders. Figures 56(b), 56(c), and 56(d) illustrate alternative locations for markers and

penetrameters for the double-wall, double-image radiography of hollow cylinders or welded pipe. These alternatives can be used for either normal or offset (corona) views. When the penetrameter is placed on the cylinder itself, as shown in Fig. 56(b), or on a short section of pipe having the same diameter and wall thickness as the pipe being inspected, as shown in Fig. 56(c), any shim that is used under the penetrameter should be only thick enough to compensate for weld reinforcement; that is, twice the nominal reinforcement for a normal view, but equal to the nominal reinforcement for an offset (corona) view. If the testpiece is a plain cylinder or if a circumferential butt weld is flush with the surface, no shim is needed. When a penetrameter block is used to provide equivalent penetrated thickness under the penetrameter, the block thickness should equal twice the nominal wall thickness of the cylinder plus twice the nominal weld reinforcement. Also, the penetrameter block should be set on a block of Styrofoam or similar nonabsorbing material so that the upper surface of the penetrameter block is aligned with the upper surface of the pipe, as shown in Fig. 56(d). When a short section of pipe is used under the penetrameter (Fig. 56c), the radiation source should be centered between the pipe being inspected and the short pipe section; otherwise, the radiation source should be centered above the pipe being inspected. To ensure that the penetrameter image is within the umbral region of the image of a shim or penetrameter block, the penetrameter should be aligned with the edge of the shim or block closest to the central beam of radiation. Figures 56(e), 56(f), and 56(g) illustrate alternative setups for the double-wall, single-image radiography of hollow cylinders or welded pipe. These alternatives are suitable for both normal and offset (corona) views. As for a double-wall, double-image technique, the penetrameter can be placed on the pipe itself--on a short section of pipe being inspected or on a penetrameter block. The setup illustrated in Fig. 56(e) can be used when there is access to the inside of the pipe for placement of the penetrameter. When a shim is used under the penetrameter (Fig. 56e and f), it should be equal to the height of nominal weld reinforcement, regardless of the view that is used. When there is no reinforcement, no shim is needed. If the penetrameter is placed on a penetrameter block, as in Fig. 56(g), the block should be equal to twice the

nominal wall thickness plus the nominal height of weld reinforcement, not plus twice the nominal reinforcement as with the double-wall, double-image technique. Radiation-source location with respect to the testpiece and the location of the penetrameter on the block or shim are the same as for a double-wall, double-image technique. In any setup for single-wall, single-image radiography, the penetrameters can be placed only on the testpiece because the film is always on one side of the wall and the source on the other side. Figures 56(h), 56(j), and 56(k) illustrate alternative arrangements for single-wall, single-image radiography. Shims, when used, need only compensate for any weld reinforcement. When the radiation source is external, as in Fig. 56(h), location markers should be placed on the outside surface for assurance that the correct overlap between adjacent exposures has been achieved. There should be a minimum of one penetrameter and one set of view and location markers per film, except that there should be three or more penetrameters and sets of markers (spaced equally around the circumference of the pipe) when a 360° simultaneous exposure is made on a single strip of film, as shown in Fig. 56(k). A minimum of three penetrameters is needed for assurance that the radiation source was actually located on the central axis of the cylinder and that equal intensity of radiation was incident on the entire circumference. When a 360° simultaneous exposure is made on overlapping pieces of film, not only should penetrameters be placed so that one appears on each piece of film, but also view markers and location markers should be placed so that they coincide with the regions of overlap between adjacent pieces of film. Radiography of Flanges. Although single-image techniques (especially the single-wall, single-image technique) are ordinarily used with a normal (vertical) viewing direction, there are applications in which an offset view is advantageous. Three setups for the single-image radiography of flanged pipe using offset views are illustrated in Fig. 56(m), 56(n), and 56(p). The principles of location-marker and penetrameter placement are similar to those previously discussed for normal views; the only difference is that extra precautions must be taken to ensure that the projected images of markers or penetrameters do not fall on the image of any region being inspected. Radiographic Inspection Revised by the ASM Committee on Radiographic Inspection*

Control of Scattered Radiation Although secondary radiation can never be completely eliminated, numerous means are available to reduce its effect. The various methods, which are discussed below in terms of x-rays, include: • • •

Use of lead screens Protection against back scatter and scatter from external objects Use of masks, diaphragms, collimators, and filtration

Most of the same principles for reducing the effect of secondary x-rays apply also to -ray radiography. However, differences in application arise because of the highly penetrating characteristics of gamma radiation. For example, a mask for use with 200-kV x-rays could be light enough in weight for convenient handling, yet a mask for use with cobalt-60 radiation would be much thicker, heavier, and more cumbersome. In any event, with either x-rays or -rays, the means for reducing the effects of secondary radiation must be selected with consideration of cost and convenience as well as the effectiveness. Lead screens placed in contact with the front and back emulsions of the film diminish the effect of scattered radiation

from all sources by absorbing the long-wavelength rays. They are the least expensive, most convenient, and most universally applicable means of combating the effects of secondary radiation. Lead screens lessen the amount of secondary radiation reaching the film or detector, regardless of whether the screens increase or decrease the intensity of detected radiation. (The intensifying effect of lead screens is discussed in the section "Image Conversion Media" in this article.) Sometimes, the use of lead screens requires increased exposure time (or image processing in the case of real-time monitoring). If high radiographic quality is desired, lead screens should not be abandoned merely because the photon energy is so low that they exhibit no intensifying action. However, at a sufficiently low photon energy, depending on the

testpiece, the absorption of transmitted image-forming radiation by the front screen will degrade image quality. Under these conditions, a front screen should not be used, but a back screen will reduce back-scattered radiation without affecting the image-forming radiation and should be used. In general, lead screens should be used whenever they improve the quality of the radiographic image. Protection Against Back Scatter. Severe back scatter can produce an image of the back of the cassette or film holder on the film, superimposed on the image of the testpiece. To prevent back scatter from reaching the film, it is customary to place a sheet of lead in back of the cassette or film holder. The thickness needed depends on radiation

quality; for example, 3 mm ( in.) of lead for 250-kV x-rays and 6 mm ( in.) of lead for 1-MeV x-rays or for Ir-192 or Co-60 -rays. At 100 kV and lower, the lead that is frequently incorporated into the back of the cassette or film holder usually provides sufficient protection from back scatter. Radiographic tables or stands can also be covered with lead to reduce back scatter. Because providing protection against back scatter can usually be done simply and conveniently, it is better to overprotect than to underprotect. For assurance of adequate protection from back-scattered radiation, a characteristic lead symbol (such as a 3 mm, or in., thick letter "B") can be attached to the back of the cassette or film holder and a radiograph made in the normal manner. If a low-density image of the symbol appears on the radiograph, it is an indication that protection against back-scattered radiation is insufficient, and additional precautions must be taken. In the event that the image of the symbol is darker than the surrounding image, the intensification effect of lead is the probable cause of the dark image of the symbol. This effect is very rarely observed, and then only when there is little or no filtration, such as in direct or fluorescent-screen exposures or when very thin lead screens are used. Masks and Diaphragms. Secondary radiation originating in sources outside the testpiece is most serious for testpieces

that have high absorption for x-rays (most metals) because secondary radiation from external sources may be large compared with the image-forming radiation that reaches the film through the testpiece. Often, the most satisfactory method of reducing this secondary radiation is by the use of cutout lead masks or some other form of lead-sheet mask mounted over or around the testpiece (Fig. 57).

Fig. 57 Use of lead-sheet masks on a testpiece for reducing secondary radiation

Copper or steel shot having a diameter of 0.25 mm (0.01 in.) or less is an effective and convenient mask. Metallic shot is also very effective for filling cavities in irregular-shape testpieces such as castings, where the normal exposure for thick areas would result in overexposure for thinner areas. Masking can also be accomplished by using barium clay, lead putty, or liquid absorbers such as a saturated solution of lead acetate plus lead nitrate. This solution is made by dissolving approximately 1.6 kg (3 lb) of lead acetate in 4 L (1 gal.) of hot water and adding approximately 1.4 kg (3 lb) of lead nitrate. Caution: Care should be exercised at all times when using liquid absorbers, because of their highly poisonous or lethal nature. When metallic shot or a liquid absorber is used as a mask, the testpiece is placed in a container made of aluminum or thin sheet steel, and the metallic shot or liquid absorber is poured in around the testpiece (Fig. 58). A form of masking called blocking, which consists of placing lead blocks at the edges of the testpiece or placing lead plugs in internal holes, also prevents side scatter from reaching the film.

Fig. 58 Setup for radiography using either metallic shot or a liquid absorber as a mask to control secondary radiation

Lead diaphragms limit the area covered by the x-ray beam. Diaphragms are particularly useful when the desired cross section of the beam is a circle, square, or rectangle. Figure 59 shows the combined use of metallic shot, a lead mask, and a lead diaphragm to control scattered radiation.

Fig. 59 Use of a combination of metallic shot, a lead mask, and a lead diaphragm to control scattered radiation

Collimators. Side scatter from walls, equipment, and other structures in the x-ray room can be greatly reduced by

improving the directionality of the x-ray beam. Directionality can be improved by the use of a collimator, which is often a thick lead diaphragm with a small hole through the middle. A collimator absorbs most of the diverging radiation that surrounds the central beam, thus eliminating most of the rays that could be scattered from nearby surfaces. Although considered good practice, removing all unnecessary equipment and other material from the x-ray room is sometimes impossible or impractical. In such cases, a collimator placed at the exit port of the radiation source can substantially reduce, if not eliminate, unwanted side scatter. Filtration. In addition to the filtering effect of lead screens, secondary x-rays can be filtered by using thin copper or lead sheets between the testpiece and the cassette or film holder. Filtration is never used in gamma radiography, because of the essentially monochromatic nature of the beam.

When the testpiece has very thin sections adjacent to thick sections or when the direct beam can strike the detector after passing around the testpiece, undercutting may be encountered. If undercutting occurs, additional filtration (that is, more than can be achieved with conventional lead screens) is necessary. Additional filtration is accomplished by placing a filter at or near the x-ray tube, as shown in Fig. 60. This may adequately eliminate overexposure in thin regions of the testpiece and also along the perimeter of the testpiece. Such a filter is particularly useful for reducing undercutting when a lead mask around the testpiece is impractical or when the testpiece may be damaged by masking with liquid absorbers or metallic shot. Filtration of the incident radiation beam reduces undercut by selectively attenuating the long-wavelength portion of the x-ray spectrum. Long wavelengths do not contribute significantly to the detection of flaws but only produce secondary radiation that reduces radiographic contrast and definition.

Fig. 60 Use of a lead diaphragm to limit the included angle of the x-ray beam, and use of a filter to reduce subject contrast and to eliminate much of the secondary radiation that causes undercutting

The choice of a filter material should be made on the basis of availability and ease of handling. For the same filtering effect, the thickness of filter required is less for those materials that have lower absorption coefficients. Often, copper or brass is the most useful because filters of these materials will be lightweight enough to handle but not so thin that they are easily bent or broken. Definite rules for filter thicknesses are difficult to formulate because the amount of filtration required depends not only on the materials and thickness range of the testpiece but also on the homogeneity of the testpiece and on the amount of

undercutting that is to be eliminated. In the radiography of aluminum, a filter of copper about 4% as thick as the thickest area of the testpiece is usually satisfactory. With a steel testpiece, a copper filter ordinarily should be about 20%, or a lead filter about 3%, as thick as the thickest area of the testpiece for optimum filtration. These values are maximum values, and depending on circumstances, useful radiographs can often be made with far less filtration. Radiation Diffraction. A special form of scattering due to x-ray diffraction is occasionally encountered in radiographic inspection. Diffraction of radiation is observed most often in the radiography of thin testpieces having a grain size large enough to be an appreciable fraction of the part thickness. Castings made of austenitic corrosion-resistant and heat-resistant stainless steel or of Inconel and other nickel-base alloys are the products most likely to exhibit diffraction in radiographs.

The radiographic appearance of this type of scattering can be confused with the mottled appearance sometimes produced by porosity, segregation, or spongy shrinkage. Diffraction patterns can be distinguished from these conditions in the testpiece by making successive radiographs with the testpiece rotated between exposures 1 to 5° about an axis perpendicular to the beam. A mottled pattern due to porosity or segregation will be only slightly changed, but a pattern due to diffraction effects will show a marked change. The radiographs of some testpieces will show mottling from both diffraction and porosity, and careful interpretation of the radiographs is needed to differentiate between them. Mottling due to diffraction can be reduced, and sometimes eliminated, by raising x-ray tube voltage and by using lead screens. Filters will usually aid in the control of diffraction. Raising the tube voltage and filtration are often of positive value even though radiographic contrast and sensitivity are reduced. Sometimes, diffraction cannot be reduced. In such cases, two radiographs made as described above can be used to identify diffraction. Scattering at High Photon Energies. Lead screens should always be used when the radiation energy exceeds 1

MeV. Use of the usual 0.13 mm (0.005 in.) thick front screen and 0.25 mm (0.010 in.) thick back screen is both satisfactory and convenient. Some users find 0.13 mm (0.005 in.) thick front and back screens adequate when filters are used both front and back of the cassette or film holder. Other users consider 0.25 mm (0.010 in.) thick front and back screens of value because of greater selective absorption of scattered radiation from the testpiece. Filtration of the incident x-ray beam offers no improvement in radiographic quality. However, filters at the film improve radiographs for the inspection of uniform sections. Lead filters are most convenient for energies above 1 MeV. Care should be taken to minimize mechanical damage to the filter because filter defects could be confused with characteristics of the material being inspected. It is important to block off all radiation except the effective beam with heavy shielding at the anode. This is usually recognized by manufacturers of high-voltage x-ray equipment. For example, in some linear accelerators, depleteduranium collimators confine the beams to a 22° included angle. Unless a high-energy x-ray beam is well collimated, radiation striking the walls of the x-ray room will generate secondary radiation and thus seriously degrade the quality of the radiograph. This will be especially noticeable if the testpiece is thick or has projecting parts that are not immediately adjacent to the filter. Radiographic Inspection Revised by the ASM Committee on Radiographic Inspection*

Interpretation of Radiographs Proper identification of both the radiograph and the testpiece, clarity of the penetrameter, suitability of radiographic techniques, adequacy of coverage, and the techniques of image processing in real-time systems or the precision and uniformity of film processing in film radiography all offset the image that is being interpreted. Film radiographs and realtime radiographs are interpreted similarly with respect to the recognition of flaws and testpiece features. The primary difference is that film radiographs are negative images, while real-time radiographs are generally positive images (which may also be enhanced with image processing).

A qualified interpreter must: • • • •

Define the quality of the radiographic image, which includes a critical analysis of the radiographic procedure and the image-developing procedure Analyze the image to determine the nature and extent of any abnormal condition in the testpiece Evaluate the testpiece by comparing interpreted information with standards or specifications Report inspection results accurately, clearly, and within proper administration channels

Proper identification of both the radiograph and testpiece is an absolute necessity for correlation of the radiograph with the corresponding testpiece. Identification includes both identification of the testpiece and identification of the view or area of coverage. Poor-quality film radiographs are usually reshot. However, reshooting radiographs increases inspection costs, not only because the original setup must be duplicated and a new exposure made but also because the testpiece must be retrieved and taken to the radiographic laboratory. With on-site radiography, which involves transporting radiographic equipment to the site and returning the exposed films to the laboratory for processing, especially high costs may be involved when poor-quality radiographs must be reshot. Table 12 lists some of the usual causes of poor quality in a radiographic image and indicates the usual corrective action required to eliminate each cause. Table 12 Probable causes and corrective action for various types of deficient image quality or artifacts on processed radiographic film Quality or artifact

Probable cause

Corrective action

Density too high

Overexposure

View with higher-intensity light. Check exposure (time and radiation intensity); if as specified, reduce exposure 30% or more.

Overdevelopment

Reduce development time or developer temperature.

Fog

See "Fog" below.

Underexposure

Check exposure (time and radiation intensity); if as specified, increase exposure 40% or more.

Underdevelopment

Increase development time or developer temperature. Replace weak (depleted) developer.

Material between screen and film

Remove material.

High subject contrast

Increase tube voltage.

High film contrast

Use a film with lower contrast characteristics.

Low subject contrast

Reduce tube voltage.

Low film contrast

Use a film with higher contrast characteristics.

Density too low

Contrast too high

Contrast too low

Poor definition

Underdevelopment

Increase development time or developer temperature. Replace weak (depleted) developer.

Testpiece-to-film distance too long

If possible, decrease testpiece-to-film distance; if not, increase sourceto-film distance.

Source-to-film distance too short

Increase source-to-film distance.

Focal spot (or large

Use smaller source or increase source-to-film distance.

-ray source) too

Screens and film not in close contact

Ensure intimate contact between screens and film.

Film graininess too coarse

Use finer-grain film.

Light leaks in darkroom

With darkroom unlighted, turn on all lights in adjoining rooms; seal any light leaks.

Exposure to safelight

Reduce safelight wattage. Use proper safelight filters.

Stored film inadequately protected from radiation

Attach strip of lead to loaded film holder and place in film-storage area. Develop test film after 2 to 3 weeks; if image of strip is evident, improve radiation shielding in storage area.

Film exposed to heat, humidity, or gases

Store film in a cool, dry place not subject to gases or vapors.

Overdevelopment

Reduce development time or developer temperature.

Developer contaminated

Replace developer.

Exposure during processing

Do not inspect film during processing until fixing is completed.

Finely mottled fog

Stale film

Use fresh film.

Fog on edge or corner

Defective cassette

Discard cassette.

Yellow stain

Depleted developer

Replace developer solution.

Failure to use stop bath or to rinse

Use stop bath, or rinse thoroughly between developing and fixing.

Depleted fixer

Replace fixer solution.

Fog

Dark circular marks

Film splashed with developer prior to immersion

Immerse film in developer with care.

Dark spots or marblelike areas

Insufficient fixing

Use fresh fixer solution and proper fixing time.

Dark branched lines and spots

Static discharge

Unwrap film carefully. Do not rub films together. Avoid clothing productive of static electricity.

Dark fingerprints

Touching undeveloped film with chemically contaminated fingers

Wash hands thoroughly and dry, or use clean, dry rubber gloves.

Light fingerprints

Touching undeveloped film with oily or greasy fingers

Wash hands thoroughly and dry, or use clean, dry rubber gloves.

Dark spots or streaks

Developer contaminated with metallic salts

Replace developer solution.

Crescent-shaped light areas

Faulty film handling

Keep film flat during handling. Use only clean, dry film hangers.

Light circular patches

Air bubbles on film during development

Agitate immediately upon immersion of film in developer.

Circular or dropshaped light patches

Water or fixer splashed on film before development

Avoid splashing film with water or fixer solution.

Light spots or areas

Dust or lint between screens and film

Keep screens clean.

Sharply outlined light or dark areas

Nonuniform development

Agitate film during development.

Reticulation (leathergrain appearance)

Temperature gradients in processing solutions

Maintain all solutions at uniform, constant temperature.

Frilling (loosening of emulsion from film base)

Fixer solution too warm

Maintain correct temperature of the fixing solution.

Fixer solution depleted

Replace fixer solution.

Personnel engaged in the interpretation of radiographs should possess certain qualifications. Some qualification recommendations are included in personnel standards published by the American Society for Nondestructive Testing (Ref 6), several governmental agencies, and many private manufacturers. Usually, a minimum level of visual acuity, minimum standards of education and training, and demonstrated proficiency are required of all interpreters of radiographs. Viewing of radiographs should be carried out in an area with subdued lighting to minimize distracting reflections

from the viewing surface. Audible distractions, which interfere with concentration, can best be avoided by locating the work area away from the main production floor or other high-noise area.

Radiographic film images are viewed on an illuminated screen. The viewing apparatus should have an opal-glass or plastic screen large enough to accommodate the largest film to be interpreted. The screen should be illuminated from behind with light of sufficient intensity to reveal variations in photographic density up to a nominal film density of at least 3.0. There may be a need for a smaller, more intensely illuminated viewer for evaluating small areas of film having densities up to 4.5 or more. Viewing screens of high-intensity illuminators should be cooled by blowers or other suitable apparatus to prevent excessive heat from damaging films and to extend lamp life. When interpreting paper radiographs or xeroradiographs, specular light as from a spotlight or high-intensity lamp should be directed onto the radiograph from the side at an angle of about 30°. Background lighting should be heavily subdued. A densitometer can be provided for accurate evaluation of small variations in photographic density or for quantitative evaluation of radiographic and processing techniques. A transmission densitometer is used with films, and a reflection densitometer is used with paper radiographs. Radiographic Acceptance Standards. Usually, a series of radiographs that exhibit various types and sizes of flaws should be selected for acceptance standards. Parts that contain similar flaws should be performance tested to determine the least acceptable condition. The radiograph of the least acceptable part then becomes the minimum acceptance standard for similar parts. Often, the acceptance standard is defined as a length or area of the image that may contain no more than a specified number of flaws of a given size and type. Certain types of flaws, such as cracks or incomplete fusion, may be prohibited regardless of size. The interpreter must determine the degree of imperfection, as related to the minimum acceptance standard, and then decide whether minimum soundness requirements have been met.

Obviously, no single standard can be applied universally to radiographic inspection. However, flaws that are frequently encountered have been reproduced in sets of reference radiographs such as those published by ASTM (Table 13). Reference radiographs depict various types of flaws that may occur in castings or weldments and are graded according to flaw size and severity. Table 13 Reference radiographs in ASTM standards ASTM standard

Subject of radiographs in standard

E 155

Aluminum and magnesium castings

E 186 Heavy-wall (50-115 mm, or 2-4

in.) steel castings

E 192

Investment steel castings for aerospace applications

E 242

Appearance of radiographic images as certain parameters are changed

E 272

High-strength copper-base and nickel-copper alloy castings

E 280 Heavy-wall (115-300 mm, or 4

-12 in.) steel castings

E 310

Tin bronze castings

E 390

Steel fusion welds

E 431

Semiconductors and related devices

E 446

Steel castings up to 50 mm (2 in.) in thickness

E 505

Aluminum and magnesium die castings

E 689

Ductile cast irons

E 802 Gray iron castings up to 115 mm (4

in.) in thickness

Codes or specifications for radiographic inspection, particularly those that have been standardized by an industry through a trade association or a professional society or those that have been adopted by a governmental agency or a prime contractor, may refer to published reference radiographs. In such cases, the code or specification should designate one or more reference radiographs in a specific set as the minimum standard for acceptance. Although it is considered most desirable to have an acceptance standard that is based on actual service data, standardized codes or specifications usually define rigid acceptance criteria that do not allow for variations in specific design features of similar products.

Reference cited in this section

6. "Nondestructive Testing Personnel Qualification and Certification, Supplement A, Radiographic Testing Method," ASNT-TC-1A, American Society for Nondestructive Testing Radiographic Inspection Revised by the ASM Committee on Radiographic Inspection*

Interpretation of Radiographs Proper identification of both the radiograph and the testpiece, clarity of the penetrameter, suitability of radiographic techniques, adequacy of coverage, and the techniques of image processing in real-time systems or the precision and uniformity of film processing in film radiography all offset the image that is being interpreted. Film radiographs and realtime radiographs are interpreted similarly with respect to the recognition of flaws and testpiece features. The primary difference is that film radiographs are negative images, while real-time radiographs are generally positive images (which may also be enhanced with image processing). A qualified interpreter must: • • • •

Define the quality of the radiographic image, which includes a critical analysis of the radiographic procedure and the image-developing procedure Analyze the image to determine the nature and extent of any abnormal condition in the testpiece Evaluate the testpiece by comparing interpreted information with standards or specifications Report inspection results accurately, clearly, and within proper administration channels

Proper identification of both the radiograph and testpiece is an absolute necessity for correlation of the radiograph with the corresponding testpiece. Identification includes both identification of the testpiece and identification of the view or area of coverage. Poor-quality film radiographs are usually reshot. However, reshooting radiographs increases inspection costs, not only because the original setup must be duplicated and a new exposure made but also because the testpiece must be retrieved

and taken to the radiographic laboratory. With on-site radiography, which involves transporting radiographic equipment to the site and returning the exposed films to the laboratory for processing, especially high costs may be involved when poor-quality radiographs must be reshot. Table 12 lists some of the usual causes of poor quality in a radiographic image and indicates the usual corrective action required to eliminate each cause. Table 12 Probable causes and corrective action for various types of deficient image quality or artifacts on processed radiographic film Quality or artifact

Probable cause

Corrective action

Density too high

Overexposure

View with higher-intensity light. Check exposure (time and radiation intensity); if as specified, reduce exposure 30% or more.

Overdevelopment

Reduce development time or developer temperature.

Fog

See "Fog" below.

Underexposure

Check exposure (time and radiation intensity); if as specified, increase exposure 40% or more.

Underdevelopment

Increase development time or developer temperature. Replace weak (depleted) developer.

Material between screen and film

Remove material.

High subject contrast

Increase tube voltage.

High film contrast

Use a film with lower contrast characteristics.

Low subject contrast

Reduce tube voltage.

Low film contrast

Use a film with higher contrast characteristics.

Underdevelopment

Increase development time or developer temperature. Replace weak (depleted) developer.

Testpiece-to-film distance too long

If possible, decrease testpiece-to-film distance; if not, increase sourceto-film distance.

Source-to-film distance too short

Increase source-to-film distance.

Focal spot (or large

Use smaller source or increase source-to-film distance.

Density too low

Contrast too high

Contrast too low

Poor definition

-ray source) too

Screens and film not in close contact

Ensure intimate contact between screens and film.

Film graininess too coarse

Use finer-grain film.

Light leaks in darkroom

With darkroom unlighted, turn on all lights in adjoining rooms; seal any light leaks.

Exposure to safelight

Reduce safelight wattage. Use proper safelight filters.

Stored film inadequately protected from radiation

Attach strip of lead to loaded film holder and place in film-storage area. Develop test film after 2 to 3 weeks; if image of strip is evident, improve radiation shielding in storage area.

Film exposed to heat, humidity, or gases

Store film in a cool, dry place not subject to gases or vapors.

Overdevelopment

Reduce development time or developer temperature.

Developer contaminated

Replace developer.

Exposure during processing

Do not inspect film during processing until fixing is completed.

Finely mottled fog

Stale film

Use fresh film.

Fog on edge or corner

Defective cassette

Discard cassette.

Yellow stain

Depleted developer

Replace developer solution.

Failure to use stop bath or to rinse

Use stop bath, or rinse thoroughly between developing and fixing.

Depleted fixer

Replace fixer solution.

Dark circular marks

Film splashed with developer prior to immersion

Immerse film in developer with care.

Dark spots or marblelike areas

Insufficient fixing

Use fresh fixer solution and proper fixing time.

Dark branched lines and spots

Static discharge

Unwrap film carefully. Do not rub films together. Avoid clothing productive of static electricity.

Dark fingerprints

Touching undeveloped film with chemically contaminated fingers

Wash hands thoroughly and dry, or use clean, dry rubber gloves.

Light fingerprints

Touching undeveloped film with oily or greasy fingers

Wash hands thoroughly and dry, or use clean, dry rubber gloves.

Fog

Dark spots or streaks

Developer contaminated with metallic salts

Replace developer solution.

Crescent-shaped light areas

Faulty film handling

Keep film flat during handling. Use only clean, dry film hangers.

Light circular patches

Air bubbles on film during development

Agitate immediately upon immersion of film in developer.

Circular or dropshaped light patches

Water or fixer splashed on film before development

Avoid splashing film with water or fixer solution.

Light spots or areas

Dust or lint between screens and film

Keep screens clean.

Sharply outlined light or dark areas

Nonuniform development

Agitate film during development.

Reticulation (leathergrain appearance)

Temperature gradients in processing solutions

Maintain all solutions at uniform, constant temperature.

Frilling (loosening of emulsion from film base)

Fixer solution too warm

Maintain correct temperature of the fixing solution.

Fixer solution depleted

Replace fixer solution.

Personnel engaged in the interpretation of radiographs should possess certain qualifications. Some qualification recommendations are included in personnel standards published by the American Society for Nondestructive Testing (Ref 6), several governmental agencies, and many private manufacturers. Usually, a minimum level of visual acuity, minimum standards of education and training, and demonstrated proficiency are required of all interpreters of radiographs. Viewing of radiographs should be carried out in an area with subdued lighting to minimize distracting reflections

from the viewing surface. Audible distractions, which interfere with concentration, can best be avoided by locating the work area away from the main production floor or other high-noise area. Radiographic film images are viewed on an illuminated screen. The viewing apparatus should have an opal-glass or plastic screen large enough to accommodate the largest film to be interpreted. The screen should be illuminated from behind with light of sufficient intensity to reveal variations in photographic density up to a nominal film density of at least 3.0. There may be a need for a smaller, more intensely illuminated viewer for evaluating small areas of film having densities up to 4.5 or more. Viewing screens of high-intensity illuminators should be cooled by blowers or other suitable apparatus to prevent excessive heat from damaging films and to extend lamp life. When interpreting paper radiographs or xeroradiographs, specular light as from a spotlight or high-intensity lamp should be directed onto the radiograph from the side at an angle of about 30°. Background lighting should be heavily subdued. A densitometer can be provided for accurate evaluation of small variations in photographic density or for quantitative evaluation of radiographic and processing techniques. A transmission densitometer is used with films, and a reflection densitometer is used with paper radiographs. Radiographic Acceptance Standards. Usually, a series of radiographs that exhibit various types and sizes of flaws

should be selected for acceptance standards. Parts that contain similar flaws should be performance tested to determine the least acceptable condition. The radiograph of the least acceptable part then becomes the minimum acceptance standard

for similar parts. Often, the acceptance standard is defined as a length or area of the image that may contain no more than a specified number of flaws of a given size and type. Certain types of flaws, such as cracks or incomplete fusion, may be prohibited regardless of size. The interpreter must determine the degree of imperfection, as related to the minimum acceptance standard, and then decide whether minimum soundness requirements have been met. Obviously, no single standard can be applied universally to radiographic inspection. However, flaws that are frequently encountered have been reproduced in sets of reference radiographs such as those published by ASTM (Table 13). Reference radiographs depict various types of flaws that may occur in castings or weldments and are graded according to flaw size and severity. Table 13 Reference radiographs in ASTM standards ASTM standard

Subject of radiographs in standard

E 155

Aluminum and magnesium castings

E 186 Heavy-wall (50-115 mm, or 2-4

in.) steel castings

E 192

Investment steel castings for aerospace applications

E 242

Appearance of radiographic images as certain parameters are changed

E 272

High-strength copper-base and nickel-copper alloy castings

E 280 Heavy-wall (115-300 mm, or 4

-12 in.) steel castings

E 310

Tin bronze castings

E 390

Steel fusion welds

E 431

Semiconductors and related devices

E 446

Steel castings up to 50 mm (2 in.) in thickness

E 505

Aluminum and magnesium die castings

E 689

Ductile cast irons

E 802 Gray iron castings up to 115 mm (4

in.) in thickness

Codes or specifications for radiographic inspection, particularly those that have been standardized by an industry through a trade association or a professional society or those that have been adopted by a governmental agency or a prime contractor, may refer to published reference radiographs. In such cases, the code or specification should designate one or more reference radiographs in a specific set as the minimum standard for acceptance. Although it is considered most

desirable to have an acceptance standard that is based on actual service data, standardized codes or specifications usually define rigid acceptance criteria that do not allow for variations in specific design features of similar products.

Reference cited in this section

6. "Nondestructive Testing Personnel Qualification and Certification, Supplement A, Radiographic Testing Method," ASNT-TC-1A, American Society for Nondestructive Testing Radiographic Inspection Revised by the ASM Committee on Radiographic Inspection*

Radiographic Appearance of Specific Types of Flaws The radiographic appearance of many of the more usual types of flaws found in castings and weldments is described in this section. The descriptions apply specifically to images on film radiographs, although paper radiographs and xeroradiographs will exhibit similar images. Real-time images of the same types of flaws will be reversed in tone (dark tones in a radiograph will be light in a fluoroscopic image and vice versa) but otherwise will be similar to the images described here. Some of these types of flaws are not unique to castings and weldments. For example, cracks can be found in any product form. Surface inspection methods such as liquid penetrant or magnetic particle inspection are more appropriate than radiography for detecting most surface cracks, yet some forms of metal separation (forging bursts, for example) are entirely internal and cannot be found by surface methods. Flaws in Castings It is possible in most cases to identify radiographic images of the usual types of flaws in castings. The main types of foundry flaws that can be identified radiographically are described in the paragraphs that follow. Specific examples can be found in reference radiographs, such as those published by ASTM (Table 13). Microshrinkage appears as dark feathery streaks or dark irregular patches, corresponding to grain-boundary shrinkage. This condition is most often found in magnesium alloy castings. Shrinkage porosity (spongy shrinkage) appears as a localized honeycomb or mottled pattern (Fig. 61). Spongy shrinkage may be the result of improper pouring temperature or alloy composition.

Fig. 61 Radiographic appearance of gross shrinkage porosity (arrow) in an aluminum alloy 319 manifold casting. Radiograph was made at 85 kV with 1-min exposure.

Gas porosity appears as round or elongated smooth, dark spots. It occurs individually or in clusters or may be

distributed randomly throughout the casting. This condition is caused by gas released during solidification or by the evaporation of moisture of volatile material from the mold surface. Dispersed Discontinuities. Although the flaws usually encountered in light-alloy castings are similar to those in

ferrous castings, a group of irregularities called dispersed discontinuities may be present in the former. These dispersed discontinuities, prevalent in aluminum and magnesium alloy castings, consist of tiny voids scattered throughout part or all of the casting. Gas porosity and shrinkage porosity in aluminum alloys are examples of dispersed discontinuities. On radiographs of sections more than 13 mm ( in.) thick, it is difficult to distinguish images corresponding to the individual voids. Instead, dispersed discontinuities may appear on film deceptively as mottling, dark streaks, or irregular patches that are only slightly darker than the surrounding regions. Tears appear as ragged dark lines of variable width having no definite line of continuity. Tears may exist in groups,

starting at a surface, or they may be internal. Tears usually result from normal contraction of the casting during or immediately after solidification. Cold cracks generally appear as single, straight, sharp dark lines and are usually continuous throughout their lengths.

Cold cracks are produced by internal stresses caused by thermal gradients and may occur upon cooling from elevated temperatures during flame cutting, grinding, or quenching operations. Cold shuts appear as distinct dark lines of variable length and smooth outline. Cold shuts are formed when two bodies

of molten metal flowing from different directions contact each other but fail to unite. Cold shuts may be produced by interrupted pouring, slow pouring, or pouring the metal at too low a temperature. Misruns appear as prominent dark areas of variable dimensions with a definite smooth outline. Misruns are produced by

failure of the molten metal to completely fill a section of casting mold, leaving the region void. Inclusions of foreign material in the molten metal may be poured into the mold. They appear as small lighter or darker

areas in a radiograph, depending on the absorption properties of the included material as compared to those of the alloy. Sand inclusions appears as gray or light spots of uneven granular texture and have indistinct outlines. Inclusions lighter than the parent metal appear as isolated irregular or elongated variations of film blackening. Occasionally, an inclusion will have absorption characteristics equivalent to those of the matrix and will go undetected, although normally an inclusion that exhibits a radiographic contrast of about 1.4 to 2.3% can be seen. A contrast of 1.4 to 2.3% corresponds to about 0.005 to 0.01 density difference between adjacent areas on the film. Dross inclusions in the outer flange of a casting are shown in Fig. 62.

Fig. 62 Radiographic appearance of dross inclusions (arrows) in the outer flange of a cast aluminum alloy 355

housing body. Radiograph made at 75 kV, 1-min exposure.

Unfused chaplets usually appear in outline conforming to the shape of the chaplet. The outline is caused by a lack of bond between the chaplet and the cast metal.

Core shift can be detected when the view makes it impossible to measure deviation from a specified wall thickness (Fig.

63). Core shift may be caused by jarring the mold, insecure anchorage of cores, or omission of chaplets.

Fig. 63 Uneven wall thickness in an internal passage of a casting caused by core shift (top right). This radiograph, of an aluminum alloy casting about 3 to 6 mm ( exposure time of 1 min.

to

in.) thick, was made at 65 kV with an

Centerline shrinkage is localized along the central plane of a wall section, irrespective of the position occupied by the section in the mold. Such shrinkage is composed of a network of numerous filamentary veinlets leading in the direction of the nearest riser and can sometimes be mistaken for tears. Shrinkage cavities occur when insufficient feeding of a section results in a continuous cavity within the section.

Shrinkage cavities appear on the radiograph as dark areas that are indistinctly outlined and have irregular dimensions. Segregation is the separation of constituents in an alloy into regions of different chemical compositions. This

condition, seen mostly in aluminum alloys, appears as lighter areas on the film that produce a somewhat mottled appearance. Surface irregularities may produce an image corresponding to any deviation from normal surface profile. It is

possible to confuse these with internal flaws unless the casting is visually inspected at the time of interpretation. Flaws in Weldments In welding, heat must be carefully controlled to produce fusion and adequate penetration. Too much heat can cause porosity, cracks, and undercutting; too little heat can cause inadequate joint penetration and incomplete fusion. Most weld flaws consist of abrupt changes in homogeneity and can be readily detected by radiographic inspection. Stresses that are induced in the metal by welding but that are not accompanied by a physical separation of material will not be detected by radiography. Also, cracks not aligned with the x-ray beam may be missed. Conditions that can be detected in welds by radiography are described below. Several reference radiographs of flaws in weldments are provided in the article "Weldments, Brazed Assemblies, and Soldered Joints" in this Volume.

Undercutting appears as a dark line of varying width along the edge of the fusion zone. A fine dark line in this darker

area could indicate a crack and should be further investigated. Incomplete fusion appears as an elongated dark line. It sometimes appears very similar to a crack or an inclusion and

could even be interpreted as such. Incomplete fusion occurs between weld and base metal and between successive beads in multiple-pass welds. Incomplete penetration along one or both sides of the weld zone has an appearance similar to that of incomplete fusion. Cracks are frequently missed if they are very small (such as check cracks in the heat-affected zone) or are not aligned

with the radiation beam. When present in a radiograph, cracks often appear as fine dark lines of considerable length but without great width. Even some fine crater cracks are readily detected. In weldments, cracks may be transverse or longitudinal and may be either in the fusion zone or in the heat-affected zone of the base metal. Figure 64 illustrates a large crack in a steel weldment.

Fig. 64 Radiograph showing a large crack in a multiple-pass butt weld in 57 mm (2 in.) thick steel plate. The crack mainly follows the edge of the weld, but both ends turn in toward the center. The weld joined a 57 mm (2 in.) thick plate to a 70 mm (2 in.) thick plate that had been tapered to 57 mm (2 the weld groove. Radiograph was made with 1-MeV x-rays on Industrex AA film.

in.) at the edge of

Porosity (gas holes) consists commonly of spherical voids that are readily recognizable as dark spots, the radiographic

contrast varying directly with diameter. These voids may be randomly dispersed, in clusters, or may even be aligned along the centerline of the fusion zone. Occasionally, the porosity may take the form of tubes (worm holes) aligned along the direction of the weld solidification front. These appear as lines with a width of several millimeters. Slag inclusions are usually irregularly shaped and appear to have some width. Inclusions are most frequently found at

the edge of the weld, as illustrated in Fig. 65. In location, elongated slag deposits are often found between the first root pass and subsequent passes or along the weld-joint interface, while spherical slag inclusions can be distributed anywhere in the weld. The density of a slag inclusion is nearly uniform throughout and has less contrast than porosity, because an inclusion can be considered as a pore with absorbing material. Tungsten inclusions or a very high density (bariumcontaining) slag will appear as white spots.

Fig. 65 Radiograph showing a crack (dark line at top) and entrapped slag inclusions (dark spots at arrows) on opposite sides of a multiple-pass butt weld joining two 180 mm (7 in.) thick steel plates. Radiograph was made with 1-MeV x-rays on Industrex AA film.

Incomplete root penetration appears as a dark straight line through the center of the weld. The width of the

indication is determined by the root gap and amount of weld penetration. Flaws in Semiconductors Voids in semiconductors occur in several different sites, depending on the type of construction. In hermetic integrated circuits and low-power transistors, voiding typically occurs at the die (semiconductor element) and header (case) interface. The voids in this area will have the same approximate density as the areas of the header that are undisturbed by the die mount. In hermetic power transistors, voids may occur at the die mount substrate to header interface and at the die to die mount substrate interface. It is not possible to differentiate between interfaces after the device is sealed. Therefore, two radiographs are required--one after substrate attachment and a second after die attachment. Orientation is critical because the radiographs must be compared to determine the total area of voiding. In plastic-encapsulated semiconductors, voids in the encapsulating material are discernible by density differences, just as a void in other materials or in a weld. Extraneous material is frequently missed because of its small size and very thin cross section. Conductive material as small as 0.025 mm (0.001 in.) in its major dimension may cause failure of a semiconductor device. In some cases, multiple views may be required to determine whether or not expulsed die mount material is attached to the device case. Because of the very small geometries used, multiple views may also be required to determine if there is adequate clearance between internal connecting wires. Electronic devices can also be inspected after they are placed on the circuit board, but the board and its other components will reduce contrast and interfere with the resolution of fine features. The advantage of inspecting after the devices are placed on the board is that connections from the device to the board can be inspected also. Because the solders used to form these joints contain elements of high atomic number, they form high contrast images. Radiographic inspection can detect flaws such as solder balls (undesired solder that has been expelled from the joint), bridging, misregistered devices, and joints without solder. Laminographic systems are available that can emphasize various planes within the circuit board (see the article "Industrial Computed Tomography" in this Volume). Radiographic Inspection Revised by the ASM Committee on Radiographic Inspection*

Appendix: Processing of Radiographic Film In the processing of radiographic film, an invisible latent image produced on the film by exposure to x-rays, -rays, or light is made visible and permanent. Film processing is an exacting and important part of the radiographic procedure. Poor processing can be just as detrimental to the quality of a radiograph as poor exposure practice. Two methods of processing can be employed: manual processing, which is carried out by hand in trays or deep tanks, and automatic processing, which is accomplished in automated equipment. Guidelines for the control and maintenance of manual and automatic radiographic film processing equipment are specified in ASTM E 999. Manual Film Processing The manual processing of radiographic film is carried out in a processing room (darkroom) under subdued light of a particular color to which the film is relatively insensitive. The film is first immersed in a developer solution that causes areas of the film that have been exposed to radiation to become dark; the amount of darkening for a given degree of development depends on the degree of exposure. After development, the film is rinsed, preferably in an acid stop bath that arrests development. Next, the film is placed in a fixing solution that dissolves the undarkened portion of the film and hardens the emulsion. The film is subsequently washed to remove the fixing chemicals and soluble salts, then dried. Although trays and other containers for photographic processing have been used, the usual method of processing industrial radiographic film by hand is the rack-and-tank method. In this method, the processing solutions and wash water are contained in tanks (Fig. 66) deep enough for the film to be hung vertically on developing hangers, or racks. The advantages to this method are: • • • • •

Processing solutions have free access to both sides of the film Both emulsion surfaces are uniformly processed to the same degree The all-important factor of temperature can be controlled by regulating the temperature of the water bath in which the tanks are immersed The equipment does not require much space There is a savings in time compared to tray processing

Fig. 66 Typical unit for the manual development of radiographic film by a rack-and-tank method. (a) Processing tanks containing developer, stop bath, and fixer. (b) Cascade (countercurrent) washing tank

Handling of Film. The processing room and all equipment and accessories must always be kept scrupulously clean and

used only for the purpose of handling and processing film. Spilled solutions should be wiped up immediately to avoid extraneous spots on the radiographs. Floating thermometers, film hangers, and stirring rods should be thoroughly rinsed in clean water after each use to avoid contamination of chemicals or streaking of film.

Film and radiographs should always be handled with dry hands. Abrasion, static electricity, water, or chemical spots will result in extraneous marks (artifacts) on the radiographs. Medicated hand creams should be avoided; rubber gloves should be used. Development Procedure. Prepared developers are ordinarily used to ensure a carefully compounded chemical that

gives uniform results. Commercial x-ray developers are of two types: automatic and manual. Both are comparable in performance and effective life, but the liquids are easier to mix. Developing time for industrial x-ray films depends mainly on type of developer. Normal developing time for all films is 8 min in a given developer at 20 °C (68 °F). More exact tables can be obtained from the manufacturers of developers. When exposed film is immersed in the developer solution, the chemicals penetrate the emulsion and begin to act on the sensitized (exposed) grains in the emulsion, reducing the grains to metallic silver. The longer the development time, the more metallic silver is formed and the blacker (more dense) the image on the film becomes. The rate of development is heavily dependent on the temperature of the solution; the higher the temperature, the faster the development. Conversely, if the developer temperature is low, the reaction is slow, and if the film were developed for 5 min at 16 °C (60 °F) instead of the normal 20 °C (68 °F), the resulting radiograph would be underdeveloped. Within certain, rather narrow, temperature limits, the rate of development can be compensated for by increasing or decreasing developing time. Exceeding these temperature limits usually gives unpredictable results. The concept of time-temperature development should be used in all radiographic work to avoid inconsistent results. In this concept, the temperature of the developer is always kept within a small range. The developing time is adjusted to temperature so that the degree of development remains essentially constant. If this procedure is not followed, the results of even the most accurate radiographic technique will be nullified. Inspection of the film at various intervals during development under safelight conditions (called sight development) should be avoided. It is extremely difficult to judge from the appearance of a developed but unfixed radiograph what its appearance will be in the dry, finished state, particularly with regard to contrast. Sight development can also lead to a high degree of fog caused by exposure to safelights during development. A major advantage of standardized time-temperature development is that the processing procedure is essentially constant, and an accurate evaluation of exposure time can be made. This alone can avoid many of the errors that can otherwise occur during exposure. Increased developing time will produce greater graininess in the radiographic image, increased film speed, and in many cases increased radiographic contrast. Although increased contrast or film speed is often desirable, maximum recommended development times should not be exceeded. Control of Temperature and Time. Because the temperature of the processing solutions has such a large influence

on their chemical activity, careful control of temperature--particularly of the developer--is extremely important. A major rule in processing is to check the developer temperature before films are immersed in the developer so that the timer can be accurately set for the correct processing time. Ideally, the developer should be at 20 °C (68 °F). At temperatures below 16 °C (60 °F), developer action is significantly retarded and is likely to result in underdevelopment. At temperatures exceeding 24 °C (75 °F), the radiograph may become fogged, and the emulsion may be loosened from the base, causing permanent damage to the radiograph. Where the water temperature in the master tanks surrounding the solution tanks may be below 20 °C (68 °F), hot and cold water connections to a mixing valve supplying the master tank should be used. In warm environments, refrigerated or cooled water may be necessary. Under no circumstances should ice be placed directly into the solution tanks for cooling purposes, because melting ice will dilute, and may contaminate, the solutions. If necessary, ice can be placed in the water bath in the master tanks for control of the solution temperature. Control of time should be done by setting a processing timer at the time the film is immersed in the developer. The film should be moved to the rinse step as soon as the timer alarm sounds. Agitation During Development. A good radiograph is uniformly developed over the entire film area. Agitating the

film during the course of development is the main factor that eliminates streaking on the radiograph.

When a film is immersed vertically in the developer and is allowed to develop without movement, there is a tendency for certain areas of the film to affect the areas directly below them. The reaction products of development have a higher specific gravity than the developer. As these products diffuse out of the emulsion, they flow downward over the surface of the film and affect the development of the areas over which they pass. As a result, uneven development of affected areas forms streaks, as shown in Fig. 67. This is sometimes referred to as bromide drag. The greater the film density from which reaction products emanate, the greater the effect on adjacent portions of the film.

Fig. 67 Streaking, or bromide drag, that can result when a film is immersed in developer solution without agitation

When the film is agitated, spent developer at the surface of the film is renewed, preventing uneven development. Immediately after hanger and film are immersed in developer, hangers should be tapped sharply two or three times to dislodge any air bells clinging to the emulsion. Although lateral movement of the hanger provides perhaps the best agitation, the size and shape of the solution tanks usually limit the extent of lateral movement, thus making this type of agitation ineffective. Vertical movement works well; it consists of lifting the hanger completely out of the solution, then immediately replacing it in the tank two or three times at intervals of about 1 min throughout the developing time. Agitation with stirrers or circulating pumps is not recommended. This type of agitation often produces a liquid flow pattern that causes more uneven development than no agitation at all. Nitrogen-burst techniques, although rarely used, can provide adequate agitation. Activity of Developer Solution. The developing power of the solution decreases when film after film is developed, partly because the developing agent is consumed in converting exposed silver bromide to metallic silver and partly because of the retarding action of accumulated reaction products. The magnitude of this decrease depends on the number of films processed and on their average size and density. Even when the developer is not used, its activity will slowly decrease because of oxidation of the developing agent. The effect of oxidation is often apparent after as little as 1 month of inactivity. Little can be done to control the effects of oxidation except using a lid on the developer tank. Although this is only partly effective in preventing oxidation, it is always good practice to cover developer tanks when not in use in order to prevent contamination. Replenishment of Developer. If the reduction of developing action is the result of the processing of many films, it is

possible to compensate for decreased chemical activity by using a replenishment technique. When done correctly, replenishment can maintain uniform development for a long period of time. Replenishment cannot be used to counteract oxidation or contamination of the developer solution. Most manufacturers of x-ray developers provide for replenishment either by supplying a separate chemical or by using the developer mixed to a different concentration from that of the original developer solution. The correct quantity of replenisher needed for maintaining consistent properties of the developer solution depends on the size and average density of the radiographs being processed. For example, a dense image over the entire radiograph will use up more developing agents and exhaust the developer to a greater degree than if the film were developed to a lower density. The quantity of replenisher required will depend on the type of subject being radiographed; the following is provided only as a guide:

Density

of

radiograph

Replenisher

(% of background exposed)

required(a)

g

oz

Low (90%)

78 2

(a) Approximate quantity for each 350 × 430 mm (14 × 17 in.) sheet of film processed

Usually, replenishers are so compounded that drainage of films back into the developer tank should be avoided. The developer being drained is essentially spent; therefore, the developer solution in the tank is more rapidly contaminated with reaction products when this spent solution is drained back. A systematic procedure should be used so that a fixed quantity of developer is removed from the tank for each film that is developed. A 350 × 430 mm (14 × 17 in.) film mounted on its hanger will normally carry with it 78 to 85 g (2 to 3 oz) of developing solution as it is removed from the tank. Because this is approximately the amount of replenisher needed for low-density and medium-density films, it is only necessary to replenish the developer tank to a given liquid level. However, for high-density and extremely high-density radiographs, it would be necessary to remove and discard some of the original developer each time replenisher is added. If replenisher is added frequently and in small quantities, fluctuations in film density due to changes in chemical activity of the developer will tend to even out. However, if replenisher is added infrequently, a fluctuation in film density will become apparent, which may lead to considerable difficulty in consistently obtaining the required image quality in successively processed radiographs. If replenishment is controlled by maintaining a specific level in the development tank, replenisher should be added when the level of the solution drops by 6 mm (

in.).

Arresting Development. After development is completed, the action of the developer absorbed in the emulsion must

be arrested by an acetic acid stop bath or at least by prolonged rinsing in clean running water. If these steps are omitted, the developing action continues for a short time in the fixer. This can produce uneven density or streaking on the radiograph and will reduce the life of the acidic fixer solution because of neutralization by the alkaline developer solution. If the smell of ammonia is detected anywhere--developer, stop bath, or fixer--the solution has become contaminated and must be changed. When development is complete, films should be removed from the developer without draining back. Films should remain in the stop bath for 30 s with moderate agitation before being transferred to the fixer solution. The stop bath acts as a replenisher for the fixer, so the films can be transferred directly to the fixer without draining back into the stop-bath solution. If a stop bath is not used, films should be rinsed in running water for at least 2 min. If the flow of water in the rinse tank is only moderate, the film should be agitated so that the rinse will effectively avoid streaks on the radiograph.

Fixing. The purpose of the fixer solution is to remove all of the undeveloped silver salts in the emulsion, leaving the

developed silver as a permanent image. Another important function of the fixer is to harden the gelatin of the emulsion so that the film will be able to withstand drying in warm air and will not be tacky to the touch when it is dry and ready for viewing. Portions of the film that have received lower amounts of radiation appear cream colored when the stop-bath procedure has been completed. In the fixing bath, this cream color gradually disappears until it is no longer visible. The interval of time required for this change to take place is called the clearing time. This is the time required to dissolve the undeveloped silver halide out of the emulsion. An equal amount of time is required to allow the dissolved silver salts to diffuse out of the emulsion and the gelatin to harden. The total fixing time, therefore, should be at least twice the clearing time but should not exceed 15 min to avoid loss of density on the film. Films should be agitated vigorously every 2 min during fixation to ensure uniform action of the fixing chemicals. In performing its function, the fixer solution accumulates soluble salts, which gradually inhibits its chemical activity. As a result of this, and possibly also because of dilution with water, the rate of fixation decreases and hardening action is impaired in proportion to the number of films processed. The usefulness of fixer solution is ended when it loses its acidity or when fixing requires an unusually long interval. At this point, the fixer solution must be discarded. Fixing films in exhausted fixer frequently causes colored stains, known as dichroic fog, to appear on the radiograph. Fixer replenisher is usually available from the manufacturer of the fixer. It is advisable not to substitute brands (or even types) of replenishers for a given brand of fixer; the replenisher may not be compatible with the fixer. Washing. After radiographic films have been fixed, they should be washed in running water, ensuring that the emulsion

area of the film receives frequent changes of water. Proper washing also requires that the hanger bar and top film clips be covered completely by running water. Effective washing of the film depends on a sufficient flow of water to rapidly carry off the fixer and to allow adequate time for fixer chemicals to diffuse out of the film. In general, the hourly flow of water in the washing tank should be from one to two times the volume of the tank. Under these conditions, and at water temperatures between 16 and 21 °C (60 and 70 °F), films require about 30 min of washing. The washing tank should be large enough to wash films as quickly as possible. Too small a tank encourages insufficient washing, which may lead to discoloration or fading of the image later when radiographs are in storage. Drying. When films are removed from washing tanks, water droplets cling to the surface of the emulsion. If the films are then dried rapidly, the areas under the droplets dry more slowly than surrounding areas. Such uneven drying causes distortion of the gelatin, changing the density of the image. Uneven drying often results in visible spots on the finished radiograph that interfere with accurate interpretation.

Water spots usually can be prevented by immersing washed films for 1 to 2 min in a wetting-agent solution and allowing them to drain for a few minutes before placing them in a drier. This procedure permits surplus water to drain off of the film more evenly, significantly reducing drying time and the likelihood that the finished radiograph will exhibit water spots. It is important to use wetting agents that are compatible with x-ray film emulsions. Some commercial wetting agents are not compatible with film and therefore should not be used. If only a few radiographs are processed daily, racks that allow the film to be air dried under room conditions of temperature and humidity can be used, although this method of drying requires considerable time. To avoid spots on the radiographs, the racks that hold the film hangers should be positioned so that films do not touch each other, and water should not be splashed on drying films. Radiographs dry best in constantly changing warm dry air. The fastest way to dry films is in commercial film driers. These driers incorporate fans and heaters, and some driers use chemical desiccants to remove water from the air. Regardless of the method of drying, a radiograph is considered dry when the film is dry under the hanger clips. When this stage is reached, the hangers are removed from the drier or rack, and the film is carefully removed from the hanger for interpretation.

In manual processing, when the film is clipped to the hanger, pins in the hanger clip penetrate the film and leave sharp projections in the corners of the film. It is usually desirable to cut off the corners containing these clip marks to prevent scratching of other radiographs during handling and reading. Automatic Film Processing Although expertly done manual processing is difficult to surpass, it is equally difficult to maintain because of human factors. Automatic processing, which accurately controls temperature, time, agitation, and replenishment, delivers a dry radiograph in a short time. In addition to the advantage of shorter processing time, automatic processing can ensure that variability of time, temperature, and activity of the solutions is eliminated. Many automatic processors incorporate a roller-transport mechanism that carries the film itself through the entire process cycle without the need for hangers. Figure 68 illustrates a typical automatic processor that incorporates a roller-transport mechanism.

Fig. 68 Cross section of an automatic film processor showing roller-transport mechanism and locations of components

Automatic processing is a carefully controlled system in which the film, processing chemicals and their replenishment, temperature of solutions, travel speed of the roller-transport mechanism, and drying conditions all work together for consistent development of latent images. The advantages of roller-type automatic processors are: •

•

Rapid processing of radiographs: Depending on many variables, particularly the kind of radiographic film and the type of processing chemicals, the elapsed time from exposed film to finished radiograph can be as short as 4 min and usually not more than 15 min. This represents a savings of at least 45 min compared to manual processing Uniformity of radiographs: Most automatic processors carefully control the time and temperature of processing. Combined with accurate automatic replenishment of solutions and constant agitation of the

•

film as it travels through the processor, accurate time-temperature control produces day-to-day consistency and freedom from processing artifacts seldom achieved in manual processing. Because processing variables are virtually eliminated, optimum exposure techniques can be established Conservation of space: Most models of automatic processors require no more than 1 m2 (10 ft2) of floor space. Some models can even be installed on a workbench. This means that hand tanks and drying facilities can be eliminated from the processing room. All that is needed is a bench for loading and unloading of film, film-storage facilities, and a small open area in front of the processor feed tray. Most processors are installed so that the film is fed from the processing room (darkroom) and emerges as a complete radiograph in the adjacent room

Automatic processors contain a developing tank, a fixing tank, a washing tank, and a drying section (Fig. 68). The stopbath and wetting-agent tanks are eliminated because automatic processors have squeegee rollers at the exit of each tank that reduce retention of residual solution by the film to a minimum, and minor solution contamination of the fixer is corrected by replenishment. The rollers at the exit of the fixer reduce the amount of residual fixer to only that absorbed in the emulsion as the film enters the wash cycle. Therefore, the running wash water is virtually free of fixer chemicals. Finally, the exit rollers of the wash tank squeeze wash water from the film, so that the film enters the drier in almost a damp-dry condition. Processing chemicals must be specially formulated for automatic processors. The developer must operate at higher temperatures than for manual processing and usually contains a hardening agent to condition the emulsion so that the film can be moved by the roller transport system without slippage. The fixer is also specially formulated to fix the emulsion in a relatively short time at higher temperatures than for manual processing and to condition the film for proper drying. It should be noted that developers and fixers formulated for manual processing usually are not suitable for automatic processing. Replenishment of Developer and Fixer. In an automatic processor, replenishment of developer and fixer is

automatically controlled. An adjustable, positive-metering device controls both developer and fixer; sometimes separate metering devices are used for developer and fixer. Usually, the replenisher pump is activated when the leading edge of the film enters a film sensor in the processor and continues to pump until the trailing edge of the film passes these rollers. Thus, the amount of replenishment is controlled by the length of film passing through the entrance rollers. Replenishment rates for developer and fixer are normally supplied by film manufacturers. Obviously, these are only guidelines because if radiographs are routinely of higher or lower densities than the average on which the manufacturer's recommendations are based, the replenishment rate may have to be adjusted upward and downward for optimum results. The procedure for checking replenishment rates and frequency of replenishment is given in the operator's manual for the processor. The accuracy of replenishment is important. Too little or too much replenishment can adversely affect film densities, lead to transport difficulties, reduce processing uniformity, and shorten the useful life of the processing solutions. Film-Feeding Procedures. Because replenisher pumps are controlled by the length of film fed into the processor, it is

obvious that feeding single, narrow-width films will cause excessive replenishment. Therefore, whenever possible, narrower films should be fed side by side. Films should be fed into the processor parallel to the side of the feed tray. Multiple films should have a space between them to avoid overlapping and should be started together into the processor to avoid excessive replenishment. Three rolls of the roller-transport system must always be in contact with the film to maintain proper travel through the processor. Thus, there is a lower limit to the length of film to be processed. Although this depends on roller diameter and spacing, the usual lower limit is about 125 mm (5 in.). In general, roll films having widths from 16 to 430 mm (0.6 to 17 in.) can be processed in most automatic processors. The processing of roll films requires a somewhat different procedure than for sheet film. Because roll film is wound on spools, it frequently has an inherent curl that can cause the film to wander out of the roller system. To avoid wandering, a sheet of leader film can be attached to the leading edge of the roll. Ideally, the leader should be unprocessed radiographic film, preferably wider than the roll being processed and at least 250 mm (10 in.) long. The leader may be attached to the roll by means of pressure-sensitive polyester tape about 25 mm (1 in.) wide. Suitable types of tape must be composed of

materials that are not soluble in the solutions. To avoid transport problems, care must be taken that adhesive from the tape does not come in contact with the rollers. It is important to feed narrow widths of roll film parallel to the edge of the feed tray. If quantities of such film are normally processed, it is usually advisable to provide a guide in the feed tray to make sure each film is parallel to the others and to the sides of the tray. If this is not done, there is a possibility of the films overlapping somewhere in the transport systems. If only one long roll of narrow film is to be processed, the replenisher pumps will keep running and the result may be excessive replenishment. To avoid this, replenisher pumps should be turned off for a portion of the feed time. Preventive Maintenance. Most of the downtime and other problems related to the operation of automatic processors stem from the lack of maintenance. Service problems can be minimized by well-established maintenance of the processor and by good housekeeping. Each processor manufacturer recommends daily and weekly cleanup procedures that take only a few minutes to perform. These procedures, usually available in the form of check lists, are necessary for reliable operation of the processor and for production of radiographs of optimum quality. Precautions. Automatic processing is a system in which the film, chemical, and processing equipment all have to work

together for optimum processing. For example, if radiographs leave the processor wet, it could well be an indication of a problem with one of the chemicals and not the result of incorrect drying temperature. Some films can be successfully processed more rapidly than others. However, changing the speed of the processor to a value other than that for which the processor was designed can unbalance the system and require adjustments in film characteristics, replenishment rates, temperatures, and perhaps other conditions in order to restore optimum processing quality. Care must be taken that the storage life of the radiograph is not impaired by changes in the processor. Table 14 lists some typical problems with automatic film processing. Table 14 Probable causes and corrective action for various types of deficient image quality or artifacts from automatic film processing See Table 12 for factors not specifically associated with automatic film processing. Quality or artifact

Probable cause

Problem and corrective action

Density too high

Overdevelopment

Developer temperature too high: Follow temperature recommendations for developer and processor used. Check temperature of incoming water. Check accuracy of thermometers used.

Improperly mixed chemicals: Follow instructions for preparations of solutions.

Density too low

Underdevelopment

Improper replenishment of developer: Check for clogged strainers or pinched tubing in developer replenishment system.

Developer temperature too low: Follow temperature recommendations for developer and processor used. Check temperature of incoming water. Check accuracy of thermometers used.

Contrast too low

Contamination

Fixer in developer tanks: Use extreme care when installing or removing fixer rack in processor. Always use splash guard when fixer rack is being removed or replaced. Do not exchange racks between fixer and developer compartments.

Underdevelopment

See Underdevelopment above.

Contamination

See Contamination above.

Fog

Overdevelopment

Developer temperature too high: Follow temperature recommendations for developer and processor used. Check temperature of incoming water. Check accuracy of thermometers used.

Poor drying

Processing

Underreplenishment of solutions (particularly fixer): Check for clogged strainers or pinched tubing in replenishment system.

Inadequate washing: Check flow of wash water.

Drying

Dryer temperature too low: Follow recommendations for film type and processor involved.

Streaks

Associated with tempo of work

Long interval between feeding of films: "Delay streaks" (uneven streaks in direction of film travel) caused by interval of 15 min or more in feeding of successive films, which results in drying of solutions on processor rollers exposed to air. Wipe down exposed rollers with damp cloth. Process unexposed 14 × 17 film before processing radiograph. (Some processors are equipped with "rewet rollers" which prevent delay streaks.)

Associated with development

Clogged developer recirculation system: Change filter cartridge regularly. Check recirculation pump. (This defect is associated with a rapid rise in developer temperature.)

Associated with dryer

Dirty tubes in dryer: (Causes regular streaks, visible by reflected light only.) Clean dryer tubes.

High dryer temperature: (Causes irregular streaks or blotches, visible by reflected light only.) Reduce temperature to recommended value.

Guide marks (regularly spaced scratches or high density lines)

Improperly adjusted guides in processor

Check clearances between guide devices and adjacent rollers or other components

Roller abrasions (random fine lines in the direction of film travel)

Stopped or hesitating rollers

Be sure all rollers are in their proper positions, and that end play is sufficient for rollers to turn freely.

Random scratches and spots

Dirt on feed tray

Clean processor feed tray frequently with soft cloth. If the atmosphere is dirty, the processing room should be fed with filtered air and kept at a pressure above that outside.

Irregular deposit (often light in color, generally elongated in direction of film

Caused by dirt or precipitate in water supplied to washing section.

If condition is temporary, clean wash rack and replace wash water in processor; drain wash tank when shutting processor down. If condition persists, use filters in incoming water lines. (Some dirt deposits can be removed from the dry radiograph by gentle rubbing

travel)

Dark lines and spots

with dry cotton or a soft cloth.)

Pressure marks caused by build-up of foreign matter on rollers or by improper roller clearances, usually in developer section.

Clean rollers thoroughly and maintain proper clearances.

"Black comets" with tails extending in direction of film travel caused by rust or other iron particles dropping on film, usually at entrance assembly.

Clean all entrance assembly components. Apply a light coat of grease to microswitch springs and terminals. If air contains iron-bearing dust, filter air supply to processing room.

"Pi lines." (So called because they occur 3.14 times the diameter of a roller away from the leading edge of a film.)

Most common in newly-installed or freshly cleaned processors. Tends to disappear with use of processor. (Some processors are equipped with buffer rollers at the exit of the wash rack, which remove the deposit before the radiograph enters the dryer.)

Tests for Removal of Fixer If radiographic films are not properly washed, fixer chemicals (largely thiosulfate salts) remain in the emulsion and affect the storage life of the radiographs. Radiographs intended for storage of 3 to 10 years are usually referred to as having commercial quality. Those to be kept for 20 years or more are known as having archival quality. Archival quality is of considerable importance in complying with certain codes, standards, and specifications. If residual thiosulfate in the radiographs exceeds a certain maximum allowable level, the radiograph is likely to become useless because of fading or a change in color during long-term storage. The American National Standards Institute has three important documents relating to this problem: ANSI PH 4.8 (1985), ANSI PH 1.41 (1984), and ANSI PH 1.66 (1985). The first document (ANSI PH 4.8) describes two methods of determining residual thiosulfate in radiographs. Both of these are laboratory procedures for evaluating unexposed but processed (clear) areas of a radiograph. The methylene blue test must be performed within two weeks of processing the film, but the silver densitometric test can be performed at any time after processing. The second document (ANSI PH 1.41) specifies, among other things, the maximum level of thiosulfate in grams per unit area for archival storage. The third document (ANSI PH 1.66) gives other helpful recommendations for storage of radiographs. Procedures given in all three ANSI documents are accepted as valid procedures by most codes, standards, and specifications. There are other tests that are easier to perform than those described in ANSI PH 4.8, but they indicate more than the residual level of thiosulfate and therefore are considered only estimates rather than accurate determinations. In one test, a solution is made of 710 mL (24 oz) of water, 120 mL (4 oz) of acetic acid (28%), and 7 mL ( oz) of silver nitrate and diluted with 950 mL (32 oz) of water prior to use. When a drop of the diluted solution is placed on a clear area of a radiograph, it turns brown. The amount of residual fixer chemicals is estimated by matching the brown spot on the film with one of the patches on a standard test strip. Several manufacturers produce estimating kits, but for accurate determination of residual thiosulfate it is necessary to perform either the methylene blue test or the silver densitometric test described in ANSI PH 4.8. Microfilming of Radiographs The microfilming of radiographs has recently been implemented on a commercial basis for the storage of industrial radiographs, and it has been in use for some time in the medical field. The microreduction of radiographs has been successfully performed by the commercial nuclear power plant industry as a solution to problems associated with radiographic film deterioration, large-volume archival storage, and general records management. Because of improper initial processing or failure to store and handle processed film using sound practices, radiographs are subject to deterioration or degradation. Specific examples of improper processing include:

• • • • •

Inadequate replenishment of chemicals (developer and fixer in particular) Failure to maintain processing temperatures within manufacturer's specified ranges Inadequate agitation Rushing processing times Failure to maintain water quality

The consequences can be failure to remove thiosulfate, resulting in a green, yellow, amber, or brown tint or image destruction, and failure to remove uncombined silver resulting in general darkening of the image. Examples of poor storage and handling practices include: • • • • • • • •

Handling radiographs without the use of gloves to protect the film from acids, chemicals, and perspiration on the hands Placing rubber bands or paper clips on the film Storing hard copy information produced using a diazo process with film Stacking film, especially film not interleaved, causing it to adhere Filing wet film Placing inventory or acceptance punches in the area of interest Failure to store the radiographs in low-humidity/room-temperature conditions Failure to store radiographs in acid-free interleaving paper

The microfilming of radiographs has been used to transfer the deteriorated or degraded image to a stable, archival medium. Microfilming is performed using a planetary camera system with illuminators to photograph the radiographic image on 35 mm roll format. Spatial resolutions in excess of 120 line pairs per millimeter (or 1500

...

...

...

Photodiode array

Good

1024

5s

1000

...

...

...

CCD

Fair

1320 × 1035

140 ms

500

...

1.5

0.14

CID

Fair

776 × 512

33 ms(b)

200

...

0.1

0.009

Vidicon (Sb2S3)

Fair

>1500

33 ms, lag = 20%

80-200

...

5

0.5

Newvicon (ZnSe)

Fair

800

33 ms, lag = 10-20%

50-200

...

0.5

0.05

Pasecon/Chalnicon (CdSe)

Fair

1600

33 ms, lag = 510%

30-60

...

1

0.09

Saticon (Se+Te+As)

Fair

1200

33 ms, lag = 3%

50-160

...

6

0.6

Plumbicon (PbO)

Fair

1200

33 ms, lag = 4%

80-160

...

2.5

0.23

SIT (Silicon)

Fair

750

33 ms, lag = 7%

60

...

0.01

0.0009

(a) Aperture.

(b) 776 × 512

Linear CCD or photodiode arrays are good choices for film digitization and can have even better dynamic ranges than those listed in Table 2. When these devices are cooled, each 7 °C (13 °F) reduction in temperature reduces the root mean square noise by a factor of two. The charge-coupled device and charge injection device are capable of good dynamic range and fair resolution, but have certain artifacts (Ref 12). Again, if the devices are cooled, the noise floor is reduced, and they can be integrated for long periods to enhance the dynamic range and sensitivity. Cooled CCDs are available that rival the low light sensitivity of silicon-intensified targets (SITs), but not the frame speed. In general, the frame rate, dynamic range, and resolution are all interrelated. The interscene dynamic range is listed as the maximum achievable for the microdensitometers and as the dynamic range that can be achieved at the given frame rate for the other devices. The faceplate illumination given for the tubes and the solid-state detectors assumes mid-level illumination (halfway between saturation and preamplifier noise) (Ref 13). This may vary among tubes and generic types by a factor of three to four (Ref 14). The interscene dynamic range will also vary greatly, but can be maximized by the proper selection of a tube such that a dynamic range of 200 can be achieved at 33 ms/field with a high resolution (>1000 lines). The quoted resolution for the tubes is at a modulation transfer function (MFT) of 5%, which means that the contrast between a black line and a white line is only 5% at the stated resolution. This, of course, is measured at optimum illumination and at the center of the tube image field. Under other conditions, the resolution will be less. For comparison, a 10242 CCD camera may have an MTF of 5% at 750 lines. The lags quoted for the tubes are typical for the particular type at 3 TV fields or 50 ms. Tube cameras are primarily used in radiation environments or in specialized applications, such as high-resolution real-time radiography, in which the frame rate is higher than that achievable by current CCD designs. Charge-coupled device camera design is rapidly evolving for high-resolution scientific use and can be expected to improve with regard to real-time frame rates (Ref 12, 15). It should be noted that a 355 × 432 mm (14 × 17 in.) film digitized at a resolution of 50 m (0.002 in.) with 12-bit accuracy will consume a 92-Mbyte file. Just writing or reading this file to or from a hard disk could take up to 8 min (optical disks take even longer). Even with high-density optical disks and data compression, the digital storing of highresolution radiographs represents a formidable problem.

References cited in this section

3. R.C. Gonzalez and P. Wintz, Digital Image Processing, Addison-Wesley, 1977 12. J.R. Janesick, T. Elliott, S. Collins, M.M. Blouke, and J. Freeman, Scientific Charge Coupled Devices, Opt. Eng., Vol 26 (No. 8), 1987, p 692-714 13. I.P. Csorba, Image Tubes, Howard W. Sams & Co., 1985 14. G.I. Yates, S.A. Jaramillo, V.H. Holmes, and J.P. Black, "Characterization of New FPS Vidicons for Scientific Imaging Applications," LA-11035-MS, US-37, Los Alamos National Laboratory, 1988 15. L.E. Rovich, Imaging Processes and Materials, Van Nostrand Reinhold, 1989 Digital Image Enhancement T.N. Claytor and M.H. Jones, Los Alamos National Laboratory

Image Processing

Prior to image processing, it may be necessary to perform some type of preprocessing on the data. Most often, an image file will need to be converted into the workstation standard from some other type of format; for example, the files may be 8-, 16-, or 24-bit uncompressed format with a variable-length header and need to be converted to 8-bit format with no header. Typically, a problem arises when a file must be read that is in an unknown size, a compressed format, or has a limited color or nonstandard palette. Other preprocessing functions operate on the raw data in some manner before inputting it for image processing. For example, the system should have the capability to average, filter, and acquire full frames and control the scanning parameters on the digitizer so that the noise level can be minimized. In other cases, more complicated operations are called for, such as tomographic reconstruction, synthetic aperture operations, or an FFT. Shown in Fig. 1(a) and (b) in the article "Use of Color for NDE" in this Volume are tomographs taken before and after oversampling by a factor of four and Wiener filtering of the input data set to correct for the point spread function (Ref 16). The preprocessing operation greatly increases the smoothness of the lines and reduces the artifacts.

Reference cited in this section

16. K. Thompson, private communication, Sandia National Laboratories, 1988 Digital Image Enhancement T.N. Claytor and M.H. Jones, Los Alamos National Laboratory

Image Enhancement There are three major scientific applications for image processing: • • •

Enhancement of an image to facilitate viewing Manipulation and restoration of an image Measurement and separation of features

These functions are listed in Table 3. Many of the functions have a dual purpose and, as will be shown, are usually combined to form other functions. Table 3 Image-processing software algorithms useful for NDE applications Image enhancement

Image operations

Information extraction

Contrast stretching

Scaling

Image statistics

Histogram equalization

Translation

Point, line, angle perimeter, area, measurement

Contouring

Rotating

FFT transformation, one and two dimensional

Thresholding

Registration

Correlation

Composite image building

Warping

Edge detection

Palette operations

Combining

Deblurring

Color model selection

Filtering

Motion restoration

True color representation

Thickening, thinning Noise cleaning Trend removal

Pattern recognition

Contrast Stretching and Histogram Equalization. The concept of contrast stretching is shown in Fig. 3(a). The

basic operation is given by the pixel value transform:

s = T(r) s = T(r) = ar + b Linear T(r) = a log(r) + b Logarithmic

(Eq 1)

where r is the original pixel value and s is the transformed pixel value.

Fig. 3 Concept of histogram stretching and equalization. (a) Histogram is stretched with a linear transformation. (b) Histogram is equalized such that the probability density is constant.

The linear stretch and offset method is the most commonly applied, while transforms of the nonlinear type can be used to convert film density to integrated dose or to linearize the film transfer function to account for base density. Although the manual contrast stretch is often used, a more automated contrast stretch is available with an operation known as histogram equalization. The general form is:

(Eq 2)

A simple linear equalization example is shown in Fig. 3(b). Occasionally, it may be desirable to use the cubic or logarithmic equalization. The cubic emphasizes the lower values, making the image darker, while the exponential makes the image much brighter. Other types of transfer functions are often used to remove nonlinear response in the imaging or camera system. For isolated defects, a manual stretch usually gives good results. On textured materials, however, histogram equalization often works well. An example of a linear histogram equalization of a low-contrast ultrasonic image of a bond line in an explosively bonded steel-to-aluminum plate is shown in Fig. 2 in the article "Use of Color for NDE" in this Volume. Contouring and Thresholding. Contouring and thresholding are often combined with palette operations to show

gradations in intensity or to highlight defects. Color should be used to detect gradients in intensity because the eye is sensitive to many more colors (approximately 104 shades) than gray levels. Thresholding can be done with a binary picture (black and white), a color scale, or a combination of gray and color. The concept of thresholding has particular application in the generation of a mask for future image-processing operations or the removal of the background. Contouring is used to change only those values between two pixel levels to a certain value. It can also be used in outlining high-definition features before further processing. Composite Image Building. The ability to display composite images is useful in comparing the difference between

images produced by different modalities, such as a radiograph and an ultrasonic or NMR image. Because image quality suffers greatly when displayed at resolutions of less than 5122 it is imperative to have at least a 10242 display or to use two monitors if detailed comparisons are to be made. Figure 4 illustrates how x-ray radiographs and neutron radiographs can be displayed together and combined to enhance voids in a composite material. As shown in Fig. 4(c), the images have simply been added to enhance the detection of voids. The absorption of neutrons in the epoxy is higher than in the alumina and the reverse is true for the absorption of x-rays. However, the voids do not absorb in either case, and the contrast is therefore improved if the images are added.

Fig. 4 Voids detected in three 60 × 60 × 10 mm (2.4 × 2.4 × 1 in.) alumina-filled epoxy tiles using three different imaging techniques. (a) Conventional radiography. (b) Neutron radiography. (c) Combined x-ray and neutron radiographic image. Building a composite image of both modalities indicates that there are voids in the sample. The addition of the two images enhances the voids because of the differential absorption in the matrix between the x-ray and neutron images.

Palette Control. Color can be displayed as a true color map or as a pseudocolor representation. In a pseudocolor

representation, the digital data values are mapped to any color that can be produced by the combination of the red, green, and blue (RGB) values specified by a look-up table (LUT). A typical 8-bit pseudocolor system is shown in Fig. 5(a). With only an 8-bit-deep pixel as shown in Fig. 2, a type of true color is possible; however, each color is represented only by eight or fewer color levels (Fig. 5b). With 8-bits per color (as in Fig. 5c), the colors can be shaded continuously and are suitable for three-dimensional displays in which depth cueing is indicated by decreasing the luminance. Eight-bit pixels with 24-bit LUTs can produce good color scenes by adjusting the LUT to include only those colors that appear in a scene. However, the palette is then image specific.

Fig. 5 Concept of a look-up table for three types of color displays. (a) 8-bit pseudocolor representation. (b) 8bit true color map. (c) 24-bit true color map. The LUT is controlled by the palette, which specifies what color composition (R,G,B) will be assigned to one of 256 levels, as in (a). The digital-to-analog converter (DAC) converts the digital signal to RGB for input to the monitor. Shown in (b) is a way to display true color with an 8bit pixel. Better true color can be obtained by using a 24-bit pixel, with each byte (8 bits) assigned to a specific primary color, as in (c).

Several types of pseudocolor and linear gray-scale palettes are shown in Fig. 4 in the article "Use of Color for NDE" in this Volume. The spectrum palette (or inverse spectrum) is preferred for range of color, but is not suitable for displaying rapid changes in pixel level. Other palettes, such as the complementary palettes shown in Fig. 4(d) and (e), can be used with good results on figures that have gradual changes in gray level. A typical threshold palette is shown in Fig. 4(b). All values below 192 will appear as a gray scale, while those equal to or above 192 will appear as red. The gray scale is still used to display lower-intensity data so that the operator can still detect unusual, yet not rejectable, anomalies. In contrast stretching, when further work on the image is anticipated, the palette (that is, LUT values) is altered, leaving the pixel values intact. Color Models and the Use of Color. There are two main color models: the red, green, blue (RGB) model and the hue, saturation, luminance (HSL) model. High-resolution imaging systems use the RGB model, while the National Television Systems Committee has adopted the HSL model for broadcast television. The colors in the RGB model are simply an additive mixture of the Commission Internationale de l'Éclairage monochromatic primary colors (red = 700 nm, or 7000 Å; green = 546.1 nm, or 5461 Å; and blue = 435.8 nm, or 4358 Å). When equal intensities of these colors are combined, a nearly white light results. In the HSL model, the primaries are transformed such that the hue value represents the color (0 to 360°; red = 0, 360; green = 120; and blue = 240), the saturation is the color intensity (values 0 to 1), and

the luminance is the amount of brightness (0 to 1). If the saturation is 0 and the luminance is 1, the color will be white. The advantage of the HSL model is that a single hue can be manipulated in intensity more easily than with the RGB model. Although both color and black-and-white images contain the same information, there are features that are much more obvious in color plots than in corresponding black-and-white plots, and vice versa. In particular, a black-and-white representation is appropriate when there is a high spatial frequency inherent in the image; in such a case, the mind-eye combination interprets this as a texture. On the other hand, color is preferred when there are a few isolated objects or low spatial frequency. This makes it easier for the image interpreter to discern small changes in image density through the use of contrast enhancement. Black-and-white images are shown in Fig. 6 and 7, and the color versions are shown in Fig. 6(c) and (d) in the article "Use of Color for NDE" in this Volume.

Fig. 6 Ultrasonic image of a hot isostatic pressed tungsten plate prepared from powder (99% dense). The black-and-white image shows small changes in density that are not easily discernible. The density changes are enhanced by the use of color, as shown in Fig. 6(c) of the article "Use of Color for NDE" in this Volume. The ring is caused by a transition from one thickness to another (the outer thickness was about 50% of the inner circle).

Fig. 7 Tungsten plate similar to the one shown in Fig. 6, except that this part was fabricated from a plate that was rolled rather than hot isostatic pressed. The black-and-white image shows a texture imparted to the material due to the inclusion of 2% porosity. The color version of the plate, Fig. 6(d) in the article "Use of Color for NDE" in this Volume, shows how difficult the texture is to interpret in a color representation using a spectrum palette.

Digital Image Enhancement T.N. Claytor and M.H. Jones, Los Alamos National Laboratory

Image Operations Listed in the second column of Table 3 are some of the more important image operations that are routinely used to manipulate images. Geometric Processes. The scaling, rotation, translation, warping, and registration of an image are classified as

geometric processes. The scaling of an image is an important function because it is often used before images are combined. Scaling is used to

magnify or shrink the images permanently to fit printer sizes or simply to view areas closely. Rotation and translation are important mainly if two images need to be registered closely and combined or

compared. The operations are described below. The scaling operation can involve interpolation, or it can merely involve a duplication of pixels. In the case of integer scaling (×2, ×3), the pixels are replicated in the x and y by the integer. For noninteger scaling (such as magnification by 1.5), every other pixel is replicated. Image interpolation and noninteger scaling should be avoided unless the image is subsequently filtered or unless interpolation and scaling are the last steps in a processing chain before printing. The printer will often perform interpolation and dithering (adding a small random number to the value) to obtain an image with a smoother appearance:

(Eq 3)

Warping is a nonlinear scaling process. The equation for a sixth-order warp is:

u = a1 + a2x + a3y + a4xy + a5x2 + a6y2 v = b1 + b2x + b3y + b4xy + b5x2 + b6y2

(Eq 4)

This transform can be of use if an object is at an angle to the viewing plane and it is necessary to measure a feature of interest or to correct the image for the viewing angle.

Image registration can be performed by trial and error with the rotation, translation, and scaling functions. The

goodness of fit can be calculated with the correlation function or by subtraction. Registration is usually a prior step to image combination. Image Combination. Image combination operations generally include the following: AND, NAND, OR, NOR, XOR,

difference, subtract, add, multiply, divide, and average. All these operations require two operands. One will be the data in the primary image, the other can be either a constant or data in a secondary image. The image combination operations are used to mask certain areas of images, to outline features, to eliminate backgrounds, to search for commonality, and/or to combine images. For example, a video of a part can be taken, edge enhanced, boosted, ANDed with 128 added to 127 and scaled and registered with a radiograph or other image, and then added to the other image. Filtering. There are many types of spatial filters. The most useful are low pass, high pass, edge, noise, and morphological, as listed in Table 4.

Table 4 Commonly used image filters Convolution filters

Other filters

High pass

Median

Low pass

Erosion

Horizontal edge

Dilation

Vertical edge

Unsharp mask

Laplacian

Roberts

Compass

Sobel or Prewitt

Wiener

Many filters are classified as convolution filters. The convolution filter uses the discrete form of the two-dimensional convolution integral to compute new pixel values. In Eq 5, the image matrix is defined as f(x,y), and the filter is a small 3 × 3 or (in general, m × n) matrix called a kernel, K. In practice, the kernel is limited to a size of less than 20 × 20 because the FFT filtering procedure becomes faster than direct convolution for large kernels. The discrete form of the convolution integral is:

(Eq 5)

Because kernels are often symmetric about a 180° rotation about the center pixel, algorithms that use the correlation integral will give identical results to the convolution operation. The discrete form of the correlation integral is:

(Eq 6) In this representation, one can visualize the kernel sliding over the image, with each pixel under the kernel being multiplied by the kernel value and then all values being summed to produce the new center pixel. It is customary to perform the convolution such that the initial and final image sizes are identical, with the outer row and column of the image not convolved with the kernel. Some of the most useful convolution kernels are shown in Fig. 8. Those kernels that add to 0, such as the Laplacian and the horizontal and vertical edge kernels, will remove most of the image except the edge components. The low-pass and noise reduction filters will blur the image and remove point noise, while the highpass filters tend to sharpen the images and accentuate noise.

Fig. 8 Various convolution kernels and their use

Other filters that are useful in image processing are the median, erosion, dilation, and unsharp mask filters. The median

filter selects the median value in an image neighborhood (usually 3 × 3) and replaces the center pixel with the median value. This filter can be very effective in removing isolated pixel noise without blurring the image as severely as the lowpass filter. The effect of a low-pass and a median filter is shown on a noisy ultrasonic image of a hot-rolled plate in Fig. 3(a), (b), and (c) in the article "Use of Color for NDE" in this Volume. The erosion and dilation filters work by replacing the center pixel value in a neighborhood with the smallest or largest index in the neighborhood, respectively. They can be used to thicken or to thin boundaries. The unsharp mask filter algorithm is:

g(x,y) = 2 · f(x,y) - L P F (f)xy

(Eq 7)

where LPF is the low-pass filter. The new image is the difference between the original image multiplied by a factor and the low-pass filtered image. This operation tends to enhance the contrast and to sharpen the edges slightly. The Roberts and Sobel filters are similar types of edge detectors. The Roberts filter algorithm is:

g(x,y) = {[f(x,y) - f(x + 1, y + 1)]2 + [(f(x,y + 1) - f(x + 1, y)]2}1/2

(Eq 8a)

This algorithm operates on a 2 × 2 neighborhood and will enhance high frequencies, while the Sobel filter operates with a 3 × 3 kernel and will produce good (but thick) outlines of images. The Sobel filter algorithm is:

g(x,y) = ({[(f(x + 1, y - 1) + 2 · f(x + 1, y) + f(x + 1, y + 1)] - [(f(x - 1, y - 1) + 2 · f(x - 1, y) + f(x - 1, y + 1)]}2 + {[f(x - 1, y - 1) + 2 · f(x, y - 1) + f(x + 1, y - 1)] - [f(x - 1, y + 1) + 2 · f(x, y + 1) + f(x + 1, y + 1)]}2)1/2

(Eq 8b)

A summary of the effect of various filters and edge detectors is shown in Fig. 9.

Fig. 9 Ultrasonic image of an adhesive bond (250 m, or 0.01 in., thick) between two opaque plastic parts. (a) Unfiltered. (b) Filtered with numerous filter types (vertical edge, high pass, Laplacian, Roberts, and Sobel)

There is a class of filters that restore image quality by considering a priori knowledge of the noise or the degradation mechanism. An example is the inverse filter. In this case, the image has been degraded and is reestimated by the use of

the inverted degrading transfer function and a noise model to minimize the least square error. This type of filter can be useful if a transfer function model is derived for the imaging chain. For noise cleaning, the filters listed in Table 4 and the median filter can be used. Fast Fourier transform techniques are particularly effective for the removal of periodic noise, especially filtering in the frequency domain. Trend removal, or field flattening, is an important processing function to perform before contrast enhancement. In

most cases, an x-ray film will not have a uniform density even if the part is uniform. In addition, variations in density may occur when the part is digitized with a camera. When the technique of contrast enhancement is attempted, parts of the image will saturate, and defects will be difficult to detect, as in the ceramic disk shown in Fig. 6(a) of the article "Use of Color for NDE" in this Volume. Listed below are techniques used to reduce the effect:

1. High-pass filter using an FFT 2. Low-pass filter using a large kernel or FFT, and subtracted from the original image 3. Polynominal fitted to image line and fitted curve subtracted from original

Most of these algorithms produce artifacts and are not entirely satisfactory, especially if there is a high spatial frequency superimposed on a low-frequency trend. Figure 6(b) of the article "Use of Color for NDE" in this Volume shows the results of using technique 2 (similar to unsharp masking) on a noisy image with a top-to-bottom trend.

Digital Image Enhancement T.N. Claytor and M.H. Jones, Los Alamos National Laboratory

Information Extraction Listed in the third column of Table 3 are image-processing functions that extract quantitative information concerning images. Image Statistics and Measurement. The most useful image statistic needed for subsequent processing is the histogram or the probability density function. This will define the dynamic range of the image and determine subsequent processing. Also of use is the line profile and frequency analysis of the line or small area of interest. The line profile is of use when quantitative analysis needs to be performed on an image--for example, measurement of the MTF of the imaging system or the variation in pixel value across a particular defect (such as a void). The line profile can also be useful in determining the amount of trend present in an image.

Simple measurement functions that are often used are: • • • •

Determination of the pixel value at a particular point (such as on a defect) Conversion of the pixel value to film density or part density The distance from one pixel to another in terms of pixel units or engineering units The angle of one line with respect to another

It is very desirable to be able to measure both the perimeter and the area of a defect. In many cases, such as a part with many voids, it may be desirable to obtain the total area of all defects above (or below) a threshold and their radius in terms of pixel values or engineering units and then to produce a histogram of the defect radii. This type of quantitative analysis is useful for material evaluation as a function of processing variables. Fast Fourier Transform. The FFT algorithm is central to many filtering and information extraction schemes. The two-

dimensional FFT is analogous to the one-dimensional FFT in that the function extracts frequency-dependent information from the waveform or image. A discrete form of the two-dimensional FFT algorithm is:

(Eq 9)

Specific methods for calculating Eq 9 are given in Ref 6 and 17. The most direct uses of the FFT and inverse FFT are as filters to eliminate periodic noise from an image. The noise may be a periodic pattern, as in the case of the filament-wound vessel shown in Fig. 10(a), or it could be 60-cycle noise. The general filter procedure is shown in Fig. 10. The image is simply transformed, edited, and inverse transformed. In Fig. 10(a), the original radiograph shows a section of a filament-wound vessel with a cut in the windings. After the transform, the signal from the filament-wound structure is seen to radiate from the center at 45° (Fig. 10b). The filament-wound structure is removed from the image by removing the frequency components of the structure as shown in Fig. 10(c). As shown in Fig. 10(c) the zero frequency component (center pixel) is left intact to restore the average gray-level value. The image is then inverse transformed to recreate the original image minus the 45° filament pattern. Figure 10(d) shows the pattern after the inverse transform; the cut section is readily apparent.

Fig. 10 General procedure used to filter an image with the FFT. (a) Original image. (b) Image transformed to the frequency domain. (c) Image edited. (d) Image inverse transformed into the filtered image

A correlation function can be used directly to detect objects or features of an image that are of interest or to align images for combination. If two images that are generated from different modalities are to be combined and a registration mark is not present, the two images can be correlated over a region of interest and moved until the correlation coefficient peaks. This is best accomplished in images having edges or texture, such as those shown in Fig. 4 and 7. The edge detection filters described in the section "Image Operations" in this article are often used to detect or

outline images; in particular, the Sobel filter works well on high-definition or noisy edges. In radiographs having noisy, fuzzy, and low-contrast edges, the Laplacian, Roberts, Sobel, and simple convolution filters will usually not produce good results. Edge followers, Hough transforms, and other ad hoc algorithms that depend on knowledge of the rough shape of

the curve (template matching) to be outlined work satisfactorily in most cases. These algorithms, although of some use, are usually not part of a processing package because of their specialized applications. Advanced Functions. General formulations of image restoration functions such as motion restoration, deblurring, maximum entropy, and pattern recognition are not commonly found in general image-processing software packages, probably because of the computation required and the limited demand for these applications. These algorithms are discussed in Ref 3 (motion restoration), Ref 1 and 3 (deblurring), Ref 9 (maximum entropy), and Ref 1 and 10 (pattern recognition). The algorithms are of use in the restoration of flash radiographs or high-speed video as well as the correction of point-spread function in tomography or ultrasonics. Maximum entropy methods may be of use in the enhancement of NMR signals and other low signal-to-noise data, while pattern recognition can be of use in automating the detection of flawed, incomplete, or misassembled parts.

References cited in this section

1. 3. 6. 9.

W.K. Pratt, Digital Image Processing, John Wiley & Sons, 1978 R.C. Gonzalez and P. Wintz, Digital Image Processing, Addison-Wesley, 1977 H.K. Huang, Element of Digital Radiology: A Professional Handbook and Guide, Prentice-Hall, 1987 B.R. Frieden, Image Enhancement and Restoration, in Picture Processing and Digital Filtering, Vol 6, Topics in Applied Physics, T.S. Huang, Ed., Springer-Verlag, 1979, p 177 10. D.H. Janney and R.P. Kruger, Digital Image Analysis Applied to Industrial Nondestructive Evaluation and Automated Parts Assembly, Int. Adv. Nondestr. Test., Vol 6, 1979, p 39-93 17. C.S. Burrus and T.W. Parks, DFT/FFT and Convolution Algorithms: Theory and Implementation, John Wiley & Sons, 1985 Digital Image Enhancement T.N. Claytor and M.H. Jones, Los Alamos National Laboratory

Image Display A wide variety of devices and techniques are available for the display and output of the processed data. CRT Displays. The CRT display is the standard means of output for all imaging applications because the range of colors, brightness, and dynamic range cannot be matched by any other display technology. Black-and-white monitors can have resolutions of up to 2000 × 1500 pixels, bandwidths of the order of 200 MHz, and a brightness of 515 cd/m2 (150 ftL) at a dynamic range in excess of 256. Color monitors may have 1280 × 1024 pixels and bandwidths of 250 MHz. Highresolution color monitors have a brightness of 85 cd/m2 (25 ft-L) or more, with a dynamic range of 256 when viewed in dim light.

Recent advances in tube design have darkened the shadow mask by a factor of four to allow increased room illumination with no degradation in dynamic range. The only other major advance in CRT technology is the recent introduction of flatscreen tubes with resolutions up to 640 × 480 pixels. The standards for video output are the RS-170, which specifies 525 lines of interlaced video at a frame rate of 30 Hz, and RS-343-A, which specifies 1023 interlaced lines also at a 30-Hz frame rate. Most modern monitors have special circuitry that enables the monitor to sync to various standards with either the sync superimposed on the green signal or as a separate input. Most high-resolution (1280 × 1024) imaging systems use a 483 mm (19 in.) tube operating in a noninterlaced mode at 60 to 64 Hz for the primary display. Printers. Various color and black-and-white printers are available for hard copy. A simple method of hard copy for

color and black-and-white is to photograph the CRT screen or a small (80 mm, or 3

in.) very flat screen directly with

an 8 × 11 in. or smaller Polaroid-type camera. The disadvantages of this method are the cost of the instant film in the large format and the small size of the image in the 4 × 5 in. format. However, the color quality of the image is excellent. The two other types of color printers for small-system color output are ink-jet printers and thermal-transfer printers. The ink-jet printers are very inexpensive and produce vivid colors at 180 dpi on paper, but have poor contrast on transparency material. The thermal copiers are more expensive but produce high-resolution copies (300 dpi) by transferring color from a Mylar sheet onto paper or film with very good color and color density. The Polaroid cameras can reproduce all the colors seen on the screen, while the ink-jet and thermal-transfer printers will produce at least 162 colors at 150 or 90 dpi, with 2 × 2 pixel enlargement. Table 5 lists the possible color and black-and-white hard copy output devices suitable for printing images. Impact printers and plotters are not listed, because of the limited number of colors and poor resolution. Also not listed in Table 5 are the color printers based on the Xerox process. Table 5 Typical characteristics of color and black-and-white hard copy output devices Device

Resolution, dpi

Image quality

Paper

Film

Color

Camera

...

Excellent

Very good

Thermal transfer

300

Very good

Very good

Ink-jet

180

Good

Fair

Camera

...

Excellent

Very good

Laser film

300

Very good

Very good

Thermal paper

300

Good

...

Thermal transfer

300

Fair

Fair

Laser printer

300

Poor

Poor

Ink-jet

180

Poor

Very poor

Black and white

There are many more black-and-white printers available than color; however, the high-quality printers are restricted to the camera type and the laser film printer. The laser film printer writes with a modulated laser directly on a special dry silver paper or film that is then developed with heat. These printers are capable of producing 64 shades of gray on paper and 128 shades on film at 300 dpi with very dense blacks. The other devices listed in Table 5 have poor resolution or low contrast.

The other black-and-white printers produce poor plots because they are binary printers (a pixel may be black or white) at 300 dpi. To achieve a gray scale, the printer must use a 2 × 2, 3 × 3 or other super pixel size, and this decreases the effective resolution. If a 2 × 2 pixel is used, the printer can print only five shades of gray at 150 dpi; if 3 × 3 pixels are used (100 dpi), ten shades are possible. The printers can actually produce many more shades by programming the device driver for a super pixel dithering. In this case, the program looks at a large super pixel of 5 × 5, determines whether or not the pixel should be smaller, and adjusts the dot density with a dither so that, effectively, about 26 or more shades of gray are perceived with a resolution of about 100 dpi. Videotape and Videodisk. A common use of videotape is to input TV frames from a dynamic test to the imageprocessing system. For example, a high-speed video can be made of a pressure vessel as it bursts. Later, the sequential frames are captured and enhanced to determine the shape of the initiation crack or the speed of crack growth. Conversely, radiographs can be made periodically during the slow evolution of a part under the influence of some external variable (for example, void coalescence in ceramics as a function of temperature). The radiographs are then digitized and interpolated to form a time-compressed video.

Videodisks can be made in a write once read many format with relatively inexpensive machines. Videodisks offer better resolution and more convenient frame access than videotape. Another alternative for the display of dynamic data is to use the hard disk in the imaging workstation. Hard disk transfer rates (300 kbytes/s) allow small (200 × 200) movies to be shown on a workstation screen for tens of seconds (depends on disk size) at seven frames per second. Ultimately, it will be possible to expand compressed data quickly enough so that it can be read off a hard disk or a removable digital optical disk in near real time for significant durations. The processing of tape can be tedious because of the amount of data required for a few seconds of video and also because of the limited resolution. However, the techniques of video data display are a very powerful enhancement tool because of the attraction of the eye to changing features. Pseudo Three-Dimensional Images. All discussion of output data display has been in terms of a flat two-

dimensional representation. This is the display of preference for processing because the algorithms are available or easily coded. However, depending on the data and the method of acquisition, other types of displays may be more appropriate. Shown in Fig. 10(a) and (b) in the article "Use of Color for NDE" in this Volume are two ways to present twodimensional data in three dimensions. In Fig. 10(a), a two-dimensional ultrasonic microscopy image of a circuit board is shown with two defective metallic conductor pads (Ref 18). The image lines are drawn, as in the terrain-mapping method, with high values elevated with respect to low values and with a hidden line algorithm. Another way to present twodimensional image data is shown in Fig. 10(b). In this case, the two-dimensional ultrasonic data were taken from a C-scan of a filament-wound sphere with Teflon shims imbedded in the matrix (Ref 19). Presenting the two-dimensional data in this manner results in a realistic display and facilitates interpretation. Tomographic, ultrasonic, and NMR data sets may consist of many two-dimensional images that can be stacked to yield a picture of a true three-dimensional object. There are two ways to represent this three-dimensional data. In the first method, the data are modeled as a geometric solid or surface; in the second method, the data are displayed as raw data, or voxels (volume elements). The first method is used in solids-modeling workstations for design engineering. It can also be used to image NDE data, especially if the number of modeled elements is very large (105 or more) or has many surfaces. Figure 11 shows a three-dimensional reconstruction of a bundle of bent fuel rods after a reactor accident test. The image was constructed of 250 individual 128 × 128 pixel tomograms. A type of surface modeling that treats surfaces as polygons was used to model the rods. An inspection of the original tomograms shows that it is impossible to quickly grasp the shape of the bent rods from simple two-dimensional plots (Ref 20). Another three-dimensional reconstruction of this failed assembly can be found in Fig. 9 in the article "Use of Color for NDE" in this Volume.

Fig. 11 Three-dimensional perspective plot representation of a bundle of failed, bent fuel rods. The representation was reconstructed from three-dimensional data sets obtained from a series of 250 individual 128 × 128 pixel tomographic slices of the rod bundle.

Figure 8 in the article "Use of Color for NDE" shows a tomogram of a turbine blade that was displayed in terms of voxel imaging (also known as volume rendering), rather than modeling the surfaces by polygons or other shapes. Volume elements are created from the raw data, and these voxels are then projected to a view surface to produce an image. The raw data are not actually displayed. Figure 8 also demonstrates how the surfaces can be made transparent and the lighting model adjusted for a realistic effect. Graphics accelerators are available so that the images shown in Fig. 8 can be rotated, lighting sources adjusted, and textures added in almost real time (see Fig. 7 of the article "Use of Color for NDE" in this Volume). Three-dimensional image processing is in its infancy, but many medical and NDE applications are envisioned (Ref 21).

References cited in this section

18. J. Gieske, private communication, Sandia National Laboratories, 1988 19. W.D. Brosey, Ultrasonic Analysis of Spherical Composite Test Specimens, Compos. Sci. Technol., Vol 24, 1985, p 161-178; private communication, Sandia National Laboratories, 1988 20. C. Little, private communication, Sandia National Laboratories, 1988 21. A.R. Smith, Geometry and Imaging: Two Distinct Kinds of Graphics, to be published 1989; private communication Digital Image Enhancement T.N. Claytor and M.H. Jones, Los Alamos National Laboratory

References 1. 2. 3. 4. 5. 6.

W.K. Pratt, Digital Image Processing, John Wiley & Sons, 1978 M.H. Jacoby, Image Data Analysis, in Radiography & Radiation Testing, Vol 3, 2nd ed., Nondestructive Testing Handbook, American Society for Nondestructive Testing, 1985 R.C. Gonzalez and P. Wintz, Digital Image Processing, Addison-Wesley, 1977 G.A. Baxes, Digital Image Processing: A Practical Primer, Prentice-Hall, 1984 W. B. Green, Digital Image Processing: A Systems Approach, Van Nostrand Reinhold, 1983 H.K. Huang, Element of Digital Radiology: A Professional Handbook and Guide, Prentice-Hall, 1987

7. 8. 9.

A. Rosenfeld and A.A. Kak, Digital Picture Processing, Academic Press, 1982 K.R. Castleman, Digital Image Processing, Prentice-Hall, 1979 B.R. Frieden, Image Enhancement and Restoration, in Picture Processing and Digital Filtering, Vol 6, Topics in Applied Physics, T.S. Huang, Ed., Springer-Verlag, 1979, p 177 D.H. Janney and R.P. Kruger, Digital Image Analysis Applied to Industrial Nondestructive Evaluation and Automated Parts Assembly, Int. Adv. Nondestr. Test., Vol 6, 1979, p 39-93 J.J. Dongarra, "Performance of Various Computers Using Standard Linear Equations Software in a Fortran Environment," Technical Memorandum 23, Argonne National Laboratory, 1989 J.R. Janesick, T. Elliott, S. Collins, M.M. Blouke, and J. Freeman, Scientific Charge Coupled Devices, Opt. Eng., Vol 26 (No. 8), 1987, p 692-714 I.P. Csorba, Image Tubes, Howard W. Sams & Co., 1985 G.I. Yates, S.A. Jaramillo, V.H. Holmes, and J.P. Black, "Characterization of New FPS Vidicons for Scientific Imaging Applications," LA-11035-MS, US-37, Los Alamos National Laboratory, 1988 L.E. Rovich, Imaging Processes and Materials, Van Nostrand Reinhold, 1989 K. Thompson, private communication, Sandia National Laboratories, 1988 C.S. Burrus and T.W. Parks, DFT/FFT and Convolution Algorithms: Theory and Implementation, John Wiley & Sons, 1985 J. Gieske, private communication, Sandia National Laboratories, 1988 W.D. Brosey, Ultrasonic Analysis of Spherical Composite Test Specimens, Compos. Sci. Technol., Vol 24, 1985, p 161-178; private communication, Sandia National Laboratories, 1988 C. Little, private communication, Sandia National Laboratories, 1988 A.R. Smith, Geometry and Imaging: Two Distinct Kinds of Graphics, to be published 1989; private communication

10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.

Acoustic Microscopy Lawrence W. Kessler, Sonoscan, Inc.

Introduction ACOUSTIC MICROSCOPY is the general term applied to high-resolution, high-frequency ultrasonic inspection techniques that produce images of features beneath the surface of a sample. Because ultrasonic energy requires continuity of materials to propagate, internal defects such as voids, inclusions, delaminations, and cracks interfere with the transmission and/or reflection of ultrasound signals. Compared to conventional ultrasound imaging techniques, which operate in the 1 to 10 MHz frequency range, acoustic microscopes operate up to and beyond 1 GHz, where the wavelength is very short and the resolution correspondingly high. In the early stages of acoustic microscopy development, it was envisioned that the highest frequencies would dominate the applications. However, because of the high-attenuation properties of materials, the lower frequency range of 10 to 100 MHz is extensively used. Acoustic microscopy is recognized as a valuable tool for nondestructive inspection and materials characterization. Acoustic microscopy comprises three different methods: • • •

Scanning laser acoustic microscopy (SLAM), which was first discussed in the literature in 1970 (Ref 1) C-mode scanning acoustic microscopy (C-SAM), which is the improved version of the C-scan instrumentation (Ref 2) Scanning acoustic microscopy (SAM), which was first discussed in the literature in 1974 (Ref 3)

Each of these methods has a specific range of utility, and most often the methods are noncompetitive with regard to applications. That is, only one method will be best suited to a particular inspection problem.

Acoustic microscopes are practical tools that have emerged from the laboratory to find useful applications within industry. They can be applied to a broad range of problems that previously had no solutions, and they have been especially useful in solving problems with new high-technology materials and components not previously available. The three acoustic microscope types can be as different from each other as are microradiography and electron microscopy. This discussion should provide the potential user with an awareness of the techniques and their distinctions in order to maximize the opportunities for the successful use of acoustic microscopy.

References

1. A. Korpel, L.W. Kessler, and P.R. Palermo, Acoustic Microscope Operating at 100 MHz, Nature, Vol 232 (No. 5306), 1970, p 110-111 2. Product Bulletin, Sonoscan, Inc. 3. R.A. Lemons and C.F. Quate, Acoustic Microscope--Scanning Version, Appl. Phys. Lett., Vol 24 (No. 4), 1974, p 163-165 Acoustic Microscopy Lawrence W. Kessler, Sonoscan, Inc.

Fundamentals of Acoustic Microscopy Methods As a general comparison between the methods, the scanning laser acoustic microscope is primarily a transmission mode instrument that creates true real-time images of a sample throughout its entire thickness (reflection mode is sometimes employed). In operation, ultrasound is introduced to the bottom surface of the sample by a piezoelectric transducer, and the transmitted wave is detected on the top side by a rapidly scanning laser beam. The other two types of microscopes are primarily reflection mode instruments that use a transducer with an acoustic lens to focus the wave at or below the sample surface. The transducer is mechanically translated (scanned) across the sample in a raster fashion to create the image. The C-mode scanning acoustic microscope can image several millimeters or more into most samples and is ideal for analyzing at a specific depth. Because of a very large top surface reflection from the sample, this type of microscope is not effective in the zone immediately below the surface unless the Rayleigh-wave mode to scan near-surface regions is used along with wide-aperture transducers. The scanning acoustic microscope uses this Rayleigh wave mode and is designed for very high resolution images of the surface and near-surface regions of a sample. Penetration depth is intrinsically limited, however, to only one wavelength of sound because of the geometry of the lens. For example, at 1 GHz, the penetration limit is about 1 m (40 in.). The C-mode scanning acoustic microscope is designed for moderate penetration into a sample, and transmission mode imaging is sometimes employed. This instrument uses a pulse-echo transducer, and the specific depth of view can be electronically gated. More detailed discussions of each acoustic microscopy technique follow. SLAM Operating Principles A collimated plane wave of continuous-wave (CW) ultrasound at frequencies up to several hundred megahertz is produced by a piezoelectric transducer located beneath the sample, as illustrated in Fig. 1. Because this ultrasound cannot travel through air (making it an excellent tool for crack, void, and disbond detection), a fluid couplant is used to bring the ultrasound to the sample. Distilled water, spectrophotometric-grade alcohol, or other more inert fluids can be used, depending on user concerns for sample contamination. When the ultrasound travels through the sample, the wave is affected by the homogeneity of the material. Wherever there are anomalies, the ultrasound is differentially attenuated, and the resulting image reveals characteristic light and dark features, which correspond to the localized acoustic properties of the sample. Multiple views can be made to determine the specific depth of a defect, as is performed by stereoscopy (Ref 4).

Fig. 1 Simplified view of the methods for producing through transmission and noncollinear reflection mode acoustic images with the scanning laser acoustic microscope

A laser beam is used as an ultrasound detector by means of sensing the infinitesimal displacements (rippling) at the surface of the part created by the ultrasound. In typical samples that do not have polished, optically reflective surfaces, a mirrored plastic block, or coverslip, is placed in close proximity to the surface and is acoustically coupled with fluid. The laser is focused onto the bottom surface of the coverslip, which has an acoustic pattern that corresponds to the sample surface. By rapid sweeping of the laser beam, the scanning laser acoustic microscope images are produced in real time (that is, 30 pictures per second) and are displayed on a high-resolution video monitor. In contrast to other less accurate uses of the term real time in industry today, the scanning laser acoustic microscope can be used to observe events as they occur--for example, a crack propagating under an applied stress. The images produced by SLAM are shadowgraph mode images of structure throughout the thickness of the sample. This provides the distinct advantage of simultaneous viewing of the entire thickness of the sample, as in x-ray radiography. In situations where it is necessary to focus on one specific plane, holographic reconstruction of the SLAM data can be employed (Ref 5, 6). Figure 2 shows a system block diagram for the scanning laser acoustic microscope. In addition to an acoustic image on a CRT, an optical image is produced by means of the direct scanned laser illumination of the sample surface. For this mode, the reflective coating on the coverslip is made semi-transparent. The optical image serves as a reference view of the sample for the operator to consult for landmark information, artifacts, and positioning of the sample to known areas. The SLAM acoustic images also provide useful and easily interpreted quantitative data about the sample. For example, the brightness of the image corresponds to the acoustic transmission level. By removing the sample and restoring the image brightness level with a calibrated electrical attenuator placed between the transducer and its electrical driver, precise insertion loss data can be obtained (Ref 7, 8). With the acoustic interference mode, the velocity of sound can be measured in each area of the sample (Ref 9, 10). When these data are used to determine regionally localized acoustic attenuation loss, modulus of elasticity, and so on, the elastic microstructure can be well characterized.

Fig. 2 Schematic showing principal components of a scanning laser acoustic microscope. The unit employs a plane wave piezoelectric transducer to generate the ultrasound and a focused laser beam as a point source detector of the ultrasonic signal. Acoustic images are produced at a rate of 30 images per second.

The simplest geometries for SLAM imaging are flat plates or disks. However, with proper fixturing, complex shapes and large samples can also be accommodated. For example, tiny hybrid electronic components, large (254 × 254 mm, or 10 × 10 in.) metal plates, aircraft turbine blades and ceramic engine cylinder liner tubes have been routinely examined by SLAM. C-SAM Operating Principles The C-mode scanning acoustic microscope is primarily a pulse-echo (reflection-type) microscope that generates images by mechanically scanning a transducer in a raster pattern over the sample. A focused spot of ultrasound is generated by an acoustic lens assembly at frequencies typically ranging from 10 to 100 MHz. A schematic diagram is shown in Fig. 3.

Fig. 3 Schematic of the C-mode scanning acoustic microscope. This instrument incorporates a reflection, pulseecho technique that employs a focused transducer lens to generate and receive the ultrasound signals beneath the surface of the sample.

The ultrasound is brought to the sample by a coupling medium, usually water or an inert fluid. The angle of the rays from the lens is generally kept small so that the incident ultrasound does not exceed the critical angle of refraction between the fluid coupling and the solid sample. Note that the focal distance into the sample is shortened considerably by the liquidsolid refraction. The transducer alternately acts as sender and receiver, being electronically switched between the transmit and receive modes. A very short acoustic pulse enters the sample, and return echoes are produced at the sample surface and at specific interfaces within the part. The return times are a function of the distance from the interface to the transducer. An oscilloscope display of the echo pattern, known as an A-scan, clearly shows these levels and their timedistance relationships from the sample surface, as illustrated in Fig. 4. This provides a basis for investigating anomalies at specific levels within a part. An electronic gate selects information from a specific level to be imaged while it excludes all other echoes. The gated echo brightness modulates a CRT that is synchronized with the transducer position.

Fig. 4 Signal/source interfaces and A-scan displays for typical sample. (a) Simplified diagram showing pulses of ultrasound being reflected at three interfaces: front surface, A; material interface, B; and rear interface, C. Main bang is the background electrical noise produced by transducer excitation. (b) Plot of amplitude versus time as transducer is switched between transmit and receive modes at each interface during a time span of under 1 s. (c) Signal at interface B gated for a duration of 30 ns

Compared to older conventional C-scan instruments, which produce a black/white output on thermal paper when a signal exceeds an operator-selected threshold, the output of the C-mode scanning acoustic microscope is displayed in full gray scale, in which the gray level is proportional to the amplitude of the interface signal (gray-scale digitization is discussed in the article "Machine Vision and Robotic Evaluation" in this Volume). The gray scale can be converted into false color, and as shown in Fig. 5, the images can also be color coded with echo polarity information (Ref 11). That is, positive echoes, which arise from reflection from a higher-impedance interface, are displayed in a gray scale having one color scheme, while negative echoes, from reflections off of lower-impedance interfaces are displayed in a different color scheme. This allows quantitative determination of the nature of the interface within the sample. For example, the echo amplitude from a plastic-ceramic boundary is very similar to that from a plastic-airgap boundary, except that the echoes are 180° out of phase. Thus, to determine whether or not an epoxy is bonded to a ceramic, echo amplitude analysis alone is not sufficient. The color-coded enhanced C-mode scanning acoustic microscope is further differentiated from conventional C-scan equipment by the speed of the scan. Here, the transducer is positioned by a very fast mechanical

scanner that produces images in tens of seconds instead of tens of minutes for typical scan areas to cover the size of an integrated circuit.

Fig. 5 Schematic and block diagram of the C-mode scanning acoustic microscope. The instrument employs a very high speed mechanical scanner and an acoustic impedance polarity detector to produce high-resolution Cscan images.

With regard to the depth zone within a sample that is accessible by C-SAM techniques, it is well known that the large echo from a liquid/solid interface (the top surface of the sample) masks the small echoes that may occur near the surface within the solid material. This characteristic is known as the dead zone, and its size is usually of the order of a few wavelengths of sound or more. Far below the surface, the maximum depth of penetration is determined by the attenuation losses in the sample and by the geometric refraction of the acoustic rays, which shorten the lens focus by the solid material. Therefore, depending on the depth of interest within a sample, a proper transducer and lens must be used for optimum results. SAM Operating Principles The scanning acoustic microscope is primarily a reflection-type microscope that generates very high resolution images of surface and near-surface features of a sample by mechanically scanning a transducer in a raster pattern over the sample (Fig. 6 and 7). In the normal mode, an image is generated from echo amplitude data over an x,y scanned field-of-view. As with SLAM, a transmission interference mode can be configured for velocity of sound measurements. In contrast to CSAM, a more highly focused spot of ultrasound is generated by a very wide angle acoustic lens assembly at frequencies typically ranging from 100 to 2000 MHz. The angle of the sound rays is well beyond the critical cutoff angle, so that there is essentially no wave propagation into the material. There is a Rayleigh (surface) wave at the interface and an evanescent wave that reaches to about one wavelength depth below the surface. As in the other techniques, the ultrasound is brought to the sample by a coupling medium, usually water or an inert fluid. The transducer alternately acts as sender and receiver, being electronically switched between the transmit and receive modes. However, instead of a short pulse of acoustic energy, a long pulse of gated radio-frequency (RF) energy is used. No range gating is possible, as in C-SAM, because of the basic design concept of the SAM system. The returned acoustic signal level is determined by the elastic properties of the material at the near-surface zone. The returned signal level modulates a CRT, which is synchronized with the transducer position. In this way, images are produced in a raster scan on the CRT. As with C-SAM, complete images are produced in about 10 s.

Fig. 6 Schematic of the scanning acoustic microscope lens used for interrogating the surface zone of a sample

Fig. 7 Block diagram of a reflective-type SAM system that uses mechanical scanning of a highly focused transducer to investigate the surface zone of a sample at high magnification

With the SAM technique operating at very high frequencies, it is possible to achieve resolution approaching that of a conventional optical microscope. This technique is employed in much the same way as an optical microscope, with the important exception that the information obtained relates to the elastic properties of the material. Even higher resolution than an optical microscope can be obtained by lowering the temperature of operation to near 0 K and using liquid helium

as a coupling fluid. The wavelength in the liquid helium is very short compared to that of water, and submicron resolution can be obtained (Ref 12). The SAM technique has also been found to be useful for characterizing the elastic properties of a sample over a microscopic-size area, which is determined by the focal-spot size of the transducer. In this method, the reflected signal level is plotted as a function of the distance between the sample and the lens. Because of the leaky surface waves generated by mode conversion at the liquid/solid interface as the sample is defocused toward the lens, an interference signal is produced between the mode converted waves and the direct interface reflection. The curve obtained is known as the V(z) curve; by analyzing the periodicity of the curve, the surface wave velocity can be determined. Furthermore, defocusing the lens enhances the contrast of surface features that do not otherwise appear in the acoustic image (Ref 13, 14). In addition, by using a cylindrical lens instead of a spherical lens, the anisotropy of materials can be uniquely characterized (Ref 15). Comparison of Methods to Optimize Use of Techniques Figure 8 illustrates the zones of application for all three types of acoustic microscopy techniques. The differences are substantial with regard to the potential for visualizing features within a sample, and they should be carefully weighed before a particular method is selected. Table 1 lists some of the major advantages and benefits of the techniques. However, generalizations are sometimes difficult to make. Proposed or laboratory-demonstrated solutions to some of the limitations of each technique may already exist. Table 1 was prepared on the basis of instrumentation and techniques that are commercially available and therefore should be used only as a general guide. Table 1 Comparison of industrial acoustic microscopy techniques Parameter

SLAM

C-SAM

SAM

General description

Utilizes CW, plane wave ultrasonic illumination of sample and scanning focused laser beam detection of ultrasound; simultaneous optical images and acoustic images are produced. SLAM produces images in real time, which is the fastest of all acoustic microscopy techniques.

High-resolution focused-beam C-scan(a). Utilizes pulse-echo mode and has full gray-scale image output. Images are produced by mechanically scanning the transducer over the sample area.

Utilizes very highly focused acoustic lenses to image exterior surfaces by means of surface acoustic wave generation. Images are produced by mechanically scanning the transducer over the sample area. SAM produces the highest resolution of all acoustic microscopy techniques.

Primary use

Nondestructive testing and materials characterization

Nondestructive testing and materials characterization

Materials characterization and research

Frequency range

10-500 MHz

10-100 MHz

100 MHz-2 GHz

Resolution

Resolution limited to the wavelength of ultrasound within the sample material

Resolution limited to the wavelength of ultrasound within the sample material, multiplied by a factor, typically 2-10, due to the lens design

Resolution limited to the wavelength of ultrasound within the coupling fluid at the surface of the sample

Imaging mode

Through transmission (primarily); off-axis reflection; orthoscopic view of sample

Reflection (primarily); orthoscopic view of sample

Reflection (only); orthoscopic view of sample

Image information

Image produced is an acoustic shadowgraph representing the entire thickness of the sample.

Image is produced at any selected depth within the sample, but is affected by wave propagation through all the levels prior to the focus location.

Image is produced only at the sample surface, and the depth of the information extends into the material a distance of one

wavelength of sound.

Ultrasound mode

Continuous-wave and frequency-modulated ultrasound waves

Short pulses of ultrasound, less than a few cycles of RF in duration

Gated RF (or tone burst) containing many cycles (10-100) of RF at frequency selected

Ultrasound signal timing

Not applicable due to CW

Echoes are spread out in time, and one is selected and electronically gated for desired image depth within sample.

Gated at the surface of the sample where virtually all the ultrasound is reflected. There are no subsurface echoes to select.

Transducer type

Plane wave transducer for sample illumination and focused laser beam for detection

Focused transducer used for transmit and receive mode. The angle of the rays is less than the critical angles between the coupling fluid and the sample. This is known as a low-numerical-aperture transducer.

Focused transducer in which the geometric angle of the rays is much greater than the critical angles in order to excite surface modes. This is known as a high-numericalaperture transducer.

Image outputs

Amplitude mode: Records level of ultrasound transmission at each x,y coordinate of the scan Interference mode: Records velocity of sound variations within sample over the field-ofview at each frequency Frequency scan mode: Similar to amplitude mode except that the insonification frequency is swept over a range to eliminate speckle and other artifacts of coherent imaging Optical mode: Laser scanned optical image produced in synchrony with the acoustic images

C-mode: Records echo amplitudes at each x,y coordinate of the scan; image output on CRT A-scan: Oscilloscope display of the echo pattern as a function of time (distance into sample) at each x,y coordinate. This can be used to measure the depth of a feature or the velocity of sound in the material. The A-scan is used to determine the setting of the echo gate, which is critical to the composition and interpretation of the images.

Normal mode: Records reflected energy from surface at each x,y position. Image output on CRT. V(z), or acoustic material signature graph, records the change in reflected signal level at any x,y coordinate as the z position is varied (Ref 12, 13, 14). This will characterize the material by means of its surface acoustic wave velocity. Interference mode: Records velocity of sound variations within sample over the field-of-view at each frequency

Depth of penetration

Penetration is limited by acoustic attenuation characteristic of sample.

In addition to attenuation by the sample, penetration is limited by focal length of lens and geometric refraction of the rays, which causes shortening of the focus position below the surface. There is also a dead zone just under the surface due to the large-amplitude front-surface echo, which masks smaller signals occurring immediately thereafter. This can be rectified with a high-numerical-aperture transducer.

Limited to a distance of one wavelength of sound below the surface. There is essentially no wave propagation into the sample.

Imaging speed

True real-time imaging: 30 frames/s; fastest of all acoustic microscopes

10 s to 30 min per frame; varies greatly among manufacturers

10 to 20 s/frame

New developments

Holographic reconstruction of each plane through the depth of the sample (Ref 5, 6)

Acoustic impedance polarity detector to characterize the physical properties of the echo producing interfaces (Ref 11)

Low-temperature liquid helium stages for extremely high resolution images (Ref 15)

(a) C-scan produces ultrasonic images by mechanically translating a pulsed transducer in an x,y plane above a sample while recording the echo amplitude within a preset electronic time gate. The transducer may be planar or focused. The frequencies of operation are typically 1-10 MHz, and the data are usually displayed as a binary brightness level on a hard copy output unit, such as thermal paper. A threshold level selected by the operator determines the transition between what amplitude of echo is displayed as a bright or dark image print.

Fig. 8 Simplified comparison of three acoustic microscopy techniques, particularly their zones of application (crosshatched area) within a sample. (a) SLAM. (b) SAM. (c) C-SAM

References cited in this section

4. L.W. Kessler and D.E. Yuhas, Acoustic Microscopy--1979, Proc. of IEEE, Vol 67 (No. 4), April 1979, p 526-536 5. Z.C. Lin, H. Lee, G. Wade, M.G. Oravecz, and L.W. Kessler, Holographic Image Reconstruction in Scanning Laser Acoustic Microscopy, Trans. IEEE, Vol UFFC-34 (No. 3), May 1987, p 293-300 6. B.Y. Yu, M.G. Oravecz, and L.W. Kessler, "Multimedia Holographic Image Reconstruction in a Scanning Laser Acoustic Microscope," Paper presented at the 16th International Symposium on Acoustical Imaging (Chicago, IL), Sonoscan, Inc., June 1987; L.W. Kessler, Ed., Acoustical Imaging, Vol 16, Plenum Press, 1988, p 535-542 7. L.W. Kessler, VHF Ultrasonic Attenuation in Mammalian Tissue, Acoust. Soc. Am., Vol 53 (No. 6), 1973, p 1759-1760 8. M.G. Oravecz, Quantitative Scanning Laser Acoustic Microscopy: Attenuation, J. Phys. (Orsay), Vol 46, Conf. C10, Supplement 12, Dec 1985, p 751-754 9. S.A. Goss and W.D. O'Brien, Jr., Direct Ultrasonic Velocity Measurements of Mammalian Collagen Threads, Acoust. Soc. Am., Vol 65 (No. 2), 1979, p 507-511 10. M.G. Oravecz and S. Lees, Acoustic Spectral Interferometry: A New Method for Sonic Velocity Determination, in Acoustical Imaging, Vol 13, M. Kaveh, R.K. Mueller, and T.F. Greenleaf, Ed., Plenum Press, 1984, p 397-408 11. F.J. Cichanski, Method and System for Dual Phase Scanning Acoustic Microscopy, Patent Pending 12. J.S. Foster and D. Rugar, Low-Temperature Acoustic Microscopy, Trans. IEEE, Vol SU-32, 1985, p 139-

151 13. K.K. Liang, G.S. Kino, and B.T. Khuri-Yakub, Material Characterization by the Inversion of V(z), Trans. IEEE, Vol SU-32, 1985, p 213-224 14. R.D. Weglein, Acoustic Micro-Metrology, Trans. IEEE, Vol SU-32, 1985, p 225-234 15. J. Kushibiki and N. Chubachi, Material Characterization by Line-Focus-Beam Acoustic Microscope, Trans. IEEE, Vol SU-32, 1985, p 189-212 Acoustic Microscopy Lawrence W. Kessler, Sonoscan, Inc.

Acoustic Microscopy Applications* It is difficult to document concisely the very broad range of applications for acoustic microscopy, but a few generalizations can be made that follow the examples of conventional ultrasonic nondestructive testing (see the articles "Ultrasonic Inspection" and "Adhesive-Bonded Joints" in this Volume). Acoustic microscopy is compatible with most metals, ceramics, glasses, polymers, and composites (made from combinations of the above materials). The compatibility of a material is ultimately limited by ultrasound attenuation caused by scattering, absorption, or internal reflection. In metals, the grain structure causes scattering losses; and in ceramics, the porosity may cause losses. The magnitude of these effects generally increases with ultrasound frequency, and the dependence is monotonic. Figure 9 shows a general guide to acoustic microscopy applications with respect to imaging. The quantitative aspects have not yet found widespread industrial acceptance, although they are extremely important and form the basis for materials characterization. The techniques of SLAM, C-SAM, and SAM all produce quantitative data in addition to images. Acoustic microscopy methods are compared below with a typical C-scan ultrasound method in terms of frequency employed:

Method

Frequency range, MHz

C-scan ultrasound

1-10

Acoustic microscopy

C-SAM

10-100

SLAM

10-500

SAM

100-2000

Fig. 9 Comparison of acoustic microscopy applications with C-scan applications, based on transducer frequency and wavelength.

The most popular application of SLAM and C-SAM is the nondestructive evaluation of bonding, delamination, and cracks in materials. These instruments are often used for process and quality control, although a significant percentage of the devices are placed in analytical and failure analysis laboratories. The most popular application of SAM utilizes its very high magnification mode and is employed as a counterpart to conventional optical and electron microscopy, that is, to see fine detail at and near surfaces. The scanning acoustic microscope, like the other acoustic microscopy methods, produces image contrast, which is a function of the elastic properties of a material, where other nonacoustic techniques may not. Composite Materials Composite materials represent an exciting challenge for the materials scientist and for nondestructive testing. With the combination of materials having different properties and with manufactured anisotropy, acoustic microscopy can clearly define the relevant property distribution of materials at the microscopic level. The examples shown in this section are polymer materials reinforced with fibers. Metal-matrix composites and ceramic-matrix composites can be similarly studied. In general, the combinations of materials having different mechanical properties result in interfaces that cause scattering of ultrasound and differential attenuation within the field-of-view. This can be important in determining the population density shifts of fibers within a sample. SLAM Images. Acoustic microscopy can be used to differentiate fibers that are bonded well to the matrix from fibers that are separated from the matrix. Excessive stress, for example, would cause such a separation. As an example of this, Fig. 10 shows two nominally identical tensile-test bars made of carbon fiber reinforced plastic (CFRP). One of the bars has been stressed to a level that produced a 1.5% strain. The other sample has not been stressed. Figures 11(a) and 11(b) compare SLAM images of these samples at 100 MHz. The lighter image, Fig. 11(a), shows texturing that follows the fiber direction. Figure 11(b) shows the much darker, attenuating characteristics of the stressed sample. The excess attenuation is due to the separation of the fiber from the matrix.

Fig. 10 Photograph of two CFRP tensile-test bars--one unstressed and one stressed. The stressed bar was subjected to a force that yielded a 1.5% strain.

Fig. 11 100-MHz SLAM acoustic micrographs of the tensile bar samples shown in Fig. 10. (a) In the unstressed bar, the texture corresponds to fiber population shifts and small areas of disbond. (b) In the stressed bar that was pulled to a 1.5% strain level, there is no visible evidence of fracture. Field of view: both 3 × 2.25 mm. However, the increased acoustic attenuation (dark zones) present in (b) indicates stress damage to the

material.

Figure 11 in the article "Use of Color for NDE" in this Volume shows a 30-MHz transmission SLAM image of a Kevlar fiber-reinforced plastic that has been cut perpendicular to the fibers, thus producing extensive delamination (red) and cracks along the axis of the sample. Defects anywhere throughout the thickness will block the ultrasound transmission. Figure 12 in this article shows a curved CFRP component. This sample is more difficult to image than the tensile-test bars because it has a complex shape; in addition, it has fibers oriented in a variety of different directions for directional strength.

Fig. 12 Complex-shaped CFRP component. The fibers are arranged to impart directional strength to certain critical areas of the sample.

In Fig. 13, a 30-MHz SLAM transmission image, a transition is seen between different fiber orientations in the sample; Fig. 14 in the article "Use of Color for NDE" in this Volume shows a large anomaly that is located in the middle of the curved portion. Scanning laser acoustic microscopy can be used to image complex-shape samples even though the curvature may cause some restriction of the field-of-view due to critical angle effects and lenslike action by the sample.

Fig. 13 30-MHz SLAM acoustic micrograph of the sample shown in Fig. 12. The left portion of the micrograph shows a more complex fiber network than the area on the right. A color image of the curved CFRP component is

shown in Fig. 14 in the article "Use of Color for NDE" in this Volume. Field of view: 14 × 10.5 mm

C-SAM Image. Figure 14 shows a 50-MHz reflection image in the C-SAM image of a CFRP tensile-test bar similar to

that shown in Fig. 10. The acoustic lens is focused near the surface of the sample, and the electronic gate was set to receive a portion of the backscattered signal. Because the sample is constructed with fibers distributed throughout the volume, acoustic energy is scattered from all depths, and distinct, time-separated echoes were not generated. When the transducer was focused deep into the sample, there was less definition of the fibers in the acoustic image, similar to what is seen in the SLAM image.

Fig. 14 50-MHz C-SAM reflection mode micrograph of a CFRP test sample. The ultrasound was focused near the top surface of the sample. Field of view: 19 × 14 mm

Ceramic Materials Ceramic materials are used in a variety of applications. Some are used in the electronics industry as substrates for delicate hybrid circuits. In structural applications, ceramics are used where high temperature and light weight are important. Silicon nitride, silicon carbide, and zirconia are receiving much attention for future engine applications. However, because of the inherent brittleness of ceramic materials, small defects are very critical to the structural integrity of the materials. The successful use of these materials necessitates careful nondestructive screening of the samples. Aluminum Oxide. Figure 15 shows an optical picture of an aluminum oxide panel with laser-machined holes. This

sample is an electronics-grade material that is 99.5% pure and has very low porosity. When the ceramic powder is first compressed in the green state, opportunities arise for segregated low-density areas to occur. After the sintering operation, these areas are usually found to be very porous. If one of these areas happens to coincide with the site of laser machining, the stresses that occur can cause the material to crack. Upon visual examination, it may be very difficult to detect fine cracks, even if they come to the surface. Dye penetrants can be used to increase the visual contrast. Unfortunately, in many applications, dyes are considered to be contaminants and therefore cannot be used for nondestructive testing. Because of the discontinuity in material property, a crack will reflect acoustic waves and produce high-contrast images.

Fig. 15 Aluminum oxide panel having laser-machined holes and slots with fine cracks around the perimeter and the circumference of the openings

Figure 16 shows a 100-MHz SLAM acoustic micrograph of a ceramic sample with a crack that can be seen as a dark line originating from one of the holes. Surrounding the area of the crack are several small, dark patches, which arise from localized increases in porosity. In the acoustic image, the pores may not be visible individually if they are smaller than the wavelength of sound. In this case, the pores are only a few microns in size, and the wavelength is about 25 m (0.001 in.). The porous areas are detectable by virtue of excess ultrasound scattering, which causes the differential attenuation. Correlative analyses have shown that pores in the 1 m (40 in.) size range cause detectable attenuation increases (Fig. 16). However, the presence of a single 1 m (40 in.) pore would be difficult to detect unless the frequency of the ultrasound was increased to 1 GHz or more.

Fig. 16 100-MHz SLAM acoustic micrograph of an alumina panel section that contains a hole showing a crack with very high contrast. A digitally enhanced image of this alumina panel is also shown in Fig. 15 in the article "Use of Color for NDE" in this Volume. Field of view: 3 × 2.25 mm

It is significant to note that attenuation changes in ceramics correlate well with variations in strength (Ref 16). Fig. 15 in the article "Use of Color for NDE" in this Volume shows a digitally enhanced pseudocolor image of a ceramic sample that is useful for quantitative analysis of the gray scale and for identifying precisely the acoustic signal levels without relying on operator interpretation of the gray scale on the CRT screen. Example 1: Use of C-SAM to Detect Internal Porosity Defects in an Alumina Ceramic Disk. Acoustic microscopy is a powerful tool for nondestructively evaluating ceramic materials and displaying internal defects and density gradients such as porosity. An acoustic microscope scanning at 50 MHz with an f/1 lens on the transducer was used to evaluate a 6.4 mm (0.25 in.) thick aluminum oxide ceramic disk. No surface preparation of the sample was required prior to scanning. The acoustic microscope, with a 25 m (0.001 in.) acoustic resolution, scanned a 12.7 × 12.7 mm (0.50 × 0.50 in.) area of the ceramic disk. A pulse-echo technique (C-SAM) displayed internal reflections. The black areas (arrows, Fig. 17), which were detected as having the highest-amplitude reflection, indicate areas of porosity of the order of 10 to 20 m (400 to 800 in.) in diameter.

Fig. 17 A portion of the 160 mm2 (0.25 in.2) area of an alumina ceramic disk scanned by an acoustic microscope showing the presence of porosity in the disk (arrows). Courtesy of G.H. Thomas, Sandia National Laboratories

Example 2: Analysis of Alumina Ceramic Disks Supplied by a Variety of Manufacturers and Subjected to SLAM, C-SAM, and SAM Evaluation. Disks of alumina were obtained from various manufacturers as part of a study on the effects of sintering temperature changes on mechanical strength and ultrasonic properties. The disks were approximately 25 mm (1 in.) diameter and 6 in.) thick. Figure 18 shows a through-transmission 10-MHz SLAM image of an alumina disk. Higher SLAM mm ( frequencies could not be used, because of excessive attenuation. This is unusual for ceramic materials unless they are very porous or otherwise defective. Most of the area of this sample is nontransmissive (dark), which indicates a large internal defect. A subsequent cross section of this part, shown in Fig. 19, reveals a large crack that correlated to the indication in the SLAM image. The crack is not parallel to the surface; this was not known prior to cross sectioning.

Fig. 18 Low-frequency 10-MHz SLAM image of a 25 mm (1 in.) diam alumina test disk. The disk is very attenuating to ultrasound because of internal defects that cover about 75% of the area (dark zones). Field of view: 35 × 26 mm

Fig. 19 Cross sectioning of the alumina test disk shown in Fig. 18 reveals a large crack that correlates with the area of low acoustic transmission. This confirms the presence of an internal defect.

Figures 16(a) and 16(b) in the article "Use of Color for NDE" in this Volume are 15-MHz reflection C-SAM images of the same part focused to different depths. The radical changes in echo pattern at different depth locations in the disk result in widely different images. This is due to the nonparallel nature of the crack. In these figures, the darkest features correspond to the greatest echo amplitude. The C-mode scanning acoustic microscope clearly shows anomalies in this part, and by sequential scanning at a variety of depths, an overall diagnosis of the part can be assembled. Figure 20 shows a SAM image produced at 180 MHz in which the lens was focused slightly below the surface (20 m, or 800 in., in water) to highlight subsurface information. The predominant features in Fig. 20 are surface scratches, which may also be associated with near-surface cracks. The grain structure and porosity of this sample are finer than can be visualized at 180 MHz.

Fig. 20 SAM image at 180 MHz of the sample shown in Fig. 18 and 19 revealing features of the ceramic at and near the surface. Color images of this disk are shown in Figures 16(a) and 16(b) in the article "Use of Color for NDE" in this Volume. Field of view: 1 × 1 mm

Another alumina sample produced under different conditions was found to be much more transparent acoustically. Figure 21 shows a 30-MHz SLAM image of a typical area of this disk. The higher frequencies are associated with higher magnifications and therefore smaller fields-of-view. The ceramic appears fairly homogeneous at this level of magnification. Figure 22, a 100-MHz SLAM image of the sample in Fig. 21, shows that some textural inhomogeneities are beginning to appear. The texture may be due to nonuniform segregation of porosity in the sample and throughout its thickness. Figure 23, a 50-MHz C-SAM image of the disk, shows large pores (white) about 1 mm (0.04 in.) below the surface. Figure 24 shows a 180-MHz SAM image made under conditions identical to Fig. 20. The fine texture at the surface of this sample is due to porosity. The differences between the two samples are clearly evident acoustically.

Fig. 21 SLAM images at 30 MHz of an alumina test disk similar in size to that shown in Fig. 18. This sample was quite transparent to the ultrasound, as evidenced by the bright, relatively uniform appearance of the acoustic image. Field of view: 14 × 10.5 mm

Fig. 22 SLAM 100-MHz transmission acoustic image of the sample shown in Fig. 21. At this higher frequency, textured variations in the sample are evident that may be due to micro (subresolution) porosity segregations, which cause differential absorption. Field of view: 3.5 × 2.6 mm

Fig. 23 C-SAM reflection mode image at 50 MHz made by setting the gate and focus to about 1 mm (0.04 in.) below the surface. The white circular spots correspond to individual pores located at this depth. Field of view: 30 × 30 mm

Fig. 24 SAM surface mode image at 180 MHz made under conditions identical to Fig. 20. In this image, the fine texture corresponds to porosity of the sample. Note the sharp contrast between the microstructure of this sample and that of Fig. 20. Field of view: 1 × 1 mm

Metals In typical optical microscopy and metallography, it is necessary to polish and etch a sample to reveal the microstructural pattern. With SAM, this may not be necessary, as evidenced by the above. As one further illustration of SAM, Fig. 25 shows a 400-MHz image, focused 23 m (920 in.) below the surface, of a manganese-zinc ferrite material. The different phases of this important material can be visualized without the use of etchants. For certain metallic samples, that is, those with smooth, polished surfaces, SAM can be used for nondestructive testing as well as for metallographic analysis.

Fig. 25 SAM surface mode image at 400 MHz of a manganese-zinc ferrite sample that was polished metallurgically but not chemically etched. The elastic property differences between the various phases of this material are responsible for the contrast shown in this image. Courtesy of Honda Electronics Company, Ltd.

Figure 26 shows the metallographic structure of low-carbon steel; the specimen was polished but not etched. Although the images are similar to optical displays, the acoustic microscope is sensitive to acoustic properties and will generate images of surface and subsurface structure. The low-carbon steel was scanned at 1.3 GHz over an area measuring 100 × 100 m (0.004 × 0.004 in.). A high-resolution scanning acoustic microscope was used to generate this image from variations in the acoustic properties of the steel sample. Details of the grain structure, grain boundaries, and impurities in the specimen are visible in Fig. 26. The darker areas indicate contaminants trapped in the metal.

Fig. 26 SAM image at 1.3 GHz of a polished but unetched manganeze-zinc low-carbon steel. Darker regions indicate the presence of typical contaminants trapped in the metal. Courtesy of G.H. Thomas, Sandia National Laboratories

Microelectronic Components In the field of microelectronics, reliability is affected by tiny defects that may be harbored within a component or assembly but that are not detectable by electrical testing. The defects may grow with time or with thermal cycling, and when they reach a critical size, electrical performance may be affected; that is, a connection may break, or an electrical characteristic may change. Integrated circuits, ceramic capacitors, and thyristors that contain defects of this type will be examined in this section. Integrated Circuits (ICs). A typical silicon IC chip (die) generates heat during its normal operation. The heat must be dissipated to stabilize the electrical behavior of the semiconductor. Therefore, the die is bonded onto a thermally conductive substrate, such as copper, aluminum oxide, or beryllium oxide. A high-reliability IC must also be hermetically sealed from ionic contamination and moisture in the atmosphere; therefore, a sealed ceramic or ceramic/metal package is used to house the die and to serve as an interconnection vehicle. Figure 27 shows a typical ceramic IC package containing a die bonded to a gold-plated surface. The die bond is also referred to as a die-attach.

Fig. 27 Typical ceramic-packaged IC showing the silicon die, which is bonded to the metallized surface

To complete the component, tiny interconnection wires are attached from the die to the corresponding metallization sites (bond pads) inside the ceramic package. A metal cover (lid) is then soldered to the top of the package. Apart from defects in the silicon chip itself, the life of the component depends on the quality of the die-attach, the integrity of the interconnection wire bonds, and the lid seal, all of which require good bonding. Conventional Radiography Limitations. Because of its wide acceptance, conventional or x-ray radiography has usually been the first method of investigation for virtually all internal examinations (see the article "Radiographic Inspection" in this Volume). However, it has finally been realized that x-ray techniques indicate only the presence, absence, and density of materials through which the beam passes. On the positive side, if a poor bond is caused by the absence of a material, for example, solder, there will be less x-ray absorption and therefore a correct indication of a defect. However, if all the materials are present and in their proper proportions, there can be no differential x-ray absorption and no defect identification.

For example, suppose that the die is uniformly coated with solder but that there is no adhesion of the solder to the substrate due to contamination of the surface or improper metallurgy. Suppose also that the solder bonds well to the die and to the metallic layer but that for some reason the metallic layer becomes detached from the substrate. Both of these conditions are common problems that cannot be detected by x-ray but, because of the thin interface air gaps, cause complete reflection of ultrasound from, and no transmission of ultrasound across, the interface (Ref 17). The U.S. military standards for the inspection of circuits now include acoustic microscopy inspection (Ref 18). Another common situation involves a die-attach material, such as epoxy, that is not very absorptive to x-rays, as is the typical lead tin solder alloy. If the die is bonded to a ceramic substrate that is also not very absorptive, low-energy x-rays can be used to detect voids (but not delaminations) in the die-attach material. However, if the die is attached to a metal substrate that absorbs x-rays, much higher energy is required for penetration. In this case, the die-attach material absorption becomes a negligible fraction of the total absorption signal, and neither disbonds nor voids can be detected. Advantages of Acoustic Microscopy. The die-attach can be inspected by SLAM techniques in those cases in which

ultrasound can be coupled to the top and bottom surfaces of the IC, that is, when there is no lid enclosing the package. When the lid is on, the air gap within the die cavity precludes the transmission of ultrasound, and a reflection method, such as C-SAM must be used. This is not true for plastic packages, for which SLAM and C-SAM can be used. Figure 28 illustrates the SLAM inspection of an IC. Figure 29 shows a transmission SLAM image of a die-attach. The dark zones represent little or no ultrasound transmission. A white line generated by the SLAM image analysis system outlines the die region.

Fig. 28 Schematic illustrating use of the SLAM through transmission technique to evaluate the die-attach bond between the silicon die and the ceramic package

Fig. 29 30-MHz transmission SLAM image of a die-attach. The dark areas of the image correspond to little or no acoustic wave transmission due to the lack of bonding between the die and the ceramic. The white rectangle outlining the die is generated by an acoustic image analysis computer, which is used to determine the percent area of disbond. A 50-MHz reflection mode C-SAM color image of this integrated circuit is shown in Fig. 13 in the article "Use of Color for NDE" in this Volume. Field of view: 14 × 10.5 mm

Figure 30 shows the SLAM data after digital processing to highlight the areas of bond (white) and disbond (black). Because of its real-time imaging speed, the scanning laser acoustic microscope would be useful for quality control inspection of the die-attach before a ceramic IC is completely assembled.

Fig. 30 Computer-analyzed SLAM acoustic image of Fig. 29 in which the disbonds are clearly displayed as

black. A 50-MHz reflection mode C-SAM color image of this IC is shown in Fig. 13 in the article "Use of Color for NDE" in this Volume.

Figure 31 illustrates the inspection of an IC within a closed, hermetic package by means of C-SAM. In this case, the ultrasound accesses the die-attach interface through the ceramic substrate. Access through the lid is precluded by the air gap. Because of the differences in acoustic impedance, Z, between the ceramic and silicon materials, an echo will be returned from the die-attach interface in the case of a good bond or a bad bond, although the amplitudes of the echoes may be different. The acoustic impedances of a few typical materials are listed below:

Material

Acoustic impedance, 106 rayl

Air, vacuum

0

Water

1.5

Plastic

2.0-3.5

Glass

15

Aluminum

17

Silicon

20

Beryllia

32

Copper

42

Alumina

21-45

Tungsten

104

Fig. 31 Schematic illustrating use of the C-SAM reflection technique to evaluate the die-attach bond between the silicon die and the ceramic package of a ceramic dual-in-line package IC. With this technique, the ultrasound access to the bond layer is obtained by applying the pulse to the ceramic package, rather than through the lid, to avoid the air gap over the die.

More complete information on acoustic impedances can be found in Ref 19. If the amplitude of an echo relative to the amplitude of the incident acoustic pulse is denoted by R, then the value of R can be easily determined by the formula:

(Eq 1)

where Z1 is the acoustic impedance of the material through which the ultrasound is traveling before reaching the interface and Z2 is the acoustic impedance of the material encountered on the other side of the interface. Using Eq 1 and considering the interface materials to be alumina (Z1 = 35) and silicon (Z2 = 20) a good bond will have an R value of 0.27. In the case of a complete disbands Z2 = 0 (for air) and the R value is 1.0. In interpreting C-SAM images, because the ratio of signal levels between good and bad will be 3.7, care must be exercised to calibrate the system on a known sample. Fig. 13 in the article "Use of Color for NDE" in this Volume shows a reflection mode C-SAM image produced at 50 MHz of the sample shown in Fig. 29. To assist the low-contrast differentiation of typical black-and-white gray-scale images, the C-SAM data have been digitally processed and presented as a pseudo-color-enhanced image in which red indicates disbond. The time gate was selected for the die-attach interface; therefore, unbonded regions of the die do not appear in the image, because from the ceramic side, the ultrasound cannot differentiate the air gap surrounding the die from the air gap under the disbonded areas of the die. In the case of unknown materials or in the case of not having an appropriate calibration sample, it may be more difficult to differentiate the bond versus the nobond condition because of nearly equal echo levels. Therefore, it is usually suggested to remove the lid from one sample of the lot and perform SLAM imaging. With SLAM, the amplitude of signal transmission across a two-material interface, T, is governed by:

T=1-R

(Eq 2)

In the case of a good bond, T = 0.73, and in the case of a bad bond T = 0. The ratio of signal levels is limited only by the dynamic range of the instrument and can be typically 1000:1. To clarify this further, the ability of any electronic instrument to measure a zero signal level is determined by the signal-to-noise ratio of the instrument, which can typically exceed 60 dB. Another important issue arises in ICs when the die is encapsulated in plastic molding compound instead of a ceramic package. In this case, both SLAM and C-SAM can be employed. When SLAM is employed, disbonds will appear nontransmitting and therefore dark (or whichever color is assigned). However, with C-SAM, more care must be taken in echo-level discrimination. The acoustic impedance of plastic is low (Z = 3) relative to ceramic; however, the echo magnitude from a plastic/air interface will still be 1.0. The echo from a plastic/silicon interface, however, will be 0.75, which is very close to 1.0, thus making the interpretation of echo level possibly difficult. Fortunately, however, the polarity of the echoes is different for air and silicon interfaces. In the case of Z2 > Z1 (plastic/silicon), the echo is positive and in the case of Z2 < Z1 (plastic/air), the echo is negative, according to Eq 2. In the color-coded enhanced C-mode scanning acoustic microscope, the images can be displayed as echo-magnitude images or as echo-magnitude and polarity images (Ref 11). This is illustrated in Fig. 32 and in the article "Use of Color for NDE" (Fig. 12) in this Volume; both of these figures show a plastic-encapsulated IC that has disbonded areas between the plastic molding compound and the metal lead frame. As shown in Fig. 12 in the article "Use of Color for NDE," the color differentiation between positive echoes (blue) and negative echoes (disbond) (red) is essential for correctly diagnosing the integrity of this device. The echo signal levels from the plastic/lead frame interface are similar to the levels from the plastic/air gap interface; therefore, the good versus bad condition cannot be analyzed in the black-and-white presentation of Fig. 32. A through transmission SLAM image, such as that shown in Fig. 33, is clearer with regard to differentiating between bond (bright) and disbond (dark). A more thorough discussion of the acoustic microscopy analysis of plastic-encapsulated ICs can be found in Ref 20.

Fig. 32 C-SAM reflection mode image at 15 MHz of a plastic-encapsulated IC showing a suspicious area of the lead frame. In this image, the brightness of the image (towards white) represents the magnitude of the echoes from the interface. There is no regard for the polarity of the echoes. The color version of this image, shown in Fig. 12 in the article "Use of Color for NDE" in this Volume, is a dramatic improvement over the black-and-white image shown here and clearly differentiates the bonded regions from the unbonded regions of the lead frame.

Fig. 33 SLAM through transmission image produced at 10 MHz of sample shown in Fig. 32 of this article and Fig. 12 in the article "Use of Color for NDE" in this Volume. In this image, the disbonded zones are presented as dark. Careful analysis of Fig. 33 indicates that the disbonded areas in black-and-white are larger than those indicated in the color micrograph. This is explained by the fact that the scanning laser acoustic microscope shows the disbonds on either side of the lead frame, that is, two interfaces, while the C-SAM image shows only one interface. When the IC was turned over and the opposide side of the lead frame examined, the SLAM and C-SAM information agreed. Field of view: 35 × 26 mm

At this point it should be mentioned that the interpretation of echoes is not always simple. Equations 1 and 2 are restricted to an interface of two thick materials. If there is a thin layer, echoes from the front and back surfaces of the thin layer will merge, and the resulting pulse shape will become distorted. More detailed analysis of the echo shape will be needed to determine the nature of the interface (Ref 21). Complete information on IC packaging is available in Ref 22. Integrated Circuit Wire Bonding. Another problem with ICs is the connection of the tiny wires to the silicon chip. High-speed, densely spaced ICs may contain well over 300 leads. Traditional techniques of wire bonding use 0.018 mm (0.0007 in.) diam gold wires that are individually point-by-point stitch bonded from pads on the chip to pads on the package. Newer methods involve gang bonding, that is, the simultaneous bonding of a network frame of leads mechanically fixed in position by a polyimide film (tape). These are created by photo etching a pattern in a solid sheet of 0.025 mm (0.001 in.) thick copper that is bonded to the film. A sample produced by this method is shown in Fig. 34. This process, known as tape automated bonding (TAB), may soon become a widely used method for assembling ICs of all types.

Fig. 34 IC bonded to a frame of tiny leads that have been simultaneously bonded by a process known as tape automated bonding. This process is being used to perform interconnections on densely packed ICs having over 300 leads per chip.

In order for the process to become acceptable, however, the bond integrity of each lead must be ensured. With conventional stitch wire bonding, each wire can be stressed up to a few grams-force in a device called a nondestructive mode pull tester. Unfortunately, the high population density of TAB leads precludes pull testing, except in the case where the device is destroyed to measure the quality of the lead bonds. Acoustic microscopy, particularly SLAM, was determined to be a very reliable nondestructive test method (Ref 23, 24). The notion behind developing the test was that if the areas of bonding of the leads could be measured and if the areas of the bonds correlated with the mechanical strength of the bonds upon destruction, then the test method could be employed nondestructively on subsequent samples (Ref 25). The good correlations described in Ref 23 and 24 are summarized and illustrated in the discussion that follows. Figures 35 and 36 show 200-MHz SLAM images of lead bonds on two IC chips. Some of the leads are bonded well, as indicated by the clear, bright areas, and other leads are obviously disbonded, completely or partially. Figure 37 shows a graph in which the area of bond, relative to a 100% maximum, is plotted for leads 1 through 68 around the perimeter of a chip. Pull strength relative to a maximum value of 100% is also given for each lead. The pull tests were performed to the point of failure of each lead, and the peak force was recorded. This acoustic microscopy procedure is under consideration as a standard test method for the military.

Fig. 35 SLAM acoustic image at 200 MHz of TAB-bonded leads on an IC. The bright areas at the tips of each lead indicate good-quality bonds. Field of view: 1.75 × 1.3 mm

Fig. 36 SLAM acoustic image at 200 MHz of TAB-bonded leads that are not of as uniform quality as those shown in Fig. 35. The poor bonds are mechanically weak and are more likely to suffer from long-term reliability problems. Field of view:1.75 × 1.3 mm

Fig. 37 Plot of bond area percentage versus pin location to verify that SLAM bond strength analysis agrees with destructive pull tests. Graph shows the similarity between the relative strength of TAB-bonded leads, as determined by destructive pull tests, and the areas of bond, as determined by 200-MHz SLAM, on the same sample plotted as a function of location around the perimeter of an IC chip. The locations are denoted by pin numbers 1 through 68. Pressure readings for inner lead bonding are averages obtained for numerous samples that were tested.

Example 3: Use of SAM to Evaluate an IC on a Silicon Wafer. Scanning acoustic microscopy offers a technique for generating high-resolution images of material structure and defects in ICs located on silicon crystal substrates. When the IC is scanned with a scanning acoustic microscope having a 2-GHz transducer, the image produced measures approximately 62 × 62 m (0.0025 × 0.0025 in.). This image displays the metallization on an IC. A high-resolution, pulse-echo scan shows 4 to 6 m (160 to 240 in.) wide aluminum lines deposited on a silicon substrate (Fig. 38). The small dark spots randomly scattered throughout the metallization are silicon nodules, which bubble up from the substrate, creating small defects in the conducting material. Depending on the focal depth, such structures as defects in the metallization, delaminations or disbonds between conductors, metal line thickness, and buried p-type material can be measured or imaged. The shaded fingers midway on the right side in Fig. 38 indicate doped material below the surface. No preparation of the sample prior to testing was necessary.

Fig. 38 Pulse-echo SAM micrograph of 4 to 6 m (160 to 240 n.) wide aluminum lines on the silicon substrate of an IC. Dark spots are silicon nodule defects, the source of which is silicon bubbling up from the substrate of the wafer. Courtesy of G.H. Thomas, Sandia National Laboratories, Livermore, CA

Ceramic capacitors are passive electronic components used in virtually every electronic circuit to store electrical

charge. They are manufactured to be physically small, yet they must have very high capacitance values, which can be obtained with specially formulated, high dielectric constant materials. Figure 39 shows various miniature multilayer capacitors. The basic construction of a capacitor involves two parallel plate electrodes separated by a small distance that is occupied by the dielectric. Capacitance values are inversely proportional to the dielectric thickness and are proportional to the total area of the electrodes.

Fig. 39 Miniature ceramic capacitors constructed of multiple layers of high dielectric constant ceramic alternated with layers of conductive metallization. Relative size of single capacitor is evident when compared to the head of a match.

To increase the area in a small space, multiple layers are used and are electrically connected together in parallel. The problems associated with capacitors that cannot be detected electrically are internal cracks in the ceramic, large voids in the dielectric layers, delaminations between the layers, and high porosity of the dielectric ceramic material. This is illustrated in Fig. 40. Most ceramic capacitors have barium titanate in the dielectric formulation and are therefore x-ray opaque.

Fig. 40 Diagram illustrating the layered construction used in multilayer ceramic capacitors and the locations of potential defects in these electronic components

Ultrasound C-scan was suggested as a method for detecting flaws in ceramic capacitors (Ref 26). However, it was later found to be useful only for the large components because of edge effects that are characteristic of all C-scan and C-SAM images. Scanning laser acoustic microscope techniques are not necessarily limited by the edge effects, as illustrated in Fig. 41.

Fig. 41 Edge effect in (a) C-SAM and (b) SLAM. In the case of C-SAM, when the transducer is too close to the left edge of a sample having thickness t, the acoustic beam becomes cut off, and the echo signal does not return from the bottom surface (side view, upper figure). Assuming a ray angle of 45°, the edge of the sample cannot be inspected to within a distance t, thus limiting the area of scan of a sample (top view, lower figure). If the sample has a lateral dimension D 2t, the focused reflection mode cannot be employed in a valid manner. By way of comparison, in the case of SLAM, because collimated energy is used, the edge effect is minor and limited only by diffraction of the illumination wave at the edges (also shown in side view, upper figure, and top view, lower figure).

Figure 42 shows a 30-MHz SLAM image of a ceramic capacitor that has a relatively uniform appearance in transmission. The dark circular spot is a delamination that was confirmed by destructive cross sectioning of the component, as shown in Fig. 43.

Fig. 42 SLAM image at 30 MHz of a plastic-encapsulated ceramic capacitor having a significant defect that can cause long-term reliability problems. This defect is not evident by electrical testing. Field of view: 14 × 10.5 mm

Fig. 43 Cross-sectional analysis of the capacitor shown in Fig. 42 revealing a delamination in the location indicated by SLAM. The cut edge of the capacitor was polished and placed under a conventional microscope.

The analysis of ceramic capacitors by SLAM has become a standard method for the nondestructive evaluation and screening of ceramic capacitors for internal defects (Ref 27). Investigations have been performed to correlate the information obtained nondestructively with that obtained by conventional destructive physical analysis. Quantitative data have been acquired and associated with the physical properties of the materials. Users of the technique have reported significant improvements in reliability as a result of the SLAM screening of parts. Silicon Thyristors. The silicon thyristor, which is basically three close coupled p-n junctions in the form of a p-n-p-n structure, is an important switching device. The thyristor has two stable states: the on state with high current and low voltage and the off state with low current and high voltage. Because of its two stable states and the low power dissipation in these two states, the thyristor has unique applications ranging from speed control in home appliances to switching and power inversion in high-voltage transmission lines. The following example illustrates the use of acoustic microscopy to evaluate whether or not a thyristor has been correctly assembled and will operate according to design specifications.

Example 4: Use of Acoustic Microscopy to Evaluate Gate-Turnoff Silicon-Controlled Thyristor Integrity. The performance of 77 mm (3.0 in.) diam, 5-kV, 2-kA silicon power thyristors is dependent on good electrical attachment between the gate array and the underlying silicon device. Gate-turnoff (GTO) thyristors are fabricated by the solid-state attachment of a conducting pattern between one large [100] cut silicon wafer and two smaller concentric annular wafer segments also having [100] cut orientations. Both the underlying wafer and the annular segments are 0.51 mm (0.020 in.) thick. The large underlying wafer is 77 mm (3.0 in.) in diameter, and the larger-diameter annulus is approximately 51 mm (2.0 in.) in diameter. The successful fabrication of this silicon-controlled power device requires that the conducting patterns be of good quality, that the wafers be intact (no cracks), and that the bonds be well attached. Acoustic microscopy was chosen to inspect the conducting patterns at the interface between the overlying annuli and the underlying silicon wafer. The surfaces presented to the interrogating ultrasonic beam were lapped and coated with approximately 1.0 m (40 in.) of gold for good electrical contact. The images shown in Fig. 44 were made by focusing a 50-MHz beam on the interface between the annuli and the underlying wafer. This produced a resolution spot size of approximately 100 m (0.004 in.). The beam was then scanned, and a 1024 × 1024 pixel image was acquired by recording and displaying the echo from the described interface.

Fig. 44 Thyristor gate arrays imaged through 0.51 mm (0.020 in.) of overlying [100] cut silicon. The annular rings combined with the underlying silicon wafer make up a 5-kV, 2-kA gate-turnoff power device. (a) This device shows well-attached and well-formed finger arrays. (b) This device shows poorly formed and poorly attached arrays. The cracked spacer ring is not critical to the operation of (a) but was included to display cracks as they would appear if the array annuli had cracked. Courtesy of R.S. Gilmore, General Electric Research and Development Center

Figure 44 shows two silicon-controlled GTO thyristors that contain most of the desirable and undesirable features that can be produced in the fabrication of a silicon power device. The device in Fig. 44(a) shows well-developed finger arrays except for the two outer fingers at the approximate three o'clock position in the inner ring. A silicon spacer ring used to keep the two annuli centered was badly cracked, but this ring will be removed in a subsequent fabrication operation. The spacer ring is included in this image to show how cracks would appear in the silicon. The device in Fig. 44(b) shows very poorly formed finger arrays, and the resulting device would probably be inoperable if it were assembled and tested. This nondestructive evaluation of GTO thyristors demonstrated that acoustic microscopy is a valuable tool for the process and quality control of silicon power devices.

References cited in this section

11. F.J. Cichanski, Method and System for Dual Phase Scanning Acoustic Microscopy, Patent Pending 16. M. Oishi, K. Noguchi, T. Masaki, and M. Mizushina, Defect Characterization of Tetragonal Zirconia Polycrystals by a Scanning Laser Acoustic Microscope, in Proceedings of the International Meeting on Advanced Materials (Tokyo), Materials Research Society MRS, 1988 17. L.W. Kessler, J.E. Semmens, and F. Agramonte, Nondestructive Die Attach Bond Evaluation Comparing Scanning Laser Acoustic Microscopy (SLAM) and X-Radiography, in Proceedings of the 35th Electronic Components Conference (Washington, D.C.), Institute of Electrical and Electronics Engineers, 1985, p 250258 18. MIL-STD-883C, Method 2030, U.S. Government Printing Office 19. A.R. Selfridge, Approximate Material Properties in Isotropic Materials, Trans. IEEE, Vol SU-32 (No. 3),1985, p 381-394 20. J.E. Semmens and L.W. Kessler, Nondestructive Evaluation of Thermally Shocked Plastic Integrated Circuit Packages Using Acoustic Microscopy, in Proceedings of the International Symposium on Testing and Failure Analysis, ASM INTERNATIONAL, 1988, p 211-215 21. S. Lees, Ultrasonic Measurement of Thin Layers, Trans. IEEE, Vol SU-18 (No. 2), 1971, p 81-86 22. Packaging, Vol 1, Electronic Materials Handbook, ASM INTERNATIONAL, 1989 23. J.E. Semmens and L.W. Kessler, Nondestructive Evaluation of TAB Bonds by Acoustic Microscopy, in Proceedings of the ISHM Conference (Seattle, WA), International Society for Hybrid Microelectronics, 1988, p 455-463 24. "Nondestructive Evaluation of Metallurgical Tape Bonds," Report 88-7114, Rome Air Development Center,

Griffiss Air Force Base (summary available from Sonoscan, Inc.), 1988 25. L.W. Kessler, Acoustic Microscopy: Nondestructive Tool for Bond Evaluation on TAB Interconnections, in Proceedings of the ISHM Symposium (Dallas, TX), International Society for Hybrid Microelectronics, 1984, p 79-84 26. G.R. Love, Nondestructive Testing of Monolithic Ceramic Capacitors, in Proceedings of the 1973 International Microelectronics Symposium, 1A3-1A8, International Society for Hybrid Microelectronics, 1973 27. L.W. Kessler and J.E. Semmens, Nondestructive SLAM Analysis of Ceramic Capacitors: An Overview of Experiences From 1980-1989, in Proceedings of the Capacitor and Resistor Technology Symposium (Orlando, FL), Components Technology Institute Note cited in this section

* Examples 1and 3 were prepared by G.H. Thomas, Sandia National Laboratories, Livermore, CA. Example 4 was prepared by R.S. Gilmore, General Electric Research & Development Center. Acoustic Microscopy Lawrence W. Kessler, Sonoscan, Inc.

References 1. 2. 3. 4. 5. 6.

7. 8. 9. 10.

11. 12. 13.

A. Korpel, L.W. Kessler, and P.R. Palermo, Acoustic Microscope Operating at 100 MHz, Nature, Vol 232 (No. 5306), 1970, p 110-111 Product Bulletin, Sonoscan, Inc. R.A. Lemons and C.F. Quate, Acoustic Microscope--Scanning Version, Appl. Phys. Lett., Vol 24 (No. 4), 1974, p 163-165 L.W. Kessler and D.E. Yuhas, Acoustic Microscopy--1979, Proc. of IEEE, Vol 67 (No. 4), April 1979, p 526-536 Z.C. Lin, H. Lee, G. Wade, M.G. Oravecz, and L.W. Kessler, Holographic Image Reconstruction in Scanning Laser Acoustic Microscopy, Trans. IEEE, Vol UFFC-34 (No. 3), May 1987, p 293-300 B.Y. Yu, M.G. Oravecz, and L.W. Kessler, "Multimedia Holographic Image Reconstruction in a Scanning Laser Acoustic Microscope," Paper presented at the 16th International Symposium on Acoustical Imaging (Chicago, IL), Sonoscan, Inc., June 1987; L.W. Kessler, Ed., Acoustical Imaging, Vol 16, Plenum Press, 1988, p 535-542 L.W. Kessler, VHF Ultrasonic Attenuation in Mammalian Tissue, Acoust. Soc. Am., Vol 53 (No. 6), 1973, p 1759-1760 M.G. Oravecz, Quantitative Scanning Laser Acoustic Microscopy: Attenuation, J. Phys. (Orsay), Vol 46, Conf. C10, Supplement 12, Dec 1985, p 751-754 S.A. Goss and W.D. O'Brien, Jr., Direct Ultrasonic Velocity Measurements of Mammalian Collagen Threads, Acoust. Soc. Am., Vol 65 (No. 2), 1979, p 507-511 M.G. Oravecz and S. Lees, Acoustic Spectral Interferometry: A New Method for Sonic Velocity Determination, in Acoustical Imaging, Vol 13, M. Kaveh, R.K. Mueller, and T.F. Greenleaf, Ed., Plenum Press, 1984, p 397-408 F.J. Cichanski, Method and System for Dual Phase Scanning Acoustic Microscopy, Patent Pending J.S. Foster and D. Rugar, Low-Temperature Acoustic Microscopy, Trans. IEEE, Vol SU-32, 1985, p 139151 K.K. Liang, G.S. Kino, and B.T. Khuri-Yakub, Material Characterization by the Inversion of V(z), Trans.

14. 15. 16.

17.

18. 19. 20.

21. 22. 23.

24. 25.

26.

27.

IEEE, Vol SU-32, 1985, p 213-224 R.D. Weglein, Acoustic Micro-Metrology, Trans. IEEE, Vol SU-32, 1985, p 225-234 J. Kushibiki and N. Chubachi, Material Characterization by Line-Focus-Beam Acoustic Microscope, Trans. IEEE, Vol SU-32, 1985, p 189-212 M. Oishi, K. Noguchi, T. Masaki, and M. Mizushina, Defect Characterization of Tetragonal Zirconia Polycrystals by a Scanning Laser Acoustic Microscope, in Proceedings of the International Meeting on Advanced Materials (Tokyo), Materials Research Society MRS, 1988 L.W. Kessler, J.E. Semmens, and F. Agramonte, Nondestructive Die Attach Bond Evaluation Comparing Scanning Laser Acoustic Microscopy (SLAM) and X-Radiography, in Proceedings of the 35th Electronic Components Conference (Washington, D.C.), Institute of Electrical and Electronics Engineers, 1985, p 250-258 MIL-STD-883C, Method 2030, U.S. Government Printing Office A.R. Selfridge, Approximate Material Properties in Isotropic Materials, Trans. IEEE, Vol SU-32 (No. 3),1985, p 381-394 J.E. Semmens and L.W. Kessler, Nondestructive Evaluation of Thermally Shocked Plastic Integrated Circuit Packages Using Acoustic Microscopy, in Proceedings of the International Symposium on Testing and Failure Analysis, ASM INTERNATIONAL, 1988, p 211-215 S. Lees, Ultrasonic Measurement of Thin Layers, Trans. IEEE, Vol SU-18 (No. 2), 1971, p 81-86 Packaging, Vol 1, Electronic Materials Handbook, ASM INTERNATIONAL, 1989 J.E. Semmens and L.W. Kessler, Nondestructive Evaluation of TAB Bonds by Acoustic Microscopy, in Proceedings of the ISHM Conference (Seattle, WA), International Society for Hybrid Microelectronics, 1988, p 455-463 "Nondestructive Evaluation of Metallurgical Tape Bonds," Report 88-7114, Rome Air Development Center, Griffiss Air Force Base (summary available from Sonoscan, Inc.), 1988 L.W. Kessler, Acoustic Microscopy: Nondestructive Tool for Bond Evaluation on TAB Interconnections, in Proceedings of the ISHM Symposium (Dallas, TX), International Society for Hybrid Microelectronics, 1984, p 79-84 G.R. Love, Nondestructive Testing of Monolithic Ceramic Capacitors, in Proceedings of the 1973 International Microelectronics Symposium, 1A3-1A8, International Society for Hybrid Microelectronics, 1973 L.W. Kessler and J.E. Semmens, Nondestructive SLAM Analysis of Ceramic Capacitors: An Overview of Experiences From 1980-1989, in Proceedings of the Capacitor and Resistor Technology Symposium (Orlando, FL), Components Technology Institute

Use of Color for NDE

Introduction MAJOR DEVELOPMENTS in digital electronics have led to the commercial availability of a wide range of specialized devices for image processing and enhancement. With digital image enhancement systems, data interpretation becomes easier, difficult-to-detect features are more readily discernible, and quantitative analysis is possible. Among the numerous image-processing operations available (see the article "Digital Image Enhancement" in this Volume), color-enhanced images produce some of the more valuable, if not spectacular, results. Use of Color for NDE

Color Images

This collection of color images is intended to assist the reader in recognizing the advantages and/or limitations of pseudocolor in NDE. This collection both complements and supplements articles in this Handbook that describe in detail the principles and applications of methods that utilize color for discontinuity characterization. Numerous comparisons are made between figures appearing in color in this section and their black-and-white counterparts found elsewhere in this Volume. The types of color images appearing in this article include:

Inspection method

Fig. No.

Ultrasonic inspection

2, 3, 6(c), 6(d), 10, 21, 23

Acoustic microscopy

5, 11, 12, 13, 14, 15, 16

Thermal analysis

17, 18, 19, 20, 22

Computed tomography

1, 7, 8, 9

Digital radiography

6(a), 6(b)

Stress analysis(a)

24

(a) The photoelastic fringe patterns shown in Fig. 24 are the result of stress/strain analysis. See the article "Strain Measurement for Stress Analysis" for details.

Fig. 1 Computed tomographic reconstruction of a 9-V battery. (a) Image reconstructed from 500 views showing evidence of undersampling. (b) Image reconstructed from 1000 views, the data filtered with a Wiener filter before backprojection to correct for the point spread of the beam. (T.N. Claytor, Los Alamos National Laboratory)

Fig. 2 Linear histogram equalization (image enhancement technique) of a low-contrast ultrasonic image of a bond line in an explosively bonded steel-to-aluminum plate. The detonation wave travels from top to bottom, with higher values of intensity indicating better bonding. (a) and (b) Digitized image with 64 levels and corresponding histogram. (c) and (d) After the histogram equalization process, the colors are more widely separated so that the subtler features of the bonding can be seen. (T.N. Claytor, Los Alamos National Laboratory)

Fig. 3 Effects of a low-pass filter and a median filter on a slightly noisy ultrasonic image of a cross-rolled tungsten plate with internal rolling tears. (a) The original image has point noise and horizontal streaks. (b) After low-pass filtering, the point noise has been eliminated, but the streaks are still visible. (c) After median filtering, the point noise is eliminated, the streaks are less prominent, and the sharpness of the original image is preserved. (T.N. Claytor, Los Alamos National Laboratory)

Fig. 4 Palettes useful for black-and-white and color presentation of data. (a) Standard linear gray scale. (b) Gray scale with a red threshold for identifying out-of-bound conditions. (c) The most colorful palette spectrum. (d) and (e) Complementary palettes. (T.N. Claytor, Los Alamos National Laboratory)

Fig. 5 50-MHz acoustic microscope image of a stainless steel-to-stainless steel diffusion bond with a silver interlayer. The blue, white, and black areas indicate poorer bonding. The pattern displayed in this image matched the ductile/brittle pattern on the fracture surface. (G.H. Thomas, Sandia National Laboratories)

Fig. 6 Examples of contrast enhancement techniques. (a) Radiograph and corresponding line density chart of a ceramic disk containing inclusions (white) and voids (dark areas). (b) After a field-flattening contrast enhancement procedure, the large voids and inclusions are more easily detected at the expense of smaller voids. (c) Ultrasonic image of a HIPed tungsten plate showing density variations enhanced by the use of color. (d) Ultrasonic image of a hot-rolled tungsten plate showing texture variations. Compare Fig. 6(c) and 6(d) with Fig. 6 and 7 in the article "Digital Image Enhancement." (T.N. Claytor, Los Alamos National Laboratory)

Fig. 7 Digital replica of a three-dimensional surface image of a turbine blade generated from computed tomography data (M.J. Dennis, General Electric NDE Systems and Services)

Fig. 8 Voxel reconstruction of a turbine blade derived from computed tomography slices showing internal cooling air passage surfaces outlined in blue (T.N. Claytor, Los Alamos National Laboratory)

Fig. 9 Three-dimensional tomographic end view of a failed nuclear fuel rod bundle showing white insulation surrounding rods cut away (with threshold operation) at an angle to expose relative positions of orange rods (T.N. Claytor, Los Alamos National Laboratory)

Fig. 10 Representations of two-dimensional data in a three-dimensional format. (a) Ultrasonic image of defects in circuit board pads. (b) C-scan of a filament-wound vessel with defects. (T.N. Claytor, Los Alamos National Laboratory)

Fig. 11 30-MHz SLAM acoustic image of a Kevlar-reinforced composite test coupon with delamination and cracks. The damage was induced by cutting the bar perpendicular to fibers. The extent of the damage is shown by the ingression of color, with red/green indicating the most extensive damage. The colors are associated with levels of acoustic transmission. A high level (white) indicates a homogeneous material, while a lower level (red/green) indicates discontinuities. (L.W. Kessler, Sonoscan, Inc.)

Fig. 12 15-MHz C-SAM reflection mode image of a plastic-encapsulated integrated circuit showing defects in the lead frame. In this image, the echo polarity information is presented by color scale changes in which disbonds are red (negative polarity) and good bonds (positive echoes) are blue. Without this polarity information, it is difficult to make a correct interpretation of the bonding condition in this sample (see Fig. 32 in the article "Acoustic Microscopy"). (L.W. Kessler, Sonoscan, Inc.)

Fig. 13 50-MHz C-SAM reflection mode image of the same die-attach sample shown in Fig. 29 and 30 in the article "Acoustic Microscopy." In this image, areas of disbond are indicated by the color red. (L.W. Kessler, Sonoscan, Inc.)

Fig. 14 30-MHz SLAM acoustic micrograph of the carbon fiber reinforced polymer component shown in Fig. 13 of "Acoustic Microscopy." Long, irregular defect (arrow) corresponds to blistering of one of the laminates. (L.W. Kessler, Sonoscan, Inc.)

Fig. 15 100-MHz SLAM micrograph of the alumina panel shown in Fig. 16 in the article "Acoustic Microscopy." Color is used to delineate a crack with a specific acoustic brightness level. (L.W. Kessler, Sonoscan, Inc.)

Fig. 16 15-MHz reflection mode C-SAM images of internal defects in the alumina test disk shown in Fig. 18 in the article "Acoustic Microscopy." In (a), the darker features correspond to higher-magnitude echo levels than the yellow. In (b), the focus and gate levels were changed slightly. Because the defects are not parallel to the surface of the part, echoes will not return at different times; therefore, a fixed-gate position will show echoes only at the depth selected. (L.W. Kessler, Sonoscan, Inc.)

Fig. 17 Thermal images of graphite-epoxy sheets with 20 × 20 mm implanted Teflon defects at 0.3 mm depth (a) and 2.25 mm depth (b). Image processing (c) improved the detectability of the 2.25 mm depth defect. (P. Cielo, National Research Council of Canada)

Fig. 18 Thermal analysis of a silicon mass airflow sensor. (a) Heater resistor (arrow) under analysis, 50×. (b) and (c) Thermal maps indicating a 100 °C rise in temperature of an exposed heat resistor chip under no-flow conditions and normal voltage (b) and under 150% voltage (c). (R.W. Gehman, Micro Switch).

Fig. 19 Emissivity display of a GaAs FET semiconductor chip. Color scale at right represents 16 emissivity levels between lowest (black) and highest (white) values. (J.O. Brown, EDO Corp.)

Fig. 20 Temperature display of a GaAs FET device showing a substrate bonding flaw as an oval-shaped anomaly (yellow) across center. Area displayed is 0.6 mm across. (J.O. Brown, EDO Corp.)

Fig. 21 50-MHz surface wave ultrasonic image of scratch and lapping marks in Si3N4. The smallest visible scratches are 25 to 50 m wide. (T. Nelligan, Panametrics Inc.)

Fig. 22 Thermal imaging of two simulated defects (Teflon inserts) in a graphite-epoxy laminate (D. Mauro, Techmarketing Inc./AGEMA Infrared Systems)

Fig. 23 Ultrasonic image of an impacted region in a 1.4 mm thick graphite-reinforced polymer sheet. The damaged region is 6 mm long. (M.C. Bhardwaj, Ultron Laboratories, Inc.)

Fig. 24 Photoelastic fringe patterns produced when a photoelastic material is strained and viewed under polarized light (a) and the photoelastic color sequence viewed with a circular polariscope (b). N, fringe order phase shift. (L.D. Lineback, Measurements Group, Inc.)

Nondestructive Inspection of Forgings

Introduction IN FORGINGS of both ferrous and nonferrous metals, the flaws that occur most often are caused by conditions that exist in the ingot, by subsequent hot working of the ingot or the billet, and by hot or cold working during forging. The nondestructive inspection (NDI) methods most commonly used to detect these flaws include visual, magnetic particle, liquid penetrant, ultrasonic, eddy current, and radiographic inspection. This article discusses the applications of these methods to forgings. Information on the equipment and techniques used in these inspection methods is available in the articles so titled in this Volume.

Nondestructive Inspection of Forgings

Flaws Originating in the Ingot Many large open-die forgings are forged directly from ingots. Most closed-die forgings and upset forgings are produced from billets, rolled bar stock, or preforms. Many, though by no means all, of the imperfections found in forgings can be attributed to conditions that existed in the ingot, sometimes even when the ingot has undergone primary reduction prior to the forging operation. Some, but again by no means all, of the service problems that occur with forgings can be traced to imperfections originating in the ingot. Chemical Segregation. The elements in a cast alloy are seldom distributed uniformly. Even unalloyed metals contain random amounts of various types of impurities in the form of tramp elements or dissolved gases; these impurities are also seldom distributed uniformly. Therefore, the composition of the metal or alloy will vary from location to location. Deviation from the mean composition at a particular location in a forging is termed segregation. In general, segregation is the result of solute rejection at the solidification interface during casting. For example, the gradation of composition with respect to the individual alloying elements exists from cores of dendrites to interdendritic regions. Segregation therefore produces a material having a range of compositions that do not have identical properties.

Forging can partially correct the results of segregation by recrystallizing or breaking up the grain structure to promote a more homogeneous substructure. However, the effects of a badly segregated ingot cannot be totally eliminated by forging; rather, the segregated regions tend to be altered by the working operation, as shown in Fig. 1. In metals, the presence of localized regions that deviate from the nominal composition can affect corrosion resistance, forging and joining (welding) characteristics, mechanical properties, fracture toughness, and fatigue resistance. In heat-treatable alloys, variations in composition can produce unexpected responses to heat treatments, which result in hard or soft spots, quench cracks, or other flaws. The degree of degradation depends on the alloy and on process variables. Most metallurgical processes are based on an assumption that the metal being processed is of a nominal and reasonably uniform composition. Ingot

Pipe

and

Centerline

Shrinkage.

A common imperfection in ingots is the shrinkage cavity, commonly known as pipe, often found in the upper portion of the ingot. Shrinkage occurs during freezing of the metal, and eventually there is Fig. 1 Microstructural bonding due to chemical insufficient liquid metal near the top end to feed the ingot. As a result, a cavity forms, usually approximating the shape of a segregation and mechanical working cylinder or cone--hence the term pipe. Piping is illustrated in Fig.

2. In addition to the primary pipe near the top of the ingot, secondary regions of piping and centerline shrinkage may extend deeper into an ingot (Fig. 3).

Fig. 2 Schematic showing piping in top-poured ingots

Primary piping is generally an economic concern, but if it extends sufficiently deep into the ingot body and goes undetected, it can eventually result in a defective forging. Detection of the pipe can be obscured in some cases if bridging has occurred. Piping can be minimized by pouring ingots with the big end up, by providing risers in the ingot top, and by applying sufficient hot-top material (insulating refractories or exothermic materials) immediately after pouring. These techniques extend the time that the metal in the top regions of the ingot remains liquid, thus minimizing the shrinkage cavity produced in this portion of the ingot. On the other hand, secondary piping and centerline shrinkage can be very detrimental because they are harder to detect in the mill and may subsequently produce centerline defects in bar and wrought products. Such a material condition may indeed provide the flaw or stress concentrator for a forging burst in some later processing operation or for a future product failure. High Hydrogen Content. A major source of hydrogen in certain metals and

alloys is the reaction of water vapor with the liquid metal at high temperatures. The water vapor may originate from the charge materials, slag ingredients and alloy additions, refractory linings, ingot molds, or even the atmosphere itself if steps are not taken to prevent such contamination. The resulting hydrogen goes into solution at elevated temperatures; but as the metal solidifies after pouring, the solubility of hydrogen decreases, and it becomes entrapped in the metal lattice. Hydrogen concentration in excess of about 5 ppm has been associated with flaking, especially in heavy sections and high-carbon steels. Hydrogen flakes (Fig. 4) are small cracks produced by hydrogen that has diffused to grain boundaries and other preferred sites, for example, inclusion/matrix interfaces. However, hydrogen concentrations in excess of only 1 ppm have been related to the degradation of mechanical properties in high-strength steels, especially ductility, impact behavior, Longitudinal section and fracture toughness. an ingot showing

Fig. 3 through extensive centerline shrinkage

Fig. 4 Hydrogen flaking in an alloy steel bar. (a) Polished cross section showing cracks due to flaking. (b) Fracture surface containing hydrogen flakes. Note the reflective, faceted nature of the fracture. (c) SEM micrograph showing the intergranular appearance of the flakes in this material

Metals can also possess a high hydrogen content without the presence of flakes or voids. In this case, the hydrogen may cause embrittlement of the material along selective paths, which can drastically reduce the resistance of a forged part to crack propagation resulting from impact loading, fatigue, or stress corrosion. In cases where hydrogen-related defects can serve as the initiation site for cracking and thus increase the likelihood of future failures, it is advisable to use a thermal treatment that can alleviate this condition. For example, slow cooling immediately following a hot-working operation or a separate annealing cycle will relieve residual stresses in addition to allowing hydrogen to diffuse to a more uniform distribution throughout the lattice and, more important, to diffuse out of the material. Nonmetallic inclusions, which originate in the ingot, are likely to be carried over to the forgings, even though several

intermediate hot-working operations may be involved. Also, additional inclusions may develop in the billet or in subsequent forging stages. Most nonmetallic inclusions originate during solidification from the initial melting operation. If no further consumableremelting cycles follow, as in air-melted or vacuum-induction products (with no remelting cycle to follow), the size, frequency, and distribution of the nonmetallic inclusions will not be altered or reduced in size or frequency during further processing. If a subsequent vacuum-remelting operation is used, the inclusions will be lessened in size and frequency and will become more random in nature. If an electroslag-remelting cycle is used, a more random distribution of inclusions will result. Two kinds of nonmetallic inclusions are generally distinguished in metals: • •

Those that are entrapped in the metal inadvertently and originate almost exclusively from particles of matter that are occluded in the metal while it is molten or being cast Those that separate from the metal because of a change in temperature or composition

Inclusions of the latter type are produced by separation from the metal when it is in either the liquid or the solid state. Oxides, sulfides, nitrides, or other nonmetallic compounds form droplets or particles when these compounds are produced in such amounts that their solubility in the matrix is exceeded. Air-melted alloys commonly contain inclusions mainly of these chemical characteristics. Vacuum- or electroslagremelted alloys more commonly contain conglomerates of any of these types, frequently combined with carbon or the hardening element or elements that precipitate during stabilization and aging cycles to form inclusions such as titanium carbonitrides or carbides. Homogenizing cycles are normally used for the ingot prior to conversion or at an early stage of conversion.

Because these compounds are products of reactions within the metal, they are normal constituents of the metal, and conventional melting practices cannot completely eliminate such inclusions. However, it is desirable to keep the type and amount of inclusions to a minimum so that the metal is relatively free from those inclusions that cause the most problems. Of the numerous types of flaws found in forgings, nonmetallic inclusions appear to contribute significantly to service failures, particularly in high-integrity forgings such as those used in aerospace applications. In many applications, the presence of these inclusions decreases the ability of a metal to withstand high static loads, impact forces, cyclical or fatigue loading, and sometimes corrosion and stress corrosion. Nonmetallic inclusions can easily become stress concentrators because of their discontinuous nature and incompatibility with the surrounding composition. This combination may very well yield flaws of critical size that, under appropriate loading conditions, result in complete fracture of the forged part. Unmelted electrodes and shelf are two other types of ingot flaws that can impair forgeability. Unmelted electrodes (Fig. 5a) are caused by chunks of electrodes being eroded away during consumable melting and dropping down into the molten material as a solid. Shelf (Fig. 5b) is a condition resulting from uneven solidification or cooling rates at the ingot surfaces.

Fig. 5 Sections through two heat-resistant alloy ingots showing flaws that can impair forgeability. (a) Piece of unmelted consumable electrode (white spot near center). (b) Shelf (black line along edge) resulting from uneven solidification of the ingot

The consumable-melting operation has occasionally been continued to a point where a portion of the stinger rod is melted into the ingot, which may be undesirable because the composition of the stinger rod may differ from that of the alloy being melted. To prevent this occurrence, one practice is to weld a wire to the stinger rod, bend the wire down in tension, and weld the other end of the wire to the surface of the electrode a few inches below the junction of the stinger rod and the electrode. When the electrode has been consumed to where the wire is attached to it, the wire is released and springs out against the side of the crucible, thus serving as an alarm to stop the melting. A disadvantage of this practice is that the wire may become detached and contaminate the melt. Nondestructive Inspection of Forgings

Flaws Caused by the Forging Operation

Flaws produced during the forging operation (assuming a flaw-free billet or bar) are the result of improper setup or control. Proper control of heating for forging is necessary to prevent excessive scale, decarburization, overheating, or burning. Excessive scale, in addition to causing excessive metal loss, can result in forgings with pitted surfaces. The pitted surfaces are caused by the scale being hammered into the surface and may result in unacceptable forgings. Severe overheating causes burning, which is the melting of the lower melting point constituents. This melting action severely reduces the mechanical properties of the metal, and the damage is irreparable. Detection and sorting of forgings that have been burned during heating can be extremely difficult. In many cases, the flaws that occur during forging are the same as, or at least similar to, those that may occur during hot working of the ingots or billets; these are described in the previous section. Internal flaws in forgings often appear as cracks or tears, and they may result either from forging with too light a hammer or from continuing forging after the metal has cooled down below a safe forging temperature. Bursts, as described above, may also occur during the forging operation. A number of surface flaws can be produced by the forging operation. These flaws are often caused by the movement of metal over or upon another surface without actual welding or fusing of the surfaces; such flaws may be laps or folds (described previously). Cold shuts often occur in closed-die forgings. They are junctures of two adjoining surfaces caused by incomplete metal fill and incomplete fusion of the surfaces. Surface flaws weaken forgings and can usually be eliminated by correct die design, proper heating, and correct sequencing and positioning of the workpieces in the dies. Shear cracks often occur in steel forgings; they are diagonal cracks occurring on the trimmed edges and are caused by shear stresses. Proper design and condition of trimming dies to remove forging flash are required for the prevention of shear cracks. Other flaws in steel forgings that can be produced by improper die design or maintenance are internal cracks and splits. If the material is moved abnormally during forging, these flaws may be formed without any evidence on the surface of the forging. Nondestructive Inspection of Forgings

Selection of Inspection Method The principal factors that influence the selection of an NDI method for forgings include degree of required integrity of the forging, metal composition, size and shape of the forging, and cost. There are sometimes other influential factors, such as the type of forging method used. For high-integrity forgings, it is often required that more than one inspection method be employed because some inspection methods are capable of locating only surface flaws; therefore, one or more additional methods are required for locating internal flaws. For example, many forgings for aerospace applications are inspected with liquid penetrants (or with magnetic particles, depending on the metal composition) for locating surface flaws, then by ultrasonics for detecting internal flaws. Certain characteristics or conditions unique to forgings can create service problems, yet these conditions are not easily detected by nondestructive inspection. Exposed end grain, which can lead to poor corrosion resistance or to susceptibility to stress-corrosion cracking, is the most prevalent of the undesirable conditions. The strength of the forging can be adversely affected if the grains flow in an undesirable direction (as in grain reversal) or if grain flow is confined to only a portion of the section being forged rather than being well distributed. Both exposed end grain and poor grain flow can be most effectively corrected by redesigning the forging or forging blank, particularly with regard to flash. It is virtually impossible to detect exposed end grain nondestructively; only rarely can poor grain flow be detected nondestructively.

Both conditions are much more readily analyzed by sectioning and macroetching sample forgings in preproduction stages of product development. When certain steels or nonferrous alloys are forged at too high a temperature or sometimes when a part cools too slowly after forging, there is a potential for grain size in the finished forging to be excessively large. Such a condition is difficult to detect nondestructively, except with ultrasonics, and then only when the grains are very large. Even with very large grains, ultrasonic inspection cannot determine grain size quantitatively, nor can it detect large grains reliably. Only the possibility that large grains are present can be inferred from excessive attenuation of the ultrasonic beam. Effect of Type of Forging Many of the types of flaws that can occur in forgings do so without particular regard to the type of forging; that is, opendie, closed-die, upset, or rolled. However, there are many cases in which a specific type of flaw is more likely to occur in one type of forging than in another. Additional information on the types of forging processes discussed below is available in Forming and Forging, Volume 14 of ASM Handbook, formerly 9th Edition Metals Handbook. Open-Die Forgings. Most forgings produced in open dies are relatively large; therefore, their size is likely to impose

some restrictions not only on the inspection method used but also on the system within a given inspection method. For large open-die forgings, NDI methods (other than visual) are generally limited to magnetic particle or liquid penetrant inspection (for surface discontinuities) or to ultrasonic inspection (for internal flaws). In general, the flaws likely to be found in open-die forgings are similar to those that may occur in other hot-worked shapes--with the exception of forging laps and cold shuts, which usually occur only in closed-die forgings. Closed-Die and Upset Forgings. The discontinuities in closed-die forgings that can be detected by liquid penetrant inspection or magnetic particle inspection (if the forging is ferromagnetic) are the following:

•

• •

• •

Forging laps, which can be caused by incorrect die design, use of incorrect size of forging stock, excessive local conditioning of forging stock for removal of surface flaws, and excessively sharp corners in the forging stock Seams, due to incomplete removal of seams from the forging stock Surface cracks, caused by incorrect forging temperature, nonductile metallic or nonmetallic segregates in the forging stock, or surface contamination from the furnace atmosphere or other contaminants in the furnace or on the forging (such as high-sulfur fuels for heating nickel alloys or leaded crayons used for marking parts before heating) Quench cracks Cold-straightening cracks

The likelihood that any of the above discontinuities will appear in a closed-die forging produced in a press or by upsetting is more prevalent than for hammer forgings, because press and upset forgings permit no opportunity to monitor the workpiece being forged during the forging operation. During hammer forging, the top surface of the forging is visible between hammer blows. The bottom surface may also be visible at times--particularly for large forgings that are raised intermittently during forging for descaling and lubricating the bottom surface or for forgings of temperature-sensitive alloys that are raised off the bottom die to permit heat recovery to the bottom surface. Any seam in the forging stock or incipient laps or cracks will probably develop into significant forging laps or cracks if not detected during formation. Consequently, in hammer forging, a large percentage of such potentially scrap forgings can be removed from the production run and can either be salvaged by removing the discontinuity prior to finishing or scrapped at that point to avoid wasted forging time. Also, multiple-cavity hammer forgings permit inspection of the parts and blending out of minor laps or superficial cracks before finish forging. Discontinuities that are not detectable by either magnetic particle or liquid penetrant inspection are noted in the following list. Most of these can be detected by ultrasonic inspection: • •

Flakes, due to the absorption of hydrogen Pipe, due to center shrinkage in the ingot and subsequent insufficient reduction of forging stock

• • • • •

Subsurface nonmetallic segregation Subsurface cracking, which may occur in certain alloys, particularly during the forging of irregular sections Weak centers in forging stock, caused by insufficient reduction from the ingot Subsurface cracks caused by forging material having comparatively cold centers, and generally occurring in large forging billets heated for insufficient time Rewelded forging laps, formed and rewelded during forging. With subsequent hammer blows, the lap forms, the scale is knocked or blown off, and the lapped metal rewelds, forming a healed lap with transverse grain flow and, possibly, entrapped scale

The presence of a rewelded forging lap in a suspected area can be checked by removing the surface metal below the decarburized layer and polishing this surface and swabbing with cold ammonium persulfate, thus revealing the decarburization at both sides of the lap (if present). The condition can be eliminated with corrections in blocker-die design. Ring-Rolled Forgings. Discontinuities in forgings produced by ring rolling may be either inherited from the ingot or

mechanically induced in forging operations, much the same as for forgings produced completely in hammers or presses. Inherited discontinuities are common to all products produced from bar or billet and can usually be traced back to the composition, cleanliness, or condition of the ingot. Although these discontinuities are not found only in ring-rolled forgings, they probably account for the vast majority of known discontinuities in ring-rolled forgings. Typical discontinuities are inclusions, porosity, hot-top remnants, and segregation. Ultrasonic inspection is a reliable method of detecting the presence of inherited discontinuities. It is always advisable to inspect the material before it is ring rolled. Extremely large billets (1200 mm, or 48 in., in diameter and larger) may have surface conditions that cause problems relative to sound entry. Large billets may exhibit structural conditions, depending on the amount of reduction from the ingot, that are too large for a complete ultrasonic inspection. Smaller, well-worked billets can be examined at 2.25 MHz. The larger billets require the use of 1.0-MHz crystals, and even then ultrasonic penetration is not always possible. Final proof that the forging is free of inherited discontinuities is accomplished through ultrasonic inspection of the completed forging. Depending on the final machined shape, certain ring-rolled forgings may require a preliminary ultrasonic inspection before final machining. When an extreme change in contour prevents a complete ultrasonic inspection of the final shape, inspection can be performed on portions of the forged ring. Externally induced mechanical discontinuities that have been found in ring-rolled forgings include surface-related laps, cracks, and exfoliations. Normally, these discontinuities can be detected visually either during the manufacturing process or in the machined condition after rolling. However, ultrasonic inspection can be valuable for determining the depth of surface-related discontinuities and for detecting them even when they have been obscured by subsequent working or metal movement. Magnetic particle inspection is also used to detect surface-related, externally induced mechanical discontinuities in ferromagnetic ring-rolled forgings. Liquid penetrant inspection has often been successfully used to detect surface flaws in nonferromagnetic rings. Mechanically induced internal discontinuities (known as strain-induced porosity) can occur in certain materials. Some nickel or titanium alloys have an inherent susceptibility to these types of discontinuities in portions of the ring stretched at critical temperatures. This may result from improper stock distribution, improper rolling techniques, or improper tooling. In extreme cases, in which the induced porosity is excessive, the rupturing may progress to one or more external surfaces. Either form of internal mechanically induced discontinuity may initiate in an area where inherited discontinuities are present. Effect of Forging Material Some types of forging flaws are unique to specific work metals, and may influence the choice of inspection method. Steel Forgings. The most common surface flaws in steel forgings are seams, laps, and slivers. Other surface flaws

include rolled-in scale, ferrite fingers, fins, overfills, and underfills. The most common internal flaws found in steel forgings are pipe, segregation, nonmetallic inclusions, and stringers.

Either magnetic particle or liquid penetrant inspection can be used for steel forgings, although magnetic particle inspection is usually preferred. Only liquid-penetrant inspection can be used for some stainless steel or nonferrous forgings. The selection of an inspection method depends on the size and shape of the forging and on whether the forging can be moved to the inspection station or the inspection equipment can be moved to the forging. For either inspection method, systems are available for inspecting forgings of almost unlimited size and weight (see articles on the specific inspection methods in this Volume). In most cases, magnetic particle inspection is less expensive and faster than liquid penetrant inspection. Heat-Resistant Alloy Forgings. Most of the flaws found in forgings of heat-resistant alloys can be categorized as those related to scrap selection, melting, or primary conversion to bar or billet or those that occur during forging or heat treatment. Tramp elements such as lead or zinc have been present in the makeup of scrap charges at levels that have caused hot shortness and a degradation of hot-tensile ductility occurring near 370 °C (700 °F) of air-melted alloys.

No NDI method would reliably evaluate the presence or absence of possible tramp-element contamination, and composition checks or hot-tensile checks could be considered to be too random for complete assurance of the presence or absence of contamination. The positive corrective action is the use of a vacuum-remelted product. Melt-related discontinuities, such as inclusions, pipe, unhealed center conditions, flakes, or voids, are the types of discontinuities that most frequently exist in heat-resistant alloy forgings. Ultrasonic inspection can detect and isolate these conditions when they exist. Segregated structures, unmelted electrodes, or portions of stringer rods are types of discontinuities that may sometimes be found in heat-resistant alloys. Flakes (internal cracks) can be produced each time the material is heated and cooled to room temperature. The random orientation does not always present a properly oriented reflector for ultrasonic inspection, but in most cases flaking can be detected ultrasonically with a high degree of reliability. Unmelted pieces of electrode or shelf conditions appear infrequently in vacuum-melted alloys. Either of these conditions can seriously degrade forgeability. Macroetching or another appropriate type of surface inspection of the machined forging or the billet is the most effective method of detecting unmelted electrodes or shelf conditions. Seams are common to rolling practices, and are readily detected by visual, magnetic particle, or liquid penetrant inspection. Grinding cracks are caused by severe grinding, which promotes network-type cracking on the surface of the material being conditioned. The network-type cracking may be present immediately after grinding or may not occur until subsequent heating for further forging. Seams and grinding cracks will cause severe surface rupturing during forging. Center bursts occur during conversion to bar or billet if reduction rates are too severe or temperatures are incorrect; they are readily detected by ultrasonic inspection. Ingot pipe, unhealed center conditions, or voids are melt-related discontinuities, but their occurrence in forgings is often a function of reduction ratio. The conversion practice must impart sufficient homogenization or healing to produce a product with sound center conditions. An example of an unsound condition that did not heal is shown in Fig. 7. Macroetching and ultrasonic inspection methods are the most widely used for identifying regions of unsoundness.

Fig. 7 Section through a heat-resistant alloy forging showing a central discontinuity that resulted from insufficient homogenization during conversion. Step machining was used to reveal the location of the rupture; original diameter is at right.

The nickel-base heat-resistant alloys are highly susceptible to surface contamination during heating for forging. Fuel oils containing sulfur will induce a grain-boundary attack, which will cause subsequent rupturing during forging. Paint or marking crayons with high levels of similar contaminants will cause similar areas of grain-boundary contamination. Surface contamination is not normally detected by NDI methods prior to heating and processing, but if present at a level high enough to cause contamination, rupturing during forging will occur that can be detected by visual inspection. If the contamination occurs after final forging, with no subsequent metal deformation, the contaminated areas will be apparent as areas of intergranular attack. Macroetching, followed by liquid penetrant inspection, should be used. Advanced forging processes, such as isothermal and hot-die forging, and the increasing use of computer modeling have greatly reduced problems associated with heat-resistant alloy forgings. These developments are outlined in Forming and Forging, Volume 14 of ASM Handbook, formerly 9th Edition Metals Handbook. Nickel Alloy Forgings. The discontinuities that occur in nickel alloy forgings are generally of the same type as those

found in heat-resistant alloy forgings; namely, cracks (external and internal), tears, seams, laps, coarse-grain wrinkles, inclusions, and pipe. Although all metals may be subject to thermal cracking during forging, the age-hardenable nickel alloys are more vulnerable than most other metals, thus requiring close temperature control during forging to avoid large temperature gradients. Internal discontinuities in nickel alloy forgings can be located by ultrasonic inspection. Liquid penetrants are most often used to inspect for surface flaws; magnetic particles can be used if the alloy is sufficiently magnetic. Aluminum Alloy Forgings. Common surface discontinuities in aluminum alloy forgings are laps, folds, chops, cracks,

flow-throughs, and suck-ins (Fig. 8(a) and 8(b)). The generation of these discontinuities is associated with the forging operation, processing practices, or design. Cracks can also result from seams in the forging stock.

Fig. 8(a) Typical discontinuities found in aluminum alloy forgings. See text for discussion.

Fig. 8(b) Band of shrinkage cavities and internal cracks in an alloy 7075-T6 forging. The cracks developed from the cavities, which were produced during solidification of the ingot and which remained during forging because of inadequate cropping. Etched with Keller's reagent. 9×

The internal discontinuities that occur in aluminum alloy forgings are ruptures, cracks, inclusions, segregation, and occasionally porosity (Fig. 8(a) and 8(b)). Ruptures and cracks are associated with temperature control during preheating or forging or with excessive reduction during a single forging operation. Cracks can also occur in stock that has been excessively reduced in one operation. Inclusions, segregation, and porosity result from forging stock that contains these types of discontinuities. The inspection of aluminum alloy forgings takes two forms: in-process inspection and final inspection. In-process inspection, using such techniques as statistical process control and/or statistical quality control, is used to determine if the product being manufactured meets critical characteristics and if the forging processes are under control. Final inspection, including mechanical property testing, is used to verify if the completed forging product conforms with all drawing and specification criteria. Typical final inspection procedures used for aluminum alloy forgings include dimensional checks, heat treatment verification, and nondestructive evaluation. Dimensional Inspection. All final forgings are subjected to dimensional verification. For open-die forgings, final dimensional inspection may include verification of all required dimensions on each forging or the use of statistical sampling plans for groups or lots of forgings. For closed-die forgings, conformance of the die cavities to the drawing requirements, a critical element in dimensional control, is accomplished prior to placing the dies in service by using layout inspection of plaster or plastic casts of the cavities. With the availability of computer-aided design (CAD) data

bases on forgings, such layout inspections can be accomplished more expediently with computer-aided manufacturing (CAM) driven equipment, such as coordinate measuring machines or other automated inspection techniques. With verification of die cavity dimensions prior to use, final part dimensional inspection may be limited to verifying the critical dimension controlled by the process (such as die closure) and monitoring the changes in the die cavity. Further, with highdefinition and precision aluminum forgings, CAD data bases and automated inspection equipment, such as coordinate measuring machines and two-dimensional fiber optics, can be used in many cases for actual part dimensional verification. Heat Treatment Verification. Proper heat treatment of aluminum alloy forgings is verified by hardness measurements and, in the case of 7xxx-T7xxx alloys, by eddy current inspection. In addition to these inspections, mechanical property tests are conducted on forgings to verify conformance to specifications. Mechanical property tests vary from the destruction of forgings to tests of extensions and/or prolongations forged integrally with the parts. Nondestructive Inspection. Aluminum alloy forgings are frequently subjected to nondestructive inspection to verify

surface or internal quality. The surface finish of aluminum forgings after forging and caustic cleaning is generally good. A root mean square (rms) surface finish of 3.2 m (125 in.) or better is considered normal for forged and etched aluminum alloys; under closely controlled production conditions, surfaces smoother than 3.2 m (125 in.) rms can be obtained. Selection of NDI requirements depends on the final application of the forging. When required, satisfactory surface quality is verified by liquid penetrant, eddy current, and other techniques. Aluminum alloy forgings used in aerospace applications are frequently inspected for internal quality using ultrasonic inspection techniques. Magnesium alloy forgings are subject to the same types of surface and internal discontinuities as aluminum alloy

forgings. In addition, surface cracks are common in magnesium alloy forgings and are usually caused by insufficient control of the forging temperature. Visual inspection and liquid penetrant inspection are used to detect surface discontinuities. Ultrasonic inspection is used to locate internal discontinuities. Titanium Alloy Forgings. Discontinuities that are most likely to occur in titanium alloy forgings are usually carried

over in the bar or billet. Typical discontinuities in titanium alloy forgings are unsealed center conditions, clean voids, and forging imperfections.

-stabilized voids, macrostructural defects,

Alpha-stabilized voids are among the most common discontinuities found in forgings of titanium alloys. Investigation

and research have determined that voids surrounded by oxygen-stabilized grains may be present in the ingot (Fig. 9). Because of the size of these voids and the coarse-grain nature of the ingot, they cannot be detected until the ingot has been suitably reduced in cross section and refined in structure. When the structure has been refined, the voids can be detected by ultrasonic inspection. Also, when the section is reduced sufficiently, radiographic inspection can be effectively used. Alpha voids do not readily deform during forging, nor do they align with the flow pattern, as do typical inclusions in carbon or alloy steel. In most cases, voids appear to be somewhat globular. Extremely small voids do not present an especially ideal target or reflector for ultrasonic energy. Attempts to correlate size with amplitude of indication obtained during ultrasonic inspections have not been completely reliable. For critical-application forgings, the material is most often inspected twice--once in the bar or billet form before forging and again after forging. Because forging further refines structure and reorients possible discontinuities in relation to the sound-entry surface, the forging operation probably enhances the possibility of detecting these discontinuities. Fig. 9 Ti-8Al-Mo-1V, as forged. Ingot void (black), surrounded by a layer of oxygenstabilized (light). The remaining structure consists of elongated grains in a dark matrix of transformed β. Etched with Kroll's reagent (ASTM 192). 25×

Macrodefects. Three principal defects are commonly found in

macrosections of ingot, forged billet, or other semifinished product forms. These include high-aluminum defects (Type II defects), high-interstitial defects (Type I defects or low-density interstitial defects), and flecks. High-aluminum defects are areas containing an abnormally high amount of aluminum. These are soft areas in the material (Fig. 10) and are also referred to as

segregation. Defects referred to as segregation are sometimes associated with segregation. These are areas in which aluminum is depleted. The high-interstitial defects (Fig. 11) are normally high in oxygen and/or nitrogen, which stabilize the phase. These defects are hard and brittle; they are normally associated with porosity, as illustrated in Fig. 9.

Fig. 10 Ti-6Al-4V 1.25×

-β processed billet illustrating the macroscopic appearance of a high-aluminum defect.

Fig. 11 Macrodefects in titanium billets. Left: Ti-6Al-4V -β processed billet illustrating macroscopic appearance of a high-interstitial defect. Actual size. Right: at 100×. The high oxygen content results in a region of coarser and more brittle oxygen-stabilized than observed in the bulk material.

Beta flecks are regions enriched in a β-stabilizing element due to segregation during ingot solidification. Figure 12 shows the macroscopic appearance of β flecks in a Ti-6Al-6V-2Sn forging billet.

Fig. 12 Ti-6Al-6V-2Sn -β forged billet illustrating macroscopic appearance of β flecks that appear as dark spots. Etched with 8 mL HF, 10 mL HF, 82 mL H2O, then 18 g/L (2.4 oz/gal.) of NH4HF2 in H2O. Less than 1×

Unsealed center conditions are associated with insufficient ingot reduction. These are more prevalent in the larger stock sizes (>230 mm, or 9 in., in diameter) and are normally removed by adequate croppage at the mill. Clean voids describe a condition that can be associated with unsatisfactory sealing of porosity elsewhere in the ingot or through center porosity formed during ingot reduction. Nondestructive Inspection. Ultrasonic inspection is the most definitive and practical method of inspecting titanium

alloy forgings. Inspection techniques are normally tailored to the rejection level indicated in the specifications and to the physical condition of the material being inspected. Surface conditions usually must be ideal, grain size must be fine, and structural conditions must be controlled. Most airframe or similar static parts are inspected with equipment settings based on a No. 3 flat-bottom-hole standard. For the examination of critical rotating forgings for aircraft gas-turbine engines, it is not uncommon to inspect to the equivalent of a No. 1 flat-bottom-hole standard. Experience with these highly critical forgings, which in service rotate at high speed in the presence of extreme temperature and pressure, has indicated that small voids can initiate cracks and have caused catastrophic failures. For satisfactory ultrasonic inspection of forgings to these stringent requirements, special techniques and equipment are usually required. Specially designed ultrasonic electronic equipment is used with focused or otherwise unique transducers. Also required are an immersion tank with rotating devices, automatic small incremental indexing devices, and automatic alarms for signal level. Special reference blocks are required, along with the usual flat-bottom-hole reference blocks. The correct indexing increment must be established, the linear alignment of the ultrasonic unit must be verified, and calibration checks must be made. All information must be recorded and retained for future reference. Nondestructive Inspection of Forgings

Visual Inspection Despite the many sophisticated inspection methods available, unaided visual inspection is still important and is often the sole method of inspecting forgings used for common hardware items. Under proper lighting conditions, the trained eye can detect several types of surface imperfections, including certain laps, folds, and seams. Visual inspection is often used first, then questionable forgings are further examined by macroetching and inspection with macrophotography or some type of nondestructive method. The only equipment necessary for visual inspection is a bench on which to place the forging and suitable cranes or hoists for forgings that are too heavy to lift by hand. Good and well-controlled lighting conditions are essential. Optical aids such as magnifying glasses that can magnify up to about ten diameters are often used to increase the effectiveness of visual inspection. Nondestructive Inspection of Forgings

Magnetic Particle Inspection Magnetic particle inspection is useful for detecting surface imperfections as well as certain subsurface imperfections that are within approximately 3 mm ( in.) of the surfaces in forgings of steel, some grades of stainless steel, and other ferromagnetic metals. Magnetic particle inspection can be used with fluorescent particles and ultraviolet light. Detailed information is available in the article "Magnetic Particle Inspection" in this Volume. The advantages of magnetic particle inspection include the following:

• • • • • • •

•

Almost instant results can be obtained in locating surface and certain subsurface imperfections Equipment can be transported to the forging, or the forging can be transported to the inspection station, as dictated by the size and shape of forging Preparation of the forging is minimal, mainly involving the removal of surface contaminants that would prevent magnetization or inhibit particle mobility Routine inspection work can be effectively done by relatively unskilled labor properly trained in interpretation For forgings that are simple in configuration, and when justified by the quantity, magnetic particle inspection can be automated For some forgings, electronic sensing can be used, thus reducing the chances of human error and increasing inspection reliability Many forgings have sufficient retentivity to permit the use of multidirectional magnetization, thus permitting the inspection of indications in all orientations with a single preparation. Retentivity must be checked for the particular forging before a decision is made to use multidirectional magnetization The cost of magnetic particle inspection is generally lower than that for several other inspection methods in terms of investment in equipment, inspection materials, and inspection time

The limitations of the magnetic particle inspection of forgings are generally the same as for inspecting other

workpieces and include the following: • • • •

•

The method is applicable only to forgings made from ferromagnetic metals Because magnetic particle inspection is basically an aided visual inspection, under most circumstances, its effectiveness is subject to the visual acuity and judgment of the inspector Magnetic particle inspection is generally limited to detecting imperfections that are within about 3 mm ( in.) of the surface of the forging Because the forging must be thoroughly magnetized, magnetic particle inspection is likely to be ineffective unless scale, grease, or other contaminants are removed from the forging. Such surface contaminants inhibit the mobility of the particles necessary to delineate the indications Following inspection, the forging usually must be demagnetized, depending mainly on the retentivity of the particular metal, subsequent shop operations, and end use

Detection of Surface Discontinuities. Magnetic particle inspection and liquid penetrant inspection are both widely

used for detecting discontinuities in steel forgings, although the former is the more widely used. As described above and in the articles that deal with the specific inspection methods in this Volume, one advantage of using the magnetic particle technique is its ability to detect certain subsurface discontinuities that are not open to the surface. Subsurface discontinuities cannot be located with liquid penetrants. Also, some surface discontinuities may be so packed with scale that liquid penetrant techniques are marginal or infeasible. Therefore, in most cases, magnetic particle inspection is preferred to liquid penetrant inspection. Continuous magnetization is usually prescribed for inspecting steel forgings, because at the stage in which the forgings are inspected they are in an annealed or semiannealed condition and consequently have poor retentivity of magnetism. Two inspection methods are available: dry powder and wet. Selection between the dry and the wet methods may sometimes be purely arbitrary, although it is usually based on the available equipment and the size of the forgings being inspected. The dry-powder method is used to a greater extent for large forgings. Similarly, selection between fluorescent and nonfluorescent particles may often be arbitrary, although the size of the forging can be a major factor, because if the fluorescent method is used the forging must usually be of such size and shape that it can be inspected under ultraviolet light, with white light substantially eliminated. Many specific procedures have been established for in-plant use. The dry-powder and wet techniques adopted in one plant for the inspection of ferromagnetic metal forgings are described below. Dry-Powder Technique. The contact method of magnetization was selected to inspect the ferromagnetic materials.

Prods are used to pass direct current or rectified alternating current through the workpiece. The magnetic particles are

nontoxic, finely divided ferromagnetic material of high permeability and low retentivity, free from rust, grease, dirt, or other materials that may interfere with the proper functioning of the magnetic particles. The particles must also exhibit good visual contrast with the forging being inspected. Inspection is by the continuous-current method; that is, the magnetizing current remains on during the period of time that the magnetic particles are being applied and also while the excess particles are being removed. Prods are spaced 150 to 200 mm (6 to 8 in.) apart, except where restricted by configuration. The magnetic field is induced in two directions, 90° apart. The current used is 4 to 5 A/mm (100 to 125 A/in.) of prod spacing and is kept on for a minimum of

s.

Dry magnetic particles are applied uniformly to the surface, using a light dustlike technique. Excess particles are removed by a dry-air current of sufficient force to remove the excess particles without disturbing particles that show indications. The nozzle is held obliquely about 35 to 50 mm (1 to 2 in.) above the test area. Nozzle size and air pressure result in a pressure (measured by a manometer) of 25 to 40 mm (1.0 to 1.5 in.) of water at an axial distance of 25 mm (1 in.) from the nozzle and 7.5 to 15 mm (0.3 to 0.6 in.) of water at 50 mm (2 in.) from the nozzle. A 100 mm (4 in.) grid pattern over the entire forging surface is normally used for evaluation (Fig. 13). The prods are placed 200 mm (8 in.) apart (Fig. 13), except where restricted by the shape of the forging, when using this grid pattern.

Fig. 13 Grid pattern and prod positions used in one plant for the magnetic particle inspection of forgings using prods and the dry-powder technique

Prods are placed on the surface to be tested in the proper position, as shown by position 1 in Fig. 13, and the current is turned on (4 to 5 A/mm, or 100 to 125 A/in., of prod spacing). The powder is applied, the excess particles are removed, and the current is turned off. Inspection is conducted during application of the powder and after removal of the excess particles. The next step is to reposition the prods 90°, as indicated by position 2 in Fig. 13; the above procedure is then repeated. When the shape of the forging does not permit a full 90° rotation with the established prod spacing, the prod spacing can be changed, provided it is not less than 50 mm (2 in.) nor more than 200 mm (8 in.) between prods.

Wet Technique. The magnetic particles selected are nonfluorescent and suspended in a liquid vehicle. The magnetizing

equipment is capable of inducing a magnetic flux of suitable intensity in the desired direction by both the circular and the longitudinal methods. Direct current from generators, storage batteries, or rectifiers is used to induce the magnetic flux. Circular magnetization is obtained by passing the current through the forging being examined or through a central conductor to induce the magnetic flux. Longitudinal magnetization is obtained by using a solenoid, coil, or magnet to induce the magnetic flux. The magnetic particles are nontoxic and exhibit good visual contrast. The viscosity of the suspension vehicle for the particles must be a maximum of 5 × 10-6 m2/s (5.0 centistokes) at any bath temperature used. The magnetic particles are limited to 28 to 40 g (1.0 to 1.4 oz) of solid per gallon of liquid vehicle. The liquid used as a vehicle for the magnetic particles may be a petroleum distillate such as kerosene. Tap water with suitable rust inhibitors and wetting and antifoaming agents can be substituted for the petroleum distillate. The water should contain about 0.3% antifoam agent, 3.9% rust inhibitor, and 12.8% wetting agent. Inspection is carried out by the continuous method. For this method, the magnetizing circuit is closed just before applying the suspension or just before removing the forging from the suspension. The circuit remains closed for approximately

to

s. For circular magnetization, an ammeter is used to verify the presence of adequate field strength. For verifying adequate field strength in longitudinal magnetization, a field indicator is useful. Typical current levels utilized to provide an adequate field strength are 4 to 12 A/mm (100 to 300 A/in.) of diameter of the surface being examined, although current levels of up to 30 A/mm (750 A/in.) of diameter have been used. The magnetizing force for longitudinal magnetization is 2000 to 4000 ampere-turns per 25 mm (1 in.) of diameter of the surface being examined. If both the inside and outside diameters of cylindrical parts are to be inspected, the larger diameter is used in establishing the current. If it is impractical to attain currents of the calculated magnitude, a magnetic field indicator is used to verify the adequacy of the magnetic field. Suspensions must be tested daily or when they appear to have become discolored by oil or contaminated by lint. Common practice is to test the suspension at the beginning of each operating shift. The suspension test is conducted as follows: • • • • • •

Let the pump motor run for several minutes to agitate a normal mixture of particles and vehicle Flow the bath mixture through the hose and nozzle for a few minutes to clear the hose Fill the centrifuge tube to the 100-mL line Place the centrifuge tube and stand in a location free from vibration Let the tube stand for 30 min for particles to settle out After 30 min, readings for settled particles should be 1.7 to 2.4 mL. If the reading is higher, add vehicle; if lower, add particle powder to the suspension

Example 1: Magnetic Particle Inspection of 1541 Steel Connecting Rods by the Multidirectional System. Connecting rods for piston engines were manufactured from 1541 steel by forging in closed dies. Because connecting rods used in piston engines are subject to considerable stress, it is necessary to ensure the integrity of these parts, especially in heavy-duty applications, such as high-performance automobiles and trucks. Accordingly, fluorescent magnetic particle inspection of the forged connecting rod was done immediately after unwanted scale and dirt were removed by shotblasting. Bidirectional (longitudinal and transverse) magnetic fields were generated rapidly and sequentially by the use of a multidirectional magnetizing machine, as shown in Fig. 14. The use of this machine allowed for a single inspection to detect all discontinuities regardless of their orientation.

Fig. 14 Setup for the inspection of forged 1541 steel automotive connecting rods by the magnetic particle method using a multidirectional magnetizing system

The process of manufacturing connecting rods usually results in only longitudinally oriented flaws, but, transversely oriented flaws occasionally do occur. Hot tears along the trim line and metal folds or laps in the web area are of primary interest in the inspection. Worn tooling, overheating, and underheating of the forging blanks, as well as improper trimming, are the primary causes of these flaws. Inspection Procedure. The inspection of these connecting rods utilized a multidirectional system, and the inspection procedure was as follows. Initially, the connecting rod was pneumatically clamped in the contact heads (Fig. 14) and then automatically flooded with a suspension of fluorescent magnetic particles. The magnetizing fields were induced while the suspension was draining from the part. Current settings for circular magnetization were 1000 A; for longitudinal magnetization, 4500 ampere-turns equivalent. The magnetizing current automatically switched back and forth, thus revealing any imperfections regardless of their orientation. Inspection of the forging was done under an ultraviolet light of no less than 2700 lx (250 ftc) at the surface of the forging. Following completion of inspection, the forging was demagnetized so that metal chips did not adhere to the surface during subsequent machining operations. The principal disadvantage of this inspection procedure was that the production rate was limited because the procedure was performed manually.

Example 2: Use of Electromagnetic, Ultrasonic, and Magnetic Particle Methods for the Inspection of Forged 1046 Steel Steering Knuckles. Automobile steering knuckles similar to the forged 1046 steel knuckle shown in Fig. 15 are regarded as extremely highintegrity pans. At one plant, to provide the required degree of quality control, 100% of all steering knuckles were nondestructively inspected using three different techniques--electromagnetic comparison for hardness and ultrasonics for internal soundness at one location, followed by in-line fluorescent magnetic particle inspection.

Fig. 15 Steering knuckle, forged from 1046 steel, that was inspected by a combination of electromagnetic, ultrasonic, and magnetic particle methods. The setup for electromagnetic and ultrasonic inspection is shown. Dimensions given in inches

Inspection Procedures. Figure 15 shows the orientation of the steering knuckle during inspection with the electromagnetic test coil and ultrasonic transducer. Electromagnetic comparator coils were used to evaluate hardness in the bearing area; the ultrasonic 60° shear wave method was used to detect internal discontinuities in the spindle portion of the steering knuckle. The spindle of the knuckle was immersed in a dielectric oil to provide couplant for the ultrasonic inspection; a 5.0-MHz transducer was used. For electromagnetic inspection, a 60-Hz frequency was used for the hardness detection comparator. Electromagnetic and ultrasonic inspections were combined both for efficiency and compatibility of the two methods.

All steering knuckles were fluorescent magnetic particle inspected following the electromagnetic-ultrasonic inspection. This inspection was accomplished by magnetizing the steering knuckle between contact plates, flooding with the particle suspension, and viewing under an ultraviolet light having an intensity of 2700 lx (250 ftc) minimum. The steering knuckles were then transported by conveyor to a magnetizing coil for magnetizing in the other direction so that circumferentially oriented indications would be shown under a subsequent second ultraviolet light inspection. These

nondestructive tests resulted in the detection of both internal and external flaws as well as the detection of variations in hardness. Nondestructive Inspection of Forgings

Liquid Penetrant Inspection Liquid penetrant inspection is a versatile NDI process and is widely used for locating surface imperfections in all types of forgings, either ferrous or nonferrous, although it is more frequently used on nonferrous forgings. There is no limitation on the size or shape of a forging that can be liquid penetrant inspected. Any of the three basic liquid penetrant systems (water-washable, postemulsifiable, and solvent-removable) can be used to inspect forgings. The product or product form is not a principal factor in the selection of a system. The fundamentals of the three systems, as well as the advantages and limitations of each for inspecting various products (including forgings), are presented in the article "Liquid Penetrant Inspection" in this Volume. Advantages. Among the advantages of liquid penetrant inspection of forgings are the following:

• • • • • •

There are no limitations on metal composition or heat-treated condition There are no limitations imposed on the size or shape of the forging that can be inspected Liquid penetrant inspection can be done with relatively simple equipment Training requirements for inspectors are minimal Inspection can be performed at any stage of manufacture Liquid penetrant materials can be taken to the forgings or the forgings taken to the inspection station, depending on the size and shape of the forgings

The limitations of the liquid penetrant inspection of forgings are basically the same as those for the inspection of other

workpieces. The characteristics of the surface of a forging sometimes impose specific limitations. The most important general limitations are: • • •

•

Liquid penetrant inspection is restricted to detecting discontinuities that are open to the surface Liquid penetrant inspection is basically a visual aid; therefore, results depend greatly on the visual acuity and judgment of the inspector Satisfactory inspection results require that the surface of the forging be thoroughly cleaned before inspection. The presence of surface scale can cause inaccurate readouts. If the surface of the forging is excessively scaled, it should be pickled or grit-blasted, preferably pickled. The forgings should also be cleaned to remove surface contaminants, such as grease and oil Liquid penetrant inspection is slower than magnetic particle inspection

Nondestructive Inspection of Forgings

Liquid Penetrant Detection of Flaws in Steel Forgings Factors affecting the selection of a special penetrant system for inspecting steel forgings include available equipment; size, shape, and surface conditions of the forgings; degree of sensitivity required; whether or not the entire forging requires inspection; and cost. Regardless of which system is used, the degree of success achieved depends greatly on the surface conditions of the forging. Rough, scaly surfaces are likely to result in either false indications or obscuring of meaningful flaws. Pickled surfaces are preferred. Abrasive blasting is usually satisfactory for cleaning forging surfaces,

although overblasting must be avoided or some flaws may be tightly closed and prevent the penetrant from entering. The postemulsifiable and solvent-removable liquid penetrant systems are most often used to inspect steel forgings. The postemulsifiable system is generally preferred to the water-washable system for forgings because of its greater sensitivity. Either the fluorescent-penetrant or the visible-dye technique can be used. Selection depends largely on whether ultraviolet-light inspection can be used. For forgings of a size and shape that can be immersed in tanks and inspected in a booth, the fluorescent technique is usually preferred. The solvent-removable system is especially well adapted to applications in which only a portion of the forging

requires inspection. Equipment for this system can be minimal and completely portable or may involve more elaborate systems used on a production basis, as described in the following examples.

Example 3: Solvent-Removable Liquid Penetrant Inspection of Type 316 Stainless Steel Forged Valve Bodies. Figure 16 illustrates the equipment layout used for the liquid penetrant inspection of 150 and 200 mm (6 and 8 in.) type 316 stainless steel forged valve bodies by the solvent-removable visible-dye system. This same equipment layout was also used for many other sizes and types of forgings requiring liquid penetrant inspection, such as the bonnet forgings described in the next example. The roller conveyor shown in Fig. 16 provided easy handling of relatively large forgings and, at the same time, allowed for the variations in time cycles that were compatible with the various sizes and shapes of forgings.

Fig. 16 Equipment layout for the liquid penetrant inspection of type 316 stainless steel forged valve bodies and bonnet forgings by the solvent-removable visible-dye system

Inspection Procedure. The processing cycle that proved satisfactory for inspection of the forged valve bodies was as

follows: •

• •

Preclean by gritblasting, followed by a light etch in a 30% HNO3, 10% HF, and 60% H2O (by volume) acid solution. Forgings were also ground and polished as required so that the surface roughness was 6 μm (250 μin.) maximum Vapor degrease (wiping with 1,1,1-trichloroethane is an acceptable alternative) Apply solvent-removable visible-dye penetrant with a spray gun at a pressure of approximately 140 kPa (20 psig). Workpiece temperature should be kept within the temperature range of 10 to 40 °C (50 to 100

• • • • • •

°F) Allow dwell time (penetration) of 15 to 20 min Rough wipe with lint-free cloth or absorbent paper Finish wipe with cloth or absorbent paper moistened with solvent remover (same group as the penetrant) Air dry for 5 to 10 min Apply nonaqueous developer, using a spray gun or pressurized spray cans, and allow drying time of 7 to 30 min prior to inspection Inspect in a booth using adequate light intensity, although at a level that will avoid producing a glare from the white surface

Acceptable workpieces were then cleaned by flushing with tap water, wiping with 1,1,1-trichloroethane, and air drying. They were then stamped and moved to the next processing operation. Forgings that failed to meet acceptance standards were marked by encircling the flaw with a crayonlike marker that did not contaminate the surface. The flaws were then ground and polished, and the forging was reinspected in accordance with the above procedure.

Example 4: Solvent-Removable Liquid Penetrant Inspection of Type 316 Stainless Steel Bonnet Forgings. The solvent-removable visible-dye system of liquid penetrant inspection and the equipment shown in Fig. 16 were used to detect surface imperfections in 340 kg (750 lb) type 316 stainless steel bonnet forgings (Fig. 17).

Fig. 17 Bonnet forging of type 316 stainless steel that was liquid penetrant inspected for surface flaws by the solvent-removable visible-dye system. Dimensions given in inches

Precleaning. Forged bonnets were delivered to the penetrant inspection operation in the machined condition. Machined

workpieces were lightly etched in a 30% HNO3, 10% HF, and 60% H2O (by volume) acid solution and thoroughly water rinsed (spray or immersion) to remove all traces of acid. Workpieces were then swabbed with a 5% ammonia solution and subsequently rerinsed in water (spray or immersion). They were then allowed to dry at room temperature. Immediately preceding application of the penetrant, the etched forgings were pre-cleaned with 1,1,1-trichloroethane and allowed to dry at room temperature for at least 5 min. Inspection Procedure. The forgings were inspected in accordance with the following procedure:

•

• • • • •

• • •

Apply solvent-removable visible-dye penetrant (group 1, MIL-I-25135) to the surfaces of the workpieces with a spray gun attached to the shop air line (brushing is an acceptable alternative). The penetrant need not be applied to the inside bore surfaces (as indicated in Fig. 17), because these surfaces are to be machined and subsequently inspected. Temperature of the workpiece for this operation as well as for subsequent operations is maintained within 15 to 50 °C (60 to 120 °F) Allow dwell time of 15 to 25 min Remove excess penetrant, using a solvent remover of the same family as the penetrant, by wiping with lint-free cloths or absorbent paper Final wipe with lint-free cloth or absorbent paper dampened slightly with solvent remover Dry at room temperature for 5 to 10 min Apply nonaqueous developer with a spray gun, pressurized spray can, or airbrush. Agitate developer periodically during application. Apply with short dusting strokes so that a uniform but thin layer is applied Perform inspection no sooner than 7 min and no later than 30 min following developer application Inspect, using adequate white light, but at a level that will avoid glare Clean inspected forgings by wiping with clean cloths moistened with 1,1,1-trichloroethane

Indications were evaluated in accordance with specifications listed as follows: •

Linear indications: None longer than 4.2 mm (

•

Rounded indications: None greater than 4.8 mm (

•

Linearly disposed rounded indications: Four or more in line separated by 1.6 mm ( in.) or less, measured edge to edge, are evaluated as linear indications Multiple indications: No more than ten within any area of 40 cm2 (6 in.2) and whose major dimension is no more than 150 mm (6 in.)

•

in.) in.) in diameter

Repair and Reinspection. Forgings containing flaws in excess of the above acceptance standards were repaired by

grinding or polishing, provided these operations did not impair dimensional requirements. Forgings that were repaired by grinding or polishing were subject to reinspection using the technique and acceptance standards outlined above. Nondestructive Inspection of Forgings

Liquid Penetrant Detection of Flaws in Heat-Resistant Alloy Forgings Because most heat-resistant alloy forgings are nonmagnetic, the use of magnetic particle inspection for detecting surface flaws cannot be considered. Liquid penetrants are extensively used for inspecting the surfaces of high-integrity forgings. Critical forgings such as these require close quality control surveillance. Following production penetrant inspection using group VI fluorescent penetrant, it is desirable to conduct quality control overchecks on samples selected from previously inspected batches. These overchecks are performed using highly sensitive, high-resolution penetrants, as described in the following examples.

Example 5: Routing Liquid Penetrant Inspection of Jet-Engine Cooling Plates Forged From René 95. Cooling plates (Fig. 18) used in the rear of several jet engines were forged from René 95, which is a nickel-base heatresistant alloy.

Fig. 18 Forged René 95 cooling plate for a jet engine that was first inspected by the postemulsifiable fluorescent penetrant system and then reinspected by a high-resolution penetrant after being accepted. Dimensions given in inches

Inspection Procedures. Forgings of this type were inspected in accordance with type 1, method B (MIL-I-6866), which is a postemulsifiable fluorescent penetrant inspection using group VI (MIL-I-25135) penetrants, emulsifiers, and dry developers. The following procedure was used:

• • • • • • • • • • • •

Clean machined forgings with solvent (1,1,1-trichloroethane) by immersion, swabbing, or wiping Lightly etch each forging with an acid mixture of 80% HCl, 13% HNO3, and 7% HF (by volume) Rinse (immerse or spray) in tap water to remove all traces of acid Air dry prior to penetrant application. Inspect visually for adequacy of etch and to ascertain that forgings are dry Using a basket-type container, immerse several forgings at a time into group VI (MIL-I-25135) fluorescent penetrant. Allow to soak in penetrant or rest on drain table for at least 30 min Immerse container in emulsifier (group VI, MIL-I-25135), or hose on emulsifier, taking care to completely cover each forging completely. Allow emulsification time of 2 min maximum Wash off excessive emulsified penetrant, using water at 40 °C (100 °F) maximum. Forgings can be immersed in agitated water wash or can be cleaned by a water spray located at the wash station Verify adequacy of removal of excess penetrant by use of ultraviolet light located at the wash station Dry in oven at 60 to 80 °C (140 to 180 °F) for 7 min maximum Visually inspect forging for adequacy of drying cycle Apply dry developer over entire surface of forging, using compressed air with a dry developerapplicator hose Allow development time of 10 to 15 min

Following development, the forgings were inspected in a darkened booth using a portable ultraviolet light (intensity: 1060 W/cm2, or 6840 W/in.2). Unacceptable indications were recorded; acceptance standards were as follows: • •

No linear indications are permitted Scattered nonlinear indications are permitted if they are not visible at one diameter in white light (1075 1x, or 100 ftc, at surface subject to inspection), if they are separated by at least 3 mm ( there are no more than ten in any 25 mm (1 in.) diam area

in.), and if

All rejected forgings were scrapped or salvaged. Forgings that were salvaged were subject to reinspection in accordance with the above acceptance standards. Following inspection, the forgings were cleaned by solvent or water (spray or immersion) to remove retained developer. They were then dried at room temperature; a warm air blast was used to speed up the drying process.

Example 6: Quality Assurance Liquid Penetrant Inspection of Jet-Engine Cooling Plates Forged From René 95. Because of the critical application for the forged jet-engine cooling plate described in Example 5 and illustrated in Fig. 18, and because of the type of flaws expected (very tight forging cracks at the periphery and excessive porosity), sample forgings were periodically selected at random from lots that had been accepted after being inspected by the procedure described in Example 5. The sample forgings were subjected to liquid penetrant inspection using a commercially available high-resolution fluorescent penetrant to evaluate the adequacy of each cycle used in the production penetrant inspection described in Example 5. In particular, the high-resolution penetrant was used to determine the effectiveness of salvage grinding and polishing operations utilized to remove unacceptable indications detected in the original inspection. It is important to recognize that the high-resolution penetrants are not normally adaptable to production liquid penetrant inspection. The following procedure was employed when conducting the quality assurance inspection: •

•

•

• • •

•

Thoroughly clean the surface of the forging, using high-resolution precleaner. When necessary, before precleaning, lightly etch with an 80% HCl, 13% HNO3, and 7% HF (by volume) acid solution; rinse; and dry. Allow precleaner to dry for 4 to 5 min Apply high-resolution fluorescent penetrant over entire surface of the forging, using a fine brush or the dauber attached to the cap of the can containing the penetrant. Allow penetrant to dry completely (approximate drying time, 1 min) Soak a piece of cheesecloth or equivalent material with high-resolution penetrant remover and remove the dried penetrant from the surface of the forging, using a circular motion. The penetrant remover will liquefy excess surface penetrant. Remove all traces of excess penetrant by scrubbing with a clean, lintfree cloth under a flood of tap water at room temperature. Finally, wipe surfaces with a clean, lint-free cloth that has been dampened with tap water at room temperature Allow surface to dry thoroughly. Drying may be aided by a blast of clean air or by means of a fan Apply nonaqueous developer from agitated spray can, in light, uniform dusting strokes. Hold can approximately 150 to 250 mm (6 to 10 in.) from the surface of the forging Allow a minimum of 10 min to elapse before inspection and evaluation of indications. Evaluate indications under ultraviolet light (intensity: 1060 W/cm2, or 6840 W/in.2) in a darkened inspection booth. Acceptance standards are the same as those listed in Example 5 Postclean, using either the high-resolution precleaner or a water wash, and then dry at room temperature

Flaws in excess of acceptance standards detected with the high-resolution penetrants determined the disposition of the lot or modifications to the production penetrant cycles and appropriate action taken. Nondestructive Inspection of Forgings

Ultrasonic Inspection Ultrasonic inspection is used to detect both large and small internal flaws in forgings. Detailed information on the fundamentals of this method can be found in the article "Ultrasonic Inspection" in this Volume. Forgings, by their nature, are amenable to ultrasonic inspection. Both longitudinal or shear wave (straight or angle beam) techniques are utilized. The size, orientation, location, and distribution of flaws influence the selection of technique and the inspection results. Consider, for example, Fig. 19, which shows the influence of flaw orientation on signal response. There are, however, some definite limitations. All ultrasonic systems currently in use generate sound electrically and transmit the energy through a transducer to the forging. Because the relationship of sound transmitted to sound received is a factor in the inspectability of a forging, particular attention must be given to the surface condition of the forging. Although techniques and couplants can enhance the energy transmission from the transducer to the forging, as-forged surfaces impair the effectiveness of ultrasonic inspection. Near-surface flaws are most difficult to detect, and a dead zone* at the entry surface often interferes. Because of the difficulty involved in detecting surface flaws by ultrasonic inspection, another method, such as magnetic particle or liquid penetrant inspection, is often used in conjunction with ultrasonic inspection to inspect high-integrity forgings thoroughly.

Fig. 19 Comparisons of discontinuities at normal orientation versus radial orientation. (a) Discontinuity normal to sound energy, almost total reflection of energy back toward transducer. (b) Discontinuity at 0° with respect to 45° sound energy path, almost no energy reflected back toward transducer

Complex shapes are difficult to inspect ultrasonically because of the problems associated with sound-entry angle. Most ultrasonic inspection of forgings uses techniques that send waves into the forging perpendicular to the surface. Radii, fillets, and similar configurations must receive special treatment if all areas of the forging must be inspected. The special treatment involves the use of a standoff that has an end contoured to fit the inspection surface or the use of a smalldiameter or focused transducer. Application. In certain cases, where the end use of a forging is considered critical, ultrasonic inspection is used to

inspect the wrought material before it is worked. Surface or internal flaws that are not detected before a billet is forged may not be detected in the final forging and will therefore be present in the finished part. Ultrasonic inspection is often used as part of a completely diagnostic inspection of a forging from newly designed dies, where use of the finished part does not warrant inspection of every part. Quality control measures often include the ultrasonic inspection of random samples from a particular forging. This provides the necessary assurance that the process is under control and that variables affecting internal quality have not been inadvertently introduced. Ultrasonic inspection is often used in the further evaluation of flaws detected by other nondestructive methods. This reduces the possibility that a particular forging will be unsuitable for its intended service. Ultrasonic inspection can be used on every forging to validate its integrity for extremely rigorous requirements. This applies in particular to forgings for nuclear and aerospace applications, where rigid standards of acceptance have been established. Standards and criteria have been set up to detect material inclusions, internal voids, laminations, and other conditions. In addition, the inspection of every forging by ultrasonics has been effective in detecting excessive grain size and other structural conditions. Ultrasonic inspection is often used to qualify a particular lot of forgings that has been subjected to certain variations in approved processing procedures. A notable instance is the use of ultrasonics to determine the presence of thermal flakes or in locating quench cracks.

Note cited in this section

* The term dead zone refers to the depth below the entry surface where any echoes from flaws cannot be detected because they are masked by the trailing edge of the transmitted pulse. The dead zone depends on

sound velocity, test frequency, and pulse length and is different for different combinations of test metal, transducer, and test instrument. Nondestructive Inspection of Forgings

Basic Procedures for Ultrasonic Inspection Many specific procedures have been developed for the inspection of forgings by ultrasonics. The basic procedures described here are based on those developed for use in a specific plant. Forgings of regular (symmetrical) shape are ultrasonically inspected at final inspection (prior to shipment). If possible, the entire forging is ultrasonically inspected. Irregularly shaped forgings, which have a configuration that precludes complete inspection prior to shipment, are ultrasonically examined at the latest stage of processing that will permit examination of the entire forging (preliminary inspection). These complex shapes are then reexamined to the extent practical before shipping (final inspection). To ensure complete inspection coverage of the forging, the transducer is indexed with at least 15% overlap with each pass. The rate of scanning does not exceed 150 mm/s (6 in./s). The surface roughness of forgings from which ultrasonic examination is conducted should be no greater than 6 μm (250 μin.), unless otherwise specified. The forging surfaces should also be free of extraneous material, such as loose scale, paint, dirt, or other contaminants. Equipment An ultrasonic, pulsed, reflection-type instrument is used to inspect forgings. The instrument is capable of operating at three or more frequencies; for example, 1.0, 2.25, and 5.0 MHz. The instrument provides a linear presentation (within ±5%) to approximately 75% of the screen height or a minimum of 40 mm (1 in.). Signal attenuators and calibrated gain control, if used, are accurate over the range used to ±10% of the nominal attenuation ratio. Forgings are examined by the longitudinal wave technique at 2.25 MHz unless adequate penetration cannot be obtained, in which case the inspection is conducted at 1.0 MHz. Longitudinal Wave Ultrasonic Inspection Where possible, all forgings are scanned in two directions perpendicular to each other. Disk (pancake) forgings are examined axially from at least one end face and radially from the circumference, whenever practical. Ring forgings are examined radially from at least the outside surface and axially from at least one end face. Cylindrical forgings are examined radially from the circumference. The axial examination of cylindrical forgings can be conducted when practical, using the back-reflection method of longitudinal wave inspection. Back-Reflection Method. Longitudinal wave inspection by back reflection is often performed with a 2.25-MHz

transducer. A higher or lower frequency can be used, when necessary, to obtain a meaningful inspection result. The sensitivity for longitudinal wave calibration is established by adjusting the instrument controls to obtain a 40 mm (1 in.) back-reflection height from the opposite side of the forging over an area free of indications (this level corresponds to approximately 75% of the useful screen height of the instrument). Calibration must be reestablished for each change of thickness greater than 10%. The calibration level and minimum-back-reflection requirements as outlined in this section apply to areas with parallel surfaces. Routine scanning is done at the calibration gain level. When specified, the instrument gain level is increased to at least two times the calibration gain level. Flaw amplitudes and back-reflection reduction are evaluated with the gain setting at

the calibration level of 40 mm (1 indications.

in.) back-reflection height in an indication-free area as close as practical to the area of

All areas of the forging are examined with sufficient gain to provide a minimum back-reflection amplitude of 40 mm (1 in.), except for areas where reduction of the back reflection is associated with ultrasonic indications. Recording of indications detected by back-reflection method is done as follows:

• •

•

•

Any indication having an amplitude that equals or exceeds 10% of the calibration back-reflection level is reported in increments of 10 Any traveling indication equal to or exceeding 5% of the calibration back reflection is recorded. Traveling indications are defined as indications whose leading edge moves a distance equivalent to 25 mm (1 in.) or more of metal depth with movement of the transducer over the surface of the forging. Any cluster of indications equal to or exceeding 5% of the calibration back reflection is recorded. A cluster of indications is defined as five or more indications located in a volume representing a 50 mm (2 in.) or smaller cube in the forging Any continuous indication equal to or exceeding 5% of the calibration back reflection is recorded. A continuous indication is defined as an indication that can be held on the same plane over a square area whose sides are greater than twice the diameter of the search unit Any reduction in back reflection (associated with a discontinuity indication) exceeding 20% of the calibration back reflection, measured in increments of 10%, is recorded

Distance-Amplitude-Correction (DAC) Curve Method. The following refers to the primary test direction, which is the radial direction for hollow and shaft-type forgings and the axial direction for upset disk-type forgings. Test sensitivity is established with reference flat-bottom-hole standards machined into either the forging itself or separate reference blocks. If separate reference blocks are used for instrument calibration, they should be of material having an attenuation coefficient within ±25% of forged production material. A DAC curve is established for longitudinal wave inspection of the forging in the primary test direction through the use of reference blocks of various thicknesses and hole sizes, as explained in the article "Ultrasonic Inspection" in this Volume. The size of the flat-bottom hole and the metal distance from entry surface to the hole bottom depend on the maximum machined thickness in the primary test direction at the time of initial ultrasonic inspection.

The following method is used to prove that the calibration blocks have attenuation characteristics within 25% of the production forging. An acoustic-decay pattern (the initial three back reflections) is determined for the calibration-block material while that material is still part of a larger section, so that there is no side-effect cancellation or reinforcement on the exponential decay pattern. An acoustic-decay pattern is also determined for the production forging at the midlength, using the same type of instrument and transducer as used for determining the acoustic-decay pattern of the calibrationblock material. The amplitude of the first back reflection should be the same for both tests and within the vertical linearity of the instrument. The calibration-block material, which is fabricated in accordance with either ASTM E 428 or E 127, is considered compatible with the production forging if the sum of the amplitudes of three successive back reflections from the production forging is within 25% of the sum of the amplitudes of three successive back reflections from the calibration-block material or if the slope of the plotted back-reflection amplitudes of the calibration-block material is within 25% of the slope of the plotted back-reflection amplitudes of the production forging at the same sound-travel distance. Recording of indications detected by the DAC-curve method is determined by the types and magnitudes of

indications. Those recorded are any indication whose amplitude equals or exceeds 100% of the established DAC curve; traveling, clustered or continuous indications (as defined above) that have amplitudes equal to or greater than 50% of the DAC curve; and any reduction in back reflection (associated with a discontinuity indication) exceeding 20% of the calibration back reflection, measured in increments of 10%. Calibration: Primary Test Direction. At the time of initial inspection, the areas of greatest and least acoustical

penetrability are determined by scanning selected passes, which are representative of the forging cross section, at a fixed instrument gain level. The amplitude of the first back reflection is monitored during these scans to determine the areas of greatest and least penetrability.

The amplitude of the reflection from the flat-bottom hole in the longest calibration block is set at 13 mm (0.5 in.) sweep to peak, and the other points of the DAC curve are obtained at this gain setting. This curve can be marked on the face of the screen or plotted on graph paper. This is the fixed DAC curve that will be used to evaluate the indications. Any portion of the DAC curve above the vertical linearity limits of the instrument is brought on screen with the attenuator for evaluation of the indications in this area. After establishing the fixed DAC curve, the amplitude of the back reflection over the area of greatest acoustical penetrability is determined at this instrument gain level. The transducer is then coupled to the forging over the area of least penetrability, and the gain setting of the instrument is increased to bring the back reflection to the same amplitude as the amplitude obtained over the area of greatest acoustical penetrability (a minimum of 40 mm, or 1 For scanning, the instrument gain level is increased a minimum of twice the final gain setting.

in., sweep to peak).

Any indication, detected while scanning, whose amplitude equals or exceeds the reporting level is evaluated. This evaluation is conducted by readjusting the back reflection in a discontinuity-free area of the forging as close as practical to the area of indication to equal the same height of back reflection obtained in the area of greatest acoustic penetrability when the DAC curve was established. The search unit is repositioned over the area of the indication, and the evaluation to the DAC curve is made. Back-reflection reduction is evaluated by adjusting the first back reflection to the maximum back reflection within the vertical linearity limits of the instrument in a discontinuity-free area of the forging as close as practical to the area of the indication to be evaluated. The search unit is repositioned over the area containing the indication, and the amplitude is recorded. For contoured forgings at final inspection, ultrasonic inspection in the primary test direction is performed using a DAC curve established from the same reference blocks used for preliminary inspection. The last point of the curve is 13 mm (0.5 in.) sweep to peak, and the other points of the DAC curve are obtained at that gain setting. This is the fixed DAC curve that will be used to evaluate any indications. Compensation for the areas of greatest and least penetrability does not apply at final inspection with a contoured forging. Calibration: Secondary Test Direction. The secondary test direction is approximately at 90° to the primary

direction. The transducer is coupled to the face of the forging, and the amplitude of the first back reflection is set within the vertical linearity limit of the instrument (minimum 40 mm, or 1 in., sweep to peak). The DAC curve is established at this gain setting, using the same blocks that were used for the longitudinal wave inspection in the primary test direction. The last point of the DAC curve is extended parallel to the sweep line to midlength plus 25 mm (1 in.). Any indications will be evaluated against this DAC curve. When the configuration of the forging at final inspection due to contour machining prevents ultrasonic examination in the axial direction to midlength plus 25 mm (1 in.), the DAC curve will be extended parallel to the sweep line to the deepest location, wherever possible, for evaluation of the indications. Shear Wave Ultrasonic Inspection Shear wave inspection is performed from the circumference of rings and hollow forgings that have axial length greater than 50 mm (2 in.) and a ratio of outside to inside diameters of less than 2:1. A 1.0-MHz, 45° angle-beam search unit is used when shear wave inspecting forgings having a ratio of outside to inside diameters of less than 1.4:1. Forgings with a ratio of outside to inside diameters of 1.4:1 or greater, but less than 2:1, will be shear wave tested at a beam angle of less than 45°. Shear wave inspection is performed by scanning in both circumferential directions (clockwise and counterclockwise) along the periphery of the forging. Inside diameter and outside diameter calibration notches are cut axially in the surface of the forging or a similar testpiece (preferably, excess metal or test metal). The sides of the notches should be smooth and parallel to the axis of the forging. The calibration notches are 25 mm (1 in.) long, V-shaped or rectangular, with a width not exceeding twice the depth. These notches should have a depth of 3% of the maximum section thickness (based on thickness at initial ultrasonic inspection) or 6 mm ( in.), whichever is smaller. If at final inspection it is necessary to place the calibration notches in the forging, these will have to be weld repaired.

The sensitivity for angle-beam inspection of the forging is established by adjusting the instrument controls to obtain a minimum 13 mm (0.5 in.) sweep-to-peak indication from the outside diameter calibration notch. The response from the inside diameter calibration notch is then obtained at this sensitivity level. The peaks of the indications from the inside diameter and outside diameter notches are connected to establish a reference line. Gain is increased two times for scanning. For contoured forgings at final inspection, the inspection is conducted to the extent that contour permits a meaningful examination. Recalibration Any realignment of the transducer with respect to the material or any change in the search unit, couplant, or instrument settings from that used for calibration necessitates recalibration. A calibration check is performed at least once every 8-h shift. If a 15% or greater decrease in the sensitivity level is observed during a calibration check, then the required calibration is reestablished, and all material examined in the preceding calibration period is reexamined. Nondestructive Inspection of Forgings

Ultrasonic Inspection of Specific Forgings** The procedure used for the ultrasonic inspection of a given forging depends greatly on the size and shape of the forging, and sometimes more importantly on the degree of inspection that is required. For example, a dome-shaped forging such as the low-alloy steel forging shown in Fig. 20 has a symmetrical shape that readily permits inspection by either the immersion or the contact ultrasonic method with no great problems other than the handling equipment needed.

Fig. 20 Symmetrical, dome-shaped forging that was ultrasonically inspected by the longitudinal wave technique on the inside and outside of the dome and on the ends, and by the circumferential shear technique on the outside of the dome and the neck portion. Dimensions given in inches

The forging illustrated in Fig. 20 was inspected by using the longitudinal wave technique on the inside diameter, outside diameter, and ends of the forging, plus the circumferential shear technique on the outside diameter. In addition, circumferential shear was conducted on the outside diameter of the neck portion of the forging. The landing-gear forging illustrated in Fig. 21 has a more intricate shape and is large, which makes it more difficult to inspect. For forgings such as this, the best procedure is thorough inspection of the billets prior to forging. A common practice is to ultrasonically inspect the billet material to an agreed-on specification. This minimizes the amount of inspection required after forging. A forging such as the one shown in Fig. 21 would be ultrasonically inspected by immersion, longitudinal wave, hand-scan techniques in all areas where the shape permits. The following examples describe procedures used for the ultrasonic inspection of a variety of steel and aluminum alloy forgings.

Fig. 21 Landing-gear forging typical of complex forgings that are preferably ultrasonically inspected in billet form because the intricate forged shape prevents thorough ultrasonic inspection

Example 7: Ultrasonic Inspection of Forged Medium-Carbon Steel Axle Shafts for the Detection of Chevrons and Burning. Automotive rear-axle shafts (Fig. 22) are made from medium-carbon steels by first forging in closed dies to form the hub end and then by cold extrusion to form the shaft section. Cold-extruded rear-axle drive shafts are prone to develop serious internal flaws known as chevrons or cup-cone ruptures. Chevrons are internal ruptures that may or may not extend to the surface of the extruded stock (most often they do not extend to the surface). There may be any number of chevrons produced during the extrusion of a single axle shaft. Therefore, it is necessary to inspect the entire length of each shaft.

Fig. 22 Setup and operating essentials for the ultrasonic inspection of the shaft portion of a forged mediumcarbon steel axle shaft

Inspection Procedure. To perform a 100% ultrasonic inspection of these axle shafts in a plant where production was

2 to 3 million units annually, an automated system was used. The test station was supplied by a conveyor. Figure 22 illustrates the position of the axle shaft when the ultrasonic test was performed. A transducer holder, with a fixed angle fabricated of hardened tool steel with no adjustment, coupled the ultrasound from the transducer to the shaft end. A detailed view of the transducer holder and V-block is shown in detail B in Fig. 22. Transducers with a frequency of 1.6 to 2.25 MHz were used to perform the ultrasonic inspection. The transducer incident angle was approximately 20°, which yielded a 45° shear wave generated within the shaft. This permitted the detection of internal flaws. The oscilloscope display in detail A in Fig. 22 illustrates the interaction of an ultrasonic beam with a chevron. Two electronic gates were utilized to monitor two separate areas of the shaft, the area from where the sound enters the shaft to

approximately 150 mm (6 in.) from the end of the flange. This was a positive gate, commonly called the chevron channel. The second gate was positioned to monitor the back-reflection signal and was a negative gate. The acceptance of any axle shaft required a strong back reflection with no signal in the chevron channel gate. In addition to detecting chevrons, this inspection procedure detected the results of any burning that had occurred during heating. A burned forging attenuated the back reflection and caused the forging to be rejected.

Example 8: Ultrasonic Inspection of a 1.7 m (68 in.) Diam 8822 Steel Billet and the Ring Into Which It Was Rolled. The large steel ring shown in Fig. 23 was produced for a nuclear application from a 1.7 in (68 in.) diam billet of 8822 steel. The large billet was first forged into a doughnut preform in a large press, then formed into the ring by ring rolling.

Fig. 23 Large rolled ring that was ultrasonically inspected as a 1.7 m (68 in.) diam billet and as a rolled ring

A ring of this type requires a high degree of integrity; therefore, inspections were performed at various stages of manufacture. Routine ultrasonic inspection of the billet revealed a gross center condition in approximately 67% of the length. Inspection was performed from the outside diameter and end faces to the maximum extent possible. On billets of this size, however, the surface condition and the lack of refinement of the internal structure have an adverse effect on the ability to perform a thorough inspection by ultrasonics. When the billet was preformed into the doughnut shape, the portion that had been the center of the billet was removed. The doughnut was then ring rolled into a large ring having a wall 460 mm (18 in.) thick (Fig. 23). It was hoped that the area indicated as defective in the billet stage would have been removed, but because it was uncertain whether all of the defective area had been removed, the ring was machined on the outside diameter and on one end face to a surface finish of 6 m (250 in.), thus permitting a more sensitive inspection. Primary consideration was given to performing a straight-beam ultrasonic inspection from the outside surface in the radial direction. During ring rolling, included foreign materials become aligned in the circumferential direction to provide a

lamellar-type discontinuity. This type of discontinuity is most readily detected from the perpendicular direction or, in this case, either from the outside or inside surfaces of the ring. Because the original discontinuities in the billet were in the center and would remain in a similar position during forging, it was anticipated that if remnants of the center condition were still present, they would be near or adjacent to the inside or bore of the ring. Therefore, a radial ultrasonic inspection from the outside using the straight-beam technique was considered to be the best method of detection. Inspection Procedure. A straight-beam inspection was instituted as indicated, and calibration was established using the back-surface-reflection method to determine whether adequate ultrasonic penetration was available. A 25 mm (1 in.) diam, 2.25-MHz transducer was used in conjunction with the ultrasonic unit.

The outside surface was scanned in overlapping paths to ensure complete coverage. During the inspection, areas of indications were noted at approximately midheight and adjacent to the bore area; one area approximately 7 m (24 ft) long in a circumferential direction was encountered during the inspection; several smaller spot-type areas were also noted. Evaluation of indications on the basis of applicable American Society of Mechanical Engineers (ASME) Boiler and Pressure Vessel Codes revealed that the indications would be acceptable. However, for closer evaluation, a straight-beam inspection from the face and a circumferential angle-beam inspection, in two 180° opposing directions, were performed. Straight-beam inspection from the face did not reveal indications at a normal gain or sensitivity level; they could be detected only at an extremely high sensitivity level. Angle-beam inspections showed very small low-level (approximately 10% maximum) occasional indications in the area previously noted. Although not required by the ASME code, an axial angle-beam inspection from the outside was performed, mainly in the area of indications. Indications were readily detectable, but were not considered serious enough to reject the forging. On the basis of results of the ultrasonic inspections, the ring was considered acceptable within the required specifications. Magnetic particle inspection revealed a few small indications in the areas tested. Because severity did not appear to be indicative of the ultrasonic results, the area was conditioned by grinding and polishing to obtain an additional inspection at a greater depth from the inside surface. Subsequent magnetic particle inspection indicated a much more severe condition, which was not acceptable. Conclusions. Metallographic investigation showed that the indications could be classified as areas of chemical

segregation and nonmetallic inclusions (predominantly sulfides). These were evidently present in the center of the billet and remained in the final ring. Therefore, the ring was considered unsatisfactory for the application. Removal of the questionable metal and repair by welding was considered, but the cost would have been excessive. Replacement of the defective ring from an acceptable billet was the most economical solution.

Example 9: Ultrasonic Inspection of a 4340 Steel Shaft Forging. The pulse-echo technique of ultrasonic inspection was selected for locating flakes, bursts, and pipe in large 4340 steel shaft forgings such as the one shown in Fig. 24.

Fig. 24 Transducer locations for radial and longitudinal ultrasonic inspection of a large steel shaft forging for

detection of flakes, bursts, and pipe. Dimensions given in inches

Equipment Specified. The ultrasonic instrument specified was a pulsed, reflection-type unit capable of testing at 1 to

5 MHz. It was also required that the instrument provide linear presentation (within ±5%) for at least 75% of the screen height (sweep line to top of screen). It was further specified that the electronic apparatus contain a signal attenuator (accurate over its useful range to ±10% of the nominal attenuation ratio), which allows measurement of signals beyond the linear range of the instrument. Transducers 25 to 28 mm (1 to 1

in.) in diameter or 25 mm (1 in.) square were used.

Couplants having good wetting characteristics, such as SAE No. 20 or No. 30 motor oil, glycerin, pine oil, water, or a nonionic water-soluble polymer, were specified. (The above couplants, however, may not be comparable to one another, and the same couplant must be used for calibration and recording of results.) Test blocks containing flat-bottom holes were used for calibrating equipment in accordance with the above apparatus requirements. Such test blocks were used to establish recording levels for longitudinal testing. Surface Preparation. It was required that all surfaces be machined so that surface roughness did not exceed 6

m (250 n.) and that surfaces contacted by transducers should be free from all extraneous material, such as scale, dirt, and paint, before the couplant was applied. Inspection Procedure. Ultrasonic inspection was performed after heat treatment, but prior to drilling holes, cutting

keyways, or otherwise machining to a contour. The entire forging was inspected using a test frequency of 2.25 MHz. Search patterns that incorporated 15% overlap between passes were scanned at a speed of less than 150 mm/s (6 in./s). Scanning from two mutually perpendicular directions (radial and longitudinal) provided complete coverage. Referring to Fig. 24, when indications were found with the transducer oriented for radial scanning (transducer location 1), the operator moved the transducer to location 2, on the end of the forging, to determine the size and axial location of the flaw. In the forging shown in Fig. 24, pipe that was 100 mm (4 in.) long by 25 mm (1 in.) thick was found in the center of the cross section and 915 mm (36 in.) from one end of the forging. Acceptance Standards. Forgings showing a single flaw indication greater than that from a 13 mm (

in.) diam flatbottom hole were not acceptable. Forgings containing clustered or traveling indications greater than 50% of that from a 13

mm ( in.) diam flat-bottom hole were not acceptable. Forgings containing a complete loss of back reflection not attributable to size, shape, or grain size were not acceptable; loss of back reflection not accompanied by an indication from a flaw was further investigated by the use of other search units. Complete loss of back reflection was assumed when the back reflection fell to less than 5% of the reference back reflection.

Example 10: Ultrasonic Inspection of an Upset-Forged 4118 Steel Shaft. Field failures were encountered in the upset portion of forged 4118 steel shafts (Fig. 25). The failures were traced to internal cracks in the upset flange area that were initiated from gross nonmetallics. These nonmetallics, which were originally located near the center of the cross section, were pushed outward into an almost radial plane in the upsetting of the flange. When the shaft was heat treated, these planes of inclusions developed internal cracks.

Fig. 25 Upset shaft forging, of 4118 steel, that was ultrasonically inspected by the pulse-echo longitudinal wave technique. Dimensions given in inches

The location of these cracks, their orientation, and the size of the shaft all indicated that ultrasonics would be the best method for detection. Two approaches were feasible. The first was to send a longitudinal beam in from the end of the shaft, and the second was to use a shear wave from the outside diameter of the shaft adjacent to the flange. The first technique was selected because it permitted testing of shafts already assembled. A 22 mm (

in.) diam, 2.25-MHz transducer was used with a plastic wear face and an oil couplant. The shaft had a

radially drilled oil hole 9 mm ( in.) in diameter (note location in Fig. 25). This radial hole was used as a landmark and standard to set the sensitivity. The signal from this hole was set to give full-screen height from a known flaw-free shaft. The end of the shaft was scanned in a zone extending from the center out 40 mm (1 in.) radially. The area around the longitudinal oil hole within the upset section was the area where the flaws occurred (Fig. 25). When the shaft was tested from the short end (measured from the upset), a portion of the test area was in the shadow of the radial oil hole. In addition, because of the distance to the radial hole, the second multiple of the radial-hole signal occurred at the same point on the screen as did the flaws. For this reason and because of variation in flaw orientation, it was desirable to test the shaft from the long end as well as from the short end. In the assembled machines, the only access was to the long end of the shaft. The same standard was used from the long end of the shaft, but an instrument having higher sensitivity was required to produce a full-screen signal from the radial hole because of the added length of beam path. The rejection level was set at 20% of full screen and was based on the size of flaws observed when the shafts were cut up. The flange portion was first cut out of the shaft and retested with a smaller-diameter transducer to better define the extent of the flaw. The section was further cut until one edge of the flaw was exposed. When possible, this reduced section was broken open through the flaw. The fracture face provided the best view of the extent and shape of the flaws. If the inclusion flaws were larger than 20 mm ( in.) in diameter, they were considered rejectable even if they had not cracked in heat treatment. An attempt was made to test the shafts in the as-forged condition, but this approach was not reliable because some inclusions developed cracks during heat treatment. Similar but slightly larger shafts were found to have the same types of flaws. These shafts were tested with the same technique as described above and initially showed a very high rate of rejection. No flaws were found when the rejected shafts were sectioned, and when the shafts were retested, no flaws were found. After some investigation, it was realized that the flaw signals were false. When a portion of the beam struck the oily surface of the longitudinal oil hole, a modeconverted wave was reflected back to the search unit. The time of travel of the mode-converted false signal coincided with the time of travel of the true flaw signals, causing rejection of the shafts. The problem was solved by removing the oil film from the longitudinal oil hole.

Example 11: Ultrasonic Inspection of B-58 Nose Landing Gear Trunnion. The development of this inspection focused on the nose landing gear trunnions (Fig. 26). Initiation sites for stresscorrosion cracking were found to originate on the trunnion-bore inside surfaces at the forging parting plane (Fig. 27). The trunnion is made from forged aluminum alloy 7075-T6. After examination of previous failures, it was shown that these cracks may propagate to a severe condition before the trunnion fails. A manual shear wave ultrasonic inspection was developed that concentrated on these initiation sites.

Fig. 26 Location of nose landing gear trunnion (arrows) on a B-58 aircraft. Courtesy of Tom Dusz, The University of Dayton Research Institute

Fig. 27 Location of parting line on trunnion. (a) Top view of parting line. (b) Bottom view of parting line. Courtesy of Tom Dusz, The University of Dayton Research Institute

Inspection Procedure. The equipment needed for this inspection includes a flaw detection instrument capable of a

test frequency of 5 MHz; a 5-MHz search unit with a special lens contoured to 45 mm (1.75 in.) cylindrical radius; a reference standard containing artificial notches 0.25, 0.50, and 1.25 mm (0.010, 0.020, and 0.050 in.) deep; and a clean oil

couplant. The reference standard is manufactured from the outer bore section of an actual trunnion. The sensitivity calibration is performed by applying sufficient couplant to the reference standard and by positioning the search unit over the 1.25 mm (0.050 in.) deep notch. A full-scale amplitude signal is obtained. The standard is scanned, and the signals are observed from the 0.25 and 0.50 mm (0.010 and 0.020 in.) deep notches. The instrument sweep delay is adjusted until the first reflected signals from the artificial defects reach their peak amplitude at the center of the screen. The sensitivity, pulse length, and reject controls are used to obtain a 95% amplitude signal from the 0.25 mm (0.010 in.) deep notch. The part is cleaned with an acceptable solvent to remove dirt, grease, tar, loose paint, and other debris. The forging parting plane and two planes parallel and 40 mm (1.5 in.) on either side of the forging parting plane are marked with ink that will not come off the part during the inspection. The areas of concern include the entire area between the parting line and the parallel lines, the bolt hole to the web on top, and the ground area on the bottom. A defect signal first appears near the center of the screen and moves to the left as the search unit is scanned closer to the parting line. When the search unit is scanned from the parting line back toward the parallel line, the defect signal moves from the left of the screen (near the front surface noise) to the center of the screen. Figure 28 shows some typical amplitude signals that may appear on the screen. Any defect indications showing a 25% and greater amplitude signal are considered for further disassembly evaluation and/or replacement.

Fig. 28 Typical signals that may appear during inspection. Dimensions given in inches. Courtesy of Tom Dusz, The University of Dayton Research Institute

Conclusion. The development of this testing technique enabled inspectors to locate very gross or severe conditions before trunnion failure occurred. This greatly reduced the amount of on-aircraft failures.

Example 12: Ultrasonic Inspection by the Immersion Technique of an Aluminum Alloy Forging Having No Parallel Surfaces. The aluminum alloy forging shown in Fig. 29(a) was ultrasonically inspected by the immersion (water) technique using a hand-held search unit. Details of the inspection conditions are given in the table accompanying Fig. 29.

Inspection technique

Immersion, hand-held unit

Mode of sound

Longitudinal wave

Frequency, MHz

5

Cable length, m (ft)

6 (20)

Water distance, mm (in.)

75 (3)

Stand-off tube, mm (in.) 25 (1) ID, 30 (1

) OD

Fig. 29 Aluminum alloy forging (a) having no parallel sides that was ultrasonically inspected by the immersion technique, using the transducer assembly and stand-off tube shown in (b) and the conditions listed in the accompanying table

Reference standards were used for calibrating the oscilloscope that corresponded to full- and half-metal thickness of the thickest section of the forging and for a metal thickness of 6 mm ( the inspection class were used in the calibration standards.

in.). Equivalent flat-bottom-hole sizes required for

The inspection was performed to determine whether internal discontinuities of any type were present. For the, type of forging shown in Fig. 29(a), because of the continuously changing metal thicknesses and surface contours, immersion testing using a hand-held search unit was the most suitable method of ultrasonic inspection. This inspection procedure is called scrubbing. The search unit consisted of a transducer fitted with a hollow plastic tube, called a stand-off tube, to maintain a constant water distance of 75 mm (3 in.) from the entry surface (see transducer assembly, Fig. 29b). The cable and scrubber housing were attached to the scrubber cap. This portion of the unit was wrapped with watertight rubber tape because the unit is partially or totally submerged during use. The stand-off tube, while submerged, was completely filled with water, with no air gaps or bubbles present. During inspection of the forging, the operator had to be aware of the metal thickness under examination at any particular moment. As a result of nonparallel surfaces, the screen presentation usually did not have a directly reflected first back reflection, because the back surface was not normal to the ultrasound beam path. There was a series of lesser-amplitude back reflections resulting from the sound wave bouncing around within the forging and ultimately being picked up by the search unit. These back reflections moved back and forth for short distances along the trace line. The operator had a good knowledge of the changing metal thicknesses and was able to determine the presence of a discontinuity relative to where an indication appeared on the trace line. Substantiation of a discontinuity was accomplished by inspecting the area from the opposite surface and establishing its depth relative to total metal thickness.

Note cited in this section

** Example 11was prepared by Carol Miller, Wright Research & Development Center, Wright-Patterson Air Force Base. The research was conducted by Automation Industries, Inc., under Air Force contract. Nondestructive Inspection of Forgings

Eddy Current and Electromagnetic Inspection In the nondestructive inspection of forgings, the eddy current method is commonly used to detect flaws, while the electromagnetic method is used to detect differences in microstructure, chemical composition, or hardness. Electromagnetic inspection, which is restricted to ferromagnetic materials, is sometimes categorized as a modification of the eddy current method because both techniques are based on electromagnetic principles. In concept, eddy current equipment functions by the introduction of relatively high-frequency alternating currents into the surface areas of conductive materials. The response of the material to the induced field is then measured by a mechanism sensitive to the induced field. Detailed information is available in the article "Eddy Current Inspection" in this Volume.

Detection of Flaws The detection of flaws in forgings by eddy current inspection is almost always done with a system consisting of a single probe that is connected to an instrument generically known as a defectometer. This system is balanced with the probe in air and is further balanced to a null value on sound material of the same composition, heat treatment, and surface condition as the forging to be inspected. Areas of the surface of the forging where flaws are suspected are scanned with the probe, which searches for an unbalance due to the flaw. Generally, the scanning is done in two directions approximately at right angles to each other. The advantages of eddy current inspection for flaw detection include the following:

•

• •

•

The unbalance level can be adjusted and calibrated with notches of known depth in the same material in the same condition, which can give a reasonable estimate of the depth of the flaw. This estimate can be helpful in reaching a decision as to the serviceability of the forging In the event that flaws are oriented in one direction only, as with seams or rolling laps in the original stock, the technique can be automated Threshold gates, which are automatic signal-monitoring networks, are available to automatically signal flaws of a sufficient magnitude to be judged defective. These signals can in turn be used to mark or otherwise identify flawed forgings or locations on the forging having flaws The automated tester can be operated by unskilled personnel once the system has been calibrated. Solidstate electronics and their adaptation to eddy current equipment permit very stable instrumentation with no need for constant adjustment

The disadvantages of eddy current inspection for detection of flaws include the following:

•

• •

The correlation between unbalance signal and flaw depth is often not linear, for a variety of reasons. Accordingly, this method is frequently used as a go/no-go device at a depth of flaw, plus or minus a band of uncertainty. The band of uncertainty must be determined by experimental methods In ferrous materials, variations in the decarburization levels can render the method invalid This method is not suited to the detection of deep subsurface flaws

Detection of Differences in Microstructure Differences in microstructure, which usually register as differences in hardness, can be detected by electromagnetic inspection using either encircling coils or spot probes as pickups. Regardless of whether a probe or a coil is used, the instrumentation must be set up and balanced in accordance with the manufacturer's directions. This is done without forgings in or near the pickup, with forgings in the pickup, or sometimes both with and without forgings. Once the test setup has been established, it is necessary to have good-quality forgings of known electromagnetic response available to ensure that the instrumentation has not varied. These forgings (which are sometimes referred to as masters) must be available to personnel who are trained in the setup and maintenance of the equipment. The advantages of electromagnetic inspection for detecting differences in microstructure include the following:

• • •

The equipment can be electronically gated based on the response of the instrumentation to the properties of the forging Electromagnetic inspection can be readily automated for properties throughout a forging (usually an encircling coil) or for properties at a specific location (using a spot probe) Once it has been properly set up, the operation can be effectively run by unskilled personnel

The disadvantages of electromagnetic inspection for detecting differences in microstructure include the following:

•

•

A given response can indicate more than one condition in the forging; for this reason, testing technique must be developed very carefully. For example, as illustrated in Fig. 30 by an electromagnetic S-curve for a medium-carbon steel forging, when two variables (hardness and microstructure) exist, the curve shows essentially the same height in power loss for 37 HRC as for 56.5 HRC (dashed line B-A, in Fig. 30) even though the microstructure is widely different Development of techniques can be done only by trained personnel; even then, a great deal of experimentation is usually required to develop procedures that will yield accurate results

Fig. 30 Relation of an electromagnetic S-curve to variations in hardness and microstructure of a mediumcarbon steel forging

Nondestructive Inspection of Forgings

Radiographic Inspection Radiography (γ-ray or x-ray) is not extensively used for the inspection of forgings for two reasons. First, the types of discontinuities most commonly located by radiography (gas porosity, shrinkage porosity, and shrinkage cavities) are not usually found in forgings. Second, for the types of internal discontinuities that are commonly found in forgings (inclusions, pipe, bursts, or flakes), ultrasonic inspection is more effective, more adaptable, and more economical. Radiographic techniques can sometimes be helpful in the further investigation of known internal discontinuities in forgings when the presence of these discontinuities has been determined earlier by ultrasonic inspection. In sections that are not too thick to penetrate with available radiographic equipment, the size, orientation, and possibly the type of discontinuities can be evaluated by radiography. Nondestructive Inspection of Castings By the ASM Committee on Nondestructive Inspection of Castings*

Introduction

INSPECTION PROCEDURES for castings are established at the foundry to ensure conformance with customer drawings and documents, which are frequently based on various government, technical society, or commercial specifications. For a foundry to ensure casting quality, inspection procedures must be efficiently directed toward the prevention of imperfections, the detection of unsatisfactory trends, and the conservation of material--all of which ultimately lead to reduction in costs. Inspectors should be able to assess on sight the probable strong and weak points of a casting and know where weaknesses and faults would most likely be found. The inspection of castings normally involves checking for shape and dimensions, coupled with aided and unaided visual inspection for external discontinuities and surface quality. Chemical analyses and tests for mechanical properties are supplemented by various forms of nondestructive inspection, including leak testing and proof loading, all of which are used to evaluate the soundness of the casting. These inspections add to the cost of the product; therefore, the initial consideration must be to determine the amount of inspection needed to maintain adequate control over quality. In some cases, this may require full inspection of each individual casting, but in other cases sampling procedures may be sufficient.

Note

* Lawrence E. Smiley, Reliable Castings Corporation; Bruce G. Isaacson, Bio-Imaging Research Inc.; R.A. Armistead, Advanced Research and Applications Corporation; I.C.H. Hughes, BCIRA International Centre for Cast Metals Technology (Great Britain); John Johnston, Krautkramer Branson; Carol Miller, Wright Research & Development Center, Wright-Patterson Air Force Base Nondestructive Inspection of Castings By the ASM Committee on Nondestructive Inspection of Castings*

Inspection Categories Methods for Determining Surface Quality. Cracks and other imperfections at the surface of a casting can be detected by a number of inspection techniques, including visual inspection, chemical etching, liquid penetrant inspection, eddy current inspection, and magnetic particle inspection (which can also reveal discontinuities situated immediately below the surface). All these inspection methods require clean and relatively smooth surfaces for effective results. Methods for Detecting Internal Discontinuities. The principal nondestructive methods used for detecting internal

discontinuities in castings are radiography, ultrasonic inspection, and eddy current inspection. Of these methods, radiography is the most highly developed technique for detailed inspection; it can provide a pictorial representation of the form and extent of many types of internal discontinuities. Ultrasonic inspection, which is less universally applicable, can give qualitative indications of many discontinuities. It is especially useful in the inspection of castings of fairly simple design, for which the signal pattern can be most reliably interpreted. Ultrasonic inspection can also be used to determine the shape of graphite particles in cast iron. Eddy current and other closely related electromagnetic methods are used to sort castings for variations in composition, surface hardness, and structure. Infrared thermography (thermal inspection) has also occasionally been proposed as a method for detecting subsurface defects. However, its successful uses have generally been restricted to the detection of larger defects because of the relatively slow rates at which heat can be put into a component and because of the relatively low sensitivity of infrared detectors. Increased use of thermal inspection may occur with the introduction of pulsed video thermography, in which a very short burst of intense heat is directed at the component. The presence of near-surface defects influences the rate at which heat is dissipated from the surface, and temperature variations are detected with a high-resolution infrared camera recorded onto videotape and presented as an image on a TV monitor. The method was developed for the detection of small defects in composites and in aerospace turbine engine blades, but some initial results obtained with cast iron test plates have proved promising (Ref 1). Methods for Dimensional Inspection. A number of techniques are used to determine the dimensional accuracy of

castings. These include manual checks with micrometers, manual and automatic gages, coordinate measuring machines,

and three-dimensional automatic inspection stations (machine vision systems). This article will discuss the use of coordinate-measuring machines. Additional information on methods for dimensional inspection can be found in the Section "Inspection Equipment and Techniques" in this Volume.

Reference cited in this section

1. P.J. Rickards, Progress in Guaranteeing Quality Through Nondestructive Methods of Evaluation, Foundryman Int., April 1988, p 196-209 Nondestructive Inspection of Castings By the ASM Committee on Nondestructive Inspection of Castings*

Casting Defects Foundrymen refer to the deviations in less-than-perfect castings as discontinuities, but these imperfections are more commonly termed casting defects. Some casting defects may have no influence on the function or the service life of cast components, but will give an unsatisfactory appearance or will make further processing, such as machining, more costly. Many such defects can be easily corrected by shotblast cleaning or grinding. Other defects that may be more difficult to remove can be acceptable in some locations. The casting designer must understand the differences and must write specifications that meet the true design needs. Classification of Casting Defects. Foundrymen have traditionally used rather unique names, such as rattail, scab,

buckle, snotter, and shut, to describe various casting imperfections (such terms are defined in the "Glossary of Terms" in Casting, Volume 15 of ASM Handbook, formerly 9th Edition Metals Handbook). Unfortunately, foundrymen may use different nomenclature to describe the same defect. The International Committee of Foundry Technical Associations has standardized the nomenclature, starting with the identification of seven basic categories of casting defects: • • • • • • •

Metallic projections Cavities Discontinuities Defects Incomplete casting Incorrect dimension Inclusions or structural anomalies

In this scheme, the term discontinuity has the specific meaning of a planar separation of the metal, that is, a crack. Table 1 presents some of the common defects in each category. In general, defects that can serve as stress raisers or crack promoters are the most serious. These include preexisting cracks, internal voids, and nonmetallic inclusions. Table 1 International classification of common casting defects No.

Description

Metallic Projections

A 100:

Metallic projections in the form of fins or flash

Common name

Sketch

A 110:

Metallic projections in the form of fins (or flash) without change in principal casting dimensions

A 111

Thin fins (or flash) at the parting line or at core prints

Joint flash or fins

A 112

Projections in the form of veins on the casting surface

Veining or finning

A 113

Network of projections on the surface of die castings

Heat-checked die

A 114(a)

Thin projection parallel to a casting surface, in re-entrant angles

Fillet scab

A 115

Thin metallic projection located at a re-entrant angle and dividing the angle in two parts

Fillet vein

A 120:

Metallic projections in the form of fins with changes in principal casting dimensions

A 123(a)

Formation of fins in planes related to direction of mold assembly (precision casting with waste pattern); principal casting dimensions change

A 200:

Massive projections

A 210:

Swells

Cracked or broken mold

A 212(a)

Excess metal in the vicinity of the gate or beneath the sprue

Erosion, cut, or wash

A 213(a)

Metal projections in the form of elongated areas in the direction of mold assembly

Crush

A 220:

Projections with rough surfaces

A 221(a)

Projections with rough surfaces on the cope surface of the casting

Mold drop or sticker

A 222(a)

Projections with rough surfaces on the drag surface of the casting (massive projections)

Raised core or mold element cutoff

A 223(a)

Projections with rough surfaces on the drag surface of the casting (in dispersed areas)

Raised sand

A 224(a)

Projections with rough surfaces on other parts of the casting

Mold drop

A 225(a)

Projections with rough surfaces over extensive areas of the casting

Corner scab

A 226(a)

Projections with rough surfaces in an area formed by a core

Broken or crushed core

Cavities

B 100:

Cavities with generally rounded, smooth walls perceptible to the naked eye (blowholes, pinholes)

B 110:

Class B 100 cavities internal to the casting, not extending to the surface, discernible only by special methods, machining, or fracture of the casting

B 111(a)

Internal, rounded cavities, usually smooth-walled, of varied size, isolated or grouped irregularly in all areas of the casting

Blowholes, pinholes

B 112(a)

As above, but limited to the vicinity of metallic pieces placed in the mold (chills, inserts, chaplets, etc.)

Blowholes, adjacent to inserts, chills, chaplets, etc.

B 113(a)

Like B 111, but accompanied by slag inclusions (G 122)

Slag blowholes

B 120:

Class B 100 cavities located at or near the casting surface, largely exposed or at least connected with the exterior

B 121(a)

Exposed cavities of various sizes, isolated or grouped, usually at or near the surface, with shiny walls

Surface or subsurface blowholes

B 122(a)

Exposed cavities, in re-entrant angles of the casting, often extending deeply within

Corner blowholes, draws

B 123

Fine porosity (cavities) at the casting surface, appearing over more or less extended areas

Surface pinholes

B 124(a)

Small, narrow cavities in the form of cracks, appearing on the faces or along edges, generally only after machining

Dispersed shrinkage

B 200:

Cavities with generally rough walls, shrinkage

B 210:

Open cavity of Class B 200, sometimes penetrating deeply into the casting

B 211(a)

Open, funnel-shaped cavity; wall usually covered with dendrites

Open or shrinkage

B 212(a)

Open, sharp-edged cavity in fillets of thick castings or at gate locations

Corner shrinkage

B 213(a)

Open cavity extending from a core

Core shrinkage

B 220:

Class B 200 cavity located completely internal to the casting

B 221(a)

Internal, irregularly shaped cavity; wall often dendritic

Internal shrinkage

B 222(a)

Internal cavity or porous area along central axis

Centerline shrinkage

B 300:

Porous structures caused by numerous small cavities

B 310:

Cavities according to B 300, scarcely perceptible to the naked eye

external

or

or

or

fillet

blind

axial

B 311(a)

Dispersed, spongy dendritic shrinkage within walls of casting; barely perceptible to the naked eye

Macroor microshrinkage, shrinkage porosity, leakers

Discontinuities

C 100:

Discontinuities, generally at intersections, caused by mechanical effects (rupture)

C 110:

Normal cracking

C 111(a)

Normal fracture appearance, sometimes with adjacent indentation marks

C 120:

Cracking with oxidation

C 121(a)

Fracture surface oxidized completely around edges

C 200:

Discontinuities caused by internal tension and restraints to contraction (cracks and tears)

C 210:

Cold cracking or tearing

C 211(a)

Discontinuities with squared edges in areas susceptible to tensile stresses during cooling; surface not oxidized

C 220:

Hot cracking and tearing

C 221(a)

Irregularly shaped discontinuities in areas susceptible to tension; oxidized fracture surface showing dendritic pattern

Hot tearing

C 222(a)

Rupture after complete solidification, either during cooling or heat treatment

Quench cracking

Breakage (cold)

Hot cracking

Cold tearing

C 300:

Discontinuities caused by lack of fusion (cold shuts); edges generally rounded, indicating poor contact between various metal streams during filling of the mold

C 310:

Lack of complete fusion in the last portion of the casting to fill

C 311(a)

Complete or partial separation of casting wall, often in a vertical plane

C 320:

Lack of fusion between two parts of casting

C 321(a)

Separation of the casting in a horizontal plane

C 330:

Lack of fusion around chaplets, internal chills, and inserts

C 331(a)

Local discontinuity in vicinity of metallic insert

C 400:

Discontinuities caused by metallurgical defects

C 410:

Separation along grain boundaries

C 411(a)

Separation along grain boundaries of primary crystallization

Conchoidal or candy" fracture

C 412(a)

Network of cracks over entire cross section

Intergranular corrosion

Cold shut or cold lap

Interrupted pour

Chaplet or insert cold shut, unfused chaplet

"rock

Defective Surface

D 100:

Casting surface irregularities

D 110:

Fold markings on the skin of the casting

D 111

Fold markings over rather large areas of the casting

Surface folds, gas runs

D 112

Surface shows a network of jagged folds or wrinkles (ductile iron)

Cope defect, elephant skin, laps

D 113

Wavy fold markings without discontinuities; edges of folds at same level, casting surface is smooth

Seams or scars

D 114

Casting surface markings showing direction of liquid metal flow (light alloys)

Flow marks

D 120:

Surface roughness

D 121

Depth of surface roughness is approximately that of the dimensions of the sand grains

Rough casting surface

D 122

Depth of surface roughness is greater than that of the sand grain dimensions

Severe roughness, high pressure molding defect

D 130:

Grooves on the casting surface

D 131

Grooves of various lengths, often branched, with smooth bottoms and edges

Buckle

D 132

Grooves up to 5.1 mm (0.2 in.) in depth, one edge forming a fold which more or less completely covers the groove

Rat tail

D 133

Irregularly distributed depressions of various dimensions extending over the casting surface, usually along the path of metal flow (cast steel)

Flow marks, crow's feet

D 134

Casting surface entirely pitted or pock-marked

Orange peel, metal mold reaction, alligator skin

D 135

Grooves and roughness in the vicinity of re-entrant angles on die castings

Soldering, die erosion

D 140:

Depressions in the casting surface

D 141

Casting surface depressions in the vicinity of a hot spot

Sink marks, draw or suck-in

D 142

Small, superficial cavities in the form of droplets of shallow spots, generally gray-green in color

Slag inclusions

D 200:

Serious surface defects

D 210:

Deep indentation of the casting surface

D 211

Deep indentation, often over large area of drag half of casting

D 220:

Adherence of sand, more or less vitrified

D 221

Sand layer strongly adhering to the casting surface

Burn on

D 222

Very adherent layer of partially fused sand

Burn in

D 223

Conglomeration of strongly adhering sand and metal at the hottest points of the casting (re-entrant angles and cores)

Metal penetration

D 224

Fragment of mold material embedded in casting surface

Dip coat spall, scab

D 230:

Plate-like metallic projections with rough surfaces, usually parallel to casting surface

Push-up, clamp-off

D 231(a)

Plate-like metallic projections with rough surfaces parallel to casting surface; removable by burr or chisel

Scabs, expansion scabs

D 232(a)

As above, but impossible to eliminate except by machining or grinding

Cope spall, boil scab, erosion scab

D 233(a)

Flat, metallic projections on the casting where mold or core washes or dressings are used

Blacking scab

D 240:

Oxides adhering after heat treatment (annealing, tempering, malleablizing) by decarburization

D 241

Adherence of oxide after annealing

Oxide scale

D 242

Adherence of ore after malleablizing (white heart malleable

Adherent material

D 243

Scaling after anneal

Scaling

Incomplete Casting

E 100:

Missing portion of casting (no fracture)

E 110:

Superficial variations from pattern shape

scab,

wash

packing

E 111

Casting is essentially complete except for more or less rounded edges and corners

Misrun

E 112

Deformed edges or contours due to poor mold repair or careless application of wash coatings

Defective coating (teardropping) or poor mold repair

E 120:

Serious variations from pattern shape

E 121

Casting incomplete due to premature solidification

Misrun

E 122

Casting incomplete due to insufficient metal poured

Poured short

E 123

Casting incomplete due to loss of metal from mold after pouring

Runout

E 124

Significant lack of material due to excessive shot-blasting

Excessive cleaning

E 125

Casting partially melted or seriously deformed during annealing

Fusion or melting during heat treatment

E 200:

Missing portion of casting (with fracture)

E 210:

Fractured casting

E 211

Casting broken, large piece missing; fractured surface not oxidized

E 220:

Piece broken from casting

Fractured casting

E 221

Fracture dimensions correspond to those of gates, vents, etc.

E 230:

Fractured casting with oxidized fracture

E 231

Fracture appearance indicates exposure to oxidation while hot

Broken casting (at gate, riser, or vent)

Early shakeout

Incorrect Dimensions or Shape

F 100:

Incorrect dimensions; correct shape

F 110:

All casting dimensions incorrect

F 111

All casting dimensions incorrect in the same proportions

F 120:

Certain casting dimensions incorrect

F 121

Distance too great between extended projections

Hindered contraction

F 122

Certain dimensions inexact

Irregular contraction

F 123

Dimensions too great in the direction of rapping of pattern

Excess pattern

F 124

Dimensions too great in direction perpendicular to parting line

Mold expansion during baking

Improper allowance

shrinkage

rapping

of

F 125

Excessive metal thickness at irregular locations on casting exterior

Soft or ramming, movement

insufficient mold-wall

F 126

Thin casting walls over general area, especially on horizontal surfaces

Distorted casting

F 200:

Casting shape incorrect overall or in certain locations

F 210:

Pattern incorrect

F 211

Casting does not conform to the drawing shape in some or many respects; same is true of pattern

Pattern error

F 212

Casting shape is different from drawing in a particular area; pattern is correct

Pattern mounting error

F 220:

Shift or Mismatch

F 221

Casting appears to have been subjected to a shearling action in the plane of the parting line

Shift

F 222

Variation in shape of an internal casting cavity along the parting line of the core

Shifted core

F 223

Irregular projections on vertical surfaces, generally on one side only in the vicinity of the parting line

Ramoff, ramaway

F 230:

Deformations from correct shape

F 231

Deformation with respect to drawing proportional for casting, mold, and pattern

Deformed pattern

F 232

Deformation with respect to drawing proportional for casting and mold; pattern conforms to drawing

Deformed mold, mold creep, springback

F 233

Casting deformed with respect to drawing; pattern and mold conform to drawing

Casting distortion

F 234

Casting deformed with respect to drawing after storage, annealing, machining

Warped casting

Inclusions or Structural Anomalies

G 100:

Inclusions

G 110:

Metallic inclusions

G 111(a)

Metallic inclusions whose appearance, chemical analysis or structural examination show to be caused by an element foreign to the alloy

Metallic inclusions

G 112(a)

Metallic inclusions of the same chemical composition as the base metal; generally spherical and often coated with oxide

Cold shot

G 113

Spherical metallic inclusions inside blowholes or other cavities or in surface depressions (see A 311). Composition approximates that of the alloy cast but nearer to that of a eutectic

Internal sweating, phosphide sweat

G 120:

Nonmetallic inclusions; slag, dross, flux

G 121(a)

Nonmetallic inclusions whose appearance or analysis shows they arise from melting slags, products of metal treatment or fluxes

Slag, dross or flux inclusions, ceroxides

G 122(a)

Nonmetallic inclusions generally impregnated with gas and accompanied by blowholes (B 113)

Slag blowhole defect

G 130:

Nonmetallic inclusions; mold or core materials

G 131(a)

Sand inclusions, generally very close to the surface of the casting

Sand inclusions

G 132(a)

Inclusions of mold blacking or dressing, generally very close to the casting surface

Blacking or refractory coating inclusions

G 140:

Nonmetallic inclusions; oxides and reaction products

G 141

Clearly defined, irregular black spots on the fractured surface of ductile cast iron

Black spots

G 142(a)

Inclusions in the form of oxide skins, most often causing a localized seam

Oxide inclusion or skins, seams

G 143(a)

Folded films of graphitic luster in the wall of the casting

Lustrous carbon films, or kish tracks

G 144

Hard inclusions in permanent molded and die cast aluminum alloys

Hard spots

(a) Defects that under some circumstances could contribute, either directly or indirectly, to casting failures. Adapted from International Atlas of Casting Defects, American Foundrymen's Society, Des Plaines, IL

Nondestructive Inspection of Castings By the ASM Committee on Nondestructive Inspection of Castings*

Common Inspection Procedures The inspection of castings is most often limited to visual and dimensional inspections, weight testing, and hardness testing. However, for castings that are to be used in critical applications, such as automotive or aerospace components, additional methods of nondestructive inspection are used to determine and to control casting quality. Visual inspection of each casting ensures that none of its features has been omitted or malformed by molding errors,

short running, or mistakes in cleaning. Most surface defects and roughness can be observed at this stage. Initial sample castings from new pattern equipment should be carefully inspected for obvious defects. Liquid penetrant inspection can be used to detect surface defects. Such casting imperfections as shrinks, cracks, blows, or dross usually indicate the need for adjustment in the gating or foundry techniques. If the casting appears to be satisfactory upon visual inspection, internal quality can be checked by radiographic and ultrasonic inspection. The first visual inspection operation on the production casting is usually performed immediately after shakeout or knockout of the casting, ensuring that major visible imperfections are detected as quickly as possible. This information, promptly relayed to the foundry, permits early corrective action to be taken with a minimum of scrap loss. The size and complexity of some sand castings require that the gates and risers be removed to permit proper inspection of the casting. Many castings that contain numerous internal cores or have close dimensional tolerances require a rapid, but fairly accurate check of critical wall dimensions. In some cases, an indicating-type caliper gage is suitable for this work, and special types are available for casting shapes that do not lend themselves to the standard types. Ultrasonic inspection is also used to determine wall thickness in such components as cored turbine blades made by investment casting. New developments in visual inspection procedures for examining component appearance are mainly based on vision systems that use electronic cameras coupled to computer-assisted image-processing systems (Ref 1). With the development of high-sensitivity cameras having exposure times of s, components can be inspected on moving belts. Flexibility for examining three-dimensional components can be achieved with an array of cameras multiplexed to a common image processor or with a computer-controlled camera scanning system. Such systems have been successfully applied to the inspection of printed circuit boards in the electronics industry and engineering subassemblies in automobile manufacture. These tests usually operate on a go/no-go basis; either the assembly is complete with connections correctly made or it is not correct. This is a far easier task than evaluating casting quality. Studies that have been carried out to assess the possibility of extending such methods to iron castings have not given encouraging results. Contrast between defective and nondefective areas is low, illumination is critical, and consistent standards of inspection are difficult to maintain because of differences in reflectivity of the casting surfaces depending on whether or not they have been recently shotblasted. Even the simple task of identifying castings to determine their type is best carried out by examining their backlighted silhouette, and this provides no advantage in examining their quality. Dimensional Inspection. Consistency of dimensions is an inescapable requirement of premium-quality castings

supplied as near-net shape components on which subsequent high-speed machining operations are to be carried out (Ref 1). Customers will not accept increased machining costs due to inconsistencies in dimensions nor will they tolerate damage to flexible machining systems or transfer times resulting from poor control and inspection in foundries. Variations in dimensions represent one of the most common complaints with regard to the machinability of iron castings. Prevention is within the control of foundries. Differences in pattern size when using multipattern plates can be virtually eliminated by the use of computer-aided design and manufacturing methods and computer numerical control machines in patternmaking (see the article "Patterns and Patternmaking" in Casting, Volume 15 of ASM Handbook, formerly 9th Edition Metals Handbook). Better process control and methoding can eliminate variations in dimensions due to changes in metal composition or feeding methods. Variations in mold rigidity, caused by inadequate compaction with green sand, or the use of cold sand or insufficient curing times with cold-setting systems, which cause casting dimensions to fall outside the preset tolerance limits, can be greatly reduced by good molding and coremaking practices (see the articles "Sand Molding" and "Coremaking" in Casting, Volume 15 of ASM Handbook, formerly 9th Edition Metals Handbook). Because the dimensions and weight of iron castings are directly related to their soundness and are dependent on mold rigidity, the measurement of size or weight provides a simple test for checking casting integrity and for monitoring the consistency of the moldmaking process (see the section "Weight Testing" in this article). Casting dimensions are usually checked with dial gages, vernier calipers, micrometers, or vertical height gages, which may be hand held or incorporated into acceptance fixtures. Wall thickness measurements can be made with small hand-

held ultrasonic thickness gages. Under ideal conditions, the accuracy of these instruments is claimed to be ±0.01 mm (±0.0004 in.), but this is rarely achieved in practice because the surfaces are not parallel and are not machined. Instruments are available that display variations in thickness from some preset standard and provide a digital readout and a permanent record of results for statistical analysis. Developments lie in the use of measuring systems employing capacitance, electrical contact, or linear displacement transducers. Such systems are capable of high accuracy and the output can be linked directly to microcomputers for data recording and statistical analysis to meet the requirements of statistical process control. The use of computer-aided dimensional control and statistical process control is discussed later in this article. Laser methods of measurement using beam displacement or time-lapse techniques are available for use in machine shops where accurate measurement is required for control of automatic machining processes. At present, they are generally not well suited for measuring castings, because of their high cost and because it is difficult to make precise measurements on components having a complex shape with curved or as-cast surfaces. As these laser methods become more widely used in other industries, lower-cost systems will become available, and these might be developed for foundry use (see the article "Laser Inspection" in this Volume). Weight Testing. Many intricately cored castings are extremely difficult to measure accurately, particularly the internal sections. It is important to ensure that these sections are correct in thickness for three main reasons:

• • •

There should be no additional weight that would make the finished product heavier than permissible Sections must not be thinner than designed so as not to decrease the strength of the casting If hollow cavities have been reduced in area by increasing the metal thickness of the sections, any flow of liquid or gases is reduced

A ready means of testing for these discrepancies is by accurately weighing each casting or by measuring the displacement caused by immersing the casting in a liquid-filled measuring jar or vessel. In certain cases in which extreme accuracy is demanded, a tolerance of only ±1% of a given weight may be allowed. Hardness testing is often used to verify the effectiveness of the heat treatment applied to actual castings. Its general

correlation with the tensile strength of many ferrous alloys allows a rough prediction of tensile strength to be made. The Brinell hardness test is most frequently used for casting alloys. A combination of large-diameter ball (5 or 10 mm) and heavy load (500 to 3000 kgf) is preferred for the most effective representation because a deep impression minimizes the influence of the immediate surface layer and of the relatively coarse microstructure. The Brinell hardness test is unsuitable for use at high hardness levels (above 600 HB), because distortion of the ball indenter can affect the shape of the indentation. Either the Rockwell or the Vickers (136° diamond pyramid) hardness test is used for alloys of extreme hardness or for high-quality and precision castings in which the large Brinell indentation cannot be tolerated. Because of the very small indentations produced in Rockwell and Vickers tests, which use loads of 150 kg or less, results must be based on the average of a number of determinations. Portable hardness testers or ultrasonic microhardness testers can be used on large castings that cannot be placed on the platform of a bench-type machine. More detailed information on hardness testing is available in Mechanical Testing, Volume 8 of ASM Handbook, formerly 9th Edition Metals Handbook. The hardness of ferrous castings can be determined from the sonic velocity of the metal if all other test conditions remain constant. This has been demonstrated on chilled rolls in determining the average hardness of the core.

Reference cited in this section

1. P.J. Rickards, Progress in Guaranteeing Quality Through Nondestructive Methods of Evaluation, Foundryman Int., April 1988, p 196-209

Nondestructive Inspection of Castings By the ASM Committee on Nondestructive Inspection of Castings*

Computer-Aided Dimensional Inspection The use of computer equipment in foundry inspection operations is finding more acceptance as the power and utility of available hardware and software increase. The computerization of operations can reduce the man-hours required for inspection tasks, can increase accuracy, and can allow the analysis of data in ways that are not possible or practical with manual operations. Perhaps the best example of this, given the currently available equipment, is the application of computer technology to the dimensional inspection of castings. Importance of Dimensional Inspection. One of the most critical determinants of casting quality in the eyes of the casting buyer is dimensional accuracy. Parts that are within dimensional tolerances, given the absence of other casting defects, can be machined, assembled, and used for their intended functions with minimal testing and inspection costs. Major casting buyers are therefore demanding statistical evidence that dimensional tolerances are being maintained. In addition, the statistical analysis of in-house processes has been demonstrated to be effective in keeping those processes under control, thus reducing scrap and rework costs.

The application of computer equipment to the collection and analysis of dimensional inspection data can increase the amount of inspection that can be performed and decrease the time required to record and analyze the results. This furnishes control information for making adjustments to tooling on the foundry floor and statistical information for reporting to customers on the dimensional accuracy of parts. Additional information on dimensional accuracy in castings can be found in the article "Dimensional Tolerances and Allowances" in Casting, Volume 15 of ASM Handbook, formerly 9th Edition Metals Handbook. Typical Equipment. A typical installation for the dimensional inspection of castings consists of an electronic

coordinate measuring machine, a microcomputer interfaced to the coordinate measuring machine controller with a data transfer cable, and a software system for the microcomputer (Fig. 1). The software system should be capable of controlling the functions and memory storage of the coordinate measuring machine as well as recalling and analyzing the data it collects. The software serves as the main control element for the dimensional inspection and statistical reporting of results. Such software can be purchased or, if the expertise is available, developed in-house for highly specialized requirements.

Fig. 1 Equipment used in a typical installation for the computer-aided dimensional inspection of castings showing a coordinate measuring machine and microcomputer

Coordinate measuring machines typically record dimensions along three axes from data points specified by the user. Depending on the sophistication of the controller, such functions as center and diameter finds for circular features and the electronic rotation of measurement planes can be performed. Complex geometric constructions, such as the intersection points of lines and planes and out-of-roundness measurements, are typically off-loaded for calculation into the microcomputer. The contact probe of the coordinate measuring machine can be manipulated manually, or in the case of direct computer controlled machines, the probe can be driven by servomotors to perform the part measurement with little operator intervention. More detailed information on these machines can be found in the article "Coordinate Measuring Machines" in this Volume. The Measurement Process. Figure 2 illustrates the general procedure that is followed in applying semiautomatic dimensional inspection to a given part. The first step is to identify the critical part dimensions that are to be measured and tracked. Nominal dimensions and tolerances are usually taken from the customer's specifications and blueprints. Dimensions that are useful in controlling the foundry process can also be selected. A data base file, including a description and tolerance limits for each dimension to be checked, is then created using the microcomputer software system.

Fig. 2 Flowchart showing typical sequence of operations for computer-aided dimensional inspection

The next step in the setup process is to develop a set of instructions for measuring the part with the coordinate measuring machine. The instructions consist of commands that the coordinate measuring machine uses to establish reference planes and to measure such features as the center points of circular holes. This measurement program can be entered in either of two ways. In the first method, the operator simply types in a list of commands that he wants the coordinate measuring machine to execute and that give the required dimensions as defined in the part data base. The second method uses a teach mode. The operator actually places a part on the worktable of the coordinate measuring machine and checks it in the proper sequence, while the computer monitors the process and stores the sequence of commands used. In either case, the result is a measurement program stored on the microcomputer that defines in precise detail how the part is to be measured. Special commands can also be included in the measurement program to display operator instructions on the computer screen while the part is being measured. In developing the measurement program, consideration must be given to the particular requirements of the part being measured. Customer prints will normally show datum planes from which measurements are to be made. When using a cast surface to establish a datum plane, it is good practice to probe a number of points on the surface and to allow the computer to establish a best-fit plane through the points. Similarly, the center points of cast holes can best be found by probing multiple points around the circumference of the hole. Machined features can generally be measured with fewer probe contacts. When measuring complex castings, maximum use should be made of the ability of the coordinate

measuring machine to electronically rotate measurement planes without physically moving the part; unclamping and turning a part will lower the accuracy of the overall layout. Once the setup process is complete, the dimensional inspection of parts from the foundry begins. Based on statistical considerations, a sampling procedure and frequency must be developed. Parts are then selected at random from the process according to the agreed-upon frequency. The parts are brought to the coordinate measuring machine, and the operator calls up the measurement program for that part and executes it. As the part is measured, the dimensions are sent from the coordinate measuring machine to the data base on the microcomputer. Once the measurement process is completed, information such as mold number, shift, date, and serial number should be entered by the operator so that this particular set of dimensions can be identified later. A layout report can then be generated to show how well the measured part checked out relative to the specified dimensions and tolerances. Figure 3 shows a sample report in the form of a bar graph, in which any deviation from print tolerance appears as a line of dashes to the left or right of center. A deviation outside of tolerance limits displays asterisks to flag its condition. Such a report is useful in that it gives a quick visual indication of the measurement of one casting.

Fig. 3 Example layout report showing all dimensions measured on a single casting, with visual indication of deviations from print mean. Note out-of-tolerance condition indicated by asterisks.

Statistical analysis permits the mathematical prediction of the characteristics of all the parts produced by measuring only a sample of those parts (the principles and applications of statistics in industrial environments are discussed in the article "Statistical Quality Design and Control" in this Volume). All processes are subject to some amount of natural variation; in most processes, this variation follows a normal distribution (the familiar bell-shaped curve) when the probability of occurrence is plotted against the range of possible values. Standard deviation, a measure of the distance from center on the probability curve, is the principal means of expressing the range of measured values. For example, a spread of six standard deviations (plus or minus three standard deviations on either side of the measured mean) represents the range within which one would expect to find 99.73% of observed measurements for a normal process. This allows the natural variation inherent in the process to be quantified. Control Charts. With statistical software incorporated into the microcomputer system, the results of numerous

measurements of the same part can be analyzed to determine, first, how well the process is staying in control, that is, whether the natural variations occurring in a given measurement are within control limits and whether any identifiable

trends are occurring. This is done by using a control chart (Fig. 4), which displays the average values and ranges of groups of measurements plotted against time. Single-value charts with a moving range can also be helpful. The control limits can also be calculated and displayed. With the computer, this type of graph can be generated within seconds. Analysis of the graph may show a developing trend that can be corrected by adjusting the tooling before out-of-tolerance parts are made.

Fig. 4 Control chart with average of groups of measurements (X values) plotted above and ranges within the groups (R) plotted below. Control limits have been calculated and placed on the chart by the computer.

Statistical Summary Report. The second type of analysis shows the capability of the process, that is, how the range of natural variation (as measured by a specified multiple of the standard deviation) compares with the tolerance range specified for a given dimension. An example of a useful report of this type is shown in Fig. 5. This information is of great interest both to the customer and the process engineer because it indicates whether or not the process being used to produce the part can hold the dimensions within the required tolerance limits. The user must be aware that different methods of capability analysis are used by different casting buyers, so the software should be flexible enough to accommodate the various methods of calculation that might be required.

Fig. 5 Statistical summary report showing the mean of measured observations, the blueprint specification for the mean, the difference between specified and measured means, the tolerance, the standard deviation of the measured dimensions, and the capability of the process. These calculations are performed for all measured dimensions on the part.

Histograms. An alternative method of assessing capability involves the use of a histogram, or frequency plot. This is a

graph that plots the number of occurrences within successive, equally spaced ranges of a given measured dimension. Figure 6 shows an example output report of this type. A graph such as this, which has superimposed upon it the tolerance limits for the dimension being analyzed, allows a quick, qualitative evaluation of the variation and capability of the process. It also allows the normality of the process to be judged through comparison with the expected bell-shaped curve of a normal process.

Fig. 6 Frequency plot for one measured dimension showing the distribution of the measurements

Other Applications for Computer-Aided Inspection. The general sequence described for semiautomatic

dimensional inspection can be applied to a number of other inspection criteria. Examples would include pressure testing or defect detection by electronic vision systems. The statistical analysis of scrap by defect types is also very helpful in identifying problem areas. In some cases, direct data input to a computer may not be feasible, but the benefits of entering data manually into a statistical analysis program should not be overlooked. The computer allows rapid analysis of large amounts of data so that statistically significant trends can be detected and proper attention paid to appropriate areas for improvement. The benefits and costs of each anticipated application of automation to a particular situation, as well as the feasibility of applying state-of-the-art equipment, need to be studied as thoroughly as possible prior to implementation. Nondestructive Inspection of Castings By the ASM Committee on Nondestructive Inspection of Castings*

Liquid Penetrant Inspection

Liquid penetrant inspection essentially involves a liquid wetting the surface of a workpiece, flowing over that surface to form a continuous and uniform coating, and migrating into cracks or cavities that are open to the surface. After a few minutes, the liquid coating is washed off the surface of the casting, and a developer is placed on the surface. The developer is stained by the liquid penetrant as it is drawn out of the cracks and cavities. Liquid penetrants will highlight surface defects so that detection is more certain. Details of this method can be found in the article "Liquid Penetrant Inspection" in this Volume. Liquid penetrant inspection should not be confined to as-cast surfaces. For example, it is not unusual for castings of various alloys to exhibit cracks (frequently intergranular) on machined surfaces. A pattern of cracks of this type may be the result of intergranular cracking throughout the material because of an error in composition or heat treatment, or the cracks may be on the surface only as a result of machining or grinding. Surface cracking may result from insufficient machining allowance, which does not allow for complete removal of imperfections produced on the as-cast surface, or it may result from faulty machining techniques. If imperfections of this type are detected by visual inspection, liquid penetrant inspection will show the full extent of such imperfections, will give some indication of the depth and size of the defect below the surface by the amount of penetrant absorbed, and will indicate whether cracking is present throughout the section. Liquid penetrant inspection is sometimes used in conjunction with another nondestructive method. Such is the case in the following example.

Example 1: Tail Rotor Gearbox Failure on the H-53 Helicopter. The H-53 tail rotor gearbox mounting lugs experienced failures that resulted in at least two aircraft accidents. The gearbox is a magnesium casting, either AZ-91 or ZE-41. Initially, only visual inspection using a 10× magnifying glass was required. Initiation sites for the cracking were found to be localized in a few areas adjacent to the attached lugs (Fig. 7). A manual eddy current inspection was developed that concentrated on the potential initiation sites. If an indication was found, then a fluorescent penetrant inspection was performed as a backup inspection.

Fig. 7 Helicopter transmission mounting lug. Note: bolt is shown removed for clarity. Bolt is not removed for eddy current inspection.

Inspection Procedure. An eddy current tester with a flawgate and shielded probes was used to perform the

inspection. The instrument is calibrated using standard magnesium alloy (either AZ-91 or ZE-41) calibration blocks. The tester is calibrated on a 0.50 mm (0.020 in.) slot to achieve a minimum 150 A deflection. The flawgate is calibrated to activate for meter deflection greater than 150 A. The part is scanned very slowly over all critical areas. As an aid for the inspection of the lug edges, a wooden toothpick is taped to the probe to maintain a constant probe-to-edge distance. A positive indication--meter deflection greater than 150 A--requires a fluorescent penetrant inspection. In preparation for the fluorescent penetrant inspection, the paint in the area to be inspected is removed. The required penetrant material sensitivity is level II or III and is used with a halogenated solvent remover and non-aqueous developer. Clean cloths moistened with solvent cleaner are required for the cleaning step. The inspection area is never flushed with solvent. An initial inspection is required before the developer is applied. If no indications are noted, developer is applied. When an indication is noted, it is wiped with a cotton swab moistened with solvent, and the area is inspected for the indication bleed-back. A 30-min penetrant dwell time is required. The minimum required dwell time between cleaning and application of the developer is 10 min before inspection. The minimum developer dwell time is half the penetrant dwell time, and a maximum developer dwell time is not to exceed the penetrant dwell time. After the inspection, all developer residue is removed from the mounting lug. Areas of greatest concern include the mount pad radius areas, lug flanges along edges, and the mounting pad. After completion of penetrant inspection, the finishes on

the mounting lug are replaced in such a way that the eddy current inspection is not hindered and the visual inspection is enhanced. The visual inspection is enhanced by painting the lugs light gray or white. Conclusions. These inspections were performed during scheduled inspection intervals until magnesium alloy castings were replaced with A356.0 aluminum alloy castings. Nondestructive Inspection of Castings By the ASM Committee on Nondestructive Inspection of Castings*

Magnetic Particle Inspection Magnetic particle inspection is a highly effective and sensitive technique for revealing cracks and similar defects at or just beneath the surface of castings made of ferromagnetic metals. The capability of detecting discontinuities just beneath the surface is important because such cleaning methods as shot or abrasive blasting tend to close a surface break that might go undetected in visual or liquid penetrant inspection. When a magnetic field is generated in and around a casting made of a ferromagnetic metal and the lines of magnetic flux are intersected by a defect such as a crack, magnetic poles are induced on either side of the defect. The resulting local flux disturbance can be detected by its effect on the particles of a ferromagnetic material, which become attracted to the region of the defect as they are dusted on the casting. Maximum sensitivity of indication is obtained when a defect is oriented in a direction perpendicular to the applied magnetic field and when the strength of this field is sufficient to saturate the casting being inspected. Equipment for magnetic particle inspection uses direct or alternating current to generate the necessary magnetic fields. The current can be applied in a variety of ways to control the direction and magnitude of the magnetic field. In one method of magnetization, a heavy current is passed directly through the casting placed between two solid contacts. The induced magnetic field then runs in the transverse or circumferential direction, producing conditions favorable to the detection of longitudinally oriented defects. A coil encircling the casting will induce a magnetic field that runs in the longitudinal direction, producing conditions favorable to the detection of circumferentially (or transversely) oriented defects. Alternatively, a longitudinal magnetic field can be conveniently generated by passing current through a flexible cable conductor, which can be coiled around any metal section. This method is particularly adaptable to castings of irregular shape. Circumferential magnetic fields can be induced in hollow cylindrical castings by using an axially disposed central conductor threaded through the casting. Small castings can be magnetic particle inspected directly on bench-type equipment that incorporates both coils and solid contacts. Critical regions of larger castings can be inspected by the use of yokes, coils, or contact probes carried on flexible cables connected to the source of current; this setup enables most regions of castings to be inspected. Equipment and techniques associated with this method can be found in the article "Magnetic Particle Inspection" in this Volume. Nondestructive Inspection of Castings By the ASM Committee on Nondestructive Inspection of Castings*

Eddy Current Inspection Eddy current inspection consists of observing the interaction between electromagnetic fields and metals. In a basic system, currents are induced to flow in the testpiece by a coil of wire that carries an alternating current. As the part enters the coil, or as the coil in the form of a probe or yoke is placed on the testpiece, electromagnetic energy produced by the coils is partly absorbed and converted into heat by the effects of resistivity and hysteresis. Part of the remaining energy is reflected back to the test coil, its electrical characteristics having been changed in a manner determined by the properties of the testpiece. Consequently, the currents flowing through the probe coil are the source of information describing the

characteristics of the testpiece. These currents can be analyzed and compared with currents flowing through a reference specimen. Eddy current methods of inspection are effective with both ferromagnetic and nonferromagnetic metals. Eddy current methods are not as sensitive to small, open defects as liquid penetrant or magnetic particle methods are. Because of the skin effect, eddy current inspection is generally restricted to depths less than 6 mm ( in.). The results of inspecting ferromagnetic materials can be obscured by changes in the magnetic permeability of the testpiece. Changes in temperature must be avoided to prevent erroneous results if electrical conductivity or other properties, including metallurgical properties, are being determined. Applications of eddy current and electromagnetic methods of inspection to castings can be divided into the following

three categories: • • •

Detecting near-surface flaws such as cracks, voids, inclusions, blowholes, and pinholes (eddy current inspection) Sorting according to alloy, temper, electrical conductivity, hardness, and other metallurgical factors (primarily electromagnetic inspection) Gaging according to size, shape, plating thickness, or insulation thickness (eddy current or electromagnetic inspection)

The following example demonstrates the use of eddy current inspection for the evaluation of an induction-hardened surface on a gray iron casting.

Example 2: Eddy Current Inspection of Hardened Valve Seats on Gray Iron Automotive Cylinder Heads. Exhaust-valve seats on gray iron automotive engine cylinder heads were induction hardened to withstand use with leadfree fuels. The seats were specified to be hardened to 50 HRC to a depth of 1.27 to 2.03 mm (0.050 to 0.080 in.). Failure to achieve these depths could lead to excessive recession of the seat, causing premature failure. To establish the feasibility of, and the conditions for, eddy current inspection of cylinder heads having valve seats of different diameters, an adjustable probe was made using a 10 mm (0.400 in.) diam standard probe. The end was machined to a 120° conical point, and the probe was mounted in a plastic holder. A scrap valve stem mounted in the holder was used to position the probe on the valve seat. The probe was centered on the valve seat by moving the probe in its holder and by the adjusting screw. After the inspection conditions were established, rugged units for use by production inspectors were made of stainless steel with the protected probe in a fixed position to suit the cylinder-head seat. In operation, the probe was brought into intimate contact with the hardened surface, and a meter reading was taken. Positioning of the probe on the valve seat is shown in Fig. 8. The meter reading, taken from an instrument calibrated from -100 to +100, was compared with a developed chart for case depth versus meter reading (a linear relationship). A meter reading of 0 was established for 50 HRC at a depth of 1.27 mm (0.050 in.) and -40 for 2.03 mm (0.080 in.). The instrumentation was set up to respond to Ar (resistive component). Inspection was performed at a frequency of 2 kHz. Extensive laboratory and production inspection has shown a standard deviation, of ±0.8 mm (±0.03 in.) and a confidence limit of 3 for the total data generated.

Fig. 8 Eddy current inspection using an adjustable probe for determining case depth of an induction-hardened gray iron valve seat.

Nondestructive Inspection of Castings By the ASM Committee on Nondestructive Inspection of Castings*

Radiographic Inspection Radiographic inspection is a process of testing materials using penetrating radiation from an x-ray generator or a radioactive source and an imaging medium, such as x-ray film or an electronic device. In passing through the material, some of the radiation is attenuated, depending on the thickness and the radiographic density of the material, while the radiation that passes through the material forms an image. The radiographic image is generated by variations in the intensity of the emerging beam. Internal flaws, such as gas entrapment or nonmetallic inclusions, have a direct effect on the attenuation. These flaws create variations in material thickness, resulting in localized dark or light spots on the image. The term radiography usually implies a radiographic process that produces a permanent image on film (conventional radiography) or paper (paper radiography or xero radiography), although in a broad sense it refers to all forms of radiographic inspection. When inspection involves viewing an image on a fluorescent screen or image intensifier, the radiographic process is termed filmless or real-time inspection. When electronic nonimaging instruments are used to measure the intensity of radiation, the process is termed radiation gaging. Tomography, a radiation inspection method adapted from the medical computerized axial tomography scanner, provides a cross-sectional view of a testpiece. All of the above terms are primarily used in connection with inspection that involves penetrating electromagnetic radiation in the form of x-rays or -rays. Neutron radiography refers to radiographic inspection using neutrons rather than electromagnetic radiation. The sensitivity, or the ability to detect flaws, of radiographic inspection depends on close control of the inspection technique, including the geometric relationships among the point of x-ray emission, the casting, and the x-ray imaging plane. The smallest detectable variation in metal thickness lies between 0.5 and 2.0% of the total section thickness. Narrow flaws, such as cracks, must lie in a plane approximately parallel to the emergent x-ray beam to be imaged; this requires multiple exposures for x-ray film techniques and a remote-control parts manipulator for a real-time system. Real-time systems have eliminated the need for multiple exposures of the same casting by dynamically inspecting parts on a manipulator, with the capability of changing the x-ray energy for changes in total material thickness. These capabilities have significantly improved productivity and have reduced costs, thus enabling higher percentages of castings

to be inspected and providing instant feedback after repair procedures. Figures 9 and 10 show real-time digital radiography images of automotive components.

Fig. 9 Digital radiography image of a die cast aluminum carburetor. Porosity appears as dark spots in the area of the center bore, through the vertical center of the image. Courtesy of B.G. Isaacson, Bio-Imaging Research, Inc.

Fig. 10 Evaluation of cast transmission housing assembly. (a) Photograph of cast part. (b) Digital radiography image used to verify the steel spring pin and shuttle valve assembly through material thicknesses ranging from 3 mm ( in.) in the channels to 25 mm (1 in.) in the rib sections of the casting. Courtesy of B.G. Isaacson, BioImaging Research, Inc.

Advances. Several advances have been made to assist the industrial radiographer. These include the computerization of

the radiographic standard shooting sketch, which graphically shows areas to be x-rayed and the viewing direction or angle at which the shot is to be taken, and the development of microprocessor-controlled x-ray systems capable of storing different x-ray exposure parameters for rapid retrieval and automatic warm-up of the system prior to use. The advent of digital image-processing systems and microfocus x-ray sources (near point source), producing energies capable of penetrating thick material sections, have made real-time inspection capable of producing images equal to, and in some cases superior to, x-ray film images by employing geometric relations previously unattainable with macrofocus x-ray systems. The near point source of the microfocus x-ray system virtually eliminates the edge unsharpness associated with larger focus devices.

Digital image processing can be used to enhance imagery by multiple video frame integration and averaging techniques that improve the signal-to-noise ratio of the image. This enables the radiographer to digitally adjust the contrast of the image and to perform various edge enhancements to increase the conspicuity of many linear indications (Fig. 11).

Fig. 11 Digital radiography images of an investment cast jet engine turbine blade showing detail through a wide range in material thickness. The trailing edge of the blade (along the top of the image) is 2 mm (0.080 in.) thick, the root section of the blade (to the far left in the image) is 19 mm (0.75 in.) thick, and the shelf area (to the right of the root section) is 25 mm (1 in.) thick. The image shown in (a) is unprocessed; the image in (b) is processed to subdue the background and to enhance edges and internal features. Courtesy of B.G. Isaacson, Bio-Imaging Research, Inc.

Interpretation of the radiographic image requires a skilled specialist who can establish the correct method of exposing the castings with regard to x-ray energies, geometric relationships, and casting orientation and can take all of these factors into account to achieve an acceptable, interpretable image. Interpretation of the image must be performed to establish standards in the form of written or photographic instructions. The inspector must also be capable of determining if the localized indication is a spurious indication, a film artifact, a video aberration, or a surface irregularity. The article "Radiographic Inspection" in this Volume provides additional information on real-time digital systems. Computed tomography, also known as computerized axial tomography (or CAT scanning), is a more sophisticated x-

ray imaging technique originally developed for medical diagnostic use (Ref 2). It is the complete reconstruction by computer of a tomographic plane, or slice of an object. A collimated (fan-shaped) x-ray beam is passed through a section of the part and is intercepted by a detector on the other side. The part is rotated slightly, and a new set of measurements is made; this process is repeated until the part has been rotated 180°. The resulting image of the slice (tomogram) is formed by computer calculations based on electronic measurement (digital sampling) of the radiation transmitted through the object along different paths during the rotating scan. The data thus accumulated are used to compute the densities of each point in the cross section, enabling the computer to reconstruct a two-dimensional visual image of the slice. The shapes of internal features are determined by their computed densities. After one slice is produced through a complete rotation, either the part or the radiation source and detector can be moved and a three dimensional image built up through the scanning of successive slices. Compared to electronic radiography, computed tomography provides increased sensitivity and detection capabilities. The contrast resolution of a good-quality tomographic image is 0.1 to 0.2%, which is approximately two orders of magnitude better than with x-ray film. In addition, images are produced in a quantitative, ready-to-use digital format (Ref 3). They provide detailed physical information, such as size, density, and composition, to aid in evaluating defects. Methods are being developed to use this information to predict failure modes or system performance under operating loads. The data can also be easily manipulated to obtain various types of images, to develop automated flaw detection techniques, and to promote efficient archiving. Figures 12 and 13 illustrate the uses of computed tomography for examining castings. More detailed information can be found in the article "Industrial Computed Tomography" in this Volume.

Fig. 12 Computed tomographic images of a die cast aluminum automotive piston. (a) Photograph of cast part. (b) Vertical slice through the piston shows porosity as dark spots in the crown area (point A) and counterbalance area (point B). (c) Transverse slice through the crown of the piston verifies the porosity; the smallest void that is visible is 0.4 mm (0.016 in.) in diameter. (d) Transverse slice through the counterbalance area also verifies porosity. Dimensional analysis of the piston walls is possible to an accuracy of ±50 m (±0.002 in.). Courtesy of B.G. Isaacson, Bio-Imaging Research, Inc.

Fig. 13 Use of computed tomography for examining automotive components. (a) Photograph of a cast aluminum transmission case with (b) corresponding tomographic image. (c) Two three-dimensional images of a cast aluminum cylinder head generated from a set of continuous tomographic scans used to view water-cooling chambers where a leak had been detected. Courtesy of R.A. Armistead, Advanced Research and Applications Corporation.

References cited in this section

2. P.M. Bralower, Nondestructive Testing. Part I. The New Generation in Radiography, Mod. Cast., Vol 76 (No. 7), July 1986, p 21-23 3. R.A. Armistead, CT: Quantitative 3-D Inspection, Adv. Mater. Process., Vol 133 (No. 3), March 1988, p 4248 Nondestructive Inspection of Castings By the ASM Committee on Nondestructive Inspection of Castings*

Ultrasonic Inspection Ultrasonic inspection is a nondestructive method in which beams of high-frequency acoustic energy are introduced into the material under evaluation to detect surface and subsurface flaws and to measure the thickness of the material or the distance to a flaw (see the article "Ultrasonic Inspection" in this Volume for details). An ultrasonic beam will travel through a material until it strikes an interface or defect. Interfaces and defects interrupt the beam and reflect a portion of the incident acoustic energy. The amount of energy reflected is a function of the nature and orientation of the interface or flaw as well as the acoustic impedance of such a reflector. Energy reflected from various interfaces or defects can be used to define the presence and locations of defects, the thickness of the material, or the depth of the defect beneath a surface. The advantages of ultrasonic tests are as follows: • • •

High sensitivity, which permits the detection of minute cracks Great penetrating power, which allows the examination of extremely thick sections Accuracy in measuring of flaw position and estimating defect size

Ultrasonic tests have the following limitations: • • •

Size-contour complexity and unfavorable discontinuity orientation can pose problems in interpreting the echo pattern Undesirable internal structure--for example, grain size, structure, porosity, inclusion content, or fine dispersed precipitates--can similarly hinder interpretation Reference standards are required

Effect of Casting Shape. Because castings are rarely simple flat shapes, they are not as easy to inspect as such

products as rolled rectangular bars. The reflections of a sound beam from the back surface of a parallel-sided casting and a discontinuity are shown schematically in Fig. 14(a), together with the relative heights and positions of the reflections of the two surfaces on an oscilloscope screen. A decrease in the back reflection at the same time as the appearance of a discontinuity echo is a secondary indication of the presence of a discontinuity. However, if the back surface of the casting at a particular location for inspection is not approximately at a right angle to the incident sound beam, the beam will be reflected to remote parts of the casting and not directly returned to the detector. In this case, as shown in Fig. 14(b), there is no back reflection to monitor as a secondary indication.

Fig. 14 Schematic of the effect of casting shapes on reflection and oscilloscope screen display of sound beams. See text for discussion.

Many castings contain cored holes and changes in section, and echoes from holes and changes in section can interfere with echoes from discontinuities. As shown in Fig. 14(c), the echo from the cored hole overlaps the echo from the discontinuity on the oscilloscope screen. The same effect is shown in Fig. 14(d), in which echoes from the discontinuity and the casting fillets at a change in section are shown overlapping on the oscilloscope. Curved surfaces do not permit adequate or easy coupling of the flat search units to the casting surface, especially with contact double search units. This can be overcome to some extent by using a suitable viscous couplant, but misleading results may be produced because multiple reflections in the wedge of fluid between the search unit and the surface can result in echoes on the screen in those positions where discontinuity echoes may be expected to appear. Because the reflections inside the couplant use energy that would otherwise pass into the casting, the back echo decreases, and this decrease might be interpreted as confirmation of the presence of a discontinuity. On cylindrical surfaces, the indication will change as a double search unit is rotated. The wedge effect is least when the division between the transmitting and receiving transducers is parallel to the axis of the cylinder. Wedge effects in the couplant are a particular problem on castings curved in two directions. One solution in this case is to use a small search unit so that the wedge is short, although the resolution and sensitivity may be reduced. If the surface of the casting to be inspected is of regular shape, such as the bore of a cylinder in an engine block, the front of the search unit can be shaped to fit the curvature of the surface. These curved shapes form an acoustic lens that will alter the shape of the sound beam, but unless the curvature is severe, this will not prevent adequate accuracy in the

inspection. Cast-on flat metal pads for application of the ultrasonic search unit are very effective and allow particular areas of the casting to be inspected. Subsurface Defects. Defects such as small blowholes, pinholes, or inclusions that occur within depths of 3 or 4 mm

(0.1 or 0.15 in.) of a cast surface are among the most difficult to detect (Ref 1). They are beyond the limits of sensitivity of conventional magnetic particle methods and are not easily identified by eddy current techniques. They usually fall within the dead zone (the surface layer that cannot be inspected) of conventional single-crystal ultrasonic probes applied directly to a cast surface, although some improvement can be obtained by using twin crystal probes focused to depths not too far below the surface. The other alternative using contact methods of ultrasonic testing is to employ angle probes, but this complicates the procedures and interpretation methods to the point at which they can only be applied satisfactorily under the close control of skilled operators. Freedom from such surface defects is, however, a very important aspect of the quality of castings. Apart from their effect in reducing bending fatigue properties, such defects are frequently revealed at late stages in the machining of a component, leading to its rejection. Ultrasonic methods for detecting subsurface defects are much more successful when the dead zone beneath the as-cast surface is virtually eliminated by using immersion methods in which the probe is held away from the cast surface at a known controlled distance, with coupling being obtained through a liquid bath. To make such methods consistent and reliable, the test itself must be automated. Semiautomatic equipment has been developed for examining castings such as cylinder heads by this method (Ref 1). With this equipment, the casting is loaded into a cradle from a roller track and is then transferred using a hoist into the immersion tank until the surface of the casting to be inspected is just submerged in the liquid. Depth of immersion is closely controlled because the customer will not permit liquid to be left in the internal passageways of the cylinder head. The immersed surface of the casting is then scanned manually using an ultrasonic probe held at a fixed distance from the casting surface. This equipment is suitable for testing any casting requiring examination over a flat surface. Internal Defects. Ultrasonic inspection is a well-established method for the detection of internal defects in castings.

Test equipment developments, automated testing procedures, and improvements in determining the size and position of defects, which is essential to assessing whether or not their presence will likely affect the service performance of the casting, have contributed to the increasing use of ultrasonic test equipment. For determining the position and size of defects, the usual method of presentation of ultrasonic data is an A-scan, in which the amplitude of the echoes from defects is shown on a time base and has well-known limitations (Ref 1). Sizing relies on measuring the drop in amplitude of the echo as the probe is passed over the boundary of a defect or measuring the reduction in the amplitude of the back wall echo due to the scattering of sound by the defect. In most cases, sizing is approximate and is restricted to one or two dimensions. Improvements in data presentation in the form of B-scans and Cscans that present a plan view through the section of the component provide a marked improvement in defining defect positions and size in two or three dimensions. Such displays have been used for automated defect characterization systems in which porosity, cracks, and dross have been distinguished. Because of the requirement to scan the probe over the surface, the application of B-scan and C-scan methods has generally been limited to simple geometric shapes having good surface finish, such as welded plate structures. Application to castings is currently restricted, but greater use of Bscan and C-scan methods is likely with either improved scanning systems or arrays of ultrasonic probes. Structure evaluation is an area of growing importance for foundry engineers. Ultrasonic velocity measurements are widely used as a means of guaranteeing the nodularity of the graphite structure and, if the matrix structure is known to be consistent, guaranteeing the principal material properties of ductile irons (Ref 1). Velocity measurements have been used to evaluate compacted graphite iron structures to ensure that the desired properties have been consistently obtained. The use of ultrasonic velocity measurements for structure evaluation is discussed in the section "Ductile Iron Castings" in this article.

Reference cited in this section

1. P.J. Rickards, Progress in Guaranteeing Quality Through Nondestructive Methods of Evaluation, Foundryman Int., April 1988, p 196-209

Nondestructive Inspection of Castings By the ASM Committee on Nondestructive Inspection of Castings*

Leak Testing Castings that are intended to withstand pressures can be leak tested in the foundry. Various methods are used, according to the type of metal being tested (see the article "Leak Testing" in this Volume). One method consists of pumping air at a specific pressure into the inside of the casting in water at a given temperature. Any leaks through the casting become apparent by the release of bubbles of air through the faulty portions. An alternative method is to fill the cavities of a casting with paraffin at a specified pressure. Paraffin, which penetrates the smallest of crevices, will rapidly find any defect, such as porosity, and will show quickly as an oily or moist patch at the position of the fault. Liquid penetrants can be poured into areas of apparent porosity and time allowed for the liquid to seep through the casting wall. The pressure testing of rough (unmachined) castings at the foundry may not reveal any leaks, but it must be recognized that subsequent machining operations on the casting may cut into porous areas and cause the casting to leak after machining. Minor seepage leaks can be sealed by impregnation of the casting with liquid or by filling with sodium silicate, a synthetic resin, or other suitable substance. As-cast parts can be impregnated at the foundry to seal leaks if there is to be little machining or if experience has shown that machining does not affect the pressure tightness. However, it is usually preferable to impregnate after final machining of the casting. Nondestructive Inspection of Castings By the ASM Committee on Nondestructive Inspection of Castings*

Inspection of Ferrous Castings Ferrous castings can be inspected by most of the nondestructive inspection methods. Magnetic particle inspection can be applied to ferrous metals with excellent sensitivity, although a crack in a ferrous casting can often be seen by visual inspection. Magnetic particle inspection provides good crack delineation, but the method should not be used to detect other defects. Nonrelevant magnetic particle indications occasionally occur on ferrous castings, especially with a strong magnetic field. For example, a properly fused-in steel chaplet can be indicated as a defect because of the difference in magnetic response between low-carbon steel and cast iron. Even the graphite in cast iron, which is nonmagnetic, can cause a nonrelevant indication. Standard x-ray and radioactive-source techniques can be used to make radiographs of ferrous castings, but the typical complexity of shape and varying section thicknesses of the castings may require digital radiography or computed tomography. Ultrasonic inspection for both thickness and defects is practical with most ferrous castings except for the high-carbon gray iron castings, which have a high damping capacity and absorb much of the input energy. The measurement of resonant frequency is a good method of inspecting some ductile iron castings for soundness and graphite shape. Electromagnetic testing can be used to distinguish metallurgical differences between castings. The criteria for separating acceptable from unacceptable castings must be established empirically for each casting lot. Gray Iron Castings Gray iron castings are susceptible to most of the imperfections generally associated with castings, with additional problems resulting from the relatively high pouring temperatures. These additional problems result in a higher incidence of gas entrapment, inclusions, poor metal structure, interrupted metal walls, and mold wall deficiencies. Gas entrapment is a direct result of gas being trapped in the casting wall during solidification. This gas may be in the

metal prior to pouring, may be generated from aspiration during pouring, or may be generated from core and mold materials. Internal defects of this type are best detected by radiography, but ultrasonic and eddy current inspection methods are useful when the defect is large enough to be detected by these methods.

Inclusions are casting defects in which solid foreign materials are trapped in the casting wall. The inclusion material

can be slag generated in the melting process, or it can be fragments of refractory, mold sand, core aggregate, or other materials used in the casting process. Inclusions appear most often on the casting surface and are usually detected by visual inspection, but in many cases the internal walls of castings contain inclusions that cannot be visually detected. Internal inclusions can be detected by eddy current, radiographic, or ultrasonic inspection; radiography is usually the most reliable method. Poor Metal Structure. Many casting defects resulting from metal structure are related to shrinkage, which is either a cavity or a spongy area linked with dendrites or a depression in the casting surface. This type of defect arises from varying rates of contraction while the metal is changing from a liquid to a solid. Other casting defects resulting from varying rates of contraction during solidification include carbide formation, hardness variations, and microporosity.

Internal shrinkage defects are best detected by radiography, although eddy current or ultrasonic inspection can be used. Soft or hard gray iron castings are usually detected by Brinell hardness testing; electromagnetic methods have proved useful on some castings. Interrupted Metal Walls. Included in this category are such flaws as hot tears, cold shuts, and casting cracks.

Cracking of castings is often a major problem in gray iron foundries because of the combination of casting designs and high production rates. Visual inspection or an aided visual method such as liquid penetrant or magnetic particle inspection is used to detect cracks and cracklike flaws in castings. Mold wall deficiencies are common problems in gray iron castings. They result in surface flaws such as scabs, rattails, cuts, washes, buckles, drops, and excessive metal penetration into the spaces between sand grains. These flaws are generally detected by visual inspection. Size and Quantity of Graphite. In cast iron, the length of the lamellae (flakes)--that is, the coarseness of the

graphite--is expressed by code numbers from 1 to 8, as described in ASTM A 247 and in Casting, Volume 15 of ASM Handbook, formerly 9th Edition Metals Handbook. These numbers correspond to lamellar lengths ranging from 1.25 to 0.01 mm (0.05 to 0.0004 in.), as viewed in micrographs of the cast iron structure. The dependence of the longitudinal wave sonic velocity on the size and quantity of graphite is shown in Fig. 15; with decreasing coarseness and quantity of graphite, the velocity approaches that in steel. In microlamellar cast iron, the amount of graphite is usually smaller. Therefore, the finer the graphite, the higher the sonic velocity. In both lamellar and spheroidal cast irons, the transverse wave sonic velocity is related to the longitudinal wave velocity, as shown in Fig. 16.

Fig. 15 Variation of longitudinal wave sonic velocity with graphite size for lamellar cast iron containing different percentages of free graphite. Scale at right indicates ratio of sonic velocity in cast iron to corresponding sonic velocity in steel.

Fig. 16 Relation of longitudinal wave and transverse wave sonic velocities for spheroidal and lamellar cast irons and for steel

Malleable Iron Castings Blowholes and spikes are defects often found in malleable iron castings. Spikes are a form of surface shrinkage not normally visible to the naked eye, but they appear as a multitude of short, discontinuous surface cracks when subjected to fluorescent magnetic particle inspection. Unlike true fractures, spikes do not propagate, but they are not acceptable where in.) long or cyclic loading could result in fatigue failure. Spikes are usually seen as short indications about 1.6 mm ( less and never more than 75 μm (0.003 in.) deep. These defects do not have a preferred orientation, but a random pattern that may or may not follow the direction of solidification. Shrinkage or open structure in the gated area is a defect often found in malleable iron castings that may be overlooked by visual inspection, although it is readily detected by either liquid penetrant or magnetic particle inspection. Ductile Iron Castings Ductile iron is cast iron in which the graphite is present in tiny balls or spherulites instead of flakes (as in gray iron) or compacted aggregates (as in malleable iron). The spheroidal graphite structure is produced by the addition of one or more elements to the molten metal. Soundness and Integrity. Cracks and fine tears that break the surface of the casting but are difficult to detect

visually can be revealed with dye penetrants or magnetic particle inspection. Modern techniques of magnetizing the casting, followed by the application of fluorescent magnetic inks, are very effective and widely used. Methods of sonic testing that involve vibrating the casting and noting electronically the rate of decay of resonant frequency or damping behavior are also used to detect cracked or flawed castings. Internal unsoundness, when not immediately subsurface, can be detected with ultrasonic inspection by the failure to observe a back wall echo when using reflected radiation or by a weakening of the signal in the transmission through the casting. Coupling of the probes and interpretation of the results involve operator skill, but methods are available that consist of partial or total immersion of the casting in a liquid, automatic or semiautomatic handling of the probes, and computer signal processing to ensure more reliable and consistent interpretation of results. Problems arise in detecting very-near-surface defects and when examining thin castings, but the use of angled probes and shear wave techniques has yielded good results. The soundness of the ductile iron can also be assessed by x-ray or γ-ray examination. The presence of graphite, especially in heavy sections, makes the method more difficult to evaluate than for steels, but the use of image intensification by electronic means offers considerable promise, especially for sections up to 50 mm (2 in.) thick.

Confirmation of Graphite Structure (Ref 4). Both the velocity of ultrasonic transmission and the resonant

frequency of a casting can be related to the modulus of elasticity. In cast iron, the change from flake graphite to nodular graphite is related to an increase in both modulus of elasticity and strength; therefore, ultrasonic velocity or resonant frequency measurement can be employed as a guide to modularity, strength, and other related properties. Because the microscopic estimation of nodularity is a subjective measurement, these other nondestructive examination methods may provide a better guide to some properties provided the matrix remains constant (Ref 5). Figure 17 illustrates how ultrasonic velocity may vary with graphite nodularity.

Fig. 17 Ultrasonic velocity versus visually assessed modularity in ductile iron castings

Ultrasonic transmission measurement is conducted with two probes on either side of the casting. This method provides a guide to the local properties. It must be coupled with a thickness measurement, and automatic equipment is commonly used, often involving immersion of the casting in a tank of fluid (Fig. 18). The calculation does not require calibration of the castings. Simple caliper devices have also been used for examining castings and simultaneously measuring their thickness to provide a calculated value of ultrasonic velocity and a guarantee of modularity (Ref 6).

Fig. 18 Ultrasonic test equipment used for determining the thickness, nodularity, and integrity of ductile iron castings. (a) Schematic of setup for the ultrasonic velocity testing for structure evaluation. (b) Photograph of test instrument used for integrity/nodularity studies showing controls and instrumentations. Courtesy of J. Johnston, Krautkramer Branson

Sonic testing involves measurement of the resonant frequency of the casting or rate of decay of resonance of a casting that has first been excited by mechanical or electrical means. This method evaluates the graphite structure of the entire casting and requires calibration of the castings against standard castings of known structure. It is also necessary to maintain casting dimensions within a well-controlled narrow range. Some foundries use sonic testing as a routine method of final inspection and structure guarantee (Ref 7). The relationship between modularity, resonant frequency, or ultrasonic transmission velocity and properties has been documented for tensile strength, proof strength, fatigue, and impact strength. Examples are shown in Fig. 19 and 20.

Fig. 19 Ultrasonic velocity versus strength in ductile iron castings

Fig. 20 Strength versus resonant frequency in the nondestructive evaluation of ductile iron test bars

The presence of carbides can also be detected with sonic or ultrasonic measurements provided enough carbides are present to reduce the graphite sufficiently to affect the modulus of elasticity. Discrimination between the effects of graphite variation and carbide amount would require an additional test, such as a hardness test.

Properties Partially Dependent on Graphite Structure (Ref 7). When the matrix structure of ductile iron varies,

this variation cannot be detected as easily as variations in graphite structure, and sonic and ultrasonic readings may not be able to reflect variations in mechanical properties. A second measurement, such as a hardness measurement, is then needed to detect matrix variations in the same way as would be necessary to confirm the presence of carbides. Eddy current or coercive force measurements can be used to detect many changes in casting structure and properties; but the indications from such measurements are difficult to interpret, and the test is difficult to apply to many castings unless they are quite small and can be passed through a coil 100 to 200 mm (4 to 8 in.) in diameter. Eddy current indications are, however, useful for evaluating pearlite and carbide in the iron matrix. Multifrequency eddy current testing uses probes that do not require the casting to pass through a coil, it is less sensitive to casting size, and it allows automatic measurements and calculations to be made; but the results remain difficult to interpret with reliability in all cases. It may, however, be a very good way to detect chill and hard edges on castings of reproducible dimensions.

References cited in this section

4. A.G. Fuller, P.J. Emerson, and G.F.Sergeant, A Report on the Effect Upon Mechanical Properties of Variation in Graphite Form in Irons Having Varying Amounts of Ferrite and Pearlite in the Matrix Structure and the Use of Nondestructive Tests in the Assessments of Mechanical Properties of Such Irons, Trans. AFS, Vol 88, 1980, p 21-50 5. A.G. Fuller, Evaluation of the Graphite Form in Pearlitic Ductile Iron by Ultrasonic and Sonic Testing and the Effect of Graphite Form on Mechanical Properties, Trans. AFS, Vol 85, 1977, p 509-526 6. P.J. Rickards, "Progress in Guaranteeing Quality Through Non-Destructive Methods of Evaluation," Paper 21, presented at the 54th International Foundry Congress, New Delhi, The International Committee of Foundry Technical Associations (CIATF), Nov 1987 7. A.G. Fuller, Nondestructive Assessment of the Properties of Ductile Iron Castings, Trans. AFS, Vol 88, 1980, p 751-768 Nondestructive Inspection of Castings By the ASM Committee on Nondestructive Inspection of Castings*

Inspection of Aluminum Alloy Castings Effective quality control is needed at every step in the production of an aluminum alloy casting, from selection of the casting method, casting design, and alloy to mold production, foundry technique, machining, finishing, and inspection. Visual methods, such as visual inspection, pressure testing, liquid penetrant inspection, ultrasonic inspection, radiographic inspection, and metallographic examination, can be used to inspect for casting quality. The inspection procedure used should be geared toward the specified level of quality. Information on casting processes, solidification, hydrogen content, silicon modification, grain refinement, and other topics related to aluminum alloy castings is provided in the articles "Solidification of Eutectic Alloys: Aluminum-Silicon Alloys," "Nonferrous Molten Metal Processes," and "Aluminum and Aluminum Alloys" in Casting, Volume 15 of ASM Handboook, formerly 9th Edition Metals Handbook. Stages of Inspection. Inspection can be divided into three stages: preliminary, intermediate, and final. After tests are

conducted on the melt for hydrogen content, for adequacy of silicon modification, and for degree of grain refinement, preliminary inspection may consist of the inspection and testing of test bars cast with the molten alloy at the same time the production castings are poured. These test bars are used to check the quality of the alloy and the effectiveness of the heat treatment. Preliminary inspection also includes chemical or spectrographic analysis of the casting, thus ensuring that the melting and pouring operations have resulted in an alloy of the desired composition. Intermediate inspection, or hot inspection, is performed on the casting as it is taken from the mold. This step is essential because castings that are obviously defective can be discarded at this stage of production. Castings that are judged unacceptable at this stage can then be considered for salvage by impregnation, welding, or other methods, depending on

the type of flaw present and the end use of the casting. More complex castings usually undergo visual and dimensional inspection after the removal of gates and risers. Final inspection establishes the quality of the finished casting through the use of any of the methods previously mentioned. Visual inspection also includes the final measurement and comparison of specified and actual dimensions. Dimensions of castings from a large production run can be checked with gages, jigs, fixtures, or coordinate measuring systems. Liquid penetrant inspection is extensively used as a visual aid for detecting surface flaws in aluminum alloy

castings. Liquid penetrant inspection is applicable to castings made from all the aluminum casting alloys as well as castings produced by all methods. One of the most useful applications, however, is the inspection of small castings produced in permanent molds from alloys such as 296.0, which are characteristically susceptible to hot cracking. For example, in cast connecting rods, hot shortness may result in fine cracks in the shank sections. Such cracks are virtually undetectable by unaided visual inspection, but are readily detectable by liquid penetrant inspection. All the well-known liquid penetrant systems (that is, water-washable, postemulsifiable, and solvent-removable) are applicable to the inspection of aluminum alloy castings. In some cases, especially for certain high-integrity castings, more than one system can be used. Selection of the system is primarily based on the size and shape of the castings, surface roughness, production quantities, sensitivity level desired, and available inspection facilities. Pressure testing is used for castings that must be leaktight. Cored-out passages and internal cavities are first sealed off

with special fixtures having air inlets. These inlets are used to build up the air pressure on the inside of the casting. The entire casting is then immersed in a tank of water, or it is covered by a soap solution. Bubbles will mark any point of air leakage. Radiographic inspection is a very effective means of detecting such conditions as cold shuts, internal shrinkage,

porosity, core shifts, and inclusions in aluminum alloy castings. Radiography can also be used to measure the thickness of specific sections. Aluminum alloy castings are ideally suited to examination by radiography because of their relatively low density; a given thickness of aluminum alloy can be penetrated with about one-third the power required for penetrating the same thickness of steel. Aluminum alloy castings are most often radiographed with an x-ray machine, using film to record the results. Real-time (digital) radiography and computed tomography are also widely used and are best suited to detecting shrinkage, porosity, and core shift (Fig. 12 and 13). Gamma-ray radiography is also satisfactory for detecting specific conditions in aluminum castings. Although the γ-ray method is used to a lesser extent than the x-ray method, it is about equally as effective for detecting flaws or measuring specific conditions. Aluminum alloy castings are most often radiographed to detect approximately the same types of flaws that may exist in other types of castings, that is, conditions such as porosity or shrinkage, which register as low-density spots or areas and appear blacker on the film or real-time image screen than the areas of sound metal. Aluminum ingots may contain hidden internal cracks of varying dimensions. Depending on size and location, these cracks may cause an ingot to split during mechanical working and thermal treatment, or they may appear as a discontinuity in the final wrought product. Once the size and location of such cracks are determined, an ingot can be scrapped, or sections free from cracks can be sawed out and processed further. Because the major dimensions of the cracks are along the casting direction, they present good reflecting surfaces for sound waves traveling perpendicular to the casting direction. Therefore, ultrasonic methods using a wave frequency that gives adequate penetration into the ingot provide excellent sensitivity for 100% inspection of that part of the ingot containing critical cracks. Because of ingot thickness (up to 400 mm, or 16 in.) and the small metal separation across the crack, radiographic methods are impractical for inspection. Ultrasonic Inspection. Aluminum alloy castings are sometimes inspected by ultrasonic methods to evaluate internal

soundness or wall thickness. The principal uses of ultrasonic inspection for aluminum alloy castings include the detection of porosity in castings and internal cracks in ingots.

Nondestructive Inspection of Castings By the ASM Committee on Nondestructive Inspection of Castings*

Inspection of Copper and Copper Alloy Castings The inspection of copper and copper alloy castings is generally limited to visual and liquid penetrant inspection of the surface, along with radiographic inspection for internal discontinuities. In specific cases, electrical conductivity tests and ultrasonic inspection can be applied, although the usual relatively large cast grain size could prevent a successful ultrasonic inspection. Visual inspection is simple yet informative. A visual inspection would include significant dimensional measurements as well as general appearance. Surface discontinuities often indicate the presence of internal discontinuities. For small castings produced in reasonable volume, a destructive metallographic inspection on randomly selected samples is practical and economical. This is especially true on a new casting for which foundry practice has not been optimized and a satisfactory repeatability level has not been achieved. For castings of some of the harder and stronger alloys, a hardness test is a good means of estimating the level of mechanical properties. Hardness tests are of less value for the softer tin bronze alloys because hardness tests do not reflect casting soundness and integrity. Because copper alloys are nonmagnetic, magnetic particle inspection cannot be used to detect surface cracks. Instead, liquid penetrant inspection is recommended. Ordinarily, liquid penetrant inspection requires some prior cleaning of the casting to highlight the full detail. For the detection of internal defects, radiographic inspection is recommended. Radiographic methods and standards are well established for some copper alloy castings (for example, ASTM E 272 and E 310). As a general rule, the method of inspection applied to some of the first castings made from a new pattern should include all those methods that provide a basis for judgment of the acceptability of the casting for the intended application. Any deficiencies or defects should be reviewed and the degree of perfection defined. This procedure can be repeated on successive production runs until repeatability has been ensured. Gas Porosity. Copper and many copper alloys have a high affinity for hydrogen, with an increasing solubility as the

temperature of the molten bath is increased. Conversely, as the metal cools in the mold, most of this hydrogen is rejected from the metal. Because all the gas does not necessarily escape to the atmosphere and may become entrapped by the solidifying process, gas porosity may be found in the casting. In most alloys, gas porosity is identified by the presence of voids that are relatively spherical and are bright and shiny inside. Visible upon sectioning or by radiography, they may either be small, numerous, and rather widely dispersed or fewer in number and relatively large. Regardless of size, they are seldom interconnected except in some of the tin bronze alloys, which solidify in a very dendritic mode. In these alloys, the gas porosity tends to be distributed in the interstices between the dendrites. Shrinkage voids caused by the change in volume from liquid to solid in copper alloys are different only in degree and

possibly shape from those found in other metals and alloys. All nonferrous metals exhibit this volume shrinkage when solidifying from the molten condition. Shrinkage voids may be open to the air when near or exposed to the surface, or they may be deep inside the thicker sections of the casting. They are usually irregular in shape, compared to gas-generated defects, in that their shape frequently reflects the internal temperature gradients induced by the external shape of the casting. Hot Tearing. The tin bronzes as a class, as well as a few of the leaded yellow brasses, are susceptible to hot shortness;

that is, they lack ductility and strength at elevated temperature. This is significant in that tearing and cracking can take place during cooling in the mold because of mold or core restraint. In aggravated instances, the resulting hot tears in the

part appear as readily visible cracks. Sometimes, however, the cracks are not visible externally and are not detectable until after machining. In extreme cases, the cracks become evident only through field failure because the tearing was deep inside the casting. Nonmetallic inclusions in copper alloys, as with all molten alloys, are normally the result of improper melting and/or

pouring conditions. In the melting operation, the use of dirty remelt or dirty crucibles, poor furnace linings, or dirty stirring rods can introduce nonmetallic inclusions into the melt. Similarly, poor gating design and pouring practice can produce turbulence and can generate nonmetallic inclusions. Sand inclusions may also be evident as the result of improper sand and core practice. All commercial metals, by the nature of available commercial melting and molding processes, usually contain very minor amounts of small nonmetallic inclusions. These have little or no effect on the casting. Inclusions of significant size or number are considered detrimental. A thorough review of copper alloy melting, refining, and casting practices is available in Casting, Volume 15 of ASM Handbook, formerly 9th Edition Metals Handbook. Nondestructive Inspection of Castings By the ASM Committee on Nondestructive Inspection of Castings*

References 1. P.J. Rickards, Progress in Guaranteeing Quality Through Nondestructive Methods of Evaluation, Foundryman Int., April 1988, p 196-209 2. P.M. Bralower, Nondestructive Testing. Part I. The New Generation in Radiography, Mod. Cast., Vol 76 (No. 7), July 1986, p 21-23 3. R.A. Armistead, CT: Quantitative 3-D Inspection, Adv. Mater. Process., Vol 133 (No. 3), March 1988, p 42-48 4. A.G. Fuller, P.J. Emerson, and G.F.Sergeant, A Report on the Effect Upon Mechanical Properties of Variation in Graphite Form in Irons Having Varying Amounts of Ferrite and Pearlite in the Matrix Structure and the Use of Nondestructive Tests in the Assessments of Mechanical Properties of Such Irons, Trans. AFS, Vol 88, 1980, p 21-50 5. A.G. Fuller, Evaluation of the Graphite Form in Pearlitic Ductile Iron by Ultrasonic and Sonic Testing and the Effect of Graphite Form on Mechanical Properties, Trans. AFS, Vol 85, 1977, p 509-526 6. P.J. Rickards, "Progress in Guaranteeing Quality Through Non-Destructive Methods of Evaluation," Paper 21, presented at the 54th International Foundry Congress, New Delhi, The International Committee of Foundry Technical Associations (CIATF), Nov 1987 7. A.G. Fuller, Nondestructive Assessment of the Properties of Ductile Iron Castings, Trans. AFS, Vol 88, 1980, p 751-768 Nondestructive Inspection of Powder Metallurgy Parts R.C. O'Brien and W.B. James, Hoeganaes Corporation

Introduction THE PROBLEM of forming defects in green parts during compaction and ejection has become more prevalent as parts producers have begun to use higher compaction pressures in an effort to achieve high-density, high-performance powder metallurgy (P/M) steels. In this article, several nondestructive inspection methods are evaluated, with the aim of identifying those that are practical for detecting defects as early as possible in the production sequence.

The most promising nondestructive testing methods for P/M applications include electrical resistivity testing, eddy current and magnetic bridge testing, magnetic particle inspection, ultrasonic testing, x-ray radiography, gas permeability testing, and -ray density determination. The capabilities and limitations of each of the techniques are evaluated in this article.

Nondestructive Inspection of Powder Metallurgy Parts R.C. O'Brien and W.B. James, Hoeganaes Corporation

Current Status of P/M Testing In the ceramics industry, the fraction of the finished-part cost that arises from scrap due to flaws introduced during processing is estimated to average 50%, and it can be as high as 75% (Ref 1). Although the ceramics industry has been mobilized for the past 15 years toward the use of nondestructive evaluation in processing, the P/M industry has built up only a scattered background of experience (Ref 2). To remain competitive, P/M parts producers have increasingly turned to simplified processing. It has been shown that the physical properties of P/M parts, especially the fatigue strength, are always improved by increasing the density (Ref 3). The need for densification by double pressing can often be avoided by pressing to high density in a single step. However, the use of higher compaction pressures requires the utmost attention to materials selection, tool design, and press setup (Ref 4). A quick, preferably nondestructive method of crack detection would be of great benefit during press setup and for testing the integrity of parts as early as possible in their production sequence. The growth of nondestructive testing in the 1980s has been explosive, and the field has benefited greatly from computerized image reconstruction techniques applied to radiography, ultrasonic, and even magnetic particle inspection. Commercial test systems are being marketed as fast as the technology is developed, and the metal powder industry should find solutions to its on-line testing needs by reviewing methods being used by other parts fabrication technologies. In preparing this article, a number of test methods presented themselves as having potential for crack detection in green (unsintered) P/M compacts, and these are recommended for further investigation. For a detailed overview of P/M technology, the reader is referred to Powder Metal Technologies and Applications, Volume 7 of the ASM Handbook.

References cited in this section

1. J.W. McCauley, Materials Testing in the 21st Century, in Nondestructive Testing of High Performance Ceramics, Conference Proceedings, American Ceramics Society/American Society for Nondestructive Testing, 1987, p 1 2. R.W. McClung and D.R. Johnson, Needs Assessment for NDT and Characterization of Ceramics: Assessment of Inspection Technology for Green State and Sintered Ceramics, in Nondestructive Testing of High Performance Ceramics, Conference Proceedings, American Ceramics Society/American Society for Nondestructive Testing, 1987, p 33 3. R.C. O'Brien, "Fatigue Properties of P/M Materials," SAE Technical Paper 880165, Society of Automotive Engineers, March 1988 4. G.F. Bocchini, "High Pressure Compaction, High Pressure Coining, and High Pressure Repressing of P/M Parts," Paper presented at the Prevention and Detection of Cracks in Ferrous P/M Parts Seminar, Metal Powder Industries Federation, 1988

Nondestructive Inspection of Powder Metallurgy Parts R.C. O'Brien and W.B. James, Hoeganaes Corporation

Summary of Defect Types in P/M Parts The four most common types of defects in P/M parts are ejection cracks, density variations, microlaminations, and poor sintering. Ejection Cracks. When a part has been pressed, there is a large residual stress in the part due to the constraint of the die and punches, which is relieved as the part is ejected from the die. The strain associated with this stress relief depends on the compacting pressure, the green expansion of the material being compacted, and the rigidity of the die. Green expansion, also known as spring out, is the difference between the ejected-part size and the die size. A typical value of green expansion for a powder mix based on atomized iron powder pressed at relatively high pressure (600 to 700 MPa, or 45 to 50 tsi) is 0.20%. In a partially ejected compact, for example, the portion that is out of the die expands to relieve the residual stress, while the constrained portion remains die size and a shear stress is imposed on the compact. When the ability of the powder compact to accommodate the shear stress is exceeded, ejection cracks such as the one shown in Fig. 1 are formed.

The radial strain can be alleviated to a degree by increasing the die rigidity and designing some release into the die cavity. However, assuming that the ejection punch motions are properly coordinated, the successful ejection of multilevel parts depends to a large degree on the use of a high-quality powder that combines high green strength with low green expansion and low stripping pressure. Density Variations. Even in the simplest tool geometry

possible--a solid circular cylinder--conventional pressing of a part to an overall relative density of 80% will result in a distribution of density within the part ranging from 72 to 82% Fig. 1 Ejection crack in sintered P/M steel. Unetched (Ref 5). The addition of simple features such as a central hole and gear teeth presents minor problems compared with the introduction of a step or second level in the part. Depending on the severity of the step, a separate, independently actuated punch can be required for each level of the part. During the very early stage of compaction, the powder redistributes itself by flowing between sections of the die cavity. However, when the pressure increases and the powder movement is restricted, shearing of the compact in planes parallel to the punch axis can only be avoided by proper coordination of punch motions. When such shear exists, a density gradient results. The density gradient is not always severe enough for an associated crack to form upon ejection. However, a low-density area around an internal corner, as shown in Fig. 2, can be a fatal flaw, because this corner is usually a point of stress concentration when the part is loaded in service.

Fig. 2 Density gradient around an internal corner in a part made with a single-piece stepped punch. Unetched

Microlaminations. In photomicrographs of unetched part cross sections, microlaminations such as those shown in Fig.

3 appear as layers of unsintered interparticle boundaries that are oriented in planes normal to the punch axis. They can be the result of fine microcracks associated with shear stresses upon ejection; such microcracks fail to heal during sintering. Because of their orientation parallel to the tensile axis of standard test bars, they have little influence on the measured tensile properties of the bars, but are presumed to be a cause of severe anisotropy of tensile properties.

Fig. 3 Microlaminations in sintered P/M steel. Unetched

Poor Sintering. When unsintered particle boundaries result from a cause other than shear stresses, they are usually

present because of insufficient sintering time or sintering temperature, a nonreducing atmosphere, poor lubricant burn-off, inhibition of graphite dissolution, or a combination of these. A severe example is shown in Fig. 4. Unlike microlaminations, defects associated with a poor degree of sintering are not oriented in planes.

Fig. 4 Poor degree of sintering in P/M compact. Unetched

Reference cited in this section

5. F.V. Lenel, Powder Metallurgy Principles and Application, Metal Powder Industries Federation, 1980, p 112 Nondestructive Inspection of Powder Metallurgy Parts R.C. O'Brien and W.B. James, Hoeganaes Corporation

Nondestructive Tests and Their Applicability to P/M Processing As described below, applicable inspection methods for P/M parts can be broadly classified into the following categories: • •

Radiographic techniques Acoustic methods

• • •

Thermal inspection Electrical resistivity inspection Visual inspection and pressure testing

The techniques covered in this article are summarized in Table 1. Additional information on these procedures can be found in the Section "Methods of Nondestructive Evaluation" in this Volume. Table 1 Comparison of the applicability of various nondestructive evaluation methods to flaw detection in P/M parts Method

Measured/detected

Applicability to P/M parts(a)

Green

Sintered

Advantages

Disadvantages

X-ray radiography

Density variations, cracks, inclusions

C

C

Can be automated

Relatively high initial cost; radiation hazard

Computed tomography

Density variations, cracks, inclusions

C

C

Can be automated; pinpoint defect location

Extremely high initial cost; highly trained operator required; radiation hazard

Gamma-ray density determination

Density variations

A

A

High resolution and accuracy; relatively fast

High initial cost; radiation hazard

Ultrasonic imaging: C-scan

Density variations, cracks

D

B

Sensitive to cracks; fast

Coupling agent required

Ultrasonic imaging: SLAM

Density variations, cracks

D

C

Fast; high resolution

High initial cost; coupling agent required

Resonance testing

Overall density, cracks

D

B

Low cost; fast

Does not give information on defect location

Acoustic emission

Cracking during pressing and ejection

C

D

Low cost

Exploratory

Thermal wave imaging

Subsurface cracks, density variations

D

C

No coupling agent required

Flat or convex surfaces only

Electrical resistivity

Subsurface cracks, density variations, degree of sinter

A

A

Low cost, portable, high potential for use on green compacts

Sensitive to edge effects

Eddy current/magnetic bridge

Cracks, overall density, hardness, chemistry

C

A

Low cost, fast, can be automated; used on P/M valve seat inserts

Under development;

Magnetic particle inspection

Surface and near-surface cracks

C

A

Simple to operate, low cost

Slow; operator sensitive

Liquid dye penetrant inspection

Surface cracks

C

D

Low cost

Very slow; cracks must intersect surface

Pore pressure rupture/gas permeability

Laminations, ejections, cracks, sintered density variations

A

A

Low cost, simple, fast

Gas-tight fixture required; cracks in green parts must intersect surface

(a) A, has been used in the production of commercial P/M parts; B, under development for use in P/M; C, could be developed for use in P/M, but no published trials yet; D, low probability of successful application to P/M

Radiographic Techniques X-Ray Radiography. Any feature of a part that either reduces or increases x-ray attenuation will be resolvable by x-ray radiography. Some types of flaws and their x-ray images are shown in Fig. 5.

Fig. 5 Schematic of flaws and their x-ray images. Defect types that can be detected by x-ray radiography are those that change the attenuation of the transmitted x-rays. Source: Ref 6

The ability to detect defects depends on their orientation to the x-ray source. A crack parallel to the x-rays will result in reduced attenuation of the rays, and the x-ray film will be darker in this region. A thin crack perpendicular to the x-ray will hardly influence attenuation and will not be detected. Historically, flaw detection by x-ray radiography has been an expensive and cumbersome process suited only to safetycritical and high added value parts. The process has been considerably improved by the development of real-time imaging techniques that replace photographic film. Real-time imaging means that parts can be tested rapidly and accepted or rejected on the spot. Real-time x-ray systems include image intensifiers or screens that convert x-rays into visible light and discrete detector arrays that convert x-rays into electronic signals (which are reconstructed by computer for video display). The image in all these systems can be recorded and digitized for image enhancement. The ability of the system to detect flaws is, however, still sensitive to defect orientation. Additional information is available in the article "Radiographic Inspection" in this Volume. Computed tomography is a recently developed version of x-ray radiography that includes highly sophisticated analysis of the detected radiation. A tomographic setup consists of a high-energy photon source, a rotation table for the specimen, a detector array, and the associated data analysis and display equipment, as shown in Fig. 6. The ability to rotate the specimen increases the chance of orienting a defect relative to the x-rays such that it will be detected. The x-ray source and detector array can be raised or lowered to examine different planes through the sample.

Fig. 6 Schematic of computed tomography, which is the reconstruction by computer of a series of tomographic planes (slices) of an object. The transmitted intensity of the fan-shaped beam is processed by computer and the resulting image is displayed on a terminal. Source: Ref 7

In a typical system, the photon source can be a radioisotope such as 60Co, depending on the energy requirements of the individual specimens. The lead aperture around the source acts as a collimator to produce a fan-shaped beam about 5 mm (0.2 in.) thick. The sample is rotated in incremental steps, and the transmitted radiation is detected at each step by computer-controlled detectors situated one every 14 mm (0.55 in.) in a two-dimensional array.

The computer then reconstructs the object using intensity data from a number of scans at different orientations. The output is in the form of a two-dimensional plan in which colors are mapped onto the image according to the intensity of the transmitted radiation. The resolution available depends on the difference in density between the various features of the object. For example, steel pins embedded in polyvinyl chloride plastic are more easily resolvable than aluminum pins of the same diameter (Ref 7). Experiments with P/M samples have shown that density can be measured to better than 1% accuracy, with a spatial resolution of 1 mm (0.04 in.) (Ref 8). The article "Industrial Computed Tomography" in this Volume contains more information on the principles and applications of this technique. Gamma-Ray Density Determination. Local variations in the density of P/M parts have been detected by measuring the attenuation of γ-rays passing through the part (Ref 9). Depending on the material and the dimensions of the part, density can be measured to an accuracy of ±0.2 to ±0.7%, and the technique has been used by P/M parts fabricators in place of immersion density tests as an aid in tool setting.

The apparatus consists of a vertically collimated -ray beam originating from a radioisotope. The beam passes through the sample as shown in Fig. 7 and reaches a detector via a 1 mm (0.04 in.) diam aperture, where the transmitted intensity is measured. The detector consists of a sodium iodide scintillation crystal, which in turn excites a photomultiplier. Exposure time is 1 to 2 min; a 4 mm (0.15 in.) aperture can reduce this time to 30 s at the expense of some resolution. The radiation source of the Gamma Densomat is Americium 241 (60 keV). For high-energy beams, Cesium 137 (660 keV) can be substituted.

Fig. 7 Schematic of the Gamma Densomat. Source: Ref 9

This method has been shown to be particularly useful in cases where the section of the part to be checked is too small for immersion density measurements (Ref 10). Tool life was extended when the method was used for part density checks in order to avoid overloading. Acoustic Methods Ultrasonic Testing. Many characteristics of solids can be determined from the behavior of sound waves propagating in

them. Ultrasonic signals impinged on a sample at one surface are transmitted at speeds and attenuated at rates determined by the density, modulus of elasticity, and continuity of the material. The sound waves are reflected from other surfaces of the sample, including cracks as well as free surfaces. They can be picked up and amplified for display on a CRT screen, as described in Ref 6 and the article "Ultrasonic Inspection" in this Volume. The height and position of the flaw/defect peak indicate its size and location. Although much effort has been directed toward relating sound propagation to the physical properties of P/M materials, little has been written on the detection of cracks by the ultrasonic inspection of porous materials.

Ultrasound Transmission in Green Compacts. The characterization of green compacts by ultrasonic techniques

appears to be hindered by problems of extreme attenuation of the incident signal. In one case, signals of 1 to 20 MHz were transmitted through an 8 mm (0.3 in.) thick compact of atomized iron with 0.2% graphite added (Ref 11). Although the attenuation did not allow back-wall echo measurement, through-transmission measurements indicated that the transmitted intensity had a maximum at 4 to 5 MHz (Fig. 8). Density was found to influence the transmitted intensity, with specimens at 95% relative density allowing some degree of transmission over the entire range of frequencies tested, while specimens at 87% relative density damped the incident signals entirely.

Fig. 8 Ultrasonic spectrum analyzer output showing change in transmitted intensity with density of green compact. Source: Ref 11

Figure 9 shows that the velocity of ultrasonic waves in green compacts is about half the velocity in sintered compacts and that it is essentially invariant with density (Ref 13). It has also been shown that the velocity of ultrasound in green parts is highly anisotropic and that the experimental reproducibility is very poor (Fig. 10). It has been proposed that the anisotropy in velocity is due to the orientation of porosity (Ref 15).

Fig. 9 Effect of density on ultrasonic velocity in green and sintered cylindrical Ancorsteel 1000-B specimens. Source: Ref 12

Fig. 10 Anisotropy of ultrasound in green transverse rupture strength bars. Source: Ref 14

The variation in the velocity of ultrasound with applied pressure during the compaction of ceramic powders has been measured in situ by fixing transducers to the ends of the punches (Ref 16). Unlike the case of finished green P/M compacts, a clear relationship was found between longitudinal wave velocity and compacting pressure (Fig. 11), probably because the constraint of the punches and die forced the individual particles together, providing an efficient acoustic coupling between particles.

Fig. 11 Ultrasonic wave velocity in ceramic powders, measured during compaction. Source: Ref 16

Ultrasound Transmission in Sintered Parts. Early work relating the physical properties of cast iron to the velocity

of sound waves suggested the potential for evaluating P/M steels in the same way (Ref 17). As expected, both the velocity of sound in P/M parts and their resonant frequencies have been related to density, yield strength, and tensile strength. Plain carbon steel P/M specimens were used in one series of tests and the correlation was found to be close enough for the test to be used as a quick check for the degree of sintering in production P/M parts (Ref 12). Other work has demonstrated the relationship between sound velocity and tensile strength in porous parts (Fig. 12). The same types of relationships have also been documented in powder forgings Ref 19.

Fig. 12 Correlation of ultrasonic velocity with tensile strength of sintered steel. Source: Ref 18

Sintered parts have been found to transmit ultrasound according to the relationships shown in Fig. 13. The highest wave velocities occurred in the pressing direction. An additional distinction was found between the velocities in the longitudinal and lateral axes of an oblong specimen, and these results were shown to be reproducible between different powder lots and specimen groups. The anisotropy of velocity diminished at higher densities and disappeared above 6.85 g/cm3.

Fig. 13 Anisotropy of ultrasound velocity in sintered transverse rupture strength bars. Source: Ref 14

Ultrasonic Imaging: C-Scan. The C-Scan is a form of ultrasonic testing in which the testpiece is traversed by the ultrasound transducer in a computer-controlled scan protocol (Fig. 14). The transmitted intensity is recorded and analyzed by computer, and a gray-mapped image is output.

Fig. 14 Schematic of a C-Scan scanning protocol for an adhesive-bonded structure. Source: Ref 20

In one trial, seeded oxide inclusions were detected in porous sintered steels using a C-Scan (Ref 21). The inclusions consisted of admixed particles of chromium oxide and alumina at concentrations of 65 to 120 particles per square centimeter. Inclusions as small as 50 μm in diameter were detected. Additional information on the C-Scan can be found in the articles "Ultrasonic Inspection" and "Adhesive-Bonded Joints" in this Volume. Ultrasonic Imaging: Scanning Acoustic Microscopy (SAM). Ultrasonic waves can be focused on a point using a

transducer and lens assembly, as shown in Fig. 15 and described in the article "Acoustic Microscopy" in this Volume. In this way, the volume of the specimen being examined is highly limited, so that reflections from defects can be closely located at a given depth and position in the specimen. In SAM, the specimen is moved by stepper motors in a raster pattern, and an image of the entire structure can be built up. Scanning acoustic microscopy has been shown to be capable of resolving small surface and subsurface cracks, inclusions, and porosity in sintered, fully dense ceramics (Ref 22).

Fig. 15 Schematic of the image-forming process in scanning acoustic reflecting microscopy. Source: Ref 22

Ultrasonic Imaging: Scanning Laser Acoustic Microscopy (SLAM). When a continuous plane wave impinges

on a sample that is roughly flat in shape, it propagates through and is emitted from the sample with relatively little scattering, retaining its planar nature. When the plane wave is emitted from the sample, it contains information on variations in properties that were encountered in the interior of the sample, which takes the form of variations in intensity with position in the plane. A scanning laser acoustic microscope detects these variations as distortions in a plastic sheet that is placed in the path of the plane wave. The information is gathered by a laser that scans a reflective coating on one side of the sheet, as shown in Fig. 16 and explained in the article "Acoustic Microscopy" in this Volume.

Fig. 16 General configuration used in scanning laser acoustic microscopy. Source: Ref 23

Therefore, although ultrasonic testing is not appropriate for evaluating green P/M parts, it is applicable to the assessment of sintered components. Optimum results dictate careful selection and placement of the transducers because the orientation of the defects influences the ability to detect them. Small defects close to the specimen surface can be masked by surface echoes. Although enhanced image analysis techniques appear beneficial, it is unlikely that the more sophisticated techniques, such as C-Scan and SLAM, will be cost effective for most ferrous P/M parts in the near future. Resonance Testing. When a structural part is tapped lightly, it responds by vibrating at its natural frequency until the

sound is damped. Both the damping characteristics and the natural frequency change with damage to the structure. Changes in the natural frequency can be detected with a spectrum analyzer, as shown in Fig. 17.

Fig. 17 Schematic of resonance test configuration. Source: Ref 24

Sintered P/M parts behave in a similar manner, and the minimum defect size that can be detected has been determined experimentally by testing the resonant frequency after milling narrow grooves of various depths in the parts (Ref 24). It was found that defects covering 2% of the cross section could always be detected and that smaller defects (down to 0.5%) could be detected under favorable conditions of part geometry. This was later shown to apply to real defects as well as machined grooves (Ref 25). There is no record of the technique having been tried on green parts. However, the extremely high sound-damping capacity of green parts would appear to preclude its use. As with ultrasonic techniques, resonance testing has been used to determine physical properties such as the elastic modulus of materials as well as their defect structure (Ref 26). Acoustic emissions are sounds generated in a material as stored elastic energy is released in a noncontinuous mode by

mechanisms such as transformation and twinning, slip, and fracture (see the article "Acoustic Emission Inspection" in this Volume). The acoustic emission spectra have been characterized for the compressive deformation of powder-forged 4600 steels with carbon contents ranging from 0.3 to 0.9% (Ref 27). If P/M tooling were monitored for acoustic emissions of the powder during compaction and ejection, it might be possible to distinguish emission peaks due to the release of stored energy as cracks are formed. However, a developmental program would be required to evaluate this concept and practical application is not anticipated. Thermal Inspection Thermal Wave Imaging. When a pulsed laser impinges on a surface, the rapidly alternating heating and cooling of the surface is conducted into the body of the specimen, as shown in Fig. 18. These thermal waves have been shown to possess many of the same characteristics as electromagnetic or mechanical waves. They can be reflected and refracted, they can form interference patterns, and they interact with irregularities contained in the transmitting medium. In coincidence with the thermal wave formation, acoustic waves are formed by the alternating expansion and contraction of the area of impingement of the laser on the surface. These photoacoustic waves have the same frequency as the thermal waves (typically 1 MHz) but have a much longer wavelength. They are also affected by scattering and reflection of the thermal waves in the volume immediately surrounding the laser impingement point, and it is this effect that allows detection of flaws. Thermal wave imaging has been used to detect delamination and microcracking in silicon integrated circuits (Ref 28).

Fig. 18 Transmission of thermal and acoustic waves in thermal wave imaging. Source: Ref 28

Another method of detecting the interactions of thermal waves with defects is optical beam deflection, or the mirage effect (Ref 29). The impingement point of the laser on the surface heats rapidly, and the air around this point is also heated. If there are no irregularities present beneath the surface, this volume of lower-density heated air is roughly hemispherical in shape. A second laser beam that transits this low-density air volume by skimming closely parallel to the specimen surface, as shown in Fig. 19, will be refracted by the density gradient in the same way as it would be by a conventional lens. A four-quadrant detector array gathers the beam deflection data as the specimen surface is scanned by the laser. Subsurface defects are detected as changes in the shape of the density gradient "lens."

Fig. 19 Detection of interactions between thermal waves and flaws by optical beam deflection (mirage effect). Source: Ref 29

Although there is no record of thermal wave imaging having been applied to P/M parts, the damping capacity of green compacts would appear to restrict the potential application of the technique to sintered components only. Full details on the principles and applications of thermal wave imaging can be found in the article "Thermal Inspection" in this Volume. Electrical Resistivity Testing Direct Current Resistivity Testing. A voltage field within a conductive solid will create currents that are influenced

by structural irregularities, including cracks and porosity. This characteristic has been used to measure carburized case depth in wrought steels (Ref 30). The arrangement shown in Fig. 20 is used to measure the voltage drop in a current field localized between two electrode probes. This method has been used to detect seeded defects in laboratory specimens. It has also been successfully applied to the production of sintered steel parts (Ref 31), as described in Examples 1, 2, and 3.

Fig. 20 Four-point probe used in the resistivity test. The outer probe pins are the current leads; the inner pins are the potential leads. Source: Ref 30

Although the resistivity of green compacts is an order of magnitude higher than that after sintering, the same technique has been shown to apply (Ref 30). Green-state specimens with laboratory-simulated cracks of the type shown in Fig. 21 have been subjected to resistivity inspection with encouraging results. If the probe electrodes span the plane containing the defects and if a series of measurements is made along the edge of the plane, the resistivity varies when defects are present, as shown in Fig. 22. Other tests on green parts are described in Ref 30 and 31.

Fig. 21 Defects in green P/M compacts. (a) Artificial defect caused by the inclusion of a fine wax sliver in the die fill. Unetched. (b) Artificial defect produced by compacting a partially filled die at 345 MPa (25 tsi), completing the fill, and carrying out final compaction of the entire part at 620 MPa (45 tsi). Unetched. See also Fig. 22.

Fig. 22 Variation in resistivity in a green compact was used to locate artificial defects of the type shown in Fig. 21(a) and 21(b). Source: Ref 30

There are two potential contributors to variability in the resistivity test. First, in addition to cracks, the edges and corners of the parts distort the current fields. The internal corners of parts are often the sites of green cracks. Testing the volume of material immediately underlying the corners necessitates the use of specially made electrode probe sets. Another variable influencing the resistivity inspection of green compacts is the nature of the oxide layers on the particles. When the oxide layer is altered with a thermal treatment, the resistivity of the green part decreases (Ref 32). Another study has yielded the relative density/conductivity relationship shown in Fig. 23, suggesting that resistivity tests could be used as a rapid check for localized density variations. As with ultrasound, the elastic modulus and the toughness of porous steels can also be distinguished by resistivity checks (Ref 34).

Fig. 23 Variation in resistivity with relative density in sintered iron. Source: Ref 33

The direct current resistivity test can be used on any conductive material; it is not limited to ferromagnetic materials. Although further development is needed, resistivity measurements appear to be one of the most promising techniques for the nondestructive evaluation of both green and sintered P/M parts. In addition to detecting cracks in green parts, as well as part-to-part density variation, studies have shown that changes in resistivity due to poor carbon pickup during sintering were also detectable (Ref 31). Resistivity testing has also been used later in the processing sequence to screen heat-treated parts for incomplete transformation to martensite. Several uses for resistivity testing are given in the following examples (Ref 31). Example 1: Automotive Air Conditioner Compressor Part. The resistivity-measuring equipment and hand-held probe are shown in Fig. 24. The part, shown in Fig. 25, was tested for green cracking in the locations marked in Fig. 25(b), which were suspect because of prior experience. The parts could then be sorted for cracks by comparing the measured resistivity with limiting resistivity values that had previously been determined using parts with cracks indicated by magnetic particle testing. The prior test method consisted of sintering, sectioning, and magnetic particle inspection, a 2-h process. This part was also the subject of a series of experiments demonstrating that the resistivity test method had high reproducibility and was not operator sensitive.

Fig. 24 Resistivity-measurement device for examining P/M parts. Courtesy of R.A. Ketterer and N.F. McQuiddy, Ferraloy

Fig. 25 Automotive air conditioner compressor part examined by resistivity measurement. (a) Actual part. (b) Cross section showing flawed areas. (c) Six location test fixture. Courtesy of R.A. Ketterer and N.F. McQuiddy, Ferraloy

Example 2: Automotive Transmission Spacer. The resistivity test was used to screen the parts shown in Fig. 26 for incomplete transformation to martensite upon heat treating. The test is based on the lower resistivity of pearlitic microstructures compared with martensitic microstructures of the same chemistry. To determine a resistivity criterion for the screening of these parts, resistivity was correlated with hardness measurements. A resistivity of 60 μΩ· cm was associated with a hardness of 30 HRC, and a go/no-go test strategy was used. The prior test methods for this part were hardness measurements and metallography.

Fig. 26 Automatic transmission spacer examined by resistivity measurement. Courtesy of R.A. Ketterer and N.F. McQuiddy, Ferraloy

Example 3: Automatic Transmission Clutch Plate. The part, shown in Fig. 27, was pressed, sintered, and sized. The resistivity test was then used to screen for part-to-part density variations to levels below 6.8 g/cm3, which was shown to be a minimum density level for achieving the radial crush strength specification for the part. Again, a limiting resistivity value was determined for the part; resistivity values below 27.5 μΩ· cm were considered acceptable.

Fig. 27 Automatic transmission clutch plate examined by resistivity measurement. Courtesy of R.A. Ketterer and N.F. McQuiddy, Ferraloy

Eddy Current Testing. Another form of resistivity testing is the eddy current test. In this test, instead of producing

currents in the part by direct contact with electrodes, eddy currents are induced in the part by an alternating electromagnetic field from an induction coil, as described in the article "Eddy Current Inspection" in this Volume. Single-Coil Tests. Disruptions in the eddy current path due to any defect that changes the resistivity of the material are

detected as extraneous induced voltages in the induction coil. Alternatively, a separate detector coil can be placed in the magnetic field around the testpiece. The alternating current in the induction coil can vary from 1 to 1000 kHz. The depth of penetration varies with frequency, with the highest frequencies yielding the smallest depths (skin effect). The way in which the eddy current varies as a function of depth is also described in the article "Eddy Current Inspection" in this Volume. The output of eddy current testing is in the form of an oscilloscope display. An eddy current inspection system can detect changes from point to point in single testpieces (for example, welded tubes) as they move through the coil. For cases where the testpieces consist of a series of discrete parts, a second coil containing a reference can be added to the system; this configuration is called a magnetic bridge comparator. Magnetic Bridge Comparator Testing. When a ferromagnetic part is placed in the core of a coil with an alternating current, a unique set of harmonics characteristic to the part can be detected in the coil. Some of the variables influencing the harmonics are alloy type, core or surface hardness, case depth, and porosity (Ref 35).

In the magnetic bridge comparator arrangement, the harmonic signals from two like coils are compared. The coils are similar and carry the same excitation waveform. One coil contains the part to be inspected and the other a reference part chosen at random from the group to be tested. Differences between the harmonic characteristics of the two parts are displayed as the displacement of a dot from the center of an oscilloscope screen; no displacement means the two parts are alike. Although the magnetic susceptibility of porous sintered steels is reduced by the pinning of domain-boundary walls by pores, P/M parts are also capable of being analyzed by the magnetic bridge comparator. In one study, 120 P/M production parts were tested in a magnetic bridge comparator. Seventeen of the parts were singled out on the basis of a displacement of the oscilloscope indication, as shown in Fig. 28. These parts were tested for chemistry, hardness, crush strength, and pressed height. For comparison, 25 parts selected at random from the remaining specimens were also tested. Statistically significant differences were found between the groups with regard to carbon content and hardness (Ref 36). The technique has also been successfully applied to powder-forged parts (Ref 37). Although there are no published trials, there is a possibility that the comparator could also be used for testing green compacts.

Fig. 28 Magnetic bridge comparator display for a set of 120 sintered parts, in which 17 parts were indicated as differing from the reference part. Source: Ref 36

Visual Inspection and Pressure Testing Magnetic Particle Inspection. Cracks that exist on or close to the surface of a ferromagnetic material in the magnetic field act as magnetic poles, creating localized increases in the field intensity. Iron particles suspended in a fluid at the surface will be preferentially attracted to these high-intensity areas, and these particles can be used to mark the locations of the flaws. The detectability of the particles themselves can in turn be improved by coating with a pigment that contrasts with the part surface or fluoresces under ultraviolet light (see the article "Magnetic Particle Inspection" in this Volume).

This method has been used to inspect finished P/M parts for cracks originating in processing, and it may also be applicable to green compacts. It is also possible to automate the inspection process by using digital image processing (Ref 38). Liquid Dye Penetrant Inspection. A liquid that wets the surface of the material being inspected will lower its

surface energy by residing preferentially in surface cracks and cavities. In the liquid penetrant inspection technique, cracks are detected by removing the dye from the flat surface of the specimen. The dye that is left behind in the cracks is then wicked out onto the surface by a fine particulate layer in which the pore radius is even lower than that of the crack. The penetrant in this particulate developer layer can be detected visually because of its high contrast with the white developer, or it can be mixed with a dye that fluoresces under ultraviolet light. This process is described in the article "Liquid Penetrant Inspection" in this Volume. The dye penetrant equipment found in P/M shops is generally used only for checking parts of the tooling and machinery for cracks. The dye does not preferentially reside at cracks in P/M parts, because the pore radius and the crack radius are equivalent. However, there might be an application for green parts because the surfaces of green parts are sealed against penetration by liquids through smearing of the metal powder against the die wall and through the formation of a thin coating of dry powder lubricant on the surface. Cracks intersecting the surface may form an opening in this layer that could be detected by the dye penetrant. Pore Pressure Rupture Testing of Green Compacts. A novel test is available for detecting ejection cracks in

green compacts (Ref 39). A pressure seal is formed around a corner or area of a part where experience has shown that cracks are likely to occur. The area is then pressurized to about 3.5 MPa (500 psi) using a fixture such as that shown in Fig. 29. If a crack is present, the gas pressure in the crack will be sufficient to propagate the crack the rest of the way

through the part. This would be classed as a proof test rather than a nondestructive test because the part is destroyed if defects are present.

Fig. 29 Pore pressure rupture test for crack detection in green parts. Source: Ref 39

The test can be used in a nondestructive manner on sintered parts. The gas permeability of the pressurized area is measured at reduced pressures, and the presence of cracks or low-density areas is indicated by high permeability, as shown in Fig. 30.

Fig. 30 Detection of flawed compact using the gas permeability technique. Source: Ref 39

References cited in this section

6. C. Rain, Uncovering Hidden Flaws, High Technol., Feb 1984 7. B. Chang et al., Spatial Resolution in Industrial Tomography, IEEE Trans. Nuclear Sci., NS30 (No. 2), April 1983 8. H. Heidt et al., Nondestructive Density Evaluation of P/M Objects by Computer Tomography, in Horizons of Powder Metallurgy, 1986 International Powder Metallurgy Conference Proceedings, Part II, p 723 9. G. Schlieper, W.J. Huppmann, and A. Kozuch, Nondestructive Determination of Sectional Densities by the Gamma Densomat, Prog. Powder Metall., Vol 43, 1987, p 351 10. C.T. Waldo, Practical Aspects of the Gamma Densomat, in Horizons in Powder Metallurgy, 1986 International Powder Metallurgy Conference Proceedings, Part II, p 739 11. J.L. Rose, M.J. Koczak, and J.W. Raisch, Ultrasonic Determination of Density Variations in Green and Sintered Powder Metallurgy Components, Prog. Powder Metall., Vol 30, 1974, p 131 12. B. Patterson, C. Bates, and W. Knopp, Nondestructive Evaluation of P/M Materials, Prog. Powder Metall., Vol 37, 1981, p 67 13. M.F. Termine, "Ultrasonic Velocity Measurements on Green and Sintered P/M Compacts," Unpublished Report, Hoeganaes Corporation, 1985

14. E.P. Papadakis and B.W. Petersen, Ultrasonic Velocity As A Predictor of Density in Sintered Powder Metal Parts, Mater. Eval., April 1979, p 76 15. A. Gallo and V. Sergi, Orientation of Porosity of P/M Materials Evaluated by Ultrasonic Method, in Horizons in Powder Metallurgy, 1986 International Powder Metallurgy Conference Proceedings, Part II, p 763 16. M.P. Jones and G.V. Blessing, Ultrasonic Evaluation of Spray-Dried Alumina Powder During and After Compaction, in NDT of High Performance Ceramics, Proceedings of the 1987 Conference, American Ceramics Society/American Society for Nondestructive Testing, 1987, p 148 17. B.R. Patterson and C.E. Bates, Nondestructive Property Prediction in Gray Cast Iron Using Ultrasonic Techniques, Paper 65, Trans. AFS, 1981, p 369 18. R.H. Brockelman, Dynamic Elastic Determination of the Properties of Sintered Powder Metals, Perspect. Powder Metall., Vol 5, 1970, p 201 19. E.R. Leheup and J.R. Moon, Yield and Fracture Phenomena in Powder Forged Fe-0.2C and Their Prediction by NDT Methods, Powder Metall., Vol 23 (No. 4), 1980, p 177 20. K. Subramanian and J.L. Rose, C-Scan Testing for Complex Parts, Adv. Mater. Process. inc. Met. Prog., Vol 131 (No. 2), 1987, p 40 21. A. Hecht and E. Neumann, Detection of Small Inclusions in P/M Alloys by Means of Nondestructive Ultrasonic Testing, in Horizons of Powder Metallurgy, 1986 International Powder Metallurgy Conference Proceedings, Part II, P 783 22. G.Y. Baaklini and P.B. Abel, Flaw Imaging and Ultrasonic Techniques for Characterizing Sintered Silicon Carbide, in Nondestructive Testing of High Performance Ceramics, Proceedings of the 1987 Conference, American Ceramics Society/American Society for Nondestructive Testing, 1987, p 304 23. E.R. Generazo and D.J. Roth, Quantitative Flaw Characterization With Scanning Laser Acoustic Microscopy, Mater. Eval., Vol 44 (No. 7), June 1986, p 864 24. P. Cawley, Nondestructive Testing of Mass Produced Components by Natural Frequency Measurements, Proc. Inst. Mech. Eng., Vol 199 (No. B3), 1985, p 161 25. P. Cawley, Rapid Production Quality Control by Vibration Measurements, Mater. Eval., Vol 45 (No. 5), May 1987, p 564 26. R. Phillips and W. Franciscovich, Free-free Resonant Frequency Testing of Powder Metal Alloys to Determine Elastic Moduli, Prog. Powder Metall., Vol 39, 1983, p 369 27. Y. Xu, S.H. Carpenter, and B. Campbell, An Investigation of the Acoustic Emission Generated During the Deformation of Carbon Steel Fabricated by Powder Metallurgy Techniques, J. Acoust. Emiss., Vol 3 (No. 2), 1984, p 81 28. A. Rosencwaig, Thermal Wave Imaging, Science, Vol 218 (No. 4569), 1982, p 223 29. L.J. Inglehart, Photothermal Characterization of Ceramics, in Nondestructive Testing of High Performance Ceramics, Proceedings of 1987 Conference, American Ceramics Society/American Society of Nondestructive Testing, 1987, p 163 30. A. Lewis, "Nondestructive Inspection of Powder Metallurgy Parts Through the Use of Resistivity Measurements," Paper presented at the Prevention and Detection of Cracks in Ferrous P/M Parts Seminar, Metal Powder Industries Federation, 1988 31. R.A. Ketterer and N. McQuiddy, "Resistivity Measurements on P/M Parts: Case Histories," Paper presented at the Prevention and Detection of Cracks in Ferrous P/M Parts Seminar, Metal Powder Industries Federation, 1988 32. E.R. Leheup and J.R. Moon, Electrical Conductivity and Strength Changes in Green Compacts of Iron Powder When Heated in Range 50-400 °C in Air, Powder Metall., Vol 23 (No. 4), 1980, p 217 33. E.R. Leheup and J.R. Moon, Electrical Conductivity Changes During Compaction of Pure Iron Powder, Powder Metall., Vol 21 (No. 4), 1978, p 195 34. E.R. Leheup and J.R. Moon, Relationships Between Density, Electrical Conductivity, Young's Modulus, and Toughness of Porous Iron Samples, Powder Metall., Vol 21 (No. 4), 1978, p 1 35. P. Neumaier, Computer-Aided Tester for Nondestructive Determination of Material Properties, Metallurg.

Plant Technol., No. 3, 1987, p 58 36. R.C. O'Brien, "Analysis of Variance of Sintered Properties of P/M Transmission Parts," Unpublished Report, Hoeganaes Corporation, 1984 37. W.B. James, "Quality Assurance Procedures for Powder Forged Materials," SAE Technical Paper 830364, Society of Automotive Engineers, Feb 1983 38. Y.F. Cheu, Automatic Crack Detection With Computer Vision and Pattern Recognition of Magnetic Particle Indications, Mater. Eval., Vol 42 (No. 11), Nov 1984, p 1506 39. I. Hawkes and C. Spehrley, Point Density Measurement and Flaw Detection in P/M Green Compacts, Mod. Develop. Powder Metall., Vol 5, 1970, p 395 Nondestructive Inspection of Powder Metallurgy Parts R.C. O'Brien and W.B. James, Hoeganaes Corporation

References 1.

2.

3. 4.

5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.

J.W. McCauley, Materials Testing in the 21st Century, in Nondestructive Testing of High Performance Ceramics, Conference Proceedings, American Ceramics Society/American Society for Nondestructive Testing, 1987, p 1 R.W. McClung and D.R. Johnson, Needs Assessment for NDT and Characterization of Ceramics: Assessment of Inspection Technology for Green State and Sintered Ceramics, in Nondestructive Testing of High Performance Ceramics, Conference Proceedings, American Ceramics Society/American Society for Nondestructive Testing, 1987, p 33 R.C. O'Brien, "Fatigue Properties of P/M Materials," SAE Technical Paper 880165, Society of Automotive Engineers, March 1988 G.F. Bocchini, "High Pressure Compaction, High Pressure Coining, and High Pressure Repressing of P/M Parts," Paper presented at the Prevention and Detection of Cracks in Ferrous P/M Parts Seminar, Metal Powder Industries Federation, 1988 F.V. Lenel, Powder Metallurgy Principles and Application, Metal Powder Industries Federation, 1980, p 112 C. Rain, Uncovering Hidden Flaws, High Technol., Feb 1984 B. Chang et al., Spatial Resolution in Industrial Tomography, IEEE Trans. Nuclear Sci., NS30 (No. 2), April 1983 H. Heidt et al., Nondestructive Density Evaluation of P/M Objects by Computer Tomography, in Horizons of Powder Metallurgy, 1986 International Powder Metallurgy Conference Proceedings, Part II, p 723 G. Schlieper, W.J. Huppmann, and A. Kozuch, Nondestructive Determination of Sectional Densities by the Gamma Densomat, Prog. Powder Metall., Vol 43, 1987, p 351 C.T. Waldo, Practical Aspects of the Gamma Densomat, in Horizons in Powder Metallurgy, 1986 International Powder Metallurgy Conference Proceedings, Part II, p 739 J.L. Rose, M.J. Koczak, and J.W. Raisch, Ultrasonic Determination of Density Variations in Green and Sintered Powder Metallurgy Components, Prog. Powder Metall., Vol 30, 1974, p 131 B. Patterson, C. Bates, and W. Knopp, Nondestructive Evaluation of P/M Materials, Prog. Powder Metall., Vol 37, 1981, p 67 M.F. Termine, "Ultrasonic Velocity Measurements on Green and Sintered P/M Compacts," Unpublished Report, Hoeganaes Corporation, 1985 E.P. Papadakis and B.W. Petersen, Ultrasonic Velocity As A Predictor of Density in Sintered Powder Metal Parts, Mater. Eval., April 1979, p 76 A. Gallo and V. Sergi, Orientation of Porosity of P/M Materials Evaluated by Ultrasonic Method, in

16.

17. 18. 19. 20. 21.

22.

23. 24. 25. 26. 27.

28. 29.

30.

31.

32. 33. 34. 35. 36.

Horizons in Powder Metallurgy, 1986 International Powder Metallurgy Conference Proceedings, Part II, p 763 M.P. Jones and G.V. Blessing, Ultrasonic Evaluation of Spray-Dried Alumina Powder During and After Compaction, in NDT of High Performance Ceramics, Proceedings of the 1987 Conference, American Ceramics Society/American Society for Nondestructive Testing, 1987, p 148 B.R. Patterson and C.E. Bates, Nondestructive Property Prediction in Gray Cast Iron Using Ultrasonic Techniques, Paper 65, Trans. AFS, 1981, p 369 R.H. Brockelman, Dynamic Elastic Determination of the Properties of Sintered Powder Metals, Perspect. Powder Metall., Vol 5, 1970, p 201 E.R. Leheup and J.R. Moon, Yield and Fracture Phenomena in Powder Forged Fe-0.2C and Their Prediction by NDT Methods, Powder Metall., Vol 23 (No. 4), 1980, p 177 K. Subramanian and J.L. Rose, C-Scan Testing for Complex Parts, Adv. Mater. Process. inc. Met. Prog., Vol 131 (No. 2), 1987, p 40 A. Hecht and E. Neumann, Detection of Small Inclusions in P/M Alloys by Means of Nondestructive Ultrasonic Testing, in Horizons of Powder Metallurgy, 1986 International Powder Metallurgy Conference Proceedings, Part II, P 783 G.Y. Baaklini and P.B. Abel, Flaw Imaging and Ultrasonic Techniques for Characterizing Sintered Silicon Carbide, in Nondestructive Testing of High Performance Ceramics, Proceedings of the 1987 Conference, American Ceramics Society/American Society for Nondestructive Testing, 1987, p 304 E.R. Generazo and D.J. Roth, Quantitative Flaw Characterization With Scanning Laser Acoustic Microscopy, Mater. Eval., Vol 44 (No. 7), June 1986, p 864 P. Cawley, Nondestructive Testing of Mass Produced Components by Natural Frequency Measurements, Proc. Inst. Mech. Eng., Vol 199 (No. B3), 1985, p 161 P. Cawley, Rapid Production Quality Control by Vibration Measurements, Mater. Eval., Vol 45 (No. 5), May 1987, p 564 R. Phillips and W. Franciscovich, Free-free Resonant Frequency Testing of Powder Metal Alloys to Determine Elastic Moduli, Prog. Powder Metall., Vol 39, 1983, p 369 Y. Xu, S.H. Carpenter, and B. Campbell, An Investigation of the Acoustic Emission Generated During the Deformation of Carbon Steel Fabricated by Powder Metallurgy Techniques, J. Acoust. Emiss., Vol 3 (No. 2), 1984, p 81 A. Rosencwaig, Thermal Wave Imaging, Science, Vol 218 (No. 4569), 1982, p 223 L.J. Inglehart, Photothermal Characterization of Ceramics, in Nondestructive Testing of High Performance Ceramics, Proceedings of 1987 Conference, American Ceramics Society/American Society of Nondestructive Testing, 1987, p 163 A. Lewis, "Nondestructive Inspection of Powder Metallurgy Parts Through the Use of Resistivity Measurements," Paper presented at the Prevention and Detection of Cracks in Ferrous P/M Parts Seminar, Metal Powder Industries Federation, 1988 R.A. Ketterer and N. McQuiddy, "Resistivity Measurements on P/M Parts: Case Histories," Paper presented at the Prevention and Detection of Cracks in Ferrous P/M Parts Seminar, Metal Powder Industries Federation, 1988 E.R. Leheup and J.R. Moon, Electrical Conductivity and Strength Changes in Green Compacts of Iron Powder When Heated in Range 50-400 °C in Air, Powder Metall., Vol 23 (No. 4), 1980, p 217 E.R. Leheup and J.R. Moon, Electrical Conductivity Changes During Compaction of Pure Iron Powder, Powder Metall., Vol 21 (No. 4), 1978, p 195 E.R. Leheup and J.R. Moon, Relationships Between Density, Electrical Conductivity, Young's Modulus, and Toughness of Porous Iron Samples, Powder Metall., Vol 21 (No. 4), 1978, p 1 P. Neumaier, Computer-Aided Tester for Nondestructive Determination of Material Properties, Metallurg. Plant Technol., No. 3, 1987, p 58 R.C. O'Brien, "Analysis of Variance of Sintered Properties of P/M Transmission Parts," Unpublished Report, Hoeganaes Corporation, 1984

37. W.B. James, "Quality Assurance Procedures for Powder Forged Materials," SAE Technical Paper 830364, Society of Automotive Engineers, Feb 1983 38. Y.F. Cheu, Automatic Crack Detection With Computer Vision and Pattern Recognition of Magnetic Particle Indications, Mater. Eval., Vol 42 (No. 11), Nov 1984, p 1506 39. I. Hawkes and C. Spehrley, Point Density Measurement and Flaw Detection in P/M Green Compacts, Mod. Develop. Powder Metall., Vol 5, 1970, p 395 Nondestructive Inspection of Steel Bar, Wire, and Billets

Introduction THE INSPECTION OF STEEL BARS will be the focus of this article, but the principles involved also apply for the most part to steel wire. In many cases, as far as nondestructive inspection is concerned, steel bars and wire are the same. Steel billets, mainly because their irregularities differ significantly from the irregularities found in bars, require special inspection techniques. Billet inspection is covered in the section "Nondestructive Inspection of Steel Billets" in this article. The primary objective in the nondestructive inspection of steel bars and wire is generally the same as for the inspection of other products, that is, to detect conditions in the material that may be detrimental to the satisfactory end use of the product. There is, however, an additional objective in attempting to detect undesirable conditions in semifinished products such as bars, namely, to eliminate unacceptable material before spending time, money, and energy in manufacturing products that will later be rejected. The nondestructive inspection of bars and other semifinished products does not impair the product, provides rapid feedback of information, and can be utilized as either an in-line or off-line system. It makes use of several devices, such as visual, audio, and electromagnetic, for the detection of flaws and of variations in composition, hardness, and grain structure. A wide range of selectivity is provided in each device, permitting acceptance or rejection at various specification levels. The most common function of nondestructive inspection in the steel industry is the detection and evaluation of flaws. It is also used for the detection of variations in composition and physical properties. No amount of nondestructive inspection can ensure an absolutely flawless bar, but it does provide a consistent specified degree of quality during everyday operation. Nondestructive Inspection of Steel Bar, Wire, and Billets

Types of Flaws Encountered The terms used for the various types of flaws discussed in this article may not be the same in various geographic areas. In many cases, different terms are applied to the same type of flaw. Therefore, this section contains a description and an illustration of each condition. The term flaw is applied to blemishes, imperfections, faults, or other conditions that may nullify acceptability of the material. The term also encompasses such terms as pipe, porosity, laminations, slivers, scabs, pits, embedded scale, cracks, seams, laps, and chevrons, as well as blisters and slag inclusions in hot-rolled products. For products that are cold drawn, die scratches may be added. Most flaws in steel bars can be traced back to the pouring of the hot metal into molds. Factors that work against obtaining a perfect homogeneous product include: • • •

The fast shrinkage of steel as it cools (roughly 5% in volume) The gaseous products that are trapped by the solidifying metal as they try to escape from the liquid and semisolid metal Small crevices in the mold walls, which cause the metal to tear during the stripping operation

•

Spatter during pouring, which produces globs of metal frozen on the mold walls because of the great difference in temperature of the mold surfaces and the liquid metal

Pipe is a condition that develops in the nominal top centerline of the ingot as the result of solidification of the molten

metal from the top down and from the mold walls to the center of the ingot (Fig. 1). Because of the metal shrinkage and lack of available liquid metal, a cavity develops from the top down and, if not completely cropped before subsequent rolling, becomes elongated and will be found in the center of the final product, as shown in ingot B in Fig. 1.

Fig. 1 Longitudinal sections of two types of ingots showing typical pipe and porosity. When the ingots are rolled into bars, these flaws become elongated throughout the center of the bars.

Porosity is the result of trapped gaseous bubbles in the solidifying metal causing porous structures in the interior of the

ingot (Fig. 1). Upon rolling, these structures are elongated and interspersed throughout the cross section of the bar product, as illustrated in Fig. 1. Inclusions may be the products of deoxidation in the ingot, or they may occur from additives for improving

machinability, such as lead or sulfur. Inclusions and their typical location in a steel bar are shown in Fig. 2(a).

Fig. 2 Ten different types of flaws that may be found in rolled bars. See text for discussion.

Laminations may occur from spatter (entrapped splashes) during the pouring of the steel into the mold. They are

elongated during rolling and are usually subsurface in the bar. Figure 2(b) illustrates a lamellar structure opened up by a chipping tool. Slivers are most often caused by a rough mold surface, overheating prior to rolling, or abrasion during rolling. Very

often, slivers are found with seams. Slivers usually have raised edges, as shown in Fig. 2(c). Scabs are caused by splashing liquid metal in the mold. The metal first freezes to the wall of the mold, then becomes

attached to the ingot, and finally becomes embedded in the surface of the rolled bar (Fig. 2d). Scabs thus bear some similarity to laminations. Pits and Blisters. Gaseous pockets in the ingot often become, during subsequent rolling, pits or blisters on the surface

or slightly below the surface of bar products. Other pits may be caused by overpickling to remove scale or rust. Pits and blisters are both illustrated in Fig. 2(e). Embedded scale may result from the rolling or drawing of bars that have become excessively scaled during prior

heating operations. The pattern illustrated in Fig. 2(f) is typical. Cracks and seams are often confused with each other. Cracks with little or no oxide present on their edges may occur

when the metal cools in the mold, setting up highly stressed areas. Seams develop from these cracks during rolling as the reheated outer skin of the billet becomes heavily oxidized, transforms into scale, and flakes off the part during further rolling operations. Cracks also result from highly stressed planes in cold-drawn bars or from improper quenching during heat treatment. Cracks created from these latter two causes show no evidence of oxidized surfaces. A typical crack in a bar is shown in Fig. 2(g). Seams result from elongated trapped-gas pockets or from cracks. The surfaces are generally heavily oxidized and

decarburized. Depth varies widely, and surface areas sometimes may be welded together in spots. Seams may be continuous or intermittent, as indicated in Fig. 2(h). A micrograph of a typical seam is shown in Fig. 3.

Fig. 3 Micrograph of a seam in a cross section of a 19 mm ( oxide and decarburization in the seam. 350×

in.) diam medium-carbon steel bar showing

Laps are most often caused by excessive material in a given hot roll pass being squeezed out into the area of the roll collar. When turned for the following pass, the material is rolled back into the bar and appears as a lap on the surface. Chevrons are internal flaws named for their shape (Fig. 2k). They often result from excessively severe cold drawing and are even more likely to occur during extrusion operations. The severe stresses that build up internally cause transverse subsurface cracks. Nondestructive Inspection of Steel Bar, Wire, and Billets

Methods Used for Inspection of Steel Bars Almost all inspection of steel bars (other than plain visual inspection) is performed by means of the following four methods, used either singly or in combination: • • • •

Magnetic particle inspection Liquid penetrant inspection Ultrasonic inspection Electromagnetic inspection

Magnetic Particle Inspection Magnetic particle inspection offers the same visual aid in the nondestructive inspection of bars as it does for castings, forgings, or machined products. The method is used for detecting seams, cracks, and other surface flaws, and, to a limited extent, subsurface flaws. As a rule, the method is not capable of detecting flaws that are more than 2.5 mm (0.1 in.) beneath the surface. The magnetic particle method utilizes a magnetic field set up in the bar. Flaws cause a leakage of flux if they are at an angle to the flux flow. This flux is due to the lower magnetic permeability of the material in the flaw (air, oxide, or dirt) compared with that of the metal. Because the flux leakage forms magnetic poles, fine iron powder sprinkled on the surface will adhere, indicating the extent of the flaw. Longitudinal Flaws. Optimum indications are obtained when the magnetic field is perpendicular to the flaws. A

similar result is obtained for flaws slightly below the surface, but the surface leakage is less and, consequently, fewer iron particles are attracted to the area, producing a less definite indication.

Various colors of iron powders are commercially available to permit the choice of a color that provides maximum contrast between the powder and the material being inspected. Fluorescent coatings on powders and ultraviolet light can be used to make the indication more vivid. The powders can be applied in dry form, or can be suspended in oil or a distillate and flowed over the workpiece during or after the magnetizing cycle. Transverse Flaws. To detect flaws transverse to the long axis of the bar being inspected, a solenoid winding or

encircling coil is used. For longitudinal-type flaws, circumferential magnetization is utilized; an electric current flowing through the bar sets up a magnetic field at right angles to the long axis of the bar. To protect the bar from arc burns when the current is turned on, electrical contact is usually made by soft metallic pads held firmly against the bar ends. Power Requirements. The power used can be direct current or alternating current. Direct current may be from

batteries or rectified alternating current. Alternating current travels near the surface and should not be used for detecting subsurface flaws. In most cases, the continuous-magnetization system is used for bars because most bars have low retentivity for magnetism; therefore, the residual-magnetism system is not suitable. Finished bars must be demagnetized; otherwise, during manufacturing operations such as machining, steel chips will adhere and possibly cause trouble. Quantity Requirements. As a rule, the magnetic particle inspection of bars is confined to the inspection of a small

quantity of bars, as in a fabricating shop. The method is, in its present state of development, considered too slow and too costly for mass-production inspection, as in a mill. Detailed information on magnetic particle inspection as it is applied to various ferromagnetic products is available in the article "Magnetic Particle Inspection" in this Volume. Liquid Penetrant Inspection Liquid penetrant inspection (another visual aid), for several practical reasons, is not extensively used for detecting flaws in steel bars. These reasons include the following: • • •

Its use is restricted to the detection of flaws that are open to the bar surface Adaptation to automation is limited compared with certain other inspection methods Time cycles are too long for the inspection of bars on a mass-production basis

There are exceptions, however, and there are cases where liquid penetrant inspection has been used advantageously for inspecting from one to a few bars, as in a fabricating shop. Specific advantages are: • •

Liquid penetrant inspection is extremely sensitive and can sometimes detect surface flaws missed by other methods The solvent-removable system (one of the several liquid penetrant systems) in particular is extremely flexible and can be used for inspecting bars or portions of bars in virtually any location, including in the field

Detailed information on liquid penetrant inspection is available in the article "Liquid Penetrant Inspection" in this Volume. Ultrasonic Inspection Ultrasonic inspection is done with high-frequency (about 1 to 25 MHz) sound waves and can successfully detect internal flaws in steel bars. Most often, the ultrasonic inspection of steel bars is restricted to large-diameter bars and to applications where high integrity is specified. Also, because of the limitations of ultrasonic inspection for detecting surface flaws, it is ordinarily used in conjunction with some other method that is more suitable for inspecting bar surfaces. An ultrasonic beam has the valuable property that it will travel for long distances practically unaltered in a homogeneous liquid or solid, but when it reaches an interface with air (for example, at a crack or at the surface of a metal body), it is almost completely reflected. The ultrasonic beam is generated by applying a high-frequency voltage to a piezoelectric crystal, which is thus brought into mechanical oscillation. This energy in turn is fed to the workpiece by a liquid couplant.

The technique most commonly used in the nondestructive inspection of bars or barlike workpieces is the pulse-echo technique. Short pulses of ultrasonic energy are passed through the bar. The sweep voltage of the time base is coordinated with the pulse-repetition frequency so that the reflections are indicated on an oscilloscope screen. A certain amount of energy is reflected at the interface between probe and specimen, giving the first transmission signal. The probe can either have two separate crystals, a transmitter and a receiver, or have only one, which is used alternately as transmitter and receiver. The ultrasonic method is characterized by high sensitivity and very deep penetration, but in addition to its surface limitations, its production speed is relatively low. A liquid couplant is necessary and can be a source of interference. This method is suitable for testing ingots, billets, plate, and tubes in addition to bars or barlike workpieces. In certain cases, ultrasonic inspection has been automated. Typical products that are ultrasonically inspected using automated equipment are forged axle shafts (which are, in effect, extruded bars) and rolled bars. Cold-Drawn Bars (Ref 1). The most effective method for the inside flaw inspection of cold-drawn bars is ultrasonic flaw detection. However, it is becoming more and more necessary to detect the smaller defects in the near-surface area in accordance with changing the production process. The conventional normal beam method (Fig. 4a) is not satisfactory, because of untested area near the surface.

Fig. 4 Schematic showing position of probe relative to flaw inside of bar and resulting wave display obtained for two methods of ultrasonic flaw detection. (a) Normal-beam method. (b) Angle-beam method. Wave display nomenclature: T, transmit pulse; S, surface reflection echo; F1, flaw echo; B1, back wall echo. Source: Ref 1

A testing method for detecting smaller flaws immediately under the surface of cold-drawn bars is the angle-beam method (Fig. 4b), which conveys ultrasonic waves into the material with an angle beam. It can detect the flaws immediately under the surface that are in the dead zone for the conventional normal-beam method. Entire cross-sectional area testing becomes possible with the angle-beam method and the conventional normal-beam method in combination. The testing method to feed the material spirally and to make the probes follow the deflection of the material feeding has already been adopted in practical use for as-rolled steel bars. It is difficult to obtain higher testing speed for cold-drawn bars because of smaller dimensions. Therefore, the following method has been developed in which the material is fed straight and the probes are simultaneously rotated at high speed. Table 1 lists the main specifications of the system, and Fig. 5 shows a schematic of the setup. For bars with smaller dimensions, guide sleeves and tripplet rollers are used to prevent the ultrasonic incident angle to the material from changing because of excessive vibration and/or bending of the material. The water circulation system also incorporates a device that stabilizes the coupling water.

Table 1 Specifications of a rotating-type ultrasonic flaw detection system Parameter

Specifications

Dimension of material, mm (in.)

15-32 (0.59-1.26)

Testing method

Normal-beam method and angle-beam method

Testing frequency, MHz

10 and 5

Number of rotations of probe, rev/min

1000

Signal transmit

Noncontact rotation transmit

Marker

One each for near-surface flaw and inside flaw

Source: Ref 1

Fig. 5 Schematic of a typical rotating-type ultrasonic flaw detection system. Source: Ref 1

For flaws located immediately under the surface, the angle-beam method record can detect flaws as small as 0.2 to 0.3 mm (0.008 to 0.012 in.). Flaw echoes this small are not detectable with the normal-beam method. Cold-Drawn Hexagonal Bars (Ref 1). Requirements for strict quality assurance are increasing for gaging inside flaws to the same level as surface flaws. The conventional testing method is manual detection with the normal-beam method. Because this method requires testing with plural directions, working efficiency is low. Furthermore, an untested zone remains at the area immediately under the surface. Therefore, a testing system using the entire cross section with higher efficiency has been sought.

Higher efficiency has been attained by incorporating an automated ultrasonic flaw detection system with probes for each face of the material to detect separately the flaws located on the inside area and the near-surface area (Fig. 6). Flaws inside the material are detected with the normal-beam method at each face of the material. In this method, the untested zone remains in the near-surface area. Therefore, surface and near-surface area flaws are detected with the angle-beam method at each face of the material. That is, six normal-beam probes and six angle-beam probes are located on the circumference of the materials to be tested, which is conveyed longitudinally. The probe positions are arranged so that the entire cross section can be detected. The probe holder is designed so that all the probes can be adjusted simultaneously by adjusting one when the material size is changed. The coupling medium is a special oil that has low ultrasonic attenuation and causes no rust on the material to be tested. Table 2 lists the specifications of the system. Table 2 Specifications of an ultrasonic flaw detection system for cold-drawn hexagonal bars Parameter

Specifications

Dimension of material, mm (in.)

12-32 (0.472-1.260)

Testing method

Normal-beam, 6 channels; angle-beam, 6 channels

Testing frequency, MHz

5

Probe position

Fixed in circumferential direction

Marker

Two for near-surface flaw and inside flaw

Source: Ref 1

Fig. 6 Dual set of six circumferentially mounted probes used to ultrasonically detect flaws in cold-drawn hexagonal bars. (a) Normal-beam method to detect flaws deep inside bar. (b) Angle-beam method to detect surface and near-surface flaws. Source: Ref 1

Flaws larger than 0.3 mm (0.012 in.) can be detected at the near-surface area. Flaws measuring at least 0.2 mm (0.008 in.) can be detected deep inside the hexagonal bar material. Ultrasonic Flaw Detection on Cold-Drawn Wires (Ref 1). Surface flaw inspection is important for drawn wires. A

rotation-type eddy current flaw detection system is used for quality assurance. However, inside flaw inspection has been urgently needed because on-line inspection has been considered impossible.

In quality assurance for drawn wires, rotating-type eddy current flaw detection has been used in combination with rotating ultrasonic flaw detection to detect surface defects and inside flaws, respectively, in a two-step process. However, the high cost and inefficiency of this method have prompted the development of a system with a rotating-type ultrasonic flaw detection unit that can also detect surface flaws. An additional die is placed behind the cold-drawing die to stabilize the vibration of the material. A detection unit, which has probes arrayed in a circumferential direction, is placed between these dies. There are three detection modes (Fig. 7): • • •

Surface wave detection mode for surface defects Angle-beam detection mode for near-surface defects Normal-beam detection mode for inside defects

Fig. 7 Principle of ultrasonic flaw detection for cold-drawn wires using three detection mode probe. Source: Ref 1

The ultrasonic incident angle can be optimized according to material dimensions. Water, the coupling medium, is always kept in full quantity even in high-speed rotation. Thus, the system has the stable mechanism to provide constant detection. One advantage of this system is that linear defects can be detected by surface wave detection at the same level as an eddy current method. Another is that the entire cross section can be covered by means of a combination angle-beam/normalbeam method. Table 3 summarizes the specifications of this setup.

Table 3 Specifications of an ultrasonic flaw detection system for cold-drawn wires Parameter

Specifications

Dimension of material, mm (in.)

15-30 (0.590-1.181)

Testing frequency

Normal beam: 10 MHz, 1 channel Angle beam: 5 MHz, 2 channels Surface wave: 5 MHz, 2 channels

Number of rotations of probe, rev/min

1000

Signal transmit

Noncontact rotation transmit

Marker

One each for near-surface flaw and inside flaw

Source: Ref 1

Results of experiments with this system showed detectability of 0.1 mm (0.004 in.) minimum flaw depth on surface defects and 0.2 mm (0.008 in.) minimum inside defect size. This system enables the user to inspect the entire cross section of cold-drawn wires to a high degree of accuracy. Detailed information on the fundamentals, equipment, and techniques for ultrasonic inspection is available in the article "Ultrasonic Inspection" in this Volume. Electromagnetic Inspection Methods Electromagnetic methods of inspection are used far more extensively for nondestructive inspection of steel bars than any of the methods discussed above. Electromagnetic methods are readily adaptable to automation and can be set up to detect flaws, as well as a number of different compositional and structural variations, in bars on a mass-production basis. Equipment can be relatively simple, but for mass-production inspection the equipment may be highly sophisticated and costly. Such equipment can not only detect flaws and indicate them on an oscilloscope or other form of readout but can also mark the location of the flaw on the bar before it emerges from the inspection equipment and can automatically sort the bars on the basis of seam depth. Eddy Current Testing of Cold-Drawn Bars (Ref 1). Surface defects on cold-drawn bars can be inspected by eddy current detection methods using an encircling coil. This method utilizes a rotating probe that detects surface defects with the probe coil rotating at high speed around the circumference of the cold-drawn bars.

The encircling coil method exhibits lower detectability on linear flaws because flaw detection depends on the difference between two test coils in which the material to be tested is encircled. On the other hand, the method of rotating the probe coil at high speed along the circumference of the material to be tested can detect linear defects because it detects bars in spiral scanning. Table 4 lists the specifications of the detection system. One of the main features is signal transmission in the probe rotation unit by the noncontact rotating transmit method, which requires no maintenance work. Guide sleeves are placed in front of and behind the probe to maintain a constant distance between the probe and the material to be tested, which is important for acceptable performance of the system (Fig. 8). Furthermore, the rotation axis of the probe and the axis of the material to be tested are kept in a line by pinch rollers placed in front of and behind the detector. On the probe, a distance sensor is used for the automatic gain control function to provide electric compensation against distance variation.

Table 4 Specifications of a rotating probe type eddy current flaw detection system

Parameter

Type I

Type II

Dimension of material, mm (in.)

5-32 (0.197-1.260)

5-25 (0.197-0.984)

Number of probes

2

4

Probe area, mm2 (in.2)

10 (0.016)

5 (0.0078)

Number of rotations of probe, rev/min

3000

6000

Testing frequency, kHz

64

512

Signal transmit

Noncontact rotation transmit

Source: Ref 1

Fig. 8 Schematic of a rotating probe type eddy current flaw detector. Source: Ref 1

Figure 9 shows the relation between flaw depth and signal output. Natural flaws produce a larger deviation in signal output than artificially introduced flaws because of the complicated cross-sectional configuration of the flaw, but the

minimum detectable flaw depth is 0.1 mm (0.004 in.). Detectable flaw length depends on the feeding speed of the material, the number of probes, and the number of rotations. For example, at a speed of 60 m/min (200 sfm), the full surface is converted, and the minimum detectable flaw length is as long as the length of the probe coil.

Fig. 9 Plot of eddy current signal output versus flaw depth to gage detectability of flaws in cold-drawn bars. Source: Ref 1

Eddy Current Flaw Detection on Cold-Drawn Hexagonal Bars (Ref 1). Cold-finished steel profiles (hexagonal

bars) are mainly used as the raw material for couplers in oil pressure piping, an application for which quality assurance is important. Surface defects on cold-drawn hexagonal bars include cracks derived from the cold-working process as well as material flaws, both of which are long, longitudinal defects. It is impossible to detect these defects by the differential method using encircling coils. The rotating probe method is also not applicable, because of the hexagonal form. An automated flaw detection system for cracks initiated by the working process was developed using the eddy current flaw detection system by a standard voltage comparison method. There are two methods for testing cold-finished steel hexagonal bars: the standard voltage comparison method with encircling probes (Fig. 10b) and the differential method with probe assembly (Fig. 10c). There is no effective difference in detectability between these two methods. For the probe assembly method, it is necessary to consider the differences in detectability of each individual probe, which is not necessary for the standard voltage comparison method.

Fig. 10 Eddy current flaw detection method for cold-drawn hexagonal bars. (a) Location of artificial flaws ranging from 0.5 to 19 mm (0.020 to in.) below probe position. (b) Schematic of setup for standard voltage comparison (encircling coil) method (left) and plot of signals obtained for the designated flaw depths (right). (c) Schematic of setup for differential (six probe coil assembly) method (left) and plot of signals obtained for the designated flaw depths (right). Source: Ref 1

The standard voltage comparison method is inferior in detectability to the rotating probe method, but is less expensive and can efficiently detect cracks resulting from the cold-working process. This method, which can detect material flaws more than 0.6 mm (0.024 in.) deep, is illustrated in Fig. 11.

Fig. 11 Plot of eddy current signal output versus flaw depth to measure detectability of flaws--specifically material flaws (open circles) and process-induced cracks (closed circles)--in cold-drawn hexagonal bars. Source: Ref 1

Eddy Current Flaw Detection of Cold-Drawn Wires (Ref 1). Surface flaw detection on wire drawing line has

been conducted by the encircling-type eddy current method. However, this method has difficulty in detecting linear flaws. A rotating probe type eddy current detection method can be effective, as illustrated in Fig. 8 for use on cold-drawn bars. It is important in the rotating probe method to maintain a constant distance between the probe and the material to be tested. The rotating unit, is positioned between dies where the smaller vibration of the material is expected. Guide sleeves are used to adjust the rotating axis and the axis of the material to be tested. Detectability is illustrated in Fig. 12. Flaws having a 0.1 mm (0.004 in.) minimum depth are detectable.

Fig. 12 Plot of eddy current signal output versus flaw depth to measure detectability of flaws, specifically cracks (open circles) and scabs (closed circles), in cold-drawn wires. Source: Ref 1

Eddy Current Flaw Detection for a Cold-Forged, High-Tensile Sheared Bolt (Ref 1). Figure 13 shows a

general view of a high-tension sheared bolt. This type of bolt has a head with a round cross section and is mainly used for general construction and bridge applications. This bolt is produced by cold forging from cold-drawn wires in the diameter similar to the outside diameter of a threaded part of the bolt. The head is the most severely processed part of the bolt. The circumferential part of the bolt head is formed between punch and die during cold forging. Therefore, cracks tend to occur on the head. Eddy current testing can detect flaws in the bolt head at high speed with the probe rotating method.

Fig. 13 Schematic of a high-tension sheared bolt.

Figure 14 shows a general view of the inspection system used. Table 5 lists the main specifications. Bolts are conveyed from hopper to line-up unit. Lined-up bolts are conveyed to the index table by straight feeder and then conveyed intermittently to the rotating detection head and further to the separator. Table 5 Specifications of an eddy current detection system for a high-tension sheared bolt Parameter

Specification

Material

M20

Testing speed, pieces/min

60

Number of rotations of detecting head, rev/min

300

Testing frequency, kHz

125

Probe type

Self induction, self comparison

Source: Ref 1

Fig. 14 Schematic of eddy current flaw detection system used to inspect sheared bolt illustrated in Fig. 13. Source: Ref 1

After the bolt heads are detected with the rotating detection head, the bolts are classified as good/no-good and separated according to detection result. Figure 15 shows the operation of the rotating detection head. The rotating detection head repeats the following operations while rotating continuously regardless of the position of the bolt head to be tested: • • • • •

A bolt stops immediately under the detection head (Fig. 15a) The detection head descends while maintaining rotation (Fig. 15b) The detection head approaches the bolt head, scans around the bolt head for two revolutions, and detects the flaws (Fig. 15c) Pincerlike probe holders release from the bolt head, and the detection head ascends Bolt is conveyed to separator while next bolt is conveyed to the position immediately under detection

head

Fig. 15 Operation of rotating eddy current detection head. (a) Shear bolt positioned under rotating detection head. (b) Rotating detection head descends to lower probes into position to inspect bolt head. (c) Probe scans bolt head as bolt undergoes two complete revolutions to detect flaws. Source: Ref 1

A detection rate of 60 pieces/min was maintained by the mechanism to keep the detection head rotating continuously. Figure 16 shows the relation between flaw depth and signal output. Noise level is high at the circumferential surface of the bolt head because of surface roughness, but the minimum detectable flaw depth is 0.3 mm (0.012 in.). Detailed information on methods of electromagnetic inspection is available in the articles "Eddy Current Inspection" and "RemoteField Eddy Current Inspection" in this Volume.

Fig. 16 Plot of eddy current signal output versus flaw depth to measure detectability of flaws in high-tensile sheared bolts. Source: Ref 1

Reference cited in this section

1. N. Matsubara, H. Yamaguchi, T. Hiroshima, T. Sakamoto, and S. Matsumoto, Nondestructive Testing of Cold Drawn Wires and Cold Forged Products, Wire J. Int., March 1986 Nondestructive Inspection of Steel Bar, Wire, and Billets

Electromagnetic Systems Electromagnetic systems include the systems that use magnetic fields generated by alternating current flowing in a solenoid. A wide range of frequencies is used. As the alternating current flows through the solenoid, the magnetic field generated induces eddy currents within the metal workpiece. These currents are affected by the electrical resistivity (more commonly expressed as electrical conductivity--the reciprocal of resistivity), magnetic permeability, configuration, homogeneity, surface irregularities, and flaws of the metal. The resistivity of the workpiece can vary because of the chemical composition, crystal orientation, structure, and history of mechanical working. Permeability will vary over a broad range, depending on the amount of stress present in the work metal. It increases slightly in the vicinity of a flaw when the bar is subjected to a stress-producing operation. Electromagnetic systems of flaw detection are broadly classified as: • •

Those depending primarily on variations in electrical conductivity Those depending primarily on variations in magnetic permeability

Both systems are capable of detecting flaws in ferromagnetic bars. The conductivity-dependent systems can also be used to detect flaws in nonferromagnetic bars. Nondestructive Inspection of Steel Bar, Wire, and Billets

Eddy Current Systems When electrical conductivity (resistivity) is the major variable relied upon, the test procedure is known as the eddy current system. The alternating-field intensity is low, permitting the use of a correspondingly small inductor. Most eddy current systems use a constant-voltage input derived from an electronic oscillator with a means of varying the output frequency through a wide range, such as from 0.5 to 1000 kHz, in discrete steps. For general flaw detection, the range of 1 to 50 kHz is widely used. For ferromagnetic bars, a means must be provided to eliminate or minimize the effects of permeability variation. This is usually accomplished by magnetically saturating the bar being tested. The means for doing this is either a dc solenoid or a strong permanent magnet. A longitudinal section of one type of eddy current coil assembly is shown in Fig. 17, a more detailed drawing of the rotating coil setup shown in Table 4.

Fig. 17 Coil assembly for the inspection of steel bars by the eddy current system. Dimension given in inches

Eddy current inspection is especially useful for detecting and evaluating seams in steel bars. With this system, depending on the circuitry used, a difference of as little as 0.025 mm (0.001 in.) in seam depth can be detected. Because of the skin effect, the ability of eddy currents to penetrate into the test metal decreases in proportion to the increase of the frequency. Eddy current inspection can be used without magnetic saturation for inspecting hot bars in the mill when the metal is above the Curie temperature, because the metal is nonmagnetic at this temperature. Therefore, it follows that the magnetic permeability system cannot be used to inspect hot bars. Nondestructive Inspection of Steel Bar, Wire, and Billets

Magnetic Permeability Systems

Systems that depend on variations in magnetic permeability can be used for detecting flaws and for detecting differences in composition, hardness, or structure. With appropriate instrumentation, both functions can be accomplished simultaneously. Flaw Detection. Magnetic permeability systems usually employ a solenoid (primary coil), which is excited by the

standard line frequency of 60 Hz with an adjustable current control to produce magnetic fields from 1000 to 30,000 ampere-turns; however, the solenoid is usually operated in the range of 10,000 to 15,000 ampere-turns. A typical coil arrangement used for permeability systems is shown in Fig. 18.

Fig. 18 Coil assembly used for the simultaneous detection of flaws and of variation in composition, structure, and hardness in steel bars. Dimension given in inches

As seen in Fig. 18, the coil arrangement consists of a primary coil (60 Hz), two null coils (zero-voltage-output coils), and two standard coils. The secondary or pickup coils (null coils) are concentric with the primary coil, connected electrically in opposition, and adjusted to a null or zero-voltage output. The null coils are usually spaced 75 to 102 mm (3 to 4 in.) apart. The reason for this spacing is that a normal seam in a bar tapers into the bar to sound material. The variation in stress level producing a measurable change in magnetic permeability is related to the change in seam depth found usually within 75 mm (3 in.) of seam length. Limitations for Flaw Detection. The detection of flaws by permeability systems depends on permeability variations resulting from changes in stress, due to cold work or heat treatment, in the adjacent area of the flaw. These changes are more or less directly proportional to the change in stress up to the elastic limit of the ferrous product.

These systems cannot be used to inspect hot-rolled or annealed bars unless they have been subjected to some uniform cold work, such as rotary straightening for round material or planar-type straightening for square, hexagonal, or flat sections. Gag-straightened bars are not suited to inspection by permeability systems, because of nonuniform high-stress concentrations wherever the ram meets the work metal. Such stresses are far in excess of those for flaws in uniformly stressed material. The efficiency of flaw detection is a function of uniform residual stress levels within the bar. The five conditions that follow are listed in order of decreasing efficiency for detection of flaws by permeability systems: • • • •

Heat treated, quenched, drawn, and machine straightened Cold drawn and machine straightened Cold drawn, annealed, and machine straightened Hot rolled and machine straightened, centerless ground

•

Hot rolled and machine straightened

After straightening, the bars should be aged 24 to 48 h at near room temperature for optimum sensitivity of flaw detection. Aging can be hastened by stress relieving at low temperature in a furnace (up to 260 °C, or 500 °F). The minimum seam depth that can be detected in cold-drawn, straightened round bars is approximately 0.025 mm (0.001 in.) for each 1.6 mm ( in.) of bar diameter; hexagonal and square bars with the same processing will be more sensitive. For example, in a 25 mm (1 in.) diam round bar, a 0.41 mm (0.016 in.) seam is readily detected, while a 0.30 to 0.33 mm (0.012 to 0.013 in.) seam can be detected in hexagonal or square bars. The reason for this difference lies in the residual stress levels imparted by the rotary and planar straighteners. Other flaws, such as laps, slivers, cracks, hard or soft spots, dimensional changes, cupping, chevrons, and pipe, are readily indicated. For subsurface-type flaws, detection is possible only if they lie within the normal penetration range and are of sufficient size to affect the inherent stress level. The penetration is approximately 6.4 mm ( 7.9 mm (

in.) for medium-carbon steels, and up to 13 mm (

in.) for low-carbon steels,

in.) for many alloy steels.

One other factor not to be overlooked is the end effect, which prevents end-to-end inspection of the bar. As the front and rear ends of the bar enter and leave the magnetic field, the field is grossly distorted, preventing inspection of the end portions of the bar. For the average inspection speed of 37 to 46 m/min (120 to 150 sfm), the noninspected lengths will be as follows:

Bar diameter

Noninspected length at each end

mm

mm

in.

102-152

4-6

152-203

6-8

in.

6.4-13 -

13-25 -1

25-50

1-2

203-305

8-12

50-75

2-3

305-406

12-16

The signal obtained for a flaw of given size, as well as the amount of end effect, will vary somewhat with the amount of draft used in drawing the bar. Using the normal 0.8 mm ( draft will increase the signal size by 50%, while a 3.2 mm (

in.) draft as the basis for comparison, a 1.6 mm (

in.)

in.) draft will produce an increase of about 90% (Fig. 19).

Fig. 19 Relationship between increase of flaw signal and increasing reduction of cross section (increasing draft) for cold-drawn steel bars. Base reference is a hot-rolled bar.

All the above values hold true only when the secondary test coil is of the proper size; that is, the inside diameter of the coil should be 3.2 to 6.4 mm (

to

in.) greater than the bar diameter. The diameters of bar stock inspected by these

systems generally range from 4.8 to 140 mm (

to 5

in.).

Equipment for Detecting Flaws. The circuitry may include three types of electronic systems: the null system for the

detection of flaws (as explained above and shown in Fig. 18) and two identical standard systems, one of which is used for detecting mixed grades in a given lot of steel and the other for indicating variations of hardness or structure. All systems are independent and provide simultaneous indications with a single pass of the bar through the coil. The null system utilizes a pair of matched windings that provides for the comparison of a section of the bar with another section spaced some distance from the first. The matched windings are connected in opposition, and the resultant voltage is therefore theoretically zero, making the wave displayed on the oscilloscope a straight line. In practice, however, such a balance is seldom obtained. A small voltage with the wave-shape showing two peaks phase displaced 180° can normally be seen on the oscilloscope screen (bar out, Fig. 20). When a bar is placed within the coils, the wave pattern changes (bar in, Fig. 20). Should a flaw of minimum depth be present, the change in the waveshape is too small for measurement, even though there is a differential voltage between the null coils. Therefore, other relationships must be used to provide the desired information.

Fig. 20 Waveshape for oscilloscope pattern of a full electrical cycle for empty coils (bar out) and loaded coils (bar in). The position of an electronic gate for viewing an established portion of the cycle is shown.

The use of an electronic gate of any desired width permits these measurements to be made in any section of the wave. For example, the test gate shown in Fig. 20 is adjustable to any position of the 360° cycle. It is normally positioned 8 to 20° on either side of the stress peaks, where experience has revealed the wild stress effects are minimal and waveform changes for flaws are readily detectable. Most systems provide a second electronic gate that can monitor the section of waveshape where flaws cause a change in the saturation level, if this can be reached for the size and grade of material under test. Deflections greater than a predetermined amount will energize a signal that indicates rejection. Use for Sorting. The two standard systems differ from the null in that only one coil winding for each is utilized on the

bar being tested (Fig. 18). The voltage derived from this coil is balanced by a voltage in the instrument that is fully adjustable to the degree that the zero-center meters can be adjusted to their midpoint while the oscilloscope presentation continually shows the distorted waveshape. Should any undesired bars appear within the lot being tested, the meter deflection will then provide power for activation of suitable alarm devices. The selectivity of the section of waveshape to

be monitored is provided by an electronic gate, adjustable through 180°. Only half of the full 360° wave is required, the remainder being the negative duplicate of the positive and not shown on the oscilloscope. Both systems (standard coils, Fig. 18) are fully independent and should be operated at different positions of the waveform to obtain as much information as possible during the test. The standard system is used to monitor each bar in a lot for composition, hardness, structure, and size and to indicate the presence of uniform-depth seams, cracks, and laps, which generally escape detection by the null system. Nondestructive Inspection of Steel Bar, Wire, and Billets

Equipment Requirements In addition to coil arrangements such as those illustrated in Fig. 17 and 18, a fairly elaborate set of electronic gear is required for inspecting steel bars. Some type of equipment for handling the bars and conveying them through the coils at the desired rate is also required. The degree of sophistication designed into the equipment depends mainly on the number of similar bars to be inspected. Typical control units are adaptable to either the eddy current or the magnetic permeability systems of inspection. Many variations are commercially available. Nondestructive Inspection of Steel Bar, Wire, and Billets

Flaw Detection Procedure Using a Permeability System The following description is presented purely as an example of setting up a procedure for detecting flaws by permeability change using the null coils with one specific type of instrumentation. Procedures vary widely for various types of equipment. Select the proper size of secondary coil for the bars to be tested, and insert it into the primary-coil unit, making sure all electrical connections are secure. Turn on the primary power, and adjust the coil current to about 8 A as shown on the electrical meter on the power-supply panel. Adjust the combined coil unit so that the material will be concentric with the inside diameter of the secondary coil. Select the appropriate feed-through speed of the conveyor system, and insert a bar into the test coil. Turn the sensitivity on the null-equipment panel clockwise to about midposition, and while the bar is moving in the forward direction, alternately adjust the balance x and y controls to bring the null waveshape, as seen on the oscilloscope, to as near a straight line as possible. Do not stop the bar while making these adjustments, because it heats rapidly and its permeability changes with increasing temperature, resulting in a flaw indication when passed through the test coil on a repeat run. Furthermore, do not adjust the controls when the feed is reversed through the test coil, because the magnetic field is dragged by the bar in the direction of travel and the balance obtained would not be correct for the normal feed direction. Assuming the oscilloscope controls are adjusted properly, the two stress peaks of the null waveshape will be located equidistant from the edge of the tube face and 180° apart. Locate the index or gate about 10 to 15° to the right of the left stress peak by use of its control knob, and adjust its height by the gain control so that it is easily recognizable. The choice of using a 60-Hz filter, either in or out, is based on past experience when testing the same type of material with a similar setup. Finally, readjust the sensitivity control so that the small fluctuations within the gate do not reach the preset trip level for the various signaling devices. Nondestructive Inspection of Steel Bar, Wire, and Billets

Sorting Procedure Using a Permeability System

The following example of a procedure for sorting involves the use of the standard coils (Fig. 18) for detecting variations in hardness, grade, and structure, employing three different electronic gates within 180° (one-half cycle). In this case, the same package of electronic gear is used for pickup from both the null and the standard coils. Position the waveshape switch to the respective A or B standard circuit, and its oscilloscope trace will appear on the face of the tube in place of the former null trace. Adjust the sensitivity control to the third or fourth step, and with a bar feeding through the test coil, position the gate to its desired location by means of the coarse-gate and fine-gate set controls. The associated zero-center meter will deflect either to the positive or negative side, and by turning the compensator knob in the appropriate direction to deflect it toward its midpoint, the meter can be balanced to zero. A rough adjustment can be made with the bar stationary in the test coil. When the bar is heated, the oscilloscope pattern continually changes. By moving the bar forward two or three times while making the above adjustments, the final balance can be obtained. The same procedure is used for the other circuit, but the gate should not be at the same position on the electrical cycle (Fig. 21). Figure 21 shows the waveshape for one-half of an electrical cycle for a bar of 1141 steel. The three gate positions for detecting variations in hardness, grade, and structure are indicated on the waveshape.

Fig. 21 Waveshape for one-half of an electrical cycle as seen on the screen of an oscilloscope with a bar of 1141 steel in the coil. Three electronic gate positions are indicated for inspecting for hardness, grade, and structure.

The bar that was used for setup purposes should be set aside until testing of the whole lot of material is completed, because every bar is being compared to the setup bar. Should the meters exhibit a deviation for the whole length of the bar, the bar could be a different grade or contain a full-length seam or crack. A varying meter reading may indicate a change of hardness or structure as well as a deep flaw. The latter will be simultaneously indicated by the null circuit. If an unwanted grade of steel bars has been separated, it is advisable to retest this group, choosing one bar from the unwanted lot for setup, because it is not unusual to find a third grade within a mixed lot of steel. Nondestructive Inspection of Steel Bar, Wire, and Billets

Nondestructive Inspection of Steel Billets

Steel billets are generally less uniform in section and straightness than steel bars. Furthermore, billet surfaces are usually less refined and therefore not as smooth as bar surfaces. These characteristics of billets make it more difficult to establish methods and procedures for nondestructive inspection. The methods described in this section are applicable to the common rolled round-cornered-square billets. Surface Preparation Regardless of the method used for billets, inspection results are greatly improved when they are free from excessive scale and blisters. Tightly adherent scale usually does not interfere with inspection unless it is thick enough to affect the sensitivity of the inspection method. Two common methods for preparing the billet surfaces for inspection are pickling in hot acid and gritblasting. Magnetic Particle Inspection The visual inspection of billets has been upgraded by the addition of the magnetic particle method. The principles involved in the magnetic particle inspection of billets are essentially the same as those for the magnetic particle inspection of other ferrous products or product forms (see the article "Magnetic Particle Inspection" in this Volume). As a rule, the wet fluorescent particle system is chosen for inspecting billets. Magnetic particle inspection has improved the ability to detect obscure flaws, and through the use of mechanical handling equipment, inspection can be accomplished quite rapidly. However, this method is subject to some of the inherent disadvantages of visual inspection. Test results depend on the alertness and eyesight of the operator and on his ability to judge the severity of the flaw. Other variables, such as magnetizing current, particle size, and contamination of the particle bath by foreign substances, can vary the intensity of the flaw indication and thus increase the difficulty of appraising its severity (Fig. 22). Another factor is that, even when the above conditions are controlled, the magnetic leakage field attracting the particles is not uniform across the flat face of a square billet. The field density is maximum at the center of the flat face and almost nonexistent at the corners of magnetized billets. Consequently, the particle concentration and the flaw indication decrease with distance from the center for any one of the four flat faces.

Fig. 22 Seam indication width versus magnetization current for a 105 × 105 mm (4 × 4 in.) 1021-1026 grade steel billet. Seams tested: center of billet face perpendicular to billet surface; seam or portion of seam with width 0.025 mm (0.001 in.) for a total depth of 0.76 mm (0.030 in.). Magnetic particle sizes: type I, 3 to 28 μm (120 μin. to 0.0011 in.) (major % = 8 to 18 μm, or 320 to 720 μin.); type II, 15 to 60 μm (600 μin. to 0.0024 in.) (major % = 28 to 48 μm, or 0.0011 to 0.0019 in.); special type II, 40 to 74 μm (0.0016 to 0.0030 in.) (major % = 44 to 62 μm, or 0.0017 to 0.0025 in.). Bath application: 76 L (20 gal.) pressurized mixing tank with hose and hand-operated applicator. Source: Ref 2

Another magnetic test for the inspection of billets also requires the billet to be magnetized. Instead of using particles, a magnetic tape placed close to the billet surface is used to detect the presence of flaws. The flux leakage resulting from a surface discontinuity is recorded on the tape. The tape is then scanned by a tiny probe coil, which detects the flux leakage recorded on the belt. The probe transforms the magnetic leakage into an electrical signal. The signal transmitted to the control cabinet by means of contactless transformers will, if it is larger than a preset value, operate an alarm or marker. The billet face is usually divided into five to ten tracks, each about 19 mm ( in.) wide. The signal operates an electronic trigger circuit, which trips the appropriate indicator for the track in which the flaw lies. This method eliminates a deficiency of the magnetic particle test in that the interpretation of results is not dependent on an operator. Also, the use of compensating circuitry reduces the differential in sensitivity to flaws from the middle to the corner of the billet. Rolled Versus Continuous-Cast Billets (Ref 2). The major flaw types for rolled-steel billets are surface or near-

surface seams, which are primarily oriented along the length of the billet. For continuous-cast products, in the as-cast state, flaws include surface cracks oriented in the transverse direction and round near-surface flaws such as slag pockets.

There are two basic methods of magnetic field application. For detecting longitudinal seams or cracks, the circular magnetization method is used. For detecting transverse seams or cracks, longitudinal (or coil) magnetization is used. For continuous-cast billets, circular and coil magnetization have been applied simultaneously. To detect multidirectional nearsurface flaws and surface flaws, it is convenient to use simultaneously a dc field for circular magnetization and an ac field for longitudinal magnetization. Generally, the billets are 3 to 14 m (10 to 45 ft) in length. The large billets are often referred to as blooms. To apply the circular field, the current is passed directly through the billet longitudinally. To apply the longitudinal field, the billet is placed along the axis of the coil, and either the billet or the coil is moved relative to the other. In the wet magnetic particle method, for economy and handling convenience, a common carrier of the particles is water. The particles are formulated so that they will not easily deteriorate in prolonged mixture with water. Other ingredients can be added to facilitate the use of such a mixture, such as a wetting agent to disperse the particles and properly wet the part surface, antifoaming chemicals to reduce the suds due to agitation, and antifreeze to keep the bath from freezing in winter. If needed, fixer or binder can also be added to the bath to make the dried indications durable enough to withstand the normal handling in the mill. Inspection of Rolled Billets. Typically, particle types with sizes ranging from less than 10 to 70 μm (400 μin. to

0.003 in.) are often adequate to find the significant seams in rolled-billet inspection. The finest-size particle, 6 μm (240 μin.) or less, is good for finding extremely minute flaws--for example, inclusion flaws as small as 25 μm (0.001 in.) wide in 0.25 mm (0.010 in.) thick steel sheets. For these flaws, larger-size particles may be ineffective; significant leakage fields are confined to very small areas. The total forces exerted on the individual particles are usually weak a short distance away, and the relatively bulky size and the momentum of large particles would prevent most of them from getting close enough to the flaws to be strongly attracted and stay there. As a result, either too few magnetic particles are retained at the flaw sites to contrast with the background, or the indications can be easily disturbed by their own weight and by the bath flow. However, very fine and shallow flaws are not important in practical billet inspection, in which flaw depths greater than 0.64 mm (0.025 in.) are often the only ones of interest. (When reheated, the outer skin of the billet with shallow flaws will be oxidized to become scale and will fall off the part during further rolling operations.) Therefore, particles with nominal diameters of 10 to 50 m (400 in. to 0.002 in.) are often used to give large indications for the significant seams and to ignore very minor ones. The ultimate choice of particle type is not determined by particle size alone. Other factors, such as the magnetic properties, color, and brightness of the particles, must also be considered. For the proper magnetization current level, a rule of thumb has been developed: Roughly 1000 Adc per 25 mm (1.0 in.) of material diameter is satisfactory for most critical inspections. The equivalent surface field is about 1.2 × 104 A · m-1 (157 Oe). For most practical large-billet inspections, less than half this value is used. Inspection of Continuous-Cast Products. Slag pockets, transverse cracks, and longitudinal cracks are the typical

flaws found in continuous as-cast products (billets, blooms, and slabs). The slag pockets are near-surface inclusions. In a few cases, they can be partially exposed to the surface. The inclusions are mainly slags. There may also be voids with coats of oxidelike chemicals on the walls. They are often approximately round, but can be elongated in depth or longitudinal direction. Depending on product type, the largest slag pockets uncovered in the laboratory can be 3.2 to 6.4 mm ( to in.) in diameter. The smallest ones may have diameters less than 0.8 mm ( in.). The pockets are generally found within a 2.5 mm (0.1 in.) deep layer under the product surface. Small slag pockets are often found near the surface. Cracks in the continuous-cast products have been observed to be deep and tight, with inclusions. The cracks can occur in the direction of the oscillating mold marks (transverse corner cracks) or perpendicular to that direction. Depending on the casting technique, they can occur in groups and branch out in all directions like a web, covering significant areas. In general, the discussions about seam indication formation in rolled billets can be applied to crack indications in the continuous-cast product. For example, in the case of slag-pocket indications, the parameter equivalent to the seam width is the indication diameter. However, there are still a few differences between rolled-billet inspection and continuous-cast bloom inspection. Differences in Magnetization Levels. The magnetization level used in continuous-cast product inspection is usually

stronger than that required for rolled billets. Because of the subsurface nature of the flaws and the wider flaw dimensions,

the attraction of the magnetic particles to the leakage fields for the flaws is not as intense as the attraction normally found in rolled billets. Therefore, the magnetization levels often must be higher. The magnetization technique for the continuous-cast product is more complicated because magnetization is needed in both circular and longitudinal directions. When simultaneously applied multidirectional magnetization is used, the combined field effect equals the oscillating field (swinging vector) on the part surface. Eddy Current Inspection One advantage of the eddy current inspection of billets over magnetic particle inspection is that uniform results can be obtained without significant involvement of operator judgment. Another advantage is that the corners of a billet can be eddy current inspected to the same degree of thoroughness as the flat surfaces. There are, however, certain problems involved in the eddy current inspection of billets. Billets are not sufficiently uniform in cross section or straightness to permit their being fed through encircling coils. Also, this type of coil is not very sensitive in detecting the longitudinal-seam type of flaw. Therefore, the eddy current inspection of billets is usually performed by using probe coils. Figure 23(a) illustrates how a probe coil induces eddy currents in the billets, and Fig. 23(b) illustrates how the flow of eddy currents is changed by the presence of a seam in a billet.

Fig. 23 Schematics of eddy current flow. (a) Eddy current flow around a probe coil for a sound billet. (b) Eddy currents flowing around the seam in a defective billet, thus altering the electrical loading on the probe coil

It is not feasible, however, because of the inherent cross-sectional shape of square billets, to inspect by merely rotating them around the surface. The probe would bounce and this would excessively vary the distance between the billet surface and the probe coil, thus degrading the accuracy of the results. Also, the rapid bounce could generate false signals. Therefore, another approach must be taken. Use of Rotary Probe Unit. One approach is the use of a rotary probe unit such as that illustrated in Fig. 24. The

significant components of this unit are search probes mounted in a housing to which is attached a tungsten carbide wear shoe and an arrangement of hardened steel rollers. This entire assembly is spring actuated. This unit and usually a second one like it are mounted 180° apart on a drum-shaped assembly. Two probes are mounted in each unit to increase the amount of inspection coverage. Electrical energy reaches the search probes through slip rings. The units rotate around the billet as it progresses through the drum assembly. As can be seen in Fig. 24, the coil in the search probe is, at all times, the same distance from the billet surface, separated only by the tungsten carbide wear shoe. The spring-actuated rollers move as required to aid in holding the search probe on the billet surface as the entire unit revolves around the billet.

Fig. 24 Rotary probe unit used for the eddy current inspection of steel billets, and graph showing effect of position on speed as the probe unit traverses radially over one quadrant of a 102 mm (4 in.) square billet

One requirement for accurate results is that the speed of the probe relative to the test surface must be maintained within certain limits. Experimental studies have established that the optimum range for this eddy current test is 25 to 43 m/min (80 to 140 sfm). A probe speed below this range results in an inconsistent measurement of flaw depths. Speeds above this range cause objectionable electronic noise. The graph in Fig. 24 shows the relative speed of the probe at different positions on one quadrant of a 102 mm (4 in.) billet section. Many of the dimensions of the probe assembly components affect the instantaneous velocity of the probe relative to the billet surface. Therefore, the dimensions of the elements must be selected so as to impart to the probe a velocity that is maximum on the flat, where bounce is least likely to occur, and minimum near and around the corner, where the coil has the greatest tendency to leave the surface. The speed fluctuates smoothly within the allowable testing range as the probe assembly is rotated around the billet.

The rotary probe unit shown in Fig. 24 can be used to test billets ranging from 75 to 152 mm (3 to 6 in.) square. For other sizes, the dimensions of the components of the unit must be changed accordingly. The forward speed of the billet conveyor is between 7.6 and 23 m/min (25 and 75 sfm). The forward speed that can be used depends on the size of the billet and the length of the flaws that must be detected. The inspection is performed on a spiral band on the surface of the billet; the pitch of the spiral (and therefore production rate) is determined by the minimum length of flaw to be detected. Production use in one mill calls for the detection of all flaws 25 mm (1 in.) and longer, resulting in a speed of 12 m/min (40 sfm) for 102 mm (4 in.) billets and 10 m/min (34 sfm) for 127 mm (5 in.) billets. At these speeds and with no delays, 41 to 54 Mg/h (45 to 60 ton/h) can be inspected in one machine. Marking the flaws is done with spray markers that are mounted near and rotate with the search-probe assemblies. One marker is used per search-probe assembly, and either of the two search probes will activate the marker. With this arrangement, the marks always lead, by a short distance, the actual longitudinal location where the flaw was detected. The circumferential position of the flaw is marked exactly, because the spray markers are displaced from the probe centerline at a distance that is equal to the distance that the faceplate rotates during the reaction time of the marking system. Calibration of the electronic circuitry is accomplished by moving the entire rotating portion of the machine out of the line of billet travel. A short test billet is placed into the test position. The probes are rotated while contacting the test billet, and the signals are recorded. The electronic circuit is then calibrated with simple dial adjustments. The test billet contains four simulated seams produced by milling slots 0.76 mm (0.030 in.) deep with a 0.15 mm (0.006 in.) wide cutter. The locations of these milled-slot seams are on the flat, before the corner, on the corner, and after the corner.

Reference cited in this section

2. L.C. Wong, Test Parameters of Wet Magnetic Particle Inspection of Steel Billets, Mater. Eval., Nov 1988 Nondestructive Inspection of Steel Bar, Wire, and Billets

References 1. N. Matsubara, H. Yamaguchi, T. Hiroshima, T. Sakamoto, and S. Matsumoto, Nondestructive Testing of Cold Drawn Wires and Cold Forged Products, Wire J. Int., March 1986 2. L.C. Wong, Test Parameters of Wet Magnetic Particle Inspection of Steel Billets, Mater. Eval., Nov 1988

Nondestructive Inspection of Tubular Products

Introduction WROUGHT TUBULAR PRODUCTS are nondestructively inspected chiefly by eddy current techniques (including the magnetic flux leakage technique) and by ultrasonic techniques. In general, the eddy current and magnetic flux leakage techniques are applied to products not exceeding 1020 mm (40 in.) in diameter or 19 mm ( other hand, ultrasonic inspection is used on tubes ranging from 3.2 to 2030 mm (

in.) in wall thickness. On the

to 80 in.) in diameter and from 0.25 to

64 mm (0.01 to 2 in.) in wall thickness. However, there are many exceptions, and the range of special techniques and applications associated with each inspection method is large. Most welded and seamless tubular products are nondestructively inspected by the manufacturer at the mill.

The many uses to which steel tubular products have been applied form a basis for classifying steel tubular products; for example, the terms casing, tube, and pipe are assigned on the basis of usage, as in water-well casing, oil-well tubing, and drill pipe. A second classification is based on methods of manufacture. Accordingly, all steel tubular products can be classified as either welded or seamless. A third classification applicable to special shapes can be considered subordinate to both of the general classifications above. The major applications of the nondestructive inspection of tubular products are: • • • •

Detection and evaluation of flaws Sorting of mixed stock Measurement of dimensions Comparative measurement of specific physical and mechanical properties

Of these, the primary application is the detection and evaluation of flaws. Sorting is often an auxiliary application employed for grade or size verification and can be based on chemical composition, dimensions, physical and mechanical properties, or other significant variables. A difficulty encountered in sorting arises when variables of little or no interest affect instrument indications to a greater degree than do the variables of interest. Several applications involving the measurement of dimensions and physical properties are described in this article. Others, such as the measurement of coating thickness and the noncontact measurement of wall thickness and variations in diameter, require highly specialized instrumentation and are discussed elsewhere in this Volume. Additional information on the inspection of pipe and pipelines, including the examination of field-welded girth welds, is available in the articles "Weldments, Brazed Assemblies, and Soldered Joints" and "Boilers and Pressure Vessels" in this Volume. Nondestructive Inspection of Tubular Products

Selection of Inspection Method The fundamental factors that should be considered in selecting a nondestructive inspection method, and in selecting from among the commercially available inspection equipment, are the product characteristics, the nature of the flaws, extraneous variables, rate of inspection, the end effect, mill versus laboratory inspection, specification requirements, equipment costs, and the operating costs. Product Characteristics. Among the product characteristics that may affect the choice of inspection method and equipment are tube or pipe diameter, wall thickness, surface condition, method of fabrication, electrical conductivity, metallurgical condition, magnetic properties (notably permeability), and degree of magnetization. Nature of Flaws. Both the nature of flaws and of potential but unallowable deviations from certain specified

dimensions or properties have a bearing on the selection of inspection methods and equipment. The nature of flaws is often markedly influenced by the method of manufacture. For example, flaws in welded pipe are usually confined to the vicinity of the weld; therefore, an inspection procedure that is confined to the weld area may be adequate. If the welds are resistance welds, the most usual flaws are located in the weld plane and are in effect two-dimensional, having length and width but negligible thickness. On the other hand, if the welds are arc welds, porosity is the most usual flaw. In all welded tubular products, cracks are the most damaging flaws. In seamless tube, the location of flaws is not restricted, but may occur anywhere in or on the tube section. Extraneous Variables. Many of the measurable variables in tubing and pipe are normal to the product and are not

cause for rejection. These extraneous or harmless factors sometimes exert a greater effect on the inspection instrument than do the flaws that must be detected. For example, variations in magnetic permeability are common in steel and generate large signals in instruments that are permeability sensitive. However, the signals often are not pertinent to the test, nor are they cause for rejection. Surface scratches may be cause for rejection in some products and yet may be acceptable in others. Consequently, the inspection method and instrument selected must ignore or minimize variables that will not affect the utility of the part in its intended application.

The rate of inspection required is a major factor in the selection of an inspection procedure. When the value of the

part or the hazard associated with its application justifies slow and thorough inspection, the procedure chosen is likely to be radically different from that selected for a mass-produced, low-cost product used in a noncritical application. End Effect. In some applications, the only portions of the tube that are genuinely critical in its ultimate application are

the ends. Unfortunately, with many nondestructive testing instruments, specific problems arise when inspection of the ends is required. End effect is encountered with the eddy current, ultrasonic, and radiographic methods. Consequently, inspection of the entire tube and the ends may require two different procedures; as a result, production speed is reduced and the total cost of inspection is correspondingly increased. Mill Versus Laboratory Inspection. Although laboratory demonstrations of nondestructive inspection techniques

may yield excellent results, subsequent mill performance may be entirely unsatisfactory because of conditions present in the mill that were not present in the laboratory. Specification requirements may also affect the choice of inspection method and equipment. When the tubular

product is covered by a flaw-size specification, all tubes with flaws larger than those specified must be rejected. However, tubes with flaws smaller than the specified rejection level should be accepted. Because many nondestructive inspection systems do not provide for linear adjustment or are incapable of making the required differentiation, this aspect of instrument performance must be carefully investigated. Equipment cost is usually a major factor in the selection of inspection method and equipment. The initial cost of

equipment may occasionally be minor, but in some cases installation may cost over $1 million in basic and related equipment. The lowest-cost equipment may be for magnetic particle or liquid penetrant inspection; high-cost installations involve automatic flaw marking, classification of product based on flaw magnitude, computer analysis of results, multiple sorting levels, and many other convenience factors. The operating cost of inspection procedures and equipment varies widely. In general, it is inversely proportional to

the cost of the installation. The more expensive installations are usually completely automatic and are incorporated in a production line whose primary function is something other than inspection. Consequently, inspection adds little to total operating cost. In contrast, the lower-cost installations usually involve a separate operation and require the services of a highly trained, skilled operator. Nondestructive Inspection of Tubular Products

Inspection of Resistance-Welded Steel Tubing Resistance-Welded Steel Tubing The diameters of resistance (longitudinal) welded steel tubing range from about 13 to 914 mm ( to 36 in.); wall thicknesses, from 0.38 to 13 mm (0.015 to 0.5 in.). Tubing of intermediate and smaller diameters is produced on a draw bench. Flaws that occur in resistance-welded steel tubing include cold welds, contact marks, cracks, pinholes, and stitching;

these are described below. The terminology used to designate such flaws varies; the terms used in this article are those adopted by the American Petroleum Institute (API) (Ref 1). Cold weld is the term widely used to indicate inadequate or brittle bonding with no apparent discontinuity in the

fracture. Cold weld cannot be detected reliably by any nondestructive inspection method currently available. Contact marks (electrode burns) (Fig. 1a) are intermittent imperfections near the weld line that result from miniature arcs between the welding electrode and the surface of the tube.

Fig. 1 Typical flaws in resistance-welded steel tubing. (a) Contact marks (electrode burns). (b) Hook cracks (upturned-fiber flaws). (c) Weld-area crack. (d) Pinhole. (e) Stitching. Views (c), (d), and (e) are mating fracture surfaces of welds.

Hook cracks (upturned-fiber flaws) (Fig. 1b) are separations within the base metal due to imperfections in the strip

edge, which are parallel to the surface and turn toward the outside or inside surface when the edges are upset during welding. Weld-area cracks (Fig. 1c) are any cracks in the weld area not due to upturned fibers. Pinholes (Fig. 1d) are minute holes located in the weld line. Stitching (Fig. 1e) comprises a regular pattern of light and dark areas that are visible when the weld is broken in the

weld line. The frequency of variation usually corresponds to weld-current variation. Increased use of ultrahigh-frequency current for welding has minimized the occurrence of stitching. The nondestructive inspection of resistance-welded tube can be performed continuously on a welding machine or on individual lengths at any stage of processing. When performed on a welding machine, test indications can be used to guide the welding-machine operator in making machine adjustments.

Eddy Current Inspection In 1929, H.C. Knerr initiated the development of eddy current inspection for Republic Steel Corporation. In 1930, in conjunction with Knerr, C. Farrow demonstrated the detection of flaws in a brass tube by the phase shift in eddy currents generated in the tube. In 1931, he demonstrated that the phase-shift concept could be applied to magnetic steel tube, provided the area being inspected was magnetically saturated (Ref 2). The eddy current inspection method replaced both hydrostatic and pneumatic pressure testing of resistance-welded tubing because of its superiority in detecting weld imperfections. Several of these typical weld imperfections are shown in Fig. 2.

Fig. 2 Mating fracture surfaces of pipe or tube welds showing imperfections detectable by eddy current inspection but not by pressure testing. (a) Unwelded spot (diagonal arrows) and a nonpenetrating pinhole (horizontal arrows). (b) Unwelded spots, probably caused by entrapped foreign matter. (c) Surface crack in weld

At present, the eddy current methods are probably the most widely used for the inspection of welded steel tubing in diameters up to 75 mm (3 in.), although these methods are not limited to the smaller diameters. In welded tubing, most flaws occur in or near the longitudinal-welded seam, and in most cases a test of a narrow band including the seam is adequate. This makes possible the use of small eddy current probe coils tangent to the seam area, eliminating the diameter limitation. Tubes with diameters up to 406 mm (16 in.) are currently being inspected in production by this method. Eddy current inspection is usually on tubing having wall thicknesses less than 3.2 mm (

in.), but successful production

testing has been reported on tubing having wall thicknesses to 13 mm ( in.) (Ref 3). Most eddy current tests use differential systems and therefore are most sensitive to flaws that involve a marked change in normal electrical characteristics. If the flaw is of considerable length and of uniform characteristics, it is sometimes necessary to use special arrangements for its detection. Small probe coils continuously compare the weld zone with the base metal, thus revealing the existence of the elongated or long flaw. When inspecting for shallow-cracklike surface flaws 3.2 mm ( in.) or less in depth, relatively close correlation between crack depth and signal magnitude has been obtained with a single coil arrangement without magnetic saturation (Ref 4). However, the limited penetration of this arrangement and its need for a surface opening for depth evaluation limit its usefulness. The speed of inspection by eddy current methods depends in part on many factors, including the size of the flaw that

must be detected, the discriminating ability of the circuit used, end-inspection requirements, and the speed of response of the signal circuit. The mathematical relationship of the test-current frequency and linear speed may automatically limit the size of flaw that can be detected. Speeds of 305 m/min (1000 sfm) have been recorded, but the usual speed is 45 to 90 m/min (150 to 300 sfm). Weld Twist. When using probe coils in eddy current inspection, the twist in the weld sometimes causes a special

problem. When the weld twists out of the zone of high sensitivity, the effectiveness of flaw detection is markedly

reduced. The problem cannot be solved by increasing the size of the arc segment covered by the probe coil, because this arrangement also reduces sensitivity. One solution involves the use of a series of small probe coils, staggered with respect to the weld line, to ensure continuous coverage (Ref 5, 6). The problem has also been solved in some installations by taking advantage of the electromagnetic difference that exists between the weld zone and the base metal. Special probe coils respond to this difference and automatically rotate the test head or the tube until the weld zone is properly located with respect to the probe coil. End effect, caused by abrupt changes in the magnetic field, becomes a problem whenever cut lengths are inspected. Various auxiliary circuits, ranging widely in effectiveness, have been developed for suppressing end effect to permit satisfactory inspection closer to the end of the tube. End effect can be minimized by keeping the tube ends in contact as they move through the test coils. Mechanical variables that may affect inspection results include transverse movement of the tube in the test coil and

changes in temperature or linear speed. The contribution of these factors to test results is sometimes difficult to determine in the laboratory, but they may create serious problems in production testing. Equipment costs for eddy current inspection can vary widely, depending on the extent of refined circuitry, automatic

handling and sorting equipment, computer analyzers, or special auxiliary equipment that may be needed. The operating costs of a well-designed eddy current system are among the lowest of any nondestructive inspection

method. After the system has been properly adjusted, it can be operated by unskilled workers. When automatic marking is provided, the inspection can frequently be combined with another operation without appreciably increasing the cost of the latter operation. Advantages and Limitations. All flaws in resistance welds except cold weld, are readily detected by eddy current

methods. Cold weld is by far the most difficult of all flaws to detect by any of the nondestructive inspection methods. Although the other types of flaws listed above can be detected by eddy current methods, it should not be inferred that all eddy current instruments will detect all of these flaws. The range of capabilities of commercial eddy current instruments is extensive, and conclusions regarding their capabilities often require actual tests. Because eddy current test coils may either surround or be adjacent to the tube being tested, the variety of coil designs, arrangements, and combinations constitutes another major group of variables affecting equipment capabilities. In general, eddy current instruments have the advantages of speed in testing and convenience in operating, marking, and sorting. Perhaps their most universal disadvantage is their inability to inspect completely to the ends of tubes. Additional information is available in the articles "Eddy Current Inspection" and "Remote-Field Eddy Current Inspection" in this Volume. Flux Leakage Inspection Flux leakage (or magnetic field perturbation) inspection is similar to eddy current inspection but requires magnetization of the tube and is limited to the inspection of ferromagnetic materials. When the tube is magnetized to near saturation, the magnetic flux passing through the flaw zone is diverted by the flaws. Detectors of various types detect the diverted flux when either the detector or the tube is moved in a direction that causes the detector to cut through the diverted flux. This in turn produces a signal to reveal the presence of the flaw. Various means are used to magnetize the tube. A current-carrying conductor inside the tube produces a circular magnetic field, magnetizing the tube in a circumferential direction. The magnetic flux is diverted by the longitudinal component of any flaws in its path. The probe, moving through the diverted flux, generates a signal roughly proportional to the size of the flaw. On a longitudinal-welded seam, an electromagnet with pole pieces on each side of the weld can be used to magnetize the weld area, with flux passing transversely across the seam. The magnetic flux is diverted by the longitudinal component of any flaw in the weld, and the flaw can be detected electronically. To detect transverse flaws, the tube may be magnetized longitudinally by an encircling conductor. The flux is then diverted by the transverse component of any flaw present, and the probe moving through the diverted or leakage flux reveals the presence of the flaw. Hall probes are the detectors ordinarily used. In all applications, there must be relative movement between the probes and the diverted flux so as to generate a signal and to indicate the presence of a flaw. The relative motion can be achieved by rotating or oscillating either the tube or the probes. As in eddy current inspection, various types of instrumentation have been developed and are available commercially (Ref 7, 8, 9, 10).

Limitations. Because of the nature of the flux leakage test, tube diameter is not a limitation, but the wall thickness that

can be tested is limited by the ability of the magnetic flux to penetrate the wall and the ability of the sensor to sense flaws at a distance from the wall. Production applications have been used on tubing having wall thicknesses up to 25 mm (1 in.), but 7.6 mm (0.3 in.) is the usual limit. At wall thicknesses in excess of 7.6 mm (0.3 in.), sensitivity becomes a serious problem. However, current developments are improving capabilities in testing thick walls (Ref 8, 9, 10). Although the flux leakage method usually detects flaws that are longitudinally oriented, the principle of the flux leakage method can be used in the design of equipment for detecting transverse flaws. Pinholes, with minimal longitudinal dimensions, and subsurface flaws are difficult to detect. For reliable detection of isolated pinholes, the pitch of the helical inspection path must be small, and the production rate is correspondingly limited. Sensitivity to subsurface flaws drops rapidly as the flaws are located farther from the surface. To detect inside-surface flaws such as cracks and gouges, flux leakage equipment requires special design features for reliable quantitative evaluation. Speed of inspection is a function of the dimensions of the elements involved and the maximum tolerable length of

flaw. Because the tube or the probe must be rotated or oscillated, only a helical or zigzag path is inspected, and the pitch of the helix or the distance between reversals must be less than the maximum tolerable length of flaw. When the tube must be fed over a central conductor for magnetization and then removed for inspection, high-speed production is hindered. The use of multiple probes reduces the actual testing time in proportion to the number of probes, but the time required to feed the tube over the conductor remains constant. Such installations are operated in production at speeds as high as 15 m/min (50 sfm). Installations using external magnetization are reported to operate at speeds to 60 m/min (200 sfm). Weld twist presents a problem in any installation in which only the weld is inspected. In flux leakage inspection, the

problem is solved by increasing the magnitude of the arc covered by the oscillating probe. End Effect. As in eddy current inspection, the error caused by end effect can be minimized by butting the ends of the

tubes together during the test. Mechanical conditions, such as tube ovality, variations in linear speed, and transverse movement of the tube, have

adverse effects on the test results and must be controlled. Equipment Costs. As with eddy current equipment, the cost of equipment for flux leakage inspection varies. An

elementary unit costs slightly more than a comparable eddy current unit because of the need for rotating devices. The addition of auxiliary equipment, such as automatic markers, recorders, computer analyzers, and special handling devices, markedly increases cost. Operating costs, which are relatively low, depend on the degree of automation and the degree to which the inspection

can be combined with other operations. Flux leakage tests can sometimes be combined with another operation--for example, welding. As the tube emerges from the welding operation, it enters the field of the electromagnetic yoke (Fig. 3), which generates a flux in the weld area. The oscillating probe detects any flux diverted by a flaw in the weld. In most cases, however, the need for movement of the probe through the diverted flux makes the combination less desirable than systems with no moving parts.

Ultrasonic Inspection Ultrasonic inspection is one of the most widely used methods for inspecting tubular products (see the article "Ultrasonic Inspection" in this Volume). Widespread use of the ultrasonic method on tubular products was made practical by the development of angle-beam shear-wave testing, immersion testing, and focused transducers. As with the eddy current and flux leakage methods, ultrasonic inspection can be applied either to the entire tube or to the weld only. Ultrasonic inspection of the entire welded tube is usually limited to small-diameter, drawn products, in which the weld cannot easily be distinguished from the remainder of the tube. The tubing may have a diameter as small as 3.2 mm ( in.) and a wall thickness of only 0.25 mm (0.01 in.). These small products are usually inspected while immersed in water (immersion inspection). They are rotated as they pass longitudinally through a glanded immersion tank. The immersed transducers must be carefully selected for tube diameter, wall thickness, and type of imperfection to be located. The transducer, focal length, response to outside diameter and inside diameter calibration notches, instrumentation pulse rate, gate adjustment for flaw alarm, and speed of tube travel are all variables to be taken into consideration. Inspection is usually performed slowly ( 0.9 m/min, or 3 sfm). Tubes must be clean, straight, round, and of uniform dimensions. All types of flaws that commonly occur in resistance welds, except cold weld, can be detected by ultrasonic inspection. Fig. 3 Setup for the flux leakage inspection of welded steel tubing

Most ultrasonic inspection of resistance-welded tubing is restricted to the weld zone and is performed immediately after the welding operation. Components and adjustments for inspecting the weld must be carefully selected and accurately controlled. The transducer must be appropriate for the size and type of flaws to be detected. Focused transducers are generally preferred. The shear-wave angle must be selected for the best evaluation of imperfections. The angle often used is 45°, but tests have revealed that angles between 50 and 70° yield signals more directly proportional to the area of flaws in the weld plane (Ref 3, 11). In the inspection of pipe, provision must be made to maintain the spacing between the transducer and the pipe constant within close tolerances as the pipe moves past the transducer. The couplant should preferably be continuously delivered to the surface of the pipe through openings in the transducer mounting. Coupling through a water jet is also used. Particular attention should be given to the detection of short flaws. Some ultrasonic pipe-inspection equipment will not detect flaws shorter than 6.4 mm (

in.), which will not satisfy the inspection requirements for most resistance-welded pipe.

A disadvantage of the ultrasonic method in tube inspection is its high sensitivity to minor scratches and to elongated

dimensional changes, such as the ridge left when the weld flash is not completely removed or rolled down. However, proper selection of inspection equipment can minimize this problem (Ref 11). An important development is the wheeltype search unit. The transducer of the wheel-type search unit is mounted on the axle of a liquid-filled wheel and is held in a fixed position as the wheel rotates. The surface of the wheel is flexible and adapts itself to the surface condition of the tube as it rolls over it. A small amount of liquid couplant, usually water, is required between the surface of the wheel tire and the surface of the tube. This arrangement provides most of the advantages of immersion testing without the necessity of immersing the tube. Speed of inspection is limited by the pulse rate of the ultrasonic equipment and by the maximum length of a tolerable

imperfection. Speeds as high as 69 m/min (225 sfm) have been reported, but unless multiple inspection heads are used, speed is ultimately dependent on the rejectable flaw size.

Weld twist can present a problem; as the weld twists away from the critical location, transducer sensitivity drops

sharply. To maintain the weld and the transducer in the correct mechanical relationship, the weld can be positioned automatically by the use of an electromagnetic control. End effect, although less of a problem than in eddy current inspection, is a factor in ultrasonic inspection, and

supplementary testing may be necessary if inspection of the tube ends is critical. The supplementary test can be made with ultrasonic equipment of special design. Mechanical variables are critical in contact ultrasonic testing. Spacing between the transducer and the surface of the

tube, angle of transducer, and sidewise movement of tube must be accurately controlled. These variables can sometimes be better controlled in immersion testing. The equipment costs of ultrasonic inspection equipment are highly dependent on the amount of auxiliary equipment

included. Accessories such as automatic marking devices, computer analyzers, and material-handling equipment can markedly increase equipment costs, especially for the inspection of heavy pipe. The operating costs of ultrasonic inspection, in accord with other inspection methods, depend on whether inspection

is operated separately or combined with another operation. For example, if inspection is incorporated into the welding line, an inspector usually is not required, and the operating costs are minimal. Example of Practice. Details of an ultrasonic inspection procedure applied to resistance-welded stainless steel tubing

that was used to clad nuclear fuel elements are given in the following example. Example 1: Ultrasonic Inspection of Welded Type 304 Stainless Steel Tubing Used To Clad Nuclear Fuel Elements. The ultrasonic method, employing the immersion pulse-echo shear-wave technique, was used in the inspection of resistance-welded type 304 stainless steel tubing that was used as cladding for nuclear fuel elements. A typical tube size was 9.88 ± 0.013 mm (0.389 ± 0.0005 in.) ID and 0.419 mm (0.0165 in.) ± 5% wall thickness. A standard length was 4.6 m (15 ft), and the tubing was inspected in the 10 to 15% cold-worked condition. The reference standard, which was a tube selected from the same production lot as that inspected, contained

longitudinal and transverse notches, 1.59 mm (0.0625 in.) long by 0.041 mm (0.0016 in.) deep by 0.10 mm (0.004 in.) wide, prepared by electrical discharge machining, in outside and inside walls. Test Conditions. The tubing, in a clean and dry condition, was inspected before being cut to final length. The tubing

was propelled through an immersion water tank on a preselected helix by a variable-speed drive. Throughput speed was 2.4 m (8 sfm). The tubing was rigidly supported to maintain accurate alignment between tubing and search unit. Direction of Testing. Normally, testing was conducted in two directions only--one a circumferential mode, searching

for longitudinal outside or inside notches, and the other a longitudinal mode, searching for circumferential or transverse notches. When required, two additional searches and search units were employed to search in directions opposite to those mentioned. Proper balance between outside and inside notches was obtained by careful alignment of search units and gate adjustments on the CRT. Search Units. Focused search units, using either spherical or cylindrical transducers, were used. Frequencies ranged

from 3.8 to 4.8 MHz, and the active beam profile was equal to, or shorter than, the notch length, thus ensuring the most reproducible test. The search unit in the longitudinal direction operated at a frequency of 3.8 MHz with a cylindrical line focus of 1.1 × 0.51 mm (0.045 × 0.020 in.). The search unit in the transverse direction operated at 4.8 MHz with a focus of 0.89 × 0.51 mm (0.035 × 0.020 in.). Commercially available electronic equipment with a pulse rate of at least 5 kHz, a gated alarm, and a pulse-stretching circuit was used. Readout was by a chart recorder. Magnetic Particle Inspection The principal use of the magnetic particle method in the inspection of resistance-welded pipe is largely limited to the inspection of pipe ends. In some pipe applications, the ends of the pipe are the sections most critically loaded, and magnetic particle inspection of the ends supplements inspection of the remainder of the pipe by other methods. In the past, the method was widely used to inspect the entire area. However, its inability to detect significant subsurface flaws,

even when the magnetic particles are coated with a fluorescent, and its dependence on human vision and judgment led to its replacement by eddy current and ultrasonic methods. The magnetic particle method is still used in the mill to help establish the precise location of flaws previously detected by other inspection methods. Additional information is available in the article "Magnetic Particle Inspection" in this Volume. Liquid Penetrant Inspection Liquid penetrants (visible-dye and fluorescent) are ordinarily used on nonferromagnetic materials, which constitute only a small fraction of resistance-welded tubular products. Testing speeds are extremely slow, and use of these methods can be justified only when the hazard involved in end use justifies extreme inspection precautions. In such cases, the penetrant methods usually supplement other methods. Additional information is available in the article "Liquid Penetrant Inspection" in this Volume. Radiographic Inspection Radiographic methods of inspection cannot be used successfully on the longitudinal seam of resistance-welded pipe, because the predominant flaws are essentially two dimensional and have little or no effect on the radiographic film. However, when the ends of resistance-welded pipe are butt welded together, arc welding is frequently used, and the method normally used to inspect arc-welded joints is radiography. Additional information is available in the articles "Radiographic Inspection," "Industrial Computed Tomography," and "Neutron Radiography" in this Volume.

References cited in this section

1. "Nondestructive Testing Terminology," Bulletin 5T1, American Petroleum Institute, 1974 2. H.C. Knerr and C. Farrow, Method and Apparatus for Testing Metal Articles, U.S. Patent 2,065,379, 1932 3. W.C. Harmon, "Automatic Production Testing of Electric Resistance Welded Steel Pipe," Paper presented at the ASNT Convention, New York, American Society for Nondestructive Testing, Nov 1962 4. W.C. Harmon and I.G. Orellana, Seam Depth Indicator, U.S. Patent 2,660,704, 1949 5. J.P. Vild, "A Quadraprobe Eddy Current Tester for Tubing and Pipe," Paper presented at the ASNT Convention, Cleveland, American Society for Nondestructive Testing, Oct 1970 6. H. Luz, Die Segmentspule--ein neuer Geber für die Wirbelstromprüfung von Rohren, BänderBlecheRohre, Vol 12 (No. 1), Jan 1971 7. W. Stumm, Tube-Testing by Electromagnetic NDT (Non-Destructive Testing) Methods: I, Non-Destr. Test., Vol 7 (No. 5), Oct 1974, p 251-258 8. F. Förster, The Nondestructive Inspection of Tubings for Discontinuities and Wall Thickness Using Electromagnetic Test Methods: I, Mater. Eval., Vol 28 (No. 4), April 1970, p 21A-25A, 28A-31A 9. F. Förster, The Nondestructive Inspection of Tubings for Discontinuities and Wall Thickness Using Electromagnetic Test Methods: II, Mater. Eval., Vol 28 (No. 5), May 1970, p 19A-23A, 26A-28A 10. P.J. Bebick, "Locating Internal and Inside Diameter Defects in Heavy Wall Ferromagnetic Tubing by the Leakage Flux Inspection Method," Paper presented at the ASNT Convention, Cleveland, American Society for Nondestructive Testing, Oct 1974 11. H.J. Ridder, "New Nondestructive Technology Applied to the Testing of Pipe Welds," Paper presented at the ASME Petroleum Conference, New Orleans, American Society of Mechanical Engineers, Sept 1972 Nondestructive Inspection of Tubular Products

Inspection of Double Submerged Arc Welded Steel Pipe Double Submerged Arc Welded Steel Pipe

Most arc-welded tubular products are produced by the double submerged arc process, in which the seam is welded in two passes, one on the outside and the other on the inside. Tube and pipe diameters range from 457 to 2030 mm (18 to 80 in.), and wall thicknesses from 6.4 to 19 mm (

to

in.). Nondestructive inspection is usually confined to the weld area.

Flaws. Some of the flaws usually encountered in double submerged arc welds are incomplete fusion, incomplete

penetration, offset of plate edges, out-of-line weld beads, porosity, slag inclusions, and weld-area cracks (Ref 1). These flaws are illustrated in Fig. 4 and described below.

Fig. 4 Typical flaws in double submerged arc welded steel pipe. (a) Incomplete fusion. (b) Incomplete penetration. (c) Offset of plate edges. (d) Out-of-line weld beads (off seam). (e) Porosity (gas pocket). (f) Slag inclusions. (g) Weld-area crack

Incomplete fusion (Fig. 4a) is a lack of complete coalescence of some portion of the metal in a weld joint. Incomplete penetration (Fig. 4b) is a condition in which the weld metal does not continue through the full thickness

of the joint. Offset of plate edges (Fig. 4c) is the radial offset of plate edges in the weld seams. Out-of-line weld beads (off-seam) (Fig. 4d) is a condition in which the inner or outer weld beads, or both, are sufficiently out of radial alignment with the abutting edges of the joint to cause incomplete penetration. Porosity (gas pocket) (Fig. 4e) consists of cavities in a weld caused by gas entrapped during solidification. Porosity

may occur as subsurface or surface cavities. Slag inclusions (Fig. 4f) are nonmetallic solid material trapped in the weld deposit or between weld metal and base

metal. Weld-area cracks (Fig. 4g) are cracks that occur in the weld deposit, the fusion line, or the heat-affected zone.

Radiographic Inspection Film x-ray was the first nondestructive inspection method applied to the quality control of double submerged arc welded steel tubular products. The film technique was expensive and was subsequently replaced by the image intensifier with fluorescent screen, an arrangement that requires the inspector, stationed in a darkened room, to observe the fluorescent screen continuously while the pipe travels over the x-ray tube. Rejection of the pipe depends on an appraisal of the degree of darkness registered on the screen. This method is still in use but is being supplemented, and sometimes replaced, by ultrasonic inspection.

Limitations. In addition to high cost and dependence on the human factor, a lesser capacity for discrimination is another

important reason for the replacement of x-ray radiographic techniques by ultrasonic inspection methods. Radiography is sensitive to flaws only when the flaws significantly alter the ability of the material to absorb radiation; this usually occurs when a flaw changes the effective metal thickness by 2% or more or when the flaw causes a change in density equivalent to at least a 2% change in metal thickness. The technique is sensitive to small gas pockets but is relatively insensitive to tightly closed cracks (Ref 12). Because cracks, especially those on the surface, are much more damaging than small totally embedded gas pockets, it is often essential to supplement radiographic inspection with another method that is more sensitive to cracks (Ref 13). Speed of inspection by the continuous radiographic method depends on the eyesight, alertness, and judgment of the

inspector and is usually in the range of 3 to 9 m (10 to 30 sfm). Weld Twist. Because of the thick cross section and the width of the weld bead characteristic of double submerged arc

welded steel pipe, weld twist is usually not a problem. End effect is an important consideration. Consequently, most specifications require that the continuous radiographic

inspection of pipe ends be supplemented by magnetic particle inspection, film x-ray inspection, or both. Equipment Costs. Most of the expenditure is directed toward the mechanical arrangements needed for handling the

large-diameter, thick-wall product. Precise positioning, uniform linear speed, and convenient controls for slowing speed to inspect flaw indications are essential for satisfactory inspection. The operating costs of continuous radiographic inspection are comparatively high because of the slow speeds and the

need for two operators in most installations. Ultrasonic Inspection In the ultrasonic inspection of double submerged arc welded pipe, the excess bead at the welded seam presents a problem in that the edges of the bead generate false indications. On the other hand, the excess bead provides a guide for maintaining the inspection head in the correct position as it tracks the weld. One solution to the problem of extraneous reflections is the use of two search units, each with two or more focused transducers and appropriate electronic accessories (Ref 14). In this method, the pipe moves into the testing station and is brushed and prewetted. An optical system locates the weld precisely. As the pipe moves forward, the two search units are lowered into the test position. As the end of a pipe leaves the tester, the search units are automatically raised from the pipe. The transducers are positioned to produce shear waves at an angle of approximately 70° in the pipe. The electronic controls are adjusted so that part of the weld zone and a 25 mm (1 in.) wide band of adjacent strip are inspected by each search unit. The inspected portions of the weld zone overlap to ensure thoroughness of inspection. Flaws exceeding a predetermined standard automatically operate the appropriate marker. The ultrasonic method is not as sensitive to gas pockets as the radiographic method but is much more sensitive to tightly closed cracks. Such cracks can be detected even when they are located at the junction of the weld bead and the base metal.

References cited in this section

1. "Nondestructive Testing Terminology," Bulletin 5T1, American Petroleum Institute, 1974 12. R.F. Lumb and G.D. Fearnebaugh, Toward Better Standards for Field Welding of Gas Pipelines, Weld. J., Vol 54 (No. 2), Feb 1975, p 63-s to 71-s 13. M.J. May, J.A. Dick, and E.F. Walker, "The Significance and Assessment of Defects in Pipeline Steels," British Steel Corporation, June 1972 14. W.C. Harmon and T.W. Judd, Ultrasonic Test System for Longitudinal Fusion Welds in Pipe, Mater. Eval., March 1974, p 45-49 Nondestructive Inspection of Tubular Products

Arc-Welded Nonmagnetic Ferrous Tubular Products Austenitic stainless steel and other nonmagnetic ferrous tubular products are, except for seamless tubing, usually fabricated from plate, sheet, and strip by forming and arc welding, frequently by the gas tungsten-arc process. The weld flaws encountered are similar to those found in double submerged arc welded products, except for those resulting from deposition of the second bead. Although tube diameters range from 3.2 to 762 mm ( to 30 in.) and wall thicknesses from 0.10 to 9.53 mm (0.004 to 0.375 in.), the small-diameter, thin-wall products predominate. Tubes having diameters from about 3.2 to 102 mm (

to 4 in.) are produced on a draw bench.

Eddy Current Inspection. Although the austenitic stainless steels are nominally nonmagnetic, some of these alloys

will develop magnetic constituents (notably ferrite) in the weld fusion zone and in cold-worked areas, thus causing variations in permeability. Consequently, magnetic saturation is required for the reliable eddy current detection of small flaws in the weld zones of these alloys. In addition, because of their lower electrical conductivity, stainless steels require higher frequencies than those used for the eddy current inspection of carbon steels. In other respects, the practices employed in eddy current inspection are the same as those applied to carbon steel products. Because diameters in the range of 3.2 to 75 mm ( to 3 in.) account for most of welded stainless steel tube production, encircling inductor and detector coils can be used in inspection. Usually, the coils are close fitting, with an allowable in.). The entire tube is inspected. For calibration purposes, a hole drilled maximum clearance of only 1.6 mm ( through the tube wall is ordinarily used. Allowable flaws are described in ASTM A 450. In general, inspection speeds range from 7.6 to 46 m/min (25 to 150 sfm), although speeds up to 140 m/min (450 sfm) have been attained. Ultrasonic inspection is performed on tubes ranging from 3.2 to 762 mm (

to 30 in.) in diameter with wall thicknesses ranging from 0.10 to 9.52 mm (0.004 to 0.375 in.). Barium titanate, lithium sulfate, quartz, or lead zirconate can be used as the transducer element; operating frequencies range from 2.25 to 15 MHz. In general, the pulse-echo shear-wave technique is used with immersed, focused transducers. For larger-diameter tubes, where the size can cause practical difficulties in immersion and where the curvature of the tube is slight, flat crystals with conventional contact coupling can be used. In immersion testing, the shear-wave angle in the testpiece ranges from 45 to 70°. When the application is critical, ultrasonic inspection must be supplemented by other methods, such as eddy current and liquid penetrant inspection. On larger-diameter tubing, only the weld is tested. Usually, the transducer is stationary and the tube is moved axially. With smaller-diameter tubing, the entire tube is tested. The transducer is usually stationary while the rotating tube is moved axially. The pitch of the helix inspected is a function of the maximum length of the flaw that can be tolerated. The calibration tube contains one or more milled slots, as required by the applicable specification. Production rate depends largely on the amount of metal being inspected--that is, the weld zone only or the entire tube-and can range from a few inches to about 1.5 m/min (5 sfm). Speeds as high as 6 m/min (20 sfm) have been attained. Liquid Penetrant Inspection. Penetrants used for the inspection of tubing may be either the fluorescent or

nonfluorescent (visible-dye) type. The fluorescent penetrants may be water soluble or an emulsion. For a reliable test, the tube must always be cleaned before testing. Because the appraisal of flaws is based on visual judgment, inspection is usually restricted to the outside surface unless the tube is large enough and short enough to permit inspection of the inside surface. Production rate is highly dependent on the mechanical handling facilities available and ranges from a few inches per minute to as much as 3 m/min (10 sfm). Radiographic Inspection. Some users specify that film x-ray must be used (Ref 15). Radiography is used to inspect

tubes ranging in diameter from 6.4 to 762 mm ( to 30 in.) and in wall thickness from 0.64 to 9.53 mm (0.025 to 0.375 in.). Calibration is by penetrameter. Either the tube or the tester may move with respect to the other. The usual flaw detection difficulties associated with radiography are compounded with the smaller diameters because the x-ray beam must pass through two wall thicknesses. Production speed ranges from the average of a few inches per minute to as much as 3.7 m/min (12 sfm).

Reference cited in this section

15. "Inspection, Radiographic," Military Standard 453A, May 1962 Nondestructive Inspection of Tubular Products

Continuous Butt-Welded Steel Pipe Most carbon steel pipe is produced by a process that continuously forms hot strip (skelp) into tubular shape, then passes it through a welding horn and welding stand, where the edges are pressed firmly together and pressure welded to provide the longitudinal seam. The strip ends are flash butt welded together without stopping the line, thus making the process truly continuous. A multiple-stand stretch-reducing mill brings the welded pipe to its finished size. As the name implies, this mill is used for reducing the diameter of the pipe being produced by simultaneously applying pressure and tension. Pipe diameters range from 22 to 114 mm (

to 4

in.) and wall thicknesses from 1.9 to 12.7 mm (0.076 to 0.500 in.).

The inspection of continuous butt-welded pipe is directed primarily toward in-plant quality control, although it is required by some specifications. The eddy current method is employed almost exclusively and is applied not only to cold pipe but also to hot pipe at a temperature above the Curie point. Encircling inductor coils and differentially wound detector coils are used with a coil-to-pipe spacing of 6.4 mm ( in.) or less and an operating frequency of 2.5 kHz. A saturating field of about 70,000 ampereturns is required for cold pipe, which is nonmagnetic at temperatures above the Curie point. The coil assembly for hot pipe must be water cooled and encased in heat-resistant material (Ref 16). Marking the hot pipe requires a special high-temperature paint. The eddy current inspection setup is usually calibrated to reject all flaws, internal and external, that extend more than 12 % through the wall. The calibration standard is usually a drilled hole. The speed of inspection corresponds with the speed of the mill and, for cold pipe, ranges from 53 to 150 m/min (175 to 500 sfm). Higher speeds are employed for inspecting hot pipe.

Reference cited in this section

16. W. Stumm, New Developments in the Eddy Current Testing of Hot Wires and Hot Tubes, Mater. Eval., Vol 29 (No. 7), July 1971, p 141-147 Nondestructive Inspection of Tubular Products

Spiral-Weld Steel Pipe Spiral-weld steel pipe is made principally in large diameters and is a relatively low-production product. The pipe is either resistance or submerged arc welded. Resistance-welded pipe is inspected by either the eddy current or the ultrasonic method. The ultrasonic or radiographic method is used to inspect submerged arc welded pipe. Usually, the ends of pipe welded by either process are also inspected by a supplementary technique, such as film x-ray or magnetic particle inspection. Nondestructive Inspection of Tubular Products

Seamless Steel Tubular Products Steels melted by various processes can be successfully converted into seamless tubes. In general, killed steels made by open-hearth, electric-furnace, and basic-oxygen processes are used. Because of the severity of the forging operation involved in piercing, the steels used for seamless tubes must have good characteristics with respect to both surface and internal soundness. A sound, dense cross section, free from center porosity or ingot pattern, is the most satisfactory for seamless tubes. Metallurgical developments have contributed greatly to the improvement of steels for seamless tubes; as a result, the seamless process has been extended to include practically all of the carbon and alloy grades of steel. Flaws in seamless tubular products may occur at any point on the outside and inside surfaces or within the tube wall (Ref

1). The flaws usually encountered are blisters, gouges, laminations, laps, pits, plug scores, rolled-in slugs, scabs, and seams. These are illustrated in Fig. 5 and described below.

Fig. 5 Typical flaws in seamless tubing. (a) Blister. (b) Gouge. (c) Lamination. (d) Lap (arrow). (e) Pit. (f) Plug scores. (g) Rolled-in slugs. (h) Scab. (j) Seam (arrow)

Blisters (Fig. 5a) are raised spots on the surface of the pipe caused by the expansion of gas in a cavity within the wall. Gouges (Fig. 5b) are elongated grooves or cavities caused by the mechanical removal of metal. Laminations (Fig. 5c) are internal metal separations creating layers generally parallel to the surface.

Laps (Fig. 5d) are folds of metal that have been rolled or otherwise worked against the surface but that have not been

fused into sound metal. Pits (Fig. 5e) are depressions resulting from the removal of foreign material rolled into the surface during manufacture. Plug scores (Fig. 5f) are internal longitudinal grooves, usually caused by hard pieces of metal adhering to the mandrel,

or plug, during plug rolling. Rolled-in slugs (Fig. 5g) are foreign metallic bodies rolled into the metal surface, usually not fused. Scabs (Fig. 5h) are flaws in the form of a shell or veneer, generally attached to the surface by sound metal. Usually,

scabs originate as ingot flaws. Seams (Fig. 5j) are crevices in rolled metal that have been closed by rolling or other work but have not been fused into

sound metal. Ultrasonic inspection is probably the method most commonly used on seamless tubular products, which range from

3.2 to 660 mm ( to 26 in.) in diameter and from 0.25 to 64 mm (0.01 to 2 in.) in wall thickness. The tube, while rotating, is usually moved longitudinally past the transducers, thus providing inspection along a helical path. In a typical installation, six transducers inspect the rotating tube as it progresses through the machine. Four transducers are below the tube and are coupled to it by water columns; two transducers are above the tube and make contact through fluid-filled plastic wheels. This machine is capable of handling tubes ranging from 50 to 305 mm (2 to 12 in.) in diameter. As with welded tubing, the smaller diameters of seamless products are inspected while immersed, and the larger diameters are inspected by direct contact. Transducer crystals may be quartz, lithium sulfate, barium titanate, or lead zirconate. The frequency ordinarily used is 2.25 MHz. The transducer may or may not be focused, depending on the tube diameter and the nature of the flaws anticipated. In most cases, the ultrasonic shear-wave angle is 45° but may be as large as 70°. The usual couplant is water, but oil is sometimes used. Because all installations involve rotation of either the tube or the transducers, the inspection invariably follows along a helical path. The pitch and width of the helical path vary widely, depending on the characteristics of the equipment and the specifications to be met. The pitch is usually between 9.5 and 13 mm ( and in.), which translates to two to three revolutions per each 25 mm (1 in.) of longitudinal travel. Almost all forms of visible and audible alarms, as well as automatic recorders, are used with the ultrasonic equipment. All types of flaws in seamless tubes can be detected by ultrasonic methods, but the minimum flaw dimensions, the degree of sensitivity, the flexibility of adjustment, and the accuracy of calibration all vary widely with the basic instrumentation and the supplementary components chosen. The flaw used most frequently for calibration is a longitudinal slot. The depth of slot may vary from 3 to 12 % of the wall thickness, depending on the end use of the product and the specification involved. The length of the slot may be as much as 38 mm (1 in.) but is usually specified as twice the width of the transducer. Widths of slots should be kept at a minimum and should never exceed twice the depth. The frequency of calibration checks depends on the criticality of the tube application. In a few cases, a calibration check is required after every tube, but once every 4 h of production is usually considered adequate. Speed of inspection also varies widely, depending on many variables, especially the maximum tolerable length of flaw and the number of transducers used. The current range of inspection speed is 0.6 to 46 m/min (2 to 150 sfm); the upper limit can be increased by increasing the number of transducers. The following example describes the setup used for the ultrasonic inspection of stainless steel tubular products, with special emphasis on the calibration procedures used for flaw detection.

Example 2: Ultrasonic Inspection of Seamless and Welded Austenitic Stainless Steel Tubular Products. Seamless and welded austenitic stainless steel tubular products were inspected by ultrasonics, using a system that rotated and moved the pipe or tube past a stationary ultrasonic search unit. The system was equipped to inspect by either contact

or immersion. The inspection unit contained a rotational and longitudinal drive mechanism and a relatively small openended and glanded immersion tank through which the tube or pipe passes (Fig. 6). The tank also contained the search units.

Fig. 6 Unit used for the ultrasonic inspection of seamless and welded stainless austenitic steel tubular products. A, tube being inspected; B, immersion tank; C, drive wheels; D, search-unit tube; E, drive mechanism. Inset shows lateral displacement of the search unit for circumferential inspection to detect longitudinal flaws.

To inspect, a reference tube or pipe was usually first placed in the unit. The reference tube or pipe was of a convenient length (usually 1.2 to 1.8 m, or 4 to 6 ft) and similar in type, size, and wall thickness to the product to be inspected. The reference tube contained notches in both outside and inside walls. Generally, the depth of the notches varied from 3 to 12 % of the wall thickness, with the shallower notch being used for seamless tube and the deeper notch for welded tube. Most notches ran in the longitudinal direction and were approximately 25 mm (1 in.) long. Typically, the width of a notch was not more than twice its depth. In some cases, transverse notches were used as well as longitudinal notches. The dimensions of the transverse notch were the same as those of the longitudinal notch. The depth of notches was never less than 0.10 mm (0.004 in.). With the reference standard placed in the inspection unit, suitable search units were selected. For the detection of longitudinal flaws, a line-focused search unit was generally used. The focal length varied from approximately 25 to 114 mm (1 to 4 in.). The short focal length was used for small-diameter tubing, and the longer focal tubing length was used for larger-diameter tubing. The length of the beam (lined up with respect to the longitudinal axis of the tube or pipe) varied from approximately 6.4 to 25 mm (

to 1 in.). The length of the beam used depended on the size of the flaw to be

detected. Shorter beam lengths were used to detect smaller (shorter) flaws. A beam 13 mm ( in.) long was normally used. To detect transverse flaws, a spherically focused search unit could be used in addition to a line-focused unit.

The search unit was centered with respect to the reference tube, ensuring that a water gap between the search unit and the tube was equal to the focal length of the search unit minus the radius of the tube. Angular adjustment and positioning were performed to obtain a maximum response from the tube wall indicative of proper centering. For circumferential inspection to detect longitudinal flaws, the search unit was displaced laterally or set off from the center of the tube to obtain the proper angle at which the sound would travel through the tube wall (inset, Fig. 6). Generally, inspection was performed using a shear wave traveling at an angle of 45° around the tube. The offset distance required was calculated using the formula:

L = (D/2)(vwvm) sin θ

(Eq 1)

where L is the offset distance, D is the diameter of the tube or pipe, vw is the velocity of sound in water, vm is the velocity of sound in the tube, and is the desired refracted angle in the tube or pipe. The value for the velocity of sound in the tube depended on whether a longitudinal or shear wave was to be used. The offset value obtained from Eq 1 was an approximation. The actual amount of offset was adjusted from that calculated to obtain best presentation and equalization of responses from outside and inside notches. When the search unit and reference tube or pipe had been properly located with respect to each other, the reference tube was rotated and driven longitudinally past the search unit. Controls of the ultrasonic instrument were adjusted to display a clear response from both inside and outside notches on an oscilloscope screen. The controls of the flaw-alarm module of the instrument were adjusted to position the gate properly to include the signals from the notches and to activate the alarm when they were detected. The pulse-repetition rate of the instrument was adjusted high enough to ensure the detection of all notable flaws at the speed of inspection. The inspection speed was controlled by the rate of rotation and longitudinal movement per revolution (pitch) of the tube. The allowable pitch was a function of the length of the line-focused ultrasonic beam and the size of the flaw to be detected. Normally, when the reference standard was passed through the unit, the controls had been adjusted to provide more than one signal from each of the notches. When this had been established, the system was properly calibrated for production inspection. Ultrasonic Inspection Precautions. Generally, a chart recorder is employed to provide a permanent record of the

inspection described in Example 2. Multiple search units can be used in the immersion tank to provide several simultaneous inspections during one pass of the pipe or tube through the unit. Specifications may call for circumferential inspection from two directions because the reflection from a flaw may vary, depending on the direction in which the ultrasonic beam strikes it. In addition to the circumferential inspection to detect longitudinal flaws, a longitudinal (axial) inspection may be required to detect transverse flaws. Also, it may be desirable to ultrasonically measure wall thickness, eccentricity, or both. All these tests can be performed at the same time by utilizing search units that are designed for the tests and that are properly positioned in the tank. Normally, rejection is based on the presence of flaw indications exceeding those from the reference notch. Reworking and reinspection are generally permitted if other requirements, such as minimum wall thickness, are satisfied. Other refinements are included in, or can be added to, the inspection system. For example, the feeding of tubes to the unit and their withdrawal can be automated. Various audible and visible alarm systems and marking devices can be added. Normally, the water used in the system is filtered and deaerated. Air entrapped in the water can produce false indications. Similarly, water on the inside surfaces of tubes will produce false signals, and these surfaces must be kept dry. Tubes are connected to each other by stoppers or by taping the ends together. The glands at each end of the water tank must be cut in a manner that will allow passage of the tubes without undue loss of water couplant. Finally, air must be prevented from being drawn into the entry gland along with the tube. This is usually accomplished by directing a stream of water over the outside of the tube just before it enters the gland. In eddy current inspection, use of an encircling detector coil is limited to a maximum tube diameter of about 75 mm

(3 in.). As tube diameter increases, the ratio of flaw size to tube diameter decreases; consequently, the flaw is increasingly more difficult to detect. This problem is overcome by using several small probe coils (Ref 6, 7, and 17) and with spinning probes (Ref 8). When probe coils are used, the flaw becomes a significantly high percentage of the zone surveyed. Because independently mounted probes ride over the tube surface, good magnetic coupling is ensured. Magnetic saturation is used to obtain maximum sensitivity to flaws close to, or on, the inside surface of the tube, and the frequency

of the test current is kept relatively low, sometimes as low as 1 kHz. Internal spinning probes can also be used if a lower production rate can be tolerated (Ref 8). In some installations, the eddy current test with magnetic saturation is supplemented by a probe-type eddy current test, which in effect provides high-sensitivity inspection of the surface (Ref 18). Four probe coils, as shown in Fig. 7, each serving as both inductor and detector, rotate about the tube as it moves longitudinally through the rotating assembly. Magnetic saturation is not required. The segment of the tube in which the flaw occurs is identified by markers. In Fig. 8, the test head is shown ready for use, with eight paint-spray guns in position for marking the proper zone.

Fig. 7 Unit used for the probe-type eddy current inspection of seamless steel tubing. A, outer cover, containing test head (Fig. 8), in open position; B, one of four rotating eddy current probe coils; C, reference-standard testpiece in position for calibration; D, one of eight paint-spray guns for marking

Fig. 8 Test head of the eddy current inspection unit shown in Fig. 7. A, orifice for test pipe or tube; B, one of eight paint-spray guns for marking; C, reference-standard testpiece

Eddy current inspection is used on seamless tubular products ranging from 3.2 to 244 mm ( to 9 in.) in diameter and from 0.25 to 14.0 mm (0.01 to 0.55 in.) in wall thickness. In most cases, magnetic saturation is used, but when the primary concern is the detection of surface imperfections, small probe coils without magnetic saturation are used. If the steel is entirely nonmagnetic, no saturation is required in any system. The frequency used ranges from 1 to 400 kHz and depends on such variables as the wall thickness of the pipe or tube, the coil design and arrangement, and the use of saturation. The spacing between the pipe or tube and the test coil(s) varies widely. However, for high sensitivity and accuracy, it should be kept to a minimum and is occasionally kept to as little as 0.25 mm (0.01 in.). This clearance is insufficient for practical use, and the usual clearance for production testing ranges from 1.6 to 19 mm (

to

in.).

Although all flaws that usually occur in seamless pipe or tube can be detected by eddy current methods, external flaws are more easily detected than internal flaws. Laminations are the most difficult flaws to detect. Some installations are intended to detect surface flaws only. Flux leakage techniques are used for the inspection of seamless tubular products ranging from 32 to 914 mm (1

to

36 in.) in diameter and from 3.2 to 19 mm ( to in.) in wall thickness. Because the flaws sought usually have a significant longitudinal dimension, transverse magnetic fields are usually used. Longitudinal fields can be used but are rarely considered necessary. As a rule, the transverse magnetic field is produced by a current passing through a conductor located in the center of the pipe or tube. In some cases, the field is produced between the poles of an electromagnet or a permanent magnet whose pole pieces are shaped to fit the pipe or tube diameters. The signal-generating movement of either the tube or the probe with respect to the other is accomplished by rotating either tube or probe. Sensitivity to inside surface flaws is a problem when using these methods; the problem becomes more serious as the wall thickness increases. In some cases, the solution to the problem is a rotating internal probe moving through the tube or kept stationary while the tube moves (Ref 8). Other installations depend on electronic filters and the difference in frequency between the signals for internal and surface flaws (Ref 10). The production rate of flux leakage inspection depends on many factors. The maximum permissible speed of probe or tube movement with respect to the other is about 90 m/min (300 sfm). The circuits will respond almost instantaneously when inspection speeds are kept below this speed limit. The principal limiting factor in production-output speed then becomes the maximum tolerable length of the flaw, which in turn governs the pitch of the helix inspected. However, the production rate can be increased to any desired level by the simultaneous inspection of several segments of the same pipe or tube. Actual inspection speeds range from 0.9 to 60 m/min (3 to 200 sfm), depending on the diameter, the system used, sensitivity required, and other variables. The methods that use external magnets can inspect at a much higher overall rate than those that depend on an internal conductor for magnetization. Magnetic particle and liquid penetrant inspection methods are simple and economical and are most useful for surface inspection in specialized, small-scale applications. When applied to welded tubing, their use can be restricted to the weld zone. However, when applied to seamless tubing, there are no surface restrictions. The inability of these methods to locate small flaws beneath the surface and their dependence on the vision, alertness, and judgment of the inspector limit their usefulness in meeting modern specifications for seamless steel tubing. Radiographic Inspection. The principal application of radiography to seamless tubing, as with welded tubing, is the

inspection of girth welds joining the ends of tubes. Even in this application, it is apparent that it should be supplemented by magnetic particle inspection (Ref 12).

References cited in this section

1. "Nondestructive Testing Terminology," Bulletin 5T1, American Petroleum Institute, 1974 6. H. Luz, Die Segmentspule--ein neuer Geber für die Wirbelstromprüfung von Rohren, BänderBlecheRohre,

Vol 12 (No. 1), Jan 1971 7. W. Stumm, Tube-Testing by Electromagnetic NDT (Non-Destructive Testing) Methods: I, Non-Destr. Test., Vol 7 (No. 5), Oct 1974, p 251-258 8. F. Förster, The Nondestructive Inspection of Tubings for Discontinuities and Wall Thickness Using Electromagnetic Test Methods: I, Mater. Eval., Vol 28 (No. 4), April 1970, p 21A-25A, 28A-31A 10. P.J. Bebick, "Locating Internal and Inside Diameter Defects in Heavy Wall Ferromagnetic Tubing by the Leakage Flux Inspection Method," Paper presented at the ASNT Convention, Cleveland, American Society for Nondestructive Testing, Oct 1974 12. R.F. Lumb and G.D. Fearnebaugh, Toward Better Standards for Field Welding of Gas Pipelines, Weld. J., Vol 54 (No. 2), Feb 1975, p 63-s to 71-s 17. F.J. Barchfeld, R.S. Spinetti, and J.F. Winston, "Automatic In-Line Inspection of Seamless Pipe," Paper presented at the ASNT Convention, Detroit, American Society for Nondestructive Testing, Oct 1974 18. T.W. Judd, Orbitest for Round Tubes, Mater. Eval., Vol 28 (No. 1), Jan 1970, p 8-12 Nondestructive Inspection of Tubular Products

Finned Tubing The production of finned tubing for use in heat exchangers, notably, the heat exchangers used in nuclear reactor installations, has been increasing. These tubes are normally designed with three or six fins, and the presence of these fins precludes the use of several of the conventional inspection techniques. Among the methods considered to be feasible are the magnetic particle, liquid penetrant, and eddy current methods. The tube wall just below the outer surface can be inspected with a high degree of efficiency and speed using special eddy current techniques (Ref 19). For this application, external coils are designed to fit the external tube contours precisely. The transmission method, in which the exiting and receiving coils are placed on different sides of the tube wall, is used. Although ultrasonic inspection can be used, it is prohibitively slow and expensive because of the finned-tube contours.

Reference cited in this section

19. F. Förster, Sensitive Eddy-Current Testing of Tubes for Defects on the Inner and Outer Surfaces, NonDestr. Test., Vol 7 (No. 1), Feb 1974, p 25-35

Nondestructive Inspection of Tubular Products

Duplex Tubing Exploratory tests have demonstrated the feasibility of applying several nondestructive inspection methods to the inspection of duplex tubing used in the atomic energy industry. The tubes, which are made by metallurgically bonding two concentric pieces of 2.25Cr-1Mo steel tubing to make a single piece, have a composite wall thickness of 3.86 mm (0.152 in.) and an outside diameter of 22 mm ( in.). The outer 22 mm ( in.) diam tube is fabricated with four equally spaced axial grooves on the inside surface. These grooves complicate testing, but laboratory experience indicates that a useful inspection of the assembled tubes can be accomplished by the ultrasonic method (Ref 20). Complete inspection requires scanning of the inside 14.5 mm ( in.) diam tube as well as the outer tube surfaces. Eddy current techniques can also be used for inspecting duplex tubing, as will be shown in the next section, "Nonferrous Tubing," which includes an example of the eddy current inspection of aluminum tubing (Example 3).

Reference cited in this section

20. K.J. Reimann, T.H. Busse, R.B. Massow, and A. Sather, Inspection Feasibility of Duplex Tubes, Mater. Eval., Vol 33 (No. 4), April 1975, p 89-95 Nondestructive Inspection of Tubular Products

Nonferrous Tubing* A wide variety of nonferrous alloy tubing, such as tubing made of brass, copper, aluminum, nickel, and zirconium, can be inspected for cracks, seams, splits, and other flaws. Eddy current inspection is the method most widely used, followed by the ultrasonic and liquid penetrant methods. Eddy Current Inspection When eddy current inspection is employed for nonferrous tubing, the range of tube diameters normally permits the use of an encircling coil. The typical flaws respond well to differential-type coils. The frequencies employed usually range from 1 to 25 kHz. The choice of frequency is generally dependent on the electrical conductivity and wall thickness of the tubing. Because magnetic saturation is not required, the inspection equipment is simpler and more compact than that used on ferromagnetic tubing. Tubes range from 3.2 to 89 mm ( to 3 in.) in diameter, with wall thicknesses from 0.25 to 14.2 mm (0.01 to 0.56 in.). Testing speeds to 370 m/min (1200 sfm) have been reported. On a limited basis, eddy current inspection is also being applied to finned copper tubing. Example 3: Corrosion of Duplex 3003-H14 Aluminum Heat Exchanger Tubes Clad With 7072 Aluminum on Inner Surface. An eddy current inspection was performed at a petrochemical complex on a heat exchanger containing 1319 tubes. It was advised that the tube material was aluminum alloy 3003 in the H14 condition. The tube side environment was salt water at a pressure of 240 kPa (35 psig) at an inlet temperature of 30 °C (85 °F) and an outlet temperature of 40 °C (100 °F). The shell side environment contained hydrocarbons, hydrogen sulfide, water, hydrogen, and ammonia. The shell side stream entered at 125 °C (260 °F) and exited at 45 °C (110 °F) at 380 kPa (55 psig). Eddy current inspection of the heat exchanger revealed that 300 out of 1319 tubes exhibited indications characteristic of broad, inner-surface pitting corrosion. The suspect tubes were randomly scattered about the tube bundle.

A tube that exhibited strong indications of inner-surface pitting was pulled, split, and visually examined. The entire outer surface of the tube was covered with a thin, uniform black scale. The inner surface of the tube was coated with lightbrown silt deposits. Broad, shallow corrosion pits were randomly scattered over the entire inner surface. Without exception, at each of the corroded areas, the depth of the corrosion attack measured 0.28 mm (0.011 in.). Figure 9(a) shows the outer and inner surfaces of the subject tube.

Fig. 9 Corrosion in duplex 3003-H14 aluminum heat exchanger tubes clad with 7072 aluminum on inner surface. (a) Macrograph showing aluminum tube samples removed from the subject heat exchanger unit after eddy current inspection. The outer surface of the tube is shown at the top. The center two sections illustrate the condition of the inner surface before cleaning, and the bottom section shows the inner surface after removal of the light silt deposits. Broad patches of corrosion attack are evident on the inner surface. (b) Micrograph of the inner surface of the duplex tube showing corrosion of the 7072 cladding (top section of photograph). The corrosive attack was limited to the clad thickness (0.28 mm, or 0.011 in.), and the substrate 3003 alloy was not affected. 50×. Courtesy of J.P. Crosson, Lucius Pitkin, Inc.

Metallographic examination revealed the tube to be of duplex design; that is, it was clad on the inner surface. The corrosive attack was limited to the cladding thickness of 0.28 mm (0.011 in.) as shown in Fig. 9(b). Qualitative spectrographic analysis revealed the aluminum alloy 3003 tube material to be clad on the inner surface with aluminum alloy 7072. The examination revealed that the subject tubes had undergone galvanic corrosion. The 7072 cladding sacrificially corroded as it electrochemically protected the substrate alloy; 7072 cladding alloy is sufficiently anodic in the tube side environment to provide cathodic protection. Consequently, the corrosive attack penetrated only to the interface, then proceeded laterally along the substrate surface, confined to the cladding. To the extent that the cladding remained over a sufficient area, the substrate material was protected. Therefore, although the tubes exhibited corrosive attack, it was apparent that considerable service life remained, and tube replacement was not recommended. Example 4: Stress-Corrosion Cracking of Copper Absorber Tubes in an Air-Conditioning Unit. Eddy current inspection was performed on a leaking absorber bundle in an absorption air-conditioning unit. The inspection revealed cracklike indications in approximately 50% of the tubes. The tube material was phosphorusdeoxidized copper. Two tubes with indications were pulled and examined visually and metallographically to determine the cause of cracking. The outer surfaces of the tubes were irregularly stained with a green-blue-black film, apparently the result of reaction with the shell side lithium bromide solution. The inner surfaces were covered with a thin, crusty green-black deposit that easily flaked off from the surface. No significant corrosive attack was observed on the inner surfaces. One of the pulled tubes exhibited a large, irregular longitudinal crack, as shown in Fig. 10(a). The other tube exhibited a very fine longitudinal crack, as shown in Fig. 10(b).

Fig. 10 Longitudinal crack and intergranular stress-corrosion cracks in copper air-conditioning absorber tubes. (a) Longitudinal crack in one of the subject absorber tubes. 0.75×. (b) Macrograph of fine, irregular crack observed on the outer surface of the second absorber tube after light acid cleaning to remove the corrosion product. 2×. (c) Micrograph showing profiles of the primary crack and two fine secondary cracks at the outer surface of the subject absorber tube. The crack profiles are typical of stress-corrosion cracking, that is, intergranular and free of any localized grain deformation. 75×. Courtesy of J.P. Crosson, Lucius Pitkin, Inc.

Metallographic examination revealed that the cracks originated at the outer surface, were intergranular in nature, and were free of any localized grain deformation. Such features are characteristic of stress-corrosion cracking in a copper heat exchanger tube. The crack path is shown in Fig. 10(c). The service contractor responsible for maintaining the absorption unit suspected that mercury contamination from a manometer was the cause of the stress-corrosion cracking. However, an electron probe microanalysis performed on a microspecimen did not reveal any mercury at or near the cracks. Chemical analysis of the lithium bromide solution revealed significant quantities of nitrates in the solution. Such nitrates are normally added to the lithium bromide solution to act as corrosion inhibitors. The results of the examination indicated that the absorber tubes failed by stress-corrosion cracking initiated by ammonia contamination in the lithium bromide solution. Cracking was from the shell side or outer surface of the tubes, where the tubes were in contact with nitrate-inhibited lithium bromide. The source of ammonia was apparently the reduction of nitrates by hydrogen evolved during corrosion of the steel shell and/or tubes. Example 5: Burst Copper Evaporator Tubes in an Absorption Air-Conditioning Unit. An eddy current survey of the copper evaporator (chiller) tubes in an absorption air-conditioning unit revealed two tubes in the evaporator bundle with indications typical of longitudinal cracks. The two tubes were pulled for visual and metallographic examination. One tube exhibited a 27 mm (1

in.) long longitudinal crack adjacent to the location of the first tube support. The

second tube exhibited a 14.5 mm ( in.) long longitudinal crack 1735 mm (68 in.) from the front end. The tube surface adjacent to the cracks appeared to be bulged and smeared. The smearing was apparently a result of removing the tubes from the unit. Splitting the tubes revealed the inner surfaces to exhibit a thin, normal, uniform oxide scale. The tube wall beneath the scale was free of any significant corrosion. Metallographic examination revealed significant necking down and grain distortion at the fracture surfaces, as shown in Fig. 11. The fracture features were characteristic of an overload failure in a ductile material. The general microstructure away from the fracture consisted of twinned equiaxed grains typical of annealed copper and was considered satisfactory.

Fig. 11 Micrograph of a transverse section of a burst copper evaporator tube showing the longitudinal rupture present in one of the failed tubes. At the fracture, grain deformation and necking down of the tube wall are evident. Such features are characteristic of overload failure in a ductile material. 55×. Courtesy of J.P. Crosson, Lucius Pitkin, Inc.

The results of the examination indicated that the ruptured tubes failed as a result of excessive internal pressure, as evidenced by the necking down of the tube wall at the fracture. The normal operating pressure of approximately 690 kPa (100 psi) produces a hoop stress in the tube wall of approximately 7.6 MPa (1100 psi), which was well below the yield strength of the material. Therefore, the failure could not be related to minor fluctuations in operation pressure. The source of the excessive internal pressure was most probably a freeze-up of the tube side water that occurred during interruption of the tube side flow or misoperation of the unit. Example 6: Worn Copper Condenser and Chiller Tubes in a Centrifugal Air-Conditioning Unit. Eddy current inspection was performed on the copper chiller and condenser tubes of a centrifugal air-conditioning unit. The results of this inspection revealed the presence of seven chiller tubes exhibiting indications characteristic of a decrease in wall thickness at a location corresponding to the forward tube support plate. Subsequent removal of these tubes confirmed the presence of tube-wall wear at the saddle (smooth portion) of the tube, which sits in the tube support. The tube-wall wear at the support saddles was measured between 10 and 60% of the original wall thickness. Two condenser tubes were also observed with wear at the support saddles. The decrease in wall thickness was measured to be 20% and 50% for the two tubes. Figure 12 shows several tubes with tube-wall wear at the tube support saddles. Results of the inspection attributed the wear to fretting corrosion attack caused by vibrational contact between the tubes and the steel tube support plate.

Fig. 12 Grooves at the tube support saddles formed by fretting corrosion attack due to vibrational contact between the copper tubes and the steel tube support plates. Courtesy of J.P. Crosson, Lucius Pitkin, Inc.

Immersion ultrasonic inspection is used on tubes ranging in diameter from 6.4 to 254 mm (

to 10 in.), with wall thicknesses as small as 0.25 mm (0.01 in.). In some installations, four channels are used, two for the detection of transverse flaws and two for longitudinal flaws. Because these tests are usually critical, they are performed at low speeds, usually less than 3 m/min (10 sfm). The liquid penetrant inspection of nonferrous tubular products is performed in the conventional manner, as

described in the article "Liquid Penetrant Inspection" in this Volume.

Note cited in this section

* Examples 3, 4, 5, and 6 were prepared by J.P. Crosson, Lucius Pitkin, Inc. Nondestructive Inspection of Tubular Products

In-Service Inspection** The demand for the in-service inspection of tubular products has been increasing in many industries, including the petroleum, chemical, nuclear, and steelmaking industries. When there is access to the outer surface of the tubing, several of the conventional inspection methods can be used, but when the outside surface is inaccessible, the problem is more complex because the test equipment must pass through the tube. This requirement sharply limits the inspection methods and equipment that can be used. Tubular Products in Commercial Applications Equipment used to inspect from the inside surface must use transducers that move with respect to the shape of the testpiece being inspected. Even with this limitation, the variety of eddy current, flux leakage, and ultrasonic devices available is large enough that commercial equipment can usually be found that is capable of meeting all requirements. Both eddy current equipment and flux leakage equipment have been successful in applications that require passage through the tube. In one application employing eddy current equipment, a combination exciting and detector probe is rotated as it passes through the tube of a reactor heat exchanger. The equipment has high sensitivity and operates at high speed (Ref 19). If the tubes are made of a nominally nonmagnetic metal that is slightly and variably magnetic, such as a Monel alloy or an austenitic stainless steel, it is necessary to magnetically saturate the tubes, using either a direct current field or a permanent magnet. A combination of the three units--inductor, detector, and saturator--built into a small probe has been successfully used to detect small flaws in the 13 mm ( power plant steam generator (Ref 21).

in.) diam by 1.24 mm (0.049 in.) wall tubes of a nuclear

Acoustic emission is a promising technique for the nondestructive inspection of buried pipe. Holes are drilled through the soil to the pipe at intervals of perhaps 120 m (400 ft) to allow sensors to be placed on the pipe. A heavy truck, driven over the area in which the pipe is buried, produces simultaneous indications on the instrumentation connected to each sensor. Analysis of the indication determines the presence or absence of cracks. Examples 7, 8, 9, 10 show typical uses of the ultrasonic inspection and eddy current inspection of tubular products. Example 7: Intergranular Attack and Root Cracks in Welded Austenitic Stainless Steel Tubing Detected Using Contact Shear-Wave Ultrasound Techniques. Contact shear-wave inspection techniques are very effective for detecting intergranular attack and/or root weld cracking in stainless steel tubing. These techniques were developed to overcome the limitations of inspecting formed and installed tubing in aircraft or other areas having limited accessibility.

This example demonstrates techniques applied to 9.5 mm ( in.) outside diameter stainless steel tubing with a 0.51 mm (0.020 in.) wall thickness. However, these techniques can be modified to enable detection of intergranular attack and/or root weld cracking in various materials and sizes. A 15-MHz Rayleigh (surface) wave transducer is machined with a 4.8 mm ( in.) radius to fit the 9.5 mm ( in.) outside diameter of the 0.51 mm (0.020 in.) wall thickness tubing to be inspected (Fig. 13a). The transducer is then positioned on the tubing, as shown in Fig. 13(b). No prior preparation of the sample was required. Mixed-mode shear wave is induced in the tubing to detect intergranular attack (Fig. 14) using a ring-pattern CRT display (Fig. 15). This transducer is also very sensitive and can be used for detecting root weld cracks (Fig. 16). The ultrasound instrument will be set up for monitoring discrete echoes from the root crack. The display produced is shown in Fig. 17.

Fig. 13 15-MHz Rayleigh surface wave transducer (90° shear) used for detecting intergranular attack and root weld cracks. (a) Transducer with radius machined on transducer shoe to allow device to conform to tubing outside diameter. (b) Transducer positioned on tube outside diameter to couple to tube using a lightweight oil couplant. Source: L.D. Cox, General Dynamics Corporation

Fig. 14 Intergranular attack of 0.51 mm (0.020 in.) wall thickness, Fe-21Cr-6Ni-9Mn stainless steel tubing inside diameter. (a) 60×. (b) 85×. Courtesy of L.D. Cox, General Dynamics Corporation

Fig. 15 Mixed-mode shear wave used to detect intergranular attack showing oscilloscope screen display for (a) transducer in air, (b) transducer coupled to an acceptable tube having no defects due to intergranular attack, and (c) transducer coupled to tube rejected because of intergranular attack. Significant attenuation of the ultrasonic signal in (c) is due to scatter. Source: L.D. Cox, General Dynamics Corporation

Fig. 16 Cross section of a tube having a crack at the root of the weld seam. Source: L. D. Cox, General Dynamics Corporation

Fig. 17 Plots obtained on oscilloscope screen with ultrasonic device set up to monitor discrete echoes from a root crack: (a) transducer in air; (b) transducer coupled to tube devoid of root crack defects; and (c) transducer coupled to tube rejected due to presence of a root crack. Signal A in (b) and (c) is due to reflection from the transducer/tube OD contact point. Signal B in (c) is the root crack signal [when the transducer is indexed circumferentially, the A signal will be stationary (no change in time-of-flight) while the B signal will shift]. Source: L.D. Cox, General Dynamics Corporation

Example 8: Eddy Current Inspection of Pitting and Stress-Corrosion Cracking of Type 316 Stainless Steel Evaporator Tubes in a Chemical Processing Operation. Eddy current inspection was performed on a vertical evaporator unit used in a chemical processing plant. The evaporator contained 180 tubes 25 mm (1 in.) in diameter. It was advised that the tube material was type 316 stainless steel. The shell side fluid was condensate and gaseous methylene chloride, while the tube side fluid was contaminated liquid methylene chloride. Eddy current inspection revealed 101 tubes that exhibited severe outer surface pitting and cracklike indications near each tube sheet. Several tubes exhibiting strong indications were pulled and examined visually and metallurgically. It was observed that the indications correlated with rust-stained, pitted, and cracked areas on the outer surfaces. The observed condition was most severe along the portions of the tubes located between the upper tube support and top tube sheet. Figures 18(a) and 18(b) show a pitted and cracked area before and after dye-penetration application.

Fig. 18 Pitting and stress corrosion in type 316 stainless steel evaporator tubes. (a) Rust-stained and pitted area near the top of the evaporator tube. Not clear in the photograph, but visually discernible, are myriads of fine, irregular cracks. (b) Same area shown in (a) but after dye-penetrant application to delineate the extensive fine cracks associated with the rust-stained, pitted surface. (c) Numerous multibranched, transgranular stresscorrosion cracks initiating from the outer surface pits. 35×. Courtesy of J.P. Crosson, Lucius Pitkin, Inc.

Metallographic examination revealed that the cracking initiated from the outer surface, frequently at pits, and penetrated the tube wall in a transgranular, branching fashion. The crack features were characteristic of chloride stress-corrosion

cracking. In many cases, the cracking, rather than penetrating straight through the tube wall, veered off in a tangential direction at or about mid-wall, suggesting the possibility of a change in the residual stress-field from tube drawing. Figure 18(c) shows stress-corrosion cracking originating from pits on the outer surface of the tube. The results of the examination indicated that the subject tube failures occurred by way of stress-corrosion cracking as a result of exposure to a wet-chloride-containing environment. Therefore, a change in tube material was recommended to avoid future failures and loss of service. Example 9: Eddy Current Inspection of a Pitted Type 316 Stainless Steel Condenser Tube. Eddy current inspection was performed on approximately 200 stainless steel tubes in a main condenser unit aboard a container ship. The stainless steel tubes comprised the upper two tube rows in the condenser. The tube material was reported to be type 316 stainless steel; this was confirmed by subsequent chemical analysis. The remaining tubes were 90Cu-10Ni. Recurring leaks had occurred in the stainless steel tubes, but no leaks had occurred in the copper-nickel tubes. Eddy current indications typical of inner surface pitting were observed in 75% of the stainless steel tubes inspected. A tube exhibiting a strong indication was pulled from the condenser and examined visually and metallographically. Visual examination of the outer surface revealed occasional patches of rust-colored deposit at the locations of the eddy current indications. No apparent defects of any type were observed on the outer surface. Subsequent splitting of the tube revealed several areas of severe pitting corrosion attack on the inner surface at locations corresponding to the eddy current indications. The corrosion progressed in such a way as to hollow out the wall thickness, and at several locations the pits had completely penetrated the wall thickness. The pitting corrosion attack tended to be close to the bottom of the tube and essentially in line along the tube sample length. Figure 19(a) shows a severely pitted location. Metallographic examination revealed the attack to be broad and transgranular in nature without any corrosion product build-up at or around the pits. Figure 19(b) shows the manner in which the pitting had penetrated into and beneath the inner surface.

Fig. 19 Pitted type 316 stainless steel condenser tube. (a) Inner surface of main condenser tube showing extensive but localized pitting corrosion attack. 1×. (b) Longitudinal section passing through a pitted area showing extensive pitting that had progressed beneath the inner surface of the main condenser tube. 55×. Courtesy of J.P. Crosson, Lucius Pitkin, Inc.

The results of the examination revealed that the subject stainless steel condenser tube had failed as a result of pitting corrosion attack, which initiated at the inner surface and progressed through the tube wall. That the pitting was essentially on the bottom of the tubes was strong evidence of deposit-type pitting corrosion attack. Deposit attack occurs when foreign material carried by the tube side fluid settles or deposits on the inner surface, generally at the bottom of the tube. The deposit shields the tube surface, creating a stagnant condition in which the fluid beneath the deposit becomes deficient in oxygen compared to the free-flowing fluid around the deposit. The difference in

oxygen content results in the formation of an oxygen concentration cell in which the smaller, oxygen-deficient sites become anodic with respect to the larger oxygenated cathodic sites. As a result, pitting corrosion attack occurs at the anodic sites. In stainless steel, the condition is further aggravated by the fact that type 316 stainless steel performs best in a service where the fluid is oxidizing and forms a passive film on the surface of the tube. If there is an interruption in the film, as may be caused by chemical breakdown through decomposition of organic materials or mechanically by abrasion, and if the damage film is not reformed, pitting corrosion will initiate and grow at the damaged site. In main condenser service, certain deposits, such as shells, sand, or decomposing sea life, can initiate breakdown of the passive film. Example 10: Eddy Current Inspection of a Magnetic Deposit Located on a Steel Tube at Tube Sheet Joint in a Centrifugal Air-Conditioning Unit. In this case, defective tubes were not detected. However, the results of the eddy current inspection were directly influenced by a previous tube failure in the unit. Eddy current inspection of the condenser bundle of a centrifugal air-conditioning unit revealed several tubes with indications typical of tube-wall wear at locations corresponding to the tube supports. One of the tubes exhibiting indications was pulled and visually examined. A tightly adherent magnetic deposit was observed at the area of the tube in contact with the first tube support. Splitting the tube revealed the deposit to be tightly packed in the fins, as shown in Fig. 20. This tube was of tru-finned rather than skip-finned design; that is, the tube did not have smooth support saddles where it was in contact with the tube support plate. Instead, the tube was finned from end to end. Therefore, although the test instrument parameters were selected to phase out the magnetically induced indications from the steel tube supports, the magnetic deposit, which was tightly embedded between the fins, was closer to the internal test probe and caused an indication that was interpreted as tube-wall wear.

Fig. 20 Magnetic corrosion product embedded in the tube fins at the tube support of a steel tube. The corrosion product caused by eddy current indication characteristic of tube-wall wear at the support. Courtesy of J.P. Crosson, Lucius Pitkin, Inc.

Through further investigation it was determined that a previous tube failure had caused the Freon on the shell side to become contaminated with water. This condition proved corrosive to the steel supports and shell and subsequently caused the magnetic corrosion deposit observed at the tube support. Oil-Country Tubular Products The application of nondestructive inspection to the tubular products of the oil and gas distribution industry is extensive and is vital to successful operation. The American Petroleum Institute has, with international cooperation and international acceptance, developed tubing and pipe specifications that include many rigorous requirements for nondestructive inspection (Ref 22). Inspection installations range from simple magnetic particle installations to complex assemblies of machinery whose continuous productivity is completely dependent on the reliability and accuracy of its nondestructive inspection equipment. The larger installations may use ultrasonic, eddy current, flux leakage, or radiographic equipment, singly or in combination, and can be supplemented by magnetic particle inspection.

The inspections of pipe or casing can be performed during manufacture, when it is received on site, while it is in service, or when it must be inspected for reuse or resale. When inspection is included in the manufacturing operation, tests are usually performed immediately after the pipe is produced and again after processing has been completed. Industry has promoted the development and use of highly sophisticated equipment for the in-service inspection of pipe in diameters of 75 mm (3 in.) and larger. In one of several different pipe crawlers available commercially, the probe travels through the gas lines and, by means of flux leakage measurements, reports on the condition of the pipe. Another type of in-service inspection unit, which is shown in Fig. 21, includes tight-fitting seals so that it can be propelled through the pipelines by the oil or gas being carried. The traveling unit includes not only the test instrumentation and a tape recorder but also a power supply so that it is completely self-sufficient, requiring no connection outside the pipe. The sections of this unit are connected by universal joints to permit passage around bends. When the unit completes its cycle, total information on the condition of the pipe is immediately available.

Fig. 21 Self-contained flux leakage inspection unit used in oil and gas pipeline for in-service inspection

One of the most important inspection procedures in this industry involves the inspection of girth welds joining the ends of pipes to each other or to fittings and bends. Although radiographic tests are widely favored for this application (Ref 23), supplementary tests are needed to detect the tightly closed flaws not detected by radiography (Ref 12). This industry also uses automated inspection of small tubular pipe couplings. One machine separates acceptable couplings from rejectable couplings automatically and requires no operator. The couplings are fed into the machine from a cutoff lathe. After automatic inspection to API specifications, rejectable couplings are diverted to a reject receptacle. Nondestructive Inspection of Steel Pipelines (Ref 24) The nondestructive inspection of welds in steel pipe is used to eliminate discontinuities that could cause failure or leakage. Most steel pipes for gas transmission are made by the hot rotary forging of pierced billets or by forming plate or strip and then welding by either the submerged arc or the resistance process. Pipes are usually made to one of the API specifications, with supplementary requirements if necessary. Submerged Arc Welded Pipe. The shrinkage of liquid metal upon solidification results in primary piping in the ingot, which can cause laminations oriented in the plane of the plate or strip rolled from the ingot. Laminations can also result from secondary piping and from large inclusions. Laminations can nucleate discontinuities during welding or propagate to form a split in the weld. They cannot be detected by radiography, because of their orientation, but they can be detected by ultrasonics or can be seen when they occur as skin laminations.

The API specifications do not require nondestructive inspection of the plate or strip before welding, but ultrasonic inspection is mandatory in some customer requirements. Generally, the periphery of the plates must be examined to ensure that edges to be welded are free of laminations.

Pulse-echo ultrasonic inspection has been used for most plate-inspection specifications. This method cannot distinguish laminations near the surface remote from the probe, because the echoes from the laminations cannot be inspected from the back echo. Some specifications require that the plate be inspected from both surfaces or that transmission methods be used. Manual scanning, although time consuming, is feasible because the echo pattern from laminations persists on the oscilloscope screen and is easy to interpret. However, if laminations are fragmented or at an oblique angle to the surface of the plate, there is no distinct flaw echo, but merely a loss of back echo. In most pulse-echo equipment, no account is taken of this loss; therefore, transmission methods are preferable for plate inspection. Such methods normally require mechanization with automatic recording of the results, and inspection systems based on these methods have been installed in some plate mills. In both transmission and pulse-echo inspection, the probe area represents the area of the plate under inspection at any instant. Shear-wave angle probes are used to detect lamination, but the method is not reliable, particularly for the thin plates used in pipe manufacture. Laminations can also be detected by ultrasonic Lamb waves, which can inspect a zone extending across some or all of the plate. Lamb waves have been used on steel sheet, but they cannot be excited in plates more than 6.4 mm ( in.) thick.

in.) thick using standard equipment. Special equipment is now available for plate up to 13 mm (

The main flaws that occur in submerged arc welds are incomplete fusion and incomplete penetration between the inside and outside weld beads or between the base metal and the filler metal, cracks, undercut or underfill, and overflow. The API standards require full-length inspection of welds by radiography or ultrasonics. Fluorescent screens or television screens are permitted for radiography, and they are often used because they are less expensive than radiographic film, although less discriminating. Fluoroscopy is inherently less sensitive to the more critical flaws, cracks, incomplete sidewall fusion, and incomplete penetration. Ultrasonic inspection is more sensitive to serious flaws and can be automated. The arrangements of transmitter and receiver probes in ultrasonic inspection of submerged arc welded pipe for detection of longitudinally and transversely oriented discontinuities are shown in Fig. 22. The region of the oscilloscope time base corresponding to the weld region is analyzed electronically, and echoes above the amplitude of the reference derived from the calibration block actuate a relay that can operate visible or audible warnings, paint sprays, or pen recorders. In some installations, only one probe is used on each side of the weld, and detection of discontinuities that are oriented transversely to the weld is not possible. To ensure correct lateral positioning of the probes on the weld, they are mounted on a frame, which is then moved along the weld; alternatively, the pipe can be moved past stationary probes.

Fig. 22 Diagram of arrangements of probes in the ultrasonic inspection of submerged arc welded pipe for the detection of (a) longitudinally oriented and (b) transversely oriented discontinuities

Accurate positioning of the probes over the welds is difficult because the width, shape, and straightness of the weld bead vary. The inspection area is limited in order to reduce confusion between echoes from flaws within the weld and the boundaries of the weld reinforcement. Acoustic coupling can be reduced by the probes riding up on weld spatter, by drifting of the scanning frame, by loss of coupling water, or by loose mill scale. General practice is to use automatic ultrasonic inspection methods and to radiograph those regions of the pipe suspected of containing discontinuities. If radiography does not reveal an objectionable flaw, the ultrasonic indication is ignored and the pipe is accepted. This procedure would accept cracks or laminations parallel to the plate surface that, because of their orientation, cannot be detected by radiography. As an alternative approach, regions that give an ultrasonic flaw indication should be inspected radiographically and by manual ultrasonics. If the original ultrasonic indication was from a discontinuity shown by the radiograph to be acceptable within the specification or if the manual ultrasonic inspection revealed that the indication was a spurious echo arising from a surface wave or from local weld shape, then the pipe was accepted; if the radiograph showed an objectionable flaw, then the pipe was rejected. If there was no explanation for the echo, it was assumed to have arisen from a discontinuity adversely oriented for radiography. Seamless Pipe. There are two sources of flaws in roll forged seamless pipe: inhomogeneities and the manufacturing

process. Inhomogeneities in the ingot such as primary and secondary ingot pipe can be carried into the roll-forged product and can cause flaws in a similar manner to the formation of laminations in steel plate. Such flaws are likely to have a

major dimension oriented in the plane of the pipe wall. In manufacturing, the rolls and the mandrel can cause surface discontinuities such as tears and laps, and such discontinuities will have substantial orientation normal to the pipe wall. In addition, pipes and tubes produced by working pierced billets are prone to eccentric wall thickness, with the eccentricity varying along the length of the pipe. The API specifications that cover seamless line pipe require neither nondestructive inspection nor wall thickness measurement away from the pipe ends. In some mills, destructive inspections are carried out on samples cut from each pipe end to determine the presence of primary pipe flaws. In some API standards (casing, tubing, and drill pipe), nondestructive inspection is optional, but in other standards (high-strength casing and tubing) nondestructive inspection of the full pipe length is mandatory. Magnetic particle, ultrasonic, or eddy current inspection methods are permitted. Magnetic particle inspection methods have little or no sensitivity to discontinuities that do not show on the surface and are likely to detect laminar discontinuities resulting from ingot piping. Although surface laps are amenable to magnetic crack detection, it would be difficult to apply the inspection method to internal-surface discontinuities. Eddy current inspection methods can be used to inspect seamless tubing. Very rapid inspection rates are possible

with the encircling-coil system. When pipe is passed through a coil fed with alternating current, the resistive and reactive components of the coil are modified; the modification depends on dimensions (and therefore indirectly on discontinuities), electrical conductivity and magnetic permeability, and the annulus between the pipe and the coil (and therefore the outside diameter of the pipe). The analysis to determine which effect is causing any modification is complex. Eddy current methods are extensively used for the inspection of small, nonferrous tubes, but ferrous material causes complications from magnetic permeability. The initial permeability is affected by residual-stress level. Roll-forged pipe may have varying amounts of residual cold work, depending on the original soaking conditions and the time taken to complete forging. The effect can be alleviated by applying a magnetically saturated field; equipment that can produce a magnetically saturated field has been installed in steel tube mills. However, saturation becomes more difficult as pipe diameter increases. Radiographic inspection methods, employing either x-ray or γ-ray transmission, can be used with a scintillation

counter to estimate the wall thickness of pipe. The accuracy of scintillation counters depends on the size of the count for a given increment of thickness; the count increases with the time the increment is in the beam. As a result, the count, and therefore the accuracy, increases with decreasing scanning rate. When large-diameter pipes are scanned at realistic rates, eccentricity is usually averaged out. Ultrasonic inspection methods can detect discontinuities oriented both in the plane of, and normal to, the pipe wall.

Discontinuities in the plane of the wall can be detected by using a compression-wave probe scanning at normal incidence. For discontinuities normal to the wall, the beam is converted to shear wave and propagated around or along the tube. The pipe is rotated and moved longitudinally relative to the probes, thus giving a helical scan. The reliability of mechanized scanning is a function of acoustic coupling, and optimum results are achieved with immersion coupling. The efficiency of acoustic coupling through large columns of water is lower but much more consistent than that through the thin liquid films used in contact scanning. Immersion methods also eliminate probe wear and the requirement for specially contoured probes to accommodate each pipe size. Alternatively, immersion coupling by a column of water flowing between the probe and the pipe can be used. With this method, probe-rotation scanning is possible. Advantage can be taken of the smaller inertia of the probes to increase the scanning rate, and therefore the speed of inspection, by about an order of magnitude. When an ultrasonic beam propagates radially through the pipe wall, the time interval between successive back echoes reflected from the bore surface is directly proportional to the wall thickness. If the first back echo is used to trigger a highspeed electronic counter whose frequency is such that it will produce a count of 100 during the time taken to receive four echoes in 25 mm (1 in.) thick plate and if a subsequent back echo is used to stop the counter, a count proportional to the wall thickness is produced. By changing the frequency of the counter oscillator, it is possible to change the thickness range inspected or to accommodate different materials. Information from the counter can be fed to a chart recorder, thus continuously recording the wall thickness. Lamination would be recorded as an abrupt localized reduction in wall thickness.

For the detection of cracks and laplike discontinuities, the display on the flaw detection oscilloscope is gated. The discontinuity can then be recorded in its position around the circumference. Chart length can be made proportional to pipe length, thus facilitating discontinuity location and extent in relation to pipe length and variation in wall thickness around and along the pipe. Alternatively, information can be monitored in a go/no-go method to provide a paint spray that identifies the locations of significant discontinuities. Resistance-Welded Pipe. The type of discontinuity usually responsible for the failure of resistance-welded pipe is incomplete fusion, with associated oxide film. The nondestructive inspection of larger-diameter resistance-welded pipes is normally restricted to inspection of the weld region. Systems similar to those used for the inspection of submerged arc welded seams can be employed, although the arrangements for tracking the probes with respect to the weld reinforcements are not applicable. Because of problems associated with accurate weld tracking, it is necessary that small variations in weld-probe separation should only cause acceptably small variations in discontinuity detection sensitivity.

Probe angles of 60 to 65° give satisfactory results, and coverage of the weld depth is achieved by using two probes on each side of the weld. Such a system is also relevant to the ultrasonic inspection of submerged arc welded pipes. Nondestructive Inspection of Pipeline Girth Welds The API specifications do not require that all girth welds be inspected; the use of radiography and the extent of coverage are optional. Generally, it has been the practice to inspect 10% of the weld length. Where the integrity of a pipeline is vital, as in high-pressure gas-transmission systems, it is advisable to consider a 100% inspection, especially when inspection is not a large proportion of the total cost of the pipeline. Where operating conditions are less critical, however, it may be possible to be less critical regarding the size and type of discontinuity permissible. Radiographic Inspection. Characteristic discontinuities in pipeline welds are slag, elongated piping in root, scattered

piping and porosity, burn-throughs in the root, incomplete root penetration, incomplete sidewall fusion, and cracks, which often break the inner surface in the heat-affected zone. Except for cracks and incomplete sidewall fusion, these discontinuities are amenable to detection by radiography. Open cracks can be detected, but tighter cracks, even though favorably oriented, are detectable only by optimum practice. Some cracks may not be revealed at all. Assuming good radiographic techniques, radiographic quality depends on the choice of conditions that control the contrast and definition of the radiograph; the detection of discontinuities improves with increasing contrast and fine definition. Contrast can be assessed in terms of the thickness sensitivity, which can be conveniently estimated by imagequality indicators. Many factors other than good radiographic techniques influence radiographic contrast. For pipeline radiography, radiation energy is probably the most important. The absorption of radiation by steel decreases with increasing energy; the absorption coefficient at 150 kV is about three times that at 700 kV. For optimum detection of small changes in thickness, the absorption should be as high as possible so that large differences in exposure, consistent with a reasonable amount of energy being transmitted to provide a realistic overall exposure, result at the film. Gamma radiation from a 192Ir source is approximately equivalent to x-rays generated at 700 kV and therefore will not be absorbed sufficiently to give good contrast sensitivity. For wall thicknesses typical of pipelines, x-rays generated at about 150 to 175 kV have reasonable absorption. For the detection and correct identification of discontinuities from the radiographic image, the delineation of the shadow must be sharp. The principal sources of unsharpness in radiographic images are geometric unsharpness, resulting front the finite size of radiation sources; unsharpness in the film resulting from the kinetic energy of the radiation, grain size of the emulsion and degree of development; and unsharpness resulting from the intensifying screen. Radiographs on pipelines are generally made under less-than-satisfactory conditions; nevertheless, it should be possible to avoid vibration and other forms of relative movement of the source and film during exposure so that the total geometric unsharpness is the penumbra effect. With piping, geometric unsharpness is not large. An important side effect of unsharpness is that when the unsharpness is greater than the width of a flaw, the contrast resulting from the flaw is reduced from its theoretical value; the greater the unsharpness, the greater the contrast reduction. For tight cracks, the contrast may be so reduced that the change in tone is below the threshold for detection. On-site conditions may reduce the capability of radiography to detect flaws. Under some conditions, it is difficult to maintain correct developer temperature. The operator is often pressured to keep pace with the welding crews; also,

weather and working conditions may be adverse. Suitable equipment and adequate planning should overcome these problems. The more difficult problem is the repetitiveness of the procedure, which causes the operators to lose concentration and gradually to devote less attention to detail. For most pipe sizes it is necessary to make three exposures to cover the circumference of the weld because the length of weld that can be covered in one exposure (the diagnostic film length) is limited by fade at each end of the film. Panoramic techniques have been used on some larger-diameter pipes. The source is held in a spider arrangement and positioned on the pipe axis; the films are placed around the outer surface of the pipe at the weld. In this manner, the entire weld can be radiographed in one exposure. This exposure is shorter than one of the exposures required in the double-wall, singleimage technique because the radiation has to propagate through only one wall of the pipe. In practice, the radiation source can be manually positioned only inside pipes having a diameter of 762 mm (30 in.) or more; even then, conditions must be good. Crawler devices are available that are mechanically propelled through the pipe, with the exposure being operated from an external control (see the section "In-Motion Radiography" of the article "Radiographic Inspection" in this Volume). On pipelines where the rate of welding is low and the investment on crawlers is not justified, x-ray sets can be clamped onto the outside of the pipe and radiography implemented by the double-wall, single-image technique. Gamma radiography has been favored for pipeline radiography because of its convenience and lower cost. Source containers are more compact and portable than x-ray generators and do not require a power supply. Panoramic x-ray radiography is barely feasible without crawler devices because of the difficulty of manually maneuvering the cumbersome x-ray sets and control units inside a pipe. Because of the potential for increased use of xray radiography on pipelines, there has been considerable effort applied to development of x-ray crawlers. Ultrasonic Inspection. Welds are usually ultrasonically inspected by a pulse-echo reflection technique. Before the inspection of a weld, the pipe should be checked for laminations that may divert the beam from its theoretical path.

Discontinuities can be identified most reliably by accurate positioning of the source of the discontinuity echo, preferably during scanning from more than one direction. Skilled operators may be able to gain additional information on type of discontinuity from the shape of the echo on the oscilloscope screen, but the display on battery-operated flaw detectors used in daylight is not sufficiently distinct for the technique to be employed on pipelines. The more significant discontinuities occur in the root of the weld, where discrimination between sources of echo reflection is more difficult. Acceptable features such as full root penetration cause echoes comparable in magnitude to those from root underbead cracks or incomplete penetration. Accurate positioning of the probe with respect to the weld centerline is necessary, and it has been suggested that the required accuracy can be achieved only by marking and machining the pipe ends before welding. Positioning from the center of the weld cap is only approximate because the weld is not necessarily symmetrical about the centerline through the root. Even with the premarking, it is difficult for an operator to locate the ultrasonic probe accurately and still be in a position to view the instrument screen. Thinner-wall pipes reduce the differences in beam path and probe position for discrimination between the various features of the root. Also, the weld must be examined with the probe-to-weld distance increased to avoid confusion between echoes from the weld and those from probe noise. This increases the effect of beam spread and may lead to extraneous echoes from the cap reinforcement. The skip distance and beam-path length vary as the wall thickness varies. Variations in wall thickness between nominally the same classes of submerged arc welded pipe range from 10 to 15%, but in seamless pipe a ±10% variation along the length or around the circumference at a given position along the length is common. Although it is possible to measure wall thickness accurately by ultrasonics, it is not feasible to measure wall thickness concurrently with scanning the weld. Surface roughness can cause considerable variations in beam angle. Weld spatter can reduce the effectiveness of the coupling, and also alter beam angle by lifting part of the probe off the pipe. Ultrasonic inspection on girth welds was originally used to determine which welds to radiograph. If the radiograph did not detect anything, it was the practice on most pipelines to accept the radiographic evidence and not that from ultrasonics. Now that pipelines are being examined 100% by radiography, the role of ultrasonics has changed to that of detecting root underbead cracks that may escape detection by radiography and of providing supplementary evidence to aid in the interpretation of radiographic images of weld-root regions.

Surface Crack Detection. Root underbead cracks break the surface of the pipe in the bore and can be detected with

liquid penetrant and magnetic particle inspection. The weld area can be magnetized using a yoke powered by permanent magnets. Both methods are sensitive under ideal conditions, but liquid penetrants require very clean surfaces. Magnetic particle crack detection is therefore preferred for pipeline applications. Interpretation of the indications is not a problem, except for the confusion that may arise from the tendency of sharp changes in root profile to give a slight crack indication.

References cited in this section

12. R.F. Lumb and G.D. Fearnebaugh, Toward Better Standards for Field Welding of Gas Pipelines, Weld. J., Vol 54 (No. 2), Feb 1975, p 63-s to 71-s 19. F. Förster, Sensitive Eddy-Current Testing of Tubes for Defects on the Inner and Outer Surfaces, NonDestr. Test., Vol 7 (No. 1), Feb 1974, p 25-35 21. V.S. Cecco and C.R. Bax, Eddy Current In-Situ Inspection of Ferromagnetic Monel Tubes, Mater. Eval., Vol 33 (No. 1), Jan 1975, p 1-4 22. "Specification for Line Pipe," API 5L, American Petroleum Institute, 1973 23. "Standard for Welding Pipe Lines and Related Facilities," API 1104, American Petroleum Institute, 1968 24. R.F. Lumb, Non-Destructive Testing of High-Pressure Gas Pipelines, Non-Destr. Test., Vol 2 (No. 4), Nov 1969, p 259-268 Note cited in this section

** Example 7 was prepared by L.D. Cox, General Dynamics Corporation. Examples 8, 9, and 10 were prepared by J.P. Crosson, Lucius Pitkin, Inc. Nondestructive Inspection of Tubular Products

References 1. 2. 3.

"Nondestructive Testing Terminology," Bulletin 5T1, American Petroleum Institute, 1974 H.C. Knerr and C. Farrow, Method and Apparatus for Testing Metal Articles, U.S. Patent 2,065,379, 1932 W.C. Harmon, "Automatic Production Testing of Electric Resistance Welded Steel Pipe," Paper presented at the ASNT Convention, New York, American Society for Nondestructive Testing, Nov 1962 4. W.C. Harmon and I.G. Orellana, Seam Depth Indicator, U.S. Patent 2,660,704, 1949 5. J.P. Vild, "A Quadraprobe Eddy Current Tester for Tubing and Pipe," Paper presented at the ASNT Convention, Cleveland, American Society for Nondestructive Testing, Oct 1970 6. H. Luz, Die Segmentspule--ein neuer Geber für die Wirbelstromprüfung von Rohren, BänderBlecheRohre, Vol 12 (No. 1), Jan 1971 7. W. Stumm, Tube-Testing by Electromagnetic NDT (Non-Destructive Testing) Methods: I, Non-Destr. Test., Vol 7 (No. 5), Oct 1974, p 251-258 8. F. Förster, The Nondestructive Inspection of Tubings for Discontinuities and Wall Thickness Using Electromagnetic Test Methods: I, Mater. Eval., Vol 28 (No. 4), April 1970, p 21A-25A, 28A-31A 9. F. Förster, The Nondestructive Inspection of Tubings for Discontinuities and Wall Thickness Using Electromagnetic Test Methods: II, Mater. Eval., Vol 28 (No. 5), May 1970, p 19A-23A, 26A-28A 10. P.J. Bebick, "Locating Internal and Inside Diameter Defects in Heavy Wall Ferromagnetic Tubing by the Leakage Flux Inspection Method," Paper presented at the ASNT Convention, Cleveland, American Society for Nondestructive Testing, Oct 1974 11. H.J. Ridder, "New Nondestructive Technology Applied to the Testing of Pipe Welds," Paper presented at the ASME Petroleum Conference, New Orleans, American Society of Mechanical Engineers, Sept 1972 12. R.F. Lumb and G.D. Fearnebaugh, Toward Better Standards for Field Welding of Gas Pipelines, Weld. J.,

13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.

Vol 54 (No. 2), Feb 1975, p 63-s to 71-s M.J. May, J.A. Dick, and E.F. Walker, "The Significance and Assessment of Defects in Pipeline Steels," British Steel Corporation, June 1972 W.C. Harmon and T.W. Judd, Ultrasonic Test System for Longitudinal Fusion Welds in Pipe, Mater. Eval., March 1974, p 45-49 "Inspection, Radiographic," Military Standard 453A, May 1962 W. Stumm, New Developments in the Eddy Current Testing of Hot Wires and Hot Tubes, Mater. Eval., Vol 29 (No. 7), July 1971, p 141-147 F.J. Barchfeld, R.S. Spinetti, and J.F. Winston, "Automatic In-Line Inspection of Seamless Pipe," Paper presented at the ASNT Convention, Detroit, American Society for Nondestructive Testing, Oct 1974 T.W. Judd, Orbitest for Round Tubes, Mater. Eval., Vol 28 (No. 1), Jan 1970, p 8-12 F. Förster, Sensitive Eddy-Current Testing of Tubes for Defects on the Inner and Outer Surfaces, NonDestr. Test., Vol 7 (No. 1), Feb 1974, p 25-35 K.J. Reimann, T.H. Busse, R.B. Massow, and A. Sather, Inspection Feasibility of Duplex Tubes, Mater. Eval., Vol 33 (No. 4), April 1975, p 89-95 V.S. Cecco and C.R. Bax, Eddy Current In-Situ Inspection of Ferromagnetic Monel Tubes, Mater. Eval., Vol 33 (No. 1), Jan 1975, p 1-4 "Specification for Line Pipe," API 5L, American Petroleum Institute, 1973 "Standard for Welding Pipe Lines and Related Facilities," API 1104, American Petroleum Institute, 1968 R.F. Lumb, Non-Destructive Testing of High-Pressure Gas Pipelines, Non-Destr. Test., Vol 2 (No. 4), Nov 1969, p 259-268

Nondestructive Inspection of Weldments, Brazed Assemblies, and Soldered Joints

Introduction THE SELECTION of a method for inspecting weldments, brazed assemblies, and soldered joints for flaws (referred to as discontinuities in welding terminology) depends on a number of variables, including the nature of the discontinuity, the accessibility of the joint, the type of materials joined, the number of joints to be inspected, the detection capabilities of the inspection method, the level of joint quality required, and economic considerations. Regardless of the method selected, established standards must be followed to obtain valid inspection results. In general, nondestructive inspection methods (NDI) are preferred over destructive inspection methods. Sections can be trepanned from a joint to determine its integrity; however, the joint must be refilled, and there is no certainty that discontinuities would not be introduced during repair. Destructive inspection is usually impractical, because of the high cost and the inability of such methods to accurately predict the quality of those joints that were not inspected. This article will review nondestructive methods of inspection for weldments (including diffusion-bonded joints) and brazed and soldered joints. More detailed information on the techniques discussed can be found in the Sections "Inspection Equipment and Techniques," and "Methods of Nondestructive Evaluation" in this Volume.

Acknowledgements The contributions of the following individuals were critical in the preparation of this article: W.H. Kennedy, Canadian Welding Bureau; Robert S. Gilmore, General Electric Research and Development Center; and John M. St. John, Caterpillar, Inc. Special thanks are also due to Michael Jenemann, Product Manager, NDT Systems, E.I. Du Pont de Nemours & Company, Inc., for supplying the reference radiographs of welds shown in Fig. 18 to 37. Finally, the efforts of the ASM Committee on Weld Discontinuities and the ASM Committee on Soldering from Volume 6 of the 9th Edition of Metals Handbook are gratefully acknowledged; material from the aforementioned Volume was used in this article.

Nondestructive Inspection of Weldments, Brazed Assemblies, and Soldered Joints

Weldments Weldments made by the various welding processes may contain discontinuities that are characteristic of that process. Therefore, each process, as well as the discontinuities typical of that process, are discussed below. Explanations of welding processes, equipment and filler metals, and welding parameters for specific metals and alloys are available in Welding, Brazing, and Soldering, Volume 6 of the ASM Handbook. Discontinuities in Arc Welds Discontinuities may be divided into three broad classifications: design related, welding process related, and metallurgical. Design-related discontinuities include problems with design or structural details, choice of the wrong type of weld joint for a given application, or undesirable changes in cross section. These discontinuities, which are beyond the scope of this article, are discussed in the Section "Joint Evaluation and Quality Control" in Welding, Brazing, and Soldering, Volume 6 of the ASM Handbook. Discontinuities resulting from the welding process include: • • • • • • • • • • • • • • •

Undercut: A groove melted into the base metal adjacent to the toe or root of a weld and left unfilled by weld metal Slag inclusions: Nonmetallic solid material entrapped in weld metal or between weld metal and base metal Porosity: Cavity-type discontinuities formed by gas entrapment during solidification Overlap: The protrusion of weld metal beyond the toe, face, or root of the weld Tungsten inclusions: Particles from tungsten electrodes that result from improper gas tungsten arc welding procedures Backing piece left on: Failure to remove material placed at the root of a weld joint to support molten weld metal Shrinkage voids: Cavity-type discontinuities normally formed by shrinkage during solidification Oxide inclusions: Particles of surface oxides that have not melted and are mixed into the weld metal Lack of fusion (LOF): A condition in which fusion is less than complete Lack of penetration (LOP): A condition in which joint penetration is less than that specified Craters: Depressions at the termination of a weld bead or in the molten weld pool Melt-through: A condition resulting when the arc melts through the bottom of a joint welded from one side Spatter: Metal particles expelled during welding that do not form a part of the weld Arc strikes (arc burns): Discontinuities consisting of any localized remelted metal, heat-affected metal, or change in the surface profile of any part of a weld or base metal resulting from an arc Underfill: A depression on the face of the weld or root surface extending below the surface of the adjacent base metal

Metallurgical discontinuities include: • • •

Cracks: Fracture-type discontinuities characterized by a sharp tip and high ratio of length and width to opening displacement Fissures: Small cracklike discontinuities with only a slight separation (opening displacement) of the fracture surfaces Fisheye: A discontinuity found on the fracture surface of a weld in steel that consists of a small pore or inclusion surrounded by a bright, round area

• •

Segregation: The nonuniform distribution or concentration of impurities or alloying elements that arises during the solidification of the weld Lamellar tearing: A type of cracking that occurs in the base metal or heat-affected zone (HAZ) of restrained weld joints that is the result of inadequate ductility in the through-thickness direction of steel plate

The observed occurrence of discontinuities and their relative amounts depend largely on the welding process used, the inspection method applied, the type of weld made, the joint design and fit-up obtained, the material utilized, and the working and environmental conditions. The most frequent weld discontinuities found during manufacture, ranked in order of decreasing occurrence on the basis of arc-welding processes, are:

Shielded metal arc welding (SMAW) Slag inclusions Porosity LOF/LOP Undercut

Submerged arc welding (SAW) LOF/LOP Slag inclusions Porosity

Flux cored arc welding (FCAW) Slag inclusions Porosity LOF/LOP

Gas metal arc welding (GMAW) Porosity LOF/LOP

Gas tungsten arc welding (GTAW) Porosity

The commonly encountered inclusions--as well as cracking, the most serious of weld defects--will be discussed in this section. Gas porosity can occur on or just below the surface of a weld. Pores are characterized by a rounded or elongated

teardrop shape with or without a sharp point. Pores can be uniformly distributed throughout the weld or isolated in small groups; they can also be concentrated at the root or toe of the weld. Porosity in welds is caused by gas entrapment in the molten metal, by too much moisture on the base or filler metal, or by improper cleaning of the joint during preparation for welding. The type of porosity within a weld is usually designated by the amount and distribution of the pores. Some of the types are classified as follows:

• • • • •

Uniformly scattered porosity: Characterized by pores scattered uniformly throughout the weld (Fig. 1a) Cluster porosity: Characterized by clusters of pores separated by porosity-free areas (Fig. 1b) Linear porosity: Characterized by pores that are linearly distributed (Fig. 1c). Linear porosity generally occurs in the root pass and is associated with incomplete joint penetration Elongated porosity: Characterized by highly elongated pores inclined to the direction of welding. Elongated porosity occurs in a herringbone pattern (Fig. 1a) Wormhole porosity: Characterized by elongated voids with a definite worm-type shape and texture (Fig. 2)

Fig. 1 Type of gas porosity commonly found in weld metal. (a) Uniformly scattered porosity. (b) Cluster porosity. (c) Linear porosity. (d) Elongated porosity

Fig. 2 Wormhole porosity in a weld bead. Longitudinal cut. 20×

Radiography is the most widely used nondestructive method for detecting subsurface gas porosity in weldments. The radiographic image of round porosity appears as round or oval spots with smooth edges, and elongated porosity appears as oval spots with the major axis sometimes several times longer than the minor axis. The radiographic image of wormhole porosity depends largely on the orientation of the elongated cavity with respect to the incident x-ray beam. The presence of top-surface or root reinforcement affects the sensitivity of inspection, and the presence of foreign material, such as loose scale, flux, or weld spatter, may interfere with the interpretation of results. Ultrasonic inspection is capable of detecting subsurface porosity. However, it is not extensively used for this purpose except to inspect thick sections or inaccessible areas where radiographic sensitivity is limited. Surface finish and grain size affect the validity of the inspection results. Eddy current inspection, like ultrasonic inspection, can be used for detecting subsurface porosity. Normally, eddy current inspection is confined to use on thin-wall welded pipe and tubing because eddy currents are relatively insensitive to flaws that do not extend to the surface or into the near-surface layer. Magnetic particle inspection and liquid penetrant inspection are not suitable for detecting subsurface gas porosity. These methods are restricted to the detection of only those pores that are open to the surface. Slag inclusions may occur when using welding processes that employ a slag covering for shielding purposes. (With other processes, the oxide present on the metal surface before welding may also become entrapped.) Slag inclusions can be found near the surface and in the root of a weld (Fig. 3a), between weld beads in multiple-pass welds (Fig. 3b), and at the side of a weld near the root (Fig. 3c).

Fig. 3 Sections showing locations of slag inclusions in weld metal. (a) Near the surface and in the root of a single-pass weld. (b) Between weld beads in a multiple-pass weld. (c) At the side of a weld near the root

During welding, slag may spill ahead of the arc and subsequently be covered by the weld pool because of poor joint fitup, incorrect electrode manipulation, or forward arc blow. Slag trapped in this manner is generally located near the root. Radical motions of the electrode, such as wide weaving, may also cause slag entrapment on the sides or near the top of the weld after the slag spills into a portion of the joint that has not been filled by the molten pool. Incomplete removal of the slag from the previous pass in multiple-pass welding is another common cause of entrapment. In multiple-pass welds, slag may be entrapped any number of places in the weld between passes. Slag inclusions are generally oriented along the direction of welding. Three methods used for the detection of slag below the surface of single-pass or multiple-pass welds are magnetic particle, radiographic, and ultrasonic inspection. Depending on their size, shape, orientation, and proximity to the surface, slag inclusions can be detected by magnetic particle inspection with a dc power source, provided the material is ferromagnetic. Radiography can be used for any material, but is the most expensive of the three methods. Ultrasonic inspection can also be used for any material and is the most reliable and least expensive method. If the weld is machined to a flush contour, flaws as close as 0.8 mm ( in.) to the surface can be detected with the straight-beam technique of ultrasonic inspection, provided the instrument has sufficient sensitivity and resolution. A 5- or 10-MHz dual-element transducer is normally used in this application. If the weld cannot be machined, near-surface sensitivity will be low because the initial pulse is excessively broadened by the rough, as-welded surface. Unmachined welds can be readily inspected by direct-beam and reflected-beam techniques, using an angle-beam (shear-wave) transducer. Tungsten inclusions are particles found in the weld metal from the nonconsumable tungsten electrode used in GTAW. These inclusions are the result of:

• • • • • •

Exceeding the maximum current for a given electrode size or type Letting the tip of the electrode make contact with the molten weld pool Letting the filler metal come in contact with the hot tip of the electrode Using an excessive electrode extension Inadequate gas shielding or excessive wind drafts, which result in oxidation Using improper shielding gases such as argon-oxygen or argon-CO2 mixtures, which are used for GMAW

Tungsten inclusions, which are not acceptable for high-quality work, can only be found by internal inspection techniques, particularly radiographic testing. Lack of fusion and lack of penetration result from improper electrode manipulation and the use of incorrect

welding conditions. Fusion refers to the degree to which the original base metal surfaces to be welded have been fused to the filler metal. On the other hand, penetration refers to the degree to which the base metal has been melted and resolidified to result in a deeper throat than was present in the joint before welding. In effect, a joint can be completely fused but have incomplete root penetration to obtain the throat size specified. Based on these definitions, LOF discontinuities are located on the sidewalls of a joint, and LOP discontinuities are located near the root (Fig. 4). With some joint configurations, such as butt joints, the two terms can be used interchangeably. The causes of LOF include excessive travel speed, bridging, excessive electrode size, insufficient current, poor joint preparation, overly acute joint angle, improper electrode manipulation, and excessive arc blow. Lack of penetration may be the result of low welding current, excessive travel speed, improper electrode manipulation, or surface contaminants such as oxide, oil, or dirt that prevent full melting of the underlying metal.

Fig. 4 Lack of fusion in (a) a single-V-groove weld and (b) double-V-groove weld. Lack of penetration in (c) a single-V-groove and (d) a double-V-groove weld

Radiographic methods may be unable to detect these discontinuities in certain cases, because of the small effect they have on x-ray absorption. As will be described later, however, lack of sidewall fusion is readily detected by radiography. Ultrasonically, both types of discontinuities often appear as severe, almost continuous, linear porosity because of the nature of the unbonded areas of the joint. Except in thin sheet or plate, these discontinuities may be too deep-lying to be detected by magnetic particle inspection. Geometric weld discontinuities are those associated with imperfect shape or unacceptable weld contour. Undercut, underfill, overlap, excessive reinforcement, fillet shape, and melt-through, all of which were defined earlier, are included in this grouping. Geometric discontinuities are shown schematically in Fig. 5. Radiography is used most often to detect these flaws.

Fig. 5 Weld discontinuities affecting weld shape and contour. (a) Undercut and overlapping in a fillet weld. (b) Undercut and overlapping in a groove weld. (c) and (d) Underfill in groove welds

Cracks can occur in a wide variety of shapes and types and can be located in numerous positions in and around a welded

joint (Fig. 6). Cracks associated with welding can be categorized according to whether they originate in the weld itself or in the base metal. Four types commonly occur in the weld metal: transverse, longitudinal, crater, and hat cracks. Base metal cracks can be divided into seven categories: transverse cracks, underbead cracks, toe cracks, root cracks, lamellar tearing, delaminations, and fusion-line cracks.

Fig. 6 Identification of cracks according to location in weld and base metal. 1, crater crack in weld metal; 2, transverse crack in weld metal; 3, transverse crack in HAZ; 4, longitudinal crack in weld metal; 5, toe crack in base metal; 6, underbead crack in base metal; 7, fusion-line crack; 8, root crack in weld metal; 9, hat cracks in weld metal

Weld metal cracks and base metal cracks that extend to the surface can be detected by liquid penetrant and magnetic particle inspection. Magnetic particle inspection can also detect subsurface cracks, depending on their size, shape, and proximity to the surface. Although the orientation of a crack with respect to the direction of the radiation beam is the dominant factor in determining the ability of radiography to detect the crack, differences in composition between the base metal and the weld metal may create shadows to hide a crack that otherwise might be visible. Ultrasonic inspection is generally effective in detecting most cracks in the weld zone. Transverse cracks in weld metal (No. 2, Fig. 6) are formed when the predominant contraction stresses are in the direction of the weld axis. They can be hot cracks, which separate intergranularly as the result of hot shortness or localized planar shrinkage, or they can be transgranular separations produced by stresses exceeding the strength of the material. Transverse cracks lie in a plane normal to the axis of the weld and are usually open to the surface. They usually extend across the entire face of the weld and sometimes propagate into the base metal.

Transverse cracks in base metal (No. 3, Fig. 6) occur on the surface in or near the HAZ. They are the result of the high residual stresses induced by thermal cycling during welding. High hardness, excessive restraint, and the presence of hydrogen promote their formation. Such cracks propagate into the weld or beyond the HAZ into the base metal as far as is needed to relieve the residual stresses. Underbead cracks (No. 6, Fig. 6) are similar to transverse cracks in that they form in the HAZ because of high

hardness, excessive restraint, and the presence of hydrogen. Their orientation follows the contour of the HAZ. Longitudinal cracks (No. 4, Fig. 6) may exist in three forms, depending on their positions in the weld. Check cracks

are open to the surface and extend only partway through the weld. Root cracks extend from the root to some point within the weld. Full centerline cracks may extend from the root to the face of the weld metal. Check cracks are caused either by high contraction stresses in the final passes applied to a weld joint or by a hot-cracking mechanism. Root cracks are the most common form of longitudinal weld metal crack because of the relatively small size of the root pass. If such cracks are not removed, they can propagate through the weld as subsequent passes are applied. This is the usual mechanism by which full centerline cracks are formed. Centerline cracks may occur at either high or low temperatures. At low temperatures, cracking is generally the result of poor fit-up, overly rigid fit-up, or a small ratio of weld metal to base metal. All three types of longitudinal cracks are usually oriented perpendicular to the weld face and run along the plane that bisects the welded joint. Seldom are they open at the edge of the joint face, because this requires a fillet weld with an extremely convex bead. Crater cracks (No. 1, Fig. 6) are related to centerline cracks. As the name implies, crater cracks occur in the weld crater

formed at the end of a welding pass. Generally, this type of crack is caused by failure to fill the crater before breaking the arc. When this happens, the outer edges of the crater cool rapidly, producing stresses sufficient to crack the interior of the crater. This type of crack may be oriented longitudinally or transversely or may occur as a number of intersecting cracks forming the shape of a star. Longitudinal crater cracks can propagate along the axis of the weld to form a centerline crack. In addition, such cracks may propagate upward through the weld if they are not removed before subsequent passes are applied. Hat cracks (No. 9, Fig. 6) derive their name from the shape of the weld cross section with which they are usually

associated. This type of weld flares out near the weld face, resembling an inverted top hat. Hat cracks are the result of excessive voltage or welding speed. The cracks are located about halfway up through the weld and extend into the weld metal from the fusion line of the joint.

Toe and root cracks (No. 5 and 8, Fig. 6) can occur at the notches present at notch locations in the weld when high

residual stresses are present. Both toe and root cracks propagate through the brittle HAZ before they are arrested in more ductile regions of the base metal. Characteristically, they are oriented almost perpendicular to the base metal surface and run parallel to the weld axis. Lamellar tearing is the phenomenon that occurs in T-joints that are fillet welded on both sides. This condition, which

occurs in the base metal or HAZ of restrained weld joints, is characterized by a steplike crack parallel to the rolling plane. The crack originates internally because of tensile strains produced by the contraction of the weld metal and the surrounding HAZ during cooling. Figure 7 shows a typical condition.

Fig. 7 Lamellar tear caused by thermal contraction strain

Fusion-line cracks (No. 7, Fig. 6) can be classified as either weld metal cracks or base metal cracks because they occur along the fusion line between the two. There are no limitations as to where along the fusion line these cracks can occur or how far around the weld they can extend.

Discontinuities Associated With Specialized Welding Processes The preceding section has dealt mainly with the discontinuities common to conventional arc-welding processes. In addition, there are certain more specialized welding methods that may have discontinuities unique to them. These methods include electron beam, plasma arc, electroslag, friction, and resistance welding. In general, the types of discontinuities associated with these processes are the same as those associated with conventional arc welding; however, because of the nature of the processes and the joint configurations involved, such discontinuities may be oriented differently from those previously described, or they may present particular problems of location and evaluation. Electron Beam Welding In electron beam welding, as in all other welding processes, weld discontinuities can be divided into two major categories: • •

Those that occur at, or are open to, the surface Those that occur below the surface

Surface flaws include undercut, mismatch, underfill, reinforcement, cracks, missed seams, and LOP. Subsurface flaws include porosity, massive voids, bursts, cracks, missed seams, and LOP. Figure 8 shows poor welds containing these flaws, and a good weld with none of them.

Fig. 8 Electron beam welds showing flaws that can occur in poor welds and the absence of flaws in a good weld with reinforcement

Surface discontinuities such as undercut, mismatch, reinforcement, and underfill are macroscopic discontinuities related to the contour of the weld bead or the joint. As such, they are readily detected visually or dimensionally. Surface discontinuities such as cracks are usually detected visually using liquid penetrant inspection or using magnetic particle inspection if the material is ferromagnetic.

When liquid penetrants are used to inspect a weld for surface discontinuities such as cracks, missed seams, and LOP, the surface to be inspected must be clean and the layers of metal smeared from machining or peened from grit- or sandblasting must be removed. Generally, some type of etching or pickling treatment works well, but the possibility of hydrogen pickup from the treatment must be considered. Occasionally, special inspection procedures must be employed to detect some types of surface discontinuities. Missed seams and LOP are often difficult to detect because they are frequently associated with complex weld joints that prevent direct viewing of the affected surface. Because of this difficulty, missed seams are often detected using a visual witness-line procedure, in which equally spaced parallel lines are scribed on both sides of the unwelded joint at the crown and root surfaces. Missed seams, which result from misalignment of the electron beam with the joint such that the fusion zone fails to encompass the entire joint, are detected by observing the number of witness lines remaining on either side of the weld bead. By establishing the relationship between the width of the fusion zone and the spacing of the witness lines, reasonably accurate criteria for determining whether the joint has been contained within the weld path (and therefore whether missed seams are present) can be developed. Lack of penetration discontinuities occur when the fusion zone fails to penetrate through the entire joint thickness, resulting in an unbonded area near the root of the joint. These discontinuities are best detected by etching the root surface and observing the macroscopic shape and width of the fusion zone for full and even penetration. An alternative method for inspecting complete weld penetration is that of immersion pulse-echo ultrasonic testing. The planet gear carrier assembly shown in Fig. 9(a) consists of three decks (plates) and eight curved spacer sections. The assembly, which is made from SAE 15B22M boron-treated structural steel, is held together by 16 welds. Weld integrity is monitored by ultrasonic methods.

Fig. 9 Planet gear carrier assembly (a) showing the four welds in one stock that connect the three decks. The top and top-center welds are tested by the top ultrasonic transducer, and the bottom and bottom-center welds are inspected by the bottom transducer. (b) Close-up of upper transducer in position to test the welds. Courtesy of John M. St. John, Caterpillar, Inc.

The parts are mounted on a turntable on a locating fixture so that the welds along the outside diameter are accessible. Two transducers, located above and below the assembly, are used. Figure 9(b) shows the upper transducer in position to test the welds. An overall view of the tank, turntable, controls, and reject/accept light panel is shown in Fig. 10. The use of such a system involves little downtime and enables a high quality level to be maintained. More information on ultrasonic test methods is presented later in this article.

Fig. 10 Overall view of the ultrasonic unit used to test the electron beam welded assembly shown in Fig. 9. Courtesy of John M. St. John, Caterpillar, Inc.

Subsurface discontinuities are generally considerably more difficult to detect than surface discontinuities because observation is indirect. The two most reliable and widely used NDI methods are radiography and ultrasonics.

Volume-type discontinuities such as porosity, voids, and bursts are detected by radiographic inspection, provided their cross sections presented to the radiating beam exceed 1 to 2% of the beam path in the metal. Discontinuities that present extremely thin cross sections to the beam path, such as cracks, missed seams, and LOP, are detectable with x-rays only if they are viewed from the end along their planar dimensions.

Ultrasonic inspection can detect most volume discontinuities as well as planar discontinuities. Planar discontinuities are best detected normal to the plane of the discontinuities, but missed seams and LOP often appear as continuous porosity when viewed looking down from the crown to the root of the weld in the plane of the discontinuity. Because of the inherent dependence of both radiographic and ultrasonic inspection on the shape and orientation of flaws and because each of the two methods can generally detect those flaws that the other misses, it is most advisable to complement one method with the other. Furthermore, to increase the likelihood of properly viewing a flaw, one of the methods should be employed in at least two (preferably perpendicular) directions. Plasma Arc Welding Discontinuities that occur in plasma arc welds include both surface and subsurface types, as shown in Fig. 11.

Fig. 11 Plasma arc welds showing flaws that can occur in poor welds and the absence of flaws in good reinforced weld

Surface discontinuities such as irregular reinforcement, underfill, undercut, and mismatch that are associated with weld bead contour and joint alignment are easily detected visually or dimensionally. Lack of penetration is also detected visually through the absence of a root bead. Weld cracks that are open to the surface are detected with liquid penetrants. Surface contamination, which results from insufficient shielding-gas coverage, is detected by the severe discoloration of the weld bead or adjacent HAZ. Subsurface discontinuities are generally more prevalent in manual than in automatic plasma arc welding and are

detected primarily by radiographic or ultrasonic inspection. Porosity is by far the most commonly encountered discontinuity. Radiographic inspection is limited to detecting pores greater than approximately 1 to 2% of the joint thickness. Visibility is greater if both the crown and root beads are machined flush. Ultrasonic inspection can detect porosity if the joint is machined flush and joint thickness exceeds approximately 1.3 mm (0.050 in.). Tunneling, as shown in Fig. 11, is a severe void along the boundary of the fusion zone and the HAZ. This discontinuity results from a combination of torch alignment and welding conditions (particularly travel speed). Tunneling is readily detectable by radiographic inspection. Lack of fusion discontinuities occur in either single-pass or multiple-pass repair welds (Fig. 11). These discontinuities result from insufficient heat input to permit complete fusion of a particular weld bead to the part. Incomplete fusion can be detected by radiographic or ultrasonic inspection. Depending on the orientation of the discontinuity, one method may have an advantage over the other, so both should be used for optimum inspection. Subsurface weld cracks, regardless of their cause, are detectable by radiographic and ultrasonic inspection. Subsurface contamination in plasma arc welding results when copper from the torch nozzle is expelled into the weld. This is caused by excessive heat, usually produced in manual repair welding when the torch nozzle is placed too close to the weld, particularly in a groove. The resulting contamination, which may be detrimental, is undetectable by conventional NDI methods. The only way of detecting copper contamination is by alerting the operator to watch for copper expulsion, which then must be machined out.

Electroslag Welding Electroslag welding involves the use of copper dams over the open surfaces of a butt joint to hold the molten metal and the slag layer as the joint is built up vertically. Wire is fed into the slag layer continuously and is melted by the heat generated as current passes through the highly resistant slag layer. Generally, electroslag welds are inspected with the same nondestructive examination (NDE) methods as other heavysection welds. With the exception of procedure qualification, all testing is nondestructive because of the sizes used. Techniques such as radiography and ultrasonic inspection are most often used, while visual, magnetic particle, and liquid penetrant testing are used also. Internal defects are generally more serious. Radiography and ultrasonic tests are the best methods for locating internal discontinuities. Because of the nature of the process, LOF is rare. If fusion is achieved on external material edges, then fusion is generally complete throughout. Cracking may occur either in the weld or the HAZ. Porosity may either take the form of a rounded or a piped shape; the latter is often called wormhole porosity. Ultrasonic inspection is probably the quickest single method for inspecting any large weldment. If defects should occur, they appear as porosity or centerline cracking. Ultrasonic inspection is effective for locating either type of defect; however, only well-qualified personnel should set up the equipment and interpret the test results. Electroslag welding results in large dendritic grain sizes because of the slow cooling rate. Inexperienced personnel often use high sensitivity and actually pick up the large coarse grains; when such welds are sectioned, usually no defects are present. Inspectors must learn to use low sensitivity to obtain good results when inspecting electroslag welds. Magnetic particle inspection is not a particularly good inspection method, because the areas examined by this technique are primarily surface or near surface. This is only a small percentage of the total weld; the only useful information is either checking the ends for craters, cracks, or centerline cracking or possibly for lack of edge fusion on the weld faces. Usually, a visual examination gives the same result unless the defect is subsurface. Visual examinations are only effective for surface defects, which are not common in this process. Friction Welding If impurities are properly dispelled during upsetting, friction or inertia welds are generally free of voids and inclusions. Incomplete center fusion can occur when flywheel speed is too low, when the amount of upset is insufficient, or when mating surfaces are concave. Tearing in the HAZ can be caused by low flywheel speed or excessive flywheel size. Cracks can occur when materials that are prone to hot shortness are joined. The penetration of a split between the extruded flash into the workpiece cross section is most prevalent during the welding of thin-wall tubing using improper conditions that do not allow for sufficient material upset. The area where LOF generally occurs is at or near the center of the weld cross section. Because this is a subsurface discontinuity, detection is limited to radiographic or ultrasonic inspection; ultrasonic inspection is more practical. The longitudinal wave test (either manual-contact or immersion method) with beam propagation perpendicular to the area of LOF gives the most reliable results. This test can be performed as long as one end of the workpiece is accessible to the transducer. Penetration of the split between the extruded flash on the outer surface of a tube is readily detected by liquid penetrant or magnetic particle inspection after the flash has been removed by machining. A split between the weld flash on the inner surface of the tube can be detected by ultrasonic inspection using the angle-beam technique with manual contact of the transducer to the outside surface of the tube. The transducer contacts the tube so that the sound propagates along the longitudinal axis of the tube through the weldment. Resistance Welding Resistance welding encompasses spot, seam, and projection welding, each of which involves the joining of metals by passing current from one side of the joint to the other. The types of discontinuities found in resistance welds include porosity, LOF, and cracks. Porosity will generally be found on the centerline of the weld nugget. Lack of fusion may also be manifested as a centerline cavity. Either of these can be caused by overheating, inadequate pressure, premature release of pressure, or late application of pressure. Cracks may be induced by overheating, removal of pressure before weld quenching is completed, improper loading, poor joint fit-up, or expulsion of excess metal from the weld.

Weld Appearance. On the surface of a resistance spot welded assembly, the weld spot should be uniform in shape and

relatively smooth, and it should be free of surface fusion, deep electrode indentations, electrode deposits, pits, cracks, sheet separation, abnormal discoloration around the weld, or other conditions indicating improper maintenance of electrodes or functioning of equipment. However, surface appearance is not always a good indicator of spot weld quality, because shunting and other causes of insufficient heating or incomplete penetration usually leave no visible effects on the workpiece. The common practice for monitoring spot weld quality in manufacturing operations is the teardown method augmented with pry testing and visual inspection (Ref 1). In visual inspection, the operator uses the physical features of the weld surface, such as coloration, indentation, and smoothness, for assessing the quality of the weld. In pry testing, a wedgeshaped tool is inserted between the metal sheets next to the accessible welds, and a prying action is performed to see if the sheets will separate in the weld zone. The teardown method consists of physically tearing apart the welded members with hammers and chisels to determine the presence and size adequacy of fused metal nuggets at the spot weld site. The specifications require that the parent metal be torn and that the weld nugget remain intact. The destructiveness and/or inadequacy of these common inspection methods has long been recognized, and as a result, nondestructive methods have been extensively studied. The pulse-echo ultrasonic inspection of spot welds is now feasible and is discussed in the section "Ultrasonic Inspection" in this article. The use of acoustic emission is also discussed below. Diffusion Bonding (Ref 2, 3, 4, 5) Diffusion bonding is a metal joining process that requires the application of controlled pressures at elevated temperatures and usually a protective atmosphere to prevent oxidation. No melting and only limited macroscopic deformation or relative motion between the faying surfaces of the parts occur during bonding. As such, the principal mechanism for joint formation is solid-state diffusion. A diffusion aid (filler metal) may or may not be used. Terms that are also frequently used to describe the process include diffusion welding, solid-state bonding, pressure bonding, isostatic bonding, and hot press bonding. Diffusion bonding has the advantage of producing a product finished to size, with joint efficiencies approaching 100%. Details of the process are given in the article "Fundamentals of Diffusion Bonding" in Welding, Brazing, and Soldering, Volume 6 of the ASM Handbook. Discontinuities in Diffusion Bonds. In the case of fusion welds, the detection of discontinuities less than 1 mm

(0.04 in.) in size is not generally expected. In diffusion bonding, in which no major lack of bonding occurs, individual discontinuities may be only micrometers in size. To understand how discontinuities form in diffusion-bonded structures, it is first necessary to consider the principles of the process. As illustrated in Fig. 12, metal surfaces have several general characteristics: • • • •

Roughness An oxidized or otherwise chemically reacted and adherent layer Other randomly distributed solid or liquid products such as oil, grease, and dirt Adsorbed gas, moisture, or both

Because of these characteristics, two necessary conditions that must be met before a satisfactory diffusion bond can be made are: • •

Mechanical intimacy of metal-to-metal contact must be achieved Interfering surface contaminants must be disrupted and dispersed to permit metallic bonding to occur (solvent cleaning and inert gas atmospheres can reduce or eliminate problems associated with surface contamination and oxide formation, respectively)

Fig. 12 Characteristics of a metal surface showing roughness and contaminants present. Source: Ref 5

For a given set of processing parameters, surface roughness is probably the most important variable influencing the quality of diffusion-bonded joints. The size of the discontinuities (voids) is principally determined by the scale of roughness of the surfaces being bonded. The degree of surface roughness is dependent on the material and fabrication/machining technique used. It has been shown that bonding becomes easier with finer surface roughness prior to bonding. Figure 13 compares the surface roughness values produced by a variety of fabrication methods. More detailed information on surface roughness can be found in the article "Surface Finish and Surface Integrity" in Machining, Volume 16 of ASM Handbook, formerly 9th Edition Metals Handbook.

Fig. 13 Surface roughness produced by common production methods. The ranges shown are typical of the processes listed. Higher or lower values can be obtained under special conditions.

The mechanism of bond formation in diffusion bonding is believed to be the deformation of the surface roughness in order to cause metal-to-metal contact at asperities, followed by the removal of interfacial voids and cracks by diffusional and creep processes. For conventional diffusion bonding without a diffusion aid, the three-stage mechanistic model shown in Fig. 14 describes bond formation. In the first stage, deformation of the contacting asperities occurs primarily by yielding and by creep deformation mechanisms to produce intimate contact over a large fraction of the interfacial area. At the end of this stage, the joint is essentially a grain boundary at the areas of contact with voids between these areas. During the second stage, diffusion becomes more important than deformation, and many of the voids disappear as the grain-boundary diffusion of atoms continues. Simultaneously, the interfacial grain boundary migrates to an equilibrium configuration away from the original plane of the joint, leaving many of the remaining voids within the grains. In the third stage, the remaining voids are eliminated by the volume diffusion of atoms to the void surface (equivalent to diffusion of vacancies away from the void). Successful completion of stage three is dependent on proper surface processing and joint processing.

Fig. 14 Three-stage mechanistic model of diffusion welding. (a) Initial asperity contact. (b) First-stage deformation and interfacial boundary formation. (c) Second-stage grain-boundary migration and pore elimination. (d) Third-stage volume diffusion and pore elimination. Source: Ref 5

Successful nondestructive evaluation of diffusion-bonded joints requires that the maximum size and distribution of discontinuities be determined. However, many conventional NDE methods and equipment are not adequate for discontinuity determination. Fluorescent penetrants provide excellent detection capability for lack of bonding as long as the interfacial crack breaks the surface. They are completely ineffective for defects that have no path to the surface. Conventional film radiography is not suitable for detecting the extremely small defects involved in diffusion bonding, but the use of x-ray microfocus techniques coupled with digital image enhancement offers an improvement in resolution. Discontinuities as small as 50 m (0.002 in.) have been detected. Conventional ultrasonic testing has some applications, although only significant lack of bonding or clusters of smaller defects can be reliably detected. High-resolution flaw detection involving frequencies approaching 100 MHz appears to have distinct advantages over conventional testing. Scanning acoustic microscopy also appears to offer excellent possibilities in diffusion bond inspection. Eddy current and thermal methods are relatively unsatisfactory for most applications. Methods of Nondestructive Inspection The nondestructive inspection of weldments has two functions: • •

Quality control, which is the monitoring of welder and equipment performance and of the quality of the consumables and the base materials used Acceptance or rejection of a weld on the basis of its fitness for purpose under the service conditions imposed on the structure

The appropriate method of inspection is different for each function. If evaluation is a viable option, discontinuities must be detected, identified, located exactly, sized, and their orientation established, which limits inspection to a volumetric technique.

Weld discontinuities constitute the center of activity with the inspection of welded constructions. The most widely used inspection techniques used in the welding industry are visual, liquid penetrant, magnetic particle, radiographic, ultrasonic, acoustic emission, eddy current, and electric current perturbation methods. Each of these techniques has specific advantages and limitations. Existing codes and standards that provide guidelines for these various techniques are based on the capabilities and/or limitations of these nondestructive methods. Selection of Technique. A number of factors influence selection of the appropriate nondestructive test technique for

inspecting a welded structure, including discontinuity characteristics, fracture mechanics requirements, part size, portability of equipment, and other application constraints. These categories, although perhaps unique to a specific inspection problem, may not clearly point the way to the most appropriate technique. It is generally necessary to exercise engineering judgment in ranking the importance of these criteria and thus determining the optimum inspection technique. Characteristics of the Discontinuity. Because nondestructive techniques are based on physical phenomena, it is useful to describe the properties of the discontinuity of interest, such as composition and electrical, magnetic, mechanical, and thermal properties. Most significant are those properties that are most different from those of the weld or base metal. It is also necessary to identify a means of discriminating between discontinuities with similar properties. Fracture

mechanics requirements, solely from a discontinuity viewpoint, typically include detection, identification, location, sizing, and orientation. In addition, complicated configurations may necessitate a non-destructive assessment of the state of the stress of the region containing the discontinuity. In the selection process, it is important to establish these requirements correctly. This may involve consultation with stress analysts, materials engineers, and statisticians.

Often, the criteria may strongly suggest a particular technique. Under ideal conditions, such as in a laboratory, the application of such a technique might be routine. In the field, however, other factors may force a different choice of technique. Constraints tend to be unique to a given application and may be completely different even when the welding process

and metals are the same. Some of these constraints include: • • • • • • • • •

Access to the region under inspection Geometry of the structure (flat, curved, thick, thin) Condition of the surface (smooth, irregular) Mode of inspection (preservice, in-service, continuous, periodic, spot) Environment (hostile, underwater, and so on) Time available for inspection (high speed, time intensive) Reliability Application of multiple techniques Cost

Failure to consider adequately the constraints imposed by a specific application can render the most sophisticated equipment and theory useless. Moreover, for the simple or less important cases of failure, it may be unnecessary. Once criteria have been established, an optimum inspection technique can be selected, or designed and constructed. The terms accuracy, sensitivity, and reliability are used loosely in NDE. Often, they are discussed as one term to

avoid distinguishing among the specific aspects of these terminologies. Accuracy is the attribute of an inspection method that describes the correctness of the technique within the limits of its

precision. In other words, the technique is highly accurate if the indications resulting from the technique are correct. This does not mean that the technique was able to detect all discontinuities present, but rather that those indicated actually exist. Sensitivity, on the other hand, refers to the capability of a technique to detect discontinuities that are small or that have

properties only slightly different from the material in which they reside. Figure 15 schematically illustrates the concepts of accuracy and sensitivity in the context of detection probability. In general, sensitivity is gained at the expense of accuracy because high sensitivity increases the probability of false alarms.

Fig. 15 Detection probability for a true positive indication. High sensitivity increases the likelihood of false indications. The minimum NDE parameter size required to establish fitness-for-purpose must lie to the right of the transition region, or reliability threshold, to achieve satisfactory reliability.

Reliability is a combination of both accuracy and sensitivity. Three factors influence reliability: inspection procedure, including the instrumentation; human factors (inspector motivation, experience, training, education, and so on); and data analysis. Uncalibrated equipment, improper application of technique, and inconsistent quality of accessory equipment (transducers, couplant, film, chemicals, and so on) may affect accuracy and, in some cases, sensitivity. Poor inspector technique, unfamiliar response, lack of concentration, and other human factors can combine to reduce reliability. Data analysis, or the lack of it, can influence reliability as well; generally, inspection is performed under conditions in which detection probability is less than 100% and is not constant with discontinuity severity. Consequently, statistics must be employed to establish the level of confidence that may be attached to the inspection results.

High sensitivity with low accuracy may be far worse, from the viewpoint of reliability, than low sensitivity with high accuracy, especially if the sensitivity level is adequate for detecting the weld discontinuities in question. As a general rule, the transition region of the detection probability curve indicates the degree of reliability. If this region occurs with the limits encompassed by inspection capabilities, which are smaller than the values required for evaluating the fitness-forpurpose of the welds being inspected, reliability is satisfactory. If, on the other hand, the region occurs at values higher than those required, reliability is unsatisfactory. The transition region can be viewed as the reliability threshold. More detailed information on the reliability of nondestructive test data can be found in the Section "Quantitative Nondestructive Evaluation" in this Volume. Visual Inspection For many noncritical welds, integrity is verified principally by visual inspection. Even when other nondestructive methods are used, visual inspection still constitutes an important part of practical quality control. Widely used to detect discontinuities, visual inspection is simple, quick, and relatively inexpensive. The only aids that might be used to determine the conformity of a weld are a low-power magnifier, a borescope, a dental mirror, or a gage. Visual inspection can and should be done before, during, and after welding. Although visual inspection is the simplest inspection method to use, a definite procedure should be established to ensure that it is carried out accurately and uniformly. Visual inspection is useful for checking the following: • • • •

Dimensional accuracy of weldments Conformity of welds to size and contour requirements Acceptability of weld appearance with regard to surface roughness, weld spatter, and cleanness Presence of surface flaws such as unfilled craters, pockmarks, undercuts, overlaps, and cracks.

Although visual inspection is an invaluable method, it is unreliable for detecting subsurface flaws. Therefore, judgment of weld quality must be based on information in addition to that afforded by surface indications.

Additional information can be gained by observations before and during welding. For example, if the plate is free of laminations and properly cleaned and if the welding procedure is followed carefully, the completed weld can be judged on the basis of visual inspection. Additional information can also be gained by using other NDI methods that detect subsurface and surface flaws. Dimensional Accuracy and Conformity. All weldments are fabricated to meet certain specified dimensions. The fabricator must be aware of the amount of shrinkage that can be expected at each welded joint, the effect of welding sequence on warpage or distortion, and the effect of subsequent heat treatment used to provide dimensional stability of the weldment in service. Weldments that require rigid control of final dimensions usually must be machined after welding. Dimensional tolerances for as-welded components depend on the thickness of the material, the alloy being welded, the overall size of the product, and the particular welding process used.

The dimensional accuracy of weldments is determined by conventional measuring methods, such as rules, scales, calipers, micrometers, and gages. The conformity of welds with regard to size and contour can be determined by a weld gage. The weld gage shown in Fig. 16 is used when visually inspecting fillet welds at 90° intersections. The size of the fillet weld, which is defined by the length of the leg, is stamped on the gage. The weld gage determines whether or not the size of the fillet weld is within allowable limits and whether there is excessive concavity or convexity. This gage is designed for use on joints between surfaces that are perpendicular. Special weld gages are used when the surfaces are at angles other than 90°. For groove welds, the width of the finished welds must be in accordance with the required groove angle, root face, and root opening. The height of reinforcement of the face and root must be consistent with specified requirements and can be measured by a weld gage.

A

Minimum allowable length of leg

B

Maximum allowable length of leg

C

1.414 times maximum allowable throat size (specifies maximum allowable convexity)

D

Maximum allowable length of leg when maximum allowable concavity is present

E

A plus B plus nominal weld size (or nominal length of leg)

F

Minimum allowable throat size (specifies maximum allowable concavity)

T

Additional tolerance for clearance of gage when placed in the fillet

Fig. 16 Gage for visual inspection of a fillet weld at a 90° intersection. Similar gages can be made for other angles. Dimension given in inches

Appearance Standards. The acceptance of welds with regard to appearance implies the use of a visual standard, such

as a sample weldment or a workmanship standard. Requirements as to surface appearance differ widely, depending on the application. For example, when aesthetics are important, a smooth weld that is uniform in size and contour may be required. The inspection of multiple-pass welds is often based on a workmanship standard. Figure 17 indicates how such standards are prepared for use in the visual inspection of groove and fillet welds. The workmanship standard is a section of a joint similar to the one in manufacture, except that portions of each weld pass are shown. Each pass of the production weld is compared with corresponding passes of the workmanship standard.

Fig. 17 Workmanship standard for visual comparison during inspection of single-V-groove welds and fillet welds. Dimensions given in inches

Discontinuities. Before a weld is visually inspected for discontinuities such as unfilled craters, surface holes, undercuts, overlaps, surface cracks, and incomplete joint penetration, the surface of the weld should be cleaned of oxides and slag. Cleaning must be done carefully. For example, a chipping hammer used to remove slag could leave hammer marks that can hide fine cracks. Shotblasting can peen the surface of relatively soft metals and hide flaws. A stiff wire brush and sandblasting have been found to be satisfactory for cleaning surfaces of slag and oxides without marring.

Additional information on the uses and equipment associated with the visual examination of parts and assemblies is available in the article "Visual Inspection" in this Volume. Magnetic Particle Inspection Magnetic particle inspection is a nondestructive method of detecting surface and near-surface flaws in ferromagnetic materials. It consists of three basic operations: • • •

Establishing a suitable magnetic field in the material being inspected Applying magnetic particles to the surface of the material Examining the surface of the material for accumulations of the particles (indications) and evaluating the serviceability of the material

Capabilities and Limitations. Magnetic particle inspection is particularly suitable for the detection of surface flaws

in highly ferromagnetic metals. Under favorable conditions, those discontinuities that lie immediately under the surface are also detectable. Nonferromagnetic and weakly ferromagnetic metals, which cannot be strongly magnetized, cannot be inspected by this method. With suitable ferromagnetic metals, magnetic particle inspection is highly sensitive and produces readily discernible indications at flaws in the surface of the material being inspected. Details on the application of this method are available in the article "Magnetic Particle Inspection" in this Volume. The types of weld discontinuities normally detected by magnetic particle inspection include cracks, LOP, LOF, and porosity open to the surface. Linear porosity, slag inclusions, and gas pockets can be detected if large or extensive or if smaller and near the surface. The recognition of patterns that indicate deep-lying flaws requires more experience than that required to detect surface flaws. Nonrelevant indications that have no bearing on the quality of the weldment may be produced. These indications are

magnetic particle patterns held by conditions caused by leakage fields. Some of these conditions are: • • • • •

• • •

Particles held mechanically or by gravity in surface irregularities Adherent scale or slag Indications at a sharp change in material direction, such as sharp fillets and threads Grain boundaries. Large grain size in the weld metal or the base metal may produce indications Boundary zones in welds, such as indications produced at the junction of the weld metal and the base metal. This condition occurs in fillet welds at T-joints, or in double-V-groove joints, where 100% penetration is not required Flow lines in forgings and formed parts Brazed joints. Two parts made of a ferromagnetic material joined by a nonferromagnetic material will produce an indication Different degrees of hardness in a material, which will usually have different permeabilities that may create a leakage field, forming indications

Operational Requirements. The magnetic particle inspection of weldments requires that the weld bead be free of

scale, slag, and moisture. For maximum sensitivity, the weld bead should be machined flush with the surface; however, wire brushing, sandblasting, or gritblasting usually produces a satisfactory bead surface. If the weld bead is rough, grinding will remove the high spots. Weldments are often inspected using the dry particle method. A powder or paste of a color that gives the best possible contrast to the surface being inspected should be used. The type of magnetizing current used depends on whether there are surface or subsurface discontinuities. Alternating current is satisfactory for surface cracks, but if the deepest possible penetration is essential, direct current, direct current with surge, or half-wave rectified alternating current is used.

The voltage should be as low as practical to reduce the possibility of damage to the surface of the part from overheating or arcing at contacts. Another advantage of low voltage is freedom from arc flashes if a prod slips or is withdrawn before the current is turned off. The field strength and flux density used must be determined for each type of weldment. An overly strong field will cause the magnetic particles to adhere too tightly to the surface and hinder their mobility, preventing them from moving to the sites of the flaws. Low field strengths result in nondiscernible patterns and failure to detect indications. Inspections can be made using the continuous-field and residual-field methods. In the continuous-field method, magnetic particles are placed on the weldment while the current is flowing. In the residual-field method, the particles are placed on the weldment after the magnetizing current is turned off. Residual magnetic fields are weaker than continuous fields. Consequently, inspections using the residual-field method are less sensitive. The need for the demagnetization of weldments after magnetic particle inspection must be given serious consideration. Where subsequent welding or machining operations are required, it is good practice to demagnetize. Residual magnetism may also hinder cleaning operations and interfere with the performance of instruments used near the weldment. Liquid Penetrant Inspection Liquid penetrant inspection is capable of detecting discontinuities open to the surface in weldments made of either ferromagnetic or nonferromagnetic alloys, even when the flaws are generally not visible to the unaided eye. Liquid penetrant is applied to the surface of the part, where it remains for a period of time and penetrates into the flaws. For the correct usage of liquid penetrant inspection, it is essential that the surface of the part be thoroughly clean, leaving the openings free to receive the penetrant. Operating temperatures of 20 to 30 °C (70 to 90 °F) produce optimum results. If the part is cold, the penetrant may become chilled and thickened so that it cannot enter very fine openings. If the part or the penetrant is too hot, the volatile components of the penetrant may evaporate, reducing the sensitivity. After the penetrating period, the excess penetrant remaining on the surface is removed. An absorbent, light-colored developer is then applied to the surface. This developer acts as the blotter, drawing out a portion of the penetrant that had previously seeped into the surface openings. As the penetrant is drawn out, it diffuses into the developer, forming indications that are wider than the surface openings. The inspector looks for these colored or fluorescent indications against the background of the developer. Details of the mechanics of this method are given in the article "Liquid Penetrant Inspection" in this Volume. Radiographic Inspection* Radiography is a nondestructive method that uses a beam of penetrating radiation such as x-rays and -rays (see the article "Radiographic Inspection" in this Volume). When the beam passes through a weldment, some of the radiation energy is absorbed, and the intensity of the beam is reduced. Variations in beam intensity are recorded on film or on a screen when a fluoroscope or an image intensifier is used. The variations are seen as differences in shading that are typical of the types and sizes of any discontinuities present. Radiography is used to detect subsurface discontinuities as well as those that are open to the surface. Penetrameters. Most U.S. codes and specifications for radiographic inspection require that the sensitivity to a specific

flaw size be indicated on each radiograph. Sensitivity is determined by placing a penetrameter (image-quality indicator) that is made of substantially the same material as the specimen directly on the specimen, or on a block of identical material of the same thickness as the specimen, prior to each exposure. The penetrameters are usually placed on the side of the specimen nearest the source, parallel and adjacent to the weld at one end of the exposed length, with the small holes in the penetrameter toward the outer end. Penetrameter thickness is usually specified to be 2% of the specimen thickness (through the weld zone), although 1 and 4% penetrameter thicknesses are also common. American Society for Testing and Materials and American Society of Mechanical Engineers plaque-type penetrameters have three small drilled holes, whose diameters equal one, two, and four times the penetrameter thickness. Image quality is determined by the ability to distinguish both the outline of the penetrameter and one or more of the drilled holes on the processed radiograph; for example, 2-2T sensitivity is achieved when a 2% penetrameter and the hole that has a diameter of twice the penetrameter thickness are visible. Also, the image of raised numbers that identify the actual thickness (not the percentage thickness) of a penetrameter should appear clearly, superimposed on the image of the penetrameter.

Surface discontinuities that are detectable by radiography include undercuts (Fig. 18 and 19), longitudinal grooves,

concavity at the weld root (Fig. 20 and 21), incomplete filling of grooves, excessive penetration (Fig. 22), offset or mismatch (Fig. 23 and 24), burn-through (Fig. 25), irregularities at electrode-change points, grinding marks, and electrode spatter. Surface irregularities may cause density variations on a radiograph. When possible, they should be removed before a weld is radiographed. When impossible to remove, they must be considered during interpretation.

Fig. 18 External undercut, which is a gouging out of the piece to be welded, alongside the edge of the top or external surface of the weld. Radiographic image: An irregular darker density along the edge of the weld image. The density will always be darker than the density of the pieces being welded. Welding process: SMAW. Source: E.I. Du Pont de Nemours & Company, Inc.

Fig. 19 Internal (root) undercut, which is a gouging out of the parent metal, alongside the edge of the bottom or internal surface of the weld. Radiographic image: An irregular darker density near the center of the width of the weld image and along the edge of the root pass image. Welding process: SMAW. Source: E.I. Du Pont de

Nemours & Company, Inc.

Fig. 20 External concavity or insufficient fill, which is a depression in the top of the weld, or cover pass, indicating a thinner-than-normal section thickness. Radiographic image: A weld density darker than the density of the pieces being welded and extending across the full width of the weld image. Welding process: SMAW. Source: E.I. Du Pont de Nemours & Company, Inc.

Fig. 21 Internal concavity (suck back), which is a depression in the center of the surface of the root pass. Radiographic image: An elongated irregular darker density with fuzzy edges, in the center of the width of the weld image. Welding process: GTAW-SMAW. Source: E.I. Du Pont de Nemours & Company, Inc.

Fig. 22 Excessive penetration (icicles, drop-through), which is extra metal at the bottom (root) of the weld. Radiographic image: A lighter density in the center of the width of the weld image, either extended along the weld or in isolated circular drops. Welding process: SMAW. Source: E.I. Du Pont de Nemours & Company, Inc.

Fig. 23 Offset or mismatch with LOP, which is a misalignment of the pieces to be welded and insufficient filling of the bottom of the weld or root area. Radiographic image: An abrupt density change across the width of the weld image with a straight longitudinal darker-density line at the center of the width of the weld image along the edge of the density change. Welding process: SMAW. Source: E.I. Du Pont de Nemours & Company, Inc.

Fig. 24 Offset or mismatch (high-low), which is a misalignment of the pieces to be welded. Radiographic image: An abrupt change in film density across the width of the weld image. Welding process: SMAW. Source: E.I. Du Pont de Nemours & Company, Inc.

Fig. 25 Burn-through, which is a severe depression or a crater-type hole at the bottom of the weld but usually not elongated. Radiographic image: A localized darker density with fuzzy edges in the center of the width of the weld image. It may be wider than the width of the root pass image. Welding process: SMAW. Source: E.I. Du Pont de Nemours & Company, Inc.

Undercuts result in a radiographic image of a dark line of varying width and density. The darkness or density of the line indicates the depth of the undercut.

Longitudinal grooves in the surface of weld metal produce dark lines on a radiograph that are roughly parallel to the weld seam but are seldom straight. These dark lines have diffused edges and should not be mistaken for slag lines, which are narrow and more sharply defined. Concavity at the weld root occurs only in joints that are welded from one side, such as pipe joints. It appears on the radiograph as a darker region than the base metal. If weld reinforcement is too high, the radiograph shows a lighter line down the weld seam. There is a sharp change in image density where the reinforcement meets the base metal. Weld reinforcements not ground completely smooth show irregular densities, often with sharp borders. When excess metal is deposited on a final pass, it may overlap the base metal, causing LOF at the edge of the reinforcement. Although there is a sharp change in image density between reinforcement and base metal, the edge of the reinforcement image is usually irregular. Irregularities at electrode-change points may be either darker or lighter than the adjacent areas. Grinding marks appear as darker areas or lines in relation to the adjacent areas in the radiograph. Electrode spatter will appear as globular and lighter on the radiograph and should be removed before radiographic inspection. As material thickness increases, radiography becomes less sensitive as an inspection method. Thus, for thick material, other NDI methods are used before, during, and after welding on both the base metal and weld metal. Subsurface discontinuities detectable by radiography include gas porosity (Fig. 26, 27, 28), slag inclusions (Fig. 29 and 30), cracks (Fig. 31, 32, 33), LOP (Fig. 34), LOF (Fig. 35 and 36), and tungsten inclusions (Fig. 37). On a radiograph, a pore appears as a round or oval dark spot with or without a rather sharp tail. The spots caused by porosity are often of varying size and distribution. A wormhole appears as a dark rectangle if its long axis is perpendicular to the radiation beam, and it appears as two concentric circles, one darker than the other, if the long axis is parallel to the beam. Linear porosity is recorded on radiographs as a series of round dark spots along a line parallel to the direction of welding.

Fig. 26 Root pass aligned porosity, which involves rounded and elongated voids in the bottom tom of the weld aligned along the weld centerline. Radiographic image: Rounded and elongated darker-density spots, which may be connected, in a straight line in the center of the width of the weld image. Welding process: GMAW. Source: E.I. Du Pont de Nemours & Company, Inc.

Fig. 27 Cluster porosity, which involves rounded or slightly elongated voids grouped together. Radiographic image: Rounded or slightly elongated darker-density spots in clusters with the clusters randomly spaced. Welding process: SMAW. Source: E.I. Du Pont de Nemours & Company, Inc.

Fig. 28 Scattered porosity, which involves rounded voids random in size and location. Radiographic image: Rounded spots of darker densities random in size and location. Welding process: SMAW. Source: E.I. Du Pont de Nemours & Company, Inc.

Fig. 29 Elongated slag lines (wagon tracks), which are impurities that solidify on the surface after welding and were not removed between passes. Radiographic image: Elongated, parallel, or single darker-density lines, irregular in width and slightly winding in the lengthwise direction. Welding process: SMAW. Source: E.I. Du Pont de Nemours & Company, Inc.

Fig. 30 Interpass slag inclusions, which are usually nonmetallic impurities that solidified on the weld surface and were not removed between weld passes. Radiographic image: An irregularly shaped darker-density spot, usually slightly elongated and randomly spaced. Welding process: SMAW. Source: E.I. Du Pont de Nemours & Company, Inc.

Fig. 31 Transverse crack, which is a fracture in the weld metal running across the weld. Radiographic image: Feathery, twisting line of darker density running across the width of the weld image. Welding process: GMAWGTAW. Source: E.I. Du Pont de Nemours & Company, Inc.

Fig. 32 Longitudinal crack, which is a fracture in the weld metal running lengthwise in the welding direction. Radiographic image: Feathery, twisting lines of darker density running lengthwise along the weld at any location in the width of the weld image. Welding process: GMAW-SMAW. Source: E.I. Du Pont de Nemours & Company, Inc.

Fig. 33 Longitudinal root crack, which is a fracture in the weld metal at the edge of the root pass. Radiographic image: Feathery, twisting lines of darker density along the edge of the image of the root pass. The twisting feature helps to distinguish the root crack from incomplete root penetration. Welding process: SMAW. Source: E.I. Du Pont de Nemours & Company, Inc.

Fig. 34 Lack of penetration, which occurs when the edges of the pieces have not been welded together, usually at the bottom of single-V-groove welds. Radiographic image: A darker-density band, with very straight parallel edges, in the center of the width of the weld image. Welding process: SMAW. Source: E.I. Du Pont de Nemours & Company, Inc.

Fig. 35 Lack of sidewall fusion, which involves elongated voids between the weld beads and the joint surfaces to be welded. Radiographic image: Elongated parallel, or single, darker-density lines, sometimes with darkerdensity spots dispersed along the LOF lines, which are very straight in the lengthwise direction and not winding like elongated slag lines. Although one edge of the LOF lines may be very straight as with LOP, lack of sidewall fusion images will not be in the center of the width of the weld image. Welding process: GMAW. Source: E.I. Du Pont de Nemours & Company, Inc.

Fig. 36 Interpass cold lap, which involves LOF areas along the top surface and edge of lower passes. Radiographic image: Small spots of darker densities, some with slightly elongated tails, aligned in the welding direction and not in the center of the width of the weld image. Welding process: GMAW. Source: E.I. Du Pont de Nemours & Company, Inc.

Fig. 37 Tungsten inclusions, which are random bits of tungsten fused into but not melted into the weld metal. Radiographic image: Irregularly shaped, lower-density spots randomly located in the weld image. Welding process: GTAW. Source: E.I. Du Pont de Nemours & Company, Inc.

Slag inclusions appear along the edge of a weld as irregular or continuous dark lines on the radiograph. Voids are sometimes present between weld beads because of irregular deposition of metal during multiple-pass welding. These voids have a radiographic appearance that resembles slag lines. The radiographic image of a crack is a dark narrow line that is generally irregular. If the plane of the crack is in line with the radiation beam, its image is a fairly distinct line. If the plane is not exactly in line with the radiation beam, a faint dark linear shadow results. In this case, additional radiographs should be taken at other angles. Lack of penetration shows on a radiograph as a very narrow dark line near the center of the weld. The narrowness can be caused by drawing together of the plates being welded, and the LOP may be very severe. Slag inclusions and gas holes are sometimes found in connection with LOP and cause the line to appear broad and irregular. The radiographic image of incomplete fusion shows a very thin, straight dark line parallel to and on one side of the weld image. Where there is doubt, additional radiographs should be made with the radiation beam parallel to the bevel face. This will increase the possibility of the LOF appearing on the radiograph. Tungsten inclusions appear either as single light spots or as clusters of small light spots. The spots are usually irregular in shape, but sometimes a rectangular light spot will appear. Real-time radiography, which involves the display of radiographic images on television monitors through the use of

an image converter and a television camera, is a rapidly developing method for weld inspection (Ref 6). One of the main advantages of real-time radiography for weld inspection is the cost savings that results from reducing the use of x-ray films. However, the possibility of expanding such an inspection system to include automatic defect evaluation by the image-processing system can yield significantly greater advantages. Automatic defect evaluation systems will result in objective and reproducible x-ray inspection, independent of human factors. Until now, the human brain has been much faster in analyzing and classifying the large range of flaw types found in welded joints. Computer programs for the efficient automated evaluation of weld radiographs are currently being developed and refined. More detailed information on the application of real-time systems for weld inspection can be found in the article "Radiographic Inspection" in this Volume. Ultrasonic Inspection

In ultrasonic inspection, a beam of ultrasonic energy is directed into a specimen, and either the energy transmitted through the specimen is measured or the energy reflected from interfaces is indicated. Normally, only the front (entry) and back surfaces plus discontinuities within the metal produce detectable reflections, but in rare cases, the HAZs or the weld itself may act as reflecting interfaces. Scanning Techniques. Figure 38 shows how a shear wave from an angle-beam transducer progresses through a flat testpiece--by reflecting from the surfaces at points called nodes. The linear distance between two successive nodes on the same surface is called the skip distance and is important in defining the path over which the transducer should be moved for reliable and efficient scanning of a weld. The skip distance can be easily measured by using a separate receiving transducer to detect the nodes or by using an angle-beam test block, or it can be calculated. Once the skip distance is known, the region over which the transducer should be moved to scan the weld can be determined. This region should

extend the entire length of the weld at a distance from the weld line of approximately to 1 skip distance, as shown in Fig. 39. A zigzag scanning path is used, either with sharp changes in direction (Fig. 39) or with squared changes (Fig. 40).

Fig. 38 Sound beam path in a flat testpiece being ultrasonically inspected with a shear wave from an anglebeam transducer, showing the skip distance between the nodes where the beam reflects from the surfaces

Fig. 39 Three positions of the contact type of transducer along the zigzag scanning path used during ultrasonic inspection of welded joints. The movement of the sound beam path across the weld is shown on a section taken along the centerline of the transducer as it is moved from the far left position in the scanning path (a), through an intermediate position (b), to the far right position (c).

Fig. 40 Ultrasonic scanning procedures to detect longitudinal and transverse discontinuities in welds that (a) are not ground flush and (b) are ground flush

To detect longitudinal discontinuities in full-penetration butt and corner welds that are not ground flush, the transducer is oscillated to the left and right in a radial motion, with an included angle of approximately 30°, while scanning perpendicularly toward the weld, as shown in Fig. 40(a). The longitudinal movement necessary to advance the transducer parallel to the weld should not exceed 75% of the active width of the transducer per transverse scan. The weld should be scanned from both sides on one surface or from one side on both surfaces to ensure that nonvertically oriented flat discontinuities are detected. This type of discontinuity can be distinguished from vertically oriented flat discontinuities because the signal amplitudes from the two sides are different. To detect transverse discontinuities in welds that are not ground flush, the transducer is placed on the base metal surface at the edge of the weld. The sound beam is directed by angling the transducer approximately 15% toward the weld from the longitudinal-weld axis, as shown in Fig. 40(a). Scanning is performed by moving the transducer along the edge of the weld either in one direction along both sides of the weld or in opposite directions along one side of the weld. To detect longitudinal discontinuities in welds that are ground flush, the transducer is oscillated to the left and right in a radial motion, with an included angle of approximately 30°, while scanning across the weld as shown in Fig. 40(b). The longitudinal movement necessary to advance the transducer parallel to the weld must not exceed 75% of the active width of the transducer per transverse scan. When possible, the weld is scanned from one surface on two sides of the weld. When this is not possible, the weld can be scanned from one side on two surfaces or from one side on one surface using at least one full skip distance.

To detect transverse discontinuities in welds that are ground flush, the transducer is oscillated to the left and right in a radial motion, with an included angle of approximately 30°, as shown in Fig. 40(b), while scanning along the top of the weld from two opposing directions. If the width of the weld exceeds the width of the transducer, parallel scans should be performed, with each succeeding scan overlapping the previous one by a minimum of 25% of the active width of the transducer. The entire volume of full-penetration welds in corner joints should be scanned with shear waves by directing the sound beam toward, or across and along, the axis of the weld, as shown in Fig. 41. If longitudinal wave testing is utilized, the weld is scanned by moving the transducer over the weld with overlapping paths. Each succeeding scan should overlap the previous scan by at least 25% of the active width of the transducer.

Fig. 41 Ultrasonic scanning procedure for full-penetration groove weld (a) and double-fillet welds (b) in corner joints

For the detection of discontinuities in the root area in T-joints (such as lack of root fusion), the width of the inspection zone should be limited to the thickness of the attachment member. The width of the inspection zone is located using ultrasonics or mechanical means and marked on the test surface. Shear-wave scanning for discontinuities in the base metal of any T-joint configuration should be performed whenever the surface opposite the attachment member is accessible. This scanning procedure can also be applied to partial-penetration welds in T-joints.

Coverage in each direction begins from the nearest section of the joint to beyond the centerline of the weld. The anglebeam transducer is directed at the particular area of interest and oscillated to the left and right in a radial motion, with an included angle of approximately 30°, while scanning perpendicularly toward the inspection zone. The inspection zone depth should be limited to the through-member-plate thickness minus 6 mm ( in.). The movement necessary to advance the transducer parallel to the inspection zone should not exceed 75% of the active width of the transducer per perpendicular scan. Discontinuity Signals. Cracks and LOF discontinuities present essentially flat reflectors to the ultrasonic beam. If the

beam is perpendicular to the place of the discontinuity, the amplitude of the signal is high; but if the beam strikes the discontinuity at an angle, most of the ultrasonic energy is reflected away from the transducer, and the reflected signal has a small amplitude that will vary with the angle. Because both cracks and sidewall LOF discontinuities produce similar reflected signals, they cannot be distinguished from one another by the signal amplitude or signal shape on the viewing screen when scanning is done from only one side. Therefore, the weld should be inspected from two sides, as shown in Fig. 42. If the discontinuity is vertically oriented, such as a centerline crack would be, the reflected signals received during a scan of each side should have approximately the same amplitude. If the discontinuity is in an inclined position, such as a sidewall LOF discontinuity would be in many joint designs, there will be an appreciable difference between the signal amplitudes.

Fig. 42 Transducer scanning positions for distinguishing between weld metal flaws that are (a) vertically oriented and (b) in an inclined position

A slag inclusion in a butt weld may produce a reflected signal with the same amplitude as that received from a crack or LOF discontinuity. However, scattered ultrasonic energy produces a relatively wide and high signal; as the transducer is manipulated around the slag inclusion, the signal height does not decrease significantly, but the edges of the signal vary. The same shape of reflected signal should be displayed when the weld is scanned from the opposite side of the weldment. The signals that are reflected from porosity (gas pockets) are usually small and narrow. The signal amplitude will vary if the transducer is manipulated around the gas pocket or if the gas pocket is scanned from the opposite side of the weld. Cluster porosity (groups of gas pockets, as shown in Fig. 27) usually produces displays with a number of small signals. Depending on the number of gas pockets and their orientation to the ultrasonic beam, the displayed signals will be stationary or will be connected with one another. Lack of fusion, weld root cracks, and LOP give essentially the same type of signal on an oscilloscope screen; the reflected signals are narrow and appear at the same location. The best way to differentiate among these flaws is to determine the extent of the flaw in the transverse direction.

Weld undercutting is distinguishable from sidewall LOF. The signals reflected from undercutting are approximately equal in amplitude when scanned from both sides. The signals produced by a sidewall LOF discontinuity vary considerably in amplitude when scanned from both sides. In many cases, a weld is made when two misaligned parts must be joined; this condition is termed weld mismatch (Fig. 23 and 24). The inspector must not confuse a signal reflected from a root crack with one reflected from the misaligned edge. A narrow signal is usually produced when the ultrasonic beam strikes the misaligned edge. In most cases, no reflected signal will be received if the misaligned edge is scanned from the opposite side. Ultrasonic Inspection of Spot Welds in Thin-Gage Steel. With the development of high-frequency transducers

(12 to 20 MHz), pulse-echo ultrasonic inspection of spot welds in very thin gage sheet metal (0.58 mm, or 0.023 in.) is now possible (Ref 1). The ultrasonic test for spot weld nugget integrity relies on an ultrasonic wave to measure the size of the nugget. The size is in three dimensions, including thickness as well as length and breadth (or diameter for a circular spot). The successful measurement of nugget size places several requirements on the ultrasonic wave path, wave velocity, and wave attenuation. Wave Path. The first requirement is that the ultrasonic wave be in the form of a beam directed perpendicular to the

faces of the metal sheets and through the center of the nugget (Fig. 43). Two diameters of nuggets are shown: larger than the beam and smaller than the beam.

Fig. 43 Schematic illustrating setup for the pulse-echo ultrasonic inspection of resistance-welded spot welds. (a) Wave paths in satisfactory weld. (b) Resulting echoes. (c) Wave paths in an unsatisfactory weld. (d) Resulting echoes. Source: Ref 1

In general, an ultrasonic wave will be reflected when it impinges on an interface where the density and/or the ultrasonic velocity change. Examples are water-to-metal and metal-to-air. In Fig. 43, reflections will occur at the outer surfaces of the two sheets and at the interface (air) between the two sheets if the nugget is small, as in Fig. 43(c). The nugget-toparent-metal boundary will not produce perceptible reflections, refraction, or scattering, because the changes in density and velocity are a tenth of a percent or less, while the air-to-steel difference is more than 99.9%. Typical oscilloscope displays showing the pulse-echo patterns for these two nugget-to-beam-diameter ratios are shown in Fig. 43(b) and 43(d). The difference in the echo patterns permits the distinction to be made between adequate and undersize welds.

Velocity/Thickness Gaging. The beam path shown in Fig. 43(a) illustrates the situation in which the ultrasonic beam

should indicate an acceptable nugget. The beam will be reflected only at the outer surfaces (1 and 3) of the pair of sheets as joined. To make this reflection sequence visible, the ultrasonic beam must consist of a short pulse that can reverberate back and forth between the outer faces and produce separate echoes when viewed on an oscilloscope. The picture observed is shown in Fig. 43(b). The pulse must be short enough to resolve the double thickness of the two joined sheets. Similarly, the beam path shown in Fig. 43(c) illustrates the situation in which the ultrasonic beam should indicate an undersize nugget. The beam will be reflected in the single thickness of the upper sheet around the perimeter of the nugget. Therefore, on the oscilloscope, echoes will appear between the principal echoes arising from the portion of the beam traversing the nugget (Fig. 43d). In terms of thickness gaging, the ultrasonic pulse in the beam must be short enough to resolve the thickness of one layer of sheet metal. Attenuation. The thickness of the nugget can be measured only indirectly because the thickness gaging function

(described above) can measure only the thickness between outer faces in the nugget area. The nugget itself is measured by the effect of its grain structure on the attenuation of the ultrasonic wave in the beam. As the wave reflects back and forth between the outer faces of the welded sheets, its amplitude is attenuated (dies out). The attenuation (rate of decay) of the ultrasonic wave depends on the microstructure of the metal in the beam. In the spot welds under consideration, the attenuation is caused principally by grain scattering. The grains scatter the ultrasonic energy out of the coherent beam, causing the echoes to die away. In most metals, coarse grains scatter more strongly than fine grains. Because a nugget is a melted and subsequently refrozen cast microstructure with coarser grains than the adjacent coldrolled parent metal, the nugget will scatter more strongly than the remaining parent metal. It follows that a nugget will produce higher attenuation than the parent metal and that a thick nugget will result in higher attenuation than will a thin nugget. Therefore, a thin nugget can be distinguished from a thick nugget by the rate of decay of the echoes in the case in which the diameters of both nuggets are equal. Typical echoes from a thick nugget area and from a thin nugget are shown in Fig. 44. It is clear that a trained observer could differentiate between the two welds on the basis of the decay patterns. Given this observation, it is obvious that the pulse-echo ultrasonic method at normal incidence could perform the required measurements on spot welds in metals with coarse-grain nuggets and fine-grain parent sheet metal.

Fig. 44 Ultrasonic thickness measurements of resistance spot weld nuggets. (a) Satisfactory weld. (b) Resulting attenuation of the ultrasonic wave. (c) Unsatisfactory weld. (d) Resulting wave attenuation. Source: Ref 1

Acoustic Emission Monitoring

Acoustic emissions are impulsively generated small-amplitude elastic stress waves created by deformations in a material. The rapid release of kinetic energy from the deformation mechanism propagates elastic waves from the source, and these are deteced as small displacements on the surface of the specimen. The emissions indicate the onset and continuation of deformation and can be used to locate the source of deformation through triangulation techniques. Details of the process are available in the article "Acoustic Emission Inspection" in this Volume. Acoustic emissions can be used to assess weld quality by monitoring during or after welding. In weldments, regions having LOP, cracking, porosity, inclusions, or other discontinuities can be identified by detecting the acoustic emissions originating at these regions. During welding processes, acoustic emissions are caused by many factors, including plastic deformation, melting, friction, solidification, solid-solid phase transformations, and cracking. In some cases, the monitoring of acoustic emissions during welding can include automatic feedback control of the welding process. In largescale automatic welding, the readout equipment can be conveniently located near the welder controls or in a qualitymonitoring area. The locations of acoustic sources on a weld line can be presented in a variety of ways. One technique displays the

number of events versus distance along the weld on an oscilloscope screen or an x-y plotter. Another technique uses a digital-line printer that gives the time of the event, its location, and its intensity. This information facilitates appraisal of the severity of each source. Once the acoustic emission sources are graded, other NDI methods can be used to study the discontinuity in detail. It is sometimes difficult to achieve a good acoustic coupling between the sensor (or sensors) and the part. This is especially true for welding processes that have a fixed arc and a moving part. For postweld acoustic emission monitoring, a stimulus must be applied after complete cooling of the weld. This involves the application of a mechanical load or sometimes a thermal stress to the part or structure. For the field inspection of in-service welds for cracks or other flaws, welds are commonly subjected to stresses that just exceed the maximum stress previously experienced by the weld. The excess stress is necessary to produce acoustic emissions due to plastic deformation at the crack tip. The application of this stress, however, is often difficult to do and in some cases is undesirable. An alternative method that can often clearly detect existing cracks in a weld is to cyclically load the weld at low-stress levels. The relative motions of the crack surfaces produce frictional excitation of stress waves. Acoustic emission monitoring for evaluation of quality and control of welding processes requires preliminary studies for each application to establish such operating conditions as the number, location, and mounting of sensors; gain settings; filtering; data presentation; and data interpretation. These studies normally include correlation with other nondestructive and destructive methods of inspection. For example, the detection of tungsten inclusions from GTAW joints by acoustic emission can also involve inclusion classification by radiography. The postweld monitoring of weldments includes both quality control inspection during the period between the

completion of the weld and additional fabrication of the part and nondestructive inspection of in-service weldments. An example of the former is the inspection of butt-welded plates, such as those fabricated in the building of a ship. The following example demonstrates the feasibility of immediate postweld acoustic emission monitoring of butt-welded plates. Example 1: Acoustic Emission Monitoring of Butt Welds in Low-Carbon Steel Plates. Using a dc arc welder, two low-carbon steel plates 460 mm (18 in.) wide by 610 mm (24 in.) long and 3.2 mm (0.125 in.) thick were butt welded to form a test plate approximately 1.2 m (48 in.) long, as shown in Fig. 45. Extra precautions were taken to produce a sound weld. After the weld metal was cool, the plate was placed in a fixture so that acoustic emission from the weld metal could be monitored.

Fig. 45 Acoustic emission monitoring of butt welds in low-carbon steel test plates. (a) Test plate. Three were prepared; the first, with a sound weld, was used to establish conditions for acoustic emission monitoring of the others. (b) Location and number of acoustic emissions in the second test plate, which had a region of intentional poor penetration and slag inclusions (shaded area). (c) Location and number of acoustic emissions in the third test plate, in which flaw indications were revealed by radiographic inspection (shaded areas).

The plate was supported horizontally by a system of rubber rollers. Weights were placed on each end of the plate to hold it firmly on the rollers. Mild bending was induced in the butt joint by raising and lowering the middle of the plate with a hydraulic jack. Acoustic emissions were monitored during loading and unloading oscillation and while the plate was held in a stress state, using sensors attached to the top side of the plate at each end of the weld. These emissions were used to calibrate the oscilloscope screen and were the basis for subsequent inspections of plates having known or suspected discontinuities in the weld metal.

The output signal of each sensor was amplified (75-dB gain), filtered (50 kHz to 1 MHz), and displayed on a dual-beam oscilloscope. Single-sweep traces of both signals were triggered at the arrival of the first signal. These were photographed from the dual-beam display, and successive photographs permitted documentation of the difference between the arrival times at the two sensors for several sources. The locations of the predominant sources were then inferred. After the preliminary experiments, a second test plate was prepared and butt welded across its entire width. The weld was made with intentional poor penetration and slag inclusions in a region centered about 100 mm (4 in.) on one side of the centerline of the plate, as indicated by the shaded area in Fig. 45(b). Acoustic emissions were recorded during bending and oscillation. The location and number of acoustic emissions are given graphically in Fig. 45(b). A good correlation existed between the tabulated source locations and the locations of known flaws. A third test plate was prepared by making a saw cut 180 mm (7 in.) long, from one edge toward the center, along the transverse centerline. This saw cut was repaired by welding, then monitored for acoustic emissions in the same manner as the second test plate. Radiographic inspection of the plate revealed two regions of discontinuities in the weld, as indicated by the shaded areas in Fig. 45(c). A region of very poor penetration between 130 and 150 mm (5 and 6 in.) from the longitudinal centerline of the plate is shown by the darker shaded area in Fig. 45(c). This corresponds to the large number of acoustic emissions occurring in the region. Also shown in Fig. 45(c) are other regions from which acoustic emissions originated, indicating discontinuities in the weld metal. From the results obtained on these test plates, it was concluded that it is feasible to use acoustic emission monitoring as a method of assessing the structural integrity of butt-welded joints. In-Service Monitoring. One application of in-service acoustic emission monitoring of welds involved the locating of defective or deteriorated welds in buried pipelines. Gas distribution pipelines that had been in service for many years needed inspections for structural integrity, especially on oxyacetylene welds. Although the location of the buried pipe was known, the locations of the welds were not. In preliminary studies, it was found that acoustic emission signals from weld discontinuities would propagate several hundred feet down the pipe. Therefore, a very efficient method of locating the defective welds was devised. The loading stimulus, an extra heavy vehicle moving slowly along the surface above the pipe, induced bending stresses that were sufficient to cause weld discontinuities to emit. Sensors were placed on either side of the suspect weld, and the signals reaching the sensors were monitored with portable equipment. By comparing the time required for the emissions to reach each pair of sensors and by comparing the intensities of the signals, the locations of the defective weld were determined. The welded joints were then excavated, and the welds were further inspected or repaired. Monitoring During Welding. Arc-welding processes are inherently ultrasonically noisy--particularly so in continuous

high-frequency welding. However, the acoustic emissions detected during proper welding without discontinuity creation have steady characteristics. When cracking, excessive slag inclusion, or a significant change in the weld conditions occurs, the acoustic emission levels change correspondingly. Therefore, online monitoring during welding gives immediate indication of variations in the quality of the weld. Cold cracking can be detected by monitoring the welded structure for minutes, or even hours, after welding. Acoustic emissions result from multiple causes during resistance spot welding. The making of a resistance spot weld consists of setdown of the electrodes, squeeze, current flow, hold time, and lift off. Many acoustic emissions are produced during these various steps. The most commonly observed signals are shown schematically in Fig. 46. The ultrasonic noise during setdown and squeeze can be related to the conditions of the electrodes and the surface of the parts. The large, but brief, signal at current initiation can be related to the initial resistance and the cleanliness of the parts. During current flow, acoustic emission results from plastic deformation, friction, melting, and expulsions. The signals associated with expulsion (spitting and flashing) are generally large in amplitude and can be easily distinguished from the rest of the emissions associated with nugget formation. When current flow ceases, some materials exhibit appreciable solidification noise that can be related to nugget size and inclusions. As the nugget cools in the hold period, acoustic emissions can result from solid-solid phase transformations and cracking. Finally, as the electrode is lifted, noise is produced by the separation of the electrode from the part. This noise, or signal, can be related to the size of the nugget as well as to the visual appearance of the weld.

Fig. 46 Schematic showing typical acoustic emission signals obtained during various stages of resistance spot welding

Any measure of the cumulative acoustic emissions during resistance spot welding cannot be expected to relate clearly to weld quality. On the other hand, by using both time discrimination and multiple detection levels, the various segments of acoustic emissions can be separately measured and related to various indicators of quality. Commercial instrumentation is available that is capable of separately monitoring several of the acoustic emission segments. For example, the expulsion count, phase-transformation count, and cracking count can be monitored and recorded for each resistance spot weld, giving a permanent record of quality. In the following example, a number of acoustic emissions during martensitic phase transformation were found to relate to the strength of the resistance spot weld. Example 2: Determination of Strength of Resistance Spot Welds by Acoustic Emission Monitoring. Twenty carbon steel coupons were spot welded at identical settings of heat and weld cycles in a 75-kVA ac spot welder. An acoustic emission sensor was attached to the lower electrode. The acoustic emission monitor was gated to begin 120 ms after the current ceased and to stop 240 ms later. Within that time interval, martensitic phase transformation occurred as the nugget cooled. Because the total volume of material experiencing the transformation is related to both the nugget size and the area of diffusion bonding, the acoustic emissions during the phase transformation are related to the strength of the weld. The specimens were pulled to failure in tension shear, and the ultimate strength was recorded. The results are shown in Fig. 47. The tests indicated that a 10% variation in weld strength corresponded to about a 30% change in acoustic emissions.

Fig. 47 Relation of strength of resistance spot welds in carbon steel coupons to number of acoustic emissions

during the martensitic phase transformation

Leak Testing Welded structures are leak tested to measure the integrity of the structure for containing gases, fluids, semisolids, and solids and for maintaining pressures and vacuums. The more common leak-testing methods used are (in order of increasing sensitivity): • • • • • •

Odor from tracer gas Pressure change Pressurized liquid (generally water) and visual observation Pressurized gas using a leak detection solution Tracer gas using thermal leak detectors Helium using a mass spectrometer during pressure and vacuum tests

Other methods less frequently used are acoustical detection of gas flow through a leak and use of radioactive tracer gas. Detailed descriptions of the various methods used are presented in the article "Leak Testing" in this Volume. Weld flaws that contribute to leakage of a structure are porosity, LOF or LOP, and cracks. Cracks are of particular concern because they may propagate when the structure is proof tested or otherwise tested for structural integrity. Therefore, it is preferred that leak testing be done after completion of the structural tests. Selection of a leak testing method depends on the environment in which the structure is used and the potential danger and economic impact involved in the event of a service failure. The acceptance criteria should include a numerical expression of the allowable leak rate; the frequently used expression "shall be free from leaks" is meaningless. Eddy Current and Electric Current Perturbation Inspection Eddy current inspection is based on the principles of electromagnetic induction and is used to identify or to

differentiate between a wide variety of physical, structural, amd metallurgical conditions in electrically conductive ferromagnetic and nonferromagnetic metals. Normally, eddy current inspection is used only on thin-wall welded pipe and tubing for the detection of longitudinal-weld discontinuities, such as open welds, weld cracks, and porosity. The application of this method to tubular goods is discussed in the articles "Eddy Current Inspection," "Tubular Products," and "Boilers and Pressure Vessels" in this Volume. The electric current perturbation method consists of establishing an electric current flow in the part to he

inspected (usually by means of an induction coil) and detecting the magnetic field associated with perturbations in the current flow around defects by using a separate magnetic field sensor (Ref 7). This technique is applicable to the detection of both very small surface cracks as well as subsurface cracks in both high- and low-conductivity, nonferromagnetic materials, such as titanium and aluminum alloys. The principles, equipment, and applications associated with this method are outlined in the article "Electric Current Perturbation NDE" in this Volume.

References cited in this section

1. T.M. Mansour, Ultrasonic Inspection of Spot Welds in Thin-Gage Steel, Mater. Eval., Vol 46 (No. 5), April 1988, p 650-658 2. P. Kapranos and R. Priestner, NDE of Diffusion Bonds, Met. Mater., Vol 3 (No. 4), April 1987, p 194-198 3. P.G. Partridge, "Diffusion Bonding of Metals," Advisory Group for Aerospace Research and Development (NATO), Aug 1987, p 5.1-5.23 4. G. Tober and S. Elze, "Ultrasonic Testing Techniques for Diffusion-Bonded Titanium Components," Advisory Group for Aerospace Research and Development (NATO), July 1986, p 11.1-11.10 5. Diffusion Welding and Brazing, Vol 3, 7th ed., Welding Handbook, American Welding Society, 1980, p 311335

6. J.J. Munro, R.E. McNulty, and W.Nuding, Weld Inspection by Real-Time Radiography, Mater. Eval., Vol 45 (No. 11), Nov 1987, p 1303-1309 7. G.L. Burkhardt and B.N. Ranganathan, Flaw Detection in Aluminum Welds by the Electric Current Perturbation Method, in Review of Progress in Quantitative Nondestructive Evaluation, Vol 4A, Plenum Press, 1985, p 483-490 Note cited in this section

* The reference radiographs discussed in this section and shown in Fig. 18 to 37 were generated from singleV-groove weld samples fabricated from 9.5 mm (0.375 in.) thick carbon steel plates using gas metal arc welding, shielded metal arc welding, and gas tungsten arc welding. Nondestructive Inspection of Weldments, Brazed Assemblies, and Soldered Joints

Brazed Assemblies Brazing is defined by the American Welding Society as a group of welding processes that produce coalescence of materials by heating them to a suitable temperature and by using a filler metal having a liquidus above 450 °C (840 °F) and below the solidus of the base metal. The filler metal is distributed between the closely fitted faying surfaces of the joint by capillary action. The temperature limitation of 450 °C (840 °F) differentiates brazing from soft soldering, which involves the use of filler metals having a liquidus below 450 °C (840 °F). To clarify the difference between brazing and conventional welding, it should be pointed out that in brazing the base materials being joined are never melted, while in most welding processes the base metals are melted (exceptions are those welding processes that utilize pressure in conjunction with heat). There are six brazing processes included under the group heading of brazing. These processes are torch brazing, furnace brazing, induction brazing, dip brazing, resistance brazing, and infrared brazing. Brazing can also be classified according to the major constituents of the more common types of filler metals used: • • • • •

Aluminum brazing Silver brazing Copper brazing Nickel brazing Precious metal brazing

There are five essential properties for brazing filler metal: • • •

• •

Ability to wet and make a strong, sound bond on the base metal Suitable melting temperature and flow properties to permit distribution by capillary attraction in properly prepared joints A composition of sufficient homogeneity and stability to minimize separation by liquation under the brazing conditions to be encountered. Excessively volatile constituents in filler metals may be objectionable Capability of producing a brazed joint that will meet service requirements, such as required strength and corrosion resistance Depending on the requirements, ability to produce or avoid interactions between base metal and filler metal

Flaws Commonly Found in Brazed Joints

Flaws exhibited by brazed joints are usually of the following types: • • • •

Lack of fill Flux entrapment Noncontinuous fillet Base metal erosion

Lack of Fill. Voids resulting from lack of fill can be the result of improper cleaning of the faying surfaces, improper

clearances, insufficient brazing temperatures, or insufficient brazing filler metal (Fig. 48).

Fig. 48 Voids resulting from lack of fill between the faying surfaces of a lap joint between two sheets of Hastelloy X brazed with BNi-1 filler metal. Unetched. 16

×.

Flux entrapment normally occurs during torch brazing, induction brazing, or furnace brazing, when reducing

atmospheres are not employed. As the term implies, flux becomes trapped within the joint by the braze metal and prevents coverage. Figure 49 is a radiograph of a torch-brazed joint in which flux entrapment was a serious problem.

Fig. 49 Radiograph showing entrapped flux (dark areas) in a low-carbon steel joint torch brazed with BAg-1 filler metal (light areas). 1×

Noncontinuous Fillet. A brazed joint in which a large void in the fillet is evident is shown in Fig. 50. Such a void is

discernible by visual examination and may or may not be acceptable, depending on requirements. For example, if the void in the fillet did not extend through the entire braze width, the joint would still be leaktight, which was the major requirement of the brazement. On the other hand, if 100% braze fillet was needed because of stress requirements, the assembly would be unacceptable.

Fig. 50 Incomplete penetration of filler metal (BAg-1) in a brazed joint between copper components. 20×

Base Metal Erosion. Certain brazing filler metals will readily alloy with the base metals being brazed, causing the

constituents of the base metal to melt and, in some cases, creating an undercut condition or the actual disappearance of the faying surfaces. This is called base metal erosion. Extreme erosion in type 304 stainless steel brazed with a nickelchromium-boron filler metal is shown in Fig. 51; a similar joint without erosion is shown in Fig. 52. Erosion may not be serious where thick sections are to be joined, but it cannot be permitted where relatively thin sections are used.

Fig. 51 Excessive erosion of type 304 stainless steel base metal by BNi-1 filler metal. Compare with the noneroded joint shown in Fig. 52. 20×

Fig. 52 Joint between type 304 stainless steel components brazed with BNi-1 filler metal, in which no base metal erosion occurred. Note characteristic sheared edge on one component and small voids in the filler metal.

50×

Three factors influencing base metal erosion are brazing temperature, time at temperature, and the amount of brazing filler metal available or used in making the joint. As the brazing temperature exceeds the melting point of the filler metal, interaction between the molten filler metal and the base metal accelerates. The brazing temperature should therefore be kept low, provided, of course, that it is sufficient for proper flow of the filler metal to fill the joint. Similarly, time at temperature should be kept to a minimum to prevent excessive interaction between the molten filler metal and the base metal. Finally, the amount of filler metal required to fill the joint and provide the necessary fillet size should be closely controlled. Filler metal present in excess of the amount required is likely to react with the base metal, creating severe or excessive erosion in proportion to the amount of excess filler metal. Joint Integrity Some form of discontinuity usually occurs in all types of brazed joints. The degree and severity vary from a minor pinhole in the filler metal to gross discontinuities. Lack of fill or flux entrapment can vary from slight to nearly 100%. Erosion of the base metal can be nonexistent or can cause complete destruction of the joint. Requirements for brazed joints are many and varied. As with other accepted joining processes, it is important that brazed joints be properly designed and engineered for the use intended. Significant factors involved are selection of proper base metals and brazing filler metal for compatibility and strength, proper fits and clearances, proper brazing process, and cleanliness of the surface to be brazed. Furthermore, it must be determined what requirements are necessary for withstanding the service conditions to which the finished brazement will be exposed. Primarily, brazed joints are designed for mechanical performance, electrical conductivity, or pressure-tightness. The braze quality requirements, therefore, should reflect the end use for which the joint was designed. Methods of Inspection Inspection of the completed assembly or subassembly is the last step in the brazing operation and is essential for ensuring satisfactory and uniform quality of the brazed unit. This inspection also provides a means for evaluating the adequacy of the design and the brazing operation with regard to ultimate integrity of the brazed unit. Destructive methods such as peel tests, impact tests, torsion tests, and metallographic examination are initially used to determine whether the braze design meets the specified requirements. In production, these methods are employed only with random selection or lot testing of brazed joints. In lot testing, samples representing a small specified percentage of all production are tested to destruction. The results of these tests are assumed to apply to the entire production, and the joints in the various lots or batches are accepted or rejected accordingly. When used as a check on an NDI method, such as visual examination, a production part can be selected at regular intervals and the joint tested to destruction so that rigid control of brazing procedures is maintained. The inspection method chosen to evaluate the final brazed component should depend on service and reliability requirements. In many cases, the inspection methods are specified by the ultimate user or by regulatory codes. In establishing codes or specifications for brazed joints, the same approach should be used as in the setting up of standards for any other phase of manufacturing. The standards should be based, if possible, on requirements that have been established by prior service or history. Visual Inspection Visual inspection is the most widely used of the nondestructive methods for evaluating brazed joints. However, as with all other methods of inspection, visual inspection will not be effective if the joint cannot be readily examined. Visual inspection is also a convenient preliminary test where other inspection methods are used. When brazing filler metal is fed from one side of the joint or replaced within the joint at or near one side so that visual examination of the opposite side of the joint after brazing shows a continuous fillet of filler metal, it can usually be assumed that the filler metal has flowed through the joint by capillary attraction and that a sound joint has been obtained. On the other hand, if the joint can be inspected only on the side where the filler metal is applied, it is quite possible that an unsatisfactory joint has been produced, even though a satisfactory fillet is in evidence to the inspector.

Visual inspection cannot reveal internal discontinuities in a brazed joint that result from flux entrapment or lack of fill. Occasionally, gross erosion can be detected. Proof Testing Proof testing is a method of inspection that subjects the completed joints to loads slightly in excess of the loads to be applied during their subsequent service life. These loads can be applied by hydrostatic methods, tensile loading, spin testing, or numerous other methods. Occasionally, it is not possible to ensure a serviceable part by any of the other nondestructive methods of inspection, and proof testing then becomes the most satisfactory method. Pressure Testing Pressure testing of brazed assemblies is a method of leak testing and is usually confined to vessels and heat exchangers where liquid, gas, or air tightness is required. Several methods of pressure testing can be employed. Most use either air or gas, depending on the application of the vessel or heat exchanger. In most cases, the test pressures are greater than those to which the assembly will be subjected in service and are specified by the user. One or more of the following three procedures are generally employed for pressure testing: •

•

•

All openings are closed. Air or gas is injected into the assembly until the specified pressure is reached. The inlet sources are closed off, the assembly is allowed to sit for a period of time, and pressure decreases are then measured on a gage All openings in the assembly are closed except one, which is fitted with an inlet-pressure line. With the assembly submerged in a tank of water, air or gas is admitted through the inlet line until a specified pressure is reached. The inspector then looks for bubbles rising through the water All openings are closed, and the assembly is pressurized to the specified pressure. Then a leak-detecting solution, of which there are several commercially available, is brushed on the joints to be inspected. If any of the joints leak, bubbling will occur

Vacuum-and-helium testing is generally used in inspecting assemblies where it is imperative that the most minute

leak be detected. This method of inspection is often employed on nuclear reactor hardware. It is also extensively used in the inspection of refrigeration equipment. The assembly to be inspected is connected to a vacuum system, and the vacuum is monitored by a mass spectrometer. Helium gas is flushed around the brazed joint; if any minute leak is present, the helium, because of its small molecule, will be pulled in by vacuum and register on the mass spectrometer, thus indicating the leak. A more sensitive technique is pressurizing the assembly with helium while the assembly is contained in a sealed plastic bag. After pressurizing for a period of time (for example, 24 h), the atmosphere in the bag is analyzed for the presence of helium. Additional information on pressure testing techniques can be found in the article "Leak Testing" in this Volume. Ultrasonic Inspection Ultrasonic inspection, although not extensively used in the evaluation of brazed joints, can be the only method applicable in certain cases. The use of ultrasonic inspection depends largely on the design of the joint and the configuration of the adjacent areas of the brazed assembly. Advancements in ultrasonic inspection may increase the utility of this process so that brazed joints can be evaluated with reliability. Radiographic Inspection Radiographic inspection is commonly used for the nondestructive evaluation of brazed joints following visual examination. In almost all cases, the radiation beam is directed at about 90° to the plane of the joint, and the radiograph is taken through the thickness of the braze metal.

X-rays readily discern the differences in density between the brazing filler metal and the base metal. Care must be exercised, however, because joints between sections of varying thicknesses can produce radiographs that are misleading and difficult to interpret. Also, it is often difficult to determine whether a joint has been penetrated fully or not at all; both situations yield radiographs in which there is a full fillet visible around the joint, and the gap in the joint itself has uniform radiographic density. By contrast, partly filled joints, voids in the braze metal, and inclusions are relatively easy to find with radiography. The filler metal in brazed joints is very thin--from 0.013 to 0.25 mm (0.0005 to 0.010 in.) in thickness. When radiographs are made of brazed joints between thick components, the process may be unable to record the braze metal as a difference in density; at least 2% difference is usually needed for good sensitivity. Liquid Penetrant Inspection Liquid penetrant inspection is another nondestructive method for determining the reliability of brazed joints and assemblies. This inspection method produces a visual image of a discontinuity in the surface of the braze and reveals the nature of a discontinuity without impairing the parent metal. Acceptable and unacceptable components or assemblies can be separated in accordance with predetermined standards. There are certain advantages obtained from the liquid penetrant inspection of brazed assemblies. However, a brazed joint or component should be visually inspected first, then inspected by a liquid penetrant method to resolve any doubt concerning joint integrity. Visual examination is restricted to those discontinuities that can be detected by the unaided eye. Liquid penetrant carries visual inspection a step further by increasing the detectability of fine cracks or openings. Discontinuities such as LOF, cracks that x-rays cannot show because of orientation, and porosity and laps become visible with this technique. Liquid penetrants do not disclose subsurface discontinuities such as voids, cracks, or flux entrapment; radiography is best used to discover such discontinuities. Selection of the specific liquid penetrant system for the inspection of brazed assemblies depends on the same factors as those that affect system selection for other workpieces. The water-washable, postemulsifiable, and solvent-removable systems have been successfully used for inspecting brazed assemblies. Inspection using liquid penetrants should not be performed prior to brazing unless adequate cleaning steps, such as vapor degreasing, are taken to remove entrapped penetrant fluid. If permitted to remain during the brazing cycle, this fluid can contaminate the furnace atmosphere and braze metal, producing flaws. Inspection With Thermally Quenched Phosphors Inspection with thermally quenched phosphors is a means of nondestructively detecting flaws such as voids and unbonded areas of laminated honeycomb brazed structures where thin (R8

>L54

>L40

>L25

>L10

3814

>R10

>L48

>L36

>L20

>L9

0.063

3814

>R12

>L42

>L30

>L18

>L8

1.8

0.071

3814

>R16

>L40

>L28

>L18

>L7

2.1

0.081

3814

>R18

>L38

>L25

>L15

>L6

mm

in.

1.0

0.040

3814

1.3

0.050

1.6

(a) Skin/stringer and skin/frame bonds shall be Class minimum; doubler/skin bonds shall be Class A/B minimum. Additional adhesive layers shall not yield readings less than the lower limits by the following factors: 2 layers (-15); 3 layers (-25); 4 layers (-30).

(b) R, right-hand side.

(c) L, left-hand side

References cited in this section

5. K.J. Rienks, "The Resonance/Impedance and the Volta Potential Methods for the Nondestructive Testing of Bonded Joints," Paper presented at the Eighth World Conference on Nondestructive Testing, Cannes, France, 1976 10. D. Hagemaier and R. Fassbender, Nondestructive Testing of Adhesive Bonded Structures, SAMPE Q., Vol 9 (No. 4), 1978 23. D.J. Hagemaier, Bonded Joints and Non-Destructive Testing: Bonded Honeycomb Structures, 2, Non-Destr. Test., Feb 1972

36. B.G. Martin, et al., Interference Effects in Using the Ultrasonic Pulse-Echo Technique on Adhesive Bonded Metal Panels, Mater. Eval., Vol 37 (No. 5), 1979 37. P.L. Flynn, Cohesive Bond Strength Prediction for Adhesive Joints, J. Test. Eval., Vol 7 (No. 3), 1979 38. "Sandwich Construction and Core Materials; General Test Methods," MIL-STD-401 Nondestructive Inspection of Adhesive-Bonded Joints* Donald J. Hagemaier, Douglas Aircraft Company, McDonnell Douglas Corporation

References 1. 2. 3. 4. 5.

6. 7. 8.

9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.

M.T. Clark, "Definition and Non-Destructive Detection of Critical Adhesive Bond-Line Flaws," AFMLTR-78-108, U.S. Air Force Materials Laboratory, 1978 R.J. Schliekelmann, Non-Destructive Testing of Adhesive Bonded Metal-to-Metal Joints, Non-Destr. Test., April 1972 R.J. Schliekelmann, Non-Destructive Testing of Bonded Joints--Recent Developments in Testing Systems, Non-Destr. Test., April 1975 P. Bijlmer and R.J. Schliekelmann, The Relation of Surface Condition After Pretreatment to Bondability of Aluminum Alloys, SAMPE Q., Oct 1973 K.J. Rienks, "The Resonance/Impedance and the Volta Potential Methods for the Nondestructive Testing of Bonded Joints," Paper presented at the Eighth World Conference on Nondestructive Testing, Cannes, France, 1976 J.C. Couchman, B.G.W. Yee, and F.M. Chang, Adhesive Bond Strength Classifier, Mater. Eval., Vol 37 (No. 5), April 1979 J. Rodgers and S. Moore, Applications of Acoustic Emission to Sandwich Structures, Acoustic Emission Technology Corporation, 1980 J. Rodgers and S. Moore, "The Use of Acoustic Emission for Detection of Active Corrosion and Degraded Adhesive Bonding in Aircraft Structures," Sacramento Air Logistics Center (SM/ALC/MMET), McClellan Air Force Base, 1980 The Sign of a Good Panel Is Silence, Aviat. Eng. Maint., Vol 3 (No. 4), April 1979 D. Hagemaier and R. Fassbender, Nondestructive Testing of Adhesive Bonded Structures, SAMPE Q., Vol 9 (No. 4), 1978 R. Botsco, The Eddy-Sonic Test Method, Mater. Eval., Vol 26 (No. 2), 1968 J.R. Kraska and H.W. Kamm, "Evaluation of Sonic Methods for Inspecting Adhesive Bonded Honeycomb Structures," AFML-TR-69-283, U.S. Air Force Materials Laboratory, 1970 N.B. Miller and V.H. Boruff, "Evaluation of Ultrasonic Test Devices for Inspection of Adhesive Bonds," Final Report ER-1911-12, Martin Marietta Corporation, 1962 J. Arnold, "Development of Non-Destructive Tests for Structural Adhesive Bonds," WADC-TR-54-231, Wright Air Development Center, 1957 R. Schroeer, et al., The Acoustic Impact Technique--A Versatile Tool for Non-Destructive Evaluation of Aerospace Structures and Components, Mater. Eval., Vol 28 (No. 2), 1970 H.M. Gonzales and C.V. Cagle, Nondestructive Testing of Adhesive Bonded Joints, in Adhesion, STP 360, American Society for Testing and Materials, 1964 D.F. Smith and C.V. Cagle, Ultrasonic Testing of Adhesive Bonds Using the Fokker Bond Tester, Mater. Eval., Vol 24, July 1966 H.M. Gonzales and R.P. Merschell, Ultrasonic Inspection of Saturn S-II Tank Insulation Bonds, Mater. Eval., Vol 24, July 1966 R.J. Botsco, Nondestructive Testing of Composite Structures With the Sonic Resonator, Mater. Eval., Vol 24, Nov 1966

20. R. Newschafer, Assuring Saturn Quality Through Non-Destructive Testing, Mater. Eval., Vol 27 (No. 7), 1969 21. "Fokker Ultrasonic Adhesive Bond Test," MIL-STD-860, 1978 22. R.J. Schliekelmann, Non-Destructive Testing of Adhesive Bonded Joints, in Bonded Joints and Preparation for Bonding, AGARD-NATO Lecture Series 102, Advisory Group for Aerospace Research and Development, 1979 23. D.J. Hagemaier, Bonded Joints and Non-Destructive Testing: Bonded Honeycomb Structures, 2, NonDestr. Test., Feb 1972 24. D. Wells, NDT of Sandwich Structures by Holographic Interferometry, Mater. Eval., Vol 27 (No. 11), 1969 25. R.M. Grant, "Conventional and Bead to Bead Holographic Non-Destructive Testing of Aircraft Tires," Paper presented at the 1979 ATA NDT Forum, Seattle, WA, Sept 1979 26. P.R. Vettito, A Thermal I.R. Inspection Technique for Bond Flaw Inspection, J. Appl. Polym. Sci. Appl. Polym. Symp. No. 3, 1966 27. C. Searles, Thermal Image Inspection of Adhesive Bonded Structures, in Proceedings of the Symposium on the NDT of Welds and Materials Joining, 1968 28. W. Woodmansee and H. Southworth, Detection of Material Discontinuities With Liquid Crystals, Mater. Eval., Vol 26 (No. 8), 1968 29. S. Brown, Cholestric Crystals for Non-Destructive Testing, Mater. Eval., Vol 26 (No. 8), 1968 30. D.H. Collins, Acoustical Holographic Scanning Techniques for Imaging Flaws in Reactor Pressure Vessels, in Proceedings of the Ninth Symposium on NDE (San Antonio), April 1973 31. A Sample of Acoustical Holographic Imaging Tests, Holosonics Inc. 32. "Adhesive Bonding (Structural) for Aerospace Systems, Requirements for," MIL-A-83377 33. "Adhesive Bonded Aluminum Honeycomb Sandwich Structure, Acceptance Criteria," MIL-A-83376 34. "Inspection Requirements, Nondestructive: For Aircraft Materials and Parts," MIL-I-6870 (ASG) 35. "Nondestructive Testing Personnel Qualification and Certification," MIL-STD-410 36. B.G. Martin, et al., Interference Effects in Using the Ultrasonic Pulse-Echo Technique on Adhesive Bonded Metal Panels, Mater. Eval., Vol 37 (No. 5), 1979 37. P.L. Flynn, Cohesive Bond Strength Prediction for Adhesive Joints, J. Test. Eval., Vol 7 (No. 3), 1979 38. "Sandwich Construction and Core Materials; General Test Methods," MIL-STD-401 Nondestructive Inspection of Boilers and Pressure Vessels

Introduction DURING THE FABRICATION of a boiler, pressure vessel, and such related components as boiling water reactor piping or steam generator tubes, various types of nondestructive inspection (NDI) are performed at several stages of processing, mainly for the purpose of controlling the quality of fabrication. In-service inspection is used to detect the growth of existing flaws or the formation of new flaws. This can be done while the vessel is in operation or down for servicing. The inspection methods used include visual, radiographic, ultrasonic, liquid penetrant, magnetic particle, eddy current, and acoustic emission inspection, as well as replication microscopy and leak testing. The assurance of component quality depends largely on the adequacy of NDI equipment and procedures and on the qualification of personnel conducting the inspection. In many cases, nondestructive inspection, both prior to and during fabrication, must be done to sensitivities more stringent than those required by specifications. The use of timely inspection and rigid construction standards results in the reduction of both the costs and delays due to rework. Quality planning starts during the design stage. For inspections to be meaningful, consideration must be given to the condition of the material, the location and shape of welded joints, and the stages of production at which the inspection is to be conducted. During fabrication, quality plans must be integrated with the manufacturing sequence to ensure that the

inspections are performed at the proper time and to the requirements of the applicable standard. In the newest nuclear plants, quality design planning includes: • •

•

•

Avoidance of complex weld geometries to facilitate attachment of ultrasonic transducers to the surface at the best positions The increased use of ring forgings for pressure vessel components; this means that there are no longitudinal welds that have to be inspected in service. The result is a reduction in the amount of inservice inspection and man-rem exposures Incorporating large numbers of access points for introducing mechanized inspection equipment, which can be operated remotely, thus avoiding exposures to operators and enabling more accurate processing than is possible with handheld inspection equipment The elimination of welds between cast austenitic components; inspection of welds through cast welds is difficult because they are opaque to ultrasonic inspection to a large degree (Ref 1)

Reference

1. Outlook on Nondestructive Examination, Nucleonics Week, 30 June 1988 Nondestructive Inspection of Boilers and Pressure Vessels

Boiler and Pressure Vessel Code and Inspection Methods* Pressure vessels--both fossil fuel and nuclear--are manufactured in accordance with the rules of the applicable American Society of Mechanical Engineers (ASME) Boiler and Pressure Vessel Code (Ref 2). For nuclear vessels, section XI of the ASME code establishes rules for continued nondestructive inspections at periodic intervals during the life of the vessel. One feature of the rules in section XI is the mandatory requirement that the vessel be designed so as to allow for adequate inspection of material and welds in difficult-to-reach areas. Section III of the code describes the material permitted and gives rules for design of the vessel, allowable stresses, fabrication procedure, inspection procedure, and acceptance standards for the inspections. Pressure vessels are constructed in various sizes and shapes, and some of the largest are those manufactured for the nuclear power industry. Some pressure vessels are more than 6 m (20 ft) in diameter and 20 m (70 ft) in length and weigh almost 900 Mg (1000 tons). Thickness of the steel in the walls of these vessels ranges from about 150 mm (6 in.) to more than 400 mm (15 in.), although many pressure vessels and components are fabricated from much thinner material. Joining of the many vessel sections is accomplished by welding. Welders of pressure vessels are qualified according to section IX of the ASME Boiler and Pressure Vessel Code, and welding is done in accordance with qualified welding procedures. Nondestructive inspection of welds is only a part of the inspection requirements; the materials themselves must be inspected prior to welding. For pressure vessels that are not constructed according to the ASME code, it is a matter of agreement between the manufacturer and the user as to whether NDI methods are to be employed and which method or methods are to be used. Nondestructive Inspection Methods. An appendix to each section of the ASME code establishes the methods for

performing nondestructive inspection to detect surface and internal discontinuities in materials. Four inspection methods are acceptable: radiographic, magnetic particle, liquid penetrant, and ultrasonic inspection. All these methods are mandatory for nuclear vessels, for section III, and for division 2 of section VIII of the code. Ultrasonic inspection is listed in division 1 of section VIII of the code as nonmandatory. Leak testing, eddy current inspection, acoustic emission inspection, and visual inspection are included in section V. Details as to which method is to be used and the required acceptance standards are specified in the appropriate articles on materials and fabrication. All NDI personnel must be qualified and certified to SNT-TC-1A procedures (Ref 3). Radiographic Inspection. Methods of radiographic inspection are extensively detailed in the ASME codes;

radiography using either x-rays or radioisotopes as the radiation source is permitted. Radiography is the oldest inspection method detailed in the codes and is probably the most understood and the most widely accepted. A principal reason for its

wide use is that radiography provides a permanent record of the results of the inspection. This record is important because the inspector can review the radiographs at any time to ensure that federal, state, or insurance requirements have been met. Acceptance standards were developed according to the limits of radiography (what can or cannot be detected by the method) and by the quality level obtainable by the manufacturing practices used to produce the vessels. Essentially, the acceptance standards do not permit the existence of indications of the following types of flaws: cracks, incomplete fusion, incomplete penetration, slag inclusions exceeding a given size that is not related to the thickness of the part, and porosity that exceeds that presented in illustrated charts provided in the codes. These standards result from the ability to distinguish among porosity, slag, and incomplete fusion in the radiograph; more important, they also mean that no indications of cracks or of incomplete fusion are permitted. Magnetic Particle Inspection. The procedures for magnetic particle inspection reference ASTM E 709 (Ref 4) or section V of the ASME code for the method. Acceptance standards permit no cracks, but rounded indications of discontinuities are permitted provided they do not exceed a certain size or number in a specified area. Magnetic particle inspection is universally used on ferromagnetic parts, on weld preparation edges of ferromagnetic materials, and on the final welds after the vessel has been subjected to the hydrostatic test. A magnetic particle inspection must be conducted twice on each area, with the lines of magnetic flux during the second application at approximately 90° to the lines of magnetic flux in the first application. Depending on the shape of the part and its location at the time of inspection, magnetization can be done by passing a current through the part or by an encircling coil and sometimes by a magnetic yoke. The acceptance level is judged by a qualified operator and is subject to review by an authorized code inspector. Liquid penetrant inspection is usually employed on nonferromagnetic alloys, such as some stainless steels and high-

nickel alloys. The acceptance standards are the same as for magnetic particle inspection and are also judged by an operator, subject to review by a code inspector. The methods are specified to those contained in ASTM E 165 (Ref 5) or section V of the ASME code. Water-washable, postemulsifiable, or solvent-removable penetrants can be used. A waterwashable color-contrast penetrant is usually employed because it is easy to handle, requires no special ventilation, and is nontoxic. Sometimes, special requirements dictate the use of either a solvent-removable color-contrast penetrant or a fluorescent penetrant. Ultrasonic inspection is used to inspect piping, pressure vessels, turbine rotors, and reactor coolant pump shafts.

Straight-beam ultrasonic inspection is specified to detect laminations in plates and to detect discontinuities in welds and forgings. This technique is described in general and specific terms in section XI of the ASME code, in the United States Nuclear Regulatory Commission Regulatory Guide 1.150 (Ultrasonic Testing of Reactor Vessel Welds During Preservice and Inservice Examinations), and in companion reports written by utility ad hoc committees (Ref, 1). Angle-beam inspection is specified for welds, and a more detailed procedure is presented, including reporting requirements, It is mandatory, however, that ultrasonic inspection, either by straight beam or angle beam, be conducted to a detailed written procedure. These procedures are usually developed by the manufacturer. Specifications and standards for steel pressure vessels are given in ASTM A 577 (Ref 6), A 578 (Ref 7), and A 435 (Ref 8). Acceptance standards for the inspection of welds by ultrasonics closely parallel the acceptance standards for radiography. Cracks, incomplete fusion, and incomplete penetration are not permitted. The size permitted for other linear indications is the same for the slag permitted by radiography. However, ultrasonic inspection can detect cracks better than radiography, but it is sometimes difficult to separate cracks from other linear indications by ultrasonics. Furthermore, ultrasonic inspection procedures refer to the amplitude of the signal obtained from a calibration notch, hole, or reflector placed in a standard reference block, but not all slag inclusions or cracks in an actual workpiece present a similar response to that obtained from the artificial calibrator. Advanced ultrasonic systems (see the section "In-Service Quantitative Evaluation" in this article) and the improvements in codes and regulations have combined to make ultrasonic inspection one of the most commonly used nondestructive methods in the power industry. Advanced ultrasonic methods are intended to ensure that the vessel remains fit for continued service by detecting and sizing defects that could degrade structural integrity. Acoustic emission (AE) inspection has been used for the following applications (Ref 9):

• • •

Inspection of chemical and petrochemical vessels Monitoring nuclear plant components or systems during hydrotests, plant operation, or preservice pressure testing of the primary system Monitoring during pressure testing of intentionally flawed vessels

• •

Monitoring fiber-reinforced plastic tanks, with the major problems being associated with poor manufacturing techniques that allow dust and other foreign objects to be mixed in with the resin Monitoring liquefied petroleum gas storage tanks. The main problems associated with this type of test program include the different propagation paths and attenuation coefficients caused by the geometry of the vessel and the correct transducer locations and spacing

Rupture tests on experimental vessels with wall thicknesses up to 150 mm (6 in.) have been monitored, and such tests help define the acoustic emission response patterns that can be used to recognize incipient vessel failures. However, many of the vessel rupture tests monitored by acoustic emission have been conducted mainly to provide fracture mechanics data, and the acoustic emission monitoring was an add-on feature (Ref 9). Acoustic emission tests are often conducted during preshutdown operations in an effort to identify areas requiring special maintenance or during special tests while under slightly varied operational conditions. Data on wave propagation and failure mechanisms have been recorded and used to develop reliable acoustic emission evaluation techniques (Ref 10). The advancement of AE inspection techniques includes the introduction of an AE methodology into Section V, Article 12, of the ASME Boiler and Pressure Vessel Code as a December 1988 Addendum. Other AE methodologies in the inspection of fiber-glass and metal pressure vessels are described in the article "Acoustic Emission Inspection" in this Volume. Eddy Current Inspection. One of the rapidly increasing applications for this method is the inspection of thin-wall

(0.9 to 1.5 mm, or 0.035 to 0.060 in.) Inconel alloy steam generator tubing (Ref 11, 12) and heat exchanger tubing (Ref 13). The focus of steam generator tube inspection by eddy current has shifted from concentrating solely on the detection and characterization of tube wall denting and wastage to more complex tube wall degradation mechanisms. The newer degradation mechanisms consist of intergranular attack (IGA), stress-corrosion cracking (SCC), mechanical wear, and pitting in the presence of copper. A summary of steam generator tubing defects is shown in Fig. 1. In response to the complexity of these newer problems, eddy current instrumentation has also evolved from a single-frequency to a multiplefrequency configuration. The analog instrumentation has been replaced by digital multiple-frequency instrumentation, offering more consistent data acquisition. The wider dynamic range offered by the digital instrumentation allows analysis of eddy current data obtained from traditionally difficult areas, such as dented tube support/tube sheet interfaces and roll transition/roll expansion areas within the tube sheet. Additional information, including typical data produced by multiplefrequency instrumentation, can be found in the article "Eddy Current Inspection" in this Volume.

Fig. 1 Locations of known tube wall degradations in recirculating steam generators. Source: Ref 11

Another advancement is that of remote-field eddy current testing, which has been used to examine nuclear fuel rods and other tubular products. The article "Remote-Field Eddy Current Inspection" in this Volume can be consulted for details and results of analyses. Replication microscopy, or field metallography, is an effective and economical means of obtaining an image of a

component surface, permitting both on-site and laboratory examination and evaluation of the metallurgical condition of the material (see the article "Replication Microscopy Techniques for NDE" in this Volume). Material sample removal from critical components is a costly and time-consuming process that often necessitates part replacement, weld repair, and stress relief. The replication examination is completed without having to cut out a portion of the component, and the metallurgical data can be determined in fine detail. Therefore, material evaluation can be performed with a nondestructive and cost-effective method that does not require the removal of material samples.

Specimen (Surface) Preparation. Critical surface locations for inspection are identified, and external insulation is

removed. The area to be replicated is polished with progressively finer grits and compounds. The final compound is a 0.25 m diamond polishing paste, which ensures a surface finish of 1 m or better. The surface is then etched to reveal the microstructural detail (Fig. 2). A plastic film is applied, cured, and carefully removed. The surface of the film in contact with the polished area retains a precise, reverse image of the etched component surface.

Fig. 2 Weld microstructures showing no evidence of microvoids in a secondary superheater outlet header made from 1.25Cr-0.5Mo steel. (a) Weld metal. (b) Heat-affected zone. (c) Base metal. Weld location: nozzle-toheader weld (header side of weld). All 400×. Courtesy of H.I. Newton, Babcock & Wilcox

Procedure/Defect Analysis. In the laboratory, the replica is examined using standard light (optical) microscopy, or it can be viewed at high magnification with an electron microscope. To improve resolution, the replica is enhanced by shadowing with sputtering or vacuum evaporation techniques.

An actual replica magnified to 400× is shown in Fig. 3. The microstructure indicates that linked creep voids exist along the grain boundaries in this component, suggesting that creep is active and that cracking is imminent. The progression of damage from microstructural voiding to linked voids to cracking is schematically illustrated in Fig. 4. Replicas can reveal all the microvoids, microcracks, and microstructural features in exact detail, just as if the metal surface itself were being viewed.

Fig. 3 Replica showing random and linked microvoids in the heat-affected zone of a secondary superheater outlet header weld. 400×. Courtesy of H.I. Newton, Babcock & Wilcox

Fig. 4 Progression of creep damage over time as depicted by replica photomicrographs. Courtesy of H.I. Newton, Babcock & Wilcox

The principal application of the replication technique is in revealing metallurgical anomalies and incipient damage

(cracking). In the absence of fatigue and/or preexisting defects, creep cracks are theorized to initiate by the formation, growth, and linkup of voids into microcracks, which in turn consolidate to form macrocracks (Fig. 5). The propagation of the macrocracks can then lead to final failure.

Fig. 5 Replicas of creep damage in pressure vessel components. (a) Random and linked microvoids in a 1.25Cr0.5Mo main steam line weld. 425×. Courtesy of H.I. Newton, Babcock & Wilcox. (b) and (c) A comparison of scanning electron micrographs of a replica from 1.25Cr-0.5Mo material with aligned creep cavities. (b) Normal contrast. (c) Reversed image contrast. Etched with 4% picral. Courtesy of E.V. Sullivan, Aptech Engineering Services, Inc.

In actuality, a creep-fatigue interaction better describes the type of damage mechanism most often encountered. It is theorized that the creep mechanism described previously is accomplished by a fatigue mechanism. In a creep-fatigue interaction, one may observe a mixture of intergranular and transgranular crack growth surfaces with the absence of voiding. The presence of preexisting flaws affords an initiation site for cracking and subsequent material degradation.

Actively growing cracks are evidence that the condition of a component is deteriorating under service loads; this warrants monitoring, repair, or replacement. Creep-fatigue interaction is discussed in detail in the article "Creep-Fatigue Interaction" in Mechanical Testing, Volume 8 of ASM Handbook, formerly 9th Edition Metals Handbook. The advantages of the replication techniques are many. It is nondestructive. An accurate image of the component

surface is also obtained. In addition, early detection of damage and the evaluation of existing failure mechanisms are possible. Replicas also reinforce decision making with regard to the need to replace or repair critical degraded components. Limitations. There are also disadvantages with the replication technique. The analysis is limited to the accessible

surface of the component. The replication technique is also site sensitive. In addition, no compositional analysis of the microstructural constituents is possible with surface replicas, although extraction replicas have been used in chemical analysis (see the section "Precipitate Analysis" in the article "Replication Microscopy Techniques for NDE" in this Volume). With these limitations in mind, it is recommended that replication be used in conjunction with other nondestructive techniques to provide a complete overall evaluation of a component.

References cited in this section

2. ASME Boiler and Pressure Vessel Code: Section II--Material Specifications, Part A--Ferrous Materials; Section III, Division 1--Nuclear Power Plant Components; Section V--Nondestructive Examination; Section VIII--Division 1--Pressure Vessels, Division 2--Alternative Rules for Pressure Vessels; Section IX-Welding and Brazing Qualifications; Section XI--Rules for Inservice Inspection of Nuclear Power Plant Components, American Society of Mechanical Engineers 3. "Recommended Practice for Nondestructive Testing Personnel Certification," SNT-TC-1A, American Society for Nondestructive Testing, 1988 4. "Standard Recommended Practice for Magnetic Particle Examination," E 709, Annual Book of ASTM Standards, American Society for Testing and Materials 5. "Standard Practice for Liquid Penetrant Inspection Method," E 165, Annual Book of ASTM Standards, American Society for Testing and Materials 6. "Standard Specification for Ultrasonic Angle-Beam Examination of Steel Plates," A 577, Annual Book of ASTM Standards, American Society for Testing and Materials 7. "Standard Specification for Straight-Beam Ultrasonic Examination of Plain and Clad Steel Plates for Special Applications," A 578, Annual Book of ASTM Standards, American Society for Testing and Materials 8. "Standard Specification for Straight-Beam Ultrasonic Examination of Steel Plates," A 435, Annual Book of ASTM Standards, American Society for Testing and Materials 9. B.R.A. Wood, Acoustic Emission Applied to Pressure Vessels, J. Acoust. Emiss., Vol 6 (No. 2), 1989, p 125-132 10. J.C. Spanner, Acoustic Emission in Pressure Vessels, in Pressure Vessel and Piping Technology--1985: A Decade of Progress, American Society of Mechanical Engineers, 1985, p 613-632 11. K. Krzywosz, Recent NDE Experiences With PWR Steam Generator Tubing Inspection, in NDE in the Nuclear Industry, ASM INTERNATIONAL, 1987, p 157-167 12. R.H. Ferris, A.S. Birks, and P.G. Doctor, Qualification of Eddy Current Steam Generator Tube Examination, in NDE in the Nuclear Industry, ASM INTERNATIONAL, 1987, p 71-73 13. V.S. Cecco and F.L. Sharp, Special Eddy Current Probes for Heat Exchanger Tubing, in NDE in the Nuclear Industry, ASM INTERNATIONAL, 1987, p 109-174 Note cited in this section

* The material on replication microscopy in this section was prepared by H.I. Newton, Babcock & Wilcox.

Nondestructive Inspection of Boilers and Pressure Vessels

Inspection of Plates, Forgings, and Tubes Plates up to 300 mm (12 in.) thick, such as those conforming to ASME SA-533, grade B (Fe-Mn-Ni-Mo) are used for pressure vessel shell sections. All plates are ultrasonically inspected 100%. Pulse-echo search units have been adapted for this inspection. The search unit consists of an immersion-type longitudinal wave or shear wave transducer encased in a container holding a liquid couplant. Because the container can conform to minor surface irregularities on the material being inspected, much of the surface conditioning previously required for contact inspection is eliminated. Good, sound penetration and a stable back reflection are obtained with this type of unit. Straight-beam inspection is conducted over the entire surface of the plate. Calibration for straight-beam inspection is accomplished by adjusting the back-reflection amplitude from the opposite surface to approximately 75% of the height of the CRT. Discontinuities causing a loss of back reflection that cannot be contained in a 75 mm (3 in.) diam circle are unacceptable. Laminar-type discontinuities, which are the most common type of imperfections encountered in plate materials, are readily detected by straight-beam inspection. The angle-beam inspection of plates is included in many specifications. Scanning is conducted in overlapping passes over the entire surface in two directions at 90° to each other. Calibration is accomplished by adjusting the signal to a square notch milled to a depth of 3% of the plate thickness. Discontinuity indications equal to or exceeding the signal amplitude from the notch are cause for rejection. During the ultrasonic inspection of plates, close attention must be given to signal patterns that indicate the presence of discontinuities, such as secondary piping, which do not normally produce significant response indications on the oscilloscope. Discontinuities of this type can result in untimely rejection or high rework costs if not detected prior to the start of fabrication. Forgings used for pressure vessel flanges, head flanges, and nozzles are commonly purchased to the requirements of ASME specifications, such as those of ASME SA-508, class 2 (Fe-Mn-Ni-Cr-Mo-V), and are heat treated and nondestructively inspected by the supplier. Ultrasonic inspection is required on two surfaces of the forging using the straight-beam inspection method; for ring or hollow round shapes, angle-beam inspection is specified. These inspections must be carried out prior to contour machining to ensure complete volumetric coverage. Inspection is repeated to the maximum extent practical after machining and heat treatment are completed. Discontinuity indications observed by straight-beam inspection that cause a loss of back reflection, and by angle-beam inspection that are equal to a 3%-ofplate-thickness notch, will result in the rejection of the component. Ferromagnetic forgings are inspected by magnetic particle inspection methods after final heat treatment and contour machining. Linear discontinuities such as laps, cracks, or inclusions are cause for rejection. Tubular products can receive ultrasonic and liquid penetrant inspections prior to release for fabrication. As stated above, eddy current inspection is common during the in-service inspection of tubular products. Nondestructive Inspection of Boilers and Pressure Vessels

Nondestructive Inspection During Fabrication Plates are formed into shell cylinder segments or into hemispherical segments for the top and bottom heads. Shell plates of thick material are hot formed by pressing to the radius required. Plates for head segments are hot formed over dies machined to the proper contour. After the hot forming is completed, the plates are allowed to cool and are then inspected to confirm that the proper radius has been obtained. The formed section is then reheated, water quenched, and tempered to obtain the required mechanical properties. Samples are cut from the heat-treated material and tested to determine the tensile and impact properties of the material before further fabrication is permitted.

Fit-Up and Welding. Edges of plates that are machined in preparation for welding are magnetic particle inspected.

Surface discontinuities, other than laminar types less than 25 mm (1 in.) long, are removed. Discontinuities with any significant thickness are further investigated by radiographic or ultrasonic inspection to determine their nature and extent. Welding of the longitudinal seams in the shell cylinders is commonly performed by the multilayer submerged arc process. Single arc and tandem arc procedures are used, with preheating and postheating when specified. Welding of head-torus sections can be done using the horizontal submerged arc process, in which the weld beads are deposited through the thickness of the material. With the horizontal submerged arc process, the axis of the weld is in a vertical position, and each weld bead is deposited from the inside to the outside of the seam progressively upward until the complete length is welded. A manual welding procedure (usually, shielded metal arc) in which the axis of the weld is in the vertical position is sometimes used instead of the submerged arc process. These processes require NDI techniques suitable for adequate interpretation of discontinuity size. After welding and interstage stress relieving of the longitudinal seams are completed, the weld is magnetic particle inspected and radiographed. The design of the weld joint and the welding procedure used are important considerations in determining the extent of nondestructive inspection and the adequacy of the results. The narrow-groove weld joint illustrated in Fig. 6 provides for good radio-graphic inspection, including crack detection. Those cracks that are more likely to occur along the sidewall of a weld are essentially parallel to the radiographic beam in the joint design. When the radiographs are made with equipment and techniques that obtain good sensitivity in these thick-section welds, cracks of this type can be readily detected. Welds having wide-angle sidewalls require additional angle radiography or ultrasonic inspection.

Fig. 6 Sequence of assembly and typical narrow-groove joint used for the welding of thick plates in pressure vessels

The depth at which slag inclusions are beneath the surface and the thickness of the inclusions usually cannot be determined by radiographs taken normal to the weld surface. Discontinuity indications observed on the film usually require additional angle radiography or ultrasonic inspection to determine the depth and thickness of the slag inclusions. Acceptance standards for welds made in the vertical position are more rigid than acceptance standards applied to other welding procedures.

Clad Vessels. Certain types of pressure vessels, such as some nuclear vessels or chemical-processing vessels, are

fabricated from carbon or alloy steels whose inside surfaces are clad with a layer of stainless steel. With these types of vessels, the inner surfaces are machined smooth to provide a suitable surface for cladding. Machining is done after the shell and head-torus subassemblies have been completed. After machining, the smoothed surfaces are magnetic particle inspected; no indications of linear discontinuities are permitted. Also at this manufacturing stage, the formed and heat-treated material is again ultrasonically inspected. This inspection is performed using the straight-beam method over the entire surface of the material. Changes in the acoustic quality of the material as a result of heat treatment may cause minor differences in the signal amplitude from those indications previously recorded. The inner surfaces of large pressure vessels are clad, oscillating multiple-wire systems or strip electrodes. Three-wire and six-wire self-shielded flux-cored wire, plasma arc processes, and electroslag overlays are also used for pressure vessel applications. Liquid penetrant inspection is performed on all clad surfaces. When properly cleaned, the as-deposited clad surface is adequate for liquid penetrant inspection. Solvent-removable red-dye penetrants are often used for this inspection. Acceptance standards do not allow linear indications of any type. Rounded indications with less than 1.5 mm ( in.) maximum dimension are acceptable. Mating surfaces of machined flanges and surfaces of O-ring seals are inspected to more rigid acceptance standards. Sealing surfaces must be free of any indications. Calibration for inspection of the cladding is accomplished by setting up the inspection equipment on a reference block clad by the same process as the material being inspected. A 10 mm ( in.) wide slot is machined through the backing material to the cladding interface. The ultrasonic pattern observed from the slot in the reference block provides the acceptance standard for the cladding. Inspections are conducted from the clad side of the material, except for small-diameter nozzles. When inspecting from the side opposite the cladding, close observation must be given to the presence of laminar discontinuities in the base material that could mask the area of inspection. Cladding on mating flange surfaces is inspected to a more rigid standard by pitchcatch echo testing. The two-transducer system is calibrated on a 1.5 mm ( in.) diam, side-drilled hole at the interface of the cladding and base material. Discontinuity indications exceeding the amplitude of the signal from the calibration hole are unacceptable. Assembly. Shell cylinders and flanges are joined together by welding the girth seams using the submerged arc process.

Back grooves, final surfaces, and surfaces prepared for welding are magnetic particle inspected. Radiography is performed after cladding the inside of the welded girth seam. Nozzle openings are cut in the shell cylinder and are then machined to the size and shape required for fitting the nozzles. The narrow-groove weld joint, through the shell thickness, is used for the nozzle-to-shell weld. Nozzle welds are made by either the manual or the automatic submerged arc process. In addition to magnetic particle and radiographic inspection, all nozzle welds are subjected to ultrasonic inspection. The techniques for the ultrasonic inspection of welds contained in the ASME code are used as the basic method of inspection. Angle-beam inspection is performed on the outer surface of the shell adjacent to the nozzle weld. The sound beam is aimed about 90° to the weld and manipulated laterally and longitudinally so that maximum coverage is obtained. Sound-beam dispersion from the effect of the clad interface prevents full-node sound transmission. Only the discontinuities encountered in the initial sound path through the material can be consistently detected by the angle beam. Straight-beam inspection is performed by directing the sound into the weld from the inside of the nozzle. The combination of straight-beam and angle-beam inspection provides full inspection coverage of the weld. For the detection of cracks in welds, ultrasonics and radiography have complementary advantages and disadvantages. Vertical, subsurface cracks in a weld that are readily detectable by radiographic methods are extremely difficult to detect by ultrasonics. Discontinuities oriented in a plane other than normal to the weld surface can be best detected ultrasonically.

Various beam angles and frequencies are used to establish the best ultrasonic procedure that could be applied to production welds. The standard 45° transducer at a frequency of 2.25 MHz is commonly used for the detection of cracks. However, to provide additional assurance, scanning should be performed at a gain setting of at least ten times the calibration sensitivity. Final Inspection. After assembly is completed and all inspections are verified, the vessel is subjected to a final stressrelieving treatment. Magnetic particle inspection of all exposed ferritic material is again performed, using the yoke method. Arc strikes that are likely to occur with the prod method are thus avoided on the final stress-relieved surface.

Final machining of the vessel and head is done after stress relieving. All machined surfaces are inspected by liquid penetrants after the vessel is dimensionally inspected. Vessel and head inside surfaces are thoroughly cleaned, and the vessel is positioned vertically for hydrostatic testing. After hydrostatic testing, all exposed ferritic welds are magnetic particle inspected. Instrumentation and attachment welds are liquid penetrant inspected. A baseline ultrasonic inspection can be performed on the completed vessel. This inspection provides a record of ultrasonic indications that can be used as a reference for future in-service inspections of the vessel. The data generated from this inspection are useful only if the method of inspection during the in-service inspection is compatible with the baseline inspection method. The effect of the cladding interface on the ultrasonic sound beam requires increased sensitivity settings on the equipment if inspection is conducted from the inside of the vessel. Beam-angle changes due to the irregular cladding interface require exact duplication of the position of the transducer and the entrance angle of the sound beam if reproducible results are to be obtained. Equipment capable of remotely controlling these variables is essential if these inspections are to be conducted. Nondestructive Inspection of Boilers and Pressure Vessels

Visual Inspection of Pressure Vessels Visual inspection can readily detect misalignment, movement of mating parts, surface contamination, cracks, and other surface discontinuities. Direct examination is possible when the eye or an equivalent detector can be placed within 600 mm (24 in.) of the surface to be inspected. Where radiation levels are too high or access is limited, remote inspections can be used; telescopes, periscopes, borescopes, fiber optics, or television monitoring systems are available (see the article "Visual Inspection" in this Volume). Visual inspection requirements are extensive in section XI of the ASME Boiler and Pressure Vessel Code. In general, all weld areas not receiving volumetric inspections must be visually inspected after hydrostatic testing. Also, comprehensive inspections are applied to bolting, piping, pumps, and valves. Nondestructive Inspection of Boilers and Pressure Vessels

Radiographic Inspection of Pressure Vessels A radiograph is basically a two-dimensional picture of the intensity distribution of a form of radiation that is projected from a source (ideally, a point source) and that has passed through a material that attenuates the intensity of the radiation. Voids, changes in thickness, or regions of different composition will, under favorable circumstances, attenuate the radiation by different amounts, producing a projected shadow of themselves. Three forms of penetrating radiation are used in radiography: x-rays, γ-rays, and neutrons. X-rays were the earliest to be used. Detailed information on radiography is available in the article "Radiographic Inspection" in this Volume. Radiography is normally used to examine welded seams in fabricated pressure vessels. The main reasons for the use of radiography are the following:

• • •

Radiography permits internal inspection of a component Radiography supplies a permanent visual record Radiography is generally sensitive to discontinuities commonly present in welds

The use of radiography is generally restricted to the inspection of welds for the following reasons: • • •

The cost of performing radiographic inspections is high when compared to other available methods Radiographic inspection of large areas is relatively slow Radiography is insensitive to laminar-type discontinuities

Radiography can be used, if the proper equipment is available, to inspect welds in steel pressure vessels with walls up to 400 mm (16 in.) thick. The only conditions that could limit the use of radiography for the inspection of welds in pressure vessels are, first, a lack of accessibility to both sides of the weld for placement of the film and the radiation source and, second, a high degree of variability in shape and thickness in the area being inspected. Both single-wall and double-wall techniques are commonly employed for the radiographic inspection of pipe. Detailed discussions of these techniques are available in the article "Radiographic Inspection" in this Volume. The following examples describe standard radiographic techniques used to inspect common welded seams in pressure vessels.

Example 1: Radiographic Inspection of Longitudinal, Girth, and Nozzle Welds in a Carbon Steel Pressure Vessel. The pressure vessel shown in Fig. 7 was used in a commercial nuclear application and was inspected in accordance with section III of the ASME Boiler and Pressure Vessel Code. The material was 65 mm (2 in.) thick carbon steel. The weld joints were single-V grooves and were welded using the submerged arc process for both the longitudinal and girth seams. Shielded metal arc welding was used to join the nozzle to the vessel.

Fig. 7 Nuclear pressure vessel, of 65 mm (2 in.) thick carbon steel, in which longitudinal, girth, and nozzle welds were radiographically inspected using 50-Ci cobalt-60 as the radiation source, Eastman Kodak type AA radiographic film, and the inspection setups shown. Dimensions given in inches

For radiographic inspection, cobalt-60 was selected as the radiation inspection because of the thickness of the material, and Eastman Kodak type AA film was selected for its relative speed to give the shortest exposure time. The combination of Co-60 and type AA film can sometimes result in marginal sensitivity that occasionally requires the reshooting of a specific region. Each longitudinal seam was radiographically inspected with eight shots spaced 350 mm (14 in.) apart using a 180 × 430 mm (7 × 17 in.) film. The penetrameters were steel ASTM No. 45 with a 1.5 mm ( in.) thick shim. The Co-60 radiation source was placed 690 mm (27 in.) from the film. The cassette was loaded with a 0.25 mm (0.010 in.) thick lead screen, two pieces of film, a 0.25 mm (0.010 in.) thick lead screen, and a 3 mm ( A in Fig. 7 shows the setup for radiographic inspection of the longitudinal seam.

in.) thick lead back shield. Section A-

The girth weld was inspected using the panoramic technique, in which the Co-60, 50-Ci radiation source was placed at the center of the vessel and 28 film cassettes were placed around the outer surface (section B-B, Fig. 7). The technique involves only one exposure of 25 min to radiograph the entire seam, but its success depends on the placement of the radiation source exactly in the center of the vessel. The disadvantages of this technique include a lengthy setup time, a fairly complicated setup, and additional film costs if an error is made. An alternative technique is to place the radiation source outside the vessel and inspect the seam with 28 separate exposures. This method would result in higher-quality radiographs, but a much longer total exposure time would be required. Three equally spaced steel ASTM No. 45 penetrameters with 1.5 mm ( in.) thick shims were used for the 28 radiographs. The film size and the types and thicknesses of screens and filters were the same as those used for the longitudinal seam welds. The nozzle seam weld was radiographed using eight 13-min exposures. As shown in section C-C in Fig. 7, the radiation source was placed outside the vessel and was offset at an angle of 7 to 10° away from the nozzle. The source-to-film distance was 900 mm (36 in.). The variation in weld thickness at the nozzle required that two penetrameters be used. One was flush to the sidewall, and the other had a 9 mm ( in.) thick shim, which compensated for the difference in weld thickness. This two-penetrameter setup qualified the sensitivity of the radiograph because of the differential thickness of the weld being inspected. A sensitivity of 2-2T and a density of 2.0 were specified for the radiographs of all three types of welds.

Example 2: Radiographic Inspection for Creep Fissures in Reformer-Furnace Tubes (Ref 14). About 1 year after start-up, two steam-methane reformer furnaces were subjected to short-time heat excursions because of a power outage, which resulted in creep bulging in the Incoloy 800 outlet pigtails, requiring complete replacement. It was thought that during this heat excursion some of the reformer tubes experienced slight bulging, about 1%, by plastic deformation. However, this type of bulging does not necessarily shorten the life of a tube in terms of creep. Each furnace had three cells, consisting of 112 vertical tubes per cell, each filled with a nickel catalyst. The tubes were centrifugally cast from ASTM A297, grade HK-40 (Fe-25Cr-20Ni-0.40C), heat-resistant alloy. The tubes had an outside diameter of 150 mm (5.85 in.) and a minimum wall thickness as-cast of 20 mm (0.802 in.), and they were more than 13.7 m (45 ft) in length. The design limits for the tubes were 2200 kPa (320 psi), with a surface temperature of 958 °C (1757 °F). Operating limits were 2140 kPa (310 psi) at a surface temperature of 943 °C (1730 °F). The furnace cell is illustrated in Fig. 8(a).

Fig. 8 Reformer-furnace cell from which cast tubes of ASTM A 297, HK-40, heat-resistant alloy were radiographically inspected for the detection of creep fissuring. (a) Schematic of furnace cell showing positions of radiographic sources and films. Dimensions given in inches. (b) Radiograph of a section removed from a failed tube that contained no catalyst showing fissures near the ruptured area. (c) Same section as (b) but containing a catalyst. Fissures are visible but less apparent. (d) Macrograph showing fissures in a tube that were detected by radiography. 6×. (e) Macrograph showing fine fissuring that was not indicated by radiography. 6×

The first tube failure occurred after 33,000 h of operation. The unit was shut down, and the failed tube was removed for metallurgical inspection. The results indicated that the tube failed from creep rupture (stress rupture). The tube failure instigated a project for detecting midwall creep fissuring. Radiography had been reported to be limited to detecting only severe third-stage creep and not the early stages. Laboratory Radiography. Preliminary studies were made to determine type of radiation source and strength, radiation source-to-film distance, exposure time, and effect of catalyst in tube. Figure 8(b) shows a laboratory radiograph of a section removed from the failed tube without catalyst. Fissures were clearly visible near the ruptured area and diminished to nondetectable fissures at a point 400 mm (16 in.) below the rupture. The same tube section radiographed under similar conditions but with catalyst in the tube is shown in Fig. 8(c). Fissures are visible, but it is apparent that the catalyst reduced the sensitivity by masking the smaller fissures.

Liquid penetrant inspection and macroexamination of the specimen tube revealed the gross fissuring that was easily detectable and those fissures that were undetectable by radiography. The largest fissures undetectable by radiography were approximately 3 mm (0.125 in.) long and 0.13 mm (0.005 in.) wide. (Fissures have been detected that were between 5 and 10 mm, or 0.200 and 0.400 in., long.) Based on the radiography and macroexamination results, it was decided that future tube replacements would be inspected by radiography, preferably with the catalyst removed. If the fissure were large enough to show on a radiograph, either with or without the catalyst, the tube could be expected to fail within 1 year. In-Service Radiographic Inspection. During shutdown of one furnace, the first full test of the radiographic

inspection method was made. All the tubes were measured for evidence of outside diameter growth, and the suspect areas

were strapped to identify the hottest zones. Growth to failure in HK-40 material having a rough exterior surface is difficult to measure, because total creep of only about 1% has occurred by the time of failure. Gaging did not show much, but strapping on some tubes clearly identified bulging in the hottest areas about 1.4 m (54 in.) above the burner terrace. This was used as the basis for radiographing all other tubes. The catalyst was removed from the tubes in preparation for radiography. Two IR-192 radiation sources having a strength of 3300 GBq (90 Ci) and one 3700 GBq (100 Ci) source were used, mounted in jigs at distances of 750 and 900 mm (30 and 36 in.), respectively, from the film. A tungsten collimator was also used to limit the emission to a single tube and permit the technicians to remain in the firebox during exposure. Exposure time was 7 to 10 min per shot, and approximately 700 shots were taken. Shortly after the first furnace was back on stream, the second furnace was shut down and the tubes radiographed. Using the same techniques but with a 3900 GBq (105 Ci) source, time to radiograph the second furnace was reduced about 25%. Twenty-four tubes in the first furnace and 53 in the second furnace showed significant fissuring. One of these fissured tubes was left in the first furnace and 15 in the second; when these tubes eventually failed, they would provide an indication of remaining life after a known radiographic examination. One of the tubes that had been removed from the furnace was sectioned at two places--through a radiographic indication of a fissure and through an area containing no indications. Macrographs of these sections are shown in Fig. 8(d) and 8(e). Figure 8(d) shows fissuring that was detected by radiography. Fine fissuring that was not seen on the radiograph is shown in Fig. 8(e). Conclusion. Radiography was a practical, economical method of detecting the creep fissuring, and it provided advance information for purchase of replacement tubes. However, because radiography was limited to detecting fissures caused by third-stage creep in tubes from which the nickel catalyst had been removed and because of the cost of removing the catalyst, ultrasonic techniques were developed for inspecting the tubes. These techniques are described in Example 4 in this article.

Reference cited in this section

14. R.R. Dalton, Radiographic Inspection of Cast HK-40 Tubes for Creep Fissures, Mater. Eval., Vol 30 (No. 12), Dec 1972, p 249-253 Nondestructive Inspection of Boilers and Pressure Vessels

Ultrasonic Inspection of Pressure Vessels Volumetric inspections are used to inspect the volume of material bounded by the surfaces of components and of piping. Theoretically, any source of energy that penetrates the volume of a material can be used. In practice, however, only xrays, -rays, and ultrasonic waves are used. In nuclear vessels, which are housed in a containment building, radiographic inspection methods for in-service inspection currently have limited use because of the need for access to both surfaces and because of the high -ray background in most areas of the containment building. When radiography is used, other inspection methods such as acoustic emission can also be used for additional monitoring. However, a large number of inspections are performed using various ultrasonic techniques. As described in the article "Ultrasonic Inspection" in this Volume, ultrasonic waves are generated by piezoelectric transducers that convert high-frequency electrical signals into mechanical vibrations. These mechanical vibrations form a wave front, which is coupled to the vessel being inspected through the use of a suitable medium. Several wave modes can be used for inspection, depending on the orientation and location of the discontinuities that exist. Longitudinal, shear, and surface waves are used separately in different techniques to reveal discontinuities that are respectively parallel to, at an angle to, and on or near the surface from which the inspection is performed. These inspections are made with longitudinal wave transducers, pulse-echo, through transmission, pitch-catch, or delta techniques. Most of the inspections are performed using pulse-echo straight longitudinal wave beams and angled shear wave beams from a single transducer.

Manual ultrasonic inspection is performed with single-transducer pulse-echo techniques. The ultrasonic wave is

coupled to the component being inspected through a couplant (usually a glycerin or light oil) that transports the ultrasound between the face of the transducer and the surface of the component. In longitudinal wave straight-beam inspection, ultrasonic waves enter normal to the surface and will detect discontinuities perpendicular to the direction of wave propagation. In angle-beam inspection, the longitudinal wave is coupled through a plastic wedge at an angle to the surface. The wave is again coupled to the component (still a longitudinal wave) and undergoes a mode conversion to a shear wave. The angle of the resulting shear wave is dependent on the ratio of the combination of ultrasonic velocities in the plastic wedge and the metal component with the angle of the incident longitudinal wave in the plastic wedge. Shear wave angles from 40 to 75° are used for inspection. Indications are recorded and plotted from reference locations on the weld metal. Automated Ultrasonic Inspection. For accurate inspections, immersion ultrasonic inspection provides the highest

level of inspection speed, accuracy, and repeatability. The coupling of ultrasonic waves is done with much less variability in couplant thickness and transducer pressure than occurs in manual techniques. The immersion method permits automated scanning and digital recording of data with a high degree of precision and repeatability. Additional advanced automated methods are described below in the section "In-Service Quantitative Evaluation" in this article. Thickness Measurements. Inspection should properly begin with thickness measurements of the shell, heads,

nozzles, and piping. Well-documented inspection records are necessary to complete a satisfactory inspection. Records can indicate where to expect metal loss or corrosion on the component and therefore enable the suspect areas to be thoroughly inspected. A comparison of measurements obtained during previous inspections will determine the amount of loss and corrosion rates. Original or nominal thicknesses taken from specifications usually have tolerances too great to make them reliable in the determination of corrosion rates. Vessels having a history of minimal corrosion require only a moderate amount of inspection, while those having a high corrosion rate or history of attack must be more thoroughly inspected. Inspection experience and records provide the key to how much coverage should be given. If the vessel is open, the thickness measurements can be made from the internal surfaces at any location the inspector desires--an approach much more flexible than dependence on fixed corrosion-gaging points. The distance between thickness measurements can vary depending on the coverage desired, but a sufficient number of readings should be obtained to ensure a correct determination of vessel condition (distances generally range from to 1 m). In determining repair areas or the size and location of a patch, a grid pattern laid out on the vessel can be useful. Rough or badly pitted surfaces should have a small area ground smooth to permit the transducer to make good contact. Transducers having a rubber membrane as a protective facing contact better on mildly pitted surfaces because the rubber will tend to conform to the surface. Removal of loose scale or dirt by scraping or filing is often sufficient surface preparation. It is not necessary to drill gage holes or use other destructive methods to determine wall thickness. Insulation, if present, must be removed at gage locations for external measurements, but the removal of insulation is not required if the measurements are to be taken from the inside. Special couplants enable many external measurements to be taken while the vessel is operating at temperatures of 370 °C (700 °F) or higher. Corrections should be made for errors introduced by the high temperature. Table 1 lists several correction factors for ultrasonic thickness measurements in carbon steels at elevated temperatures. Additional information can be found in Ref 15. Table 1 Correction factors for ultrasonic thickness measurements in carbon steels at elevated temperatures Temperature

°C

°F

Correction factor

Measurement

Ultrasonic

Micrometer

mm

mm

in.

in.

40

100

13.11

0.516

13.11

0.516

0

95

200

13.15

0.518

13.13

0.517

0.998

150

300

13.26

0.522

13.14

0.5175

0.991

205

400

13.33

0.525

13.15

0.518

0.987

260

500

13.36

0.526

13.16

0.5185

0.986

315

600

13.41

0.528

13.18

0.519

0.983

370

700

13.46

0.520

13.20

0.520

0.981

400

750

13.51

0.532

13.26

0.522

0.981

Source: Ref 15 Flaw detection prior to fabrication is usually concerned with the inspection of steel or alloy plates for internal laminations, segregates, and similar discontinuities because they can be quite harmful under certain operating conditions. For example, hydrogen blistering can occur at a lamination, and stress-corrosion cracking, fatigue, intergranular corrosion, and hydrogen embrittlement are often accelerated by such discontinuities.

For flaw detection, a continuous-scan method is often used, in which the transducer is carried on a rigid carriage and the ultrasound waves are coupled into the plate by a stream of water. In some applications, the plate can be immersed, which simplifies the operation. To eliminate some of the human errors, a flaw-alarm system is set up so that a flaw of predetermined size will trigger a warning of some kind. For the inspection of a weld, the weld shape may be such that the use of longitudinal waves is impossible, and a flaw such as a crack at the fusion zone probably will not present enough reflective area to be detected from directly above it. Nozzles located in hemispherical heads present special problems, and special techniques must be used to ensure a thorough inspection. Using shear waves, the sound can be directed into a suspect area from almost any angle. Large-scale drawings of the suspect area are needed to lay out correct angles and distances for the transducers (Fig. 9). For components having a complicated design, a plastic model can be an aid to better visualization of the sound path and angles of reflection. Thought given to the best approach and techniques can save hours in inspection and can ensure proper coverage. Angle transducers of 45 and 60° will detect almost any weld discontinuity, particularly if both internal and external surfaces are available for inspection. If the internal surface is not available, the nozzle can be filled with water and a search unit lowered to the correct depth to inspect the weld. The search unit can be adjusted to any angle and turned from side to side while in use. A plastic wedge on the search unit can be used to direct the sound beam through the nozzle body. Correctly angled, the beam will be intercepted by any radial flaw in the nozzle body and can also reach the attachment weld on the far side. The wedge can be shaped on a test block duplicating a nozzle section, and the sound path can be determined by through transmission (one transducer transmitting the signal and another, placed opposite, receiving the signal).

Fig. 9 Typical layout used on large-scale drawings to determine transducer angles and distances for inspection of difficult-to-reach weld areas

Shear wave search units intended for inspection at high temperatures may have the plastic wedge made of a special heatresistant material. Such materials usually increase attenuation, and calibration must be made at temperatures close to that of the surface of the part being inspected if detection of a flaw of given size is to be ensured. Examples of Application. The use of ultrasonics in the detection of discontinuities in a pressure vessel is discussed in

the following examples. Additional information on in-service inspection is presented in the following section of this article.

Example 3: Use of Ultrasonic Inspection to Detect Creep Rupture in Stainless Steel Headers of an Ammonia-Plant Reformer Furnace (Ref 16). A leak was detected in one of the coils in the radiant section of a primary reformer furnace used in an ammonia plant. Furnace temperatures were reduced immediately, but it was necessary to continue firing of the furnace at about 540 to 595 °C (1000 to 1100 °F) for 16 h to reduce the catalyst. Subsequent shutdown inspection revealed that the bottom of one of seven outlet headers had ruptured, causing a section about 100 mm (4 in.) wide by 460 mm (18 in.) long to fall to the furnace floor. The outlet headers were 100 mm (4 in.) nominal diameter schedule 120 (11 mm, or 0.438 in., wall) pipe about 10 m (34 ft) long and were made of ASME SA-452, grade TP316H, stainless steel. In the absence of hot spots, the surface of the outer tube ranged from 845 to 855 °C (1550 to 1575 °F). The unit had been on stream about 29,000 h prior to failure. To get the unit back in service quickly, a section about 1.4 m (4 ft) in length was replaced. The amount of header metal replaced was based on results from both visual and liquid penetrant inspections of the outer surface of the header. The inner surface, where accessible, was also checked visually and by liquid penetrant inspection. The six other headers were inspected at selected locations, and no surface cracks were detected. Investigation of the Removed Section. Metallographic examination of the 1.4 m (4

ft) section of the header that had been removed revealed that it had failed as a result of intergranular fissuring and oxidation, commonly termed creep rupture. Severe intergranular fissuring was found throughout the cross section of the header sample. Primary cracking occurred intergranularly through the grain boundaries, typical of high-temperature fissuring. Close observation of the crack paths linking some of the larger fissures, however, did reveal local areas having grains that were much smaller than

those of the header metal. The presence of these very small grains could only be the result of cold deformation and subsequent recrystallization at elevated temperatures. Therefore, some plastic strain and cracking had occurred at temperatures well below the operating temperature, apparently as a result of thermal expansion and contraction stresses during startups and shutdowns, and then the cold-strained areas had recrystallized subsequently when they were reheated to the service temperature. Therefore, it follows that all of the fissuring did not take place during any one operating period. Also, the microstructure in the ruptured area gave the following evidence that the header metal had not been subjected to gross overheating in service: •

•

The header metal still had a general grain structure that would be classified as being of fine size, while severe grain coarsening would have occurred if the header metal had been heated at temperatures between 1095 and 1315 °C (2000 and 2400 °F) The microconstituents present in the grains would have been dissolved at grossly elevated temperatures and would have been of much finer size if they had subsequently precipitated out of solid solution during the final 16 h of firing at 595 °C (1100 °F) to reduce catalyst

It was concluded from this investigation of the ruptured area that the header had failed by conventional long-time creep rupture as a result of exposure to operating temperatures probably between 900 and 955 °C (1650 and 1750 °F), rather than by short-time exposure at grossly elevated temperatures. It was suspected that the furnace headers might not be as sound as the visual and liquid penetrant inspections of the surface metal had indicated; therefore, three ring sections from the removed 1.4 in (4 ft) long header section were selected for further study. The sections were taken from the ruptured area (sample A), from a slightly bulged but nonruptured area (sample B), and from visually sound metal about 0.6 m (2 ft) from the rupture (sample C). Inspection of the three samples revealed the presence of pinhead-size intergranular fissures throughout the cross sections of samples B and C (Fig. 10). This indicated that the header metal at these locations had been in an advanced stage of creep rupture even though the fissures or voids were of small size. Sample C had fewer pinhead voids than sample B. With continued service, the pinhead voids would have increased in number and grown in size and eventually would have linked up and caused failure. Generally, the microstructure of the inner surface of the header in the area of rupture (sample A, Fig. 10a) contained intergranular fissures (black areas), a relatively fine grain structure, precipitated microconstituents in the grains, and considerable amounts of carbides in the nonfissured grain boundaries. The presence of the carbides was evidence of carbon pickup in service.

Fig. 10 Unetched microscopic appearance of three samples of failed pipe from a reformer-furnace header of TP316H stainless steel pipe and corresponding ultrasound responses of the three samples compared to the response of unused pipe. Micrographs of (a) sample A, (b) sample B, and (c) sample C. All 100×. Black voids are intergranular fissures. Ultrasound responses for (d) unused pipe, (e) sample C, (f) sample B, and (g) sample A. Strong reflections were obtained from the unused pipe, and attenuation increased as the number of fissures increased in damaged pipes.

Sample C had been removed from a visually undamaged header area located about 0.6 m (2 ft) from the rupture and about 250 mm (10 in.) from the sound end of the header section. It was considered possible that the header metal that was in service and only 250 mm (10 in.) away from the area of sample C was also in an advanced stage of creep rupture. It appeared probable that localized areas of the other six headers were in critical condition or that the problem was of a general nature. Therefore, it was decided that the remaining headers in service should be inspected as soon as possible because obtaining replacement parts might require several months of lead time. An ultrasonic attenuation method was considered to be the only possible nondestructive method for this evaluation, and development of a suitable system for use in the field was undertaken. Ultrasonic Inspection Equipment and Standards. It was determined that optimum results that permitted the

determination of changes in attenuation as a function of the number of cracks present required the use of a 22-MHz search unit in conjunction with a pulse-echo instrument. To minimize surface effects and curvature, a short water column was used to couple the sonic energy to the headers. This water column was enclosed in a Lucite cylinder 32 mm (1.25 in.) long with a 6 mm (0.25 in.) thick polyurethane-foam gasket at the bottom. The pliable gasket conformed to the curved

surfaces of the header. The search unit was held in position by a setscrew. Water flowed through the unit by gravity and escaped through a small hole in the Lucite cylinder at a point just above the front surface of the transducer. A section of unused type 316H stainless steel pipe and the three sections of header metal from the rupture area were used for reference. Strong reflections were obtained from the undamaged pipe (Fig. 10); increased attenuation was obtained from damaged piping as the number of cracks increased, as indicated by the ultrasonic responses in Fig. 10(e) to 10(g). The results of the test correlated well with the characterization of the number of cracks observed by optical microscopy (Fig. 10a to 10c). Approximately 1 year later, with 37,336 h on stream, the plant was shut down, and in-service inspection was conducted using the ultrasonic technique developed. In-Service Ultrasonic Inspection. The standards and equipment used to develop the ultrasonic technique were used for the field inspection of the reformer furnace. Ultrasonic readings were taken on the horizontal header metal between each vertical tube and on both sides of any header welds. Over 350 readings were made. In addition, at some locations, readings were taken on the top, bottom, and sides. It was found that the headers all contained internal voids throughout their lengths, of varying numbers between those of samples B and C. The inspection results indicated that all headers were in an advanced stage of creep rupture (stress rupture) but that no areas had fissured to a degree that they needed immediate replacement. The replacement piping that had been installed during the previous shutdown 13 months earlier gave a "good" signal similar to unused material. The technique developed for the field survey was not adaptable for determining the conditions of the welded joints, so none of the readings represents welds.

On the basis of results of the in-service inspection, two conclusions were reached. First, the furnace was deemed serviceable, and second, in the absence of local hot spots, the headers would survive for a reasonable period of time.

Example 4: Ultrasonic Inspection for Creep Fissures in Reformer-Furnace Tubes (Ref 17). Preliminary studies were made to establish whether sound would transmit through cast heat-resistant alloy HK40 tubes from a reformer furnace. Several tests were conducted, and the most likely method was found to be the measurement of attenuation losses in a dual sensor using the through transmission method in an immersion tank. The tests showed that sound would transmit in HK-40 tubes with the use of a low-frequency transducer. Also, the rough outer surface of a centrifugal casting was the significant factor, rather than poor transmission in the large-grain cast microstructure. Immersion methods minimized the surface roughness condition and provided sufficient acoustic energy in the tubes. A dual search unit, with a sending transducer and a receiving transducer to produce a refracted-angle sound beam, was the best means for passing sound through the tube and across the plane of fissure formation. The attenuation of this beam, measured by the receiver, was a measure of fissure density. Figure 11 shows the results of ultrasonic C-scan amplitude recordings on three sample tubes in an immersion tank. By moving the dual search unit at a constant speed longitudinally along the tube and indexing at the end of each traverse, and continuously recording the received ultrasonic signal, attenuation patterns in fissured tubes were recorded. Sample tubes with known fissures showed that some or all of the sound transmission was interrupted, depending on the density of the fissures (compare Fig. 11a and 11b). Sample tubes with no fissures showed full sound transmission (Fig. 11c).

Fig. 11 Ultrasonic C-scan recording of three sample alloy HK-40 tubes from a steam-methane reformer furnace. (a) Severely fissured tube. (b) Tube with small fissures. (c) Tube with no fissures. Light areas represent sound attenuation (no through signal); black areas represent no sound attenuation (through signal).

Strip chart recordings of sound attenuation were made on sections of sample tubes. By evaluation of the strip charts, the tubes were categorized as to sound attenuation and a grading scale of 1 to 5 was arbitrarily established. A rating of 1 represented good sound transmission through unfissured tubes, and a rating of 5 represented little or no sound transmission through severely fissured tubes. Intermediate values represented various degrees of fissuring. Field Tests. On the basis of the preliminary tests, a field unit was built for subsequent furnace inspections that could be

clamped on a tube and mechanically operated. The basic principle of the field unit is shown in Fig. 12.

Fig. 12 Creep fissures in a centrifugally cast HK-40 reformer-furnace tube that are detectable by ultrasonic inspection and by radiography with nickel catalyst in tube. (a) Tube cross section. 0.45×. (b)Tube wall. 2.5×. (c) Enlargement of inside diameter portion of wall shown at bottom in (b). 7.5×

The first full plant test with the unit was conducted with catalyst in the tubes. At that time, the furnace had operated over 50,000 h. The unit was calibrated on sample tubes with known conditions. Two men were in the furnace with voice communications to a third man outside the furnace operating the supporting ultrasonic equipment. With the unit clamped to the tube, a single scan was made at the critical hot area at each level. Ultrasonic scanning time at each location was less than 30 s, during which the attenuation information was recorded on a strip chart for later interpretation. About 32 h was required to inspect the entire furnace. The tubes were graded on the scale of 1 to 5 established during testing of the sample tubes. Of the 295 tubes tested, about 15% had ratings of 5. About 30% of these tubes were replaced; the remainder were left in the furnace for a time-to-failure test. Of the 15 known fissured tubes left in the furnace at the previous shutdown (see Example 2), only 3 showed severe creep fissures by ultrasonic inspection and were replaced. The remaining 12 tubes had ratings of 4. Metallographic examination was made of specimens taken from the removed tubes to determine the reliability of the ultrasonic unit. Results indicated that the unit was so sensitive that it could detect mild third-stage creep (not detectable by radiography) as well as severe fissures (detectable by radiography). Because of the sensitivity of the unit, there was some difficulty in differentiating between mild and severe fissuring. Radiography with catalyst in the tubes can detect only the most severe creep fissures, but it was used as the basis on which tubes were replaced. The tubes with indications of severe creep fissures (ratings of 5) were radiographed. Seven tubes showed definite fissures on the radiographs, even with catalyst in place. These seven tubes and seven additional tubes that showed questionable fissures were replaced. One of the known fissured tubes left in the furnace during the previous shutdown showed questionable fissures on the radiograph taken with catalyst and had an ultrasonic rating of 5. This tube was again allowed to remain in the furnace for a time-to-failure test. Metallographic features of a tube with severe creep fissures that had a

rating of 5 with the ultrasonic unit, subsequently confirmed with radiography (with catalyst in the tube), are shown in Fig. 12. Conclusions. The ultrasonic unit is a good field inspection tool for centrifugally cast alloy HK-40 tubes. Further refinement of the device is necessary to discriminate between mild and severe fissures and thus eliminate the need for radiography to determine the most severe fissures.

An ultrasonic inspection system provides the following advantages: • • • • •

Higher speed Increased coverage at lower cost Increased sensitivity Need for fewer radiographs Elimination of the need for removal of catalyst to effect inspection

In addition, maintenance work need not be interrupted while inspection is in progress, as it must be for radiography.

References cited in this section

15. D.J. Evans, Field Application of Nondestructive Testing in the Petroleum and Petrochemical Industries, in Materials Engineering and Sciences Division Biennial Conference, American Institute of Chemical Engineers, 1970, p 484-487 16. B. Ostrofsky and N.B. Heckler, Detection of Creep Rupture in Ammonia Plant Reformer Headers, in Materials Engineering and Sciences Division Biennial Conference, American Institute of Chemical Engineers, 1970, p 472-476 17. R.R. Dalton, Ultrasonic Inspection of Cast HK-40 Tubes for Creep Fissures, Mater. Eval., Vol 32 (No. 12), Dec 1974, p 264-268 Nondestructive Inspection of Boilers and Pressure Vessels

In-Service Quantitative Evaluation (Ref 18, 19, 20) The structural integrity of the reactor pressure vessel receives considerable attention because the vessel is the primary containment for the reactor coolant. In the United States, periodic in-service examination of the vessel is performed according to section XI of the ASME Boiler and Pressure Vessel Code. Ultrasonic methods of quantitative nondestructive evaluation (NDE) are those most commonly used to accomplish in-service examinations. Nearly all of the examinations are performed with remote-controlled equipment. Many innovative devices and specialized ultrasonic techniques have been developed to examine components with complex geometry, which are often extremely difficult to access. Some of the more difficult areas to examine are the under-clad region nozzle inner radii, nozzle-to-shell welds, dissimilar-metal welds in the safe-ends, and seam welds in areas of complex shape (Ref 18). Fracture mechanics is used to evaluate indications detected during in-service examinations. Accurate measurements of the sizes and locations of all defects are required. Furthermore, the probabilistic failure prediction methodology now being used requires additional information on the probability of detecting flaws with each NDE technique. These requirements are the driving force in the current trend toward additional regulatory requirements for the quantitative demonstration of NDE performance (Ref 18). The probability of detection as a function of flaw size and the accuracy of flaw size measurement are the important characteristics of each NDE technique that are to be measured in performance demonstration (Ref 18). Much of the work to date in the area of performance demonstrations and quantitative NDE of pressure vessels has been carried out at the Electric Power Research Institute (EPRI). Some of the published EPRI work is contained in the extensive list of Selected

References found at the end of this article. Additional information on the principles of quantitative analysis can be found in the Section "Quantitative Nondestructive Evaluation" in this Volume. In designing NDE performance demonstrations, both the examination sensitivity and the mechanical precision

of the scanning device must be addressed to determine if they are adequate to detect and size the defects of concern (Ref 18). Accordingly, the scope of a demonstration can be broken into three conveniently separate parts: • • •

The mechanical handling system The ultrasonic system The data-recording system

Mechanical Handling (Ref 18). The intent of testing the mechanical system is to measure and document the accuracy,

backlash, and repeatability of the complete remote positioning system over its full range of operation. The tolerances of the mechanical system, when combined with those of the ultrasonic system, must give adequate flaw sizing and location capability for any required fracture mechanics analysis. Tests of Ultrasonic and Data-Recording Systems (Ref 18). The intent of these tests is to demonstrate that the

intended inspection procedures are capable of detecting and sizing all flaws of potential concern to the safety of the reactor pressure vessel. The vessel regions to be addressed can be categorized as follows: • • •

The region of the vessel in the vicinity of the cladding/base metal interface The nozzle regions, including nozzle inner radius, nozzle-to-vessel welds, and safe-end welds (up to and including the pipe-to-safe-end weld) The remaining welds that can be further categorized as circumferential or longitudinal welds. These are separately identified because of the different relative directions of weld, cladding, and curvature

Full-sized specimens representing the appropriate component are required. Intentional flaws are introduced of the size and type of concern. For example, these may be thermal fatigue cracks in the nozzle inner radius or intergranular stresscorrosion cracks in the nozzle-to-safe-end weld. The number of defects required in the demonstration is dependent on the purpose of that demonstration. Tests can be classified as either performance demonstration or validation. Additional information can be found in Ref 18 and in the Selected References that follow this article. NDE of Clad Vessels (Ref 19). In the early 1980s, there was much concern in the United States about pressurized

thermal shock, a series of events that started with a considerable primary fluid loss, the addition of cold make-up water, and the subsequent repressurization of the system. Under these conditions, small cracks (6 mm, or 0.25 in., in depth), if present immediately under the stainless steel vessel clad, could act as initiation sites for crack growth. This concern caused many reactor pressure vessel inspection development efforts to be focused on the detection of small cracks in the innermost region of the vessel inner surface. The following sections discuss several advanced systems developed at EPRI for the detection and sizing of flaws, the automatic discrimination of flaw signals, and computer-aided sizing from signals using crack tip diffraction methods. The ultrasonic data recording and processing system (UDRPS) is a high-speed, general-purpose device that

consists of a large minicomputer, high-speed data channel processor, color video display, and disk/tape storage devices. The UDRPS uses a detection criterion that is based on the signal-to-local-noise ratio threshold and the apparent motion of the target within the field of view of the moving transducer. Resulting patterns of indications are color coded for signalto-noise ratio and are viewed by an analyst for the presence of formations suggestive of defects. Crack length and depth are also estimated from several image display modes that are available to the operator. Figure 13 shows the UDRPS results on a test block. Crack length and depth measurement capabilities are shown in Fig. 14.

Fig. 13 UDRPS flaw detection result on a heavy-section test block. Source: Ref 19

Fig. 14 UDRPS estimates of the flaw depths (a) and flaw lengths (b) in a heavy-section test block. Inspection

technique: 45° shear and 60° shear. Source: Ref 19

Flaw discriminators provide the ability to distinguish among ultrasonic signal types. Feasibility studies have been

conducted to demonstrate the use of integrated ultrasonic inspection and pattern recognition systems for distinguishing among slag inclusions, cracks, and spurious clad-noise signals. Pressure Vessel Imaging Systems. The present inspection method for weld zones of nuclear reactor pressure

vessels uses a pulse-echo ultrasonic technique for both preservice and in-service inspections. Pulse-echo inspection data are not sufficiently accurate to satisfy the demands of structural analysis by fracture mechanics methods. This has resulted in the development of imaging systems that combine conventional ultrasonic inspection techniques (B-scan, C-scan, pulse echo) with acoustic holography, which provides real-time three-dimensional estimations of flaw size and depth and more accurate information for fracture mechanics analysis. Computer-aided sizing through crack tip diffraction involves the use of advanced digital processing

methodologies that utilize spectral features of signals diffracted from the under-clad crack tips to distinguish more accurate depth estimates from less reliable ones. This technology is based on the development of sizing algorithms that are available as computer codes.

References cited in this section

18. A.J. Willets, F.V. Ammirato, and J.A. Jones, Objectives and Techniques for Performance of In-Service Examination of Reactor Pressure Vessels, in Performance and Evaluation of Light Water Reactor Pressure Vessels, American Society of Mechanical Engineers, 1987, p 79-86 19. G.J. Dav and M.M. Behravesh, U.S. Developments in the Ultrasonic Examination of Pressure Vessels, Int. J. Pressure Vessels Piping, Vol 28, 1987, p 3-17 20. P.C. Riccardella, J.F. Copeland, and J. Gilman, Evaluation of Flaws or Service Induced Cracks in Pressure Vessels, in Performance and Evaluation at Light Water Reactor Pressure Vessels, American Society of Mechanical Engineers, 1987, p 87-94 Nondestructive Inspection of Boilers and Pressure Vessels

Acoustic Emission Monitoring of Pressure Vessels When discontinuities exist in a metal, a stress concentration occurs at the tips of discontinuities; under increasing applied stress, deformation occurs first at these discontinuities. This deformation, which may be plastic flow, microcracking, or even large-scale cracking, produces signals that, by means of suitable amplification, can be recorded as acoustic emission. The emission is converted to an electrical signal by means of a piezoelectric transducer, which is mounted on the pressure vessel. The transducer is contained in a simple housing, which can be bonded to the vessel with glue or with a film of grease. The signal is then amplified (often using a preamplifier close to the transducer), filtered to remove low-frequency extraneous mechanical and electrical noise, and then recorded. Different types of analysis are used, and the signals can be shaped as pulses to aid quantitative interpretation of data. Stressing of a discontinuity can produce a continuous emission and bursts of high amplitude. Analysis of these signals can be made in any of several ways; for example, each ring of the transducer above a set threshold can be counted or, alternatively, each high-amplitude burst can be counted as a discrete pulse. The number of counts per second or the integrated count is then compared with vessel pressure for strain. Application. A series of search units is used for monitoring the emission of sound energy from stressed pressure vessels. All the outputs can be digitally stored for subsequent analysis. Additionally, some of the circuits can be continuously monitored to give an immediate indication of high acoustic activity and therefore the existence of a severe discontinuity that could cause fracture. Any source of acoustic emission can be located by measurement of the time taken for the signal

to reach different search units at known locations. On-line or subsequent location techniques can be used, preferably in conjunction with a small computer. The number of search units necessary depends on the degree of accuracy required in defining the sources of the emissions. Most of the information in any signal generated from a source within a metal occurs in the 1 to 2 MHz frequency range; however, at these frequencies, the signal attenuation is high, so that the higher frequencies in the stress wave are soon attenuated to near the background noise over relatively short distances (a few meters). The lower-frequency components of the wave in the 50 to 500 kHz frequency range can transmit over larger distances, because of less attenuation. The lower-frequency signals in the 50-kHz range can be detected over relatively large distances on an uncoated surface. This means that some general source location can be performed using the lower-frequency components of the acoustic emission signal for large transducer separations, provided the signal attenuation (which increases when a vessel is coated and/or buried) does not reduce the surface wave to the extent that the true signal is lost in the background noise (Ref 9). Source identification, however, requires the use of wide-band transducers operating at high frequencies and located as close as possible to the source of the emission. Only then is it possible to obtain some indication of the original waveform, and even then it will have been modified by its passage through the material to the surface so that care must be used in interpretation. Other factors influencing the interpretation of results include: • • • •

Defect type, size, shape, orientation, and location Chemical composition of the vessel material Vessel geometry and its associated pipework Epoxy-type coatings

Details on these factors can be found in Ref 9 and in the article "Acoustic Emission Inspection" in this Volume. Proof Testing. Because the acoustic emission technique depends on a changing state of stress, especially around a

discontinuity, the most convenient time for application to most pressure vessels is at the first proof test. With a sufficient number of search units, it is possible to monitor the entire vessel and locate the areas where discontinuities exist. Subsequent ultrasonic inspection can confirm the existence of very small acceptable discontinuities in the position indicated by the acoustic emission technique. Test to Failure (Ref 9). In an attempt to improve the interpretation of acoustic emission data, tests have been conducted

on operating vessels, static vessels, and vessels deliberately tested to failure. Data on wave propagation and failure mechanisms have been recorded and used to develop a reliable acoustic emission integrity evaluation technique. The following example illustrates the use of these tests.

Example 5: Test to Failure of a 10-Year-Old Pressure Vessel (Ref 9). A vessel that had been in service for over 10 years and was operated at a normal pressure of 9.6 MPa (1400 psi) at 550 °C (1020 °F) was tested by acoustic emission and analyzed by subsequent fractographic examination. The defects causing its removal from service were extensive cracks in the nozzle reinforcement. These cracks ran circumferentially around the nozzle penetration, and those on the inside surface of the vessel were up to 25 mm (1 in.) deep. The test program was designed to pressurize the vessel to failure after many pressure cycles designed to accelerate the growth of flaws in the nozzle by cyclic fatigue. Numerous pressurizations were carried out at progressively increasing pressures, including a group of 995 cycles in the range of 7 to 24 MPa (1000 to 3500 psi). The last four cycles were monitored with acoustic emission. The last three cycles were up to 58 MPa (8400 psi), and in the final pressurization, the failure of the vessel occurred at 62 MPa (9000 psi). There was considerable acoustic emission activity detected in the early stages of the test program, but this is thought to have resulted from movement of the vessel supported by observations using television monitors. After the pressure had exceeded 56 MPa (8100 psi), this acoustic emission activity diminished concurrently with the vessel becoming more settled on its supports. Two active source areas were localized at the bottom of adjoining nozzles. It was noted in this test that some of the location patterns were distorted due to some propagation paths being around nozzle holes in the vessel. During the final stages of the test, the acoustic emission event rate further diminished, which was partly due to the reduced pumping rate caused by the vessel expansion.

More acoustic emission activity was detected from the base of the central nozzle (No. 3, Fig. 15) than most of the other locations, and this primary initiation point for failure was later confirmed by independent fractograhy examination.

Fig. 15 Schematic of pressure vessel tested to failure and monitored by acoustic emission. Dimensions given in millimeters. Source: Ref 9

The failure was almost entirely brittle in character and appeared to be from localized exhaustion of ductility, with no apparent involvement of any significant preexisting defect. It appeared to propagate axially along the vessel from the initiation point and then to branch circumferentially between nozzle Nos. 2, 3, and 4, as shown in Fig. 15. Half of the vessel then lifted and separated, causing an axial fracture diametrically opposite the underside of these nozzles. This fracture acted as a hinge, which allowed this section to lift. The only evidence of ductile fracture was in a narrow shear lip adjacent to the external surface of the vessel. The absence of a flaw at the point of initiation was somewhat surprising; however, the highly localized acoustic emission activity and subsequent failure may have resulted from a steep strain gradient adjacent to the reinforced and highly restrained nozzle penetration. The failure of the preexisting cracks to propagate confirms predictions of stress analysis that they were not located in highly stressed regions during simple pressurization. Their formation during service resulted from loadings imposed by external supports and pipework, rather than from internal pressure. The conditions encountered in this test are clearly different from those that would exist in a nondestructive proof test. The vessel suffered extensive plastic deformation at pressure considerably higher than any conceivable proof test pressure. Also, the failure occurred by the local exhaustion of ductility in a highly strained region, rather than by the initiation of yielding in a highly stressed region. However, the ability of the equipment to identify localized acoustic emission sources was clearly demonstrated. Inspection During Fabrication. During fabrication, acoustic monitoring can be used to detect cracking during or after the welding process. Stress-relief cracking can be identified as it occurs, although provision must be made for keeping the transducers at a low temperature. This is normally done by the use of waveguides, the extremities of which hold the transducers outside the stress-relieving furnace. In-service inspection may consist of monitoring during periodic proof testing, during normal pressure cycles, or

continuously during normal operation. When the vessel is pressurized to a level less than that to which it has been previously subjected (during, for example, the proof tests), little or no acoustic emission occurs. Therefore, on subsequent pressurizations a quiet vessel will be obtained unless a crack has extended in service because of corrosion or fatigue. On pressurizing after crack growth, the stress system at the enlarged crack will be changed from that in the proof test, and further emission will be obtained.

Flaw Location. In many cases, especially on large pressure vessels, it becomes necessary not only to detect acoustic

emissions but also to locate the source of the signals. This can be accomplished by uniformly spacing multiple search units over the surface area of a pressure vessel and monitoring the time of arrival of the signals to the various search-unit locations. Because of the high velocity of sound and the relatively close spacing of search units on a steel vessel, time resolutions must be made in microseconds to locate the source to within a centimeter. In most cases, inspection requirements are such that data must be available in a short period of time. Therefore, most systems of this type utilize a computer for handling and displaying the data.

Reference cited in this section

9. B.R.A. Wood, Acoustic Emission Applied to Pressure Vessels, J. Acoust. Emiss., Vol 6 (No. 2), 1989, p 125132 Nondestructive Inspection of Boilers and Pressure Vessels

References 1. 2.

3. 4. 5. 6. 7.

8. 9. 10. 11. 12. 13. 14. 15.

Outlook on Nondestructive Examination, Nucleonics Week, 30 June 1988 ASME Boiler and Pressure Vessel Code: Section II--Material Specifications, Part A--Ferrous Materials; Section III, Division 1--Nuclear Power Plant Components; Section V--Nondestructive Examination; Section VIII--Division 1--Pressure Vessels, Division 2--Alternative Rules for Pressure Vessels; Section IX--Welding and Brazing Qualifications; Section XI--Rules for Inservice Inspection of Nuclear Power Plant Components, American Society of Mechanical Engineers "Recommended Practice for Nondestructive Testing Personnel Certification," SNT-TC-1A, American Society for Nondestructive Testing, 1988 "Standard Recommended Practice for Magnetic Particle Examination," E 709, Annual Book of ASTM Standards, American Society for Testing and Materials "Standard Practice for Liquid Penetrant Inspection Method," E 165, Annual Book of ASTM Standards, American Society for Testing and Materials "Standard Specification for Ultrasonic Angle-Beam Examination of Steel Plates," A 577, Annual Book of ASTM Standards, American Society for Testing and Materials "Standard Specification for Straight-Beam Ultrasonic Examination of Plain and Clad Steel Plates for Special Applications," A 578, Annual Book of ASTM Standards, American Society for Testing and Materials "Standard Specification for Straight-Beam Ultrasonic Examination of Steel Plates," A 435, Annual Book of ASTM Standards, American Society for Testing and Materials B.R.A. Wood, Acoustic Emission Applied to Pressure Vessels, J. Acoust. Emiss., Vol 6 (No. 2), 1989, p 125-132 J.C. Spanner, Acoustic Emission in Pressure Vessels, in Pressure Vessel and Piping Technology--1985: A Decade of Progress, American Society of Mechanical Engineers, 1985, p 613-632 K. Krzywosz, Recent NDE Experiences With PWR Steam Generator Tubing Inspection, in NDE in the Nuclear Industry, ASM INTERNATIONAL, 1987, p 157-167 R.H. Ferris, A.S. Birks, and P.G. Doctor, Qualification of Eddy Current Steam Generator Tube Examination, in NDE in the Nuclear Industry, ASM INTERNATIONAL, 1987, p 71-73 V.S. Cecco and F.L. Sharp, Special Eddy Current Probes for Heat Exchanger Tubing, in NDE in the Nuclear Industry, ASM INTERNATIONAL, 1987, p 109-174 R.R. Dalton, Radiographic Inspection of Cast HK-40 Tubes for Creep Fissures, Mater. Eval., Vol 30 (No. 12), Dec 1972, p 249-253 D.J. Evans, Field Application of Nondestructive Testing in the Petroleum and Petrochemical Industries, in Materials Engineering and Sciences Division Biennial Conference, American Institute of Chemical

16.

17. 18.

19. 20.

Engineers, 1970, p 484-487 B. Ostrofsky and N.B. Heckler, Detection of Creep Rupture in Ammonia Plant Reformer Headers, in Materials Engineering and Sciences Division Biennial Conference, American Institute of Chemical Engineers, 1970, p 472-476 R.R. Dalton, Ultrasonic Inspection of Cast HK-40 Tubes for Creep Fissures, Mater. Eval., Vol 32 (No. 12), Dec 1974, p 264-268 A.J. Willets, F.V. Ammirato, and J.A. Jones, Objectives and Techniques for Performance of In-Service Examination of Reactor Pressure Vessels, in Performance and Evaluation of Light Water Reactor Pressure Vessels, American Society of Mechanical Engineers, 1987, p 79-86 G.J. Dav and M.M. Behravesh, U.S. Developments in the Ultrasonic Examination of Pressure Vessels, Int. J. Pressure Vessels Piping, Vol 28, 1987, p 3-17 P.C. Riccardella, J.F. Copeland, and J. Gilman, Evaluation of Flaws or Service Induced Cracks in Pressure Vessels, in Performance and Evaluation at Light Water Reactor Pressure Vessels, American Society of Mechanical Engineers, 1987, p 87-94

Nondestructive Inspection of Boilers and Pressure Vessels

Selected References Replication Microscopy • J.J. Balaschak and B.M. Strauss, Field Metallography in Assessment of Steam Piping in Older Fossil Power Plants, in Microstructural Science, Vol 15, ASM INTERNATIONAL, 1987, p 27-36 • C.J. Bolton, B.F. Dyson, and K.R. Williams, Metallographic Methods of Determining Residual Creep Life, Mater. Sci. Eng., Vol 46, 1980, p 231-239 • P.B. Ludwigsen, The Replica Method for Inspection of Material Structures and Crack Detection, Structure, No. 15, Sept 1987, p 3-5 • B. Neubauer and U. Wedel, NDT: Replication Avoids Unnecessary Replacement of Power Plant Components, Power Eng., Vol 88 (No. 5), May 1984, p 5 • N. Nilsvang and G. Eggeler, A Quantitative Metallographic Study of Creep Cavitation in a 12% Chromium Ferritic Steel (X20 CrMoV 12 1), Pract. Metallogr., Vol 24, 1987, p 323-335 • E.V. Sullivan, Field Metallography Equipment and Techniques, in Microstructural Science, Vol 15, ASM INTERNATIONAL, 1987, p 3-11 Quantitative NDE (1980) • A.J. Boland et al., "Development of Ultrasonic Tomography for Residual Stress Mapping," EPRI NP1407, Electric Power Research Institute, May 1980 • A.E. Holt, "Defect Characterization by Acoustic Holography--Volume 1: Imaging in Field Environments," EPRI NP-1534, Electric Power Research Institute, Sept 1980 • P.H. Hutton, "Development of an Acoustic Emission Zone Monitor and Recorder for BWR PipeCracking Detection," EPRI NP-1408, Electric Power Research Institute, June 1980 • W. Lord, "Magnetic Flux Leakage for Measurement of Crevice Gap Clearance and Tube Support Plate Inspection," EPRI NP-1427, Electric Power Research Institute, June 1980 • W. Lord and R. Palanisamy, Magnetic Probe Inspection of Steam Generator Tubing, Mater. Eval., Vol 38 (No. 5), May 1980 • V.I. Neeley et al., "Technology Transfer Phase of Advanced Ultrasonic Nuclear Reactor Pressure Vessel Inspection System," EPRI NP-1535, Electric Power Research Institute, Sept 1980 • G.P. Singh and J.L. Rose, "Ultrasonic Field Analysis Program for Transducer Design in the Nuclear Industry," EPRI NP-1335, Electric Power Research Institute, Feb 1980 • S. Wenk et al., "NDE Characteristics of Pipe Weld Defects," EPRI NP-1590-SR, Electric Power Research Institute, Sept 1980

Quantitative NDE (1981) • D.G. Eitzen et al., "Fundamental Development for Quantitative Acoustic Emission Measurements," EPRI NP-2089, Electric Power Research Institute, Oct 1981 • D.G. Eitzen et al., "Summary of Fundamental Development for Quantitative Acoustic Emission Measurements," EPRI NP-1877, Electric Power Research Institute, June 1981 • T.D. Jamison et al., "Studies of Section XI Ultrasonic Repeatability," EPRI NP-1858, Electric Power Research Institute, May 1981 • W. Lord, "Development of a Finite Element Model for Eddy-Current NDT Phenomena," EPRI NP-2026, Electric Power Research Institute, Sept 1981 • W.R. McDearman et al., "Steam Generator Support Plate Radiography," EPRI NP-2042, Electric Power Research Institute, Sept 1981 • E.J. Parent et al., "Profilometry for Steam Generator Tube Dents," EPRI NP-2141, Electric Power Research Institute, Nov 1981 • C.O. Ruud, "Review and Evaluation of Nondestructive Methods for Residual Stress Measurements," EPRI NP-1971, Electric Power Research Institute, Sept 1981 Quantitative NDE (1982) • S. Brown, "Field Experiences With Multifrequency-Multiparameter Eddy Current Technology," EPRI NP-2299, Electric Power Research Institute, March 1982 • W.E. Cramer et al., "Application of an Eddy Current Technique to Steam Generator U-Bend Characterization," EPRI NP-2339, Electric Power Research Institute, April 1982 • G.J. Dau et al., Automatic Analysis of Eddy Current Signals, in Proceedings of the 5th International Conference on Inspection of Pressurized Components, Institute of Mechanical Engineers, 1982 • E.S. Furgason and V.L. Newhouse, "Evaluation of Pulse-Echo Ultrasound for Steam Generator Tube-toSupport Plate Gap Measurement," EPRI NP-2285, Electric Power Research Institute, June 1982 • M.E. Lapides, Factors Influencing Detection, Location, and Sizing of Flaws: Intergranular Stress Corrosion Cracking (IGSCC), in Proceedings of the 5th International Conference on Inspection of Pressurized Components, Institute of Mechanical Engineers, 1982 • M.E. Lapides and T.U. Marston, In-Service Inspection of Heavy Section Castings: Techniques, Results, Implication, in Proceedings of the Seminar on Improvements in Power Plant Casting Quality (St. Charles, IL), American Society for Metals, 1982 • J.L. Rose et al., "A Physically Modelled Feature Based Ultrasonic System for IGSCC Classification," Paper presented at the ASNT Spring Conference (Boston), American Society for Nondestructive Testing, March 1982 Quantitative NDE (1983) • C. Bradshaw, Benefits of Automatic Steam Generator Tube Inspection Data Acquisition, in Proceedings of the SMIRT-7 Post Conference (Monterey, CA), 1983 • S.D. Brown, "Eddy Current NDE for Intergranular Attack," EPRI NP-2862, Electric Power Research Institute, Feb 1983 • S.D. Brown, "Steam Generator U-Bend Eddy Current NDE," EPRI NP-3010, Electric Power Research Institute, April 1983 • M.E. Lapides, "Radiographic Detection of Intergranular Stress Corrosion Cracking: Analysis, Qualification, and Field Testing," EPRI NP-3164-SR, Electric Power Research Institute, Oct 1983 • W. Lord, "Magnetic Flux Leakage for Measurement of Crevice Gap Clearance and Tube Support Plate Inspection," EPRI NP-2857, Electric Power Research Institute, Feb 1983 • W.R. McDearman et al., "Steam Generator Support Plate Radiographic Evaluation System," EPRI NP3253, Electric Power Research Institute, Jan 1983 • V.I. Neeley, "Development of a Production Prototype Pressure Vessel Imaging System," EPRI NP-3253, Electric Power Research Institute, Oct 1983

•

R.B. Thompson et al., "A Prototype EMAT System for Inspection of Steam Generator Tubes," EPRI NP2836, Electric Power Research Institute, Jan 1983

Quantitative NDE (1984) • F.L. Becker, Effective Demonstrations for Under Clad Crack Detection, in Non-Destructive Examination for Pressurised Components, Elsevier, 1984 • R.L. Beverly and R.A. Baker, "Evaluation of Nondestructive Examinations of Intergranular Stress Corrosion Cracking Countermeasures," EPRI NP-3324-LD, Electric Power Research Institute, March 1984 • C.L. Bradshaw, Automatic Denting Analysis of Steam Generator Tubing, in Non-Destructive Examination for Pressurised Components, Elsevier, 1984 • G.J. Dau and M.M. Behravesh, Status of Intergranular Stress Corrosion Crack Depth Sizing, in NonDestructive Examination for Pressurised Components, Elsevier, 1984 • M.E. Lapides, The Challenge of Continuous Flaw Monitoring in Electric Utilities Practice, J. Pressure Vessel Technol., Vol 106, Aug 1984 • M.E. Lapides, MINAC: Portable, High Energy Radiographic Source; Experience and Applications, in Non-Destructive Examination for Pressurised Components, Elsevier, 1984 • M.E. Lapides, Radiographic Detection of Crack-Like Defects in Thick Sections, Mater. Eval., Vol 42 (No. 6), May 1984 • C.R. Mikesell and S.N. Liu et al., Detection of Intergranular Stress Corrosion Cracking Using Automated Ultrasonic Techniques, in Proceedings of the Nondestructive Testing and Electrochemical Methods of Monitoring Corrosion in Industrial Plants, ATME E-7/G-1 Symposium, Montreal, Canada, May 1984 Quantitative NDE (1985) • M.L. Fleming, Inspection of Pipe, Tubing and Plate, in Proceedings of the 1985 ASNT Spring Conference (Washington, D.C.), American Society for Nondestructive Testing, 1985 • M.L. Fleming, Field Experience in Automated Ultrasonic Inspection of Stainless Steel, in Proceedings of the 1985 Pressure Vessels and Piping Conference (New Orleans, LA), American Society of Mechanical Engineers, 1985 • V.I. Neeley, "Application of Medical Ultrasonic Testing Technology to the Utility Industry," EPRI NP4034, Electric Power Research Institute, May 1985 • R.B. Thompson et al., "Ultrasonic Scattering From Intergranular Stress Corrosion Cracks: Derivation and Application of Theory," EPRI NP-3822, Electric Power Research Institute, Jan 1985 Quantitative NDE (1986) • F.V. Ammirato et al., "Development of Improved Procedure for Examination of Dissimilar-Metal Welds in BWR Nozzle-to-Safe-End Welds," EPRI NP-4606-LD, Electric Power Research Institute, May 1986 • M.J. Avioli, Jr., Modeling Ultrasonic Flaw Detection, EPRI J., Sept 1986 • F.L. Becker et al., NDT of Steam Piping and Headers, in Proceedings of the Fossil Plant Inspections Workshop, Electric Power Research Institute, 1986 • E.S. Furgason et al., "Digital Techniques to Improve Flaw Detection by Ultrasound Systems," EPRI NP4878, Electric Power Research Institute, Oct 1986 • B.P. Hildebrand, "Investigation of Advanced Acoustic and Optical Nondestructive Evaluation Techniques," EPRI NP-4897, Electric Power Research Institute, Oct 1986 • K. Krzywosz, Trends and Recent Developments in NDE of Steam Generator Tubes, in Proceedings of the Fifth Annual Steam Generator NDE Workshop (Myrtle Beach, SC), Steam Generator Owners Group II, 1986 • S.M. Walker, Ultrasonic Examination of Corrosion-Resistant Clad Weldments, in Proceedings of the Southwest Research Institute 14th Nuclear Power Educational Seminar, Southwest Research Institute, 1986 • A.J. Willetts et al., "Evaluation of the Ultrasonic Data Recording and Processing System (UDRPS)," EPRI NP-4397, Electric Power Research Institute, Jan 1986

Quantitative NDE (1987) • F. Ammirato and S. Walker et al., Examination of Dissimilar Metal Welds in BWR Nozzle-to-Safe-End Joints, in Proceedings of the 8th International Conference on NDE in the Nuclear Industry (Orlando, FL), ASM INTERNATIONAL, 1987 • L. Becker and S. Walker et al., NDE of Fossil Plant Steam Piping, in Proceedings of the 8th International Conference on NDE in the Nuclear Industry (Orlando, FL), ASM INTERNATIONAL, 1987 • M.M. Behravesh et al., Status of Advanced UT Systems for the Nuclear Industry, Nucl. Eng. Des., No. 102, 1987, p 265-273 • D. Kedem, "Computed Tomography for Thick Steel Pipe and Castings," EPRI NP-5107-LD, Electric Power Research Institute, March 1987 • K. Krzywosz, Recent NDE Experience With PWR Steam Generator Tubing Inspection, in Proceedings of the 8th International Conference on NDE in the Nuclear Industry (Orlando, FL), ASM INTERNATIONAL, 1987 • D. MacDonald and E.K. Kietzman, Comparative Evaluation of Acoustic Holography Systems, in Proceedings of the 8th International Conference on NDE in the Nuclear Industry (Orlando, FL), ASM INTERNATIONAL, 1987 • D. MacDonald and S.M. Walker, "Effects of Ultrasonic Equipment Variations on Crack Length Measurements," EPRI NP-5485, Electric Power Research Institute, Oct 1987 • G. Selby, "Ultrasonic Examination of BWR Replacement Pipe Joint MockUps," EPRI NP-5438-LD, Electric Power Research Institute, Aug 1987 • R.B. Thompson et al., "Modeling Ultrasonic Inspection of Nuclear Components--Beam Models and Applications," EPRI NP-5330, Electric Power Research Institute, Aug 1987 • A.J. Willetts and E.K. Kietzman, UT Techniques for Detection and Sizing of Under-Clad Cracks in Reactor Pressure Vessels, in Proceedings of the 8th International Conference on NDE in the Nuclear Industry (Orlando, FL), ASM INTERNATIONAL, 1987 Quantitative NDE (1988) • F. Ammirato, "NDE of Cast Piping in the Nuclear Industry," Paper presented at the IAEA Specialists Meeting on the Inspection of Austenitic and Dissimilar Metal Welds, Espoo, Finland, International Atomic Energy Agency, June 1988 • F. Ammirato et al., Ultrasonic Examination of Dissimilar-Metal Welds in PWR and BWR Plants, in NonDestructive Examination in Relation to Structural Integrity, Proceedings of the Post-SMiRT Seminar #3, Elsevier, 1988 • L. Goldberg, "Reliability of Magnetic Particle Inspection Performed Through Coatings," EPRI NP-5919, Electric Power Research Institute, July 1988 • J.D. Heald, "Evaluation of a New Gamma-Scanning Method for Detection and Sizing of Intergranular Stress Corrosion Cracking," EPRI NP-5759-LD, Electric Power Research Institute, May 1988 • B.P. Hildebrand, "Stepped Frequency Imaging for Flaw Monitoring," EPRI NP-6033, Electric Power Research Institute, Sept 1988 • B.P. Newberry and R.B. Thompson, Prediction of Surface Induced Ultrasonic Beam Distortions, in Review of Progress in Quantitative NDE-8, Plenum Press, 1988 • J.L. Rose, "Wave-Propagation Studies for Improved Ultrasonic Testing of Centrifugally Cast Stainless Steel," EPRI NP-5979, Electric Power Research Institute, Aug 1988 • J. Saniee, T. Wong, and N.M. Bilgutay, Optimal Ultrasonic Flaw Detection Using a Frequency Diversity Technique, in Review of Progress in Quantitative NDE-8, Plenum Press, 1988 • T. Sasahara and F. Ammirato, "Automated Ultrasonic Pipe Examination and Interpretation," EPRI NP5760, Electric Power Research Institute, April 1988 • R. Shankar, Field Application of Integrated Ultrasonic Feature-Based and Imaging-Based Analysis, in Proceedings of the 9th International Conference on Non-Destructive Evaluation in the Nuclear Industry, ASM INTERNATIONAL, 1988

•

R. Shankar et al., Feature-Enhanced Ultrasonic Imaging--Application of Signal Processing and Analysis, in Non-Destructive Examination in Relation to Structural Integrity, Proceedings of the Post-SMiRT Seminar #3, Elsevier, 1988

Introduction to Quantitative Nondestructive Evaluation Vicki E. Panhuise, Allied-Signal Aerospace Company, Garrett Engine Division

Introduction THE RELIABILITY of a nondestructive inspection (NDI) procedure was defined in the article "Reliability of Flaw Detection by Nondestructive Inspection" in Volume 11 of the 8th Edition of Metals Handbook as a quantitative measure of the efficiency of that procedure in finding flaws of specific type and size. During the years since that article was published, many inspection reliability programs have been conducted, and various quantitative measurements have been cited to express the procedure capabilities. To establish the basis for NDI method reliability, it is necessary to review the history of NDI and its relation to reliability methods.

Historical Development of Quantitative Measurement Techniques Nondestructive inspection methods are specified for material and/or component inspection requirements to maintain the necessary quality for the final service life of the material/component. In most industries, the inspection requirements are defined in a specification that describes the sensitivity level of the inspection method as well as the rejectable flaw size. An example of a specification requirement is given in Table 1; the specification is based on longitudinal wave inspection using flat-bottom holes (FBH). It defines the ultrasonic inspection requirements for product over 13 mm (0.5 in.) thick. Table 1 indicates the defect detection/rejection limits for each quality class material. For example, quality class AA requires that single discontinuities be detected at a sensitivity level equivalent to a No. 3 (1.2 mm, or

in., diam) FBH.

In addition, multiple discontinuities* shall be detected at a sensitivity level equivalent to a No. 1 (0.4 mm, or FBH.

in., diam)

Table 1 Typical NDI acceptance criteria Five classes of ultrasonic quality are established for longitudinal wave inspection. These classes are defined for inspection involving flat-bottom reflectors in ultrasonic references. Quality class

Single discontinuity(a), FBH No.

Multiple discontinuities(a), FBH No.

Maximum linear discontinuity, in.

Maximum loss of back reflection, %

AA

3

1(b)

No. 1 response for 0.12

50

A1

3

2(c)

No. 2 response for 1.00

50

A

5

3

No. 3 response for 1.00

50

B

8

5

No. 5 response for 1.00

50

C

As established by purchaser and vendor for specific part

Source: Ref 1 (a) FBH numbers indicate diameter in multiples of 0.4 mm (

in.) of FBH in ultrasonic reference.

(b) 11% of a No. 3 FBH is equivalent to a No. 1 FBH and can be used in place of the response from the No. 1 FBH.

(c) 44% of a No. 3 FBH is equivalent to a No. 2 FBH and can be used in place of the response from the No. 2 FBH.

Each specification also defines a procedure for demonstrating the capability to detect defects at the required sensitivity levels. In AMS 2630A, this defined procedure is a calibration technique used to set up the ultrasonic instrumentation. Ultrasonic standards have been designed to establish the performance of the inspection system. Procedures have been defined by the American Society for Testing and Materials for the manufacture of these ultrasonic reference blocks for longitudinal wave testing (Ref 2, 3). Finally, the specifications require specific training for the inspectors. For conformance with AMS 2630A, personnel must be certified to and function within the limits of their levels of certification as specified by the American Society for Nondestructive Testing (Ref 4). Further details of ultrasonic inspection procedures are described elsewhere in this Volume. These specification requirements were designed to control the inspection processes and the quality of the inspection results. However, several catastrophic failures of major engineering systems (such as the F-111, space shuttle, nuclear reactors, and the Alaskan pipeline) and the development of advanced materials such as composites were major forces in the development and application of NDI technology (Ref 5). In concert with these events, a new design method using linear elastic fracture mechanics (LEFM) required inspection methodology for detecting defects in production and in service (for a description of LEFM, see Mechanical Testing, Volume 8 of ASM Handbook, formerly 9th Edition Metals Handbook). Linear elastic fracture mechanics design assumes the presence of structural defects and then allows the designer to answer the following questions: • • • •

What is the critical flaw size that will cause failure for a given component subject to service stress and temperature conditions? How long can a precracked structure be safely operated in service? How can a structure be designed to prevent catastrophic failure from preexisting cracks? What inspections must be performed to prevent catastrophic failure?

The ability to answer these questions forms the basis of nondestructive evaluation (NDE), which involves damagetolerant design approaches and is centered on the philosophy of ensuring safe operation in the presence of flaws. The U.S. Air Force has used damage tolerance analysis, as shown in Fig. 1. The initial component in the as-manufactured condition is assumed to have a flaw of length a0. This flaw length is based on the manufacturing inspection capability or material flaw size distribution. The growth of the flaw is predicted for service usage and will reach a critical flaw size, af, after t0 flight hours. The current Air Force philosophy requires an inspection at half the time required for the potential crack to grow to critical size. This inspection is assumed to detect and remove any flaw of size larger than aNDE. The assumption creates the requirement to determine NDE probability of detection (POD) for this current design practice, as described in the articles that follow in this Section.

Fig. 1 Crack growth life curve for damage-tolerant design

The driving functions described above caused the NDE/NDI community to evaluate the inspection capabilities. The first evaluation in the aerospace industry was conducted by Lockheed in the 1970s and was called the "Have Cracks, Will Travel" program (Ref 6). This program was established to determine if airframe inspection would repeatedly detect cracks so that engineers could use LEFM philosophy for design. The study concluded that the overall reliability of NDI performed by the Air Force and evaluated in the program falls below that which has been previously assumed by established guidelines (Ref 6). The most significant result of the program was that the 90/95 reliability criteria could not be attained for any flaw size with typical inspection techniques applied by the average technician. Many advancements have been made in NDE/NDI technology to improve the inspection capability since this study. These technologies are described throughout this Volume.

References cited in this section

1. "Ultrasonic Inspection of Product over 0.5 in. (13 mm) Thick," Aerospace Material Specification 2630A, Society of Automotive Engineers, 1980 (original release 1960) 2. "Standard Practice for Fabricating and Checking Aluminum Alloy Ultrasonic Standard Reference Blocks," E 127, Annual Book of ASTM Standards, American Society for Testing and Materials 3. "Standard Recommended Practice for Fabrication and Control of Steel Reference Blocks Used in Ultrasonic Inspection," E 428, Annual Book of ASTM Standards, American Society for Testing and Materials 4. "Recommended Practice, Personnel Qualification and Certification in NDT," SNT-TC-IA, American Society for Nondestructive Testing 5. W. Rummel, Recommended Practice for a Demonstration of Nondestructive Evaluation (NDE) Reliability on Aircraft Production Parts, Mater. Eval.,Vol 40 (No. 9), Aug 1982, p 922 6. W.H. Lewis, et al., "Reliability of Nondestructive Inspections--Final Report," Report SA-ALC/MME 76-638-1, United States Air Force, Air Logistics Center, Kelly Air Force Base, 1978 Note cited in this section

*

Multiple discontinuities are defined in AMS 2630A as two or more indications above the level established for the class that occur within 16 cm3 (1.0 in.3) of the inspected surface.

Introduction to Quantitative Nondestructive Evaluation Vicki E. Panhuise, Allied-Signal Aerospace Company, Garrett Engine Division

NDE Reliability The following article in this Section--"Fracture Control Philosophy"--describes the U.S. Air Force philosophy for current design practices. Damage-tolerant design requires knowledge of the reliability of the inspection technique used to detect flaws or damage. In the initial stages of using this design approach, a one-number characterization of NDE capability was in use. The one-number characterization was the minimum crack (flaw) length for which there is a fixed degree of confidence that at least a fixed probability of cracks will be detected. Typically, the minimum crack length was chosen such that, at the 95% confidence level (CL), at least 90% of all cracks greater than this length will be detected. The number is referred to as the 90/95 (POD/CL) crack length. Probability of detection functions for describing the reliability of an NDE technique have been the subject of many studies. The ideal inspection system POD function is shown schematically in Fig. 2. All flaws larger than aNDE would be detected all of the time, while all flaws smaller than aNDE would not. Obviously, no ideal system exists, and the POD functions used produce a continuous curve. Figure 3 shows a POD curve for an automated eddy current inspection of titanium bolt holes. Two confidence level POD curves are shown, 50 and 95%. Berens and Hovey completed a study that compared POD analysis techniques to determine the optimum calculation method (Ref 7). Additional information concerning the analysis of NDE data for the determination of reliability is available in the article "NDE Reliability Data Analysis" in this Section.

Fig. 2 Schematic of probability of detection curves

Fig. 3 POD curve for the automated eddy current inspection of titanium bolt holes

The accuracy of the POD function is dependent on the demonstration program design. The data acquired during the demonstration program testing can influence the resulting prediction of reliability. For example, consider the cases presented in Fig. 4, which graphically presents the specimen/crack depth histogram for two case studies. If the design requirement is such that aNDE (that is, 90/95 crack depth) must be 250 μm (10 mils) or less, only the specimens in case study No. 2 could demonstrate this capability. That is, to demonstrate 90/95 crack size, the specimens fabricated for the demonstration program must have cracks less than aNDE and some larger. If it is assumed that all the flaws were detected in both cases, the 90/95 crack sizes predicted are as follows:

Case study

1

90/95 crack size

μm

mils

417

16.40

Fig. 4 Specimen crack depth distribution for two case studies. (a) Case study No. 1. (b) Case study No. 2

This case study, which demonstrates the criticality of planning the reliability demonstration program, showed only the importance of specimen flaw size to the final outcome of the reliability study. In the following articles in this Section, fracture control philosophy is discussed, reliability demonstrations that have been completed are reviewed, and the analysis of NDE data is examined. The final article in this Section, "Models for Predicting NDE Reliability," provides information on advanced modeling studies for predicting reliability for the inspection of a particular component. The objective of this Section is to provide the necessary information and resources to conduct an effective NDE reliability program.

Reference cited in this section

7. A.P. Berens and P.W. Hovey, "Flaw Detection Reliability Criteria Volume I--Methods and Results, Final Report," Report AFWAL-TR-84-4022 Volume I, United States Air Force Material Laboratory, WrightPatterson Air Force Base, 1984 Introduction to Quantitative Nondestructive Evaluation Vicki E. Panhuise, Allied-Signal Aerospace Company, Garrett Engine Division

References 1. "Ultrasonic Inspection of Product over 0.5 in. (13 mm) Thick," Aerospace Material Specification 2630A, Society of Automotive Engineers, 1980 (original release 1960) 2. "Standard Practice for Fabricating and Checking Aluminum Alloy Ultrasonic Standard Reference Blocks," E 127, Annual Book of ASTM Standards, American Society for Testing and Materials

3. "Standard Recommended Practice for Fabrication and Control of Steel Reference Blocks Used in Ultrasonic Inspection," E 428, Annual Book of ASTM Standards, American Society for Testing and Materials 4. "Recommended Practice, Personnel Qualification and Certification in NDT," SNT-TC-IA, American Society for Nondestructive Testing 5. W. Rummel, Recommended Practice for a Demonstration of Nondestructive Evaluation (NDE) Reliability on Aircraft Production Parts, Mater. Eval.,Vol 40 (No. 9), Aug 1982, p 922 6. W.H. Lewis, et al., "Reliability of Nondestructive Inspections--Final Report," Report SA-ALC/MME 76-6-38-1, United States Air Force, Air Logistics Center, Kelly Air Force Base, 1978 7. A.P. Berens and P.W. Hovey, "Flaw Detection Reliability Criteria Volume I--Methods and Results, Final Report," Report AFWAL-TR-84-4022 Volume I, United States Air Force Material Laboratory, Wright-Patterson Air Force Base, 1984 Fracture Control Philosophy William D. Cowie, United States Air Force, Aeronautical System Division, Propulsion Directorate

Introduction FRACTURE CONTROL PHILOSOPHIES are being used in the design, development, and life management of United States Air Force (USAF) turbine engine and airframe components. This article describes the fracture control program for turbine engine components. The section "Applications (Case Studies)" of the article "Applications of NDE Reliability to Systems" in this Volume provides an overview of fracture control programs for both airframe and turbine engine components. The establishment of a fracture control philosophy and the implementation of a fracture control program have been integral components of the USAF turbine engine development process since 1978. They have been applied to new engine programs as part of the USAF Engine Structural Integrity Program (ENSIP) described in military standard MIL-STD1783 (issued formally in 1984). This military standard was reviewed and approved by the Aerospace Industries Association of America in 1982. Fracture control philosophy has also been applied to existing inventory USAF engines through structural durability and damage tolerance assessments. In all, fracture control programs have been applied or are being applied to the F-100, TF34, F100-PW-220, F100-PW-229, F110-GE-100, F110-GE-129, F101-GE-102, F109-GA-100, F-119, F120, and T406 engines and have resulted in the implementation of enhanced nondestructive evaluation (NDE) methods (for example, eddy current inspection) at manufacturing and at field/depot. These inspections have been successful in detecting early cracking and in accelerating corrective actions. Several developmental efforts in the last 5 years have identified fluorescent penetrant inspection process improvements that must be implemented within industry and Air Force depots to improve flaw detection reliability. The need to quantify detection reliability for imbedded defects is also identified. The engine development process has been evolutionary in terms of the application of upgraded requirements. The new process of fracture control, sometimes referred to as damage tolerance, is contained in ENSIP, and it involves material selection as well as design and life management. Recent experience clearly demonstrates that the damage tolerance requirement is cost effective when assessed on a life cycle basis. Fracture Control Philosophy William D. Cowie, United States Air Force, Aeronautical System Division, Propulsion Directorate

Overview of ENSIP

In the past 16 years, a large number of structural problems have occurred in USAF gas turbine engines. Many of these were safety problems that resulted in loss of aircraft, and an even greater number affected durability, causing a high level of maintenance and modification costs. All of these problems have adversely affected fleet readiness. The Engine Structural Integrity Program was intended to reduce these problems substantially and was developed based on the following specific lessons: • • • • • • • •

It is unrealistic (and can be dangerous) to assume defect-free structure in safety-of-flight components Critical parts (and part details) and potential failure modes must be identified early and appropriate control measures implemented Internal thermal and vibratory environments must be identified early in the engine development Predicted analytical stresses must be verified by test for complex components Materials and processes must be adequately characterized (particularly, the fracture properties) Design stress spectra, component test spectra, and full-scale engine test spectra must be based on the anticipated service usage of the engine, that is, accelerated mission-related testing Potential engine/airframe structural interactions must be defined and accounted for Management procedures (such as individual engine tracking procedures and realistic inspection and maintenance requirements) must be defined and enforced

The Engine Structural Integrity Program was established by the Air Force to provide an organized and disciplined approach to the structural design, analysis, development, production, and life management of gas turbine engines, with the goal of ensuring engine structural safety, increased service readiness, and reduced life cycle costs. The five major tasks associated with ENSIP are the development of design information; design analysis and component and material characterization; component and core engine testing; ground and flight engine testing; and production quality control and engine life management. Each major task is subdivided into a number of subtasks (Table 1) that guide the development process. Table 1 Tasks of the engine structural integrity program Task I: Design information Development plans

ENSIP master Durability and damage control Material and process characterization Corrosion prevention and control Inspection and diagnostics

Operational requirements

Design service life and usage requirements Design criteria

Task II: Design analysis and material characterization and development tests Design duty cycle Material characterization Design development tests Structural/thermal analysis Installed engine inspectability Manufacturing and quality control

Task III: Component and core engine tests Component tests

Strength Vibration Damage tolerance Durability

Core engine tests

Thermal survey Vibration strain and flutter boundary survey

Task IV: Ground and flight engine tests Ground engine tests

Strength Damage tolerance Accelerated mission test Thermal survey Vibration strain and flutter boundary survey

Flight engine tests

Fan strain survey Thermal survey Installed vibration Deterioration

Task V: Engine life management Updated analyses Structural maintenance plan Operational usage survey Individual engine tracking Durability and damage tolerance control actions (production)

The Engine Structural Maintenance Plan represents the output of the ENSIP program. This plan identifies and defines individual part life limits, the necessary inspection periods for each fracture-critical part, and the inspection procedure. The basic components of the Engine Structural Maintenance Plan are as follows:

• • •

Structural safety is obtained in ENSIP by requiring a structure with a damage tolerance that is capable of accommodating flaws induced either in manufacture or service Durability design requirements stipulate that the economic life of the engine must exceed the specified design service life of the aircraft when flown to the design usage spectra Maintainability criteria require that old parts fit and function with new parts, that repair life be defined,

• • • •

and that inspectability and structural diagnostics be designed into the engine and its components A materials and process characterization plan controls materials development through key engine development points Environmental definition requirements specify the thermal, dynamic, and steady-state stress; the stress spectra; and the component sensitivities A comprehensive ground test policy is utilized to ensure compliance with safety, durability, and maintainability requirements A usage and tracking policy is used to form the basis of an engine life management program

Fracture Control Philosophy William D. Cowie, United States Air Force, Aeronautical System Division, Propulsion Directorate

Overview of ENSIP In the past 16 years, a large number of structural problems have occurred in USAF gas turbine engines. Many of these were safety problems that resulted in loss of aircraft, and an even greater number affected durability, causing a high level of maintenance and modification costs. All of these problems have adversely affected fleet readiness. The Engine Structural Integrity Program was intended to reduce these problems substantially and was developed based on the following specific lessons: • • • • • • • •

It is unrealistic (and can be dangerous) to assume defect-free structure in safety-of-flight components Critical parts (and part details) and potential failure modes must be identified early and appropriate control measures implemented Internal thermal and vibratory environments must be identified early in the engine development Predicted analytical stresses must be verified by test for complex components Materials and processes must be adequately characterized (particularly, the fracture properties) Design stress spectra, component test spectra, and full-scale engine test spectra must be based on the anticipated service usage of the engine, that is, accelerated mission-related testing Potential engine/airframe structural interactions must be defined and accounted for Management procedures (such as individual engine tracking procedures and realistic inspection and maintenance requirements) must be defined and enforced

The Engine Structural Integrity Program was established by the Air Force to provide an organized and disciplined approach to the structural design, analysis, development, production, and life management of gas turbine engines, with the goal of ensuring engine structural safety, increased service readiness, and reduced life cycle costs. The five major tasks associated with ENSIP are the development of design information; design analysis and component and material characterization; component and core engine testing; ground and flight engine testing; and production quality control and engine life management. Each major task is subdivided into a number of subtasks (Table 1) that guide the development process. Table 1 Tasks of the engine structural integrity program Task I: Design information Development plans

ENSIP master Durability and damage control Material and process characterization Corrosion prevention and control

Inspection and diagnostics

Operational requirements

Design service life and usage requirements Design criteria

Task II: Design analysis and material characterization and development tests Design duty cycle Material characterization Design development tests Structural/thermal analysis Installed engine inspectability Manufacturing and quality control

Task III: Component and core engine tests Component tests

Strength Vibration Damage tolerance Durability

Core engine tests

Thermal survey Vibration strain and flutter boundary survey

Task IV: Ground and flight engine tests Ground engine tests

Strength Damage tolerance Accelerated mission test Thermal survey Vibration strain and flutter boundary survey

Flight engine tests

Fan strain survey Thermal survey Installed vibration

Task V: Engine life management Updated analyses Structural maintenance plan Operational usage survey Individual engine tracking Durability and damage tolerance control actions (production)

The Engine Structural Maintenance Plan represents the output of the ENSIP program. This plan identifies and defines individual part life limits, the necessary inspection periods for each fracture-critical part, and the inspection procedure. The basic components of the Engine Structural Maintenance Plan are as follows:

• • • • • • •

Structural safety is obtained in ENSIP by requiring a structure with a damage tolerance that is capable of accommodating flaws induced either in manufacture or service Durability design requirements stipulate that the economic life of the engine must exceed the specified design service life of the aircraft when flown to the design usage spectra Maintainability criteria require that old parts fit and function with new parts, that repair life be defined, and that inspectability and structural diagnostics be designed into the engine and its components A materials and process characterization plan controls materials development through key engine development points Environmental definition requirements specify the thermal, dynamic, and steady-state stress; the stress spectra; and the component sensitivities A comprehensive ground test policy is utilized to ensure compliance with safety, durability, and maintainability requirements A usage and tracking policy is used to form the basis of an engine life management program

Fracture Control Philosophy William D. Cowie, United States Air Force, Aeronautical System Division, Propulsion Directorate

ENSIP and Fracture Control Philosophy Policy Damage tolerance is defined as the ability of the engine to resist failure due to the presence of flaws, cracks, or other

damage for a specified period of usage. The damage tolerance or fracture control philosophy used in ENSIP is shown in Fig. 4. Components are designed for crack growth so that the safety limit exceeds two times the required inspection interval. The safety limit or residual life is the time for assumed initial flaws to grow and cause failure. Because the requirement is to inspect at one-half the safety limit, the design goal for the safety limit is two times the required design life (that is, no inspections). The minimum design requirement for the safety limit is two times the planned depot visit interval. An important aspect of the damage tolerance requirement is that it applies only to fracture-critical components.

Fig. 4 Damage tolerance approach to life management of cyclic-limited engine components. The safety limit or residual life is the time for the initial flaw to grow and cause failure. The size of the initial flaw, ai, is based on the inspection method or material defect distribution (for imbedded defects).

Fracture-critical components are defined as those components whose failure will result in probable loss of the aircraft due to noncontainment or, for single-engine aircraft, power loss that presents sustained flight because of direct part failure or by causing other progressive part failures. Damage tolerance requirements are applied only to fracturecritical components (that is, components that must maintain their integrity during flight) and not, in general, to durabilitycritical components (that is, components that affect maintenance schedules). As expected, component classification is affected by aircraft engine configuration (single engine or multiengine). Component classification is established early and is identified in the contract. Initial Flaw Size. Initial flaws are assumed to exist in fracture-critical components. Experience has shown that

premature cracking (that is, crack initiation prior to the LCF limit) occurs at high-stress areas and where components initially contained both material- and manufacturing-related quality variations (voids, inclusions, machining marks, scratches, sharp cracks, and so on). The fracture control or damage tolerance requirement assumes a sharp crack as the initial flaw when characterizing these abnormal initial conditions. The assumed initial imbedded flaw sizes are based on the intrinsic material defect distribution or the NDE methods to be used during manufacture. The assumed surface flaw size also depends on the NDE capability. An inspection reliability of 90% probability of detection (POD) at the lowerbound 95% confidence level (CL) is required for the assumed initial flaw sizes. The assumed initial flaw size to account for intrinsic material defect distribution should encompass 99.99% of the defect population if a scatter factor of two is used to establish the inspection interval, or 99.9% if a scatter factor of one is used. If embedded defects cannot be inspected in service, the 99.99 percentile (or the 99.9 percentile) is used to satisfy the design life requirement. An initial flaw size not less than 0.75 mm (0.030 in.) in length (for surface flaws) or 0.4 × 0.4 mm (0.015 × 0.015 in.) in size (corner cracks) for nonconcentrated stress areas (bores, webs, and so on) is required. Initial flaw sizes for other surface locations (holes, fillets, scallops, and so on) will be consistent with the demonstrated capability (90% POD/95% CL) of the inspection systems proposed for use. It is recommended that the initial design and sizing of components be based on 0.75 mm (0.030 in.) long surface flaws or 0.4 × 0.4 mm (0.015 × 0.015 in.) corner cracks at all locations. This design recommendation is based on the initial flaw size that can be detected with fluorescent penetrant inspection. This includes fully automated fluorescent penetrant inspection systems that are being developed to meet the 0.75 mm (0.030 in.) and 0.4 × 0.4 mm (0.015 × 0.015 in.) inspection criteria. These flaw sizes are intended to represent the maximum size of the damage that can be present in a critical location after manufacture and/or inspection. The specification of these flaw sizes is based on the demonstrated flaw detection capability of the nondestructive inspection (NDI) method. During design of the components, the assumed initial flaw size that is appropriate for various NDI methods is:

• • • • • •

0.75 mm (0.030 in.) surface length where the NDI method is fluorescent penetrant inspection 0.25 mm (0.010 in.) surface length where the NDI method is eddy current or ultrasonic inspection 1.3 mm2 (0.002 in.2) area for imbedded defects utilizing ultrasonic inspection 5 mm (0.200 in.) surface length and imbedded sphere = 0.2 × thickness for weldments When initial flaw sizes are based on material defect distribution, selected size shall encompass 99.99% of the distribution Demonstration that assumed flaw sizes can be reliably detected with a 90% POD and a 95% CL

The capabilities of the NDI method must be demonstrated by the contractor. The design of NDE reliability experiments is discussed in the article "NDE Reliability Data Analysis" in this Volume. Residual strength is defined as the load-carrying capability of a component at any time during the service exposure

period, considering the presence of damage and accounting for the growth of damage as a function of exposure time. The requirement is to provide limit load residual strength capability throughout the service life of the component. In other words, the minimum residual strength for each component (and location) must be equal to the maximum stress that occurs within the applicable stress spectra based on the design duty cycle. Normal or expected overspeed due to control system tolerance and engine deterioration is included in the residual strength requirement, but fail-safe conditions, such as burst margin, are excluded. The residual strength requirement is illustrated in Fig. 5. Inspection Intervals. It is highly desirable to have no

damage tolerance inspections required during the design lifetime of the engine. This in-service noninspectable classification requires that components be designed such that the residual life or safety limit be twice the design life. Designing components as in-service noninspectable is a requirement for those components or locations that cannot be inspected during the depot maintenance cycle. However, the weight penalty incurred to achieve a safety limit/residual life/damage growth interval twice the design life may be prohibitive on some components/locations. Therefore, in-service inspections will be allowed on some components subject to justification. The basis for the justification is characterization of the costs as a function of the requirements as established by trade studies. Cost is usually expressed in terms of weight or life cycle cost, and the requirement in terms of safety limit/residual life/damage growth interval.

Fig. 5 Diagram of the residual strength requirement

The depot or base-level inspection interval for damage tolerance considerations should be compatible with the overall engine maintenance plan. Once again, it is highly desirable that the inspection interval be equal to the design service life of the parts in the hot gas path (that is, the hotpart design service life, which is equal to one-half the design lifetime of the engine) because this is the expected minimum depot or maintenance interval for the engine or module. It is required that the minimum damage tolerance inspection interval be contained in the contract

specification. Flaw Growth. It is required that the assumed initial flaw sizes will not grow to critical size and cause failure of a

component due to the application of the required residual strength load in two times the inspection interval. The flaw growth interval is set equal to two times the inspection interval to provide a margin for a variability that exists in the total process (that is, inspection reliability, material properties, usage, stress predictions, and so on). Factors other than two should be used when individual assessments of the variables that affect crack growth can be made (for example, to account for observed scatter in crack growth during testing).

It is important that the effects of vibratory stress on unstable crack growth be accounted for in establishing the safety limit. Experience shows that the threshold crack size can be significantly less than the critical crack size associated with the material fracture toughness, depending on the material, the major stress cycle, and the vibration stress. As shown in Fig. 6, the conventional Goodman diagram may not disclose the true sensitivity of initial defects to vibratory stresses. The threshold crack size must be established at each individual sustained-power condition (idle, cruise, intermediate) using the appropriate values of steady stress and vibratory stress. The smallest threshold crack size will be used as a limiting value in calculating the safety limit if it is less than the critical crack size associated with the material fracture toughness.

Fig. 6 Interaction of vibratory stress and initial flaws. (a) Large vibratory stress required to initiate crack. (b) Low vibratory stress will propagate cracks. The crack growth threshold, At, represents the threshold of vibratory motion that will cause the growth of a given crack size.

Fracture Control Philosophy William D. Cowie, United States Air Force, Aeronautical System Division, Propulsion Directorate

Fracture Control Verification Verification that the fracture control policy is met is accomplished by the development and implementation of a Damage Tolerance Control Plan, by analysis and test, and by the implementation of reliable inspection methods during manufacture and field/depot maintenance. A Damage Tolerance Control Plan is prepared that identifies and schedules each of the tasks and interfaces in the functional areas of design, materials selection, tests, manufacturing control, and inspection. Specific tasks that are addressed in the Damage Tolerance Control Plan are:

• • • • • • • • •

Trade studies for design concepts/material/weight/performance/cost Analysis Development and qualification tests Fracture-critical parts list Zoning of drawings Basic materials fracture data Material properties controls Traceability NDI requirements

Most of the tasks to be contained in the Damage Tolerance Control Plan have been accomplished by engine manufacturers in past development and production programs. However, the damage tolerance requirement now established by the Air Force imposes the need for new tasks as well as tighter controls and more involvement among the functional areas. Experience indicates that the development and implementation of a Damage Tolerance Control Plan are difficult, but experience also shows very strongly that the development of a plan results in an improved understanding of what must be done. The importance of having a plan rests on the involvement of multiple functional areas and the criticality of having assigned responsibilities for each task. A particularly important part of the Damage Tolerance Control Plan is the requirement for early trade studies for design concepts/materials/weight/performance/cost. These trade studies are critical to defining cost versus requirement (for example, weight impact versus inspection interval). Analysis and Test Correlation Requirement. The procedure for the analysis of damage tolerance includes the

development of stress spectra (Fig. 7) and the analysis and testing of residual life (Fig. 8). Particular emphasis is placed on establishing a correlation between analytical predictions and test measurements for the growth of cracks in critical areas. Refined analysis models that predict the stress state at and away from the surface, as well as multiple cyclic tests of coupons, subcomponents, and full-scale components in the presence of initial damage, are required. As with the overall ENSIP development philosophy, damage tolerance analysis and tests are conducted early with minimum impact. Test requirements include: •

• • •

Specimen tests to define basic material fracture data, such as the plane-strain fracture toughness (KIc), the plane-stress fracture toughness (Kc), the threshold stress intensity for stress-corrosion cracking (KIscc), and the crack growth curve, which defines the crack growth rate da/dN (where a is the crack length and N is the number of cycles) Subcomponent tests to evaluate crack growth at typical critical features, such as bolt holes and fillet radii Spin pit test of full-scale components, such as disks, spacers, or rotors Engine testing of preflawed components

Fig. 7 Development of stress spectra

Fig. 8 Residual life analysis and test procedure

A typical test configuration for a cyclic spin test is shown in Fig. 9, and typical empirical data are shown in Fig. 10, 11, 12. Verification requirements may also include engine test with components that are preflawed or cracked in critical locations to determine the effects of the real environment (temperature and gradient, vibration, and so on). Such tests must be closely controlled and monitored using the inspection requirements planned for service engines to ensure safety of the test engine.

Fig. 9 Typical setup for a cyclic spin test (heated spin pit) of turbine rotor. Arrows indicate critical locations that

were preflawed to simulate the worst expected damage.

Fig. 10 Crack progagation as influenced by increases in three parameters. (a) Frequency, f. (b) Stress ratio, R. (c) Temperature, T. The variable K is the calculated stress at the crack tip.

Fig. 11 Actual and predicted flaw growth for an engine test on a second-stage high-pressure turbine disk forward cooling air hole

Fig. 12 Methodology components)

correlation

(specimens

and

Nondestructive evaluation requirements are implemented on fracture-critical parts during manufacture and during

field/depot inspection to ensure safety. Specific inspection requirements are derived through design analysis trade-offs among initial flaw size assumption, stress level, and material properties for a given usage (stress environment spectrum). As discussed in the section "Initial Flaw Size" in this article, a flaw size assumption (for surface flaws) of less than 0.75 mm (0.030 in.) requires implementation of enhanced NDE (that is, eddy current inspection). The ability of eddy current inspection to reliably detect surface flaws having depths of 0.13 mm (0.005 in.) has been demonstrated in several applications. Primary emphasis on the use of eddy current inspection is for stress concentration areas in which a small flaw size assumption is required to achieve the necessary residual life without excessive weight penalty. Typical probability of detection data for eddy current inspection in Fig. 13.

Fig. 13 POD curves with eddy current inspection. All curves are for lower 95% confidence limit.

In general, fluorescent penetrant inspection is specified for areas in which the detection of a flaw with a surface length of 0.75 mm (0.030 in.) or larger is required to achieve the specified residual life. However, the ability of current fluorescent penetrant inspection processes to reliably detect 0.75 mm (0.030 in.) flaws is not clear. Therefore, in some cases, eddy current inspection may be specified for large areas if susceptibility data (that is, probabilistic data on the capability of fluorescent penetrant inspection) indicate the need. Data generated on numerous demonstration programs clearly indicate that the fluorescent penetrant inspection process can be significantly improved through upgraded training, equipment, and procedures (proper cleaning, including etch, hydrophilic emulsifier, and wet developer). These demonstration programs have been conducted on several engine development programs (F100, F101, F110, F110-GE-129, F100-PW-229) and laboratory technology programs (Air Force Wright Research and Development Center). Some typical detection improvements that have been demonstrated for upgraded fluorescent penetrant inspection processes are shown in Fig. 14. The critical need is to implement the best fluorescent penetrant inspection process within industry and within the Air Force logistics centers because this method will likely remain the most widely used for inspecting large areas for cracklike damage.

Fig. 14 Probability of detection with fluorescent penetrant inspection. LCL, lower confidence limit

Another critical NDE need is to quantify the POD of ultrasonics to detect imbedded defects in bulk volumes and to develop inspection methods for finished shapes. Very limited data indicate that reliable detection limits may be as large as 1.3 mm2 (0.002 in.2) (approximately equal to a planar disk of 1.2 mm, or in., diameter). The goal is to develop and implement ultrasonic inspection methods such that a residual life equal to two times the required life or the inspection interval can be achieved, assuming the largest undetectable flaw size without excessive impact on weight. Retirement-for-Cause (RFC). Traditionally, components whose dominant failure mode is low-cycle fatigue have

been designed to a crack initiation criterion. With this approach, only 1 component in a population of 1000 would have actually initiated a crack, and the remaining 999 components would be discarded with substantial undefined useful life to crack initiation remaining. Figure 15 shows that the difference between the number of cycles to reach the designallowable curve and the population average curve for an average component would have consumed only 10% or less of its potential useful life-to-crack initiation. Under the initiation criterion, there is no way to utilize this potential life without accepting a higher probability of failure of the remaining components.

Fig. 15 Stress versus loading cycles to crack initiation for Inconel 718. Temperature is 540 °C (1000 °F), and the ratio of alternating stress to mean stress (A ratio) = 1.

Under the RFC concept, this additional useful life can be utilized by adopting a rejection criterion that uses each component in a population until it specifically initiates a crack, rather than rejecting the entire population on the behavior of the statistical minimum. The development of fracture mechanics concepts over the last several years has permitted the degree of predictability for crack progagation rates necessary to implement such an approach on a safe basis. The RFC concept would apply the fracture control philosophy (Fig. 4) to life management. In using RFC as an operating system, all components would be inspected first at the end of a safety limit period divided by an appropriate safety margin, and only those components containing detectable cracks equal to or greater than ai would be retired or repaired. All others would be returned for additional service (with the assumption that if a flaw existed, it would be smaller than ai for another inspection interval). In this way, the crack progagation residual life is continually reset to a safe value. By following this approach, components are rejected only for cause (cracks), and each component is allowed to operate for its own specific crack initiation life. It should be noted that if a crack is missed at the first inspection interval, another chance should exist to find a larger crack. It is clear that not all fatigue-limited components can be handled in this way and that each component must be evaluated individually to determine the technical feasibility of RFC. Low-cycle fatigue is a real physical phenomenon that is not directly associated with the presence of defects. Any criterion, initiation or otherwise, that allows components to run beyond a design-allowable life (-3 , for example) or beyond the average life will inherently result in a significant increase in the probability of a large number of cracks and possible failures. Such an increase in risk may be acceptable, but must be understood and evaluated. Any inspection process has associated with it a finite probability of detection and therefore a finite probability of missing real cracks (not just defects assumed to be there and assumed to be in the form of sharp cracks). Missing real cracks and presuming that crack growth knowledge is sufficient to detect cracks at the next inspection has significantly more risk associated with it than concern over possible defects. The economic feasibility of RFC must also be evaluated. The inspection interval must be such that it does not place undue constraints on the operation of the component or that the cost of the necessary tear-down and inspection does not negate the advantage of the life extension. It seems unlikely that RFC can be applied to components limited by high-cycle fatigue considerations, but for many high-cost components limited by low-cycle fatigue, such as engine disks, this approach does offer significant economic advantages. It is also clear that in applying RFC, nondestructive evaluation becomes a critical factor. The crack length determines the residual life of the component, and its detection is limited by the resolution and reliability of the inspection system employed. In many cases, the decision as to whether or not RFC can be applied to a component will be predicated upon the ability of available NDE approaches to detect the initial flaw, ai, with sufficient sensitivity and reliability. Because RFC procedure is based on fracture control concepts, the NDE techniques can be selected, refined, and focused on a particular local area, rather than attempting to critically inspect large areas.

Fracture Control Philosophy William D. Cowie, United States Air Force, Aeronautical System Division, Propulsion Directorate

Selected References •

•

•

W.D. Cowie and T.A. Stein, "Damage Tolerant Design and Test Considerations in the Engine Structural Integrity Program," Paper presented at the propulsion session of the 21st Structures, Structural Dynamics, and Materials Conference, Seattle, WA, American Institute of Aeronautics and Astronautics, May 1980 T.T. King (ASD/EN), W.D. Cowie (ASD/YZEE), and W.H. Reimann (AFWL/MLLN), "Damage Tolerance Design Concepts for Military Engines," Paper presented at AGARD Conference 393, San Antonio, TX, Advisory Group for Aerospace Research and Development, April 1985 C.F. Tiffany and W.D. Cowie, "Progress on the ENSIP Approach to Improved Structural Integrity in Gas Turbine Engine/An Overview," Paper 78-WA/GT-13, presented at the Winter Annual Meeting (San Francisco), American Society of Mechanical Engineers, Dec 1978

Applications of NDE Reliability to Systems Ward D. Rummel, Martin Marietta Astronautics Group; Grover L. Hardy and Thomas D. Cooper, Wright Research & Development Center, Wright-Patterson Air Force Base

Introduction THE SUCCESS of a reliable NDE application depends greatly on the expertise and thoroughness of the NDE engineering that is performed. This involves comprehensive analyses to define the relationships between the NDE measurements to be made and the impact on the system functions being assessed and on the capability to implement the NDE measurements to discriminate to the expected level of acceptance applied. Most failures in NDE systems applications and in the automation of an NDE system can be attributed to failures in NDE engineering and to unrealistic NDE performance expectations. All modern engineering is based on comprehensive applications of principles that can be implemented by qualitative measurements and predictive modeling and on comprehensive procedural applications of principles based on prior art and systems test data. Modern engineering methods are equally applicable to the use of NDE on a material, component, structure, or system. The principles and data available are, however, not as well defined, generally recognized, or understood as those of other engineering disciplines. Applications of NDE Reliability to Systems Ward D. Rummel, Martin Marietta Astronautics Group; Grover L. Hardy and Thomas D. Cooper, Wright Research & Development Center, Wright-Patterson Air Force Base

General Considerations Prior Art. Particular concern must be given to NDE assessment and analyses based on prior art. Although NDE methods

and principles have been applied since the beginning of time and have been applied specifically to quantitative materials evaluation during the last decade, the perceived performance level is often considerably different from the actual performance level. Differences can be attributed to the primitive level of understanding of materials, engineering, and NDE engineering sciences and principles; excessive optimism in effecting early application of an NDE procedure; economic and social pressure to solve problems with troublesome engineering systems; and the attitudes and practices of our legal system. Progressive developments in the evaluation of NDE reliability have established a new dimension for the assessment of NDE performance. Applications of prior NDE art must be judiciously examined to determine suitability for new applications. Quantification of the performance level, calibration (or process control) methods, and acceptance criteria are

particularly important in extrapolating the applicability of prior NDE data to a new NDE engineering problem. Conversely, the lessons learned in applications of prior art can be very useful in design and calibration and process control procedures, in establishing the characteristics and performance of materials and equipment used, in anticipating the problems and controls necessary to effect application in a production environment, and in assessing human factors relevant to the application. Incorporation and integration of the qualitative factors and considerations in the application of prior art are essential for making the transition from laboratory test data to production line use. However, careful analysis and criteria must be applied to quantitative data from application of prior art. Quantitative assessment of ongoing NDE process applications has shown that performance levels may vary considerably in NDE applications to established specification/process requirements. Performance variations are rarely integrated into overall system reliability estimates and management is rarely accurate in recognizing and identifying superior performance by human operators. Quantitative data must be supported by actual NDE system measurements and accurate descriptions of NDE materials, equipment, procedures, and human operator qualifications to be seriously considered in NDE system design or qualification by similarity. It must be emphasized that quantitative data are not necessary for the multitude of NDE applications used to add confidence to routine process control and other applications in which the NDE procedure does not constitute final acceptance of performance characteristics. Quantitative data are required when the NDE measurement/acceptance is integral to design acceptance and/or performance acceptance. NDE Response. The response from an NDE system or process may take the form of a signal output (or outputs) or a

direct or indirect image. Acceptable conditions can be differentiated from unacceptable conditions by threshold discrimination from the electronic output or by pattern recognition and threshold discrimination by image analyses. Discrimination can be automated or performed by a human operator. Discrimination of threshold electronic signals can be automated or gated to alert the human operator. The consistency and reliability of electronic signal discrimination can often be improved by automating the discrimination process. Superior consistency and reliability of pattern recognition and discrimination level for images are usually achieved by the human operator. The feasibility of application of NDE to a system is dependent on the establishment and characterization of a relationship between the response from an NDE output and a desired engineering system performance parameter. A direct or indirect relationship between an NDE response and a system performance characteristic may be functional under laboratory conditions, but may be impractical in applications under production or service conditions. Factors such as calibration, acceptance criteria, component accessibility, surface condition, inspection material compatibility, and inspection environment must be assessed to determine that a positive relationship between NDE response and system performance can be maintained. NDE System Management and Schedule. The implementation of a reliable NDE procedure is dependent on

allowing time to collect data, perform the critical analyses, apply required resources, and develop operator (personnel) skills. Many critical NDE procedures have been implemented as a result of unanticipated engineering system failures. The economic and social pressures resulting from an engineering system failure must be judiciously balanced against the required time and resources necessary to perform disciplined and thorough NDE engineering analyses, procedure development, and procedure validations. After the required procedures have been implemented, NDE system/process control must be maintained to ensure a consistent level of discrimination. Shortcuts in NDE engineering, NDE procedure development, and NDE system/process control increase risks in system performance, may not reduce the risk of engineering system failure, and may contribute to a false confidence level in system performance. Applications of NDE Reliability to Systems Ward D. Rummel, Martin Marietta Astronautics Group; Grover L. Hardy and Thomas D. Cooper, Wright Research & Development Center, Wright-Patterson Air Force Base

NDE Engineering The difference between NDE engineering and the classical engineering disciplines result from the variety of problems and the indirect nature of NDE measurements on engineering system performance. The functional performance of most NDE methods can be measured, controlling parameters can be documented, and performance output can be modeled; however, the interaction of the NDE method with the test object necessitates the generation of new response parameters and

characteristics for many new applications. In addition, variations in material properties, geometry, surface condition, access, or environmental conditions may modify the NDE responses. Therefore, NDE engineering is an essential element of critical engineering system design, qualification, acceptance, and life cycle management. A critical design is not complete until the NDE engineering has been performed and NDE system/process performance validated to functional design requirements and acceptance criteria levels. Procedure Selection/Development. Trade studies to identify and select candidate NDE procedures are needed for establishing the most economical and reliable procedure that meets acceptance requirements. The process may be satisfied by the assessment of prior art and applications to similar problems or may require research and development of totally new procedures. For demanding applications, a combination of complementary NDE procedures may be required to meet the acceptance criteria objectives. One or more methods can be further characterized and assessed to ensure that the performance objectives, NDE performance margins, NDE costs, engineering system performance risks, and the risks of NDE system/process false alarms can be balanced in overall engineering system management. System/Process Performance Characteristics. Although care, discipline, and control measures are applied to

ensure a consistent output from an NDE system or process, the output will vary within the established control parameters and as a result of slight variations in engineering hardware materials properties, geometry, surface condition, and so on. If repetitive applications are made, a probability density distribution of signal/image output will be generated. This distribution is similar to that obtained by repetitive measurements of a dimension such as a hole diameter or the length of a bolt. Nondestructive measurements are usually indirect, and positive signals may be generated from nonrelevant sources, such as surface roughness, grain structure, and geometry variations. Such signals constitute the application noise inherent in a specific NDE process or procedure. Discrimination of NDE signal/image outputs must be derived from those signal levels/amplitudes that exceed the level of the application noise (Fig. 1). Analysis of signal and signal plus noise are common in electronic devices, optics, and other discrimination processes. Similarly, the signal-to-noise margin (ratio) is a primary factor in establishing the level of discrimination of an NDE procedure. Signal/noise reduction procedures can be used to enhance the overall performance of an NDE procedure. However, it is important to recognize that the dominant noise source in an NDE process is not electronic noise that may be reduced by filtering, multiple sampling, and averaging techniques, but is instead the noise due to nonrelevant signals generated in applying the NDE procedure to a specific hardware element.

Fig. 1 Signal/noise density distribution for a large flaw (a), a medium flaw (b), and a small flaw (c)

Conditional Probability in NDE Discrimination. Nondestructive evaluation involves the measurement of complex

parameters with inherent variations in both the measurement process and the test object. The output from such a measurement/decision process can be analyzed as a problem in conditional probability. When an NDE assessment is performed for the purpose of crack detection, the outcome is not a simple accept/reject (binary) process, as is frequently envisioned. It is actually the product of conditional acceptance due to the interdependence of the measurement and decision responses. Figure 2 shows the four possible outcomes that result from the application of NDE procedure for crack detection. As shown in Fig. 2, the possible outcomes from an inspection process are: •

True positive (TP): A crack exists and is detected, where M(A,a) is the total number of true positives and P(A,a) is the probability of a true positive

• • •

False positive (FP): No crack exists but one is identified, where M(A,n) is the total number of false positives and P(A,n) is the probability of a false positive False negative (FN): A crack exists but is not detected, where M(N,a) is the total number of false negatives and P(N,a) is the probability of a false negative True negative (TN): No crack exists and none is detected, where M(N,n) is the total number of true negatives and P(N,n) is the probability of a true negative

The interdependence of these matrix quantities can be expressed as:

M(A,a) + M(N,a) = (TP) and (FN) outcomes giving the total opportunities for positive calls (Total number of defects) and

M(A,n) + M(N,n) = Total opportunities for false alarms from the possible (FP) and (TN) outcomes

Fig. 2 Matrix of four possible outcomes from an NDE procedure for flaw detection

Because of the interdependent relationship, only two independent probabilities need be considered to quantify the inspection/decision task. The probability of detection (POD) or probability for a true positive P(A,a) can be expressed as:

Similarly, the probability of false alarms (POFA) or the probability for a false positive P(A,n) can be expressed as:

Signal/Noise Relationships. The desired results of the application of NDE procedure are crack detection (signal present) or crack nondetection (signal absent). The basis for detection is that of sensing a signal response and determining that the signal response is above a predetermined threshold. Both sensing and interpretation are dependent on the signal (plus noise) and the noise (application background or response to nonrelevant parameters) that are subjected to the discrimination media (programmed machine discriminator or human operator).

If an NDE procedure is repetitively applied to a single flaw of a given size (in a part with a fixed geometry, surface condition, and so on), the output responses can be used to plot probability density distributions of both flaw signal and application noise responses. Under ideal conditions, such as the response from a large flaw, the signal and noise distributions will be well separated, as shown in Fig. 1(a). The discrimination of flaw responses from application noise responses is a simple process; POD will be high, and the POFA will be low. In practical engineering applications, the flaw size is not fixed (and is rarely large), and the discrimination process is more complex. Indeed, the discrimination process is applied to a continuous range of flaw sizes, where the capability for discrimination is dependent on the inherent performance characteristic of the NDE procedure and on the separation of the signal (plus noise) from the noise response of the process. If the NDE procedure is applied to a single flaw of intermediate size (in a part with the same fixed surface finish, geometry, and so on), the output responses can be used to generate probability density distributions for signal and noise, as shown in Fig. 1(b). For this flaw size, the distributions overlap (in part), and the capability for discrimination is dependent on the response from a single set of output signals within these distributions. If the single set has output signals that are well separated (that is, signals at the outer extremes of the distributions), the output response will be interpreted as acceptable (no flaw condition) for those cases where the threshold response acceptance level is located between the signal and noise signals. If the single set of outputs lies at the inner extremes of the distributions, the output response may be interpreted as acceptable (no flaw or undetected flaw condition) or may be interpreted as unacceptable (false alarm condition) for the same threshold response acceptance level. For this condition, the POD will be lower and the POFA will be higher than for the case of discrimination with positive signal/noise separation margins. If the process is repeated for a small flaw (under the same operating conditions), the signal and noise response distribution will approach coincidence, as shown in Fig. 1(c). The POD will be low, and the POFA will be high. It is clear that the performance capability of a given NDE procedure is dependent on the nature and distribution of the signal outputs generated under the conditions of application. It is also clear that the threshold acceptance criterion applied in the discrimination process is an important factor in the successful application of a procedure. Consider the application of an NDE procedure to a large flaw under conditions that produce a significant separation of probability density distributions of signal and noise, as shown in Fig. 3. If the threshold acceptance criterion (represented by the vertical arrow) is placed at too high a level (Fig. 3a), some of the flaws will be missed (reduce POD). If the acceptance criterion is placed at a proper level (Fig. 3b), clear discrimination will result (high POD). If the acceptance criterion is placed too low (Fig. 3c), all of the flaws will be rejected; but some false alarms will result, and good parts will be rejected (high POFA).

Fig. 3 Influence of acceptance criterion (vertical arrow) on process discrimination. (a) Acceptance criterion too high. (b) Acceptance criterion at proper level. (c) Acceptance criterion too low

The NDE procedure performance characteristics of primary importance are the signal-to-noise ratio (separation margin) and the threshold acceptance criteria applied in the discrimination process. Optimum NDE procedure performance can be obtained by characterizing an NDE procedure and by matching the threshold acceptance criteria to the performance capabilities of the NDE procedure. Such characterization also enables the assessment and quantification of risks that result from changes in acceptance criteria. Reference Standards. Historically, reference standards for NDE methods of defect detection have been used to ensure

the reproducibility of the application of the method(s) and to establish an acceptable quality of the process rather than to establish the dimensions or other applicable parameter of the defect or anomaly. Some methods (such as those used for thickness gaging or electrical conductivity determination) were able to provide an extremely accurate measurement of the appropriate parameter, usually by extrapolation between two known and closely spaced reference standards representative of the condition to be determined. The remaining methods, as typified by radiography, offer at best only a crude estimate of the dimensions, orientation, shape, or other characteristics of the detected defects. The primary reason for not using reference standards for the quantitative evaluation of defects was, and still is, that the NDE methods respond to most of the parameters of a defect simultaneously; in most cases, there is no way to separate the response from a single parameter, or there is not an accurate response to a single parameter in other cases. Consider, for example, the case of penetrant inspection. The indications formed are usually greater in length, width, or area than the discontinuity present because of the flow of the penetrant material out of the discontinuity during development. When this excess material is removed and the indication is viewed as it starts to appear again, the full length of the discontinuity, if it is linear, will initially not be revealed, because the ends provide little or no penetrant for formation of the indication. As the indication continues to form, it will eventually reach the same length as the discontinuity and will then continue to grow as additional penetrant flows to the surface. This lack of response to the extremities of a discontinuity is common to all NDE methods and illustrates the difficulties associated with using NDE methods for sizing defects. Quantitative NDE, however, requires that a good estimate be made of the defect size that is detected or, more important, the size of the largest defect that might be left in the part. Because the NDE methods, by the laws of physics, are

inherently inaccurate in sizing, the only available approach is to make a conservative estimate of the size of the defects that can remain. The approach requires a second type of standard used along with the conventional reference standards. This second type, called a qualification standard, contains defects that represent the worst case for both flaw detection and crack growth (generally a surface fatigue crack). Fatigue cracks have the advantages that they can be grown in the laboratory and, when produced under well-controlled conditions, have predictable geometries. The qualification standards are then used in sets to define the lower limit of the flaw size that a given NDE process can reliably detect. The conventional reference standards are used to control the NDE process; therefore, once the qualification standards establish the process sensitivity, the reference standards can be used to ensure that the sensitivity is maintained. Consequently, there are certain requirements that reference standards must meet. First, reference standards must produce a response comparable to that produced by the smallest qualification standard flaw that is considered reliably detectable. This comparability includes not only response to the flaw itself but also the geometry in which the flaw is contained. In the ultrasonic method, for example, the response to a fatigue crack located in the center of a flat plate can be far different from the response to the same size crack that is located in the bore of a largediameter hole. Consequently, application of the ultrasonic method for inspection of the two geometries requires qualification standards as well as reference standards for both geometries. Second, reference standards for a specific inspection must be relatable to other reference standards used for the same inspection. That is, when several reference standards are available, the responses for each one must be known, and more important, the differences in responses for the standards must be known so that adjustments can be made in the inspection to ensure that a uniform process sensitivity can be maintained. Ideally, the responses of all reference standards for a given inspection should be identical; however, from a practical standpoint this is impossible to achieve. As stated previously, the response of the NDE methods is from a multitude of parameters associated with a given discontinuity. For example, if an electrically discharge machined slot is selected as a defect for a reference standard for an ultrasonic inspection, exact control over the size of the slot in each standard is not sufficient to guarantee identical responses. Slight variations in the orientation of the slot with respect to the surface of the standard and in the surface finish of the slot itself can cause noticeable differences in the ultrasonic response. This is the worst case; other NDE methods vary in their response to subtle geometric parameters, with the magnetic particle and penetrant methods probably being the most tolerant. However, even these methods are highly sensitive to the width of the flaw used in a reference standard. Another important property of reference standards is durability. Both the material and the type of flaw in a reference standard must be selected so that the standard will not readily deteriorate or change in the environment in which it will be used. These selections are affected by the NDE method for which the standard is intended. Both ultrasonic and penetrant methods are very sensitive to the presence of foreign material inside the flaw. For ultrasonics, this can affect the amount of energy that is reflected from the flaw. For penetrant inspection, the quantity of penetrant that can enter the defect, and consequently the brightness of the indication, will be reduced. The foreign material may be fluids, soils, or corrosion products. For these reasons, magnesium, ferritic steels, and aluminum are particularly poor choices for reference standards for penetrant inspection and require some type of protection if used as ultrasonic standards. Eddy current methods are not affected by foreign material in the flaw, but are very sensitive to such surface conditions as scratches, pitting, and corrosion. Magnetic particle methods are sensitive to the width of the defect and to anything, such as cold working, that may change the magnetic permeability of the standard. Radiographic methods are sensitive to the thickness of the standard and to changes in the shape of the flaw. In general, a good choice of material for any method is one that is reasonably hard and forms an adherent, tough, and stable surface oxide layer (such as a titanium or nickel-base alloy), thus providing protection against mechanical damage and the gradual buildup of corrosion products. Personnel. Unless the inspection process is fully automated, the proficiency of the inspection personnel is the largest

variable affecting inspection reliability. This proficiency varies widely not only from inspector to inspector but also with the same inspector, depending on his working environment and his mental condition. For fully automated inspections, the proficiency of the inspector in operating the equipment is important but has little or no impact on inspection reliability. The first task in addressing the contribution of the inspector to inspection reliability is to ensure that he is knowledgeable of the specific techniques to be used and has the basic proficiency to perform the inspection to the required reliability. Experience has demonstrated that the previous qualifications of the inspector--for example, certification to MIL-STD410D (Ref 1)--are not sufficient to ensure the desired performance with a new inspection that must be performed with high reliability. Therefore, the most straightforward way to assess proficiency is to require inspector participation in the demonstration of inspection reliability. All inspectors that will be required to perform the inspection should also participate. This not only establishes the reliability of the proposed inspection but also identifies those personnel requiring

additional training or experience before they can be expected to perform adequately. Careful observation of the inspector during the demonstration and of the results obtained is necessary to identify the additional training or experience needed. After training and/or additional experience is acquired, the demonstration effort can be repeated to indicate if the inspector has become sufficiently proficient in the inspection technique. After basic inspection proficiency has been demonstrated, it becomes a supervisory task to ensure that this proficiency is maintained. Control of the work environment of the inspector is important. Distractions such as noise, extremes in temperature, and other irritants should be eliminated to the extent possible. Break periods should be frequent enough to reduce fatigue. Personnel who are ill or otherwise physically impaired should be temporarily assigned other tasks to the extent possible. Other efforts that improve or maintain a good mental attitude are excellent ways to ensure sustained inspection reliability. These include providing acceptable materials and equipment with which to conduct the inspection. Finally, when it is not possible to provide a consistently conducive environment for a highly reliable inspection, two inspectors can perform the same inspection independently to achieve higher reliability than can be obtained with a single inspector. Two inspectors generally will not make identical mistakes.

Reference cited in this section

1. "Nondestructive Testing Personnel Qualification and Certification," MIL-STD-410D, 25 June 1974 Applications of NDE Reliability to Systems Ward D. Rummel, Martin Marietta Astronautics Group; Grover L. Hardy and Thomas D. Cooper, Wright Research & Development Center, Wright-Patterson Air Force Base

NDE Engineering The difference between NDE engineering and the classical engineering disciplines result from the variety of problems and the indirect nature of NDE measurements on engineering system performance. The functional performance of most NDE methods can be measured, controlling parameters can be documented, and performance output can be modeled; however, the interaction of the NDE method with the test object necessitates the generation of new response parameters and characteristics for many new applications. In addition, variations in material properties, geometry, surface condition, access, or environmental conditions may modify the NDE responses. Therefore, NDE engineering is an essential element of critical engineering system design, qualification, acceptance, and life cycle management. A critical design is not complete until the NDE engineering has been performed and NDE system/process performance validated to functional design requirements and acceptance criteria levels. Procedure Selection/Development. Trade studies to identify and select candidate NDE procedures are needed for establishing the most economical and reliable procedure that meets acceptance requirements. The process may be satisfied by the assessment of prior art and applications to similar problems or may require research and development of totally new procedures. For demanding applications, a combination of complementary NDE procedures may be required to meet the acceptance criteria objectives. One or more methods can be further characterized and assessed to ensure that the performance objectives, NDE performance margins, NDE costs, engineering system performance risks, and the risks of NDE system/process false alarms can be balanced in overall engineering system management. System/Process Performance Characteristics. Although care, discipline, and control measures are applied to

ensure a consistent output from an NDE system or process, the output will vary within the established control parameters and as a result of slight variations in engineering hardware materials properties, geometry, surface condition, and so on. If repetitive applications are made, a probability density distribution of signal/image output will be generated. This distribution is similar to that obtained by repetitive measurements of a dimension such as a hole diameter or the length of a bolt. Nondestructive measurements are usually indirect, and positive signals may be generated from nonrelevant sources, such as surface roughness, grain structure, and geometry variations. Such signals constitute the application noise inherent in a specific NDE process or procedure. Discrimination of NDE signal/image outputs must be derived from those signal

levels/amplitudes that exceed the level of the application noise (Fig. 1). Analysis of signal and signal plus noise are common in electronic devices, optics, and other discrimination processes. Similarly, the signal-to-noise margin (ratio) is a primary factor in establishing the level of discrimination of an NDE procedure. Signal/noise reduction procedures can be used to enhance the overall performance of an NDE procedure. However, it is important to recognize that the dominant noise source in an NDE process is not electronic noise that may be reduced by filtering, multiple sampling, and averaging techniques, but is instead the noise due to nonrelevant signals generated in applying the NDE procedure to a specific hardware element.

Fig. 1 Signal/noise density distribution for a large flaw (a), a medium flaw (b), and a small flaw (c)

Conditional Probability in NDE Discrimination. Nondestructive evaluation involves the measurement of complex

parameters with inherent variations in both the measurement process and the test object. The output from such a measurement/decision process can be analyzed as a problem in conditional probability. When an NDE assessment is performed for the purpose of crack detection, the outcome is not a simple accept/reject (binary) process, as is frequently envisioned. It is actually the product of conditional acceptance due to the interdependence of the measurement and decision responses. Figure 2 shows the four possible outcomes that result from the application of NDE procedure for crack detection. As shown in Fig. 2, the possible outcomes from an inspection process are: • • • •

True positive (TP): A crack exists and is detected, where M(A,a) is the total number of true positives and P(A,a) is the probability of a true positive False positive (FP): No crack exists but one is identified, where M(A,n) is the total number of false positives and P(A,n) is the probability of a false positive False negative (FN): A crack exists but is not detected, where M(N,a) is the total number of false negatives and P(N,a) is the probability of a false negative True negative (TN): No crack exists and none is detected, where M(N,n) is the total number of true negatives and P(N,n) is the probability of a true negative

The interdependence of these matrix quantities can be expressed as:

M(A,a) + M(N,a) = (TP) and (FN) outcomes giving the total opportunities for positive calls (Total number of defects) and

M(A,n) + M(N,n) = Total opportunities for false alarms from the possible (FP) and (TN) outcomes

Fig. 2 Matrix of four possible outcomes from an NDE procedure for flaw detection

Because of the interdependent relationship, only two independent probabilities need be considered to quantify the inspection/decision task. The probability of detection (POD) or probability for a true positive P(A,a) can be expressed as:

Similarly, the probability of false alarms (POFA) or the probability for a false positive P(A,n) can be expressed as:

Signal/Noise Relationships. The desired results of the application of NDE procedure are crack detection (signal present) or crack nondetection (signal absent). The basis for detection is that of sensing a signal response and determining that the signal response is above a predetermined threshold. Both sensing and interpretation are dependent on the signal (plus noise) and the noise (application background or response to nonrelevant parameters) that are subjected to the discrimination media (programmed machine discriminator or human operator).

If an NDE procedure is repetitively applied to a single flaw of a given size (in a part with a fixed geometry, surface condition, and so on), the output responses can be used to plot probability density distributions of both flaw signal and application noise responses. Under ideal conditions, such as the response from a large flaw, the signal and noise distributions will be well separated, as shown in Fig. 1(a). The discrimination of flaw responses from application noise responses is a simple process; POD will be high, and the POFA will be low. In practical engineering applications, the flaw size is not fixed (and is rarely large), and the discrimination process is more complex. Indeed, the discrimination process is applied to a continuous range of flaw sizes, where the capability for discrimination is dependent on the inherent performance characteristic of the NDE procedure and on the separation of the signal (plus noise) from the noise response of the process.

If the NDE procedure is applied to a single flaw of intermediate size (in a part with the same fixed surface finish, geometry, and so on), the output responses can be used to generate probability density distributions for signal and noise, as shown in Fig. 1(b). For this flaw size, the distributions overlap (in part), and the capability for discrimination is dependent on the response from a single set of output signals within these distributions. If the single set has output signals that are well separated (that is, signals at the outer extremes of the distributions), the output response will be interpreted as acceptable (no flaw condition) for those cases where the threshold response acceptance level is located between the signal and noise signals. If the single set of outputs lies at the inner extremes of the distributions, the output response may be interpreted as acceptable (no flaw or undetected flaw condition) or may be interpreted as unacceptable (false alarm condition) for the same threshold response acceptance level. For this condition, the POD will be lower and the POFA will be higher than for the case of discrimination with positive signal/noise separation margins. If the process is repeated for a small flaw (under the same operating conditions), the signal and noise response distribution will approach coincidence, as shown in Fig. 1(c). The POD will be low, and the POFA will be high. It is clear that the performance capability of a given NDE procedure is dependent on the nature and distribution of the signal outputs generated under the conditions of application. It is also clear that the threshold acceptance criterion applied in the discrimination process is an important factor in the successful application of a procedure. Consider the application of an NDE procedure to a large flaw under conditions that produce a significant separation of probability density distributions of signal and noise, as shown in Fig. 3. If the threshold acceptance criterion (represented by the vertical arrow) is placed at too high a level (Fig. 3a), some of the flaws will be missed (reduce POD). If the acceptance criterion is placed at a proper level (Fig. 3b), clear discrimination will result (high POD). If the acceptance criterion is placed too low (Fig. 3c), all of the flaws will be rejected; but some false alarms will result, and good parts will be rejected (high POFA).

Fig. 3 Influence of acceptance criterion (vertical arrow) on process discrimination. (a) Acceptance criterion too high. (b) Acceptance criterion at proper level. (c) Acceptance criterion too low

The NDE procedure performance characteristics of primary importance are the signal-to-noise ratio (separation margin) and the threshold acceptance criteria applied in the discrimination process. Optimum NDE procedure performance can be obtained by characterizing an NDE procedure and by matching the threshold acceptance criteria to the performance capabilities of the NDE procedure. Such characterization also enables the assessment and quantification of risks that result from changes in acceptance criteria.

Reference Standards. Historically, reference standards for NDE methods of defect detection have been used to ensure

the reproducibility of the application of the method(s) and to establish an acceptable quality of the process rather than to establish the dimensions or other applicable parameter of the defect or anomaly. Some methods (such as those used for thickness gaging or electrical conductivity determination) were able to provide an extremely accurate measurement of the appropriate parameter, usually by extrapolation between two known and closely spaced reference standards representative of the condition to be determined. The remaining methods, as typified by radiography, offer at best only a crude estimate of the dimensions, orientation, shape, or other characteristics of the detected defects. The primary reason for not using reference standards for the quantitative evaluation of defects was, and still is, that the NDE methods respond to most of the parameters of a defect simultaneously; in most cases, there is no way to separate the response from a single parameter, or there is not an accurate response to a single parameter in other cases. Consider, for example, the case of penetrant inspection. The indications formed are usually greater in length, width, or area than the discontinuity present because of the flow of the penetrant material out of the discontinuity during development. When this excess material is removed and the indication is viewed as it starts to appear again, the full length of the discontinuity, if it is linear, will initially not be revealed, because the ends provide little or no penetrant for formation of the indication. As the indication continues to form, it will eventually reach the same length as the discontinuity and will then continue to grow as additional penetrant flows to the surface. This lack of response to the extremities of a discontinuity is common to all NDE methods and illustrates the difficulties associated with using NDE methods for sizing defects. Quantitative NDE, however, requires that a good estimate be made of the defect size that is detected or, more important, the size of the largest defect that might be left in the part. Because the NDE methods, by the laws of physics, are inherently inaccurate in sizing, the only available approach is to make a conservative estimate of the size of the defects that can remain. The approach requires a second type of standard used along with the conventional reference standards. This second type, called a qualification standard, contains defects that represent the worst case for both flaw detection and crack growth (generally a surface fatigue crack). Fatigue cracks have the advantages that they can be grown in the laboratory and, when produced under well-controlled conditions, have predictable geometries. The qualification standards are then used in sets to define the lower limit of the flaw size that a given NDE process can reliably detect. The conventional reference standards are used to control the NDE process; therefore, once the qualification standards establish the process sensitivity, the reference standards can be used to ensure that the sensitivity is maintained. Consequently, there are certain requirements that reference standards must meet. First, reference standards must produce a response comparable to that produced by the smallest qualification standard flaw that is considered reliably detectable. This comparability includes not only response to the flaw itself but also the geometry in which the flaw is contained. In the ultrasonic method, for example, the response to a fatigue crack located in the center of a flat plate can be far different from the response to the same size crack that is located in the bore of a largediameter hole. Consequently, application of the ultrasonic method for inspection of the two geometries requires qualification standards as well as reference standards for both geometries. Second, reference standards for a specific inspection must be relatable to other reference standards used for the same inspection. That is, when several reference standards are available, the responses for each one must be known, and more important, the differences in responses for the standards must be known so that adjustments can be made in the inspection to ensure that a uniform process sensitivity can be maintained. Ideally, the responses of all reference standards for a given inspection should be identical; however, from a practical standpoint this is impossible to achieve. As stated previously, the response of the NDE methods is from a multitude of parameters associated with a given discontinuity. For example, if an electrically discharge machined slot is selected as a defect for a reference standard for an ultrasonic inspection, exact control over the size of the slot in each standard is not sufficient to guarantee identical responses. Slight variations in the orientation of the slot with respect to the surface of the standard and in the surface finish of the slot itself can cause noticeable differences in the ultrasonic response. This is the worst case; other NDE methods vary in their response to subtle geometric parameters, with the magnetic particle and penetrant methods probably being the most tolerant. However, even these methods are highly sensitive to the width of the flaw used in a reference standard. Another important property of reference standards is durability. Both the material and the type of flaw in a reference standard must be selected so that the standard will not readily deteriorate or change in the environment in which it will be used. These selections are affected by the NDE method for which the standard is intended. Both ultrasonic and penetrant methods are very sensitive to the presence of foreign material inside the flaw. For ultrasonics, this can affect the amount of energy that is reflected from the flaw. For penetrant inspection, the quantity of penetrant that can enter the defect, and consequently the brightness of the indication, will be reduced. The foreign material may be fluids, soils, or corrosion

products. For these reasons, magnesium, ferritic steels, and aluminum are particularly poor choices for reference standards for penetrant inspection and require some type of protection if used as ultrasonic standards. Eddy current methods are not affected by foreign material in the flaw, but are very sensitive to such surface conditions as scratches, pitting, and corrosion. Magnetic particle methods are sensitive to the width of the defect and to anything, such as cold working, that may change the magnetic permeability of the standard. Radiographic methods are sensitive to the thickness of the standard and to changes in the shape of the flaw. In general, a good choice of material for any method is one that is reasonably hard and forms an adherent, tough, and stable surface oxide layer (such as a titanium or nickel-base alloy), thus providing protection against mechanical damage and the gradual buildup of corrosion products. Personnel. Unless the inspection process is fully automated, the proficiency of the inspection personnel is the largest

variable affecting inspection reliability. This proficiency varies widely not only from inspector to inspector but also with the same inspector, depending on his working environment and his mental condition. For fully automated inspections, the proficiency of the inspector in operating the equipment is important but has little or no impact on inspection reliability. The first task in addressing the contribution of the inspector to inspection reliability is to ensure that he is knowledgeable of the specific techniques to be used and has the basic proficiency to perform the inspection to the required reliability. Experience has demonstrated that the previous qualifications of the inspector--for example, certification to MIL-STD410D (Ref 1)--are not sufficient to ensure the desired performance with a new inspection that must be performed with high reliability. Therefore, the most straightforward way to assess proficiency is to require inspector participation in the demonstration of inspection reliability. All inspectors that will be required to perform the inspection should also participate. This not only establishes the reliability of the proposed inspection but also identifies those personnel requiring additional training or experience before they can be expected to perform adequately. Careful observation of the inspector during the demonstration and of the results obtained is necessary to identify the additional training or experience needed. After training and/or additional experience is acquired, the demonstration effort can be repeated to indicate if the inspector has become sufficiently proficient in the inspection technique. After basic inspection proficiency has been demonstrated, it becomes a supervisory task to ensure that this proficiency is maintained. Control of the work environment of the inspector is important. Distractions such as noise, extremes in temperature, and other irritants should be eliminated to the extent possible. Break periods should be frequent enough to reduce fatigue. Personnel who are ill or otherwise physically impaired should be temporarily assigned other tasks to the extent possible. Other efforts that improve or maintain a good mental attitude are excellent ways to ensure sustained inspection reliability. These include providing acceptable materials and equipment with which to conduct the inspection. Finally, when it is not possible to provide a consistently conducive environment for a highly reliable inspection, two inspectors can perform the same inspection independently to achieve higher reliability than can be obtained with a single inspector. Two inspectors generally will not make identical mistakes.

Reference cited in this section

1. "Nondestructive Testing Personnel Qualification and Certification," MIL-STD-410D, 25 June 1974 Applications of NDE Reliability to Systems Ward D. Rummel, Martin Marietta Astronautics Group; Grover L. Hardy and Thomas D. Cooper, Wright Research & Development Center, Wright-Patterson Air Force Base

Applications (Case Studies) Airframes. In the aerospace industry, the first application of quantitative NDE in a production facility was on airframe

structures. Loss of an F-111 aircraft by the propagation of an undetected manufacturing defect led to the incorporation of fracture control in the design and qualification of aircraft structure by the United States Air Force. This was first applied rigorously to the B-1 aircraft and consisted of: •

The identification of critical structural components whose failure would cause the loss of aircraft

• •

The identification of those areas in the critical structural components experiencing the highest stresses defined The estimation of the maximum size of a rogue defect in these areas that could exist without growing to failure in twice the estimated design life

The maximum acceptable size of a rogue defect was established as 6.4 mm (0.250 in.) long by 3.2 mm (0.125 in.) deep for a surface flaw, or 1.3 × 1.3 mm (0.050 × 0.050 in.) for a corner crack in a bolt hole or a structural edge. For an embedded flaw, the size was a 6.4 mm (0.250 in.) diam circle. These sizes were selected because experience had shown that they were large enough to be readily detected during production inspection. If the crack sizes that the designer selected for his structure were to be smaller, the capability of the production NDE facility to find the smaller cracks reliably would have to be demonstrated. For the B-1, the crack sizes for many of the critical structural members was smaller than those suggested. Redesign of the components to tolerate the larger, suggested flaws would have resulted in unacceptable weight penalties; consequently, the B-1 production NDE facility was subjected to the first quantitative NDE capability evaluation. A conservative approach was taken in the design of the NDE reliability demonstration program. Fatigue cracks were selected to represent manufacturing defects for surface and corner flaws, and voids created by diffusion bonding were to serve as embedded defects. Flat panels were the geometries selected, either single panels for surface defect areas or panel stackups for fastener hole areas, and rectangular blocks housed the embedded defects. The materials were aluminum, titanium, and steel. The surface finish for the specimens was the same as for production parts. The methods to be demonstrated were penetrant inspection for all materials with surface flaws, magnetic particle for steel panels with surface flaws, eddy current for all materials with bolt holes, and ultrasonic for titanium and steel embedded flaws. The surface and corner flaws were divided into four groups based on size (Table 1). Because of the difficulty of manufacturing embedded flaws, only two sizes were selected: 1.3 mm (0.050 in.) diam circle and a 1.3 × 2.5 mm (0.050 × 0.100 in.) ellipse. Table 1 Groups of manufactured flaws used to demonstrate the reliability of NDE methods in the B-1 program Group

Length range

Depth range

mm

in.

mm

in.

Surface flaws

1

0.75-1.91

0.03-0.075

0.25-0.90

0.01-0.035

2

1.92-2.55

0.076-0.100

0.50-1.25

0.020-0.050

3

2.56-3.81

0.101-0.150

0.50-1.80

0.020-0.070

4

3.83-6.35

0.151-0.250

0.50-3.0

0.020-0.12

Group

Bore length range

Radial depth range

mm

mm

Corner flaws

in.

in.

1

0.75-1.53

0.03-0.060

0.25-0.76

0.01-0.030

2

1.54-2.55

0.061-0.100

0.78-1.27

0.031-0.050

3

2.56-3.56

0.101-0.140

1.29-1.78

0.051-0.070

4

3.58-5.0

0.141-0.20

1.80-2.5

0.071-0.10

For every flawed specimen presented to an inspector, at least two unflawed specimens would also be included. Sizing of the defects was not required, and no penalties were assessed if the inspector identified a flaw that did not exist. In addition, if the desired reliability was achieved for a given size range, it also had to be met at the next higher range. The reliability criterion was established to be a 90% probability of detection at the 95% confidence level (CL). This would be considered to be met for the largest flaw size in the given ranges if the following detection requirements were met:

Number of observations

Misses

29

0

46

1

61

2

75

3

89

4

103

5

A variety of lessons were learned during the demonstration program. The production inspection processes used at the time were unable to meet the design requirements of 1.3 mm (0.050 in.) depth for corner cracks. The processes were refined and the documentation was completely rewritten before the requirements were met. Second, not all inspection personnel were capable of achieving the required reliability, even with the refined procedures and documentation. As a result, the demonstration program also became a qualification requirement for personnel that would inspect fracture-critical components. Etching was also required before the penetrant inspection could meet the requirements for any of the materials. Embedded flaws could only be produced for titanium. Table 2 summarizes the results demonstrated for the B-1 production NDE facilities at the reliability criterion of a 90% POD at 95% CL. Table 2 Results of production NDE facilities in the B-1 program Method

Material

Flaw type

Flaw depth

Penetrant

mm

in.

Aluminum

Surface

0.90

0.035

Titanium

Surface

0.90

0.035

Steel

Surface

1.0

0.040

Magnetic particle

Steel

Surface

1.25

0.050

Ultrasonic, shear

Titanium

Embedded

0.9

0.035

Steel

Embedded

1.25

0.050

Ultrasonic, longitudinal

Titanium

Embedded

1.17

0.046

Eddy current

All

Corner

0.75

0.030

Fracture control, with the resultant quantitative NDE requirement, has been incorporated into the design of every Air Force aircraft since the B-1. In addition, designs already in existence at the time have been analyzed to determine which inspections would be required in service to ensure attainment of the design life of a particular system. Generally, where the suggested flaw sizes mentioned above could not be tolerated, the assumed design flaw for new designs has been a 0.75 mm (0.030 in.) corner crack and a surface flaw with a depth of 1.25 mm (0.050 in.) and a length of 2.5 mm (0.100 in.). The analysis of existing designs, however, has required the assumption of even smaller design flaws. These instances of having to detect smaller flaws in service have been few, and fortunately, identification of the requirement has occurred in sufficient time to allow adequate NDE engineering to address the problems and explore various noninspection options. Gas Turbine Engines. Following the successful application of damage tolerance concepts and designs to Air Force

aircraft structures, attention began to focus on aircraft gas turbine engines. The reason is that the failure of a high-energy (rotating) component in an aircraft gas turbine engine usually results in catastrophic consequences for that engine. Although most engines are designed to contain failure of the blades, the fracture of a disk or spacer will result in destruction of the engine and can cause significant damage to adjacent structures, such as fuel tanks, major structural members, or other engines. Even the contained failure of a blade can cause immediate engine shutdown, which can also have catastrophic consequences for a high-performance, single-engine fighter. Therefore, the integrity of many highperformance components in gas turbine engines is critical to aircraft safety. As part of an effort to increase the reliability and reduce the costs of operating and maintaining gas turbine engines for the U.S. Air Force, a program known as the Engine Structural Integrity Program (ENSIP) has evolved over the last few years. Its objective is to establish an approach to defining the structural performance, design, development, verification, and life management requirements for new engines for Air Force aircraft. Military standard MIL-STD-1783 defines ENSIP and is currently being applied to the development of all new engines for the Air Force (Ref 2). The document is written in a generic format so that it can be tailored for use by specific System Program Offices to define an engine that will satisfy their own needs. An accompanying handbook is attached as an appendix to provide specific guidance on the rationale, background criteria, lessons learned, and instructions necessary to tailor specific sections of the standard for application. The technical approach is similar to the Aircraft Structural Integrity Program (ASIP), which is defined in MIL-STD-1530 and which has been successfully used for several years in the design of airframes for Air Force systems (Ref 3). One of the most significant differences between ENSIP and the traditional approaches used in designing engine structures is the requirement to apply damage tolerance and durability criteria to critical components. This requires the designer to assume that flaws exist in the engine structure as manufactured and then to design the critical parts so that the flaws

cannot grow to the size that will cause failure in the lifetime of the part or at least within some predetermined inspection interval. It also establishes life management requirements and procedures to ensure that the necessary inspections capable of finding flaws in the size range used in design are conducted and that the engine parts are sufficiently durable so that the economic life of the engine is acceptable. The impact of this approach on the inspection community is very significant. Based on the estimated capability of stateof-the-art inspection methods and procedures, many improvements have had to be made both in manufacturing and depot practice to satisfy the intent of ENSIP. Furthermore, to allow the full implementation of this approach, the problem of defining available inspection capability in quantitative terms must continue to receive attention. Acceptable procedures are being established so that it is possible to define exactly how sensitive the inspection methods are on a statistical basis. The first application of this technology by the Air Force was made to an already designed and operational aircraft gas turbine engine. The F-100 engine, designed and built by Pratt & Whitney, was already widely used in the dual-engine F15 and the single-engine F-16 aircraft when, in 1978, a durability and damage tolerance assessment effort was initiated. This engine was selected because it was to be purchased in large numbers for Air Force applications for many years to come and because it was a high-performance machine that was very demanding of materials and designs. The costs of owning and operating this system could quickly become untenable if problems developed that significantly limited the life of critical parts or caused significant down-time for repairs. Following the pattern that had been established for applying ASIP to airframes, a joint Air Force/Pratt & Whitney team was formed to work on-site at the facilities of the contractor to complete the analysis and to provide a viable Force Structural Maintenance Plan that could be implemented at the San Antonio Air Logistics Center, where maintenance responsibility for the engines resided. After the team was in place and the necessary analytical studies were started, it was recognized early in the program that a quantitative understanding of the inspection procedures used in the manufacture of the engine was lacking and that a reliable definition of the largest flaw that could escape detection during manufacture had never been determined. Qualitative statements were made expressing confidence in the inspection methods used, based on the good performance of the engines to that time. It was acknowledged, however, that the fleet was still young and that a quantitative definition of the inspection process was urgently needed. To fill that need, a joint Air Force Materials Laboratory/Pratt & Whitney effort was established to prepare specimens with known flaws in selected size ranges that would allow quantitative determination of the capability of the inspection methods to be established. The specimens were designed to contain flawed areas in geometrical features that simulated the real areas in the actual hardware. These included holes, the edges of holes, radii, and flat surfaces. Small flaws were generated by damaging the surface, initiating and growing a fatigue crack, and then removing the damaged area until only the desired depth of the flaw that remained was used to produce the specimens. During 1979, some 39 sets of specimens with the desired geometries were fabricated from nickel- and titanium-base alloys for the program. Target crack depths of 0.13, 0.25, and 0.50 mm (0.005, 0.010, and 0.020 in.) with a nominal 3:1 aspect ratio were prepared. Similar specimens with no flaws were also included in each set. The specimens were to be used not only to determine the capability of the manufacturing inspection methods but also to provide guidance concerning the establishment of inspection methods to be used in the depot during maintenance. Fluorescent penetrant inspection and eddy current procedures were evaluated using these specimens. The specimen sets were evaluated by both laboratory and production inspectors. Because this was the first documented attempt to fabricate flawed specimens of this complexity, a great deal of experience was gained, not only in specimen preparation but also in evaluating the effectiveness of the inspection methods. Fluorescent penetrant inspection was included because, at the time, it was the most extensively applied inspection method used in both manufacturing and depot maintenance. Eddy current methods were included because they had the best potential for finding the very small surface-connected flaws of concern. Based on the work done in this program, the following conclusions were reached: •

• •

The capability of fluorescent penetrant inspection to find small flaws with confidence was affected by many variables, including surface condition, nature of the flaw, the process used and the extent to which it was controlled, and the skills and abilities of the inspectors. The process did not have the necessary reliability to detect very small flaws in many of the critical areas that had been defined Eddy current methods appeared to have the best potential for detecting small flaws with the confidence level required Eddy current technology and procedures developed in the program could also be adequately automated to make this method viable for use in the depot inspection environment. They could also be adapted for

manufacturing

As a result of this study, the decision was made to implement semiautomated eddy current inspection methods at the San Antonio Air Logistics Center to provide reliable inspection of defined areas in selected critical parts. A facility was established at the depot to allow the inspection of critical components as they were cycled through the depot. The eddy current equipment designed by Pratt & Whitney and incorporated into the facility has been demonstrated to have the capability of finding 0.13 mm (0.005 in.) deep flaws with a reliability of 90% POD at 95% CL. The second application of damage tolerance analysis involved the TF-34 engine, which powers the A-10 ground-support attack aircraft and the S-3A antisubmarine aircraft. In this case, an Air Force/General Electric team was formed at the General Electric Aircraft Engine Business Group facility in May 1981. Once again, a critical part of the assessment activity was to establish the level of quality built into the parts during the years the engines were manufactured. This was needed to provide an indication of the largest flaw that could have been missed by the inspection methods being used at the time. Because production of the engine was essentially completed by the time the assessment started, the only impact that establishing improved NDE procedure could have was on the methods being used at the maintenance and overhaul depot. In the case of the TF-34 engine, depot responsibility had been assigned to the Navy and was conducted at the Naval Air Newark Facility in Alameda, CA. Once again, the assessment determined that to achieve acceptable inspection intervals for certain critical rotating components, inspection methods more sensitive and reliable than fluorescent penetrant inspection would be required. A special clean room containing eddy current equipment was established at Alameda to allow inspection of the engine disks and spacers that were considered to be the most critical. The equipment selected for this room was the Eddy Current II system developed under an Air Force Materials Laboratory manufacturing technology contract by the General Electric Company. In addition, the fluorescent penetrant inspection facility at Alameda was upgraded to make it more reproducible and reliable. Inspections have been conducted with these systems since 1984. Other applications of this technology by the U.S. Air Force have been made to all new gas turbine engines now being used. This includes the F-101 engine in the B-1B and the F-110 engine in the F-16. New engines now being designed for Air Force applications will also incorporate ENSIP technology and will therefore have quantitative inspection capabilities as an integral part of their development. Additional information on ENSIP is available in the article "Fracture Control Philosophy" in this Volume. Space Shuttle Program. Design requirements for the space shuttle program of the National Aeronautics and Space

Administration included the use of fatigue and fracture mechanics principles in all systems designs. The implementation of fatigue and fracture mechanics requires knowledge of the quantitative performance capabilities of the materials, components, and systems acceptance methods. For this application, Rummel et al. introduced the concepts of statistical assessments of NDE process performance capabilities (Ref 4). Fatigue cracks in 2219-T87 aluminum alloy were selected as the test specimens for the assessment of various NDE methods. Test specimens were prepared by inducing fatigue cracks of varying size in aluminum alloy sheet specimens (fatigue crack growth from electrodischarge machined starter notches), by machining the specimen surfaces to produce a surface that was representative of production conditions, and by passing the specimens through various NDE procedures and measuring the success in crack detection for the various procedures and operating conditions. The concept of probability of detection as a function of flaw size was introduced, and the performance level of the various NDE procedures was quantified as summarized below. The experimental test sequence for data gathering, including nondestructive testing (NDT), is shown in Fig. 5. The test specimens were given a precrack (starter notches) that enabled the growth of fatigue cracks of varying size and aspect ratio. A total of 328 cracks were grown in 118 specimens, with flaw length ranging from 0.3 to 18 mm (0.012 to 0.700 in.). The specimens were subjected to x-ray radiographic, liquid penetrant, ultrasonic, and eddy current procedures in the as-machined, after-etch, and after-proof-test conditions. The resulting POD curves for machined and etched surfaces are also shown in Fig. 6.

Fig. 5 Experimental test analysis sequence

Fig. 6 POD plots for four different NDE methods on the same set of specimens. (a) Penetrant inspection. (b) Ultrasonic inspection. (c) Eddy current inspection. (d) X-ray inspection

The precision in crack sizing was also measured, and the results are shown in Fig. 7. The composite threshold detection results were reduced to a plot, as shown in Fig. 8. These data were further simplified to produce the design limits, as shown in Fig. 9. Figure 9 contains design limits for standard NDE and special NDE. The use of special NDE to meet difficult design constraints requires actual demonstration of the performance capabilities of the proposed NDE procedures and a system of controls to ensure that the performance conditions are maintained. These design limits, set by actual NDE performance demonstration, were used as the basis for design and risk management for all space shuttle system components. Assessments and NDE performance capability demonstrations have been continued with space shuttle contractors for all systems production and revalidation.

Fig. 7 Actual versus NDE estimated crack length for etched specimens. (a) Penetrant inspection. (b) X-ray inspection

Fig. 8 Combined threshold detection results

Fig. 9 Plot of design limits

Special Inspection Systems. As a result of the evolution of damage tolerance requirements, special equipment has

been developed in the last few years for increasing the reliability of inspection operations by automating the process and by incorporating computerized data generation and control. This has significantly reduced the dependence on human operators, thus eliminating one of the major sources of error in the process. Most of these automation efforts have been directed toward the inspection of critical aircraft gas turbine engine hardware, primarily in the maintenance depot environment. The systems could be used equally effectively in manufacturing operations. The following are examples of the systems that have been developed over recent years. Structural Assessment Testing Applications. The NDE performance requirements for overhaul and revalidation of the structural integrity of aircraft engine components (United States Air Force facilities) required the use of special, controlled methods to meet meantime between overhaul and life cycle performance requirements. Special equipment, fixturing, procedures, and personnel training are implemented at the San Antonio Air Logistics Center at Kelly Air Force Base to approach the imposed design constraints. Assessment, demonstration, and validation of the system performance capabilities and reliabilities were necessary to ensure that design requirements were being met.

The special NDE methods were successful in meeting design requirements, and new levels of understanding and performance were gained by the implementation of advanced NDE methods and controls in this special facility. The most significant output from the assessment of these processes was the demonstration of the need for excellence in NDE engineering in validating NDE procedures and in setting acceptance limits. The performance capability of an eddy current method with an acceptance threshold limit set at 3.0 mV is shown in Fig. 10(a). The performance level of the same procedure (same data) with an acceptance limit set at 0.5 mV is shown in Fig. 10(b). This example clearly illustrates the importance of fully characterizing the NDE procedure and managing the NDE procedure within achievable acceptance limits.

Fig. 10 POD plotted (a) at a threshold of 3.0 mV and (b) at a threshold of 0.5 mV

Integrated Blade Inspection System. The quantitative assessment of process performance capabilities and process

characterization is absolutely necessary in implementing automated NDE systems. At a gas turbine overhaul facility, for example, quantitative assessment methods were applied to the implementation of an integrated blade inspection system with an automated fluorescent penetrant inspection module (Fig. 11). The processed blades are introduced into a robotic

handling system that manipulates the blade in a high-gain optical-laser scan readout system to produce a digitized image of the fluorescent penetrant indications. A computerized data processing and image analysis system provides the readout and decision processing to accept or reject the blades.

Fig. 11 Flow diagram of an automated fluorescent penetrant inspection system

The optical performance capabilities of the readout system enable the detection of very small indications of varying brightness level. The performance of the system is therefore limited by the fluorescent process capabilities and by the pattern recognition capabilities of the system. Signal/noise distributions must be addressed by the decision discrimination process, as shown in Fig. 12. Actual signal/noise distributions measured for the system are shown in Fig. 13. If the discrimination threshold is set too low, a high false call rate will reduce the effectiveness of the automated process. A balance of discrimination at an acceptable false call rate is used to achieve the figure of merit of performance level (Fig. 14) for the system. Signal/noise distribution data can be used to manage the overall process at any desired discrimination/false call level. The process can be modeled to establish performance at varying discrimination levels, as shown in Fig. 15. Quantification and characterization of process parameters are essential for the design, implementation, and management of automated NDE processes.

Fig. 12 Probability distribution of a noise signal and flaw indication

Fig. 13 A sample of the signal/noise distribution for an integrated blade inspection system

Fig. 14 POD curve for an automated fluorescent penetrant inspection system

Fig. 15 Performance curves for various discrimination levels

Retirement-For-Cause (RFC) Inspection Equipment. As described previously, the concept of removing critical

rotating components from gas turbine engines because of the initiation of an actual flaw rather than on the basis of an arbitrarily selected time period has already saved the Air Force significant sums of money. One of the key technologies

that has allowed the RFC concept to be accepted has been the development of an accurate, repeatable, reliable inspection system capable of finding very small flaws in the parts. The RFC system, developed under an Air Force Manufacturing Technology contract, has been in operation at the San Antonio Air Logistics Center at Kelly Air Force Base since late 1986. It was developed under a contract with Systems Research Laboratories, and although applied specifically to parts for the F-100 engine (used in the F-15 and F-16 aircraft), it was designed to be sufficiently flexible to inspect parts from any gas turbine engine. This capability was achieved through the use of a team of subcontractors on the program that included Pratt & Whitney, General Electric, Garrett, and Allison. Eddy current methods for surface flaw detection and ultrasonic methods for detecting embedded flaws were the two inspection techniques selected for the RFC system. Significant advances had been made in automating eddy current inspection through the development of the Eddy Current II system by the General Electric Company on earlier Air Force Manufacturing Technology contracts. These stand-alone automated eddy current devices are in use at the Alameda Naval Air Rework Facility, as previously discussed; the Oklahoma City Air Logistics Center; the General Electric Manufacturing facility in Evandale, OH; and other commercial overhaul facilities. The technology developed and proved in these systems provided an excellent base for the evolution of the more complex RFC system. The RFC inspection system, which is housed in the special facility shown in Fig. 16, consists of an operator console, a system computer, and eddy current and ultrasonic inspection stations. The operator console is used to monitor the operational status of the system, to track inspection status at each NDE station, and to generate inspection data reports. The system computer performs advanced data processing, system-wide communication, and sophisticated high-speed mathematical and scientific data analyses critical to the inspection process. The NDE inspection stations perform the automated part inspections, flaw detection, and signal-preprocessing activities. The system functions essentially independent of any human operator input, with the exception of loading and unloading the parts to be inspected onto the stations.

Fig. 16 Retirement-for-cause inspection facility

When the system was installed, a series of critical tests was conducted to establish its flaw size detection capability and reliability. Tests included automatic scans of engine disks and a statistically significant number of representative fatiguecracked test specimens. Rivet hole inspection data showed a 90% POD with a 95% CL at the 100 m (4 mil) crack depth range. Bolt hole and flat surface data indicated reliable detection in the desired 125 to 250 m (5 to 10 mil) depth range. A strong correlation between apparent versus actual flaw depth data was seen in all test data. The ultrasonic inspection data were similarly encouraging. Examples of POD at 95% CL generated for various geometrical configurations are shown in Fig. 17. A similar facility is to be installed at the Air Force engine maintenance depot at the Oklahoma City Air Logistics Center at Tinker Air Force Base.

Fig. 17 RFC system POD curves for various geometrical features

References cited in this section

2. "Engine Structural Integrity Program (ENSIP)," MIL-STD-1783 (USAF), 30 Nov 1984 3. "Aircraft Structural Integrity Program, Airplane Requirements," MIL-STD-1530A (USAF), 11 Dec 1975 4. W.D. Rummel, P.H. Todd, Jr., S.A. Frecska, and R.A. Rathke, "The Detection of Fatigue Cracks by Nondestructive Testing Methods," CR-2369, National Aeronautics and Space Administration, Feb 1974 Applications of NDE Reliability to Systems Ward D. Rummel, Martin Marietta Astronautics Group; Grover L. Hardy and Thomas D. Cooper, Wright Research & Development Center, Wright-Patterson Air Force Base

References 1. 2. 3. 4.

"Nondestructive Testing Personnel Qualification and Certification," MIL-STD-410D, 25 June 1974 "Engine Structural Integrity Program (ENSIP)," MIL-STD-1783 (USAF), 30 Nov 1984 "Aircraft Structural Integrity Program, Airplane Requirements," MIL-STD-1530A (USAF), 11 Dec 1975 W.D. Rummel, P.H. Todd, Jr., S.A. Frecska, and R.A. Rathke, "The Detection of Fatigue Cracks by Nondestructive Testing Methods," CR-2369, National Aeronautics and Space Administration, Feb 1974

NDE Reliability Data Analysis Alan P. Berens, University of Dayton Research Institute

Introduction INSPECTION SYSTEMS are inevitably driven to their extreme capability for finding small flaws. When applied at this extreme, not all flaws of the same size will be detected. In fact, repeat inspections of the same flaw will not necessarily produce consistent hit or miss indications, and different flaws of the same size may have different detection probabilities. Because of this uncertainty in the inspection process, capability is characterized in terms of the probability of detection (POD) as a function of flaw size, a. At present, the function POD(a) can be estimated only through inspection reliability experiments on specimens containing flaws of known size. Further, statistical methods must be used to estimate the parameters of the POD(a) function and to quantify the experimental error in the estimated capability. The methods for analyzing NDE reliability data have undergone a considerable evolution since the middle of the 1970s. Formerly, a constant probability of detection of all flaws of a given size was postulated, and binomial distribution methods were used to estimate this probability and its lower confidence bound (Ref 1, 2). This nonparametric method of analysis produced valid statistical estimates for a single flaw size, but required very large sample sizes to obtain reasonable lower confidence bounds on the probability of detection. In the absence of large numbers of representative specimens with equal flaw sizes, various methods were devised for analyzing data based on grouping schemes. Although the resulting POD appeared more acceptable using these schemes, the lower confidence bounds were no longer valid. In recent years, an approach based on the assumption of a model for the POD(a) function was devised (Ref 3, 4, 5, 6, 7). Analyses of data from reliability experiments on nondestructive inspection (NDI) methods indicated that the POD(a) function can be reasonably modeled by the cumulative log normal distribution function or, equivalently, the log-logistics (log odds) function. The parameters of these functions can be estimated using maximum likelihood methods. The statistical uncertainty in the estimate of NDI reliability has traditionally been reflected by a lower (conservative) confidence bound on the POD(a) function. The asymptotic statistical properties of the maximum likelihood estimates can be used to calculate this confidence bound. Details of the mathematics for these maximum likelihood calculations are presented in this article.

References

1. B.G.W. Yee, F.H. Chang, J.C. Couchman, G.H. Lemon, and P.F. Packman, "Assessment of NDE Reliability Data," NASA CR-134991, National Aeronautics and Space Administration, Oct 1976 2. W.D. Rummel, Recommended Practice for Demonstration of Nondestructive Evaluation (NDE) Reliability on Aircraft Production Parts, Mater. Eng., Vol 40, Aug 1982, p 922-932 3. W.H. Lewis, W.H. Sproat, B.D. Dodd, and J.M. Hamilton, "Reliability of Nondestructive Inspections--Final Report," SA-ALC/MME 76-6-38-1, San Antonio Air Logistics Center, Kelly Air Force Base, Dec 1978 4. A.P. Berens and P.W. Hovey, "Evaluation of NDE Reliability Characterization," AFWAL-TR-81-4160, Vol 1, Air Force Wright-Aeronautical Laboratories, Wright-Patterson Air Force Base, Dec 1981 5. A.P. Berens and P.W. Hovey, Statistical Methods for Estimating Crack Detection Probabilities, in Probabilistic Fracture Mechanics and Fatigue Methods: Applications for Structural Design and Maintenance, STP 798, J.M. Bloom and J.C. Ekvall, Ed., American Society for Testing and Materials, 1983, p 79-94 6. D.E. Allison et al., "Cost/Risk Analysis for Disk Retirement--Volume I," AFWAL-TR-83-4089, Air Force Wright-Aeronautical Laboratories, Wright-Patterson Air Force Base, Feb 1984 7. A.P. Berens and P.W. Hovey, "Flaw Detection Reliability Criteria, Volume I--Methods and Results," AFWAL-TR-84-4022, Air Force Wright-Aeronautical Laboratories, Wright-Patterson Air Force Base, April 1984

NDE Reliability Data Analysis Alan P. Berens, University of Dayton Research Institute

Statistical Nature of the NDE Process In the application of an NDE method there are many factors that can influence whether or not the inspection will result in the correct decision as to the absence or presence of a flaw. In general, NDE comprises the application of a stimulus to a structure and the interpretation of the response to the stimulus. Repeated inspections of a specific flaw can produce different magnitudes of stimulus response because of minute variations in setup and calibration. This variability is inherent in the process. Different flaws of the same size can produce different response magnitudes because of differences in material properties, flaw geometry, and flaw orientation. The interpretation of the response can be influenced by the capability of the interpreter (manual or automated), the mental acuity of the inspector as influenced by fatigue or emotional outlook, and the ease of access and the environment at the inspection site. All these factors contribute to inspection uncertainty and lead to a probabilistic characterization of inspection capability. There are two related approaches to a probabilistic framework for analyzing inspection reliability data. Originally, inspection results were recorded only in terms of whether or not a flaw was found. Data of this nature are called hit/miss data, and an analysis method for this data type evolved from the original binomial characterization (Ref 3, 4, 5). It was later observed that there is more information in the NDE signal response from which the hit/miss decision is made (Ref 6). Because the NDE signal response can be considered to be the perceived flaw size, data of this nature are called data. A second analysis method was developed based on data (Ref 7). Although the analysis frameworks are based on data of different natures, the hit/miss data can be obtained from data. Both methods are based on the same model for the POD(a) function, but different results will be obtained if the two analysis methods are applied to the same data set. The following sections present the two approaches to formulating the POD(a) function. POD(a) From Hit/Miss Data. In typical NDE reliability studies, relatively few inspections are performed on each

flaw in the specimen set. Table 1 presents an example of hit/miss results from fluorescent penetrant inspections by 3 inspectors on 35 cracks in flat plate specimens. Because there are only three inspections, it is impossible to obtain more than a general impression of any change in the chances of crack detection as the cracks get larger. In a study for the Air Force (Ref 3), inspections of cracked specimens were independently conducted by large numbers of inspectors around the country. The data from this program provided considerable insight into the nature of the probability of crack detection. Table 1 Example of summary data sheet of hit/miss results The example is based on the fluorescent penetrant inspection of flat plates by three inspectors. Crack size

Inspector(a)

mm

in.

A

B

C

1

2.21

0.087

0

1

1

2

1.63

0.064

1

0

0

3

0.38

0.015

0

0

0

4

2.84

0.112

1

1

0

5

0.99

0.039

0

0

0

Crack identification

6

4.42

0.174

1

1

1

7

2.97

0.117

1

1

1

8

2.06

0.081

0

0

0

9

0.46

0.018

0

0

0

10

4.22

0.166

1

0

1

11

2.54

0.100

1

0

0

12

1.98

0.078

1

1

0

13

0.64

0.025

0

1

0

14

2.18

0.086

1

0

0

15

2.64

0.104

0

0

1

16

2.49

0.098

0

1

1

17

2.41

0.095

1

0

0

18

1.42

0.056

0

0

0

19

0.20

0.008

0

0

0

20

0.58

0.023

0

0

0

21

6.99

0.275

1

1

1

22

3.30

0.130

0

0

1

23

6.20

0.244

1

1

1

24

2.03

0.080

0

0

0

25

1.85

0.073

0

0

0

26

4.95

0.195

1

1

1

27

0.23

0.009

0

0

0

28

2.13

0.084

0

1

1

29

5.59

0.220

1

1

1

30

1.02

0.040

0

0

0

31

0.99

0.039

0

0

0

32

2.18

0.086

0

0

1

33

0.25

0.010

0

0

1

34

4.09

0.161

0

0

1

35

0.51

0.020

0

0

0

(a) 1 indicates crack was found; 0 indicates crack was not found.

Figure 1 shows the results of eddy current inspections by 60 different inspections of a set of 41 cracks around countersunk fastener holes in a 1.5 in (5 ft) segment of a C-130 center wing box (Ref 3). Each data point represents the proportion of times that the crack was found. This data set, which is representative of such available data sets, clearly indicates that: • • •

The chances of detection are correlated with crack size Different cracks of the same size can have significantly different crack detection probabilities Factors other than size are affecting the chances of detection

Data of this nature also provide the analysis framework for characterizing NDE reliability for the general hit/miss data set.

Fig. 1 Example crack detection probabilities from 60 eddy current inspections of each crack

The POD(a) function is defined as the proportion of all cracks of size a that will be detected in a particular application of an NDE system. Assume that each crack of size a in the potential population of cracks has its own distinct crack detection probability, p, and that the probability density function of the detection probabilities is given by fa(p). Figure 2 shows a schematic representation of this density. The conditional probability of a randomly selected crack from the population having detection probability of p and being detected at the inspection is given by p fa(p) dp. The unconditional probability of a randomly selected crack from the population being detected is the sum of the conditional probabilities over the range of p, that is:

(Eq 1)

Therefore, POD(a) is the average of the detection probabilities for cracks of size a.

Fig. 2 Schematic of distribution of detection probabilities for cracks of fixed length

Equation 1 implies that the POD(a) function is the curve through the averages of the individual density functions of the detection probabilities. This curve is the regression equation and provides the basis for testing assumptions about the applicability of various POD(a) models. In Ref 4, seven different functional forms were tested for applicability to available POD data, and it was concluded that the log-logistics (log odds) function best modeled the data and provided an acceptable model for the data sets of the study. Note that the log odds model is commonly used in the analysis of binary (hit/miss) data because of its analytical tractability and its close agreement with the cumulative log normal distribution (Ref 8). Two mathematically equivalent forms of the log odds model have subsequently been used. The earliest form is given by:

(Eq 2) This parametrization can also be expressed as:

(Eq 3)

In the Eq 3 form, the log of the odds of the probability of detection (the left-hand side of Eq 3) is expressed as a linear function of ln (a) and is the source of the name of the log odds model. Note that given the results of a large number of independent inspections of a large number of cracks, the parameters of the model can be fit with a regression analysis. As an example, Fig. 3 shows Eq 3 fit to the data of Fig. 1. This regression approach will not be discussed further, because the maximum likelihood estimates (see the section "Analysis of Hit/Miss Data" in this article) can be applied to much smaller samples of inspection results and can give equivalent answers for large sample sizes.

Fig. 3 Example linear relation between log odds of crack detection and log crack size

Although the parametrizations of Eq 2 and 3 are sensible in terms of estimation through regression analyses, and are not easily interpretable in physical terms. A mathematically equivalent form of the log odds POD(a) model is given by (Ref 8):

(Eq 4)

In this form, μ= ln a0.5, where a0.5 is the flaw size that is detected 50% of the time, that is, the median detectable crack size. The steepness of the POD(a) function is inversely proportional to ; that is, the smaller the value of σ, the steeper the POD(a) function. The parameters of Eq 2 and 4 are related by:

(Eq 5) (Eq 6)

The log odds POD(a) function is practically equivalent to a cumulative log normal distribution with the same parameters, μ and σ of Eq 4. Figure 4 compares the log odds and cumulative log normal distribution functions for μ= 0 and σ= 1. Equation 4 is the form of the log odds model that will be used in the section "Analysis of Hit/Miss Data" in this article.

Fig. 4 Comparison of log odds and cumulative log normal models

POD(a) From Signal Response Data. The NDE flaw indications are based on interpreting the response to a

stimulus. In eddy current or ultrasonic systems, the response might be a peak voltage referenced to a calibration. In fluorescent penetrant inspections, the response would be a combination of brightness and size of the indication. Assume the response can be quantified and recorded in terms of a parameter, , that is correlated with flaw size. Then summarizes the information for determining if a positive flaw indication will be given. Only if exceeds a defined decision threshold, dec, will a positive indication be given. As an example of the concept, Table 2 summarizes the results of highly automated eddy current inspections of 28 cracks in flat plate specimens. The three data sets resulted from the use of three probes, with all other factors held constant. The values in Table 2 are the depth of each crack and the peak voltage in counts recorded by the system. Figure 5 shows a plot of the versus a data for probe A. No signal was recorded for 2 of the cracks, because their values were below the recording signal threshold, th. These points are indicated by a down arrow at th, indicating that the response was at an indeterminable value below the recording signal threshold. Similarly, for 5 of the cracks exceeded the saturation limit, sat, of the recording system. These points are indicated by an up arrow at sat, indicating that the response was at an indeterminable value above the recording saturation limit. In Fig. 5, the decision threshold is set at 250 counts. Only those cracks whose value is above 250 would have been flagged (detected). Table 2 Example of a summary data sheet of versus a data The example is based on eddy current inspections of flat plates.

Crack depth

Peak voltage in counts

mm

in.

Probe A

Probe B

Probe C

11

0.33

0.013

1052

884

1282

30

1.40

0.055

4095

4095

3831

42

0.38

0.015

1480

1182

1699

2

0.25

0.010

723

624

840

21

0.74

0.029

4095

4095

2249

13

0.48

0.019

2621

2401

1101

19

0.30

0.012

377

809

350

26

0.23

0.009

223

205

277

15

0.56

0.022

1654

3319

1289

29

1.65

0.065

4095

4095

2648

33

0.08

0.003

(a)

(a)

(a)

25

0.25

0.010

669

565

824

32

0.18

0.007

374

379

407

34

0.03

0.001

(a)

(a)

(a)

39

0.18

0.007

409

387

586

12

0.28

0.011

895

690

677

38

0.20

0.008

374

301

549

20

0.79

0.031

4095

4095

1778

28

0.23

0.009

638

454

782

Crack identification

27

0.15

0.006

533

385

631

1

0.08

0.003

150

136

135

35

0.28

0.011

749

660

989

40

0.20

0.008

433

378

591

31

0.36

0.014

879

888

1402

3

0.23

0.009

286

211

352

7

0.23

0.009

298

163

215

16

0.41

0.016

1171

1110

1628

37

2.54

0.100

4095

4095

4095

(a) Peak voltage below the recording level threshold

Fig. 5 Example inspection signal response as a function of crack depth

The POD(a) function can be obtained from the relation between values for fixed crack size a, then:

and a. If ga( ) represents the probability density of the

(Eq 7)

This calculation is illustrated in Fig. 6, in which the shaded area under the density functions represents the probability of detection.

Fig. 6 Schematic of POD(a) calculation from

versus a relation

In general, the correlating function between and a defines the mean of ga( ), that is:

= (a) +

(Eq 8)

where (a) is the mean of ga( ) and is a random error term accounting for the differences between distributional properties of δ determine the probability density ga( ) about μ (a), as will be shown.

and

(a). The

In the data analyzed to date, a linear relation between ln ( ) and ln (a) with normally distributed deviations has proved satisfactory (for example, Fig. 5). This model is expressed by:

ln ( ) =

0

+

1

ln (a) +

(Eq 9)

where δ is normally distributed with zero mean and constant standard deviation, . Data have been observed that flatten at the large crack sizes. However, because the decision threshold was far below the non-linear range, restricting the range of cracks to smaller sizes permitted the application of Eq 9. The normality of has proved to be an acceptable assumption. Assuming that the versus a relation is modeled by Eq 9 and that is normally distributed with zero mean and standard deviation of , the POD(a) function is calculated as:

(Eq 10)

where

is the standard normal distribution function. Using the symmetry properties of

, Eq 10 can be reduced to:

(Eq 11)

Equation 11 is a cumulative log normal distribution function with mean and standard deviation of log crack length given by:

(Eq 12) (Eq 13) In the section "Signal Response Analysis" in this article, maximum likelihood methods for estimating β0, β1, and σ from versus a data will be presented. Note that the values below the recording threshold and above the saturation limit must be properly accounted for in these analyses. Note also that data from multiple inspections of the same cracks require analysis methods that are dependent on the design of the reliability experiment. Methods for placing lower confidence bounds on the estimated POD(a) function using the sampling distributions of the maximum likelihood estimates of β0, β1, and are also included in the section "Signal Response Analysis."

References cited in this section

3. W.H. Lewis, W.H. Sproat, B.D. Dodd, and J.M. Hamilton, "Reliability of Nondestructive Inspections--Final Report," SA-ALC/MME 76-6-38-1, San Antonio Air Logistics Center, Kelly Air Force Base, Dec 1978 4. A.P. Berens and P.W. Hovey, "Evaluation of NDE Reliability Characterization," AFWAL-TR-81-4160, Vol 1, Air Force Wright-Aeronautical Laboratories, Wright-Patterson Air Force Base, Dec 1981 5. A.P. Berens and P.W. Hovey, Statistical Methods for Estimating Crack Detection Probabilities, in Probabilistic Fracture Mechanics and Fatigue Methods: Applications for Structural Design and Maintenance, STP 798, J.M. Bloom and J.C. Ekvall, Ed., American Society for Testing and Materials, 1983, p 79-94 6. D.E. Allison et al., "Cost/Risk Analysis for Disk Retirement--Volume I," AFWAL-TR-83-4089, Air Force Wright-Aeronautical Laboratories, Wright-Patterson Air Force Base, Feb 1984 7. A.P. Berens and P.W. Hovey, "Flaw Detection Reliability Criteria, Volume I--Methods and Results," AFWAL-TR-84-4022, Air Force Wright-Aeronautical Laboratories, Wright-Patterson Air Force Base, April 1984 8. D.R. Cox, The Analysis of Binary Data, Methuen and Co., 1970

NDE Reliability Data Analysis Alan P. Berens, University of Dayton Research Institute

Design of NDE Reliability Experiments An NDE reliability experiment comprises a test matrix of inspections on a set of specimens with known flaw locations and sizes. The specimens are inspected under conditions that simulate as closely as practical the actual application conditions. The experimental design determines the test matrix, and there are four major analysis concerns to be addressed in the experimental design. These are: • • • •

The method of controlling the factors to be evaluated in the experiment The method of accounting for the uncontrolled factors in the experiment The number of flawed and unflawed inspection sites The sizes of the flaws in the specimens

These topics are addressed in the following sections. Controlled and Uncontrolled Factors The primary objective of NDE reliability experiments has been to demonstrate efficacy for a particular application by estimating the POD(a) function and its lower 95% confidence bound. (Although NDE reliability experiments can also be conducted to optimize a system, analyses to meet this objective are beyond the scope of this article.) To demonstrate capability, it is assumed that the protocol for conducting the inspections is well defined for the application, that the inspection process is under control (hit/miss decisions are stable over time), and that all other factors introducing variability into the inspection decision will be representative of the application. The representativeness of these other factors can be ensured either by controlling the factors during the inspection or randomly sampling the factors to be used in the experiment. The methods of accounting for these factors are important aspects of the statistical design of the experiment and significantly influence the statistical properties of the estimates of the POD(a) function parameters. Of particular note in this regard is that k inspections on n flaws is not equivalent to inspections on n · k different flaws, even if the inspections are totally independent. The most important of the factors introducing variation are: • • • •

Differences in physical properties of cracks of nominally identical sizes The basic repeatability of the magnitude of the NDE signal response when a specific crack is independently inspected by a single inspector using the same equipment The summation of all the human factors associated with the particular inspectors in the population of interest Differences introduced by changes in inspection hardware

These factors must be addressed explicitly or implicitly in every NDE reliability experiment. In general, the specimens used in NDE reliability experiments are very expensive to obtain and characterize in terms of the sizes of the flaws in the specimens. Therefore, each experiment is based on one set of specimens containing flawed and unflawed inspection sites. Because the results are significantly influenced by the specimens, it must be assumed that the flaws are representative of those that will be present in the structural application. If other factors are to be included in the experiment, they will be based on repeated inspections of the same flaws. From a statistical viewpoint, this restriction on the experimental design limits the sample size to the number of flaws in the specimen set. Because different cracks of the same size can have significantly different crack detection probabilities, multiple inspections of the same crack provide information about the detection probability of only that crack.

The generality of the capability characterization is limited to the application for which the experiment is representative. Either important factors must be typical of the application or random samples must be chosen from the population of interest and repeat inspections performed for these factors. For example, if a single inspector is used to characterize a fluorescent penetrant inspection, it must be assumed that this inspector is typical of all the inspectors in the shop. An alternative might be to choose a random sample of inspectors from the total pool and have each of the selected inspectors perform the experiment. Depending on the application of the results of the experiment, stratified sampling may be required to obtain a representative sample. For instance, if the capability will apply to two facilities and one of them inspects twice as many components as the other, then that facility should have twice as many inspectors in the experiment. An alternative method is to characterize each facility independently. Care is then required in combining the results for the joint characterization. Factorial Experiments for Hit/Miss Data. The analysis for the hit/miss data requires that all factors be balanced in any one analysis. When practical, this can be most easily achieved by performing complete factorial experiments. For example, Table 1 contains the results of a two-factor experiment, with the factors being cracks and inspectors. These data can be analyzed as one data set with three inspections per crack. The resulting POD(a) function and its confidence bound would be representative of the population of inspectors from which the sample was drawn.

If the effect of a third factor, for example, different lots of penetrants, were to be included, the entire experiment would be repeated for each of the lots chosen at random from the population of all lots. If three lots were sampled, a total of nine inspections would be performed on each of the flawed specimens, and the resulting POD(a) would apply to the entire inspection process. Suppose, however, that the second and third samples of penetrant were used only by Inspector A. In this case, the two additional sets of inspection data cannot be combined with the other three in a single analysis, because the triple representation of Inspector A would bias the resulting POD(a) function toward his specific capability. The three sets of inspection results for Inspector A can be combined, but the range of applicability of the answer is limited to Inspector A (unless it can be shown or assumed that Inspector A is representative of the entire population). When many factors must be considered, the number of possible combinations in a factorial experiment can easily become prohibitive. More sophisticated experimental designs (fractional replications, for example) may then be required. In such cases, the assistance of a professional statistician is recommended. Experimental Design for

Data. Inspection-result data in the form contain considerably more information than hit/miss data and, as a consequence, permit more flexibility in the design of the experiment. In analysis, the parameters of the POD(a) function are estimated from the slope, intercept, and standard deviation of residuals of the ln ( ) versus ln (a) relation, as given by Eq 9, 12, and 13. In Eq 9, can be considered to be the sum of random effects, and experiments can be designed to estimate the components of the total variation in . For example, operators, probes, and repeatability can be jointly evaluated in a factorial experiment and their effects accounted for in the estimate of POD(a). The statistical model for this experiment would be:

ln ( ) = β0 + β1 ln (a) + Ci + Oj + Pk + Rl + (interaction terms)

(Eq 14)

where Ci, Oj, Pk, and Rl are the random effects due to cracks, operators, probes, and repeats, respectively. The random term, δ of Eq 9, is the sum of all random effects. It can be assumed that the mean and variance of random effect X are zero and

, respectively. Then:

=

+

+

+

+...

Therefore, β0 and β1 can be estimated from a regression analysis, and variance using the expected mean squares for the random effects.

(Eq 15) can be estimated from the components of

In principle, any statistical design from which the components of variance can be estimated can be used in an NDE reliability experiment. However, the analysis methods would be specific to the particular design, and it is beyond the scope of this article to address the general problem. In the section "Signal Response Analysis" in this article, it will be

assumed that only the variation due to cracks and one other factor is being investigated. It is recommended that the assistance of a qualified statistician be obtained for more sophisticated experimental designs. Sample Sizes and Flaw Sizes Sample sizes in NDE reliability experiments are driven more by the economics of specimen fabrication and characterization than by the desired degree of precision in the estimate of the POD(a) function. Although apparently reasonable POD(a) functions can often be obtained from applying the maximum likelihood analysis to relatively few test results, the confidence bound calculation is based on asymptotic (large sample) properties of the estimates. It should be emphasized that the calculations can also produce totally unacceptable results from the relatively few test results or from data that are not reasonably represented by the assumptions of the models. Therefore, there are minimal sample size requirements that must be met to provide a degree of reasonable assurance in the characterization of the capability of the system. Larger sample sizes in NDE reliability experiments will, in general, provide greater precision in the estimate of the POD(a) function. However, the sample size is determined from the number of cracks in the experiment, and there is a coupling with the flaw sizes that must also be considered. The effect of this coupling manifests itself differently for the hit/miss and analyses. Sample Size Requirements for Hit/Miss Analysis. Data from hit/miss experiments are generally not amenable to testing assumptions regarding the form of the POD(a) model. These tests require either large numbers of independent inspections on each flaw of a specimen set or inspection results from an extremely large number of compatible specimens (Ref 3, 4). Number and size considerations in hit/miss experiments are directed at their effect on the sampling properties of the parameters of the POD(a) function (Ref 9).

In the hit/miss analysis, the output of an inspection states only whether or not a crack of known length was found in the inspection (Table 1). There are probabilities associated with the outcomes, and the analysis assumes that this probability increases with flaw size. Because it has been assumed that the inspection process is in a state of control, there is a range of flaw sizes over which the POD(a) function is rising. In this flaw size range of uncertainty, the inspection system has limited discriminating power in the sense that detecting or failing to detect would not be unusual. Such a range might be defined by the interval (a0.10, a0.90), where ap denotes the flaw size that has probability of detection equal to p; that is:

POD(ap) = p

(Eq 16)

Flaws smaller than a0.10 would then be expected to be missed, and flaws greater than a0.90 would be expected to be detected. In a hit/miss reliability experiment, flaws outside the range of uncertainty do not provide as much information concerning the POD(a) function as cracks within this range. Cracks in the almost-certain detection range and almost-certain miss range provide very little information concerning probability of detection. Therefore, in the hit/miss experiment, not all flaws convey the same amount of information, and the effective sample size is not necessarily the total number of flaws in the experiment. Adding a large number of very large flaws does not increase the precision in the estimate of the parameters of the POD(a) function. In a reliability experiment, the location of the increasing range of the POD(a) function is not known. Further, the same sets of specimens are often used in many different experiments. Therefore, it is not possible to fabricate a set of specimens with optimal flaw sizes for a particular experiment. To minimize the chances of completely missing the crack size range of maximum information and to accommodate the multiple uses of specimens, flaw sizes should be uniformly distributed between the minimum and maximum of the sizes of potential interest. A minimum of 60 flaws should be distributed in this range, but as many as economically possible should be used. Sample Size Requirements for

Analysis. The recorded signal response, , provides significantly more information for analysis. In particular, the POD(a) model is derived from the correlation of the versus a data, and the assumptions concerning the POD(a) model can be tested using the signal response data. Further, the pattern of responses can indicate an acceptable range of extrapolation. Therefore, the range of crack sizes in the experiment is not as critical in an analysis as in a hit/miss analysis. For example, if the decision threshold in Fig. 5 were set at 250 counts, all but four of the cracks would have been detected. The larger cracks would have provided little information about the

POD(a) function in a hit/miss analysis. In an analysis, however, all of the recorded values provided full information concerning the relation between signal response and crack size, and the values at the signal threshold and saturation limit provided partial information. The linearity of the fit, the normality of the deviations, and the constancy of the residual variation can all be easily evaluated from the versus a plot. Because of the added information in the data, it is recommended that at least 30 flaws be present in experiments whose results can be recorded in this form. Increasing the number of flaws increases the precision of the estimates, so the test set should contain as many flawed specimens as economically feasible. Unflawed Inspection Sites. In the context of the analyses presented in this section, sample size refers to the number

of known flaws in the specimens to be inspected. The total specimen set should also contain at least twice this number of unflawed inspection sites. The unflawed sites are necessary to ensure that the NDE procedure is discriminating between flawed and unflawed sites and to provide an estimate of the false call rate. Although the false call rate can have important economic consequences, the NDE reliability analyses in this section were dictated by the requirements of damage tolerance analyses. The primary objective was to estimate the chances of missing flaws that might lead to structural failures. The concepts of these NDE reliability analyses can be generalized to include a non-zero probability of a flaw indication when no flaw is present at an inspection site, that is, POD(a = 0) > 0.

References cited in this section

3. W.H. Lewis, W.H. Sproat, B.D. Dodd, and J.M. Hamilton, "Reliability of Nondestructive Inspections--Final Report," SA-ALC/MME 76-6-38-1, San Antonio Air Logistics Center, Kelly Air Force Base, Dec 1978 4. A.P. Berens and P.W. Hovey, "Evaluation of NDE Reliability Characterization," AFWAL-TR-81-4160, Vol 1, Air Force Wright-Aeronautical Laboratories, Wright-Patterson Air Force Base, Dec 1981 9. A.P. Berens and P.W. Hovey, The Sample Size and Flaw Size Effects in NDI Reliability Experiments, in Review of Progress in Quantitative Nondestructive Evaluation 4B, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1985 NDE Reliability Data Analysis Alan P. Berens, University of Dayton Research Institute

Maximum Likelihood Analysis Parameter estimation based on the principle of maximizing the likelihood of an observed sample of data is a standard statistical technique and is amply described in the literature (Ref 10, 11). The purpose of this section is to summarize the method and its asymptotic sampling distribution properties in the context of analyzing NDE reliability data. Further, a method for using this information to calculate lower confidence bounds on the POD(a) function is also presented. Parameter Estimation. Let Xi represent the outcome of the ith inspection and f(Xi;θ) represent the probability of obtaining Xi, where θ= (θ1, θ2,. . .θk)' is the vector of the k parameters in the probability model. For example, in a hit/miss experiment, Xi would be 0 or 1 with probability defined by Eq 4, where a is the size of flaw i and θ= (μ,σ)'. In an versus a experiment, Xi is the log of the signal response, and f(Xi;θ) is a normal density function with mean and standard deviation given by θ= (β0 + β1 ln a, σ )', as defined in Eq 9. Let X1,. . ., Xn represent the results of independent inspections of n flaws. The likelihood, L, of a specific result is given by the likelihood function:

(Eq 17)

For a given outcome of the experiment, Xi is known and Eq 17 is a function of θ. The maximum likelihood estimate is the value, , which maximizes L(θ). For the models considered here, it is more convenient to work with the log L(θ):

log L( ) = which is also maximized at equations:

log f(Xi; )

(Eq 18)

. The maximum likelihood estimates are given by the solution of the k simultaneous

(Eq 19) Asymptotic Sampling Distribution Properties. For the models being used in NDE reliability studies, the

maximum likelihood estimates are invariant, consistent, and efficient. Further, they are asymptotically joint normally distributed with means given by the true parameter values, θi, and the variance-covariance matrix defined by:

V = I-1

(Eq 20)

where I is the information matrix whose elements Iij are the expected (E) values:

(Eq 21)

In application, the maximum likelihood estimate, , is substituted for θ in Eq 21. Therefore, given the results of inspecting a large number of flaws and a specific function for the POD(a) model, the parameters of the model can be estimated, and the sampling distribution of the parameters will be joint normal with the known variance-covariance matrix. Examples of these equations for the hit/miss and response signal models are given in the sections "Analysis of Hit/Miss Data" and "Signal Response Analysis," respectively, in this article. In these applications, the assumed models will be the log odds and cumulative log normal distribution functions. However, other models can also be used if evidence is available to support their selection. Confidence Bounds on the POD(a) Function. Because the POD(a) function is equivalent to a cumulative

distribution function and the parameters are being estimated by maximum likelihood, a procedure developed by Cheng and Iles (Ref 12 and 13) can be used to place lower confidence bounds on the POD(a) function. Such bounds are calculated from the variance-covariance matrix of the estimates and reflect the sensitivity of the experiment to both the number and sizes of flaws in the specimens of the experiment. The assumed POD(a) model is a cumulative log normal distribution function with parameters θ= (μ,σ)'. For distribution functions defined by location and scale parameters (as is the case of the log normal distribution), the information matrix can be written in the form:

(Eq 22)

where n is the number of cracks in the experiment. The lower one-sided confidence bound of the POD(a) function is given by:

POD (a) = where

( - h)

(z) is the standard cumulative normal distribution, and:

(Eq 23)

(Eq 24)

(Eq 25)

where

is obtained from Table 3 for the number of cracks in the experiment and the desired confidence level.

Table 3 Values of Sample size

for lower confidence bounds on the POD(a) function

Confidence level, %

90

95

99

20

3.903

5.243

8.401

25

3.884

5.222

8.376

30

3.871

5.208

8.359

40

3.855

5.191

8.338

50

3.846

5.180

8.325

60

3.839

5.173

8.317

80

3.831

5.165

8.306

100

3.827

5.159

8.300

3.808

5.138

8.273

References cited in this section

10. H. Cramer, Mathematical Methods of Statistics, Princeton University Press, 1946 11. J.F. Lawless, Statistical Models and Methods for Lifetime Data, John Wiley & Sons, 1982 12. R.C.H. Cheng and T.C. Iles, Confidence Bands for Cumulative Distribution Functions of Continuous Random Variables, Technometrics, Vol 25 (No. 1), Feb 1983, p 77-86 13. R.C.H. Cheng and T.C. Iles, One Sided Confidence Bands for Cumulative Distribution Functions, Technometrics, Vol 32 (No. 2), May 1988, p 155-159

NDE Reliability Data Analysis

Alan P. Berens, University of Dayton Research Institute

Analysis of Hit/Miss Data Estimation of the parameters of the log odds model for hit/miss data is based directly on the probability of each 0 or 1 result of an inspection. Assume that a balanced experiment has produced k inspections on each of n cracks. For this application, the likelihood function is given by:

(Eq 26)

where Zij = 0 or 1 for the jth inspection of the ith flaw producing a miss or a find, respectively, and the probability of detecting a flaw of size ai is given by:

(Eq 27)

(Eq 28)

This form of the POD(a) function is simply a more convenient algebraic form of Eq 4. The vector of parameters to be estimated is defined by θ= (μ,σ)'. The log likelihood equation is:

(Eq 29) Parameter Estimation in Hit/Miss Analysis. The maximum likelihood estimates are given by the solution to:

(Eq 30)

(Eq 31) Taking the derivatives and simplifying yields:

(Eq 32) (Eq 33)

Any standard computational method, such as the Newton-Rhapson iterative procedure (Ref 14), can be used to find the solutions to Eq 32 and 33.

Because iterative techniques converge to local maxima, the solution to Eq 32 and 33 may be sensitive to the initial values. A set of initial values based on the method of moments has been found to be useful (Ref 7). These are given by:

(Eq 34)

(Eq 35)

where X1,. . .,Xn are the ordered values of the natural logs of the flaw sizes and pi is the observed percentage of detections of the ith ordered flaw size. If convergence is not obtained, increasing the initial estimate of a has often provided convergence. However, Eq 32 and 33 are not always solvable. This will be discussed further in the section "Comments on Hit/Miss Analysis" in this article. Confidence Bound Calculation in Hit/Miss Analysis. The information matrix is estimated from Eq 21, using

and

for and

. For this POD(a) model, the elements of the information matrix are given by:

(Eq 36)

(Eq 37)

(Eq 38)

Note that k0, k1, and k2, the parameters required in the calculation of the lower confidence bound on the POD(a) function, are also defined by Eq 36, 37, and 38. All of the parameters required by Eq 23, 24, and 25 to calculate the lower confidence bound on the POD(a) function are available. Hit/Miss Analysis Examples. As examples of the application of the method to real data, the parameters of the log

odds POD(a) function were obtained for the data in Table 1. Table 4 presents a summary of the parameters of the POD(a) function for each data set of Table 1 and the combination of the three data sets in a single analysis. Figure 7 shows the POD(a) function and the lower 95% confidence bound for Inspector A and the same information when the data from the three inspectors are combined. Adding inspections of the same cracks did not increase the precision of the estimate of the POD(a) function. Figure 8 compares the POD(a) functions for the three inspectors and the composite. Table 4 POD(a) parameters for the hit/miss data in Table 1 Parameter

Composite

Inspector

A

B

C

0.96

1.11

0.82

0.96

0.59

1.04

0.87

0.88

a50, mm (in.)(a)

2.62 (0.103)

3.03 (0.119)

2.27 (0.089)

2.61 (0.103)

a90, mm (in.)(b)

5.34 (0.210)

10.6 (0.417)

6.54 (0.257)

7.18 (0.283)

a90/95, mm (in.)(c)

21.6 (0.850)

232 (9.13)

38.8 (1.53)

51.0 (2.01)

(a) a = exp( ) = estimate of crack size at 50% POD. 50

(b) a = exp( 90

+ 1.282

) = estimate of crack size at

90% POD.

(c) a90/95 = upper 95% confidence bound on the estimate of a90.

Fig. 7 POD(a) function and lower 95% confidence bound from hit/miss analysis of the data in Table 1 for one inspection per crack (from Inspector A) and for three inspections per crack (from the composite result of Inspectors A, B, and C)

Fig. 8 POD(a) functions from hit/miss analysis of the data in Table 1

Comments on Hit/Miss Analysis. In a well-designed experiment of sufficient sample size for which the log odds

model is a reasonable representation of the POD(a) function, the maximum likelihood hit/miss analysis will provide a valid solution. Conversely, lacking any of these elements, it is possible that either no solution or an unacceptable solution can result. If there is no overlap in the flaw size ranges of the detections and misses, Eq 32 and 33 will not yield a solution. More flaws are needed in the region of increase of the POD(a) function. It is also possible to obtain an estimate of a POD(a) function that decreases with flaw size if the inspection system is poorly designed or not in control and if large flaws tend to be missed more often than small flaws. Both of these types of results are readily apparent, albeit disconcerting. A third type of unacceptable result is an apparently acceptable POD(a) function but a confidence bound that eventually decreases with flaw size. This situation is most easily understood in terms of the log odds versus log flaw size plot. If the slope is positive, the POD(a) function will appear reasonable, but if it is not significantly greater than zero, the lower confidence bound will eventually decrease with flaw size. Therefore, a decreasing confidence bound is evidence of lack of fit of the log odds model. Finally, lack of fit of the model is often manifest in large values of coupled with small values of or extremely wide confidence intervals. Although there are, in general, insufficient data in hit/miss experiments to test hypotheses about the POD(a) model, as a minimum each fit should be subjectively judged. For example, in Fig. 9, the observed detection proportions of each crack in the data of Table 1 are superimposed on the composite POD(a) function and confidence limit from Fig. 7. The uncertainty in the POD(a) function as indicated by the width of the confidence bound seems justified by the plot of the raw data. In this example, if greater precision (narrower confidence bounds) were desired, more cracks in the 2 to 8 mm (0.08 to 0.3 in.) range would be needed in the experiment. Such plots provide an indication of the fit of the model to the data as well as the range of flaw sizes that are contributing to the information from which the POD(a)

function is being estimated. This is true even for experiments in which there is only one inspection per crack and all detection probabilities are plotted at 0 or 1.

Fig. 9 Example fit of hit/miss POD(a) function and lower 95% confidence bound to observed detection probabilities (three inspections per crack)

References cited in this section

7. A.P. Berens and P.W. Hovey, "Flaw Detection Reliability Criteria, Volume I--Methods and Results," AFWAL-TR-84-4022, Air Force Wright-Aeronautical Laboratories, Wright-Patterson Air Force Base, April 1984 14. A. Ralston, A First Course in Numerical Analysis, McGraw-Hill, 1965 NDE Reliability Data Analysis Alan P. Berens, University of Dayton Research Institute

Signal Response Analysis In signal response data analysis, the parameters of the POD(a) function are calculated from parameters of the versus a relation. If all the values are between the signal recording threshold and the saturation limit, a simple regression analysis of ln ( ) versus ln (a) will yield the necessary information to estimate the POD(a) function and its lower confidence

bound. In fact, the least squares estimates from the regression analysis also happen to be the maximum likelihood estimates. The analysis presented in this section is applicable to the more general case in which some of the values are censored at the recording threshold or the saturation limit. This more general analysis will give answers identical to those of the regression analysis if all values are available for all of the flaws (Ref 15). In the response signal analysis, it is assumed that the and standard deviation given by:

=

0

+

1

values for a flaw of size a have a normal distribution, with mean

(Eq 39)

ln (a) (Eq 40)

=

where does not depend on the crack size. To simplify the notation, let Yi = ln ( i) and Xi = ln (ai). The random variable:

(Eq 41)

has a standard normal distribution. Let

(z) represent the density function of the standard normal distribution:

(Eq 42)

and

(z) represent the cumulative normal distribution:

(Eq 43)

The likelihood function is partitioned into three regions: • • •

Region R, for which values were recorded Region T, for which only a maximum value is known (the values fall below the recording signal threshold and cannot be recorded) Region S, for which only a minimum value is known (the values fall above the saturation limit and cannot be recorded)

These regions are identified by the open circles, the down arrows, and the up arrows, respectively, in Fig. 5. The likelihood function for the entire sample is the product of the likelihood functions for the three regions:

L(

0,

1,

But (suppressing the dependency of L on

(Eq 44)

)= 0,

1,

and

):

(Eq 45)

(Eq 46)

(Eq 47)

(Zi) dz is the probability of observing i for the ith flaw in R, i(ath) is the probability of obtaining an ln because 1/ value i value below the recording threshold for the ith flaw in S, and 1 i (asat) is the probability of obtaining an above the saturation limit for the ith flaw in T. The log of the likelihood function is:

(Eq 48)

where r is the number of cracks in R, that is, the number of cracks for which values were recorded. Parameter Estimation in

Analysis. The maximum likelihood estimates are given by the solutions to:

(Eq 49)

(Eq 50)

(Eq 51)

where:

(Eq 52) (Eq 53)

Standard numerical methods, such as the Newton-Rhapson iterative procedure (Ref 14), can be used to find the solutions to Eq 49, 50, and 51. Excellent choices for the initial estimates of iterative procedures are the intercept, slope, and standard deviation of residuals obtained from a standard regression analysis of only those values for which a valid response was recorded.

Confidence Bound Calculation in

Analysis. Because the POD(a) parameters are calculated from the estimates of the versus a data, the calculation of the lower confidence bound is a five-step process:

•

The information matrix for the estimates of

•

The variance-covariance matrix of 0, 1, and is obtained by inverting the information matrix (Eq 20) The variance-covariance matrix of the estimates of and are calculated based on a first-order Taylor series expansion of the equations relating 0, 1, and to and (Eq 12 and 13) The information matrix for and is obtained by inverting the variance-covariance matrix to obtain Eq 22 The calculated values are substituted into Eq 23, 24, and 25 to obtain the lower confidence bound

• • •

The elements of the information matrix for = (

0,

1,

0,

1,

and

is obtained using Eq 21

) are given by (dropping the subscripts):

(Eq 54)

(Eq 55)

(Eq 56)

(Eq 57)

(Eq 58)

(Eq 59)

(Z) = V(Z) [V(Z) - Z]

(Eq 60)

(Z) = -W(Z) [W(Z) + Z] Let V (

0,

1,

) represent the variance-covariance matrix of the maximum likelihood estimates of the ln ( ) versus ln

(a) analysis. The value V ( 1,

(Eq 61)

0,

1,

) is obtained from the inverse of the information matrix. Let the elements of V (

0,

) be defined by:

(Eq 62)

Using a Taylor series expansion about the true values of and variance-covariance matrix of and is given by (Ref 16):

to linearize the relations expressed by Eq 12 and 13, the

(Eq 63)

and the transformation matrix T is defined by:

(Eq 64)

Multiplying the matrices yields the variances and covariance of

and

as:

(Eq 65) (Eq 66) (Eq 67)

Inverting this variance-covariance yields the values of k0, k1, and k2 required in Eq 25 to calculate the lower confidence bound on the POD(a) function.

Multiple Inspections Per Flaw. Repeat values for the same flaw can be analyzed to estimate the magnitude of the total variability being introduced by factors other than the flaws in the experiment. In essence, the random term, , of Eq 9, can be partitioned into components that can be estimated if the experiment is properly designed. The relative magnitude of the components of variance indicates potential areas for improving the system. However, the methods for using the values to generate POD(a) functions and confidence bounds from complex experiments with censored values are still under development.

If there are no censored values, the following analysis provides valid estimates of both the POD(a) function and its confidence bound. If there are censored values, the POD(a) parameter estimates are valid, but the confidence bound is approximate. Much of the data recorded in the format is from automated systems. In these systems, the variability in values has been dominated by that associated with different flaws. This variability is correctly analyzed in the following analysis. In applications to date, the approximation to account for the secondary sources of variability has been judged to be negligible. To model an additional source of variability, Eq 9 is rewritten as:

Yij = ln ( ij) = 0 + 1 ln (ai) + ci + rij i = l, . . ., n j = l, . . ., k

(Eq 68)

where ij represents the jth observation on the ith flaw, ci is the random deviation of flaw i from the fit, and rij is the random deviation about ci introduced by replicate j of the source of the repeated observations. For this model of ln ( ) versus ln (a):

= ci + rij

(Eq 69)

Because it is reasonable to assume that the variability introduced by flaws is independent of that introduced by other factors:

=

(Eq 70)

+

where can be estimated as the pooled-within variance of were recorded (Ref 17):

i

values for each flaw using those flaws for which

values

(Eq 71)

where n* is the number of flaws with uncensored values and ki is the number of uncensored values for flaw i. Between-crack variability, , cannot be estimated directly, but can be estimated indirectly from a censored regression analysis. First, the mean log response for each flaw is calculated. This mean response may be a simple average if all values are available, but a mean based on an analysis of censored data (as previously discussed) will be required for the flaws for which values were censored at the decision threshold or saturation limit. The analysis in the section "Parameter Estimation in Analysis" in this article is then used on the model:

Yi = to obtain estimates

0,

1,

and

0

+

1

Xi + ci +

* . However:

(Eq 72)

(Eq 73)

Solving Eq 73 for

and substituting into Eq 70 yields:

(Eq 74)

The value ( * )2 is obtained from the censored regression analysis, and Equation 74 provides the estimate of to be used in the POD(a) function.

is obtained directly from the ln ( ) values.

To date, the lower confidence bound has been placed on the POD(a) function using the information matrix derived from the censored regression through the average of the ln ( ) values for each crack. This procedure does not account for the added uncertainty resulting from the estimate of . When is significantly greater than , the error introduced by neglecting this variation is judged to be negligible. (Most of the applications of the analysis have been on highly automated eddy current systems. In these applications, the variability of ln ( ) values within cracks has been significantly less than that between cracks.) The method of constructing exact confidence bounds is under development. This process can be extended to account for more than one component of the variability of ln ( ) values. Examples of Analysis. The data in Table 2 resulted from three eddy current inspections of 28 fatigue cracks in flat plates simulating the web/bore of an aircraft engine disk. The three inspections are from the use of three different probes, with all other factors being held fixed. Table 5 presents a summary of the parameters of the POD(a) function for each data set in Table 2 and the combination of the three data sets in a single analysis. The ln ( ) versus ln (a) data from probe A are presented in Fig. 5. This data set had two inspection results below the recording threshold and five results above the saturation limit. The POD(a) function for probe A and its lower 95% confidence bound are shown in Fig. 10. The combined results from the three probes yield the composite POD(a) function and its lower 95% confidence bound, which are also shown in Fig. 10. If the probes were in some sense selected at random from the population of all probes applicable to this equipment and if other factors were assumed to be representative of the application, this composite POD(a) function would be representative of the inspection system. Figure 11 shows the individual POD(a) functions for all three probes and their composite.

Table 5 Example of POD(a) parameters determined from the data in Table 2 The decision threshold is 250 counts. Parameter

Composite

Probe

A

B

C

-1.89

-1.80

-2.12

-1.96

0.24

0.23

0.40

0.33

a50, mm (in.)(a)

0.15 (0.006)

0.17 (0.007)

0.12 (0.005)

0.14 (0.0055)

a90, mm (in.)(b)

0.21 (0.0083)

0.22 (0.009)

0.20 (0.008)

0.21 (0.0083)

(a) a = exp( ) = estimate of crack size at 50% POD. 50

(b) a = exp( 90

+ 1.282

) = estimate of crack size at 90% POD.

(c) a90/95 = upper 95% confidence bound on the estimate of a90.

Fig. 10 POD(a) function and lower 95% confidence bound from signal response analysis of the data in Table 2 with one inspection per crack (with probe A) and with three inspections per crack (from the composite result of probes A, B, and C)

Fig. 11 POD(a) functions from signal response analysis of the data in Table 2

Comments on

Analysis. There are several advantages of the analysis over the hit/miss analysis in estimating the POD(a) function. These accrue primarily because of the added information contained in the values.

It is not as critical in the experiment to have flaws with sizes in the range of increase of the POD(a) function. Because the analysis does not depend on whether or not flaws were detected, the decision threshold can be arbitrarily set after the experiment. Even if the recording threshold is close to the decision threshold, the method permits extrapolation of the results from larger flaw sizes to the smaller sizes of interest. The extrapolation is reasonable over the flaw size range for which the ln ( ) versus ln (a) relation is linear, and deviations from the fit are normally distributed with constant variance. A principal advantage of the analysis is that statistical tests of the underlying assumptions are readily available (Ref 17). In most of the experiments that have been analyzed by this method, the assumptions could not be rejected. However, for very large flaws, a tendency for the ln ( ) versus ln (a) relation to bend down has been observed on occasion. In this region, the values are very large, and all of the flaws are easily detected. One method of linearizing the ln ( ) response is to restrict the analysis to a range of smaller flaw sizes by censoring values at a lower saturation limit. This is equivalent to deleting data points and may reduce the sample size to an unacceptable number. In this case, the flaw sizes in the experiment were not appropriate for the application. It should be noted that ignoring this type of nonlinearity tends to produce a nonconservative POD(a) function, because the effect is to produce a smaller value of the median detectable crack size, . When the data do not fit the model, it is possible to obtain results that have no relation to reality. This can occur if ln ( ) is not an increasing linear function of ln (a) or if there are outliers in the data set that have a significant influence on the analysis. The effects of these anomalies are sometimes manifested in unreasonable values of or a or by a lower

confidence bound that eventually decreases for large crack sizes. As a minimum, it is recommended that a plot of ln ( ) versus ln (a) be obtained for all experiments. This will permit at least a subjective judgment concerning the assumptions of the analysis. If the data do not fit the model, it is necessary to ensure that the process is in control and that the experiment was properly designed and executed. It may also be possible to use other relations between and a as the basis of analysis, but these have not yet been explored.

References cited in this section

14. A. Ralston, A First Course in Numerical Analysis, McGraw-Hill, 1965 15. M. Glaser, Regression Analysis With Dependent Variable Censored, Biometrics, Vol 21, June 1965, p 300307 16. S.R. Searle, Linear Models, John Wiley & Sons, 1971 17. W.J. Dixon and F.J. Massey, Jr., Introduction to Statistical Analysis, McGraw-Hill, 1957 NDE Reliability Data Analysis Alan P. Berens, University of Dayton Research Institute

References 1. 2. 3.

4. 5.

6. 7.

8. 9.

10. 11. 12. 13.

B.G.W. Yee, F.H. Chang, J.C. Couchman, G.H. Lemon, and P.F. Packman, "Assessment of NDE Reliability Data," NASA CR-134991, National Aeronautics and Space Administration, Oct 1976 W.D. Rummel, Recommended Practice for Demonstration of Nondestructive Evaluation (NDE) Reliability on Aircraft Production Parts, Mater. Eng., Vol 40, Aug 1982, p 922-932 W.H. Lewis, W.H. Sproat, B.D. Dodd, and J.M. Hamilton, "Reliability of Nondestructive Inspections-Final Report," SA-ALC/MME 76-6-38-1, San Antonio Air Logistics Center, Kelly Air Force Base, Dec 1978 A.P. Berens and P.W. Hovey, "Evaluation of NDE Reliability Characterization," AFWAL-TR-81-4160, Vol 1, Air Force Wright-Aeronautical Laboratories, Wright-Patterson Air Force Base, Dec 1981 A.P. Berens and P.W. Hovey, Statistical Methods for Estimating Crack Detection Probabilities, in Probabilistic Fracture Mechanics and Fatigue Methods: Applications for Structural Design and Maintenance, STP 798, J.M. Bloom and J.C. Ekvall, Ed., American Society for Testing and Materials, 1983, p 79-94 D.E. Allison et al., "Cost/Risk Analysis for Disk Retirement--Volume I," AFWAL-TR-83-4089, Air Force Wright-Aeronautical Laboratories, Wright-Patterson Air Force Base, Feb 1984 A.P. Berens and P.W. Hovey, "Flaw Detection Reliability Criteria, Volume I--Methods and Results," AFWAL-TR-84-4022, Air Force Wright-Aeronautical Laboratories, Wright-Patterson Air Force Base, April 1984 D.R. Cox, The Analysis of Binary Data, Methuen and Co., 1970 A.P. Berens and P.W. Hovey, The Sample Size and Flaw Size Effects in NDI Reliability Experiments, in Review of Progress in Quantitative Nondestructive Evaluation 4B, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1985 H. Cramer, Mathematical Methods of Statistics, Princeton University Press, 1946 J.F. Lawless, Statistical Models and Methods for Lifetime Data, John Wiley & Sons, 1982 R.C.H. Cheng and T.C. Iles, Confidence Bands for Cumulative Distribution Functions of Continuous Random Variables, Technometrics, Vol 25 (No. 1), Feb 1983, p 77-86 R.C.H. Cheng and T.C. Iles, One Sided Confidence Bands for Cumulative Distribution Functions, Technometrics, Vol 32 (No. 2), May 1988, p 155-159

14. A. Ralston, A First Course in Numerical Analysis, McGraw-Hill, 1965 15. M. Glaser, Regression Analysis With Dependent Variable Censored, Biometrics, Vol 21, June 1965, p 300-307 16. S.R. Searle, Linear Models, John Wiley & Sons, 1971 17. W.J. Dixon and F.J. Massey, Jr., Introduction to Statistical Analysis, McGraw-Hill, 1957 Models for Predicting NDE Reliability J.N. Gray, T.A. Gray, N. Nakagawa, and R.B. Thompson, Center for NDE, Iowa State University

Introduction FRACTURE CONTROL PHILOSOPHIES, as discussed in previous articles in this Section, depend on a damage-tolerant design. The essential feature of such an approach is the incorporation of redundant load paths so that, even if local failures do occur, the structure will be safe for a period of time and can either be removed from service or repaired. Parts containing readily detectable damage can be used if that damage will remain in stable condition until the next inspection opportunity. The implementation of such an approach requires the knowledge and integration of stress, flaw size, and failure mechanisms. As shown in Fig. 1 for the case in which fatigue can be modeled by linear-elastic fracture mechanics, cyclic stress excursions, Δσ, and flaw size, a, are required as inputs to fracture mechanics to predict a stress intensity range, ΔK. If the stress intensity exceeds a critical value known as the plane-strain fracture toughness, KIc, catastrophic failure is imminent. Otherwise, crack growth laws of the form da/dN = A(ΔK)m (where N is the number of stress excursions and A and m are constants) can be used to estimate the safe life available before catastrophic failure.

Fig. 1 Methodology of lifetime prediction for metal parts undergoing cyclic fatigue. Source: Ref 1

Implementation of this damage-tolerant design approach rests on three methodologies: stress analysis, nondestructive evaluation (NDE), and failure modeling. Incorporation of these methodologies in the design of damage-tolerant

components would ideally take advantage of analytical or numerical computations to model the expected lifetime performance of a component. At present, extensive capabilities are in place for modeling stresses and failures, and these are widely used in the design process. However, the modeling of nondestructive evaluation is not nearly as widely accepted. Instead, frequent use is made of empirical rules based on extensive demonstration programs. For both economic and time reasons, there is a significant need to develop a model base for estimating NDE reliability (which is often measured in terms of the probability of flaw detection at given confidence level). It is the purpose of this article to present the current status and future directions of efforts to develop such a capability. This is not intended as a review of international efforts in NDE reliability modeling but rather as a summary of the authors' experience in modeling the inspectability of aerospace components, with emphasis on engine components. Therefore, attention is given to ultrasonic, eddy current, and radiographic inspection. Broader sets of references for the case of ultrasonics can be found in recent review articles (Ref 2, 3). Of particular note is the work performed by the Central Electricity Generating Board in modeling the inspectability of nuclear power generating components (Ref 4). The details of models for NDE reliability are partially dictated by their envisioned uses (Ref 5, 6), which are conceptually illustrated in Fig. 2. One would like to have a model that predicts the probability of detecting flaws of various sizes. This would clearly require as inputs the design of the component, its history of processing and service, and a specification of the inspecting methodology to be used. Given the specifications of these input parameters, it could first be asked whether the predicted probability of flaw detection of the NDE system is adequate to meet the demands imposed by the required performance of the component. Should the expected probability of detection (POD) be inadequate, the model could be exercised to modify the specifications of the inspection, the design of the component, or the processing or service profiles. The NDE models thus become an integral part of a broader concept known as unified life cycle engineering or simultaneous engineering (Ref 7). The essential feature is that one should consider all aspects of the life of a component in the design process, including the ability to inspect and maintain the component, rather than just the initial costs. This will ultimately lead to more sophisticated networking of models, as shown in Fig. 3. Figure 3 illustrates a number of factors that must be added to the traditional computer-aided design and manufacturing (CAD/CAM) methodologies to produce a design optimized for life cycle performance. Multiple interactions among the various factors must be considered to allow the design to address simultaneously all the issues associated with damage tolerance (Ref 8).

Fig. 2 Diagram of probability of detection model and its application to NDE system qualification and optimization and to computer-aided design for inspectability. POD, probability of detection

Fig. 3 Schematic of possible linkages needed for unified life cycle engineering

Given these objectives, it is obvious that an NDE model must exhibit certain characteristics. First, it must predict the response of a real measurement system, as influenced by the specific characteristics of commercially available probes and instruments, rather than an idealized response based on assumptions such as plane wave illumination. Second, the models should give as outputs the information obtained by real inspection protocols. For example, if a signal strength is compared to a threshold as a criterion for detection, this operation should be simulated by the model. If separate protocols are followed in detection and sizing, these should be described by separate models. Third, the models should be used to develop more reliable standardization approaches. This is necessary to ensure that inspections specified in the design process are uniformly implemented by NDE units at the various manufacturing and maintenance departments encountered by the component in its lifetime. Fourth, it is desirable in certain industries for the models to be configured such that they can be integrated with standard CAD packages. This article discusses some ultrasonic, eddy current, and x-ray radiography models that have been developed to exhibit the characteristics mentioned above. As noted previously, primary emphasis is placed on formulations that have been developed in response to the particular needs of the aerospace industry. This article also presents a broader discussion of possible future applications of a reliability modeling capability.

References

1. S.T. Rolfe and J.M. Barson, Fracture and Fatigue Control in Structures: Application of Fracture Mechanics,

Prentice-Hall, 1977 2. R.B. Thompson and T.A. Gray, Use of Ultrasonic Models in the Design and Validation of New NDE Techniques, Philos. Trans. R. Soc. (London) A, Vol 320, 1986, p 329-340 3. R.B. Thompson and H.N.G. Wadley, The Use of Elastic Wave-Material Structure Interaction Theories in NDE Modeling, CRC Crit. Rev. Solid State Mater. Sci., in press 4. J.M. Coffey and R.K. Chapman, Application of Elastic Scattering Theory for Smooth Flat Cracks to the Quantitative Prediction of Ultrasonic Defect Detection and Sizing, Nucl. Energy, Vol 22, 1983, p 319-333 5. R.B. Thompson, D.O. Thompson, H.M. Burte, and D.E. Chimenti, Use of Field-Flaw Interaction Theories to Quantify and Improve Inspection Reliability, in Review of Progress in Quantitative Nondestructive Evaluation, Vol 3A, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1984, p 13-29 6. T.A. Gray and R.B. Thompson, Use of Models to Predict Ultrasonic NDE Reliability, in Review of Progress in Quantitative Nondestructive Evaluation, Vol 5, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1986, p 911 7. H.M. Burte and D.E. Chimenti, Unified Life Cycle Engineering: An Emerging Design Concept, in Review of Progress in Quantitative Nondestructive Evaluation, Vol 6B, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1987, p 1797-1809 8. D.O. Thompson and T.A. Gray, The Role of NDE in Global Strategies for Materials Synthesis and Manufacturing, Proceedings of the 1988 Fall Meeting, Materials Research Society, in press Models for Predicting NDE Reliability J.N. Gray, T.A. Gray, N. Nakagawa, and R.B. Thompson, Center for NDE, Iowa State University

Ultrasonic Inspection Model Empirical determinations of ultrasonic inspectability based on demonstration experiments are of limited utility because their predictions cannot, in general, be extrapolated to new situations beyond the bounds of the data set upon which they are based. Additional data are needed in order to apply them to other cases, and the costs and time required to develop such results for all (or many) possible component designs and scan plans are prohibitive. However, the physical principles upon which many different ultrasonic inspection techniques are based, as applied to a variety of components, are quite similar. Therefore, a mathematical model, whose validity has been proved against a relatively small amount of empirical data, can be used to accurately predict inspectability beyond the bounds of the experimental evidence. The foundation of a physically based mathematical or computer model of ultrasonic inspectability is an analytical formalism, a numerical algorithm, or a combination of the two incorporating the physical principles of the measurement. Such a model consists of descriptions of the waves radiated by the probes, their modification by the geometry of the testpiece, the wave propagation and scattering from defects, and the effects of signal processing and display. The first four sections below review the technical details of each of these elements as developed to describe measurements made in aircraft engine components. In the fifth section, the use of the models to predict the POD of flaws is discussed. The prediction of the POD for flaws is emphasized because of the importance of that parameter in damage-tolerant design. Other uses of models, such as assisting in the interpretation of data during flaw characterization and sizing, are also important but are not explicitly discussed. Some early applications of POD models are summarized in the sixth section. Reciprocity Relation. The ultrasonic NDE simulation models described in this section are based on the formalism of

the electromechanical reciprocity relationship of Auld (Ref 9). This relationship, when specialized to the case of elastic wave scattering, can be expressed as follows. Assume that two identical ultrasonic transducers, a and b, are placed in a fluid to be used in an immersion, pitch-catch measurement (a single probe, or pulse-echo, configuration is a special case) of a component containing a flaw, F. Let be the ratio of the electrical signal radiated into coaxial line b by the receiving transducer to the electrical signal incident on the transmitting probe from coaxial line a. Then the change, F, in this signal induced by the presence of the flaw in the insonified region of the component is given by:

(Eq 1)

where a and a are the displacement vector field and the stress tensor, respectively, that would be produced in the presence of the flaw when probe a is excited by an electrical power P, b and b are the fields that would have been established in the absence of the flaw if probe b had been excited, the overhead dot denotes time differentiation, SF is an arbitrary closed surface enclosing the flaw, and n is the inward normal to that surface. Equation 1 is an exact result known as the electromechanical reciprocity relationship. Its simplicity, however, belies its intractability, except in a very few special cases of limited practical use. Measurement Model. A number of approximations and simplifications are necessary in deriving an accurate yet

computationally efficient inspection measurement model from Eq 1. One situation for which a useful and accurate model can be extracted from Eq 1 through rather remarkable simplifications is the pulse-echo inspection of isotropic, homogeneous elastic materials containing small flaws of fairly simple shape. This is applicable, for example, to typical ultrasonic inspections of gas turbine aircraft engine components, in which fracture-critical flaws are quite small because of the high stresses created during engine operation. The assumption of small in this case means that the dimensions of the defect are small relative to the variations in the transverse profile of the ultrasonic beam. The ultrasonic fields can be approximated locally as plane waves whose displacement and stress fields are the same as those of the true fields (Ref 10). Scattered fields can also be simply modeled, provided their variation is not significant over the face of the receiving transducer. These fields can be represented by the product of a spherically spreading wave times the far-field, unbounded medium scattering amplitude of the flaw (Ref 10). This scattering amplitude A is formally defined as:

(Eq 2) where us is the displacement amplitude of the fields at a distance r from the center of the flaw, u0 is the amplitude of an illuminating plane wave, and ks is the wavenumber of the scattered wave mode. A time-harmonic term of the form exp (i t) is implicitly assumed in this discussion. However, the model has the capacity to deal with pulses, as described later. Applying these approximations to Eq 1, the a fields (those produced when probe a is fired and the flaw is present) are expressed as the sum of a plane wave illuminating field and a scattered field due to that plane wave, using the scattering amplitude as in Eq 2. The b field is simply a plane wave term representing illumination by probe b in the absence of the flaw. After some manipulation of the resulting integral and neglecting higher-order terms, a measurement model is derived that represents the measured signal as a product of factors describing the effects of transducer efficiency, transmission through interfaces, attenuation and beam spread, and scattering (Ref 10). Specifically, one finds:

(Eq 3)

where β is an efficiency factor; TaCa and TbCb represent local plane wave amplitudes at the flaw depth for the a and b fields (these are the product of interface transmission, T, and beam diffraction, C, factors); A is the scattering amplitude of the flaw; ρ1Vb and ρ0V0 are the acoustic impedances of the solid and fluid media, respectively; kb is the wavenumber for the received wave mode; aT is the transducer radius (which is assumed to be the same for both a and b); and the two exponential terms represent the ultrasonic phase change and attenuation, respectively. Equation 3 represents only one frequency. A frequency spectrum can be obtained via superposition of individual frequency terms. Thus, time-domain waveforms, such as would be observed on an oscilloscope screen, can be obtained by an inverse Fourier transform of this measurement model spectrum. (Further details of this derivation and the determination of the efficiency factor in Eq 3 can be found in Ref 10.) Work at a number of institutions has led to formulas, algorithms, and data bases for the ultrasonic beam and scattering amplitude factors in Eq 3, so that a variety of useful simulations can be made, including

the use of planar or focused probes (Ref 11, 12, 13, and 14), inspection through planar or curved liquid/solid interfaces (Ref 11, 12, 14, and 15), and scattering from both volumetric (Ref 16, 17, 18) and cracklike (Ref 17, 19) defects. These aspects of the measurement model will be described in subsequent sections in this article. An alternative formulation, appropriate when the flaw is not small, is found in Ref 4. As an example, Fig. 4 shows a comparison of an experimental pulse-echo radio-frequency (RF) waveform obtained from a circular flat crack versus two model-based simulations. The magnitude of the Fourier spectrum of each RF signal is also shown. The simulated crack was a circular, disk-shaped cavity 0.8 mm (0.032 in.) in diameter and 0.08 mm (0.003 in.) thick located in a diffusion bond plane of a specimen of a nickel-base powder metal alloy (IN 100). The specimen was a 25 mm (1.0 in.) thick plate. The face of the crack was parallel to the surface of the sample, and a 10-MHz, 6.35 mm (0.25 in.) diam unfocused transducer was tilted approximately 7° from normal to the sample surface to generate a 30° refracted longitudinal wave in the sample.

Fig. 4 Comparisons between experimental and model-predicted RF waveforms (top) and their Fourier spectra (bottom) for 30° longitudinal wave backscatter from a 0.8 mm (0.03 in.) diam circular crack in IN100. (a) Experimental measurements of scattering amplitudes (top) and their Fourier spectra (bottom). (b) Model of scattering with method of optimal truncation. (c) Model of scattering with the Kirchhoff approximation

The scattering amplitudes used in Fig. 4 were either results of the method of optimal truncation (MOOT) (Ref 17), which is a computationally intensive algorithm, or the elastodynamic Kirchhoff approximation (Ref 19). The experimental and MOOT results are very nearly identical in both the time and frequency domains. Because the MOOT results are in quasiexact agreement with the measured scattering amplitudes in this case, this comparison highlights the accuracy of the measurement model. The Kirchhoff model result fails to reproduce some of the detailed wiggles of the experimental and MOOT results, but quite accurately reproduces the overall signal amplitude (voltage). This amplitude is the measured quantity that is routinely used in ultrasonic flaw detectors. Therefore, the approximation, which in this case yields a simple and computationally efficient model, can simulate a practical inspection problem. Note, however, that the error of the Kirchhoff model will depend on the flaw size and the orientation and polarization of the ultrasonic wave. Care must be taken to ensure that sufficient accuracy is obtained in particular applications by validating the model through comparison with controlled experiments and/or more exact theories for special cases.

Beam Models. To perform the preceding comparison, one essential element in the simulation was the representation of

the ultrasonic fields in the vicinity of the flaw. Because of the finite size of any realistic transducer, these fields will be quite complex, exhibiting peaks and valleys along the axis of the beam and side-lobes away from the axis. The simulations shown in Fig. 4 represent the case of a scatterer on the axis of the beam, for which a number of approximate beam models have been generated (Ref 11, 13, 14). In a typical automated ultrasonic scan of a component, however, a defect in that component will not generally lie along the axis of the ultrasonic beam. The degree of misalignment will depend, for example, on the coarseness of the scan mesh used to inspect the part. To simulate such an inspection situation, it is necessary to incorporate the full fields of an ultrasonic transducer into the model. This is a formidable task because of the elastic (tensor) nature of wave propagation in a solid and because of the need to consider the interaction of the probing fields with a possibly curved liquid/solid interface at the component surface. However, two approximations have emerged as useful models of transducer radiation patterns: the Gaussian model (Ref 12) and the Gaussian-Hermite model (Ref 13, 14, 20). The former, and simpler, of these models assumes that the transverse profile of the radiation profile is Gaussian in shape at all distances from the probe. This Gaussian beam model provides a set of simple algebraic formulas that predict diffraction effects (beam spread only), effects of lenses, and refraction/focusing due to transmission through curved liquid/solid interfaces (Ref 12). However, typical ultrasonic transducers do not generate Gaussian radiation patterns. For example, typical piston-type radiators exhibit side-lobes and peaks and nulls along the axis of the probe in the near field. However, in the far field (that is, several times the near-field distance), the Gaussian model, if suitably normalized, does accurately predict the amplitude and width of the main lobe in the radiation pattern of a typical piston-type transducer (Ref 12). One application of this approximation, therefore, is the simulation of the fields near a focal region, which can occur either as a result of an acoustic lens on the probe or the focusing effect of a curved component surface. The Gaussian-Hermite model is based on a series expansion of the radiated fields of a transducer in terms of a complete set of orthogonal solutions to a reduced wave equation (Ref 13, 14, 20). These functions are products of a Gaussian factor and a Hermite polynomial. The coefficients in the series expansion are obtained by integrating the product of the Gaussian-Hermite functions and the velocity distribution on the face of a probe over its area. This distribution and the shape of the probe face are arbitrary, so that virtually any probe shape, lens type, and so on, can be modeled. The laws for transmission through curved liquid/solid interfaces and propagation in elastic isotropic media are implemented as simple algebraic operations. The primary disadvantage of the Gaussian-Hermite model is that it is a series solution and therefore can require significantly longer computation times than the Gaussian approximation because of the need for a large number of terms, especially in the near field. However, this becomes less of a disadvantage as computational speeds continue to increase as a result of advances in computer hardware. Scattering Approximations. Another key element in the simulation of ultrasonic inspection of structures is the

model, or models, for representing the interaction of the probing ultrasonic fields with defects. In the most general case, this is represented by a complicated and computationally intractable integral, such as Eq 1. In some cases, however, the effects of the probing ultrasonic fields can be separated from the scattering effects. Specifically, under the assumptions that led to the measurement model (Eq 3), the ultrasonic beam can be described by one of the models just mentioned. For example, elastic wave scattering can be modeled through the use of a far field, unbounded medium scattering amplitude, whose definition was given in Eq 2. Fortunately, considerable research effort has been directed over the past several years toward the development of various models, approximations, and solutions for scattering amplitudes of both volumetric and cracklike defects (Ref 3). For application to the ultrasonic inspection of jet aircraft engine components, a reasonable inventory of scattering models includes formalisms for both volumetric and cracklike flaws. For volumetric flaws, of ellipsoidal shape and arbitrary orientation, both voids and inclusions can be represented by an elastodynamic Kirchhoff approximation (Ref 18). This approximation is exact in its treatment of the strength of the front surface reflection ( function). It is valid for both longitudinal and shear wave backscatter, such as would be used to simulate pulse-echo inspections. The limitation of this model, however, is that it is accurate only for early-time events in the scattering. Therefore, it does not predict the amplitude of scattered fields that have reverberated within an inclusion; in some cases, these scattered fields can be of higher amplitude than the initial front surface reflection. The Kirchhoff approximation is therefore a conservative model for scattering from volumetric flaws. It does have the benefit of simplicity and computational efficiency. For cracklike flaws, an elastodynamic Kirchhoff approximation to scattering from internal flat cracks of elliptical shape has been implemented for both pulse-echo and pitch-catch techniques and for longitudinal and shear wave modes (Ref 19). This Kirchhoff approximation accurately predicts the specular content (mirror reflection) of scattering, but does not properly include edge diffraction contributions or surface wave modes. It also does not contain any provision for surface roughness or partial closure of the crack faces. It does, however, yield reasonably accurate predictions of signal

amplitudes in the near-specular regime, as can be inferred from Fig. 4. One fairly significant limitation of the model is its inaccuracy in the nonspecular regime. It predicts, for example, that the scattering amplitude is identically zero for edge-on incidence from a crack, which is inconsistent with established crack scattering results (Ref 17, 21). However, the model is of significant utility because an inspection system for detecting cracks would be set up to take advantage of specular orientations, if possible. Furthermore, in this case, the Kirchhoff approximation is a simple and computationally fast algorithm. Probability of Detection Models. One important application of ultrasonic measurement modeling is the simulation of probability of detection. Detection is appropriately described in terms of a probability for several reasons. A given size and type of defect may occur at random positions and with a range of orientations within each of a set of nominally identical components. The detailed shape of the defect may vary in a way that influences its ultrasonic response differently from its fracture response. Variations in grains, porosity, surface roughness, and so on, as well as the electronic equipment in a detection system, will cause noise, which will interfere with the signals from a flaw. Therefore, a given size and type of defect will exhibit a distribution of signal amplitudes if measured in a population of components containing such defects. Because these signal amplitudes are compared to a preset amplitude threshold in typical flaw detectors, some of the flaws will be missed, while others will be detected. The POD is the ratio of the number of flaws that are detected to the total number of flaws in the inspected components.

A formalism to predict the probability, p(S, N, T), that a given signal, S (video envelope), in the presence of noise with total power N2 will be detected by exceeding a threshold, T, is given by:

(Eq 4)

where io(z) = exp (-z)I0(z), with I0(z) being the modified Bessel function of the first kind and order zero. This approach, based on work performed by Rice (Ref 22), was developed for a narrow-band signal, as is typical in radar analysis. Equation 4 represents only the probability of detecting a single signal level, S, and does not take into account the distribution of signal amplitudes for variations in the size, shape, and type of flaw. A POD model for ultrasonics has been developed by using the measurement model to calculate the variability of signal levels as influenced by the position and orientation of the flaws. Then, for a given root-mean-square noise level, N, Eq 4 can be used to represent the probability that the video signal from a specific defect (a given size, shape, type, location, orientation, and so on) in a given component (material, geometry, and so on) and using a specific inspection system will exceed a threshold amplitude. As an example, let px and p represent the probability distribution functions for the location, x, and orientation, θ, of a given size, shape, and type of defect, and let S(x, θ) represent its signal amplitude, as simulated by the measurement model. An expression for POD, assuming a noise level N and a detection threshold amplitude T, can then be written formally as:

(Eq 5)

where the integrals are taken over the range of orientations and positions of possible flaws (Ref 15). Obvious generalizations of Eq 5 can treat the effects of flaw shape, type, and so on. Equation 5 does not contain an explicit factor to represent the probability distribution of the presence of flaws; therefore, it predicts the probability that an ultrasonic indication (signal plus noise) will be detected assuming that a flaw is, in fact, present. The use of the POD model to analyze reliability issues, such as the probabilities of falsely accepting flawed components or of falsely rejecting good ones, would require the incorporation into the model of the probability distribution function for flaws as a function of defect size, location within the part, and so on. The result of Eq 5 would then need to be further integrated with respect to that probability distribution. Applications. Figure 5 shows the results of simulating the detectability (POD) of circular cracks at three different

depths below a cylindrical component surface and for two different scan plans (Ref 6). Figure 5(a) illustrates the use of

the POD model to quantify the detection capability of an NDE system. For the specific parameters in that simulation, cracks that are otherwise identical have significantly different detectability levels, depending on their depth below the surface of the part. In this example, the curved surface of the component behaves like an acoustic lens that happens to focus the beam at the middle depth (25 mm, or 1 in.). Because of the relatively coarse scan mesh and the reduced beam width in this focal region, there is a significant likelihood that a flaw will be located in a low-amplitude portion of the beam profile and, therefore not be detected. For the other two depths, the beam width is greater than the scan mesh distance. The plot in Fig. 5(b) shows the result of sufficiently refining the scan mesh so that the beam width at the focal region is greater than the distance between scan points. In this case, the POD curves are nearly the same for the three different depths. This example illustrates the use of the POD model for quantifying the capability of a flaw detection system and for suggesting improvements in this area in the system or its operation.

Fig. 5 Predicted influence of scan plan on POD. (a) POD at three depths for axial and circumferential scan increments of 2.5 mm (0.1 in). (b) POD at three depths for axial and circumferential scan increments of 5 and 1.3 mm (0.2 and 0.05 in.), respectively

Another example of the use of the POD model is shown in Fig. 6 (Ref 6). The POD curve is expressed in a rather nonstandard manner, because it does not represent the typical POD versus flaw size plot. Instead, Fig. 6 shows the variation in POD due to modification of a geometrical parameter of a component. Specifically, the component is a simulated turbine disk assumed to contain radially oriented circular cracks below a bicylindrical fillet. The abscissa of Fig. 6 is the in-plane radius of curvature of that fillet. The flaw size is assumed to be constant, representing, for example, the critical flaw size as predicted by fracture mechanics. Most important, Fig. 6 shows that flaw detectability can be improved by modifying the geometry of the component. This information is important for the definition of sonic near-net shapes of components in production, for example, and ensuring in-service inspectability during maintenance. Moreover, because the characteristics of the inspection system are contained explicitly in the POD model, the scan plan required to achieve the necessary detectability levels are easily determined. This concept of predicting component inspectability at the design stage, of determining the component design parameters that favorably influence flaw detectability, and of incorporating the requisite scan procedure into the design data base is perhaps the ultimate application of inspectability modeling.

Fig. 6 Simulated POD for a 0.8 mm (0.03 in.) diam circular crack below a bicylindrical fillet as a function of fillet radius of curvature

References cited in this section

3. R.B. Thompson and H.N.G. Wadley, The Use of Elastic Wave-Material Structure Interaction Theories in NDE Modeling, CRC Crit. Rev. Solid State Mater. Sci., in press 4. J.M. Coffey and R.K. Chapman, Application of Elastic Scattering Theory for Smooth Flat Cracks to the Quantitative Prediction of Ultrasonic Defect Detection and Sizing, Nucl. Energy, Vol 22, 1983, p 319-333 6. T.A. Gray and R.B. Thompson, Use of Models to Predict Ultrasonic NDE Reliability, in Review of Progress in Quantitative Nondestructive Evaluation, Vol 5, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1986, p 911 9. B.A. Auld, General Electromechanical Reciprocity Relations Applied to the Calculation of Elastic Wave Scattering Coefficients, Wave Motion, Vol 1, 1979, p 3 10. R.B. Thompson and T.A. Gray, A Model Relating Ultrasonic Scattering Measurements Through LiquidSolid Interfaces to Unbounded Medium Scattering Amplitudes, J. Acoust. Soc. Am., Vol 74, 1983, p 1279 11. R.B. Thompson and T.A. Gray, Analytical Diffraction Corrections to Ultrasonic Scattering Measurements, in Review of Progress in Quantitative Nondestructive Evaluation, Vol 2, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1983, p 567 12. R.B. Thompson and E.F. Lopes, The Effects of Focusing and Refraction on Gaussian Ultrasonic Beams, J. Nondestr. Eval., Vol 4, 1984, p 107 13. R.B. Thompson, T.A. Gray, J.H. Rose, V.G. Kogan, and E.F. Lopes, The Radiation of Elliptical and Bicylindrically Focused Piston Transducers, J. Acoust. Soc. Am., Vol 82, 1987, p 1818 14. B.P. Newberry and R.B. Thompson, A Paraxial Theory for the Propagation of Ultrasonic Beams in

Anisotropic Solids, J. Acoust. Soc. Am., to be published 15. B.P. Newberry, R.B. Thompson and E. F. Lopes, Development and Comparison of Beam Models for TwoMedia Ultrasonic Inspection, in Review of Progress in Quantitative Nondestructive Evaluation, Vol 6, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1987, p 639 16. C.F. Ying and R. Truell, Scattering of a Plane Longitudinal Wave by a Spherical Obstacle in an Isotropically Elastic Solid, J. Appl. Phys., Vol 27, 1956, p 1086 17. J.L. Opsal and W.M. Visscher, Theory of Elastic Wave Scattering: Applications of the Method of Optimal Truncation, J. Appl. Phys., Vol 58, 1985, p 1102 18. J.-S. Chen and L.W. Schmerr, Jr., The Scattering Response of Voids--A Second Order Asymptotic Theory, in Review of Progress in Quantitative NDE, Vol 7, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1988, p 139 19. L. Adler and J.D. Achenbach, Elastic Wave Diffraction by Elliptical Cracks: Theory and Experiment, J. Nondestr. Eval., Vol 1, 1980, p 87 20. B.D. Cook and W.J. Arnoult III, Gaussian-Laguerre/Hermite Formulation for the Nearfield of an Ultrasonic Transducer, J. Acoust. Soc. Am., Vol 59, 1976, p 9 21. J.D. Achenbach, A.K. Gautesen, and H. McMaken, Ray Methods for Waves in Elastic Solids, Pittman Publishing, 1982 22. S.O. Rice, Mathematical Analysis of Random Noise, Bell Syst. Tech. J., Vol 23, 1944, p 282; Vol 24, 1945, p 96 Models for Predicting NDE Reliability J.N. Gray, T.A. Gray, N. Nakagawa, and R.B. Thompson, Center for NDE, Iowa State University

Eddy Current Inspection Model The eddy current NDE method has a long history of use (Ref 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, and 34). Because of its conceptually straightforward design principle, the method allows measurements to be made in a noncontacting, single-sided inspection geometry. Eddy current is therefore a cost-effective and truly nondestructive method of inspection. However, as a fully quantitative method, eddy current inspection has not achieved the level of sophistication found in other methods, such as ultrasonics and x-ray radiography. The reason for the delay in its development is associated with the fundamental nature of the measurement. In essence, the measurement response is a complex function of the probe fields and their interaction with the flaw, and it is difficult to isolate these two contributions. Therefore, one cannot develop simple models in which various effects can be described by separate factors, as in the case of ultrasonics. A significant problem has also been obtaining material property data (such as permeability and conductivity) for complex engineering materials. Welds are especially difficult. From a modeling perspective, it is necessary to obtain electromagnetic field solutions for a given probe/flaw system by solving Maxwell's equations. When written in the form suitable to eddy current problems, the equations show that the basic dynamics are of a dissipative, nonscattering nature. One well-known consequence is that only near-surface flaws, confined within a region of a finite skin depth, are detectable. Putting this obvious limitation aside, there is a subtler problem posed by the basic principle. Namely, electromagnetic fields spread over a wide region outside the specimen, permitting no simple way of focusing the probe sensitivity to flaw regions. (In the other inspection methods, the beam focusing, for example, can be used for this purpose). The impedance signal, being an integrated quantity, contains not only flaw information but also redundant environment information (such as probe lift-off), which is uninteresting in terms of flaw detection and characterization and may contribute significant noise. Extracting flaw information from impedance signals therefore becomes highly dependent on component geometry and is a computationally intensive task. Many efforts have been devoted to overcoming this difficulty, and the results have been promising (Ref 25, 26, 27, 28, 29, 30, 31, and 32). Rapid progress has occurred recently in computational methods with regard to both hardware and software. It appears that the new-generation desktop workstations are sufficiently capable of handling the computational requirements of eddy current data analyses. One may find an even better situation when using eddy current models to

assist in design processes, for which state-of-the-art supercomputers may be available. In addition to progress in hardware, researchers are applying certain new software techniques to eddy current problems. With new software running on powerful computers, the difficulty in numerical analyses is becoming obsolete. Recent experimental advances have also proved beneficial. Emphasis is placed, more than before, on making characterizable probes and on operating them in broad frequency ranges. In fact, no fully quantitative eddy current methods can be established without these two capabilities. Incidentally, all existing eddy current POD models discussed in the literature rely on some type of experimental input for noise characterizations (Ref 30, 34). Reciprocity Relation. On the theoretical side, the appearance of a reciprocity formula, proposed by B.A. Auld (Ref 31), has been regarded as a breakthrough in that it provides a formalism for interpreting measurements in realistic experimental geometries. The formula gives the impedance change ΔZ due to the presence of a flaw in the form:

(Eq 6)

where E and H denote electric and magnetic fields in the absence of the flaw, and E' and H' denote the fields in the presence of the flaw. The integration in Eq 6 is carried out over a surface enclosing the flaw, with n being the innerdirected surface normal. The attractive feature of Eq 6 is that flaw signals can be evaluated as soon as electromagnetic fields are obtained in the vicinity of a flaw region. Equation 6 has been proved effective in impedance evaluations and has been used in many applications. Auld's theory, which was built around a combination of the reciprocity relation and the Born approximation, has been playing a central role in the existing eddy current POD studies.

A word of caution is necessary concerning the reciprocity formula. Although its use is highly recommended, the reciprocity formula is not the ultimate method for conducting impedance evaluations. In fact, the most versatile method is the direct use of Faraday's law, which works under any circumstances. However, this method usually results in the most computationally intensive efforts. A third possibility is to use the energy conservation law. This method, however, is the least versatile because it is not always possible to estimate an amount of energy escaping from the system. There are no general criteria for method selection. A case-by-case comparison is required, taking cost-effectiveness and the required degree of accuracy into account. Probe-Flaw Interaction Models. To make the eddy current method quantitative, one must be able to predict flaw

signals for any given system. Earlier, equivalent circuit methods were attempted (Ref 23). The present consensus is that detailed evaluations of electromagnetic fields are needed to make sufficiently precise predictions. One therefore needs to solve Maxwell's equations, assuming that all the system parameters (that is, a probe, a part, a flaw, and the inspection geometry) are given. Such mathematical techniques as exact methods, integral equation methods, perturbation theory, and variational methods are frequently used. Various approximation schemes can also be employed. In the literature, the integral equation methods are used most frequently, where the basic differential equations are cast into integral equations using appropriate boundary conditions. Particularly popular are the three-dimensional finiteelement method and the boundary integral equation (BIE) method. Again, selection of the most suitable method for a given situation should be done on a case-by-case basis. The threedimensional finite-element method is the most versatile, but is also the most expensive. For crack problems, the BIE method may be favored in several aspects. Typically, an eddy current system is piece. wise homogeneous, consisting of several spatial regions separated by sharp boundaries. Moreover, the boundaries are rigid surfaces that are not subject to change over time. Clearly, this is an ideal situation for applying the BIE method. Another favorable aspect is that the aforementioned reciprocity formula uses only field values on boundaries. As long as it is applicable, the BIE approach gives a faster algorithm than the three-dimensional finite-element method. There are, however, situations in which the BIE method fails to work well. For example, it is not effective when dealing with flaws caused by corrosion where conductivity may vary continuously in space. Presumably, this type of flaw is still confined in a local region. If so, it is possible to solve the problem with a combination of the BIE and three-dimensional finite-element methods. To the best of the authors' knowledge, such a combined approach has never appeared in the literature. Perturbation theory may be useful if the variation is small. The perturbation method should certainly be used more frequently in dealing, for example, with a conductivity fluctuation, which is an important noise source to consider in POD models.

Probability of Detection Models. In this section, definitions of probability of detection and probability of false alarm

(POFA) will be given in a form suitable for the eddy current inspection method. Various, inequivalent definitions are seen in the literature, but the intention here is to illustrate the basic idea. Suppose that an eddy current inspection system is specified and that a series of impedance measurements is carried out over a flaw at a given frequency. By nature, the measured signals will scatter around a mean value in a complex impedance plane. Distribution A shown in Fig. 7 illustrates a typical data distribution from a flawed area, while distribution B represents a data set taken under the same controlled conditions but over an unflawed part. Strictly speaking, a flaw signal is given by the difference between the on-flaw and off-flaw signals; the flaw signal corresponds graphically to the shift of the two peak positions, A and B. The fluctuations around the peaks reflect the noise characteristics of the given inspection system.

Fig. 7 Illustration of signal distributions. Measured impedance values will follow a certain distribution. Distribution A illustrates an on-flaw signal distribution (signals measured on a flaw), while B illustrates an offflaw signal distribution.

Once signal/noise distributions are given, one can begin to study the inspectability of the system, as follows. Consider a window in the impedance plane (for example, C in Fig. 8), and make a hypothesis that any signal falling in this window is actually a flaw signal. Then, POD is defined by the probability of this hypothesis being true, which is equal to the integral of distribution A over window C (provided the total integral of A is normalized to unity). Similarly, POFA is given by the integral of B over C, that is, the probability of off-flaw signals being mistaken as flaw signals. The POD and POFA

values thus evaluated are functions of C, and when evaluated for various windows, they will form a domain in a PODPOFA plane. Such POD-POFA plots can be used as a measure of inspectability. Alternatively, representatives of POD and POFA values can be used for specification, such as those in which the maximum POD/POFA ratio is achieved.

Fig. 8 The same distributions as in Fig. 7 viewed from above. Window C will be used to define POD and POFA as integrals of the distributions over C.

Therefore, the fundamental task is to determine the on-flaw and off-flaw signal distributions for a given system. One important theoretical task, for example, is to predict the peak positions A and B with model calculations. Noise distributions are more difficult to predict because, by nature, any measurement system will contain certain unpredictable noise sources. Nonetheless, to make eddy current POD packages an effective design aid, as much effort as possible should be devoted to the complete identification and characterization of conceivable noise sources. References 30 and 34 contain specific examples of POD calculations, and Fig. 9 illustrates an example of a POD-POFA plot.

Fig. 9 Sample plot of POD-POFA curves for different flaw sizes. Source: Ref 12

Discussions and Future Directions. Comparing the goal just stated with the current developmental status in the literature (Ref 30, 34) will lead one to expect improvements in eddy current POD models in the following areas. First, theoretical models of limited capabilities have been used for impedance calculations. These limitations inevitably restrict the applicability of the existing POD models. For example, the models have been probe specific, limited to work only for certain measurement geometries, and so on. Any progress in forward calculations will provide more flexible POD models.

Second, noise characterizations have been accomplished almost exclusively by relying on experimental input. In view of using POD models as a design aid, this is not an ideal situation, because one needs to build a prototype system before calculating POD. Unfortunately, this trend will continue until sufficient data have been accumulated so that reliable general formulas can be extracted empirically. Very little has been done to find purely theoretical noise simulations. Finally, as a part of software development, a program shell of a POD package should be established, following the general guideline of POD evaluations. Ideally, the package will be written based on a modular programming concept. It may consist of three basic modules: impedance calculation, noise calibration, and POD calculation. Once the shell is written, any progress in either forward calculation or noise characterization can be implemented quickly into the package by replacing the appropriate modules. There are also two technical recommendations. First, although POD itself can be calculated for a fixed frequency, it is of crucial importance to enable the system to handle a broad range of frequencies. Without this capability, no precise determinations of flaw sizes and shapes can be expected. This applies equally to hardware and software developments. Second, few of the previous measurement techniques and associated POD models fully exploit the complex nature of impedance. Both the real and imaginary parts should be treated on an equal footing as described above in order to make full use of valuable information. It should be emphasized that, so far, the concept of inspectability has been defined for systems as a whole, so that one can discuss the POD of a system but not the POD of a single component. To validate the notion of the POD of a given probe, it is necessary to integrate the probe into an inspection system in which all the other system components are strictly controlled to meet certain specifications. In the future, such a standardization may become an important issue for comparison purposes.

References cited in this section

12. R.B. Thompson and E.F. Lopes, The Effects of Focusing and Refraction on Gaussian Ultrasonic Beams, J. Nondestr. Eval., Vol 4, 1984, p 107 23. H.L. Libby, Introduction to Electromagnetic Nondestructive Test Methods, Wiley-Interscience, 1971 24. M. Burrows, "A Theory of Eddy Current Flaw Detection," Ph.D. thesis, University of Michigan, 1964 25. C. Dodd and W.E. Deeds, Analytical Solution to Eddy-Current Probe-Coil Problem, J. Appl. Phys., Vol 39, 1968, p 2829 26. A.H. Kahn, R. Spal, and A. Feldman, Eddy-Current Losses Due to a Surface Crack in a Conducting Material, J. Appl. Phys., Vol 48, 1977, p 4454 27. W.D. Dover, F.D.W. Charlesworth, K.A. Taylor, R. Collins, and D.H. Michael, in The Measurement of Crack Length and Shape during Fatigue and Fracture, C.J. Beevers, Ed., EMAS, 1980 28. B.A. Auld, Theoretical Characterization and Comparison of Resonant-Probe Microwave Eddy-Current Testing With Conventional Low-Frequency Eddy-Current Methods, in Eddy Current Characterization of Materials and Structures, STP 722, G. Birnbaum and G. Free, Ed., American Society for Testing and Materials, 1981 29. T.G. Kincaid, A Theory of Eddy Current NDE for Cracks in Nonmagnetic Materials, in Review of Progress in Quantitative NDE, Vol 1, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1982 30. A.J. Bahr, System Analysis of Eddy-Current Measurements, in Review of Progress in Quantitative NDE, Vol 1, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1982; A.J. Bahr and D.W. Cooley, Analysis and Design of Eddy-Current Systems, in Review of Progress in Quantitative NDE, Vol 2A, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1983; J.R. Martinez and A.J. Bahr, Statistical Detection Model, in Review of Progress in Quantitative NDE, Vol 3A, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1984 31. B.A. Auld, F.G. Muennemann, and D.K. Winslow, J. Nondestr. Eval., Vol 2, 1982, p 1; B.A. Auld, F.G. Muennemann, and M. Riaziat, Quantitative Modeling of Flaw Responses in Eddy Current Testing, in Research Techniques in Nondestructive Testing, Vol 7, R.S. Sharpe, Ed., Academic Press, 1984 32. R.E. Beissner, Boundary Element Model of Eddy Current Flaw Detection in Three Dimensions, J. Appl. Phys., Vol 60, 1986, p 352-360; Analytical Green's Dyads for an Electrically Conducting Half-Space, J. Appl. Phys., Vol 60, 1986, p 855-858 33. J.C. Moulder, P.J. Shull, and T.E. Capobianco, Uniform Field Probe: Experiments and Inversion for Realistic Flaws, in Review of Progress in Quantitative NDE, Vol 7B, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1987 34. R.E. Beissner, K.A. Bartels, and J.L. Fisher, Prediction of the Probability of Eddy Current Flaw Detection, in Review of Progress in Quantitative NDE, Vol 7B, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1988 Models for Predicting NDE Reliability J.N. Gray, T.A. Gray, N. Nakagawa, and R.B. Thompson, Center for NDE, Iowa State University

Radiographic Inspection Model In radiographic modeling, one is essentially trying to predict how the two-dimensional projection of an irregularly shaped object changes as the position of the object changes with respect to the x-ray beam and detector and, further, to study the degree to which various parameters control image quality (Ref 35, 36). The requirements for a quantitative model of the radiographic process consist of the ability to predict the output of an x-ray generator, the interaction of the x-ray beam with the sample, and the detector characteristics. The capability to handle complex part geometries and non-trivial flaw morphologies is required if the simulation is to have wide application. Gray et al. have developed a full three-dimensional radiography model that generates a two-dimensional (x,y) film image (Ref 37, 38) rather than the one-dimensional source-sample-detector dimension previously modeled (Ref 39). The full

shadowing effects and the radial divergence of the x-ray beam in the projections are included, allowing a realistic treatment of complex part geometries. The model consists of:

(Eq 7) D(x,y,E) = D0(1 - e

)

(Eq 8)

In Eq 7, E is the energy, I is the intensity immediately above an xy point on the detector, I0, is the initial intensity produced from the x-ray generator, is the energy-dependent linear absorption coefficient, is the x-ray path length through the sample, and r is the distance from the source to the detector. In Eq 8, D is the film density, is the interaction cross section of an x-ray with a film grain, is the coefficient for the xray scattering, t is the time of the film exposure, is the natural film fog density, and D0 is the maximum film density. The integration goes over the area of the source, thus including the affect of source properties on image quality. This model of the x-ray radiographic process has the capability of modeling a very large number of parameters that affect the image quality of an x-ray radiograph. As noted in Eq 7 and 8, the major elements of the model include an x-ray beam model (I0), the interaction of the beam with a nontrivial sample, and a detector model. These must be combined with a detectability criterion in order to complete a POD prediction. Beam Model. The model of the initial x-ray beam involves a first-principles prediction of the energy-dependent x-ray

intensity spectrum calculated from the bremsstrahlung and characteristic interaction cross sections (Ref 40). These cross sections are calculated from the interaction of the relativistic electron beam with the bound atomic electrons of the target atom and are based on a one-photon production process. The angular dependence of the bremsstrahlung cross section is integrated over all angles. This simplification implies that the calculated intensities will require a scaling factor to match the experimentally measured values. The energy dependence of the spectrum is calculated assuming a thick target attenuation of the electron beam. This beam model has features that allow variation in the beam focal-spot size, shape, and uniformity. Further, the target material and angle of incidence to the electron beam are variables. The model accounts for the inherent filtration of the generated x-ray beam by the target material and the vacuum window of the exit port. Features that the beam model does not address are the generation of off-axis x-rays, the interaction cross sections that involve the generation of more than one photon, the physical mechanics of electron beam focusing, and the variation of the voltage from power supply. These features influence the intensity of the beam and at present are handled in the beam model with an intensity scaling factor. Sample Interaction Model. The interaction of the x-ray beam with the sample monitors the energy-dependent

attenuation through the sample (Ref 41). This portion of the model includes the effects of the sample shape, the flaw morphology, and the configuration of the experimental setup. The sample shape currently used in the simulation is a flat plate with surface roughness. The simple part shape allows the effects of many parameters to be isolated from those involving the complex part shapes. Two types of flaws are available for study in the simulation. The first is an ellipsoid with nine independent parameters characterizing its shape, orientation, and position. The position of the origin, the three major axes, and three Euler angles give complete, arbitrary control of the position and shape of the ellipsoid. The ellipsoid can be a void or an inclusion. Further, the sample position relative to the source or detector, the thickness, and the material can be chosen as desired, thus providing a wide range of setup geometries. The second flaw is a truncated cone whose apex angle, truncation distance, composition material, and xyz position in the sample can all be varied. Additional features are the ability to model density variations in the material and a wide choice of sample materials ranging from low atomic number carbonbase materials to high atomic number materials such as tantalum. As noted above, the sample shape is limited. However, future integration of the simulation into a CAD package will remedy the lack of a wide variety of shapes because any part geometry that the CAD package can generate can be used in the simulation. A note on the introduction of complex CADgenerated part shapes is that the calculation of the POD can take significant amounts of time, at least with present computer capabilities. Detector Model. If film is used as the detector, a model of the film simulates a number of film speeds, maximum

densities, and the energy-dependent effects of the x-ray interaction with the silver halide grains. The film model is derived by considering the number of grains that have absorbed x-ray photons, which absorb an additional photon as a result of the flux incident on the film. The number of photons required for a silver halide grain to be developed is energy

dependent. For very low energy (10 keV), several photons are required, while at high energy (1 MeV), one photon can activate several grains. Unfortunately, the energy dependence of grain activation is also dependent on the likelihood of xray absorption, a feature that is dependent on the size of the grain and the energy of the x-ray photon. These energydependent features are reflected in the x-ray interaction cross section, . This cross section, together with the grain size, gives the film model the capability to model the speed of the film (Ref 42). The total amount of silver halide grain present, as represented by the weight of the emulsion, controls the maximum film density, D0. The fog density, , of the unexposed film can be accounted for in the final density. The film model tracks the random counting statistics noise and in a limited set of sample geometries tracks the noise due to Compton scattering. For simple geometries, the Compton scattering is a uniform intensity at the film and is modeled through the scattering coefficient, . This is a simplistic approximation to a complex problem. Indeed, the Compton scattering for energies above several hundred kiloelectron volts is the dominant scattering mechanism and can contribute more to the film density than the information-carrying portion of the beam. The additional complexities of part geometry nullify the uniformity of the Compton scattering at the detector plane, making a detailed tracking of the Compton component necessary. The addition of this important effect is the subject of ongoing research. Detectability. The detectability criterion accounts for the physiological limitation of the human eye to detect, under

normal conditions, variations in gray film scale levels. The determination of the limits of gray-scale variation that the human eye can perceive is based on studies of uniform gray background with a circle that has a different gray-scale level (Ref 35). Although this is not the general condition under which radiographs are viewed, this is the criterion currently incorporated into the model. Further modification involves a mechanism to account for degradation of the contrast due to a poor signal-to-noise ratio. One of the features of the simulation is the capability of turning off the noise, thus allowing examination of the noise degradation of the signal. The noise in the film image is, in most cases, dominated by the random fluctuations of the number of photons impinging on a unit area. The result is that the noise magnitude is related to the density of the film; higher noise is associated with low densities, while low noise corresponds to high density. This has led to the incorporation of contrast degradation due to noise and therefore to the inclusion of its influence on flaw signal detectability. This detectability model falls into the amplitude model class of detectability, and although a simple model, it has the major features for adequately describing detectability (Ref 43). Limitations. At present, there are a number of limitations to the model. They include a limited accounting for the effect

of Compton scattering, the need for an intensity scaling factor for the initial beam, and a simple detectability model. Experimental verification of the major components is under way, and the preliminary results are very good. The initial beam model gives results that only need scaling to accurately model the observed x-ray spectrum. The film model can characterize the film response to a photon flux for a wide range of film types. For the full simulation, with only the scaling for the initial beam intensity, the results, when compared to an actual radiograph of a machined cone-shaped void, are in the worst case within a factor or two of the actual film densities. In the best cases, the model results are within 20% of the observed results. The refinement of the model is work that is ongoing. Applications. With this highly flexible model of the radiographic process, several factors affecting flaw detectability

can be studied in a quantitative manner. A sample of the types of effects studied can be seen in Fig. 10 and 11, in which the loss of detectability was simulated for changes in beam hardness and flaw orientation, respectively. The simulation predicts a range of kilovoltages for which adequate contrast and noise levels are present. To perform adequate x-ray inspections, the data must be collected in these plateau regions. It should be noted that the sizes of these optimal regions are also strongly controlled by the flaw shape and size. The loss of detectability was calculated quantitatively based on the physiological limitation of human vision to discern variations in gray shades, the noise level present in the radiograph, and the contrast available.

Fig. 10 Simulation results showing the effects on thickness sensitivity of the x-ray beam hardness and the relative size of the flaw to sample thickness

Fig. 11 Detectability of different ellipsoidal flaws as a function of their orientation and shape. The shape of the ellipsoid is changed by varying the aspect ratio, where two of the major axes are kept equal and the third is decreased. The ratio of the constant major axes to the smaller is the aspect ratio. As the aspect ratio increases, the ellipsoid flattens and becomes cracklike. The scale for angular orientation is chosen such that a crack is optimally oriented at 90°.

Figure 10 shows a plot of the effect of beam hardness (the white spectrums of a harder beam have a higher average energy) on the detectability of a flaw where the film type, the sample composition, the x-ray beam shape, and the physical geometric setup were kept the same. The time of the exposure was increased to attain the same film densities for each simulated exposure. As can be seen, the harder beam has less sensitivity but a shorter exposure time, while the softer beam yields greater sensitivity at the cost of longer exposure times. Thus, the trade-off between sensitivity and speed of inspection can be studied. Figure 11 shows the effect of flaw morphology on detectability. The flaws represented in Fig. 11 have widely varying aspect ratios, thus simulating varying degrees of crack tightness. As the orientation of these cracks with respect to the incident x-ray beam changes, their detectable signal drops abruptly. There are a number of applications of the forward model of the x-ray radiographic process in the field of production inspection. The first is the collapsing of the full three-dimensional aspect of the model to a one-dimensional line path. This allows quick calculation of the kilovoltage, current, and time required for the exposure of a type of film to the density required. Another is in the area of inspection validation studies. The use of seeded defects and image quality indicators and the sensitivity testing of various parameters are time consuming and expensive, so much so that a full study in which all parameters are varied is rarely done. The use of a simulation program coupled with a set of benchmark trials can greatly expand the scope of this type of certification of an inspection procedure. Modest extensions of the model would facilitate an even wider range of applications. The prediction of a full POD requires only the addition of a probability distribution function describing the flaw distribution, as described in the section "Ultrasonic Inspection Model" in this article. As can be seen from Fig. 10 and 11, the current model is very close to the generation of POD curves for different system parameters. A second area for extension is the simulation of tomographic scans. The addition

of a detector model typical of those used in tomographic scans will allow the generation of the projection data required for the reconstruction algorithms.

References cited in this section

35. R. Halmshaw, Physics of Industrial Radiography, Applied Science, 1966 36. P. DeMeester and W. Aerts, in Research Techniques in Nondestructive Testing, Vol VI, R.S. Sharpe, Ed., Academic Press, 1982 37. J.N. Gray, Three Dimensional Modeling of Projection Radiography, in Review of Progress in Quantitative Nondestructive Evaluation, Vol 7A, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1988, p 343348 38. J.N. Gray, F. Inanc, and B.E. Shull, Three Dimensional Modeling of Projection Radiography, in Review of Progress in Quantitative Nondestructive Evaluation, Vol 8A, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1989 39. D.E. Rundquist, in Research Techniques in Nondestructive Testing, Vol IV, R.S. Sharpe, Ed., Academic Press, 1982 40. C.M. Lee, L. Kissell, R.H. Pratt, and H.K. Tseng, Phys, Rev., Vol A13 (No. 5), 1976, p 1714-1727 41. W.H. McMaster, N. Kerr Del Grande, J.H. Mallett, and J.H. Hubbell, "Compilation of X-ray Cross Sections," UCRL-50174 Section II Rev. 1, Los Alamos National Laboratory, 1974 42. H.H. Barrett and W. Swindell, in Radiological Imaging: The Theory of Image Formation, Detection, and Processing, Vol 1, Academic Press, 1981, p 206 43. L.D. Loo, K. Doi, and C.E. Metz, Phys. Med. Biol., Vol 29 (No. 7), 1984, p 837-856 Models for Predicting NDE Reliability J.N. Gray, T.A. Gray, N. Nakagawa, and R.B. Thompson, Center for NDE, Iowa State University

Future Impact of Models The three preceding sections describe computer models that can be used to generate quantitative predictions of inspectability for ultrasonic, eddy current, and x-ray film methods. These are all physically based models that explicitly consider the characteristics of the realistic flaws, the typical component geometries and materials, and the practical inspection methods and practices. These or similar models may: • • •

Replace costly and time-consuming experimental demonstration programs in the prediction of NDE reliability Improve the validation and optimization of inspection procedures Improve component design and the definition of the life cycles

Another potential application of NDE reliability models may involve integration with CAD systems. This approach could allow the assessment of inspectability during the design stage just as form, fit, and function are analyzed during design. The integration of NDE reliability models with a CAD system may also allow quantitative estimates of inspection reliability as a function of design changes. The following example, illustrated in Fig. 12 and 13, describes this application of NDE/CAD integration for ultrasonic inspectability (Ref 8). The inspectability (POD) model for ultrasonic NDE has achieved the highest state of technical maturity, and initial strides have been made toward its incorporation into CAD systems. However, the general idea is applicable to NDE reliability models of inspection methods other than ultrasonics.

Fig. 12 Example of a CAD-generated display of inspectability (POD) as a function of position within the cross section of a simulated disk. The POD scale ranges from black for the lowest to white for the highest POD values. The low POD in the fillet region is due to the combination of a coarse scan mesh and a tight fillet radius.

Fig. 13 A display similar to the one shown in Fig. 12, except that the scan mesh has been refined and the fillet has a larger radius of curvature. The combination of scan plan and component design modification has significantly improved inspectability throughout the disk.

Example 1: Integration of an Ultrasonic Reliability Model with a CAD System. Figures 12 and 13 show the result of a computer simulation of ultrasonic POD for oblate spheroidal inclusions in a rotationally symmetric component, such as a jet engine turbine disk. The display shows the results obtained from a cross section of the component, and the gray shades refer to the POD value, in the range from zero to unity, with black being the lowest and white the highest POD values. In the POD simulation, the inclusions were assumed to have a preferential orientation with their largest face parallel to the component surface, and the component was assumed to be scanned in pulse-echo mode with a focused probe normal to the surface of the part and focused at the surface. The interior points of the component, at which the POD values were calculated, were defined to be the nodes of an automatically generated finite-element mesh that was imposed on the solid model of the part. This artifice allowed the built-in postprocessing graphical interface of the CAD package to be used to display the inspectability results in a format that would be very familiar to a designer who is experienced in the use of the package. Figure 12 shows the result of a nominal scan plan and component design. In particular, the scan mesh was quite coarse, causing poor detectability in the web section to the right of the cross section and near the surface of the entire component. The fillet region, at the juncture between the web and the thicker bore area to the left of the cross section, was also assumed to have a small radius of curvature. Below this fillet region, the POD values are quite low, which is a result of the combination of the ultrasonic beam being focused on the surface of the part and of the focusing effect of the surface curvature. This caused the fields in the region below the fillet to be rapidly diverging and therefore of low amplitude. To improve the POD levels throughout the component, two questions must be considered. The first is whether the inspection technique was optimized with respect to the design of the component, and the second is whether the design itself is inspection limited. The first question is crucial because a simple modification of the scan plan will almost certainly be a better solution than the change in operational characteristics of the component created by a modification of its design. In some cases, however, both the scan plan and the design will require optimization. Figure 13 shows the result

of both options; the scan mesh was refined, which caused the enhanced POD in the near-surface and web regions, and the fillet radius of curvature was increased, which improved the POD in the fillet region.

Example 2: Implications on Standardization. As the design applications of reliability models become more widespread, the questions of quantitative standards will become more important. Because the manufacturer will usually be at a remote location from the designer and may be affiliated with a different organization, it will be necessary to communicate the inspection procedure at the time that manufacturing procedures are specified. Absolute calibration standards will then be essential to ensure that the inspection, including specified thresholds, is properly set up and implemented. This may require a move away from the use of reference reflectors, such as flat-bottom holes (Ref 44, 45). Recent work in ultrasonics has shown how the measurement model, Eq 3, can be used as the basis for such a procedure (Ref 46).

Example 3: Uses of Artificial Intelligence in Optimization. An important part of the consideration of inspectability in CAD procedures is optimization of the inspection and/or design. This was done intuitively by a highly trained operator in Example 1. However, in the future, it may be desirable to eliminate the operator through the use of artificial intelligence techniques, such as expert systems. Although this has not yet been accomplished, a step in that direction has been the use of the ultrasonic measurement model in the development of an expert system to differentiate cracklike flaws from volumetric flaws (Ref 47).

References cited in this section

8. D.O. Thompson and T.A. Gray, The Role of NDE in Global Strategies for Materials Synthesis and Manufacturing, Proceedings of the 1988 Fall Meeting, Materials Research Society, in press 44. B.R. Tittmann and D.O. Thompson, Approach to a Self-Consistent Calibration Procedure of an Ultrasonic System, Mater. Eval., Vol 35, 1977, p 96-102 45. B.R. Tittmann, D.O. Thompson, and R.B. Thompson, Standards for Quantitative Nondestructive Evaluation, in Nondestructive Testing Standards--A Review, M. Berger, Ed., STP 624, American Society for Testing and Materials, 1977, p 295-311 46. D.D. Bennink and A.L. Pate, Investigation of Scatter in Ultrasonic Responses Caused by Variability in Transducer and Material Properties, in Review of Progress in Quantitative Nondestructive Evaluation, Vol 7A, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1988, p 621-628 47. S.W. Nugen, K.E. Christensen, L.-S.Koo, and L.W. Schmerr, FLEX-An Expert System for Flaw Classification and Sizing, in Review of Progress in Quantitative Nondestructive Evaluation, Vol 7A, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1988, p 445-451 Models for Predicting NDE Reliability J.N. Gray, T.A. Gray, N. Nakagawa, and R.B. Thompson, Center for NDE, Iowa State University

References 1. 2. 3. 4. 5.

S.T. Rolfe and J.M. Barson, Fracture and Fatigue Control in Structures: Application of Fracture Mechanics, Prentice-Hall, 1977 R.B. Thompson and T.A. Gray, Use of Ultrasonic Models in the Design and Validation of New NDE Techniques, Philos. Trans. R. Soc. (London) A, Vol 320, 1986, p 329-340 R.B. Thompson and H.N.G. Wadley, The Use of Elastic Wave-Material Structure Interaction Theories in NDE Modeling, CRC Crit. Rev. Solid State Mater. Sci., in press J.M. Coffey and R.K. Chapman, Application of Elastic Scattering Theory for Smooth Flat Cracks to the Quantitative Prediction of Ultrasonic Defect Detection and Sizing, Nucl. Energy, Vol 22, 1983, p 319-333 R.B. Thompson, D.O. Thompson, H.M. Burte, and D.E. Chimenti, Use of Field-Flaw Interaction Theories

6.

7.

8. 9. 10. 11.

12. 13. 14. 15.

16. 17. 18.

19. 20. 21. 22. 23. 24. 25. 26. 27.

to Quantify and Improve Inspection Reliability, in Review of Progress in Quantitative Nondestructive Evaluation, Vol 3A, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1984, p 13-29 T.A. Gray and R.B. Thompson, Use of Models to Predict Ultrasonic NDE Reliability, in Review of Progress in Quantitative Nondestructive Evaluation, Vol 5, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1986, p 911 H.M. Burte and D.E. Chimenti, Unified Life Cycle Engineering: An Emerging Design Concept, in Review of Progress in Quantitative Nondestructive Evaluation, Vol 6B, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1987, p 1797-1809 D.O. Thompson and T.A. Gray, The Role of NDE in Global Strategies for Materials Synthesis and Manufacturing, Proceedings of the 1988 Fall Meeting, Materials Research Society, in press B.A. Auld, General Electromechanical Reciprocity Relations Applied to the Calculation of Elastic Wave Scattering Coefficients, Wave Motion, Vol 1, 1979, p 3 R.B. Thompson and T.A. Gray, A Model Relating Ultrasonic Scattering Measurements Through LiquidSolid Interfaces to Unbounded Medium Scattering Amplitudes, J. Acoust. Soc. Am., Vol 74, 1983, p 1279 R.B. Thompson and T.A. Gray, Analytical Diffraction Corrections to Ultrasonic Scattering Measurements, in Review of Progress in Quantitative Nondestructive Evaluation, Vol 2, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1983, p 567 R.B. Thompson and E.F. Lopes, The Effects of Focusing and Refraction on Gaussian Ultrasonic Beams, J. Nondestr. Eval., Vol 4, 1984, p 107 R.B. Thompson, T.A. Gray, J.H. Rose, V.G. Kogan, and E.F. Lopes, The Radiation of Elliptical and Bicylindrically Focused Piston Transducers, J. Acoust. Soc. Am., Vol 82, 1987, p 1818 B.P. Newberry and R.B. Thompson, A Paraxial Theory for the Propagation of Ultrasonic Beams in Anisotropic Solids, J. Acoust. Soc. Am., to be published B.P. Newberry, R.B. Thompson and E. F. Lopes, Development and Comparison of Beam Models for TwoMedia Ultrasonic Inspection, in Review of Progress in Quantitative Nondestructive Evaluation, Vol 6, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1987, p 639 C.F. Ying and R. Truell, Scattering of a Plane Longitudinal Wave by a Spherical Obstacle in an Isotropically Elastic Solid, J. Appl. Phys., Vol 27, 1956, p 1086 J.L. Opsal and W.M. Visscher, Theory of Elastic Wave Scattering: Applications of the Method of Optimal Truncation, J. Appl. Phys., Vol 58, 1985, p 1102 J.-S. Chen and L.W. Schmerr, Jr., The Scattering Response of Voids--A Second Order Asymptotic Theory, in Review of Progress in Quantitative NDE, Vol 7, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1988, p 139 L. Adler and J.D. Achenbach, Elastic Wave Diffraction by Elliptical Cracks: Theory and Experiment, J. Nondestr. Eval., Vol 1, 1980, p 87 B.D. Cook and W.J. Arnoult III, Gaussian-Laguerre/Hermite Formulation for the Nearfield of an Ultrasonic Transducer, J. Acoust. Soc. Am., Vol 59, 1976, p 9 J.D. Achenbach, A.K. Gautesen, and H. McMaken, Ray Methods for Waves in Elastic Solids, Pittman Publishing, 1982 S.O. Rice, Mathematical Analysis of Random Noise, Bell Syst. Tech. J., Vol 23, 1944, p 282; Vol 24, 1945, p 96 H.L. Libby, Introduction to Electromagnetic Nondestructive Test Methods, Wiley-Interscience, 1971 M. Burrows, "A Theory of Eddy Current Flaw Detection," Ph.D. thesis, University of Michigan, 1964 C. Dodd and W.E. Deeds, Analytical Solution to Eddy-Current Probe-Coil Problem, J. Appl. Phys., Vol 39, 1968, p 2829 A.H. Kahn, R. Spal, and A. Feldman, Eddy-Current Losses Due to a Surface Crack in a Conducting Material, J. Appl. Phys., Vol 48, 1977, p 4454 W.D. Dover, F.D.W. Charlesworth, K.A. Taylor, R. Collins, and D.H. Michael, in The Measurement of Crack Length and Shape during Fatigue and Fracture, C.J. Beevers, Ed., EMAS, 1980

28. B.A. Auld, Theoretical Characterization and Comparison of Resonant-Probe Microwave Eddy-Current Testing With Conventional Low-Frequency Eddy-Current Methods, in Eddy Current Characterization of Materials and Structures, STP 722, G. Birnbaum and G. Free, Ed., American Society for Testing and Materials, 1981 29. T.G. Kincaid, A Theory of Eddy Current NDE for Cracks in Nonmagnetic Materials, in Review of Progress in Quantitative NDE, Vol 1, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1982 30. A.J. Bahr, System Analysis of Eddy-Current Measurements, in Review of Progress in Quantitative NDE, Vol 1, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1982; A.J. Bahr and D.W. Cooley, Analysis and Design of Eddy-Current Systems, in Review of Progress in Quantitative NDE, Vol 2A, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1983; J.R. Martinez and A.J. Bahr, Statistical Detection Model, in Review of Progress in Quantitative NDE, Vol 3A, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1984 31. B.A. Auld, F.G. Muennemann, and D.K. Winslow, J. Nondestr. Eval., Vol 2, 1982, p 1; B.A. Auld, F.G. Muennemann, and M. Riaziat, Quantitative Modeling of Flaw Responses in Eddy Current Testing, in Research Techniques in Nondestructive Testing, Vol 7, R.S. Sharpe, Ed., Academic Press, 1984 32. R.E. Beissner, Boundary Element Model of Eddy Current Flaw Detection in Three Dimensions, J. Appl. Phys., Vol 60, 1986, p 352-360; Analytical Green's Dyads for an Electrically Conducting Half-Space, J. Appl. Phys., Vol 60, 1986, p 855-858 33. J.C. Moulder, P.J. Shull, and T.E. Capobianco, Uniform Field Probe: Experiments and Inversion for Realistic Flaws, in Review of Progress in Quantitative NDE, Vol 7B, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1987 34. R.E. Beissner, K.A. Bartels, and J.L. Fisher, Prediction of the Probability of Eddy Current Flaw Detection, in Review of Progress in Quantitative NDE, Vol 7B, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1988 35. R. Halmshaw, Physics of Industrial Radiography, Applied Science, 1966 36. P. DeMeester and W. Aerts, in Research Techniques in Nondestructive Testing, Vol VI, R.S. Sharpe, Ed., Academic Press, 1982 37. J.N. Gray, Three Dimensional Modeling of Projection Radiography, in Review of Progress in Quantitative Nondestructive Evaluation, Vol 7A, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1988, p 343348 38. J.N. Gray, F. Inanc, and B.E. Shull, Three Dimensional Modeling of Projection Radiography, in Review of Progress in Quantitative Nondestructive Evaluation, Vol 8A, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1989 39. D.E. Rundquist, in Research Techniques in Nondestructive Testing, Vol IV, R.S. Sharpe, Ed., Academic Press, 1982 40. C.M. Lee, L. Kissell, R.H. Pratt, and H.K. Tseng, Phys, Rev., Vol A13 (No. 5), 1976, p 1714-1727 41. W.H. McMaster, N. Kerr Del Grande, J.H. Mallett, and J.H. Hubbell, "Compilation of X-ray Cross Sections," UCRL-50174 Section II Rev. 1, Los Alamos National Laboratory, 1974 42. H.H. Barrett and W. Swindell, in Radiological Imaging: The Theory of Image Formation, Detection, and Processing, Vol 1, Academic Press, 1981, p 206 43. L.D. Loo, K. Doi, and C.E. Metz, Phys. Med. Biol., Vol 29 (No. 7), 1984, p 837-856 44. B.R. Tittmann and D.O. Thompson, Approach to a Self-Consistent Calibration Procedure of an Ultrasonic System, Mater. Eval., Vol 35, 1977, p 96-102 45. B.R. Tittmann, D.O. Thompson, and R.B. Thompson, Standards for Quantitative Nondestructive Evaluation, in Nondestructive Testing Standards--A Review, M. Berger, Ed., STP 624, American Society for Testing and Materials, 1977, p 295-311 46. D.D. Bennink and A.L. Pate, Investigation of Scatter in Ultrasonic Responses Caused by Variability in Transducer and Material Properties, in Review of Progress in Quantitative Nondestructive Evaluation, Vol 7A, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1988, p 621-628 47. S.W. Nugen, K.E. Christensen, L.-S.Koo, and L.W. Schmerr, FLEX-An Expert System for Flaw

Classification and Sizing, in Review of Progress in Quantitative Nondestructive Evaluation, Vol 7A, D.O. Thompson and D.E. Chimenti, Ed., Plenum Press, 1988, p 445-451

Statistical Quality Design and Control Richard E. DeVor, University of Illinois, Urbana-Champaign; Tsong-how Chang, University of Wisconsin, Milwaukee

Introduction A MAJOR REVOLUTION in the industrial sector has been taking place in America during the 1980s as manufacturers strive to regain the competitive position they once held in the world marketplace. One element of this revolution has centered around a renewed emphasis on quality, with an approach aimed at preventing defective materials from being manufactured through improved process monitoring and diagnosis and at designing quality into the product from the very beginning. The concepts and methods of Deming and others have had a profound impact on the way quality is viewed from the manufacturing/process perspective. The simple but powerful statistical methods for process control developed by Shewhart some time ago have been successfully revived and applied on a very broad basis. In the engineering design arena, the methods of Taguchi, referred to as off-line quality control, have been successfully used for more than 30 years to provide a sound basis for improved product/process design. From the total system point of view, the concept of company-wide quality control, which has been practiced in Japan for some time, is now receiving considerable attention. In particular, recent emphasis has been placed on quality function deployment as a means to transmit customer needs through the organization both vertically and horizontally. The work of Shewhart in the 1920s led to a sound approach to the scrutiny of process variation and the diagnosis and removal of process faults. However, the statistical approach to the sampling of process output prior to shipping to determine the extent to which it conformed to specifications dominated the quality field from the 1930s through the 1970s. Unfortunately, it has only recently been recognized that this product control approach to quality control contributes little to the enhancement of competitive position. The recognition that quality and productivity can move together in the right direction only when the process is evaluated, finding the root cause of process faults and taking action to remove them, is today reshaping the meaning and intent of quality control. The dramatic change in the meaning and application of quality control as a discipline has both a philosophical and an analytical side. These two aspects of the problem must be understood together. This article will present some fundamental elements of statistical thinking and methods for quality design and control. Commonly used techniques employing control charts and the design of experiments are discussed with illustrations.

Acknowledgement This article was adapted from R.E. DeVor and T-H. Chang, Statistical Quality Design and Control: Contemporary Concepts and Methods, Macmillan. Statistical Quality Design and Control Richard E. DeVor, University of Illinois, Urbana-Champaign; Tsong-how Chang, University of Wisconsin, Milwaukee

Quality and Productivity Fundamentals Dr. W. Edwards Deming is the man largely responsible for bringing statistical thinking and methods for quality improvement to Japan after World War II. An eminent statistician in America during the 1930s and 1940s, Deming was one of the men in the forefront of the statistical quality control scene during World War II. After the war, Deming's philosophies and teachings fell on the deaf ears of American management, and the quality effort fires that burned so brightly during the war years slowly went out. However, the Japanese became aware of Deming's work and invited him to Japan, where he met with and caught the attention of top management. When Deming first traveled to Japan in the late 1940s, he found a climate that was quite conducive to the promotion of his concepts and methods. On the one hand, Japan appeared to have a solid base of statistical expertise, although its energies had been directed primarily toward mathematical theory and the application of that theory in nonmanufacturing

environments such as agriculture. On the other hand, Deming found an industrial leadership base in Japan very eager to listen to what he had to say. With the lessons learned from his experience in attempting to promote quality improvement in America, Deming's first order of business when he arrived in Japan was to conduct a series of top management training seminars in which he laid out what needed to be done to place quality improvement on an institutional basis within any organization. The obligations and responsibilities of management that he spelled out in these seminars came to be known as his Fourteen Points for management. Perhaps it is the combination of upper management perspective and a firmly rooted background in mathematical statistics that has enabled Deming to sustain his efforts in a leadership position over more than 50 years. He has been able to tell management what they ought to do and then to provide the rigorous analytical tools and methods needed to carry out his directions. His teachings are heavily oriented toward the use of statistical thinking and methods to identify opportunities for quality and productivity improvement. Deming has developed a road map for management to follow to enhance competitive position over the long run. This road map, referred to as his Fourteen Points (management's obligations), is summarized below. It should be noted that in the spirit of his Fourteen Points, Deming continues to this day to refine and improve these tenets. As a result they themselves are continually changing. An excellent discussion of Deming's Fourteen Points is provided in Ref 1. Deming's Fourteen Points are listed below: • • • • • • • • • • • • • •

Create constancy of purpose for the improvement of product or service Adopt the new philosophy of process control and variation reduction Cease dependence on mass inspection for quality control End the practice of awarding business on the basis of price Improve constantly and forever the system of production and service in order to improve quality and productivity and thus continuously decrease costs Institute thorough and better job-related trainings Institutionalize leaderships Drive out fears, so that everyone may work effectively for the company Break down barriers between departments Eliminate slogans, exhortations, and targets for the workforce that ask for zero defects and new levels of productivity Eliminate work standards on the factory floor Remove the barriers that rob employees at all levels in the company of their right to pride of workmanship Institute a vigorous program of education and self-improvement Put everybody in the organization to work to accomplish the transformation

The Fourteen Points are clearly the responsibility of management. They define the essential elements of the institutionalization of quality and productivity improvement through statistical thinking and methods. Other pioneers in quality planning and management include Dr. J.M. Juran (Ref 2) and Dr. A.V. Feigenbaum (Ref 3). Quality and Engineering Specifications In today's economic age, it is simply impractical to use concepts and measures of quality that do not relate the achievement of function in the field to the engineering design process. Definitions of quality that promote improvement to some acceptable plateau of performance will inhibit the continual pursuit of never-ending improvement and will therefore have a weak and perhaps opposing relationship to process performance in terms of efficiency/productivity. Instead, to improve competitive position, the definition of quality should emphasize: • • • •

A design criterion to provide a quantitative basis to move the quality issue upstream to engineering design Prevention rather than containment to promote focus on the process, not the product, in a manufacturing sense Continual pursuit of never-ending improvement to be strongly tied to, and therefore promote, the issue of productivity Consumer versus producer orientation to quantify/measure the loss imparted to the customer as a result

of poor quality rather than the loss imparted to the producer

The traditional definitions of quality often fall short in terms of their ability to articulate quality in a way that can foster improvement in competitive position. In particular, the association of quality with conformance to the engineering specification puts the measurement of quality on an attribute basis and provides little more than a shipping criterion when it is essential that quality be articulated in a manner that enables it to be used as a design criterion. The view of quality as conformance to specifications promotes the product control approach to quality control and therefore stands as a significant inhibitor to the adoption of a process control approach to manufacturing and the integration of quality and the design process. Figure 1 illustrates the traditional interpretation of the engineering specification. Figure 1(a) and 1(b), two representations of a process as a statistical distribution of measurements are shown relative to a certain bilateral specification. Although the bell-shaped process distribution in Fig. 1(a) might be preferred to the more loaf-shaped process distribution in Fig. 1(b), there generally would not be much distinction between the two cases in terms of quality; that is, in both cases, virtually all the processes/parts are conforming to the specifications.

Fig. 1 Graphical depiction of three traditional interpretations of engineering specifications. (a) Desirable. (b) Acceptable. (c) Nominal

When one begins to consider the dots labeled A, B, and C as representing three different manufactured parts in Fig. 1(c), part A is generally interpreted as unacceptable because it is outside the specifications, while parts B and C are considered acceptable because they are within the specifications. The crucial point in this interpretation is the fact that a distinction is generally not made between parts B and C so far as quality is concerned. In fact, it is unlikely that there is very much difference between parts A and B in terms of functional performance, while part C will likely perform considerably better than either A or B.

Loss Function Concept The methods of Taguchi have provided a whole new perspective on the way the engineering design process is viewed and have provided further support to the methods of design of experiments for quality design and improvement (Ref 4, 5). The design of experiments as a discipline has a rich and extensive history that finds its origin in the work of Sir Ronald Fisher during the first quarter of the 20th century (Ref 6). From an industrial point of view, the work of Dr. G.E.P. Box and his coworkers has left an indelible mark on the design of experiments in particular and the theory and practice of industrial statistics in general (Ref 7, 8). The Taguchi approach to quality engineering has a number of significant strengths. In particular, Taguchi has placed a great deal of emphasis on the importance of minimizing variation as the primary means of improving quality. The concepts of designing products such that the performance of each is insensitive to environmental conditions and making this happen at the design stage through the use of design of experiments have been cornerstones of the Taguchi methodology for quality engineering. Taguchi suggests that it is important to think of quality in terms of the loss imparted to society during product use as a result of functional variation and harmful effects. Harmful effects refer to the side effects that are realized during the use of the product and are unrelated to product function. Functional variation refers specifically to the deviation of product performance from that intended by design, that is, from the nominal. Specific sources of functional variation will be discussed later in this article. Taguchi uses a loss function concept to quantify quality as loss due to functional variation. Figure 2 illustrates the idea of the loss function.

Fig. 2 Plot of quality loss versus quality characteristic to illustrate concept of loss function of quality

The loss function concept as a way to measure quality clearly suggests that the goal of design and manufacturing is to develop products and processes that perform on target with smallest variation. It places tremendous importance on the reduction of variation to achieve the most consistently performing products and processes. To drive performance toward this overall objective, more must be learned about the way in which the important parameters of the product/process influence performance. Over the years, the methods of design of experiments have been proved to be an effective way to do this. Figure 3 provides an interpretation of the loss function concept when overlaid on the traditional engineering specification definition of quality. In this way, no loss would be considered as long as the quality characteristic in question lies within the lower and upper specification. Outside of the specifications, the product is considered to be unacceptable, and a constant loss would therefore be realized. This loss would likely be measured in terms of scrap cost or rework cost. The following case study suggests that this interpretation is illogical.

Fig. 3 Plot of quality loss versus quality characteristic to illustrate loss function of engineering specification

Variation Reduction and the Loss Function. The loss function view of quality suggests that there exists a clear

economic advantage to reducing variation in the performance of a product. The case study described below clearly indicates how such reduced manufacturing imperfection can lead to a reduction in costs and therefore an improvement in competitive position. Several years ago an automobile company performed a study to compare the manufacturing variations evident in certain transmission components for comparable transmissions made in the United States and in Japan. Random samples of transmissions were selected in each case. The transmissions were disassembled, and a number of critical dimensions were measured and recorded. Figure 4 illustrates the general findings of the study. In particular, in the case of the U.S. transmissions, the critical dimensions generally consumed about the middle 75% of the tolerance range specified. Normally, one would conclude that the capabilities associated with the manufacturing processes are therefore well within normal expectations. However, the same critical dimensions for the Japanese counterpart of this transmission consumed only about the middle 25% of the same tolerance range. The question that begs to be asked is, Why would the Japanese strive to make the parts to such tight tolerances?

Fig. 4 Comparison of critical dimensions for transmission components manufactured in the United States versus

those manufactured in Japan

The warranty cost bars shown in Fig. 5 have been plotted at the extremes of the two distributions of the critical dimensions to depict the relative costs associated with variability. It is clear that for the U.S. transmissions the cost associated with repair and replacement is significantly larger than that for their Japanese counterparts. This comparison of the economic data associated with the two transmissions would suggest that there exists a definite relationship between variability and cost in terms of loss incurred due to functional variation. One could think of the warranty costs as literally mapping out the loss function.

Fig. 5 Loss function interpretation of the engineering specifications in Fig. 4 in terms of warranty costs

Robust Design Approach to Quality Design The design process in general and the design/manufacturing interface in particular have received considerable attention over the past 20 years. Computer-aided design and manufacture has dealt primarily with the computerization of the design and manufacturing processes and in particular with the translation of design specifications into manufacturing procedures and activities. The computer has greatly facilitated this translation, and as a result, lead times have been reduced. Design modeling and analysis have been strengthened and extended. Standardization and rationalization of both design characteristics and manufacturing process characteristics have been advanced, and the efficiency of the interactive process at the design/manufacturing interface has been improved. The concepts of design for manufacturability and design for assembly have been the subject of considerable research and development. Numerous specific models have been proposed and refined and will not be discussed here. This work is aimed at overcoming the difficulties precipitated by the traditional over-the-wall design philosophy. The magnitude of the problem with the over-the-wall approach to design and manufacturing can perhaps be measured in terms of the number of times the part drawing is thrown back and forth over the wall. As the initial design concept progresses from prototype testing and development to final design detailing and ultimately to the initiation of production and beyond, the number of design interactions has in the past been too high. In Ref 9, typical Japanese and U.S. companies are contrasted in this regard with an illustration similar to Fig. 6.

Fig. 6 Comparison of Japanese and American product design life cycles

Taguchi methods have strong engineering orientation and focus primary attention on the engineering design process, particularly the projection of a three-stage design process model of system design, parameter design, and tolerance design. Figure 7 illustrates the Taguchi design process model (Ref 4, 5).

Fig. 7 Block diagram illustrating Taguchi's three-stage design process model

Systems Design Stage. In the initial stage, systems design (the available science, technology, and experience bases)

is used to develop and select the basic design alternative to meet customer needs. A variety of techniques may be useful in specifically mapping the relationship between customer needs and the selection of design configuration and parameters that will effectively meet those needs. The methods of quality function deployment may be most useful at this stage of the design process. At the parameter design stage, interest focuses on the selection of the specific nominal values for the important design parameters. The overarching selection criterion is to identify those nominal values that minimize the transmission of functional variation to the output performance as a result of the presence of noise factors. It is at the parameter design stage that Taguchi strongly advocates the use of design of experiments methods. The tolerance design stage of the design process concentrates on the selection of allowable tolerances for the

important design parameters. The loss function concept of quality is used to provide a basis for striking the proper economic balance in the selection of design. Product/Process Performance Factors. Taguchi suggests an engineering interpretation of the varying roles that

the important system factors/variables play in influencing product/process performance. Taguchi emphasizes the importance of evaluating quality performance as part of the design process. Taguchi classifies the factors that can influence product/process performance into four categories: • •

• •

Signal factors: Factors that can be adjusted by the user to attain the target performance. Steering angle, for example, is a signal factor for the steering mechanism of an automobile Control factors: Product/process design parameters whose values are to be determined during the design process. One design activity is to select the optimum levels of the control factors according to an appropriate design criterion Noise factors: Factors that are either inherently uncontrollable or impractical to control because of technological/economic reasons. Taguchi further classifies these into outer and inner noises Scaling/leveling factors: Special cases of control factors. They are factors that can be easily adjusted to achieve a desired functional relationship between a signal factor and the output response. For example, the gearing ratio in a steering mechanism can be easily adjusted to achieve a desired relationship between the turning radius and the steering angle

Signal factors are those variables adjusted to attain the target/nominal performance. The control factors are those variables under the control of the designer. Selection of the nominal values for the control factors is the primary role of parameter design. Noise factors describe those variables that are difficult or impossible to control but whose variation is analyzed to understand the way it is transmitted to the output quality. To minimize the effects of outer and inner noises on product quality, certain countermeasures can be taken. The most important of these countermeasures is that of design. A product is said to be robust if its design is such that its performance is minimally influenced by uncontrollable noise factors. An automobile design can be said to be robust if, for example, the fuel economy remains fairly constant over a wide range of speeds, road conditions, and wind velocities. This is shown graphically in Fig. 8.

Fig. 8 Effect of three noise factors on automotive fuel efficiency (mpg) to demonstrate a robust product (Car A). Car A offers peak performance over a wide range of conditions for all three factors, while Car B offers peak

performance over a limited range of conditions for each individual factor.

References cited in this section

1. W.W. Scherkenbach, The Deming Route to Quality and Productivity: Road Maps and Roadblocks, Mercury Press/Fairchild Publications, 1987 2. J.M. Juran, Juran on Planning for Quality, The Macmillan Free Press, 1986 3. A.V. Feigenbaum, Total Quality Control, McGraw-Hill, 1983 4. G. Taguchi and Y. Wu, Introduction to Off-Line Quality Control, Central Japan Quality Control Association, 1979 5. G. Taguchi, On-Line Quality Control During Production, Japanese Standards Association, 1981 6. R.A. Fisher, Design of Experiments, 8th ed., Hafner Publishing, 1966 7. G.E.P. Box and J.S. Hunter, The 2k-p Fractional Factorial Designs, Part I and Part II, Technometrics, Vol 3, 1961 8. G.E.P. Box, W.G. Hunter, and J.S. Hunter, Statistics for Experimenters, John Wiley & Sons, 1978 9. L.P. Sullivan, Reducing Variability: A New Approach to Quality, Qual. Prog., Vol 17 (No. 7), 1985, p 15-21 Statistical Quality Design and Control Richard E. DeVor, University of Illinois, Urbana-Champaign; Tsong-how Chang, University of Wisconsin, Milwaukee

Sources of Variation and Their Countermeasures To reduce functional variation, that is, to increase the consistency of product/process performance, it is essential to identify the basic sources of functional variation so that appropriate countermeasures can be formulated and implemented. Taguchi suggests that variations in product or process function (also known as functional variations) arise from three basic sources (Ref 4): •

• •

Outer noise: External sources or factors that are operating in the environment in which the product is functioning and whose variation is transmitted through the design to the output performance of the product. Examples of outer noise factors are temperature, humidity, contaminants, voltage fluctuations, vibrations from nearby equipment, and variations in human performance Inner noise: Internal change in product characteristics such as drift from the nominal over time due to deterioration. Inner noise can be precipitated by such factors as mechanical wear and aging Variational noise: Variation in the product parameters from one unit to another as a result of the manufacturing process. For example, the design nominal for a resistor may be 200 Ω; however, one manufactured resistor may have a resistance of 202 Ω, while another may have a resistance of 197 Ω

The significance of the recognition of the above sources of variation becomes evident as one begins to think in terms of the fundamental countermeasures one might invoke to mitigate the forces of these sources of variation. It quickly becomes clear that the forces of outer noise and inner noise can only be effectively dealt with upstream by the engineering design process. Variational noise is a matter of manufacturing imperfection and can therefore be dealt with, in part, at the process with such techniques as statistical process control. However, mitigating the forces of variational noise should also be considered to be a product and process design issue. In fact, it is likely that variational noise can be dealt with in a more significant and fundamental way if it is thought of as a process and product design problem. Certainly the concept of design for manufacturability and the current emphasis on the simultaneous engineering of products and processes have bearing on this issue.

Figure 9 illustrates how the sources of variation defined above are transmitted to the quality response of the product. At the product design stage, the nominal values of the critical design parameters are selected to produce a prespecified level of performance in the product in terms of one or more quality responses. Sources of variation active in the manufacturing processes for the product cause the design parameters to vary from those values intended by the design process, and this introduces variation into the quality response. Once the product is put into field use, outer noise sources of variation active in the environment in which the product is being asked to function are transmitted through the design and introduce further variation in the quality response. For example, variation in temperature may cause product performance to vary. As time goes by and the use of the product continues, the forces of inner noise are transmitted through the design and introduce still further variation into the quality response. For example, as a result of wear, a critical design parameter (size, resistance, viscosity, and so on) will actually change, thus causing a change in quality performance. From Fig. 9 it is clear that sources of variation enter the picture and affect the ultimate performance of the product during manufacturing and field use.

Fig. 9 Block diagram illustrating the impact of numerous sources of variation on the quality performance of a product from conception to eventual discard

Shewhart (Ref 10), Deming (Ref 11), and Juran (Ref 12) all clearly point out that quality and productivity problems at the process fall into two basic categories. Shewhart described the variation in the process as arising from either chance causes or assignable causes. Deming refers to systems faults and local faults, while Juran refers to chronic problems and sporadic problems. Each provides a lucid description of the overall situation. Assignable causes, local faults, or sporadic problems arise in a somewhat unpredictable fashion and can usually be dealt with effectively at the workstation by the immediate level of supervision or the operator. On the other hand, system faults/chance causes/chronic problems are ever present, influencing all of the production until the system is changed. It is the attack on the system problems that Deming stresses, because about 80 to 85% of all the problems encountered are of this nature, while only 15 to 20% are of a local nature. All processes are subject to local faults and system faults:

Local faults Special causes Sporadic problems

System faults Common causes Chronic problems

Broken tools

Poor supervision

Jammed machine

Poor training

Material contamination

Inappropriate methods

Machine setting drift

Poor workstation design

Correctable locally, at the machine, by the operator or first level of supervision

Requires a change in the system; only management can implement changes

In the case of common causes, it is clear by their very nature that management must take the responsibility for their removal. Only management can take the actions necessary to improve the training and supervision of the workers, only management can take the responsibility for the redesign of a poor workstation layout, and only management can establish new methods or procedures. Clearly, it is essential that these two types of problems be properly differentiated so that responsibility can be assigned for solving them. Process Behavior Over Time A succession of parts emanating from a process under statistical control will exhibit variability in their measurements because of a constant set of common causes. These variable measurements tend to collect into a predictable pattern of variation that can be easily described by a few simple statistical measures; namely, a mean, a standard deviation, and a frequency distribution. These measures stand as a model that predicts how the process will behave if subject only to a constant set of common causes. Figure 10 illustrates how such a statistical model may emerge in the manufacture of automobile head gaskets. The gasket thickness (Fig. 10a) is the quality characteristic of interest.

Fig. 10 Schematic and statistical analysis of an automobile head gasket to demonstrate variability of the manufacturing process. (a) Gasket thickness specification. (b) Gasket thickness measurements taken of samples produced at 1-h intervals. (c) Histogram of mean gasket thickness. (d) Plot showing standard deviation of gasket thickness. (e) Plot showing distribution of gasket thickness measurements

Measurements of thickness can be recorded for a succession of gaskets produced by the process (Fig. 10b), and these data can be summarized by statistical measures, such as an average (Fig. 10c), a standard deviation (Fig. 10d), and a distribution (Fig. 10e). If a model can be developed for the process measurements when subject only to a constant system of common causes, then when a major disturbance affects the process measurement it will be seen not to conform to this model. It will clearly stand out in the common cause variability pattern. Time Order of Production. Although the above statistical description and the interpretation of data are useful, they ignore a crucial characteristic of the data: the time order of production. To properly indicate statistical control or lack of it, the model must be considered as it evolves over time of production. Figure 11 shows the time order evolution of the process. In Fig. 11(a), the process mean and variation stay constant over time, and the process is said to be in statistical control. In Fig. 11(b), a process mean shift occurs at 11:30 a.m. The process is said to have gone out of control. A process can be subject to several types of disturbances (special causes) that can produce a variety of unstable behaviors with respect to either its mean level or the level of variability or both.

Fig. 11 Changing process behavior over a 5-h time interval. (a) Process in statistical control. (b) Stable variability but out-of-control process due to sustained shift in process mean at 11:30 a.m.

Process Control and Process Improvement Many years ago, Dr. Walter Shewhart showed how such data from a manufacturing process could be developed and interpreted through the use of very simple but profound statistical methods. In his benchmark book on the subject, Economic Control of Quality of Manufactured Product, Shewhart establishes from the very beginning the overarching philosophy that drives the control chart concept (Ref 10). The first paragraph of the Preface of this work clearly sets the foundation for what is known today as statistical control: "Broadly speaking, the object of industry is to set up economic ways and means of satisfying human wants and in so doing to reduce everything possible to routines requiring a minimum amount of human effort. Through the use of the scientific method, extended to take account of modern statistical concepts, it has been found possible to set up limits within which the results of routine efforts must lie if they are to be economical. Deviations in the results of a routine process outside such limits indicate that the routine has broken down and will no longer be economical until the cause of trouble is removed" (Ref 10). From the above statement, several factors become immediately obvious: • • •

•

•

The fundamental focus is on the process--"ways and means of satisfying human wants" The overarching objective is economic operation of the process--"reduce everything possible to routines requiring a minimum amount of human effort" During normal operation, process behavior falls within predictable limits of variation--"It has been found possible to set up limits within which the results of routine efforts must lie if they are to be economical" Deviations in performance outside these limits signal the presence of problems that are jeopardizing the economic success of the process--"Deviations in the results of a routine process outside such limits indicate that the routine has broken down and will no longer be economical" Improvement in quality and productivity requires that attention be directed at the process to find the root cause of the trouble and remove it--"the routine has broken down and will no longer be economical until the cause of trouble is removed"

In addition, there is no mention of the product and the conformance of the product to specifications in the above statement from Shewhart. The total emphasis is on the process and its economic operation. Bringing a process into a state of statistical control does not necessarily mean that a fundamental improvement has been achieved. Clearly, a bad situation has been rectified by bringing the process into control, and quality and productivity are enhanced. However, bringing a process into control simply means that the process is back to where it should have been to begin with. At this point, it is then possible to begin to assess the present ability of the process to realize the potential it was initially intended to have. It may be failing to realize this potential because the implementation of the process is flawed or because the design of the process itself is flawed. In either case, the root cause(s) of the chronic common cause problem must be identified and removed at the system level. This constitutes a breakthrough in performance; that is, an improvement in the process has taken place. The results of the essential steps leading to such a breakthrough are shown in Fig. 12.

Fig. 12 Impact of having process initially in a state of statistical control versus improvement resulting from a breakthrough in performance

Fundamental Purposes of Shewhart Control Charts There are two fundamental purposes that Shewhart control charts can serve: • •

As a tool to provide a sound economic basis for making a decision at the machine regarding whether to look for trouble, adjust the process, or leave the process alone As a means of assisting in the identification of both improvement opportunities through the detection of sporadic and chronic faults in the process and to help provide the basis to formulate corrective and improvement actions

Although these two purposes may appear to be essentially the same, they are not. The former is strictly an on-line activity, while the latter is an off-line activity. On line, the operator can only recognize and take action against the presence of special causes of variation. If the process, as monitored on line, is free of special causes, the operator is doing the best that he can. In short, if the process is in statistical control, the operator should do nothing at all to disturb the routine behavior of the process. On the other hand, the fact that the process is in statistical control in no way guarantees that it is a capable process (able to meet customer expectations as they may be communicated through a specification). However, the operator cannot be expected to be responsible for this, though he may have some valuable insight into the problem. A lack of process capability is basically a product/process design and planning problem. For the techniques of statistical process control (specifically Shewhart control charts) to be successfully employed as an off-line problem identification and problem-solving tool, it is essential to keep in mind that it is a three-step process, as follows: • • •

Use statistical signals to find improvement opportunities through the identification of process faults Use experience, technical expertise, and fault diagnosis methods to find the root cause of the fault that has been identified Develop an action plan to correct the fault in a manner that will enable any gains that are realized to be

held

This three-step process can be explained by using the classical feedback control system perspective, as shown in Fig. 13. There are five distinct stages in the generic control loop shown in Fig. 13. These five stages facilitate the three-step process in the following way: • • •

Step 1: Use of statistical signals (observation and evaluation) Step 2: Fault diagnosis (diagnosis) Step 3: Action plan (decision and implementation)

Fig. 13 Classical feedback control system view of statistical process control implementation

References cited in this section

4. G. Taguchi and Y. Wu, Introduction to Off-Line Quality Control, Central Japan Quality Control Association, 1979 10. W.A. Shewhart, The Economic Control of Quality of Manufactured Product, Van Nostrand, 1931 (ASQC, 1980) 11. W.E. Deming, Quality, Productivity, and Competitive Position, MIT Center for Advanced Engineering, 1982 12. J.M. Juran, Quality Problems, Remedies, and Nostrums, Ind. Qual. Con., Vol 22, 1966, p 647-653 Statistical Quality Design and Control Richard E. DeVor, University of Illinois, Urbana-Champaign; Tsong-how Chang, University of Wisconsin, Milwaukee

Construction and Interpretation of

and R Control Charts

This section focuses on the construction and interpretation of Shewhart control charts for variable data, in particular, and R control charts. All the necessary equations and procedures for the calculation and graphical representation of the basic elements of the charts are presented. In addition, this section presents some probabilistic rules to aid in the interpretation of these charts. Shewhart Control Chart Model

The control chart stands much as a jury in a court of law. The information from each sample is judged in order to determine whether or not it indicates the presence of a special cause disturbance. Unless the evidence is overwhelmingly in favor of the occurrence of a special cause disturbance, a verdict of not guilty is entered. In other words, there is no strong reason to believe that forces other than those of common causes are at work. Such an interpretation is, in fact, based on and consistent with the hypothesis-testing approach to making statistical inferences about the population/process based on information contained in the sample. If a process under surveillance by periodic sampling maintains a state of good statistical control, this means that its mean level and level of variability remain constant over time. If such is the case, the sample means would have a distribution that follows the normal distribution, and nearly all the sample means would fall within the band of plus or minus three standard deviations of , or ±3σ , where σ is the standard deviation of about the established process average. Because the chance of realizing a point outside the 3σ limits is so small if good statistical control is evident, the occurrence of such a sample result must be interpreted as signaling some special/sporadic departure from expected behavior; that is, something other than common cause variation must be present. Figure 14 illustrates the presence of such a special cause. The 3σ limits about the process average/centerline are referred to as the upper control limit (UCL) and the lower control limit (LCL).

Fig. 14 Use of probabilistic limits to identify unusual process behavior over a period of time

Suppose a sample mean, , falls above the upper control limit. Because the chance of this occurring is so small (about of 1%) if the process mean is truly at the centerline, it must be assumed that some special cause is present that shifted the process mean to a different level. Unfortunately, the fact that all points fall within the 3 limits does not mean that special causes are not present. Departures from expected process behavior may not always manifest themselves on the control chart immediately or for that matter at all. Although narrower limits (for example, 2σ limits) would allow easier and faster detection of special causes, such limits would also increase the chances of false alarms, that is, times when the process is actually in good statistical control but a sample mean, , falls outside the control limits. Appropriate placement of the upper and lower control limits is an economic issue. The intent would be to fix the limits in such a way as to balance the economic consequences of failing to detect a special cause when it does occur and wrongly identifying the presence of a special cause when it has really not occurred. Experience has led to the use of 3σ limits as a good balance of these risks. Generally, chaotic disturbances manifest themselves in two possible ways: • •

Shifts or changes in the mean level of the process Shifts or changes in the amount of process variability

Figure 14 represents only one side of the coin; thus, a fundamental rule has been violated--that an -chart should not be shown without the associated range chart. By monitoring the process through sampling, the sample mean, , can be used

to aid in determining when a change in the mean level has occurred. By determining the range, R, for each sample and tracking its behavior over time, a basis is provided for identifying the presence of changes in process variability. Establishing a control chart for ranges R comparable to the -chart requires more specific information about the distribution of the process. In particular, it will be assumed that the individual measurements x are normally distributed. Doing so establishes the relationship between the standard deviation of ranges R and the standard deviation of individual process observations or measurements. For simplicity, ±3 R limits will be placed about the average range to define the upper and lower control limits for R even though the frequency distribution of ranges is not a symmetric distribution. Obtaining σR will be discussed later in this article. Setting Up and R Control Charts. Once the statistical basis for Shewhart control charts has been established, it remains to define the necessary elements of the charts mathematically and to establish a standard graphical representation. This section deals with the construction of and R control charts. All the necessary equations and general procedures for the calculation and graphical representation of the basic elements of the charts are presented in this section.

The first step in setting up and R control charts is the selection of the samples. It is important that all samples be rational samples. Rational samples are groups of measurements whose variation is attributable only to one constant system of common cause. Sampling from different machines, sampling over extended periods of time, and sampling from product combined from several sources are all nonrational sampling methods and must be avoided. Rational samples will be discussed in more detail later in this article. As a rule of thumb, 25 to 50 samples should be selected to provide a solid basis for the initiation of the control charts. This will help to ensure more precise estimation of the process mean and standard deviation. The sample/subgroup size should be relatively small (between n = 3 and n = 6). With k rational samples of n each, the following steps can be used as a guide when constructing 3 limits: • • •

•

, R control charts with

Calculate the sample mean and sample range for each sample using = x/n and R = xmax - xmin Calculate the grand mean of the n sample means and the average range using = /k and = R/k Calculate the control limits for the R-chart. Although the true distribution of sample ranges is not normal and not symmetric, the symmetric limits are conventionally used for the R-charts. With assumed normal distribution for the individual measurements, the following formulas can be used for the calculation of the control limits: UCLR = D4 and LCLR = D3 Calculate the control limits for the -chart. Although the required standard deviation (the standard deviation of the sample mean, ) for setting the limits is x, this value is conveniently estimated by /(d2

), where d2 is a function of n. For 3

3/(d2 LCLx =

limits, one uses a factor called A2, which is equal to

) and can be found in Table 1. Thus, the control limits are calculated by: UCLx = - A2

The appropriate values for d2, A2, D3, and D4 are obtained from Table 1. Table 1 Factors for Sample size, n

and R control chart limits

Factors for control limits

-chart A2

R-chart

D3

D4

Factor for calculating σx from range, R d2

+ A2

and

2

1.880

0

3.267

1.128

3

1.023

0

2.573

1.693

4

0.729

0

2.282

2.059

5

0.577

0

2.114

2.326

6

0.483

0

2.004

2.534

7

0.419

0.076

1.924

2.704

8

0.373

0.136

1.864

2.847

9

0.337

0.184

1.816

2.970

10

0.308

0.223

1.777

3.078

11

0.285

0.256

1.744

3.173

12

0.266

0.283

1.717

3.258

13

0.249

0.307

1.693

3.336

14

0.235

0.328

1.672

3.407

15

3.472

0.223

0.347

1.653

16

0.212

0.363

1.637

3.532

17

0.203

0.378

1.622

3.588

18

0.194

0.391

1.608

3.640

19

0.187

0.403

1.597

3.699

20

0.180

0.415

1.585

3.735

Example 1: Construction of

and R Control Charts of an Engine Block Cylinder Bore.

The inside diameter of the cylinder bore in an engine block was measured after a boring operation. Measurements were made to 0.0001 in. Samples of size n = 5 were taken to obtain some data to initiate and R control charts for this process. The samples were taken roughly every

h. The sample measurements were all taken on the same number cylinder in the

block. The results of the first 20 samples are given in Table 2. The actual measurements are of the form 3.5205 in., 3.5202 in., and so on. Table 2 provides only the last three digits in the measurement. Table 2

and R control chart engine block cylinder boring process data of Example 1

Sample number, k

Individual measurements, x(a)(b)

(b)

R(b)

1

2

3

4

5

1

205

202

204

209

205

205

7

2

202

196

201

198

202

199.8

6

3

201

202

199

197

196

199

6

4

205

203

196

201

197

200.4

9

5

199

196

201

200

195

198.2

6

6

203

198

192

217

196

201.2

25

7

202

202

198

203

202

201.4

5

8

197

196

196

200

204

198.6

8

9

199

200

204

196

202

200.2

8

10

202

196

204

195

197

198.8

9

11

206

204

202

210

205

205.4

8

12

200

201

199

200

201

200.2

2

13

205

196

201

197

198

199.4

9

14

202

199

200

198

200

199.8

4

15

200

200

201

205

201

201.4

5

16

201

187

209

202

200

199.8

22

17

202

202

204

198

203

201.8

6

18

201

198

204

201

201

201

6

19

207

206

194

197

201

201

13

20

200

204

198

199

199

200

6

(a) Sample/subgroup size, n = 5.

(b) Although actual cylinder bore data ranged from 3.5187 to 3.5217 in., only the last three digits in the measurement for x,

, and R are

used.

Determination of Trial Control Limits. Based on these first 20 samples, each sample mean

as listed in the last two columns of Table 2. The grand average, For -chart:

, and average range,

, are

and R was calculated = 200.62 and = 8.5.

UCL = + A2 = 200.62 + (0.58) (8.5) = 205.55 LCL = - A2 = 200.62 - (0.58) (8.5) = 195.69 For R-chart:

UCL = LCL =

D4 = 8.5 (2.11) = 17.935 D3 = 8.5 (0) = 0.0

With the above calculations completed, the and R control charts can be constructed. It will be important to start with the R-chart and to get it under statistical control first. This is necessary because the limits of the -chart depend on the magnitude of the common cause variation of the process measured by . If initially some points on the -chart exceed the upper control limit (special causes present), the limits in the -chart will be inflated and will need to be recalculated after such special cause data are removed. Interpretation of Initial Charts. Figure 15 shows the

and R control charts based on the calculations given above. The R-chart is examined first, and it is found that two points exceed the 3 limit. This indicates that there are special causes (excess process variability) at least at these points. These points--samples 6 and 16--are then examined to identify the reasons for these special causes. The records show that at these points the regular operator was absent and a lessexperienced relief operator ran the production for a short time. As a result, samples 6 and 16, taken over the time the relief operator ran the process, exhibited greater variability, perhaps because of inexperience. Because special causes can be identified for samples 6 and 16, these sample values (both and R) are removed and new limits are calculated. The revised centerlines and limits for and R control charts based on the data after samples 6 and 16 have been deleted are shown in Fig. 16.

Fig. 15 Control charts for inside diameter measurements of engine block cylinder bores in Table 2. (a)

-chart.

(b) R-chart. Data are for k = 20, n = 5.

Fig. 16 Revised control charts for Table 2 data with sample 6 and sample 16 data eliminated (both have R values above UCL in Fig. 15) because they are sources of excess process variability. (a) -chart.(b) R-chart. Data now have k = 18, n = 5.

Interpretation of Revised Charts. Again, the R-chart is examined first. No points are now outside the control limits, and no other unusual patterns of variability appear to be present. The R-chart seems to be in good statistical control. However, examination of the -chart reveals that there are two points above the upper control limit. The investigation of these points reveals the fact that these two samples (1 and 11) occurred at 8:00 a.m. and 1:00 p.m., roughly corresponding to the start-up of the boring machine in the morning and directly after the lunch hour. It was found that the samples were taken from the first few parts made in each case. Once the machine warmed up (approximately 10 min), the problem disappeared. It was decided as a standard policy not to initiate production until 10 min after start-up of the machine.

With this special cause identified and the corresponding values removed, the new limits are calculated. The control charts in Fig. 17 now both show good statistical control. The limits can now be extended, and monitoring of the process can be continued.

Fig. 17 Second revision control charts for Table 2 data with two more sample deletions (samples 1 and 11, both of which exceed UCL in Fig. 16) resulting from workpieces produced prior to machine being properly warmed up. (a) -chart. (b) R-chart. Data now have k = 16, n = 5.

Importance of Using Both and R Control Charts. This example points to the importance of maintaining both and R control charts and the significance of first focusing attention on the R-chart and establishing its stability. Initially, no points fell outside the -chart control limits and one could be led to believe that this indicates that the process mean exhibits good statistical control. However, the fact that the R-chart was initially not in control caused the limits on the chart to be somewhat wider because of two inordinately large R values. Once these special causes of variability were removed, the limits on the -chart became narrower, and two values now fall outside these new limits. Special causes were present in the data, but initially were not recognizable because of the excess variability as seen in the R-chart.

This example also points strongly to the need to have 25 or more samples before initiating control charts. In this case, once special causes were removed, only 16 subgroups remained to construct the charts. This is simply not enough data. Importance of Rational Sampling Perhaps the most crucial issue to the successful use of the Shewhart control chart concept is the definition and collection of the samples or subgroups. This section will discuss the concept of rational sampling, sample size, sampling frequency, and sample collection methods and will review some classic misapplications of rational sampling. Also, a number of practical examples of subgroup definition and selection will be presented to aid the reader in understanding and implementing this central aspect of the control chart concept. Concept of Rational Sampling. Rational subgroups or samples are collections of individual measurements whose

variation is attributable only to one unique constant system of common causes. In the development and continuing use of control charts, subgroups or samples should be chosen in a way that provides the maximum opportunity for the measurements within each subgroup to be alike and the maximum chance for the subgroups to differ from one another if special causes arise between subgroups. Figure 18 illustrates the notion of a rational sample. Within the sample or subgroup, only common cause variation should be present. Special causes/sporadic problems should arise between the selection of one rational sample and another.

Fig. 18 Graphical depiction of a rational subgroup illustrating effect of special causes on mean. (a) Unshifted. (b) Shifted

Sample Size and Sampling Frequency Considerations. The size of the rational sample is governed by the

following considerations: •

•

•

Subgroups should be subject to common cause variation. The sample size should be small to minimize the chance of mixing data within one sample from a controlled process and one that is out of control. This generally means that consecutive sample selection should be used rather than distributing the sample selection over a period of time. There are, however, certain situations where distributed sampling may be preferred Subgroups should ensure the presence of a normal distribution for the sample means. In general, the larger the sample size, the better the distribution is represented by the normal curve. In practice, sample sizes of four or more ensure a good approximation to normality Subgroups should ensure good sensitivity to the detection of assignable causes. The larger the sample size, the more likely that a shift of a given magnitude will be detected

When the above factors are taken into consideration, a sample/subgroup size of four to six is likely to emerge. Five is the most commonly used number because of the relative ease of further computation. Sampling Frequency. The question of how frequently samples should be collected is one that requires careful thought.

In many applications of and R control charts, samples are selected too infrequently to be of much use in identifying and solving problems. Some considerations in sample frequency determination are the following: •

If the process under study has not been charted before and appears to exhibit somewhat erratic behavior, samples should be taken quite frequently to increase the opportunity to quickly identify improvement opportunities. As the process exhibits less and less erratic behavior, the sample interval can be

•

•

lengthened It is important to identify and consider the frequency with which occurrences are taking place in the process. This might include, for example, ambient condition fluctuations, raw material changes, and process adjustments such as tool changes or wheel dressings. If the opportunity for special causes to occur over a 15-min period is good, sampling twice a shift is likely to be of little value Although it is dangerous to overemphasize the cost of sampling in the short term, clearly it cannot be neglected

Common Pitfalls in Subgroup Selection. In many situations, it is inviting to combine the output of several parallel

and assumed-to-be-identical machines into a single sample to be used in maintaining a single control chart for the process. Two variations of this approach can be particularly troublesome: stratification and mixing. Stratification of the Sample. Here each machine contributes equally to the composition of the sample. For example,

one measurement each from four parallel machines yields a sample/subgroup of n = 4, as seen in Fig. 19. In this case, there will be a tremendous opportunity for special causes (true differences among the machine) to occur within subgroups.

Fig. 19 Block diagram depicting a stratified sample selection

When serious problems do arise, for example, for one or more of the machines, they will be very difficult to detect because of the use of stratified samples. This problem can be detected, however, because of the unusual nature of the chart pattern (recall the previous pattern analysis) and can be rectified provided the concepts of rational sampling are understood. The R-charts developed from such data will usually show very good control. The corresponding control chart will show very wide limits relative to the plotted values, and their control will therefore appear almost too good. The wide limits result from the fact that the variability within subgroups is likely to be subject to more than merely common causes (Fig. 20).

Fig. 20 Typical control charts obtained for a stratified sample selection. (a)

-chart. (b) R-chart

Mixing Production From Several Machines. Often it is inviting to combine the output of several parallel machines/lines into a single stream of well-mixed product that is then sampled for the purposes of maintaining control charts. This is illustrated in Fig. 21.

Fig. 21 Block diagram of sampling from a mixture

If every sample has exactly one data point from each machine, the result would be the same as that of stratified sampling. If the sample size is smaller than the number of machines with different means or if most samples do not include data

from all machines, the within-sample variability will be too low, and the between-sample differences in the means tend to be large. Thus, the -chart would give an appearance that the values are too far away from the centerline. Statistical Quality Design and Control Richard E. DeVor, University of Illinois, Urbana-Champaign; Tsong-how Chang, University of Wisconsin, Milwaukee

Zone Rules for Control Chart Analysis Special causes often produce unnatural patterns that are not as clear cut as points beyond the control limits or obvious regular patterns. Therefore, a more rigorous pattern analysis should be conducted. Several useful tests for the presence of unnatural patterns (special causes) can be performed by dividing the distance between the upper and lower control limits into zones defined by , 2 , and 3 boundaries, as shown in Fig. 22. Such zones are useful because the statistical distribution of follows a very predictable pattern--the normal distribution; therefore, certain proportions of the points are expected to fall within the ± boundary, between and 2 , and so on. The following sections discuss eight tests that can be applied to the interpretation of and R control charts. Not all of these tests follow/use the zones just described, but it is useful to discuss all of these rules/tests together. These tests provide the basis for the statistical signals that indicate that the process has undergone a change in its mean level, variability level, or both. Some of the tests are based specifically on the zones defined in Fig. 22 and apply only to the interpretation of the -chart patterns. Some of the tests apply to both charts. Unless specifically identified to the contrary, the tests/rules apply to the consideration of data to one side of the centerline only. Fig. 22 Control chart zones to aid chart interpretation

When a sequence of points on the chart violates one of the rules, the last point in the sequence is circled. This signifies that the evidence is now sufficient to suggest that a special cause has occurred. The issue of when that special cause actually occurred is another matter. A logical estimation of the time of occurrence may be the beginning of the sequence in question. This is the interpretation that will be used here. It should be noted that some judgment and latitude should be given. Figure 23 illustrates the following patterns: • • • • • • •

•

Test 1(extreme points): The existence of a single point beyond zone A signals the presence of an out-ofcontrol condition (Fig. 23a) Test 2 (2 out of 3 points in zone A or beyond): The existence of 2 out of any 3 successive points in zone A or beyond signals the presence of an out-of-control condition (Fig. 23b) Test 3 (4 out of 5 points in zone B or beyond): A situation in which there are 4 out of 5 successive points in zone B or beyond signals the presence of an out-of-control condition (Fig. 23c) Test 4 (runs above or below the centerline): Long runs (7 or more successive points) either strictly above or strictly below the centerline; this rule applies to both the and R control charts (Fig. 23d) Test 5 (trend identification): When 6 successive points on either the or the R control chart show a continuing increase or decrease, a systematic trend in the process is signaled (Fig. 23e) Test 6 (trend identification): When 14 successive points oscillate up and down on either the or R control chart, a systematic trend in the process is signaled (Fig. 23f) Test 7 (avoidance of zone C test): When 8 successive points, occurring on either side of the centerline, avoid zone C, an out-of-control condition is signaled. This could also be the pattern due to mixed sampling (discussed earlier), or it could also be signaling the presence of an over-control situation at the process (Fig. 23g) Test 8 (run in zone C test): When 15 successive points on the -chart fall in zone C only, to either side of the centerline, an out-of-control condition is signaled; such a condition can arise from stratified

sampling or from a change (decrease) in process variability (Fig. 23h)

The above tests are to be applied jointly in interpreting the charts. Several rules may be simultaneously broken for a given data point, and that point may therefore be circled more than once, as shown in Fig. 24

Fig. 23 Pattern analysis of violates a specific rule.

-charts. Circled points indicate last point in a sequence of points on a chart that

Fig. 24 Example of simultaneous application of more than one test for out-of-control conditions. Point A is a violation of tests 3 and 4; point B is a violation of tests 2, 3, and 4; and point C is a violation of tests 1 and 3. See text for discussion.

In Fig. 24, point A is circled twice because it is the end point of a run of 7 successive points above the centerline and the end point of 4 of 5 successive points in zone B or beyond. In the second grouping in Fig. 24, point B is circled three times because it is the end point of: • • •

A run of 7 successive points below the centerline 2 of 3 successive points in zone A or beyond 4 of 5 successive points in zone B or beyond

Point C in Fig. 24 is circled twice because it is an extreme point and the end point of a group of five successive points, four of which are in zone B or beyond. Two other points (D, E) in these groupings are circled only once because they violate only one rule. Statistical Quality Design and Control Richard E. DeVor, University of Illinois, Urbana-Champaign; Tsong-how Chang, University of Wisconsin, Milwaukee

Control Charts for Individual Measurements In certain situations, the notion of taking several measurements to be formed into a rational sample of size greater than one simply does not make sense, because only a single measurement is available or meaningful at each sampling. For example, process characteristics such as oven temperature, suspended air particulates, and machine downtime may vary during a short period at sampling. Even for those processes in which multiple measurements could be taken, they would not provide valid within-sample variation for control chart construction. This is so because the variation among several such measurements would be primarily attributed to variability in the measurement system. In such a case, special control charts can be used. Commonly used control charts for individual measurements include: • • •

x, Rm control charts Exponentially weighted moving average (EWMA) charts Cumulative sum charts (CuSum charts)

Both the EWMA (Ref 13, 14, 15, 16) and the CuSum (Ref 17, 18, 19, 20, 21) control charts can be used for charting sample means and other statistics in addition to their use for charting individual measurements.

x and Rm (Moving-Range) Control Charts. This is perhaps the simplest type of control chart that can be used for

the study of individual measurements. The construction of x and Rm control charts is similar to that of and R control charts except that x stands for the value of the individual measurements and Rm for the moving range, which is the range of a group of n consecutive individual measurements artificially combined to form a subgroup of size n (Fig. 25). The moving range is usually comprised of the largest difference in two or three successive individual measurements. The moving ranges are calculated as shown in Fig. 25 for the case of three consecutive measurements used to form the artificial samples of size n = 3.

Fig. 25 Examples of three successive measurements used to determine the moving range

Because the moving range, Rm, is calculated primarily for the purpose of estimating common cause variability of the process, the artificial samples that are formed from successive measurements must be of very small size to minimize the chance of mixing data from out-of-control conditions. It is noted that x and Rm are not independent of each other and that successive sample Rm values are overlapping. The following example illustrates the construction of x and Rm control charts, assuming that x follows at least approximately a normal distribution. Here, Rm is based on two consecutive measurements; that is, the artificial sample size is n = 2.

Example 2: x and Rm Control Chart Construction for the Batch Processing of White Millbase Component of a Topcoat. The operators of a paint plant were studying the batch processing of white millbase used in the manufacture of topcoats. The basic process begins by charging a sandgrinder premix tank with resin and pigment. The premix is agitated until a homogeneous slurry is obtained and then pumped through the sandgrinder. The grinder output is sampled to check for fineness and gloss. A batch may require adjustments by adding pigment or resin to achieve acceptable gloss. Through statistical modeling of the results of some ash tests, a quantitative method was developed for determining the amount of pigment or resin to be added when necessary, all based on the weight per unit volume (lb/gal.) of the batch. Therefore, it became important to monitor the weight per unit volume for each batch to achieve millbase uniformity. Table 3 lists weight per unit volume data for 27 consecutive batches. Table 3 x and Rm control chart data for the batch processing of white millbase topcoat component of Example 2 Rm (a)

Batch

x, lb/gal.

1

14.04

2

13.94

0.10 (14.04 - 13.94 = 0.10)

3

13.82

0.12 (13.94 - 13.82 = 0.12)

4

14.11

0.29 (14.11 - 13.82 = 0.29)

5

13.86

0.25

6

13.62

0.24

7

13.66

0.04

8

13.85

0.19

9

13.67

0.18

10

13.80

0.13

11

13.84

0.04

12

13.98

0.14

13

13.40

0.58

14

13.60

0.20

15

13.80

0.20

16

13.66

0.14

17

13.93

0.27

18

13.45

0.48

19

13.90

0.45

20

13.83

0.07

21

13.64

0.19

22

13.62

0.02

23

13.97

0.35

24

13.80

0.17

25

13.70

0.10

26

13.71

0.01

27

13.67

0.04

= 13.77

m

= 0.19

(a) Calculated, n = 2

In the calculation of averages in Table 3, is an average of all 27 individual measurements, while m is an average of 27 - 1 = 26 Rm values because there are only 26 moving ranges for n = 2. If the artificial samples were of size n = 3, there would be only 27 - 2 = 25 moving averages. Once the and m values are calculated, they are used as centerline values of x and Rm control charts, respectively. The calculation of upper and lower control limits for the Rm control chart is also the same as in , R control charts, using the artificial sample size n to determine D3 and D4 values. However, the upper and lower control limits for the x chart should always be based on a sample size of one, using m the same way as in control charts. These calculations are shown below for the example data. For the Rm-chart, for n = 2, D3 = 0, and D4 = 3.27:

CL UCL LCL

= m = 4.99/26 = 0.192 = D4 m = (3.27)(0.19) = 0.62 = D3 m = 0

For the x-chart, an estimate of the standard deviation of x is equal to Thus, 3 = (3/d2) m = (3/1.128) m = 2.66 m:

CLx UCLx = + 2.66 = LCLx = - 2.66 = 13.26

m/d2,

where d2 = 1.128 from Table 1 using n = 2.

= 13.77 m = 13.77 + (2.66)(0.19) 14.28 m = 13.77 - (2.66)(0.19)

When the individual x values follow a normal distribution, the patterns on the x-chart are analyzed in the same manner as the Shewhart -charts. A sample x is circled as a signal of out-of-control values if it falls outside a control limit, if it is the end point of a sequence that violates any of the zone rules, or if it simply indicates a nonrandom sequence. However, tests for unnatural patterns should be used with more caution on an x-chart than on an -chart because the individual chart is sensitive to the actual shape of the distribution of the individuals, which may depart considerably from a true normal distribution. Figures 26 and 27 show the x and Rm control charts for the example data.

Fig. 26 Rm control chart obtained for white millbase data in Table 3. Data are for k = 27, n = 2.

Fig. 27 x control chart obtained for white millbase data in Table 3. Data are for k = 27, n = 2.

References cited in this section

13. S.W. Roberts, Control Charts Based on Geometric Moving Averages, Technometrics, Vol 1, 1959, p 234250 14. A.L. Sweet, Control Charts Using Coupled Exponentially Weighted Moving Averages, Trans. IIE, Vol 18 (No. 1), 1986, p 26-33 15. A.W. Wortham and G.F. Heinrich, Control Charts Using Exponential Smoothing Techniques, Trans. ASQC, Vol 26, 1972, p 451-458 16. A.W. Wortham, The Use of Exponentially Smoothed Data in Continuous Process Control, Int. J. Prod. Res., Vol 10 (No. 4), 1972, p 393-400 17. A.F. Bissell, An Introduction to CuSum Charts, The Institute of Statisticians, 1984 18. "Guide To Data Analysis and Quality Control Using CuSum Techniques," BS5703 (4 parts), British Standards Institution, 1980-1982 19. J.M. Lucas, The Design and Use of V-Mask Control Scheme, J. Qual. Technol., Vol 8 (No. 1), 1976, p 1-12 20. J. Murdoch, Control Charts, Macmillan, 1979 21. J.S. Oakland, Statistical Process Control, William Heinemann, 1986

Statistical Quality Design and Control Richard E. DeVor, University of Illinois, Urbana-Champaign; Tsong-how Chang, University of Wisconsin, Milwaukee

Shewhart Control Charts for Attribute Data Many quality assessment criteria for manufactured goods are not of the variable measurement type. Rather, some quality characteristics are more logically defined in a presence-of or absence-of sense. Such situations might include surface flaws on a sheet metal panel; cracks in drawn wire; color inconsistencies on a painted surface; voids, flash, or spray on an injection-molded part; or wrinkles on a sheet of vinyl. Such nonconformities or defects are often observed visually or according to some sensory criteria and cause a part to be defined simply as a defective part. In these cases, quality assessment is referred to as being made by attributes. Many quality characteristics that could be made by measurements (variables) are often not done as such in the interest of economy. A go/no-go gage can be used to determine whether or not a variable characteristic falls within the part specification. Parts that fail such a test are simply labeled defective. Attribute measurements can be used to identify the presence of problems, which can then be attacked by the use of and R control charts. The following definitions are required in working with attribute data: • • • • •

Defect: A fault that causes an article or an item to fail to meet specification requirements. Each instance of the lack of conformity of an article to specification is a defect or nonconformity Defective: An item or article with one or more defects is a defective item Number of defects: In a sample of n items, c is the number of defects in the sample. An item may be subject to many different types of defects, each of which may occur several times Number of defectives: In a sample of n items, d is the number of defective items in the sample Fractional defective: The fractional defective, p, of a sample is the ratio of the number of defectives in a sample to the total number of items in the sample. Therefore, p = d/n

Operational Definitions The most difficult aspect of quality characterization by attributes is the precise determination of what constitutes the presence of a particular defect. This is so because many attribute defects are visual in nature and therefore require a certain degree of judgment and because of the failure to discard the product control mentality. For example, a scratch that is barely observable by the naked eye may not be considered a defect, but one that is readily seen is. Furthermore, human variation is generally considerably larger in attribute characterization (for example, three different caliper readings of a workpiece dimension by three inspectors and visual inspection of a part by these same individuals yield anywhere from zero to ten defects). It is therefore important that precise and quantitative operational definitions be laid down for all to observe uniformly when attribute quality characterization is being used. The length or depth of a scratch, the diameter of a surface blemish, or the length of a flow line on a molded part can be specified. The issue of the product control versus process control way of thinking about defects is a crucial one. From a product control point of view, scratches on an automobile grille should be counted as defects only if they appear on visual surfaces, which would directly influence part function. From a process control point of view, however, scratches on an automobile grille should be counted as defects regardless of where they appear because the mechanism creating these scratches does not differentiate between visual and concealed surfaces. By counting all scratches, the sensitivity of the statistical charting instrument used to identify the presence of defects and to lead to their diagnosis will be considerably increased. A major problem with the product control way of thinking about part inspection is that when attribute quality characterization is being used not all defects are observed and noted. The first occurrence of a defect that is detected immediately causes the part to be scrapped. Often, such data are recorded in scrap logs, which then present a biased view of what the problem may really be. One inspector may concentrate on scratch defects on a molded part and will therefore tend to see these first. Another may think splay is more critical, so his data tend to reflect this type of defect more frequently. The net result is that often such data may then mislead those who may be using it for process control purposes.

Figure 28 shows an example of the occurrence of multiple defects on a part. It is essential from a process control standpoint to carefully observe and note each occurrence of each type of defect. Figure 29 shows a typical sample result and the careful observation of each occurrence of each type of defect. In Fig. 29, the four basic measures used in attribute quality characterization are defined for the sample in question.

Fig. 28 Typical multiple defects present on an engine valve seat blank to illustrate defect identification in an attribute quality characterization situation

Fig. 29 Analysis of four basic measures of attribute quality characterization used to illustrate the typical defects present in the engine valve seat blank shown in Fig. 28. Out of ten samples tested, four had no defects, three had single defects, and three had multiple defects.

p-Chart: A Control Chart for Fraction Defective Consider an injection-molding machine producing a molded part at a steady pace. Suppose the measure of quality conformance of interest is the occurrence of flash and splay on the molded part. If a part has so much as one occurrence of either flash or splay, it is considered to be nonconforming, that is, a defective part. To establish the control chart, rational samples of size n = 50 parts are drawn from production periodically (perhaps, each shift), and the sampled parts are inspected and classified as either defective (from either or both possible defects) or nondefective. The number of defectives, d, is recorded for each sample. The process characteristic of interest is the true process fraction defective p'. Each sample result is converted to a fraction defective:

(Eq 1)

The data (fraction defective p) are plotted for at least 25 successive samples of size n = 50. The individual values for the sample fraction defective, p, vary considerably, and it is difficult to determine from the plot at this point if the variation about the average fraction defective, , is solely due to the forces of common causes or special causes. Control Limits for the p-Chart. It can be shown that for random sampling, under certain assumptions, the occurrence

of the number of defectives, d, in the sample of size n is explained probabilistically by the binominal distribution. Because the sample fraction defective, p, is simply the number of defectives, d, divided by the sample size, n, the occurrence of values for p also follows the binominal distribution. Given k rational samples of size n, the true fraction defective, p', can be estimated by:

(Eq 2)

or

(Eq 3) Equation 3 is more general because it is valid whether or not the sample size is the same for all samples. Equation 2 should be used only if the sample size, n, is the same for all k samples. Therefore, given , the control limits for the p-chart are then given by:

UCLp = + 3 LCPp = - 3

(Eq 4a) (Eq 4b)

Thus, only has to be calculated for at least 25 samples of size n to set up a p-chart. The binomial distribution is generally not symmetric in quality control applications and has a lower bound of p = 0. Sometimes the calculation for the lower control limit may yield a value of less than 0. In this case, a lower control limit of 0 is used. Example 3: A p-Chart Applied to Evaluation of a Carburetor Assembly (Ref 22). This example illustrates the construction of a p-chart. The data in Table 4 are inspection results on a type of carburetor at the end of assembly; all types of defects except leaks were noted, and n = 100 for all samples. Samples taken numbered k = 35. Table 4 p-chart data for carburetor assembly of Example 3 Sample, k(a)

d

p

1

4

0.04

2

5

0.05

3

1

0.01

4

0

0.00

5

3

0.03

6

2

0.02

7

1

0.01

8

6

0.06

9

0

0.00

10

6

0.06

11

2

0.02

12

0

0.00

13

2

0.02

14

3

0.03

15

4

0.04

16

1

0.01

17

3

0.03

18

2

0.02

19

4

0.04

20

2

0.02

21

1

0.01

22

2

0.02

23

0

0.00

24

2

0.02

25

3

0.03

26

4

0.04

27

1

0.01

28

0

0.00

29

0

0.00

30

0

0.00

31

0

0.00

32

1

0.01

33

2

0.02

34

3

0.03

35

3

0.03

(a) n = 100

Using this sample data to establish the p-chart:

Therefore:

UCLp = 0.02086 + 0.04287 = 0.06373 LCLp = 0.02086 - 0.04287 = -0.02201 That is:

LCLp = 0 The plot of the data on the corresponding p-chart is shown in Fig. 30. The process appears to be in statistical control, although eight points lie on the lower control limit. In this case, results of p = 0 that fall on the lower control limit should not be interpreted as signaling the presence of a special cause. For a sample size of n = 100 and a fraction defective p' = 0.02, the binomial distribution gives the probability of d = 0 defectives in a sample of 100 to be 0.133. Therefore, a sample with zero defectives would be expected about one out of seven times.

Fig. 30 p control chart obtained for the evaluation of the carburetor assembly data in Table 4. Data are for k = 35, n = 100.

In summary, the p-chart in this example seems to indicate good statistical control, having no extreme points (outside the control limits), no significant trends or cycles, and no runs of sizable length above or below the centerline. At least over this period of data collection, the process appears to be operating only under a common cause system of variation. However, Fig. 30 shows that the process is consistently operating at a 2% defective rate. Variable Sample Size Considerations for the p-Chart. It is often the case that the sample size may vary from

one time to another as data for the construction of a p-chart are obtained. This may be the case if the data have been collected for other reasons (for example, acceptance sampling) or if a sample constitutes a day's production (essentially

100% inspection) and production rates vary from day to day. Because the limits on a p-chart depend on the sample size n, some adjustments must be made to ensure that the chart is properly interpreted. There are several ways in which the variable sample size problem can be handled. Some of the more common approaches are the following: •

•

•

Compute separate limits for each individual subgroup. This approach certainly leads to a correct set of limits for each sample, but requires continual calculation of the control limits and a somewhat messylooking control chart Determine an average sample size, , and a single set of control limits based on . This method may be appropriate if the sample sizes do not vary greatly, perhaps no more than about 20%. However, if the actual n is less than , a point above the control limit based on may not be above its own true upper control limit. Conversely, if the actual n is greater than , a point may not show out of control when in reality it is A third procedure for varying sampling size is to express the fraction defective in standard deviation units, that is, plot (p - )/ p on a control chart where the centerline is zero and the control limits are set at ±3.0. This stabilizes the plotted value even though n may be varying. Note that (p - )/ p is a familiar form; recall the standard normal (Z) distribution. For this method, the continued calculation of the stabilized variable is somewhat tedious, but the chart has a clean appearance, with constant limits of always ±3.0 and constant centerline at 0.0

c-Chart: A Control Chart for Number of Defects The p-chart deals with the notion of a defective part or item where defective means that the part has at least one nonconformity or disqualifying defect. It must be recognized, however, that the incidence of any one of several possible nonconformities would qualify a part for defective status. A part with ten defects, any one of which makes it a defective, is on equal footing with a part with only one defect in terms of being a defective. Often it is of interest to note every occurrence of every type of defect on a part and to chart the number of defects per sample. A sample may only be one part, particularly if interest is focusing on final inspection of an assembled product, such as an automobile, a lift truck, or perhaps a washing machine. Inspection may focus on one type of defect (such as nonconforming rivets on an aircraft wing) or multiple defects (such as flash, splay, voids, and knit lines on an injectionmolded truck grille). Considering an assembled product such as a lift truck, the opportunity for the occurrence of a defect is quite large, perhaps to be considered infinite. However, the probability occurrence of a defect in any one spot arbitrarily chosen is probably very, very small. In this case, the probability law that governs the incidence of defects is known as the Poisson law or Poisson probability distribution, where c is the number of defects per sample. It is important that the opportunity space for defects to occur be constant from sample to sample. The Poisson distribution defines the probability of observing c defects in a sample where c' is the average rate of occurrence of defects per sample. Construction of c-Charts From Sample Data. The number of defects, c, arises probabilistically according to the Poisson distribution. One important property of the Poisson distribution is that the mean and variance are the same value. Then given c', the true average number of defects per sample, the 3 limits for the c-chart are given by:

CLc = c' ± 3

(Eq 5)

Note that the standard deviation of the observed quantity c is the square root of c'. The Poisson distribution is a very simple probability model, being completely described by a single parameter c'. When c' is unknown, it must be estimated from the data. For a collection of k samples, each with an observed number of defects c, the estimate of c' is:

(Eq 6)

Therefore, trial control limits for the c-chart can be established, with possible truncation of the lower control limit at zero, from:

UCLc = + 3 LCLc = - 3 Example 4: c-Chart Construction for Continuous Testing of Plastic-Insulated Wire at a Specified Test Voltage. Table 5 lists the results of continuous testing of a certain type of plastic-covered wire at a specified test voltage. This test causes breakdowns at weak spots in the insulation, which are cut out before shipment. Table 5 c-chart data for plastic-insulated wire of Example 4 Sample number, k

Number of breakdowns

1

1

2

1

3

3

4

7

5

8

6

1

7

2

8

6

9

1

10

1

11

10

12

5

13

0

14

19

15

16

16

20

17

1

18

6

19

12

20

4

21

5

22

1

23

8

24

7

25

9

26

2

27

3

28

14

29

6

30

8

The original data consisted of the number of breakdowns in successive lengths of 1000 ft each. There may be 0, 1, 2, 3, . . ., breakdowns per length, depending on the number of weak spots in the insulation. However, so few defects were obtained during a short period of production by using the 1000 ft length as a unit and the expectancy in terms of the number of breakdowns per length was so small that a longer length of 10,000 ft was used for the unit size for the corresponding c-chart. In general, it is desirable to select the sample size for the c-chart application such that on average ( ) at least one or two defects are occurring per sample. In most applications, the centerline of the c-chart is based on the estimate of the average number of defects per sample. This estimate can be calculated by:

The resulting c-chart in Fig. 31 shows the presence of special causes of variation.

Fig. 31 c control chart obtained for the evaluation of the plastic-insulated wire data (k = 30) in Table 5

u-Chart: A Control Chart for the Number of Defects per Unit Although in c-chart applications it is common for a sample to consist of only a single unit or item, the sample or subgroup can be comprised of several units. Further, from subgroup to subgroup, the number of units per subgroup may vary, particularly if a subgroup is an amount of production for the shift or day, for example. In such cases, the opportunity space for the occurrence of defects per subgroup changes from subgroup to subgroup, violating the equal opportunity space assumption on which the standard c-chart is based. Therefore, it is necessary to create some standardized statistic, and such a statistic may be the average number of defects per unit or item where n is the number of items per subgroup. The symbol u is often used to denote average number of defects per unit, that is:

(Eq 7) where c is the total number of defects per subgroup of n units. For k such subgroups gathered, the centerline on the uchart is:

(Eq 8)

The trial control limits for the u-chart are then given by:

CLu =

±3

(Eq 9)

Example 5: Use of the u-Chart to Evaluate Leather Handbag Lots. Table 6 lists inspection results in terms of defects observed in the inspection of 25 consecutive lots of leather handbags. Because the number of handbags in each lot was different, a constant sample size of n = 10 was used. All defects were counted even though two or more defects of the same or different type occurred on the same bag. The u-chart data are as follows (Fig. 32):

Table 6 u-chart data for leather handbag lot production of Example 5 Sample number, k(a)

Total number of defects

Defects per unit

1

17

1.7

2

14

1.4

3

6

0.6

4

23

2.3

5

5

0.5

6

7

0.7

7

10

1.0

8

19

1.9

9

29

2.9

10

18

1.8

11

18

1.8

12

5

0.5

13

8

0.8

14

11

1.1

15

18

1.8

16

13

1.3

17

22

2.2

18

6

0.6

19

23

2.3

20

22

2.2

21

9

0.9

22

15

1.5

23

20

2.0

23

20

2.0

25

24

2.4

Total

382

38.2

(a) n = 10

Fig. 32 u control chart obtained for the evaluation of leather handbag lot data in Table 6. Data are for k = 25, n = 10. Datum for sample 9 is an extreme point because it exceeds value of UCL.

Reference cited in this section

22. I. Burr, Statistical Quality Control Methods, Marcel Dekker, 1976 Statistical Quality Design and Control Richard E. DeVor, University of Illinois, Urbana-Champaign; Tsong-how Chang, University of Wisconsin, Milwaukee

Process Capability Assessment This section presents both traditional and more modern views of process capability and how it is assessed. The presentation stresses the relationship between the control/stability of a process and its capability. The clear distinction between the engineering specification and the statistical control limits in terms of their use and interpretation is emphasized. Using the traditional conformance to the specifications view of process capability, this section illustrates the consequences of a lack of statistical control in terms of the manner and extent to which a process produces parts that meet design intent. This section also presents the loss function approach to the articulation of process capability. Process Capability Versus Process Control There are two separate but vitally important issues that must be addressed when considering the statistical representation of process data. These are: • •

The ability of the process to produce parts that conform to specifications The ability of the process to maintain a state of good statistical control

These two process characteristics are linked in the sense that it will be difficult to assess process capability with respect to conformance to specifications without being reasonably assured of having good statistical control. Although control certainly does not imply conformance, it is a necessary prerequisite to the proper assessment of conformance. In a statistical sense, conformance to specifications involves the process as a whole; therefore, attention will be focused on the distribution of individual measurements. In dealing with statistical control, summary statistics from each sample, mainly and R, are used; as a result, this involves the distribution of these statistics, not individual measurements. Because of the above distinction between populations and samples, it is crucial not to compare or confuse part specifications and control limits. In fact, tolerance/specification limits should never be placed on a control chart. This is so because the control chart is based on the variation in the sample means , while it is the individual measurements in the sample that should be compared to specifications. Placing specifications on the control chart may lead to the mistaken impression that good conformance exists when, in fact, it does not. This is illustrated in Example 6. A process may produce a large number of pieces that do not meet the specified production standards, even though the process itself is in a state of good statistical control (that is, all the points on and R control charts are within the 3 limits and vary in a random manner). This may be because the process is not centered properly; in other words, the actual mean value of the parts being produced may be significantly different from the specified nominal value of the part. If this is the case, an adjustment of the machine to move the mean closer to the nominal value may solve the problem. Another possible reason for lack of conformance to specifications is that a statistically stable process may be producing parts with an unacceptably high level of common cause variation. In summary, if a process is in statistical control but not capable of meeting the tolerances, the problem may be one of the following: • • •

The process is off-center from the nominal The process variability is too large relative to the tolerances The process is off-center and with large variation

Example 6: Statistical Assessment of Process Capability for a Workpiece. For a certain part, a dimension of interest was specified by the engineering department as 0.140 ± 0.003 in. Many parts were being rejected on 100% inspection using a go/no-go gage because they failed to meet these tolerances. It was decided to study the capability of the process using and R control charts. These data were taken from the same machine and operator and at the rate of about one sample per hour. Both the and R control charts seem to indicate good statistical control with no points exceeding the 3 limits, a reasonably normal distribution of points between the limits, and no discernible trends, cycles, and so on. Therefore, the calculated sample mean = 0.1406 in. and the sample standard deviation σx = 0.0037 in. are good estimates. The process can then be evaluated with respect to its conformance to specifications. To obtain a clear picture of the statistical nature of the data, a frequency histogram was plotted that resembled a normal distribution, but the mean appears to be slightly higher than the nominal value of 0.140 set by the engineering department. Figure 33 shows this population distribution curve centered at with a spread of x = 0.0037 in. The specifications are also shown on this plot.

Fig. 33 Normal distribution model for process capability for the data of Example 6. LSLx = 0.137, USLx = 0.143, = 0.1406, and σx = 0.0037.

The shaded areas in Fig. 33 represent the probability of obtaining a part that does not meet specifications. To compute the probability of a part failing below the lower specification, the standard normal distribution, Z, is calculated, and a normal curve table is used:

Z = (x -

)/

where x is the value of either the lower or upper specification, estimate of the process standard deviation.

x

is the estimate for the population mean, and σx is the

To find the probability of a point below the lower specification limit, LSLx = 0.137, with

= 0.1406, and

x

= 0.0037:

Looking this value up in the normal table produces Prob (0.137 or less) = 0.1660. This means that there is a 16.6% chance of an individual part falling below the specified tolerance. To find the probability of an individual part falling above the upper specification limit, USLx = 0.143, 0.0037, and:

= 0.1406,

x

=

The probability of a part being above the upper specification limit is equal to 0.2578, based on Z = 0.65. The process that does not meet the specifications is therefore 16.6% + 25.78% = 42.37%. It might be asked whether centering the process at the nominal value of 0.140 would help. To check, a normal curve is constructed, centered at 0.140. The probability of a point below the lower specification is found by computing Z using the nominal value as the population mean:

Looking up the area for this Z value in the normal curve table produces a value of 0.209, which is the probability of getting a value below the lower specification limit. The probability of getting a value above the upper specification is found by:

which from the normal curve table gives an area = 0.209. The total probability of a part not meeting specification is the sum of these, or 0.209 + 0.209 = 0.41, or 41%. Therefore, recentering the process will not be of much help. The process is in control; no special causes of variability were indicated. However, about 42% of the parts were outside the tolerances. Possible remedies include the following: • • • •

Continue to sort by 100% inspection Widen the tolerances, for example, 0.140 ± 0.006 Use a more precise process; reduce process variation Use statistical methods to identify variation reduction opportunities for the existing process

Too often, the strategy used (or at least urged) is the second remedy listed above. Clearly, a stronger consideration should be given to the final remedy listed above. Comparison of Tolerances and Control Limits It is important to clearly differentiate between specification limits and control limits. The specification limits or tolerances of a part are: • • • •

Characteristic of the part/item in question Based on functional considerations Related to/compared with an individual part measurement Used to establish the conformability of a part

The control limits on a control chart are: • • • •

Characteristic of the process in question Based on process variability Dependent on sampling parameters, namely, sample size Used to identify presence/absence of special cause variation in the process

Control limits and tolerances must never be compared numerically and should not appear together on the same graph. Tolerances are limits on individual measurements and as such can be compared against the process as a whole as represented by many individual measurements collected in the form of a statistical distribution, as was done in Example 6 to assess overall process capability. Process Capability Indices It is common to measure process capability in the units of process standard deviations. In particular, it is common to look at the relationship between the process standard deviation and the distance between the upper and lower specification:

(Eq 10)

The minimum acceptable value for Cp is considered to be 6. Recently, many companies have begun to use a capability index referred to as Cpk. For bilateral specifications, Cpk is defined in the following manner. First, the relationship between the process mean and the specification limits in the units of standard deviations is determined:

(Eq 11a) (Eq 11b)

Then the minimum of these two values is selected:

Z min = min[ZUSL, - ZLSL]

(Eq 12)

The Cpk index is then defined by dividing this minimum value by 3:

(Eq 13)

Commonly, Cpk must be

1.00.

Statistical Process Control and the Statistical Tolerance Model The issue of part tolerancing and, in particular, the statistical assignment and assessment of tolerances are excellent examples of the need for design and manufacturing to understand what each other is doing and why. The best intentions of the design process can go unmet if the manufacturing process is not operated in a manner totally consistent with design intent. To more clearly appreciate the marriage of thinking that must exist between the design and manufacturing worlds, some of the basic assumptions of the tolerancing activity and their relationship to the manufacturing process will be examined. The following sections clearly point to the importance of statistical process control relative to the issue of process capability. The key concepts in statistical tolerancing are: • • •

The use of a statistical distribution to represent the design characteristic and therefore the process output for the product/part in question relative to the design specifications The notion of random assembly, that is, random part selection from these part process distributions when more than one part is being considered in an assembly The additive law of variances as a means to determine the relationship between the variability in individual parts and that for the assembly

To assume that the parts can be represented by a statistical distribution of measurements (and for the assumption to hold in reality), the part processes must be in a state of statistical control. The following example illustrates the importance of statistical process control in achieving design intent in a tolerancing problem. Example 7: Statistical Tolerance Model for Optimum Fit of a Pin Assembly in a Hole Machined in a Plate.

Figure 34 shows two simple parts: a plate with a hole and a pin that will ultimately be assembled to a third part but must pass through the hole in the plate. For the assembly, it is desired for function that the clearance between the plate hole and the pin be at least 0.015 in. but no more than 0.055 in.

Fig. 34 Machined components statistically analyzed in Example 7. (a) Plate with hole. (b) Pin assembly. Dimensions given in inches

To achieve the design requirement stated above, the nominal values and tolerances for the plate hole and pin were statistically derived and are shown in Fig. 34. To arrive at these tolerances, it was assumed that: • • •

The parts would be manufactured by processes that behave according to the normal distribution The process capabilities would be at least 6σ, the processes would be centered at the nominal values given in Fig. 34, and the processes would be maintained in a state of statistical control Random assembly would prevail

If these assumptions are met, the processes for the two parts, and therefore the clearance associated with assembled parts, would be as shown in Fig. 35, and the design intent would be met.

Fig. 35 Statistical basis for satisfying design intent for the hole/pin assembly clearance in Fig. 34. (a) Distribution of hole. (b) Distribution of pin. (c) Distribution of clearance

Suppose that despite the assumptions made and the tolerances derived, the processes manufacturing the pin and plate hole were not maintained in good statistical control. As a result, the parts actually more nearly follow a uniform/rectangular distribution within the specifications, as shown in Fig. 36. Such could have arisen as a result of sorting or rework of a more variable process(es), in which case the results are doubly distressing, that is, poorly fitting assemblies and increased cost to the system.

Fig. 36 Clearance implications of poor process control of plate hole and pin dimensions for components of Fig. 34. (a) Distribution of hole. (b) Distribution of pin. (c) Distribution of clearance

Figure 36 shows the distribution of the clearance if the hole and pin dimensions follow the uniform distribution within the specifications. The additive law of variances has been used to derive the variation in the clearance distributions but assuming the uniform distribution for the individual part processes. Some assemblies may not go together at all, some will fit quite tightly and may later bind if foreign matter gets into the gap, and others will fit together with a much larger clearance than desired. The problem here is not a design problem. The plate hole and pin tolerances have been derived using sound statistical methods. However, if the processes are not in good statistical control and therefore not capable of meeting the assumptions made during design, poor-quality assemblies will follow. It should be noted that the altogether too common process appearance of a uniform distribution of measurements within the specifications can arise in several different ways: • • •

From processes that have good potential with regard to variation, but are not kept in good statistical control From unstable and/or large variation processes that require sorting/rework to meet the specifications From processes that are intentionally allowed to vary over the full range of the specifications to take advantage locally of wide specifications relative to the process variation

In all of the three cases mentioned above, additional costs will be incurred and product quality will be eroded. Clearly, statistical process control is crucial to the tolerancing issue in engineering design. Taguchi's loss function model, which is an essential element in tolerance design, assumes similarly that quality characteristics can be represented by a statistical distribution of measurements. Again, for this assumption to be met at the process and therefore in the ultimate product in the field, the manufacturing processes must be maintained in a state of statistical control.

Statistical Quality Design and Control Richard E. DeVor, University of Illinois, Urbana-Champaign; Tsong-how Chang, University of Wisconsin, Milwaukee

Design of Experiments: Factorial Designs The process of product design and its associated manufacturing processes and tolerance designs often involve many experiments to better understand the various cause-effect relationships for quality performance of the product and for ease in process control. The sections that follow will present some of the basic concepts and methods for the planning, design, and analysis of experiments. The purpose of most experimental work is to discover the direction(s) of change that may lead to improvements in both the quality and the productiveness of a product or process. Such endeavors can be referred to as process improvement because product improvement can only be meaningfully measured through its use and that is of course a process. Historically, design of experiments methods have tended to focus more attention on process improvement as contrasted with product design. In this regard, the different view that might be taken toward design of experiments in product design versus processing is probably overstated. In this section, the role of design of experiments in product design is emphasized, as is its use in the simultaneous engineering of products and processes. In investigating the variation in performance of a given product or process, attention focuses on the identification of those factors that, when allowed to vary, cause performance to vary in some way. Some such factors are qualitative in nature (categorical variables), while others are quantitative (possessing an inherent continuity of change). The situations examined below may consider both qualitative and quantitative variables simultaneously. In fact, an important advantage of the two-level factorial designs that will be discussed is their ability to consider both types of variables within the same test plan. Mathematical Model A fundamental problem of design of experiments is that of selecting the appropriate arrangement of test points within the space defined by the design/control and noise variables. Although many different considerations must come into play in selecting a test plan, none can be more fundamental than the notion of the mathematical model. Whether or not explicitly recognized as such, most experimental studies are aimed either directly or indirectly at discovering the relationship between some performance response and a set of candidate variables influencing that response. In general, this relationship can be written as:

Y = f(X1, X2, . . ., Xk) + e

(Eq 14)

where Y is the response of interest, f is some unknown functional relationship, X1, . . ., Xk are the independent variables, and e is a random error. The functional form f can be thought of as a transfer function. In Taguchi's framework, the variables X1, X2, . . ., Xk are generally partitioned into signal, control, and noise variables. Sometimes enough is known about the phenomenon under study to use theoretical considerations to identify the form of f. For example, a chemical reaction can be described by a differential equation, which when solved produces a theoretical relationship between the dependent and independent variables. More often than not, however, the knowledge is more sparse, and empirical models must be relied upon that act as mathematical french curves describing relationships through the data; for example:

Y = b0 + b1X1 + b2X2 + e Y = b0 + b1X1 + b11

+e

(Eq 15a) (Eq 15b)

Model Building. In most studies, the experimenter begins with a tentative hypothesis concerning the plausible model

forms that are to be initially entertained. He must then select an experimental design having the ability to produce data that will:

• •

Be capable of fitting proposed model(s) Be capable of placing the model in jeopardy in the sense that inadequacies in the model can be detected through analysis

The second consideration above is of particular importance to ensure that through a series of iterations the most appropriate model can be determined, while others may be proved less plausible through the data. If, for example, a curvilinear relationship between temperature and reaction time in a chemical process is suspected, an experiment that studies the process at only two temperatures will be inadequate to reveal this possibility. An experiment with three levels of temperature would, however, allow this possibility to be considered. Figure 37 illustrates several scenarios that emphasize the importance of the relationship between the math model and the associated design of experiment. In Fig. 37, the following points should be noted: • • • •

The relationship is actually curvilinear, but such will never be detected by the data A poor model (straight line) has been hypothesized, but model checking can reveal this and help propose a better model If the relationship is known to be a straight line, many levels of temperature in the experiment are unnecessary If it is known a priori that the relationship is a straight line, the best test plan would be to study only two relatively extreme levels of temperature and to use additional tests for replication to observe the amount of experimental error

Fig. 37 Comparison of typical time-temperature relationships for true relationship (a) compared to experimental and fitted models (a through d. See text for more details.)

Sequential and Iterative Experimentation. There is always the temptation to carefully design one large experiment that will consider all aspects of the problem at hand. Such a step is dangerous for the following reasons:

•

•

•

•

If erroneous decisions and hypotheses about the state of affairs are made, considerable time and experimental resources may be wasted, and the end product may provide little useful information or direction in terms of what to do next If knowledge of the underlying situation is limited a priori, many factors may be suspected as being important, requiring a very large experiment in terms of number of tests. Ultimately, only a small subset of variables will be found to be of major significance In the early stages of experimentation, knowledge of the specific ranges of variables that ought to be studied is not always available. Furthermore, the metrics to be employed to define the variables, responses, or even what responses to observe may not always be clear in the early stages of experimental work One large experiment will necessarily cause the testing period to be protracted in time, making it more

difficult to control the forces of nuisance variation

For these reasons, it is much more desirable to conduct an experimental program through a series of smaller, often interconnected experiments. This provides the opportunity to modify hypotheses about the state of affairs concerning the situation at hand, to discard variables that are shown to be unimportant, to change the region of study of some or all of the variables, and/or to define and include other measures of process performance. Experimental designs that can be combined sequentially are very useful in this regard. This is often referred to as the sequential assembly of experimental designs. Revelation of Variable Effects. Often, the variables of importance are not clearly known a priori. It is desirable to be

able to study several variables together but to independently observe the effect of a change in each one of the variables. Furthermore, it may be deemed important to know if such a variable effect varies with the conditions of the process, that is, when other variables take on varying levels. An arrangement of the tests is called a design, which provides for the opportunity to learn much about the relationships between the variables and the response. In particular: • •

The effect of changing any of the variables alone can be observed The possibility that the effects measured above can vary as conditions of other variables vary can be observed, that is, the existence of variable interactions

System Noise/Variation The experimental study of any phenomenon is made difficult by the presence of noise or experimental error. Many factors, not directly under study, are varying over the course of the experiment. These are often referred to as the forces of common cause system variation. Such variation may cloud or mask the effect of change of the factors under study in an experiment. The forces of noise or variation can be better understood or mitigated by several approaches, some of which are strictly experimental design issues. Statistical Control/Stability Analysis. If the phenomenon under study is already a viable and ongoing process, the

pursuit of improvement opportunities through experimentation can be considerably enhanced by employing the techniques of statistical process control. In this way, spurious or sporadic sources of variation can be identified and, through remedial action, removed. Achievement of a stable process will greatly contribute to the ability to more readily observe the effects of purposeful process change. Thus, continued study will further enhance the ability to observe the persistence of changes that might be introduced. Once a process is stabilized, continued attack on the common cause system will lead to a progressively quieter process, further heightening the ability to observe the forces of purposeful process change through experimentation. Experimental Design Strategies. In many situations, the notion of a stable, ongoing process has little meaning. In

the early stages of product or process design or prototype or pilot-plant testing, a stable, consistent process is not present. It is perhaps for this reason (among others) that the body of knowledge known as experimental design was cultivated. Under such situations, the following factors are significant: •

•

• •

Attempt to identify major sources of variation and take action to ensure that their presence is blocked out from the comparisons made within an experiment. The technique of blocking is useful for this purpose Counteract the forces of unknown systematic variation over the period of the experiment by randomization of the tests so that such variation is uniformly and randomly distributed across the trials conducted Include replication in the experimental test plan. Multiple tests at the same conditions will provide comparisons that directly measure the amount of variation/experimental error Include confirmatory testing as part of the experimental strategy. It will be important that additional trials are run under specific conditions determined from the analysis to verify the improvement opportunities revealed from the experiment

The parameter design method is specifically directed at mitigating the forces of noise variation as it may be transmitted through the product/process design to the output performance. Nature of Variable Interactions For many products and/or processes, the effects that the important design/control factors have on the system performance responses of interest do not act independently of each other. That is, the effect a certain factor has on the response may be different for different levels of a second factor. When this occurs, the two factors are said to interact or to have an interdependency relationship; that is, a two-factor interaction effect is present. Figure 38 summarizes the nature of the two-factor interaction effect. Figure 38(a) shows that the effect of pressure on time (the slope of the line) is the same regardless of the level of temperature. Therefore, no interaction is present. However, in Fig. 38(b), the effect of pressure on time is clearly seen to vary with temperature. Therefore, a two-factor interaction is present.

Fig. 38 Graphical depiction of the absence (a) and presence (b) of a two-factor interaction effect

Simple Yet Powerful Experimental Design. Many of the problems created by ad hoc testing methods and/or

methods such as the one-variable-at-a-time approach can be overcome by using an experimental design structure referred to as the two-level factorial design. For such test plans, each factor/variable is studied over only two levels or settings, and all possible combinations are examined. Therefore, the total number of unique tests required for such a test plan is 2k, where k is the number of variables; for example, for two variables, 22 = 4 test conditions define the test matrix. Figure 39 shows a graphical representation of the two-level factorial design when two and three variables are under study. The geometric representation is useful from the standpoints of interpreting the variable effects and communicating the purpose and results of the test plan to others. The corners of the square and the cube represent geometrically the conditions of each unique combination of the variable settings.

Fig. 39 Two-level factorial design for two (four tests required) (a) and three (eight tests required) (b) variables

Tables 7(a) and 7(b) provide for a more algebraic way to represent the test conditions for a two-level factorial design. The two levels for each factor are often simply referred to as the high and low levels and are represented in coded form as + or +1 and - or -1. This facilitates the determination of the variable effects, given the data. Each row in Tables 7(a) and 7(b) represents the recipe for a particular test. For example, in the 23 factorial in Table 7(b), test 3 is run with variable 1(X1) at its low level, variable 2(X2) at its high level, and variable 3(X3) at its low level. Table 7(a) Test matrix for simple two-variable, two-level factorial design Test

X1

X2

1

-

-

2

+

-

3

-

+

4

+

+

Table 7(b) Test matrix for simple three-variable, two-level factorial design Test

X1

X2

X3

1

-

-

-

2

+

-

-

3

-

+

-

4

+

+

-

5

-

-

+

6

+

-

+

7

-

+

+

8

+

+

+

Statistical Quality Design and Control Richard E. DeVor, University of Illinois, Urbana-Champaign; Tsong-how Chang, University of Wisconsin, Milwaukee

Experimental Study Using a 23 Factorial Design Example 8: Use of 23 Factorial Design to Determine Need for Preheating or Postheating of Welded High-Carbon Steel Rail Bars (Ref 23). High-carbon steel, because of its high strength and low cost, has been extensively used for railway track. The Rail Steel Bar Association sponsored a research project at the Welding Research Laboratory of the University of Wisconsin around 1960 to study whether or not preheating and postheating were necessary for the high-carbon steel to have good-quality welds. After a preliminary investigation by manual arc-welding tests, it appeared that there were three variables affecting the ultimate tensile stress of a weld. Based on the needs of the study and the available funds, it was suggested to run 16 tests that were defined according to a 23 factorial design. The three selected variables and their high and low levels are shown below:

Variable

Low level

High level

Ambient temperature (T), °F

0

70

Wind velocity (V), mph

0

20

4

11

Bar size (B),

in.

First, the variables are coded such that the high level will be denoted by +1 and the low level by -1. By so doing, regardless of the physical conditions represented by the two levels, the basic design of any two-level factorial design becomes a simple arrangement of plus or minus ones. Writing down three columns of the plus and minus ones next to one another produces the desired 23 factorial design, which consists of the 8 distinct sets of coded test conditions, as indicated

in Table 8. The 8 sets of coded test conditions are given by the 8 rows corresponding to test numbers 1 to 8. These constitute the recipes for running the tests. For example, the actual welding conditions for test number 1 (which is denoted by -1, -1, -1) were a temperature of 0 °F, a wind velocity of 0 mph, and a bar size of

in.

Table 8 Coded test conditions based on actual test conditions for the steel rail bars of Example 8 Test number

Coded test conditions

Actual test conditions

Temperature, °F

Wind velocity, mph

Bar size,

X1

X2

X3

in.

1

0

0

4

-1

-1

-1

2

70

0

4

+1

-1

-1

3

0

20

4

-1

+1

-1

4

70

20

4

+1

+1

-1

5

0

0

11

-1

-1

+1

6

70

0

11

+1

-1

+1

7

0

20

11

-1

+1

+1

Geometric Representation of 23 Factorial Design. If the three variables are considered to be three mutually

perpendicular coordinate axes X1, X2, and X3, the 23 factorial design can be represented geometrically as a cube, as shown in Fig. 40. The circled numbers at the 8 corner points of the cube represent the corresponding test numbers in standard order. The 8 actual test conditions are given in brackets. This geometric configuration will be useful later when the average and interaction effects of the variables are calculated and when the average/main effect and two- and three-factor interaction effects are interpreted.

Fig. 40 Geometric representation of the 23 factorial design

Test Conduct: Random Order and Replication. In performing experiments, randomization of the test order should

be exercised wherever possible. This is important because the standard order discussed previously is a very systematic ordering of the tests in terms of the patterns of the levels of the variables from test to test. In this example, there are 8 test conditions, each of which was replicated once for a total of 16 tests. The main purpose of conducting replicated tests is to provide an estimate of the experimental error of the test method. The observed responses of these welding experiments were the ultimate tensile stresses of the welds. Because of replication, 2 responses (yai, ybi) were observed at each of the 8 distinct sets of test conditions. Therefore, there were a total of 16 responses (Table 9). Table 9 Results of the 16 welding experiments for the steel rail bars of Example 8 Because of replication, two responses (yai, ybi) were observed for each of the eight distinct sets of test conditions. i

1

X1

-1

X2

-1

X3

-1

Replicate 1 Order of

Replicate 2 Order of

yai

yai

ybi

ybi

6

84.0

3

91.0

Average,

87.5

2

+1

-1

-1

8

90.6

7

84.0

87.3

3

-1

+1

-1

1

69.6

5

86.0

77.8

4

+1

+1

-1

2

76.0

4

98.0

87.0

5

-1

-1

+1

5

77.7

8

80.5

79.1

6

+1

-1

+1

3

99.7

1

95.5

97.6

7

-1

+1

+1

4

82.7

2

74.5

78.6

8

+1

+1

+1

7

93.7

6

81.7

87.7

The tests were performed in a random fashion, as indicated by the test order. The average of the 2 responses observed at each set of test conditions (that is, yai + ybi/2, denoted by i) was determined, and the 8 average responses ( i, where i = 1, 2, . . ., 8) are given in the far right-hand column of Table 9. These average responses will be used to calculate the average main and average interaction effects for the variables under study. Calculation of Variable Effects on Weld Strength. Given the results in Table 9, it is possible to evaluate the

individual and joint influences of ambient temperature, wind velocity, and bar size on the ultimate tensile stress of the weld. To help determine what these variable effects are, what they mean, and how they can be calculated, the geometrical representation of the experimental design is used. The cube shown in Fig. 41 depicts the 23 design geometrically. The average response (ultimate tensile strength) for each test is given at the corners of the cube.

Fig. 41 Geometric representation of the effect on ultimate tensile stress, given at the corners of the cube, of changes in ambient temperature when wind velocity and bar size are held constant

Calculation of Average Main Effects. As Fig. 41 shows, there are four comparisons of test results or contrasts that

indicate how ultimate tensile stress changes when ambient temperature changes, with wind velocity and bar size being held constant. The effect of ambient temperature alone can be determined by looking across the cube from left to right. The differences in the results within each of the 4 pairs of tests reflect the effect of ambient temperature alone on ultimate tensile stress. The differences in units of ksi are: • • • •

Test No. 1 and 2: Test No. 3 and 4: Test No. 5 and 6: Test No. 7 and 8:

4682

= 87.3 - 87.5 = -0.2 3 = 87.0 - 77.8 = 9.2 5 = 97.6 - 79.1 = 18.5 7 = 87.7 - 78.6 = 9.1 1

The average effect of ambient temperature, designated by E1, is defined as the average of the above four differences. Note that this average effect is also commonly referred to as a main effect. That is:

E1 =

[(

2

-

1)

+(

4

-

3)

+(

6

-

5)

+(

8

-

7)]

E1 = [(-0.2) + (9.2) + (18.5) + (9.1)] E1 = 9.15 ksi Geometrically, the average (main) effect of ambient temperature, E1, is the difference between the average test result on plane II (high level of ambient temperature) and the average test result on plane I (low level of ambient temperature), as shown in Fig. 41. The average effect of ambient temperature indicated that, on the average, over the ranges of the variables studied in this investigation, the effect of changing the ambient temperature from its low level to its high level is to increase the ultimate tensile stress by 9.150 ksi. However, the individual differences (0.2, 9.2, 18.5, and 9.1 ksi) are actually quite erratic. The average effect, therefore, must be interpreted with considerable caution because this effect is not particularly consistent over the four unique combinations of wind velocity and bar size. Again, referring to Fig. 41, the following pairs of tests can be compared or contrasted to determine the effect of wind velocity. The average effect of wind velocity, E2 can be obtained by taking the average of the four individual differences, which are in units of ksi: • • • •

Test No. 1 and 3: Test No. 2 and 4: Test No. 5 and 7: Test No. 6 and 8:

4783

= 77.8 - 87.5 = -9.7 2 = 87.3 - 87.0 = -0.3 5 = 78.6 - 79.1 = -0.5 6 = 87.7 - 97.6 = -9.9 1

The value E2 for tests 1 to 8 is calculated as follows:

E2 =

[(

3

-

1)

+(

4

-

2)

+(

7

-

5)

+(

8

-

6)]

E2 = [(-9.7) + (-0.3) + (-0.5) + (-9.9)] = -5.1 ksi Similarly, the average effect of bar size, E3, is 0.85 ksi. Interpretation of Average/Main Effects. The average or main effect of a variable has been defined as the amount

of change observed in the response, on the average, when only that variable changes from its low to high value or level. It is an average effect because within the experimental design there are generally several comparisons or contrasts that can

be made that measure how much the response changes when only a certain variable changes. The sign and the magnitude of an average/main effect have the following meaning: • •

The sign indicates the direction of the effect The magnitude indicates the strength of the effect

If the numbers comprising this average are quite similar in magnitude (and of course, sign), the effect of that particular variable would appear to be independent of the particular level(s) of other variable(s). However, if the numbers comprising this average are quite different in magnitude (or even sign), the effect of that particular variable depends on the level(s) that the other variable(s) assumes. In this case, the average/main effect does not hold much significance, and the response of interest must be examined when two or more variables change simultaneously. Meaning of Variable Interactions. In calculating the average/main effect of ambient temperature, the amount (and

even direction) of change in weld strength with change in temperature seemed to depend quite heavily on the particular levels of wind velocity and bar size:

Wind velocity, mph

Bar size, in.

Effect of temperature on weld strength, ksi

0

-0.2

20

+9.2

0

+18.5

20

+9.1

What is the particular nature of the dependency of the effect of temperature on wind velocity and bar size? The degree of dependency of the temperature effect on the particular levels of wind velocity and bar size can be appreciated via the following graphical summaries. Referring to the geometrical representation of the 23 factorial in Fig. 41, the interaction between temperature and wind velocity can be examined by compressing the cube in the bar size direction as shown in Fig. 42(a). Compressing the cube means that the response values for a given temperature and wind velocity combination-for example, (-), (-)--are averaged across the high and low levels of bar size. The result is that the cube (Fig. 42a) becomes a square (Fig. 42b). Similarly, the other two interactions are found to be:

E13 = [(92.65 - 78.85) - (87.15 - 82.65)]/2 = 4.65 ksi E23 = [(83.15 - 82.40) - (88.35 - 87.40)]/2 = -0.10 ksi

Fig. 42 Two-factor interaction between temperature and wind velocity obtained by compressing threedimensional cube (a) in the bar size direction to obtain two-dimensional square plane (b)

A Simplified Method for Main and Interaction Effects. Although the geometrical representation of the

experimental design has been quite useful, a simpler and more general method will be needed for calculation of the variable main and interaction effects. A simplified calculation procedure, which is easily extended for analyzing two-level factorial designs in any number of variables, is described below. The mathematical model form associated with a 22 factorial is:

Y = b0 + b1X1 + b2X2 + b12X1X2 + e where the coefficients b1 and b2 correspond to the average/main effects of variables 1 and 2, respectively, and the coefficient b12 corresponds to the interaction effect between variables 1 and 2.

Referring to the design matrix for Example 8 (Table 9), estimates of the interaction effects E12, E13, E23, and E123 can be obtained by forming the cross-product columns X1X2, X1X3, X2X3, and X1X2X3, as indicated in Table 10. This complete seven-column matrix will be referred to as the calculation matrix. The cross-product columns of ± signs are simply the inner-products of the individual columns, for example, X1X2 = (X1)(X2). Table 10 Calculation matrix for the data obtained for the steel rail bars of Example 8 Test

Main effects

Interactions

X1

X2

X3

X1X2

X1X3

X2X3

X1X2X3

1

-1

-1

-1

+1

+1

+1

-1

87.5

2

+1

-1

-1

-1

-1

+1

+1

87.3

3

-1

+1

-1

-1

+1

-1

+1

77.8

4

+1

+1

-1

+1

-1

-1

-1

87.0

5

-1

-1

+1

+1

-1

-1

+1

79.1

6

+1

-1

+1

-1

+1

-1

-1

97.6

7

-1

+1

+1

-1

-1

+1

-1

78.6

To calculate any one of the average effects or interactions, merely multiply, element by element, the appropriate column by the column of average responses, sum algebraically, and divide the sum by 4 (that is, 23/2; in general, for a 2k factorial design, the sum should be divided by 2k/2). For example, to calculate the average effect of ambient temperature, E1, it is necessary to proceed as indicated in Table 11. The sum is 36.6 × 103 psi or 36.6 ksi, which, when divided by 4, yields 9.150 ksi for the average effect of temperature, the same answer obtained previously. Table 11 2k factorial design calculation of the average effect of ambient temperature on steel rail bar ultimate tensile strength in Example 8 X1

X1

i

X1

i

i

-1

87.5

(-1)

(87.5)

-87.5

+1

87.3

(+1)

(87.3)

+87.3

-1

77.8

(-1)

(77.8)

-77.8

+1

87.0

(+1)

(87.0)

=

=

+87.0

-1

79.1

(-1)

(79.1)

-79.1

+1

97.6

(+1)

(97.6)

+97.6

-1

78.6

(-1)

(78.6)

-78.6

+1

87.7

(+1)

(87.7)

+87.7

Sum

+36.6

=

The two-variable and three-variable interactions can be calculated by the same method (Table 12). For example, to calculate the interaction between ambient temperature and wind velocity, E12, multiply the column X1X2 by the column of 3 i, sum algebraically, and then divide the sum by 2 /2 = 4. Dividing the sum by 4 yields the answer obtained previously (namely, 0) for this two-variable interaction. Table 12 2k factorial design calculation of interaction between ambient temperature and wind velocity on steel rail bar ultimate tensile strength in Example 8 X1X2

X1X2

i

X1X2

i

+1

87.5

(+1)

(87.5)

+87.5

-1

87.3

(-1)

(87.3)

-87.3

-1

77.8

(-1)

(77.8)

-77.8

+1

87.0

(+1)

(87.0)

+1

79.1

(+1)

(79.1)

+79.1

-1

97.6

(-1)

(97.6)

-97.6

-1

78.6

(-1)

(78.6)

-78.6

+1

87.7

(+1)

(87.7)

+87.7

=

Sum

=

=

i

+87.0

0.0

Judging the Importance of Average Effects. To judge the relative importance of a variable effect based on the

calculated average effects and interactions, it is necessary to make use of the test replications to estimate the error expected in an effect estimate. In particular, it is necessary to: •

Estimate the experimental error via the variance of an individual observation

• •

Estimate the error/variance associated with an average effect and/or interaction effect, either of which is simply a linear combination (sum) of several individual observations Construct the statistical distribution of effect estimates and examine the range of estimates that could arise

Figure 43 shows a graphical depiction of what could arise from the above-mentioned test replications. In case 1 (Fig. 43a), the effect estimates could easily arise from a distribution with a mean of zero. In case 2 (Fig. 43b), the implication is that they could not. The above procedures come under the heading of what is commonly referred to as a statistical test of significance. The two most common approaches in this regard are the hypothesis-testing method and the confidence interval method. Additional information is available in Ref 8 and 24.

Fig. 43 Judging the relative importance of a variable effect. (a) Effect not important. (b) Effect important. Open circles represent effect estimates based on multiple experiment results.

Confidence Interval Method. Each of the 8 tests in Example 8 was replicated once (Table 9), so that there were

actually 16 individual observations on the ultimate tensile stress. The error or variance of each of these 16 observations will now be estimated. The amount of error in an observation is assumed to be about the same for all observations; that is, the true variance is the same for all 16 observations, and the observations are independent. For test 1, the two observations are 84.0 ksi and 91.0 ksi. A sample variance for this test condition, designated as be calculated in the usual way:

In this example, 8 different and completely independent sample variances ( unique test condition. The 8 sample variances are calculated to be:

= 24.50

= 21.78

= 134.48

,

,...

, can

) are calculated, 1 for each

= 242.00

= 3.92

= 8.82

= 33.62

= 72.00

Because there is assumed to be a common variance for all 16 observations, an estimate for the variance is the pooled sample variance

, of the 8 estimated variances

,

,...

. In this case:

(Eq 16)

In general, an effect is determined from:

(Eq 17) where N is the total number of unique test/trial results. In this general case, the variance of an effect is given by:

(Eq 18)

An estimate of the variance of an effect

is obtained by substituting the pooled sample variance,

, for

.

A confidence interval for a certain parameter can be calculated on the basis of the sample statistic. Here the statistics of interest are the average effects, E1, E2, E3, and the interactions, E12, E13, E23, and E123. Because the sample variances of the average effects and interactions are all estimated by Student's t-statistic, is:

/4, the result, for 100(1 - a)% confidence intervals using the

(Eq 19)

where v = 8 is the degree of freedom and a = 0.05 is the level of significance. Therefore, the 95% confidence intervals for each of the sample average effects are therefore:

Table 13 gives the seven statistics (E1, E2, E3, E12, E13, E23, and E123) together with their corresponding 95% confidence intervals. From Table 13, it would appear that none of the estimated effects is important/significant. Table 13 Calculation of 95% confidence levels for average effect and interaction statistics of steel rail bars of Example 8 Average effects

95% confidence interval, psi

Ambient temperature (E1)

9150 ± 9480

Wind velocity (E2)

-5100 ± 9480

Bar size (E3)

850 ± 9480

Ambient temperature × wind velocity (E12)

0 ± 9480

Ambient temperature × bar size (E13)

-4650 ± 9480

Wind velocity × bar size (E23)

-100 ± 9480

Ambient temperature × wind velocity × bar size (E123)

4700 ± 9480

Unreplicated Factorial Designs. There are times when it is either not feasible or not desirable to include replication in a two-level factorial design. In such cases, it is therefore not possible to obtain a direct estimate of the experimental error; as a result, it becomes more difficult to assess the relative importance of the variable effects. There are, however, at least two methods that can be employed to aid in the assessment of the relative importance of variable effects estimated from the results of an unreplicated two-level factorial design. These two methods are:

• •

The use of normal probability plots of the effect estimates The use of higher-order interaction effect estimates as a means to obtain an estimate of the experimental error

These methods are discussed in Ref 7 and 25.

References cited in this section

7. G.E.P. Box and J.S. Hunter, The 2k-p Fractional Factorial Designs, Part I and Part II, Technometrics, Vol 3, 1961 8. G.E.P. Box, W.G. Hunter, and J.S. Hunter, Statistics for Experimenters, John Wiley & Sons, 1978 23. S.M. Wu, Analysis of Rail Steel Bar Welds By Two-Level Factorial Design, Weld. J. Research Supplement,

April 1964 24. R.E. DeVor and T.H. Chang, Quality and Productivity Design and Improvement: Module 2, in Statistical Process Control, Ford Motor Company, Plastic Products Division, 1988 25. R.E. DeVor and T.H. Chang, Statistical Methods for Quality and Productivity Design and Improvement, in The Tool and Manufacturing Engineer's Handbook, Vol 4, Society of Manufacturing Engineers, 1987 Statistical Quality Design and Control Richard E. DeVor, University of Illinois, Urbana-Champaign; Tsong-how Chang, University of Wisconsin, Milwaukee

Two-Level Fractional Factorial Designs Although the class of two-level factorial designs appears to be an efficient way to deal simultaneously with several factors, this efficiency quickly disappears as the number of variables to be studied grows. Because the two-level factorial requires the consideration of all possible combinations of k variables at two levels each, a ten-variable experiment would require 210 = 1024 tests. In dealing with phenomena involving continuous variables, the relationship that describes the influence of the variables on a response of interest often constitutes a relatively smooth response surface. When dealing with qualitative (discrete) variables, it is usually the case that the responses are similar at different levels of such variables. Therefore, the higherorder variable interactions are often very small or negligible in magnitude. In fact, interaction effects involving three factors or more can be ignored for the purpose of screening out the much more important main effects and two-factor interaction effects from a limited number of tests. Experimentation with large numbers of variables generally arises out of uncertainty as to which variables have the dominant influence on the response of interest. In the end, however, only a subset of all these variables will be proved to be important. This phenomenon is referred to as the sparsity of variable effects (Ref 8). If this is the case, one should be able to conduct fewer tests than required for the full factorial without much loss of relevant information. Consequences of Fractionation In the rail steel bar problem, three variables (ambient temperature, wind velocity, and bar size) were studied to determine their possible effect on the ultimate tensile strength of the welded bars. A 23 full factorial design was run, and the three main effects, three two-factor interaction effects and the one three-factor interaction effect could all be estimated separately. Suppose now that the investigator had wished to consider a fourth variable, type of welding flux, but that the full factorial, 24 = 16 tests, could not be considered. Rather, only 8 tests can be run. Is it possible to conduct an experiment on 4 factors with only 8 tests, and if so, what useful information can be obtained from the results? Based on the assumption about the negligible third and higher-order interaction effects, one could consider assigning the column of plus and minus signs associated with the 123 interaction to be the fourth variable and estimate the average main effect of variable 4 using this column. In using the 123 column to introduce a fourth variable into the experiment, the new design matrix that defines the 8 tests to be conducted becomes:

Test

1

Variables

1

2

3

4 (123)(a)

-

-

-

-

2

+

-

-

+

3

-

+

-

+

4

+

+

-

-

5

-

-

+

+

6

+

-

+

-

7

-

+

+

-

8

+

+

+

+

(a)

123 column replaced by variable 4

Once the tests are conducted in accordance with the test recipes defined by the design matrix, the calculation matrix is determined to provide for the estimation of the interaction effects. By expanding the design matrix above, Table 14 provides the calculation matrix obtained by forming all possible products of columns 1 through 4. Table 14 Calculation matrix obtained by expanding the design matrix Test number

Average

1

2

3

4

12

13

14

23

24

34

123

124

134

234

1234

1

+

-

-

-

-

+

+

+

+

+

+

-

-

-

-

+

2

+

+

-

-

+

-

-

+

+

-

-

+

-

-

+

+

3

+

-

+

-

+

-

+

-

-

+

-

+

-

+

-

+

4

+

+

+

-

-

+

-

-

-

-

+

-

-

+

+

+

5

+

-

-

+

+

+

-

-

-

-

+

+

+

-

-

+

6

+

+

-

+

-

-

+

-

-

+

-

-

+

-

+

+

7

+

-

+

+

-

-

-

+

+

-

-

-

+

+

-

+

8

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

Examination of the calculation matrix in Table 14 reveals that many of the columns are identical. In particular, of the 16 columns, only 8 are unique; each unique column appears twice. The following pairs of variable effects are represented in the calculation matrix by the same column of plus and minus signs:

1 and 234

2 and 134

3 and 124

4 and 123

12 and 34

13 and 24

23 and 14

Ave and 1234

What does all this mean? For example, when the 12 column is multiplied by the data, summed, and divided by 4, is the result an estimate of the two-factor interaction 12? Or the two-factor interaction 34? Or both? The interactions 12 and 34 are said to be confounded or confused. They are aliases of the unique column of plus and minus signs defined by (+ - - + + - - +). Use of this column for effect estimation produces a number (estimate) that is actually the combined total of the two-factor interaction effects 12 and 34. Similarly, 1 and 234 are confounded effects, 2 and 134 are confounded effects, and so on. It seems that the innocent act of using the 123 column to introduce a fourth variable into a 23 full factorial scheme has created a lot of confounding among the variable effects, but if all the three-factor interactions are negligible, one obtains clear estimates of all 4 main effects from only 8 tests. The 2k-p Fractional Factorial Designs. The four-variable, eight-test, two-level experiment discussed thus far is

referred to as a two-level fractional factorial design because it considers only a fraction of the tests defined by the full factorial. In this case, a one-half fraction design has been created. It is commonly referred to as a 24-1 fractional factorial design. It is a member of the general class of 2k-p fractional factorial designs. In these designs, the following factors must be considered: • • • •

k variables are examined in 2k-p tests Require that p of the variables be introduced into the full factorial in k-p variables Assign them to interaction effects in the first k-p variables. This assignment is done through the use of relationships known as generators (Ref 7, 8) These generators can then be used to establish the defining relationship (Ref 7, 8), which completely reveals the confounding/alias structure of the experimental design

Resolution of Two-Level Fractional Factorial Designs. As discussed previously, the introduction of additional

variables into full two-level factorials gives rise to confounding or aliasing of variable effects. It would be desirable to make this introduction in such a way as to confound low-order effects (main effects and two-factor interactions)--not with

each other but with higher-order interactions. Then, under the assumption that third and higher-order interactions can be neglected, the low-order effects become, in a sense, unconfounded by this assumption. To illustrate, consider the study of 5 variables in just 16 tests (the full factorial would require 25 = 32 tests). One additional variable--the fifth variable--must be introduced into a 24 = 16 run base design. Any of the interactions in the first 4 variables could be used for this purpose: 12, 13, 14, 23, 24, 34, 123, 124, 134, 234, 1234. If any one of the twofactor interactions are used, for example, 5 = 12, then the design generator becomes I = 125, which is also the defining relationship. Therefore, at least some of the main effects will be confounded with two-factor interactions, namely, 1 = 25, 2 = 15, 5 = 12. Selecting Preferred Generators. If any one of the three-factor interactions is used to introduce the fifth variable, the situation is greatly improved, at least for the estimation of main effects. For example, if 5 =123, then I = 1235 is the generator and defining relationship; therefore, some main effects are confounded with, at worst, three-factor interactions, while two-factor interactions are confounded with each other, for example:

1 = 235

2 = 135

3 = 125

5 = 123

12 = 35

13 = 25

23 = 15

If the four-factor interaction in the first 4 variables is used to introduce the fifth variable, an even more desirable result is obtained (the best under these circumstances): 5 = 1234. The generator and defining relationship is I = 12345. Therefore:

1 = 2345

2 = 1345

3 = 1245

4 = 1235

5 = 1234

12 = 345

13 = 245

14 = 235

15 = 234

23 = 145

24 = 135

25 = 134

34 = 125

35 = 124

45 = 123

In this last case, all main effects are confounded with four-factor interactions, and all two-factor interactions are confounded with three-factor interactions. Concept of Resolution. The varying confounding structures produced by using different orders of variable

interactions to introduce the fifth variable in the example above are described by the concept of the resolution of fractional factorial designs. The resolution of a two-level fractional factorial design is defined to be equal to the number of letters (numbers) in the shortest-length word (term) in the defining relationship, excluding I. If the defining relationship of a certain design is I = 124 = 135 = 2345, then the design is of Resolution III. If the defining relation of a certain design is I = 1235 = 2346 = 1456, then the design is of Resolution IV. The last design examined above, which had the defining relationship I = 12345, is a Resolution V design. These concepts can be summarized as follows: • •

•

If a design is of Resolution III, this means that at least some main effects are confounded with twofactor interactions If a design is of Resolution IV, this means that at least some main effects are confounded with threefactor interactions, while at least some two-factor interactions are confounded with other two-factor interactions If a design is of Resolution V, this means that at least some main effects are confounded with fourfactor interactions, and some two-factor interactions are confounded with three-factor interactions

Example 9: Design Resolution/Selection of Generators. A 26-2 fractional factorial design is set up by introducing variables 5 and 6 via 5 = 123, 6 = 1234. What is the resolution of this design?

The design generators are I = 1235 = 12346, with the defining relationship I = 1235 = 12346 = 456. Therefore, the design is of Resolution III. What would the resolution be if the generators were 5 = 123 and 6 = 124? Because the defining relationship is I = 1235 = 1246 = 3456, the design is of Resolution IV. It is clear that the selection of the proper design generators is very important. Additional Observations on Design Resolutions. From the above discussion on design resolution, several

observations can be made: •

• • •

Higher-resolution designs seem more desirable because they provide the opportunity for low-order effect estimates to be determined in an unconfounded state, assuming higher-order interaction effects can be neglected The more variables considered in a fixed number of tests, the lower the resolution of the design becomes There is a limit to the number of variables that can be considered in a fixed number of tests while maintaining a prespecified resolution requirement No more than (n - 1) variables can be examined in n tests (n is a power of 2, for example, 4, 8, 16, 32, . . .) to maintain a design resolution of at least III. Such designs are commonly referred to as saturated designs. Examples are 23-1, 27-4, 215-11, 231-26. For saturated designs, all interactions in the base design variables are used to introduce additional variables

Importance of Sequential Experimentation In the early stages of an investigation, it often seems that many variables are of potential importance. A project group or task force can draw up a list of 5 to 15 or more variables. In the final analysis, perhaps only 2 or 3 of these variables will prove to be important. The problem is, Which ones? The first task at hand is to conduct some experiments that will quickly reduce the number of variables under study to the few seemingly important ones that will then be the focus of further experimentation. For this screening task, two-level fractional factorial designs constitute a powerful and efficient tool. If the investigator attempts to get his arms around the entire problem by designing one comprehensive experiment, the resource requirements will probably be extensive, and the final results of the experiment may be inconclusive because of poor selection of variable levels and/or a poorly controlled experimental environment precipitated by a large experiment. It would be wise to take a sequential approach, building up knowledge more gradually through a series of related experiments. In this regard, two-level fractional factorial designs serve as useful building blocks in a course of sequential experimentation. This notion is important to ensure that a lack of knowledge early in an investigation will not lead to the waste of experiment resources--for example, inappropriate variables or variable levels chosen in the context of one large experiment. It is possible to identify families of experiments that combine or piece together well. A key point here is that these related experiments can provide several different alternatives for a second experiment, depending on the results and inferences drawn from the first experiment. The sequential assembly aspect of related experiments is discussed below. Example 10: Notion of Families of Fractional Factorials. All 16 one-sixteenth fractions of the 27 factorial are related through the generators I = ±124, I = ±135, I = ±236, I = ±1237. The terms remain the same, only the signs are changed for different fractions. The original design (all + signs on the generators) is called the principal fraction. The remaining 15 are the alternate fractions. The key point is that, depending on the interpretation of the results of the principal fraction, any one of several other alternate fractions can be chosen to achieve a particular result when the two fractions are combined. A general strategy used in running successive experimental designs is to choose judiciously among the members of a given family of fractional factorials. Clearly, given the principal fraction, several alternatives are available among the remaining alternate fractions, each providing a different set of information/effect estimates from the combined design. Table 15 provides some relevant information concerning the family of fractional factorials associated with a seven-variable experiment using eight-test designs.

Table 15

family of fractional factorials

Fraction

Generators

When combined with principal fraction gives:

Principal

I = 124 I = 135 I = 236 I = 1237

...

A1

I = -124 I = -135 I = -236 I = -1237

All main effects

A2

I = -124 I = - 135 I = 236 I = -1237

1, 12, 13, 14, 15, 16, 17

A3

I = -124 I = 135 I = -236 I = -1237

2, 12, 23, 24, 25, 26, 27

A4

I = 124 I = -135 I = -236 I = -1237

3, 13, 23, 34, 35, 36, 37

A5

I = -124 I = 135 I = 236 I = 1237

4, 14, 24, 34, 45, 46, 47

A6

I = 124 I = -135 I = 236 I = 1237

5, 15, 25, 35, 45, 56, 57

A7

I = 124 I = 135 I = -236 I = 1237

6, 16, 26, 36, 46, 56, 67

A8

I = 124 I = 135 I = 236 I = -1237

7, 17, 27, 37, 47, 57, 67

Orthogonal Arrays The theory of fractional factorial designs was first worked out by Finney (Ref 26) and Rao (Ref 27). Many highly fractionated designs were introduced by Tippett (Ref 28), Plackett and Burman (Ref 29), and others. Some of these were referred to as magic squares and orthogonal arrays. Two-level and three-level fractional factorial designs gained widespread attention and industrial application, beginning in the 1950s. Box and Hunter (Ref 7) provided much useful guidance to the practitioner in the adroit use of these experimental design structures. In the late 1970s, the use of orthogonal arrays for quality design and improvement by Taguchi (Ref 4) and others gained widespread acceptance in industry. This has led to considerable discussion concerning the relative merits of these designs and the methods of design selection vis-à-vis the methodology of the more general class of fractional factorial designs. In particular, it is important to examine the philosophical framework and interpretation of these similar (often identical) design structures. Issue of Confounding. To many, the most significant property of the designs is orthogonality, that is, the ability to

separate out the individual effects of several variables on a response of interest. The term orthogonal arrays as used by Taguchi implies having this property of orthogonality or producing effects that are not confounded. In fact, whenever anything less than a full factorial is under consideration, confounding or mixing of effects is present by definition. Such confounding can only be removed by assumptions made about the physical system and confirmed through some analyses of the results. Similarly, when two-level factorials are fractionated, the ability to determine all the interactions among a set of factors is lost. However, judicious fractionation leads to the ability to obtain knowledge on low-order interactions under the assumption that higher-order interactions are of negligible importance. Because the orthogonal arrays, as applied to Taguchi methods, are highly fractionated factorial designs, the information these arrays produce is a function of two elements: the nature of their confounding patterns and the assumptions made about the physical system they are applied

to. Often, Taguchi's use of linear graphs for design selection fails to produce the highest possible degree of resolution, as the following example will illustrate. Example 11: Examination of the Use of an L16(215) Orthogonal Array (Ref 4). The notation for various orthogonal arrays by Taguchi and Wu is generally of the following form for designs without mixing levels Ln(rm), where n is number of tests, r is number of levels of each factor, and m is the maximum number of factors, including selected interactions, that the design can study. For example, L16 (215) refers to a design with 16 tests to study up to 15 two-level factors and/or interactions. An L16(215) orthogonal array is given in Table 16. Table 16 An L16 (215) orthogonal array F

A

e(a)

B

e(a)

A×B

E

C

H

e(a)

B×D

e(a)

A×D

G

D

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

1

1

1

1

1

1

1

1

1

1

1

1

1

1

2

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

3

1

1

1

2

2

2

2

1

1

1

1

2

2

2

2

4

1

1

1

2

2

2

2

2

2

2

2

1

1

1

1

5

1

2

2

1

1

2

2

1

1

2

2

1

1

2

2

6

1

2

2

1

1

2

2

2

2

1

1

2

2

1

1

7

1

2

2

2

2

1

1

1

1

2

2

2

2

1

1

8

1

2

2

2

2

1

1

2

2

1

1

1

1

2

2

9

2

1

2

1

2

1

2

1

2

1

2

1

2

1

2

10

2

1

2

1

2

1

2

2

1

2

1

2

1

2

1

11

2

1

2

2

1

2

1

1

2

1

2

2

1

2

1

12

2

1

2

2

1

2

1

2

1

2

1

1

2

1

2

13

2

2

1

1

2

2

1

1

2

2

1

1

2

2

1

14

2

2

1

1

2

2

1

2

1

1

2

2

1

1

2

Test

15

2

2

1

2

1

1

2

1

2

2

1

2

1

1

2

16

2

2

1

2

1

1

2

2

1

1

2

1

2

2

1

(a) e = estimate of error

In this application of the L8(215) orthogonal array, the following assignments have been made: • • •

Eight columns have been assigned to variable main effects A, B, C, D, E, F, G, H Three columns have been assigned to two-factor interactions A × B, B × D, A × D Four columns have been assigned to the estimation of error

Close examination of the L16(215) orthogonal array shows that the columns assigned to E, G, H, and C are the following products of the columns assigned to D, B, A, and F:

5 = 234 (E = B × A × F)

7 = 123 (G = C × B × A)

6 = -14 H = -(C × F)

8 = -1234 D = -(C × B × A × F)

It follows then from the associated defining relationship that this L16(215) orthogonal array is not only a 28-4 fractional factorial but also a Resolution III 28-4 design. This is surprising because it is possible to design a 16-run, two-level fractional factorial to examine eight variables (Resolution IV) using generators such as:

5 = 234 (I = 2345)

6 = 134 (I = 1346)

7 = 123

(I = 1237)

8 = 124 (I = 1248)

The L16(215) orthogonal array under examination, when viewed as a 28-4 fractional factorial design, produces the following 16 linear combinations of effects, which can be estimated from the data. It is assumed that three-factor interactions and higher can be neglected:

l0 estimates

I (mean)

l1 estimates

1 - 46 - 58

l2 estimates

2

l3 estimates

3

l4 estimates

4 - 16 - 78

l5 estimates

12 + 37

l6 estimates

13 + 27

l7 estimates

6 - 14 - 57

l8 estimates

23 + 45 + 17 + 68

l9 estimates

24 + 35

l10 estimates

34 + 25

l11 estimates

7 - 48 - 56

l12 estimates

26 + 38

l13 estimates

36 + 28

l14 estimates

5 - 18 - 67

l15 estimates

8 - 15 - 47

It is interesting to note the following for this particular 28-4 fractional factorial: • • •

Only two of the eight main effects are not confounded with two-factor interactions. The other six main effects are each confounded with two two-factor interactions One of the desired two-factor interactions (23) is confounded with three other two-factor interactions The four columns assigned to the estimation of error each represent pairs of two-factor interactions confounded with each other

An Alternate L16(215) Orthogonal Approach. It would appear that the L16(2 ) orthogonal array using the 15

generators given below is more suitable for the purposes of Taguchi methods (or any methods for that matter), particularly given the assumptions generally made about interaction effects:

5 = 234

6 = 134

7 = 123

8 = 124

From this 28-4 fractional factorial design, the linear combination of effects (ignoring interactions of three or more factors) given in Table 17 would result. Table 17 Linear combination of effects obtained from 28-4 fractional factorial design Effect name in Taguchi experiment

l0 estimates

I (mean)(a)

l1

D

l2

B

l3

A

l4

F

l5

+ 37 + 48 + 56

B×D

l6

+ 46 + 37 + 58

A×D

l7

14 + 36 + 28 + 57

l8

+ 45 + 17 + 68

e(b)

A×B

l9

24 + 35 + 18 + 67

e(b)

l10

34 + 25 + 16 + 78

e(b)

l11

H

l12

C

l13

G

l14

E

15 + 25 + 47 + 38

l15

e(b)

(a) Underscored effects identify corresponding effects. For example,

is D,

is B, and so on, in

Taguchi's experiment.

(b) e = estimate of error

The following conclusions can be drawn from Table 17: • • •

All eight main effects are confounded with three-factor and higher--only two previously All three desired two-factor interactions are confounded with three other two-factor interactions--only one previously All four error estimates are a set of four confounded two-factor interactions; previously, they were sets of two confounded two-factor interactions. This is a trade-off for the estimation of two of the desired two-factor interactions: B × D and A × D

Choice of Variables in Calculation Matrix. When fewer than k = n - 1 variables are studied in 2

k-p

= n tests, some choice exists in terms of how variables (main effects) are assigned to columns in the Ln(2 ) array (the calculation matrix). Some choices are often better than others in terms of the resulting design resolution. Under the Taguchi approach to design of experiments, the philosophy toward interaction effects is quite different. As a result, the bottom line is that unless they are expressly identified and assigned to the orthogonal array, they are assumed to be neglectable. If such an attitude is adhered to, it is perhaps of even greater importance to place emphasis on design resolution and the careful examination of the alias structure of the experimental design. When higher-resolution designs (arrays) are available, it is k

of utmost importance to seek them out in an effort to protect the assumptions used. Again, Example 11 demonstrates the importance of design resolution and the judicious choice of design generators. In the context of Taguchi's approach to design of experiments, this is even more important because the sequential assembly concept is not used; that is, single experiments tend to be employed on any given problem.

References cited in this section

4. G. Taguchi and Y. Wu, Introduction to Off-Line Quality Control, Central Japan Quality Control Association, 1979 7. G.E.P. Box and J.S. Hunter, The 2k-p Fractional Factorial Designs, Part I and Part II, Technometrics, Vol 3, 1961 8. G.E.P. Box, W.G. Hunter, and J.S. Hunter, Statistics for Experimenters, John Wiley & Sons, 1978 26. D.J. Finney, The Fractional Replication of Factorial Arrangements, Ann. Eugen., Vol 12, 1945, p 291-301 27. C.R. Rao, Factorial Experiments Derivable From Combinatorial Arrangements of Arrays, J. Royal Stat. Soc., Vol B9, 1947, p 128-140 28. L.H.C. Tippett, Applications of Statistical Methods to the Control of Quality in Industrial Production, Manchester Statistical Society, 1934 29. R.L. Plackett and J.P. Burman, Design of Optimal Multifactorial Experiments, Biometrika, Vol 23, 1946, p 305-325 Statistical Quality Design and Control Richard E. DeVor, University of Illinois, Urbana-Champaign; Tsong-how Chang, University of Wisconsin, Milwaukee

Implementing Robust Design Recently, a number of different approaches have been proposed as possible ways to implement the robust design concept of Taguchi. These vary from purely analytical approaches, to computer simulation using product/process mathematical models and/or Monte Carlo methods, to the use of physical experimentation. In most of these approaches, the use of experimental design strategies, including two-level and multilevel factorial and fractional factorial designs and orthogonal arrays, has been extensive. Taguchi views the design process as evolving in three distinct phases or steps: • • •

System design Parameter design Tolerance design

It is perhaps this broad umbrella that he places over his concepts and methods for quality design and improvement that makes his approach so widely accepted by the engineering community. As discussed previously, Taguchi considers engineering design as the central issue and statistical methods as just one of several tools to accomplish his objectives in engineering. In addition to the different approaches to generating data on product/process performance, the issue of the specific measures of performance to use in the facilitation of parameter design needs to be considered. Taguchi and his colleagues make extensive use of the signal-to-noise ratio as a measure of performance. Others evaluate the mean and the variance or standard deviation of performance separately. The relative merits of these varying approaches with regard to performance evaluation will be discussed later in this article. The separate analysis of mean response and variation in response is discussed below.

Degree of Control The issue of the control of a certain factor is actually one of economics; therefore, any factor might be thought of in terms of its degree of control rather than as controllable or not controllable. Suppose that an injection-molding process is under consideration because of a part shrinkage problem. A brainstorming session might lead to the following list of factors as potentially having an influence on part shrinkage: • • • • • • • • • •

Cycle time Percentage of regrind versus virgin raw material Gate configuration Holding pressure Screw speed Mold temperature Cavity surface finish Raw material moisture content Raw material melt flow index Mold cooling water temperature

Although any or all of these factors can influence part shrinkage, they are not all controlled or manipulated to the same degree or with the same ease, economically. Some of these factors can be controlled at the engineering design process-for example, the gate configuration or the cavity surface finish of the mold. Others are controllable at the molding machine--for example, screw speed or cycle time. Still others are properties of the incoming raw material and are thus under the control of the supplier--for example, melt flow index. Finally, certain factors can be best categorized as defining part of the environment in which the process must function--for example, mold cooling water temperature and percentage of regrind versus virgin raw material. Two case studies will be discussed that exemplify some of the different ways in which the concept of parameter design can be implemented to develop robust product and process designs. Analytical Approaches to Parameter Design It is important to keep a broad perspective on the contributions of Taguchi and the wide variety of ways in which they can be utilized. Perhaps the single overarching concept that drives the philosophies of Taguchi is the importance placed on the consistency of performance of products and processes, that is, the importance of variation reduction. In essence, as designs are formulated and output performances evaluated, either analytically or experimentally, as much or more emphasis should be placed on variation in performance as on performance on average. There are many ways to invoke this overarching concept. Although much of the knowledge used in a given design situation is derived from experimentation, an ever-expanding baseline of knowledge exists that is being derived from first principles/physical laws and the utilization of sophisticated modeling methods such as finite-element modeling. As closed form solutions become available, it becomes possible to evaluate them analytically or numerically using classical optimization methods. The performance function should be examined in terms of both its expected value and its variance function. Searching the variance function to find those values for the design/control variables that minimize the variance in performance then amounts to invoking the parameter design method. Example 12: Parameter Design Methods Applied to Analysis of Effort Required to Close an Automobile Door (Ref 30). An explicit model is used for the mean square error of design performance in an automobile door-closing-effort problem to identify the values of the design parameters that provide the most consistent design configuration (minimum variance) while maintaining an acceptable level of average performance. In this case study, the following logic/methodology is employed. Problem Formulation. The first step in any design optimization problem is to formulate the design performance

measure of interest. In this case study, automobile door fit is the issue, particularly the sensitivity of the closing effort to variations in door positioning as it is hung in place. Figure 44 illustrates a car door and the locations of the three positioning points used to hang the door. The closing effort is a function of the compression of the weatherstripping

around the door periphery. The closing effort can be expressed analytically as a function of the weatherstripping diametral stiffness, the weatherstripping diameter and length, and the nominal values of the door position points that define the door/frame gap. The basic design problem is then to find the nominal values for the door position points that minimize the variation in closing effort as a result of the tolerance range assigned to the door position points. This problem was formulated in an attempt to exploit the underlying nonlinearities in the system to improve performance consistency.

Fig. 44 Schematic of an automobile door showing locations (solid circles) of three positioning points used to hang the door

Model Derivation. Under the assumption that the door gap is linearly proportional to the door position point locations, it is possible to derive the door-closing effort (in stored energy) as a quadratic function of the position point locations. The position location points are the control factors in this problem. The quadratic function for closing effort as a function of the three hanging positions has several unknown parameter values. Although the closing effort will clearly vary as a function of variation from car to car in the weatherstripping thickness, car body, size, and so on, only the variation in the door hanging position from car to car was examined in this case study. Model Fitting. The parameters of the quadratic function for closing effort were estimated by first calculating the door

gap at 15 points around the door as a function of variations in the door hanging positions and then using the calculated gap values to determine the closing-effort values. The variations in the hanging position control factors were introduced using a three-level factorial design, which required a total of 27 evaluations of the gap and the stored energy at each of the 15 locations around the door. These data were then used to estimate the coefficients of the quadratic closing-effort function. Design Optimization. The variance of the closing-effort quadratic function can now be derived analytically, and this

variance function can be subjected to a classical optimization routine. Such an optimization algorithm will seek the values of the three door position locations that minimize the variance of the closing effort. This optimization is subject to the constraint that the mean closing effort is less than or equal to a prespecified target value. The result is the determination of the nominal values for the design/control factors, that is, the door positioning locations that minimize the car-to-car closing-effort variation because of variation in these locations. In this design optimization problem, the control factors are the nominal values for the door position locations, and the noise factors are the tolerance/variations in the door position locations. Figure 45 shows a plot of the derived functional relationship between closing effort and the location of point 1 for positioning the door for fixed values of location points 2 and 3. Figure 45 clearly shows how the parameter design method exploits a basic system nonlinearity by seeking a new nominal value for the location of point 1 that reduces the transmitted variation to the closing effort due to the variation in the location of point 1 about its nominal value.

Fig. 45 Plot of closing effort versus point 1 location for positioning the door shown in Fig. 44 with fixed values of location points 2 and 3

Experimental Approaches to Parameter Design The parameter design concept can be implemented through physical experimentation in many different ways. The following example illustrates the explicit use of an inner/outer design for robust design. The example was developed from the results of a larger experiment and is presented to demonstrate how simple experiments such as two-level factorial designs can be used to study and improve the performance of an assembly process using the concept of robust design. Example 13: Optimization of Automotive Fuel Gage Reading to Actual Fuel Tank Quantity. Customer complaints regarding the difference between the fuel gage reading and the actual amount of fuel in the tank led to an investigation of the process of gage calibration. Seven variables were tentatively identified as being of possible importance to the position (% deflection or fallback) of the indicator in the fuel gage. Initial Design of Experiments. A 2

7-3

fractional factorial design experiment was constructed and carried out. A portion of the results will be examined that involve four of the seven variables determined to be the most important. These four variables are:

Variable No.

Variable

Low level

High level

1

Spring tension

Loose

Tight

2

Method of pointer location

Visual

Other (standard)

3

Bimetal hook twist

Vendor A

Vendor B

4

Bimetal bracket bend

90°

93°

Variables 1 and 2 describe adjustments that can be made during the calibration process on the assembly line. Variables 3 and 4 describe the condition (considered undesirable) of certain purchased parts that are assembled into the gage and may

affect the calibration process. The 27-3 fractional factorial design was collapsed into a 24 full factorial design to study the effects of these variables, as they are purposely changed, on the response in question. Table 18, the design matrix, lists the settings for the 16 tests conducted in terms of the four variables defined above. The tests are listed in standard order, although they were actually run in a randomized test sequence. Table 18 Design matrix for fuel gage experiment of Example 13 Test number

x1

x2

x3

x4

Y, %

1

Loose

Visual

Vendor A

90

14

2

Tight

Visual

Vendor A

90

10

3

Loose

Other

Vendor A

90

7

4

Tight

Other

Vendor A

90

8

5

Loose

Visual

Vendor B

90

18

6

Tight

Visual

Vendor B

90

9

7

Loose

Other

Vendor B

90

7

8

Tight

Other

Vendor B

90

10

9

Loose

Visual

Vendor A

93

15

10

Tight

Visual

Vendor A

93

9

11

Loose

Other

Vendor A

93

14

12

Tight

Other

Vendor A

93

7

13

Loose

Visual

Vendor B

93

7

14

Tight

Visual

Vendor B

93

12

15

Tight

Other

Vendor B

93

9

16

Tight

Other

Vendor B

93

12

The main effects of each of the four factors, as well as the two-, three-, and four-factor interactions, were estimated from the data and are listed below:

E1 = -2.50 *E2 = 0.00 E3 = -1.75 E4 = 3.00

*E12 = 0.50 E13 = 1.75 E14 = -1.00 E23 = 2.25 *E24 = 0.50 E34 = -1.25

*E123 = 0.75 E124 = -2.50 *E134 = 0.75 *E234 = 0.25

E1234 = -2.25

The results of the larger experiment, in fact, showed that these four variables tended to exhibit the stronger effect estimates, both main effects and interactions. A normal probability plot of the estimated effects could be interpreted as suggesting that only the effect estimates indicated by the asterisk might be considered insignificant. This interpretation leaves the investigator somewhat perplexed as to what exactly to do next. Many interactions seem important, yet it is not at all clear how the process can best be adjusted/altered. It would appear that the main effects of variables 1 and 4 might be two of the more important effects: • •

E1 = -2.50, implies that, on the average, tighter spring tension reduces fallback (deflection) E4 = +3.00, implies that larger bimetal bracket bends (93°) increase deflection, which is undesirable

The latter inference suggests that closer adherence to the 90° bend, which is the desired nominal value, may have to be requested from the bimetal bracket supplier. This essentially means that the supplier would be required to tighten the tolerance on this angle. Rather than taking this more costly approach, it might be useful to consider the variation in this part dimension as a noise variable and then determine how the transmission of the variation in the bimetal bracket bend could be reduced through the manipulation of control factors during the calibration process. Alternative Experimental Design and Analysis. A more thoughtful examination of the four variables under study seems to indicate that they fall into two basic categories. The first category consists of:

• •

Variable 1: Spring tension Variable 2: Method of pointer location

The second category consists of: • •

Variable 3: Bimetal hook twist Variable 4: Bimetal bracket bend

Variables 1 and 2 can be easily adjusted at the discretion of the operator/setup person. Variables 3 and 4 describe certain aspects of the condition of two purchased parts. Although the vendors have been given certain requirements, strong control is not exercised over these factors (which determine the condition of the parts coming in) without tightening the

specification requirements--a costly proposition. Therefore, these two variables might be considered to be noise variables, and perhaps the robust design/parameter design concept can be used to see how adjustments in variables 1 and 2 could be made to reduce the transmitted variability due to variables 3 and 4. As a result, instead of thinking in terms of a 16-run experiment design that considers four variables at two levels each, one should think in terms of two 22 factorial experiments: • •

The inner design is a 22 factorial in the two controllable variables (spring tension and method of pointer location) The outer design is a 22 factorial in the two noise variables (bimetal hook twist and bimetal bracket bend)

The outer design is conducted at each of the four unique variable settings of the inner design. A graphical representation of this experiment design is shown in Fig. 46. The numbers in the corners of each of the four outer design squares represent the fallback results for each of the 16 unique trails. Referring to the data in Fig. 46, and Y, the mean and standard deviation of the % deflection, are determined at each of the four test combinations of the inner design:

Test no.

X1

X2

1

-

-

12.0

2.94

2

+

-

9.0

3.37

3

-

+

11.5

4.8

4

+

+

9.5

2.08

Y

Fig. 46 Inner/outer experimental design structure for the fuel-gage study data in Table 18. Numbers at corners of outer design squares are fallback (% deflection) values, Y.

By analyzing the above responses (average fallback and standard deviation of fallback) separately, the settings/values can be determined for the control variables that minimize average fallback and give rise to minimum variation in fallback, that is, the most robust process design. Results. Figures 47 and 48 illustrate the results of the analyses. The interpretation of the results is as follows:

•

•

In terms of average response, % deflection is reduced by increasing spring tension (Fig. 47). This reduction is about the same for both methods of pointer location. On the average, the pointer location method has little effect on % deflection In terms of standard deviation, it is best to use the other (standard) method of pointer location along with tighter spring tension (Fig. 48)

It appears that the nominal response can be controlled by the manipulation of spring tension, while the variation in response is controlled by manipulation of the method of pointer location.

Fig. 47 Effect of control factors on average performance. Numbers at corners of square are the means of the % deflection,

.

Fig. 48 Effect of control factors on variation in performance. Numbers at corners of square are the standard deviations of the % deflection, Y.

Reference cited in this section

30. H.L. Oh, Variation Tolerant Design, in ASME Proceedings on Quality, Design, Planning, and Control, American Society of Mechanical Engineers, 1987, p 137-145

Statistical Quality Design and Control Richard E. DeVor, University of Illinois, Urbana-Champaign; Tsong-how Chang, University of Wisconsin, Milwaukee

References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

W.W. Scherkenbach, The Deming Route to Quality and Productivity: Road Maps and Roadblocks, Mercury Press/Fairchild Publications, 1987 J.M. Juran, Juran on Planning for Quality, The Macmillan Free Press, 1986 A.V. Feigenbaum, Total Quality Control, McGraw-Hill, 1983 G. Taguchi and Y. Wu, Introduction to Off-Line Quality Control, Central Japan Quality Control Association, 1979 G. Taguchi, On-Line Quality Control During Production, Japanese Standards Association, 1981 R.A. Fisher, Design of Experiments, 8th ed., Hafner Publishing, 1966 G.E.P. Box and J.S. Hunter, The 2k-p Fractional Factorial Designs, Part I and Part II, Technometrics, Vol 3, 1961 G.E.P. Box, W.G. Hunter, and J.S. Hunter, Statistics for Experimenters, John Wiley & Sons, 1978 L.P. Sullivan, Reducing Variability: A New Approach to Quality, Qual. Prog., Vol 17 (No. 7), 1985, p 1521 W.A. Shewhart, The Economic Control of Quality of Manufactured Product, Van Nostrand, 1931 (ASQC, 1980) W.E. Deming, Quality, Productivity, and Competitive Position, MIT Center for Advanced Engineering, 1982 J.M. Juran, Quality Problems, Remedies, and Nostrums, Ind. Qual. Con., Vol 22, 1966, p 647-653 S.W. Roberts, Control Charts Based on Geometric Moving Averages, Technometrics, Vol 1, 1959, p 234250 A.L. Sweet, Control Charts Using Coupled Exponentially Weighted Moving Averages, Trans. IIE, Vol 18 (No. 1), 1986, p 26-33 A.W. Wortham and G.F. Heinrich, Control Charts Using Exponential Smoothing Techniques, Trans. ASQC, Vol 26, 1972, p 451-458 A.W. Wortham, The Use of Exponentially Smoothed Data in Continuous Process Control, Int. J. Prod. Res., Vol 10 (No. 4), 1972, p 393-400 A.F. Bissell, An Introduction to CuSum Charts, The Institute of Statisticians, 1984 "Guide To Data Analysis and Quality Control Using CuSum Techniques," BS5703 (4 parts), British Standards Institution, 1980-1982 J.M. Lucas, The Design and Use of V-Mask Control Scheme, J. Qual. Technol., Vol 8 (No. 1), 1976, p 112 J. Murdoch, Control Charts, Macmillan, 1979 J.S. Oakland, Statistical Process Control, William Heinemann, 1986 I. Burr, Statistical Quality Control Methods, Marcel Dekker, 1976 S.M. Wu, Analysis of Rail Steel Bar Welds By Two-Level Factorial Design, Weld. J. Research Supplement, April 1964 R.E. DeVor and T.H. Chang, Quality and Productivity Design and Improvement: Module 2, in Statistical Process Control, Ford Motor Company, Plastic Products Division, 1988 R.E. DeVor and T.H. Chang, Statistical Methods for Quality and Productivity Design and Improvement, in The Tool and Manufacturing Engineer's Handbook, Vol 4, Society of Manufacturing Engineers, 1987 D.J. Finney, The Fractional Replication of Factorial Arrangements, Ann. Eugen., Vol 12, 1945, p 291-301 C.R. Rao, Factorial Experiments Derivable From Combinatorial Arrangements of Arrays, J. Royal Stat. Soc., Vol B9, 1947, p 128-140

28. L.H.C. Tippett, Applications of Statistical Methods to the Control of Quality in Industrial Production, Manchester Statistical Society, 1934 29. R.L. Plackett and J.P. Burman, Design of Optimal Multifactorial Experiments, Biometrika, Vol 23, 1946, p 305-325 30. H.L. Oh, Variation Tolerant Design, in ASME Proceedings on Quality, Design, Planning, and Control, American Society of Mechanical Engineers, 1987, p 137-145

Abbreviations and Symbols

o •

Abbreviations and Symbols

a

•

A

•

A

•

•

crack length; flaw size; depth; level of significance

•

ampere

•

amplitude; area

•

angstrom

•

alternating current

•

ac

•

A/D •

analog-to-digital converter

•

acoustic emission

•

AE

•

AFS •

•

AGC

•

AIA

•

AIPD

• • •

•

AISI

•

AMS

•

• •

American Foundrymen's Society automatic gain control Aerospace Industries Association of America acoustic impedance polarity detector American Iron and Steel Institute Aerospace Material Specifications

API •

American Petroleum Institute

•

ASIP

•

ASTM

• •

•

atm

•

B

•

B

standard atmosphere

•

bar size; magnetic flux density

•

magnetic flux vector

•

body-centered cubic

bcc

•

bcd •

•

BIE

•

Bq

•

BWO c

•

C

American Society for Testing and Materials

•

•

•

Aircraft Structural Integrity Program

binary coded decimal

•

boundary integral equation

•

becquerel

•

backward wave oscillator

•

speed of light; specific heat

•

coulomb

•

CAD/CAM

•

CARP

•

CCD

• •

computer-aided design/computer-aided manufacturing Committee for Acoustic Emission in Reinforced Plastics

•

charge-coupled device

•

candela

•

cd

•

CFA •

•

CFRP

•

Ci

•

CID

crossed-field amplifier

•

carbon fiber-reinforced plastic

•

curie

•

•

CIE

•

CL

charge-injected device

•

Commission Internationale de l'Eclairage (International Commission on Illumination)

•

control limit; confidence level

•

CMM

•

CMS

•

COD

• • •

coordinate measuring machine coordinate measuring system crack opening displacement

•

CRT

•

C-SAM

•

cathode ray tube

•

C-mode scanning acoustic microscopy

•

centiStokes

•

computed tomography

•

cSt

•

CT

•

CuSum •

•

CW

•

d

•

d

•

continuous wave

•

deuteron

•

used in mathematical expressions involving a derivative (denotes rate of change); depth; diameter

•

diameter; duration; dislocation density

•

D

•

da/dN •

•

DAC

•

dB

•

dBae

fatigue crack growth rate

•

digital-to-analog converter

•

decibel

•

decibel of acoustic emission amplitudes

•

direct current

•

dc

•

DCC

•

cumulative sum

•

diam

direct computer control

• •

DIN

•

DIP

•

DLM

• •

•

dpi

•

DR

•

e

•

diameter Deutsche Industrie-Normen (German Industrial Standards) dual-in-line package

•

digital light meter

•

dots per inch

•

digital radiography; digital radiograph

•

natural logarithm base, 2.71828; random error

•

modulus of elasticity; electric field strength; emissive power; electromotive force; energy counts (acoustic emission); applied voltage

•

eddy current

E

•

EC

•

ECP •

•

EDM

•

EMA

• •

electric current perturbation electrical discharge machining electromagnetic acoustic (transducer)

•

ENSIP

•

EPA

•

Eq

•

Engine Structural Integrity Program

•

Environmental Protection Agency

•

equation

•

et al.

•

ET

•

and others

•

eddy current testing

•

electron volt

•

eV

•

EWMA

•

f

•

exponentially weighted moving average

•

frequency; transfer function

•

eddy current inspection test frequency

•

F

•

FERPIC •

•

FET

•

FFT

•

FID

• • •

•

Fig.

•

fm

•

FM

•

ferroelectric photoconductor image camera field-effect transistor fast Fourier transform flame ionization detector

•

figure

•

femtometer

•

FMR

frequency modulation

•

ferromagnetic resonance

•

felicity ratio

•

FR

•

FRP

•

ft

•

fiber-reinforced plastic

•

foot

•

footcandle

•

ftc

•

ft-L

•

g

•

G

•

GD

•

footlambert

•

gram

•

gauss

•

film gradient

•

GBq

•

GPa

•

GTO

• •

•

Gy

•

h

•

h

•

H

•

H

gigabecquerel gigapascal

•

gate turnoff

•

gray (unit of absorbed radiation)

•

convection heat transfer coefficient

•

hour

•

magnetic field strength

•

magnetic field vector

•

HAZ

•

HB

•

heat-affected zone

•

Brinell hardness

•

HIP

•

HNDT

•

HR

• • •

•

HSL

•

i

•

i

•

I

• •

holographic nondestructive test Rockwell hardness (requires scale designation, such as HRC for Rockwell C hardness) hue, saturation, luminance for a complex variable; a summation index

•

unit vector along the x-axis

•

intensity; electrical current

•



•

IACS

•

hot isostatic pressing

• •

IBIS

average intensity International Annealed Copper Standard

•

IC

•

ID

•

IF

•

integrated blade inspection system

•

integrated circuit

•

inside diameter

•

intermediate frequency

•

IGA

•

IMPATT

•

in.

• •

impact avalanche transit time

•

inch

•

IQI

•

IR

•

intergranular attack

•

image quality indicator

•

infrared

j •

•

j

•

J

•

k

•

k

•

K

•

unit vector along the y-axis

•

joule

•

thermal conductivity; wave number

•

unit vector along the z-axis

•

stress intensity factor; kernel

•

plane-stress fracture toughness

•

plane-strain fracture toughness

•

Kc

•

KIc

•

KIscc

•

kg

•

km

•

threshold stress intensity for stress-corrosion cracking

•

kilogram

•

kilometer

•

kPa

•

ksi

•

kV

•

kilopascal

•

1000 lb (1 kip) per square inch

•

kilovolt

•

kVp

•

l

•

•

kilovolt peak

•

average rate of occurrence of defects per sample; length

•

length; likelihood (function)

•

pound

L

•

lb

•

LCC

•

for a complex variable; a summation index

•

LCF

life cycle cost

• •

LCL

•

LED

•

LEL

• • •

low cycle fatigue lower control limit; lower confidence limit light-emitting diode lower explosive limit

•

LINAC

•

LLW

•

lm

• •

leaky Lamb wave

•

lumen

•

natural logarithm (base e)

•

ln

•

LNG •

liquified natural gas

•

local oscillator

•

LO

•

LPF

•

lpm

• •

•

LSP

•

LUT

•

•

lx

•

m

•

m

•

Ma

•

Mt

linear accelerator

low-pass filter line pairs per millimeter laser speckle photography

•

look-up table

•

lux

•

meter

•

maximum number of factors, including selected interactions, that the design can study

•

molecular weight of air

•

molecular weight of tracer gas

•

MARSE

•

MDL

•

MFD

• •

measured area under the rectified signal envelope minimum detectable leakage

•

magnetic field disturbance

•

milligram

•

megagram (metric tonne)

•

mg

•

Mg

•

MIG •

metal inert gas (welding)

•

MIL-STD

•

MIVC

• •

•

min

•

mL

• •

military standard magnetically induced velocity change minimum; minute milliliter

•

MLCD

•

mm

•

MOOT

• • •

•

MPa

•

mpg

• •

•

mrem

•

mSv

•

MTF

• • •

•

mV

•

MVp

•

•

n

•

N

•

n

n

•

N

•

millimeter method of optimal truncation megapascal miles per gallon millirem millisievert modulation transfer function millivolt

•

megavolt peak

•

neutrons

•

newton

•

index of refraction; fringe number; number of tests

•

average sample size

•

unit vector normal to the surface

•

number of cycles; summation index; fringe order; length of the near field

•

Knudsen number

•

Reynolds number

• •

microwave liquid crystal display

NK

•

NRe

•

NBS •

•

NDE

•

NDI

•

NDT

• • •

National Bureau of Standards nondestructive evaluation nondestructive inspection nondestructive testing

•

Nd: YAG

•

nm

•

neodymium: yttrium-aluminum-garnet (laser)

•

nanometer

•

NMR

•

No.

•

NTSC

•

nuclear magnetic resonance

•

number

•

•

OD

•

Oe

•

National Television Systems Committee outside diameter

• •

OSHA

•

oz

•

p

•

P

•

Pa

•

PC

•

Occupational Safety and Health Administration

•

ounce

•

probability

•

pressure

•

pascal

•

printed circuit

•

PCB

•

pH

•

printed circuit board

•

negative logarithm of hydrogen ion activity

•

precipitation hardening

•

PH

•

PIND

•

PLC

• •

•

P/M

•

PMT

• •

•

POD

•

POFA

•

ppm

• •

•

ppt

•

psi

probability of false alarm

pounds per square inch

r

Ra

probability of detection

•

•

•

photomultiplier tube

parts per trillion

Q

R

powder metallurgy

•

•

•

programmable logic control

parts per million

psig

R

particle impact noise detection

•

•

•

oersted

•

gage pressure (pressure relative to ambient pressure) in pounds per square inch

•

calibrated leakage flow

•

number of levels of each factor; original pixel value; radius; range

•

roentgen

•

resistance; reflection coefficient; rise time; gas constant; radius

•

arithmetic average roughness

•

effective shunt resistance

•

Rs

•

rad •

•

radac

•

rbe

radiation absorbed dose

•

rapid digital automatic computing

•

relative biological effectiveness

•

ref

•

rem

•

RF

• •

roentgen equivalent man; remainder

•

radio frequency

•

RFEC

•

RGB

• •

•

RHM

•

RMM

•

rms

• • •

•

ROC

•

RSSS

• •

•

RTR

•

s

•

S

•

S

•

S

reference

remote-field eddy current red, green, blue roentgen per hour at one meter roentgen per minute at one meter root mean square relative operating characteristic radiographic standard shooting sketch

•

real-time radiography

•

second

•

siemens

•

sensitivity; speed; spacing

•

signal input

•

SAE

•

SAM

•

SAT

• • •

•

SDD

•

SDP

•

SEC

• • •

•

SEM

•

sfm

• •

Society of Automotive Engineers scanning acoustic microscopy structural assessment testing source-detector distance standard depth of penetration secondary electron-coupled (vidicon) scanning electron microscopy surface feet per minute

•

SiIMPATT

•

SIT

•

SLAM

• • •

•

SLR

•

S/N

•

• •

SOD

silicon impact avalanche transit time silicon intensified target scanning laser acoustic microscopy single lens reflex signal-to-noise (ratio)

• •

SOF

•

SPC

•

SPDT

• •

•

Sv

•

std

•

t

•

T

•

T

•

sievert

•

standard

•

time; thickness

•

tesla

•

temperature; transmission coefficient

•

TE

•

TEM

•

tape-automated bonding

•

transverse electric (mode)

•

•

TEM

•

TM

• •

•

TIG

•

TOF

•

tsi

•

tungsten inert gas (welding)

tons per square inch

•

geometric unsharpness

UCL •

•

ULCE

•

USAF

•

UST

• •

V

transverse magnetic (mode)

•

•

•

transverse electromagnetic (wave)

time of flight

Ug

v

transmission electron microscopy

•

•

•

statistical process control single pole double throw

TAB

UT

shape and orientation factor of the flaw

•

•

•

source-object distance

upper control limit unified life cycle engineering United States Air Force

•

underground storage tank

•

ultrasonic testing

•

degree of freedom; velocity

•

volume; velocity; variance

•

VHS

•

VHSIC

• •

•

VLSI

•

vol

very high speed very high speed integrated circuit

•

very large scale integration

•

volume

•

W

•

WORM

•

• •

yr

•

Z

•

°

•

°C

•

°F

• •

÷

•

=

• • •

• • • •

•

sample mean

•

yttrium iron garnet

•

year

•

impedance; atomic number; standard normal distribution

•

angular measure; degree

•

degree Celsius (centigrade)

•

degree Fahrenheit

•

direction of reaction

•

divided by

•

equals

•

approximately equals

•

not equal to

•

identical with

•

greater than

•

much greater than

•

greater than or equal to

•

infinity

•

is proportional to; varies as

•

integral of

•

less than

•

much less than

•

less than or equal to

•

maximum deviation

•

minus; negative ion charge

•

diameters (magnification); multiplied by

•

multiplied by

•

per




•

•

•

YIG

•

•

watt

/

•

%

•

+

• • • • • • • • • • • • •

•

percent

•

plus; positive ion charge

•

square root of

•

approximately; similar to

•

partial derivative

•

thermal diffusivity; angle of incidence

•

angle of refraction

•

shear strain

•

skin depth

•

change in quantity; increment; range; phase shift

•

normal strain; emissivity; dielectric coefficient

•

viscosity

•

angle

•

wavelength; mean free path

•

linear attenuation coefficient; magnetic permeability; the mean (or average) of a distribution

•

in.

•

m

• • • • • • • • •

o • • •

•

microinch

•

micron (micrometer)

•

Poisson's ratio; frequency

•

pi (3.141592)

•

density; resistivity

•

stress; standard deviation; electrical conductivity

•

summation of

•

shear stress

•

angle of refraction

•

circular frequency (angular velocity)

•

ohm

Greek Alphabet

A, B, ,

•

alpha

•

beta

•

,

•

E,

•

Z,

•

H,

•

,

•

•

gamma

•

delta

•

epsilon

•

zeta

•

eta

•

theta

•

iota

•

kappa

•

lambda

•

mu

•

nu

•

xi

I,

•

K,

•

,

•

M,

•

N,

•

,

•

O, o

• •

• • •

•

pi

•

rho

•

sigma

•

tau

•

upsilon

•

phi

•

chi

•

psi

•

omega

P, , T,

• •

omicron

,

• •

•

, , X, , ,