Delamination in Wood, Wood Products and Wood-Based Composites
Voichita Bucur Editor
Delamination in Wood, Wood Products and Wood-Based Composites
123
Editor Dr. Voichita Bucur CSIRO Clayton Laboratories Materials Science and Engineering Bayview Avenue 3168 Clayton Victoria Australia
[email protected] ISBN 978-90-481-9549-7 e-ISBN 978-90-481-9550-3 DOI 10.1007/978-90-481-9550-3 Springer Dordrecht Heidelberg London New York Library of Congress Control Number: 2010938326 © Springer Science+Business Media B.V. 2011 No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work. Cover image: Intra-ring internal checking in sample (100 × 50 mm – width × thick) of regrowth Victorian Ash (Eucalyptus delegatensis or E. regnans). Photo taken by Philip Blakemore. Printed on acid-free paper Springer is part of Springer Science+Business Media (www.springer.com)
Foreword
It is with great pleasure that I prepare this foreword. The senior author, Professor Voichita Bucur, is one the preeminent wood scientists in the world today. She is well known for her excellent research on acoustics, especially the acoustic properties of wood and wood-based materials. Her previous books Acoustics of Wood and Nondestructive Characterization and Imaging of Wood are outstanding reference documents; they provide a summary of much of the world’s research and development efforts in these two important technical areas. Professor Bucur has contacted widely respected technical authorities and asked them to prepare chapters dealing with various aspects of the formation and detection of separations and delaminations in wood-based materials. P. Blackmore – CSIRO Australia, S Blumer Holzinnovationzentrum, Austria, G Daian Melbourne University, Australia, BSW Dawson – SCION New Zealand, F Divos – Faculty of Wood Science Sopron, Hungary , L. Donaldson – SCION New Zealand, T. Gereke ETH Zürich, Switzerland, P.J. Gustafsson Lund University Sweden, N. Haque – CSIRO Australia, CL Huang – Weyerhauser USA, S. KazemiNajafi – Tarbiat Modares University, Iran, C. Mueller – ETH Zürich, Switzerland, J. Neuenschwander – Empa Switzerland, P. Niemz – ETH Zürich, Switzerland, K. Persson – Lund University Sweden, M.S.J. Sanabria Empa, Switzerland, U. Sennhauser Empa, Switzerland, A. P. Singh – SCION New Zealand, all graciously agreed and provided excellent technical contributions. This book is organized into three parts. Part I, General Aspects, presents much needed basic information, including terminology, the theoretical basis for evaluation of delamination in wood and wood-based materials, and mechanical stress development in the woody cell wall in response to various stressors. A vibrationbased approach is proposed to evaluate delamination with ultrasonics or with low frequency vibrations. Crack initiation and growth of delamination is studied with a fracture mechanics approach. A theoretical model for collapse recovery is proposed. Part II, Methodology for Delamination Detection and Factors Inducing and Affecting Delamination, begins by examining a variety of methods for detecting delamination in wood products, then delves into discussion of the formation of delamination or separations at several levels – from the microscopic, anatomical level within solid wood sections to examination of the interface of wood and surface
v
vi
Foreword
coatings. The techniques presented for observing separations include confocal laser scanning microscopy, light microscopy, scanning electron microscopy and ultrasonics. Excellent discussions of delamination caused by moisture induced stresses, including those that form during the drying of wood and lumber products and those observed with weathered wood surfaces, are included. Part III, Delamination in Different Products, focuses on practical aspects of delamination in a wide range of wood products. An excellent discussion of the industry’s perspective is presented. Practical discussions dealing with the formation, detection, and performance problems associated with delamination in trees, logs, laminated panels, composites, glued laminated timbers, and parquet floors are presented in detail. The authors prepared this book to serve as a primary reference on subject of delamination in wood-based materials and products. It was prepared to provide a concise source of information on the topic to manufacturers and users of wood products, as well as research scientists. It was made possible through the efforts of dedicated scientists who spent countless hours in laboratories developing technical information on this important subject. This book is a tribute to their efforts and a significant contribution. This book is a significant contribution to the wood science and technology literature. Professor Bucur has completed another significant contribution to the wood science literature. Project Leader USDA Forest Products Laboratory October, 2009
Robert J. Ross, Ph.D.
Preface
Delamination occurs in all man made composite materials as well as in natural composites like wood, bones or rocks. Many groups of specialists with widely different backgrounds and interests need knowledge of factors influencing delamination in wood, wood products and wood based composites. I was amazed with the lack of information on the subject and particularly with the way in which the available information is scattered in the literature. Out of this amazement arose the idea to write and edit this book. Part I of the volume deals with general aspects of delamination, the terms used for defining delamination in wood science and technology and with the theoretical aspects in the evaluation of delamination. Part II is directed at the methodology developed for delamination detection. Factors that induce and affect delamination are analyzed. Part III is a study of delamination in different products. Extensive reference is made to the literature. An attempt has been made to select the most important references for the corresponding chapter. Thus, for any given topic, it should be easy for the reader to quickly acquaint himself with what has been done by looking up the listed references. It is also the hope of the authors that this volume will be a valuable source of information for the practitioner who mostly deal with the design or evaluation of structures subjected to delamination. In recent years manufacturers are becoming more aware of the importance of delamination and other factors that affect the performance of their finished products. Thus there is an evident need for this type of book. Experts called upon to render opinions on structure safety are faced with not only the daunting task of discovering and quantifying structural defects such as delamination, but also translating those observations into the probability of failure and determining levels of “unacceptable risk”. Even though the mechanics of wood failure is better understood today than two decades ago, and the tools for nondestructive identification of defects are more accurate and powerful, the fact remains that deciding what level of defect represent an “unacceptable risk” continues to be a subjective judgment. This is particularly true for structures with significant but not severe defects such as delamination and on sites that present high levels of risk (i.e. snow). The bibliography of this book is intended to be comprehensive and we hope, an important contribution of this book (near 1000 references) is to accurately identify vii
viii
Preface
the evolution of ideas in the last decades. All references cited in the text are included in the reference section at the end of each chapter. At present, no comparable book exists covering the entire field of delamination in wood, wood products and wood based composites. The editor would like to firstly acknowledge the contributions by colleagues acting as authors of the individual chapters, who gave their time and energy to prepare this excellent text. I would like to express my sincere thanks to all colleagues and organizations that have made possible the publication of this volume, to the CSIRO – Commonwealth Scientific and Industrial Research Organisation – Australia and SCION- Forest Research Institute, New Zealand, who supported this idea. In preparing such a text it is very difficult to acknowledge all the help given to the editor. I am indebted to the three main scientific communities, wood science, mechanical and acoustical communities who have undertaken research and development that is reflected in the cited publications. This book encompasses a variety of recent research result, a number of unpublished results and refinement of older material. This book would certainly not have been possible without the help of my colleague Nick Ebdon, CSIRO – Clayton, who work very hard on the preparation and formatting all figures. Last but not least, I would also thank my family and my Australian friends who followed with interest and enthusiasm the progress of the manuscript of this book. Working for this book was for me an extraordinary opportunity to discover the natural splendors of Australia and the atmosphere of this country, which is a proud modern civilization. Melbourne, Victoria October 2009
Voichita Bucur
Acknowledgements
Permission for the figures cited in this book have been granted by Copyright Clearance Center, (http://www.copyright.com), by different organisations and colleagues cited in the corresponding chapters of this book. The authors are very thankful for their kind permission to reproduce figures. As editor of this book, I own special thanks to Ms Danila Durante, Information Specialist, CSIRO Australia, Information Management & Technology Division, in Melbourne for numerous hours spent together for copyright permissions with the new electronic system required by Copyright Clearance Center. Many, many thanks are also addressed to Ms Bee Thia, Information Specialist, CSIRO Australia, Information Management & Technology Division, for her continuous and enthusiastic help in collecting documents and books cited in this volume. Melbourne, Australia
Voichita Bucur
ix
Contents
Part I
General Aspects
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Voichita Bucur
3
2 Terms for Delamination in Wood Science and Technology . . . . . Voichita Bucur
17
3 Delamination Detection – A Vibration-Based Approach . . . . . . Voichita Bucur
33
4 Initiation and Growth of Delamination in Wood and Wood-Based Composites, a Fracture Mechanics Approach . . Voichita Bucur 5 A Theoretical Model of Collapse Recovery . . . . . . . . . . . . . . Philip Blakemore Part II
51 101
Methodology for Delamination Detection and Factors Inducing and Affecting Delamination
6 Delamination of Wood at the Microscopic Scale: Current Knowledge and Methods . . . . . . . . . . . . . . . . . . . . . . . . Lloyd Donaldson
123
7 Probing the Wood Coating Interface at High Resolution . . . . . . Adya P. Singh and Bernard S.W. Dawson
145
8 Delamination in Timber Induced by Microwave Energy . . . . . . Georgiana Daian
159
9 Delaminations Induced by Weathering in Wood and Wood-Based Composites Panels . . . . . . . . . . . . . . . . . Voichita Bucur 10
Delamination in Timber Induced by Drying . . . . . . . . . . . . . Nawshad Haque
173 197
xi
xii
Contents
Part III
Delamination in Different Products
11
Industry Prospective of Delamination in Wood and Wood Products Chih Lin Huang
215
12
Internal Checking During Eucalypt Processing . . . . . . . . . . . Philip Blakemore
237
13
Acoustic Tomography for Tension Wood Detection in Eucalypts . . Voichita Bucur
255
14
The Hygroscopic Warping of Cross-Laminated Timber . . . . . . . Thomas Gereke, Per Johan Gustafsson, Kent Persson, and Peter Niemz
269
15
Acoustic Emission Activity Induced by Delamination and Fracture of Wood Structure . . . . . . . . . . . . . . . . . . . Voichita Bucur
287
Delamination Detection in Wood – Based Composites Panel Products Using Ultrasonic Techniques . . . . . . . . . . . . . . . . Voichita Bucur and Saeed Kazemi-Najafi
307
Delamination Evaluation of in-Service Glulam Beams and other Structural Members Via Ultrasonics . . . . . . . . . . . Ferenc Divos
353
16
17
18
Moisture Induced Stresses and Deformations in Parquet Floors . . Samuel Blumer, Erick Serrano, Per Johan Gustafsson, and Peter Niemz
19
Glue Line Nondestructive Assessment in Timber Laminates with an Air-Coupled Ultrasonic Technique . . . . . . . . . . . . . . Sergio J. Sanabria, Christian Müller, Jürg Neuenschwander, Peter Niemz, and Urs Sennhauser
20
365
379
From Present Researches to Future Developments . . . . . . . . . Voichita Bucur
389
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
393
Contributors
Philip Blakemore Division of Materials Science and Engineering, CSIRO, Clayton laboratories, Clayton South, VIC 3169, Australia,
[email protected] Samuel Blumer b-h-e GmbH, Holzinnovationszentrum 1a, 8740 Zeltweg, Austria,
[email protected] Voichita Bucur CSIRO, Materials Science and Engineering Div. Bayview Avenue, Clayton, Victoria 3168, Australia,
[email protected] Georgiana Daian The University of Melbourne, Department of Forest and Ecosystem Science, Melbourne, VIC 3010, Australia,
[email protected] Bernard S.W. Dawson Wood and Biofibre Technologies, Scion Rotorua Te Papa Tipu Innovation Park, 49 Sala Street Whakarewarewa, 3010, Bay Of Plenty, New Zealand,
[email protected] Ferenz Divos Faculty of Wood Science, University of West Hungary, Sopron, Hungary,
[email protected];
[email protected] Lloyd Donaldson Bioproduct Development, Scion - Next Generation Biomaterials, 49 Sala St. Rotorua, Private Bag 3020, Rotorua 3046, New Zealand,
[email protected] Thomas Gereke Composites Group, Department of Civil Engineering & Department of Materials Engineering, The University of British Columbia, 6250 Applied Science Lane, Vancouver, B.C., Canada V6T 1Z4,
[email protected] Per Johan Gustafsson Division of Structural Mechanics, Lund University, P.O. Box 118, SE-221 00 Lund, Sweden,
[email protected] Nawshad Haque Division of Minerals, CSIRO Clayton, Bag 312, Clayton South, VIC 3169, Australia,
[email protected] Chih Lin Huang Weyerhaeuser Technology Center, 32901 Weyerhaeuser Way S, Federal Way, WA 98001, USA,
[email protected] xiii
xiv
Contributors
Saeed Kazemi–Najafi Wood & Paper Science & Technology Department, Tarbiat Modares University, Noor, Iran,
[email protected] Christian Müller Institute for Building Materials, Wood Physics, ETH Zürich, Schafmattstrasse 6, CH-8093, Zürich, Switzerland,
[email protected] Jürg Neuenschwander Electronics/Metrology/Reliability Laboratory, Swiss Federal Laboratories for Materials Science and Technology, Empa, Überlandstrasse 129, CH-8600, Dübendorf, Switzerland,
[email protected] Peter Niemz Institute for Building Materials, Wood Physics, ETH Zürich, Schafmattstrasse 6, CH-8093, Zürich, Switzerland,
[email protected] Kent Persson Division of Structural Mechanics, Lund University, P.O. Box 118, SE-221 00 Lund, Sweden,
[email protected] Robert J. Ross Forest Products Research Laboratory One Gifford Pinchot Drive Madison, Madison, WI 53726, USA,
[email protected] Sergio J. Sanabria Electronics/Metrology/Reliability Laboratory, Swiss Federal Laboratories for Materials Science and Technology, Empa, Überlandstrasse 129, CH-8600, Dübendorf, Switzerland,
[email protected] Urs Sennhauser Department of Electronics/Metrology/Reliability Laboratory, Swiss Federal Laboratories for Materials Science and Technology, Empa, Überlandstrasse 129, CH-8600, Dübendorf, Switzerland,
[email protected] Erik Serrano University of Vaxjo, Lucklings plats 1 SE 35195 Vaxjo, Sweden,
[email protected] Adya P. Singh Wood and Biofibre Technologies, Scion Te Papa Tipu Innovation Park, Rotorua 3010, New Zealand,
[email protected] Part I
General Aspects
Chapter 1
Introduction Voichita Bucur
Contents 1.1 Background . . . . . . 1.2 Solid Wood . . . . . . 1.3 Wood-Based Composites 1.4 Summary . . . . . . . References . . . . . . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
. . . . .
3 5 9 12 13
1.1 Background In order to improve the quality of mass produced wood-based composites and in order to undertake quality assessment of adhesive interfaces in these materials it is first necessary to develop the theoretical basis describing both qualitatively and quantitatively, the quality parameters of the composite, and secondly to develop new non-destructive techniques for their testing and evaluation. Mechanical integrity of interfaces in wood-based composites plays a major role in determining the serviceability of structures and their components. New advanced materials (i.e. parallel-strand lumber, laminated veneer lumber, etc.) are designed with specialty interfaces to increase fracture resistance of wood-based composite materials and to accommodate residual stresses. Of particular note is that the mechanical properties of wood-based composites, used mainly in civil engineering, may degrade severely in the presence of damage, often with tragic consequences. Therefore damage detection is a very important issue in the context of structural health monitoring for mechanical engineering infrastructure with elements in wood and wood-based composites. Wood-based composites are complex materials exhibiting important anisotropic properties. Commonly observed damage in these materials are: delamination V. Bucur (B) CSIRO, Materials Science and Engineering Div. Bayview Avenue, Clayton, Victoria 3168, Australia e-mail:
[email protected] V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_1,
3
4
V. Bucur
Fig. 1.1 Delamination observed on cross sections of Douglas-fir laminated lumber. Note delamination occurs predominantly in wood elements in the direction of medulary rays, frequently starting or finishing at the interface between the earlywood and latewood (Vick and Okkonen 2000, Figure 5a)
between plies, debonding of wood–adhesive layers, or wood fibre fracture. Delamination, which is a debonding of two adjoining layers in the laminated wood-based composite, is probably the most frequently observed damage. Delamination can occur at several scales: Fig. 1.1 shows the cross section of Douglas – fir lumber laminates with macroscopic delamination, while at a submicroscopic scale, delamination can be observed between the S1 and S2 layers in spruce latewood tracheids, as can be seen in Fig. 1.2. Delamination may result from manufacturing errors, by imperfect bonding, by separation of adjoining piles, etc., or, during in service loading such as by accidentally excessive loading produced for example by snow or, by fatigue in cyclical environmental conditions of temperature and humidity.
Fig. 1.2 Delamination in spruce latewood tracheids between S1 and S2 layers (Zimmermann et al. 1994, Figure 3)
1
Introduction
5
As noted by Garg (1988) many years ago, the prediction of delamination in composites is a challenge for both scientists and manufacturers. This is due to the large number of parameters involved in the design of composites and, on the other hand, to the complexity of the stress state which leads to the initiation and propagation of delamination. For the initiation aspect of delamination, the tolerance prediction is based on semi-empirical criteria, such as point-stress or average stress criteria. Due to the use of such criteria, industries are led performing numerous tests in order to ensure the safety margins for delamination failure are not exceeded. The non-propagation certification relies on fracture mechanics analyses, which are very complex and introduce difficulties for the characterization of the initial delamination pattern (Srinivasan 1996; Murata and Masuda 2006). The last 30 years there have been several important advances toward a better understanding of the mechanics of laminated composites and of the damage mechanisms, because of their intensive utilisation in aerospace engineering. This progress concerns the analysis and identification on the micro, macro and meso scales, as well as the development of advanced anisotropic material models. To be able to rely on computational models, both academics and manufacturers recognize that a prerequisite is to develop a detailed material model with a clear identification procedure and to validate this model by means of representative experimental tests. The physics of delamination is governed by interactions among different damage mechanisms, such as fibre breakage, transverse microcracking and debonding of the adjacent layers of the cell wall. To understand the physics of delamination in composite biological materials and more specifically in wood, wood based products and wood-based composites, it is necessary to have detailed knowledge about the microstructure of these materials. As noted by Kelly (1989) in the Concise Encyclopedia of Composite Materials, “plant cells are a good example of laminated composite material; the shape of the cells is roughly tubular with various laminae of cellulose microfibrils glued together to form a wall. Each lamina has a characteristic fibre orientation which can be random, cross-helical or single-helical. . . . . . These biomaterials are grown under stress; this means that the loading conditions of the structure as a whole can be used effectively as blueprints for the most efficient use of fibre reinforcement. By their very nature, natural fibrous composites are better materials in tension than in compression and their use in many applications is often limited by this fact. The excess of tensile strength available can be profitably used to pre-stress in tension the regions of the structure which are more vulnerable into compressive loads. Also the presence of water as compression members will result in lighter structures”.
1.2 Solid Wood Wood is a natural fibrous, layered composite which exhibits a remarkable combination of properties related to strength, stiffness and toughness (Vincent and Currey 1980; Schniewind 1989). As noted by Schniewind (1981) “wood is composed from
6
V. Bucur
a complex aggregation of cells, of tubes shape, which during the life of the tree had biological function. The structural features of wood are oriented following the principal directions of growth of the tree, namely longitudinal-parallel to the axis of the tree, radial and tangential – versus the annual rings” Several models have been proposed to represent wood structure in relation to its mechanical behaviour, starting with Price (1929), who modelled the cell structure as an array of parallel cylindrical tubes, of isotropic structure, oriented in the stem direction. Another version, proposing also a tubular model, useful for modelling the cell wall as a laminated composite material is presented in Fig. 1.3. A softwood or conifer wood cell is essentially a hollow tube of about 30 μm diameter with a multi layered laminated wall composed generally from four layers – primary wall, S1 , S2 and S3. The S2 layer, is the principal load bearing component of the cell wall and is close to 80% of the total cell wall area. It contains cellulose components in the form of microfibrils of about 10–20 nm in diameter. In most cases the microfibrils lie at an angle to the cell axis and form a steep helix at an angle, ranging between 0◦ and 25◦ and 0◦ and 50◦ for hardwood and softwood respectively. Fibres with low microfibril angle (10◦ ) posse high tensile strength (400 MPa) and low elongation (1%). The cells are parallel to the grain direction and are bonded to each other by an amorphous matrix containing mostly lignin. Nearly 90% of the cells are aligned in one direction forming a honeycomb structure with highly anisotropic mechanical properties. The alternation of spring and summer growth (earlywood and latewood layers in the annual ring) in softwood and ring porous hardwood species from temperate climates produces well known ring patterns which introduce a further element of complexity.
Fig. 1.3 Layered structure of the cell wall modelled as a laminated composite material (Mark 1967, Figure 1-7)
1
Introduction
7
A more complex model was proposed by Mark (1981), and consists of a matrix and framework. The corresponding mechanical properties of the cell wall material can be derived from the natural polymer constituent (cellulose, hemicellulose and lignin) properties by the rule of mixture. The stiffness and strength of cellulose itself are considerable, the theoretical value for Young’s modulus and tensile strength being in the order of 250 GPa and 25 GPa (Mark 1967) respectively. Wood mechanical properties are considerably inferior to those of pure cellulose. Figure 1.4 shows the degradation of Young’s modulus from cellulose to wood. To E was analysed for several situations and the illustrate this aspect, the ratio σrupture smaller the ratio, the better the material will be in resisting crack propagation. In an ideal solid this ratio is in the order of 10, however this ratio is about 100 in longitudinal anisotropic direction of wood. The reduction of the Young’s modulus E from cellulose to wood is due to largely to the very complex structural arrangement of this material in which the microfibrillar angle plays a very important role. The development of computation techniques in the last 25 years, and the progress achieved in mechanical characterisation of solids in general and of composite materials in particular, affected positively the development of modelling of the wood structure. Gibson and Ashby (1988) proposed a cellular structure model with hexagonal cell shape and used for calculation the principles of cellular solid mechanics. Some improvements of this approach were given by Kahle and Woodhouse (1994) and Watanabe et al. (2000, 2002), which considered the cell wall material as transversely isotropic. Significant progress in Wood Science has been achieved using multiscale models which were elaborated by using three-dimensional finite element simulation of representative softwood related cellular models. In addition data related to the microstructural characteristics such as the micrifibril angle and the chemical composition of the cell wall such as lignin, hemicelluloses, water and crystalline cellulose were also integrated into their models (Harrington et al. 1998; Astley et al. 1998; Yamamoto 1999; Persson 2000; Watanabe and Norimoto 2000; Yamamoto et al. 2005; Hofstetter et al. 2005, 2006; Fritsch and Hellmich 2007). Using the experimental observations of wood behaviour at different scales, Hofstetter et al. (2007) proposed a very original approach considering simultaneously the continuum mechanics for the solid-type behaviour of the cell wall and on the other hand, the unit cell method, for the plate-type behaviour of the softwood microstructure. It was stated that the activation of different load-carrying mechanisms of cellular structure depends on the loading state of wood, such as for example: – the plate-type bending and shear deformations of the cell walls which are dominant in tangential direction, when the transverse shear loading and longitudinal compression straining are applied on solid wood specimens. – the solid-type deformations are dominant in longitudinal and radial directions when longitudinal shearing loading straining are induced on wood specimens.
8
V. Bucur
Fig. 1.4 Schematic representation of the degradation of mechanical properties of wood (expressed by Young’s modulus) compared to those of pure cellulose (Jeronimidis 1980, Figure 2)
At a cellular scale the plate-like deformation modes were studied combining random/periodic multi-step homogenisation with corresponding values obtained from continuum micromechanics modeling. The average predictive capacity of this model is low, about 8%, with very large variations depending on the value of the elastic constants. The highest errors were observed on GRT (error can be as high as 290%) and on Poisson’s ratios (error of about 75%). It is very likely that the
1
Introduction
9
predictive capacity of this model could be substantially improved by using more accurate values of the elastic constants at a microscopic scale, which can be obtained with the development of specific acoustic microscopic technique as suggested by Bucur (2003). All these studies related to the modelling of wood structure clearly suggest that delamination can occur between different layers at submicroscopic, microscopic and macroscopic structural levels.
1.3 Wood-Based Composites With regards to the wood-based composites, the mechanical behaviour of two groups of products must be analysed: the laminated wood products such as glulam, plywood, laminated veneer lumber (LVL), parallel-strand timber (PSL), structural particleboard, oriented strandboard (OSB), the fibre-based products such as fibreboard particleboard, paper and fiber reinforced composite such as fibre-cement boards, carbon fibre-reinforced plywood, and wood and glass-fibre composites, paper, etc. Performance criteria for wood-based composites relate directly to product end use. Laminated products are frequently used for structural purposes. This requires consideration of engineering strength needs, safety and short and long term response of the material to the service environment. Structural, exterior-grade products have the most demanding bond-quality requirements, since glue line failure could be catastrophic to these structures. In these situations glue line strength, durability and reliability must be assured, by computational analysis and bond quality testing programs. Computational models to simulate mechanical behaviour of new woodbased composites are critically needed because of cost-effectiveness. The effects of varying raw material characteristics on the mechanical properties of prospective new products can be thoroughly analysed. The intensive and expensive bond quality testing programs also can be improved by modeling. The factors affecting the quality of adhesion in wood-based composites are related to the heterogeneous and anisotropic character of wood reflected in the anatomical characteristics, permeability, density and moisture content, fibre bonding sites, and on the other hand in the nature of adhesives (thermosetting or thermoplastic). As noted by Schniewind (1981) “bond formation depends upon the development of physical and chemical interactions both within the bulk adhesive polymer and at the interface between adhesive and wood. Interactions within the adhesive accumulate to give cohesive strength while the forces between adhesive and wood provide adhesive strength. Both should exceed the strength of the wood allowing substantial wood failure during destructive testing of high-quality bond”. Optimum bond formation requires intimate contact between adhesive and wood substrates to ensure macromolecular interaction over a large area. Different techniques (X-ray, NMR, microindentation, etc.) were developed for the mechanical characterisation of the wood-adhesive interface. Figure 1.5 shows the light microscopy image of a spruce parallel-strand lumber specimen which contains fractured
10
V. Bucur
Fig. 1.5 Transverse section of spruce parallel-strand lumber which contains fractured and delaminated zones. The arrows indicate the zones tested via nanoindentation (Konnerth and Gindl 2006, Figure 1)
and delaminated zones tested with the nanoindentation technique developed by Konnerth and Gindl (2006). Modelling of the mechanical behaviour of laminated wood composites for predicting elasticity and strength has been reported for more than 30 years in numerous articles. A small snapshot of these include: Hunt and Suddarth (1974) who predicted the Young’s modulus and the shear modulus of medium-density flakeboard, Okuma (1976) studied the plywood properties influenced by the glue line, Gerrard (1987) proposed an equivalent orthotropic elastic model for the properties of plywood, Shaler and Blakenhorn (1990), Wang and Lam (1998) or Lee and Wu (2003) predicted the mechanical properties of oriented flakeboard. The mechanical behaviour of laminated veneer lumber, LVL, has been studied by Bejo and Lang (2004), Castro and Paganini (2003), Hata et al. (2001), Kamala et al. (1999), Lang et al. (2003), Park and Fushitani (2006). Finite element modelling of laminated wood composites as a multilayer system was proposed by several authors (Triche and Hunt 1993; Suo and Bowyer 1995; Clouston et al. 1998; Morlier and Valentin 1999; Nafa and Araar 2003; Wu et al. 2004) for predicting tensile, compression or bending strength and stiffness using failure criteria. Clouston and Lam (2001, 2002) and Clouston (2007) proposed an advanced methodology for analysing the multiaxial stress states in small specimens of parallel wood-strand composites, using a 3D non-linear stochastic finite element model and Monte Carlo simulations. The Tsai-Wu strength theory to predict the ultimate load carrying capacity of a centre point off-axis bending member made from Douglas fir laminated veneer, incorporating the size effect was reported by Clouston et al. (1998).
1
Introduction
11
The behaviour of wood cement composites has been reported from the beginning of there presence on the market, over 70 years ago as low-density and insulation material. Today the cement bonded structural flakeboards offer high, fire, insect and fungal resistance. In addition the quality has improved resulting in better weatherability and acoustic insulation (Lee et al. 1987; Mosemi and Pfister 1987; Fan et al. 1999). References relating to the modelling of mechanical behaviour of fibre-based composites are as abundant as those for laminated wood-based composites, but only several are cited here (Smulski and Ifju 1987; Claisse and Davis 1998; Lopez-Anido et al. 2000; Moulin et al. 1990; Ogawa 2000; Pirvu et al. 2004; Rowlands et al. 1986; Tascioglu et al. 2003; Tsai et al. 2005; Xu 2002; Xu et al. 2005; Chakraborty et al. 2006). Mechanical properties of fibre-based composites are influenced by factors such as: fibre geometry, orientation and distribution, fibres packing in flake of different orientation, random distribution of flakes, moisture content, adhesive-type, etc. Single layer flake models and multilayer mat structures were suggested (Bodig and Jayne 1982; Steiner and Dai 1993; Dai and Steiner1994; Lenth and Kamke 1996) to explain the mechanical behaviour of fibre based composites. Several authors (Ogawa 2000; Tascioglu et al. 2003) reported successful utilisation of hybrid fiberreinforced polymer composites – glulam products for structural applications in civil infrastructures such as beams for bridges stringers, panels for bridge and pier decks. It was noted that these composites are very resistant to delamination tests during accelerated exposure to wetting and drying (Pirvu et al. 2004) Mechanical defibering action produces important structural modifications such as: internal fibrillation observed as a helical wraps of fibres, cell wall delamination, external fibrillation which is the peeling off of the fibrils from the fibre surface, with formation of fines, fibrils or fibrillar lamellae attached to the exterior fibre surface and fibre shortening, depending on the refining conditions, the fibre type – hardwood or softwood – and the pulp type – mechanical or chemical. It is appropriate to mention here that the hydroxyl groups available on the surface of the cellulose molecule are the prime means by which fibres and cement, or other material used as matrix, bond together. The increasing environmental concern about the wastes from wood, wood products, forest waste and construction waste materials has given rise to the development of new or improved technological processes such as the water vapour explosion process. This process rapidly defibrates wood wastes producing a new raw material for novel wood cement composites (Wei et al. 2004). Figure 1.6 shows the interfacial zone between cement and wood fibres, with a delamination of the cell wall near the wood-cement interface. As noted by Schneider (1994) the development of fibre-based composites testing methodology was encouraged as part of the efforts being made to control the performance of low cost building materials for use in developing countries. The renewed interest in producing new composites with wood fibre began almost inadvertently in 1960, and Australia was a leading country in this field as noted by Coutts (2005). In the 21st century a great need still remains to improve the durability of fibre-based products and to study new, cheaper methods of fibre production and
12
V. Bucur
Fig. 1.6 Interface zone between cement and wood fibres, with delamination of the cell wall near the interface wood-cement. (Wei et al. 2004, Figure 1C)
low cost processes. Durability of these products is related to matrix formulations, processing methods and curing regimes. If natural fibre reinforced cement products are to be readily available for low cost housing much research still remains to be conducted for improving the durability of the products.
1.4 Summary Commonly observed damage in wood products and wood-based composites are: wood fibre fracture, delamination between plies or debonding of wood–adhesive layers. Delamination which is probably the most frequently observed damage, may be produced during manufacturing or, during in service loading such as accidental excessive loading produced for example by snow or, by fatigue in highly variable environmental conditions of temperature and humidity. Damage detection in general and delamination in particular is a very important issue in the context of structural health monitoring for mechanical engineering infrastructure with elements in wood and wood-based composites. The development of computational techniques in the last 25 years, and the progress achieved in mechanical characterisation of solids in general and of composites in particular, affected positively the development of the modelling of wood mechanical behaviour in function of its structure. Related studies clearly suggest that delamination in solid wood can occur between different layers of the cell wall at submicroscopic, microscopic and macroscopic structural levels.With respect to wood-based composites, the behaviour of two groups of products has been analysed: the laminated products (plywood, laminated veneer lumber, parallel-strand timber, structural particleboard, oriented strandboard, etc.) and the fibre-based products (fibreboards, fibres-cement composites, carbon fibre-reinforced plywood, particleboard, wood and glass-fibre composites). Finite element modelling of laminated wood composites as a multilayer system was
1
Introduction
13
proposed. More recently analysis of the multiaxial stress states in parallel woodstrand composites, has been proposed using a 3D non-linear stochastic finite element model and Monte Carlo simulations. The development of fibre-based composites testing methodologies must be encouraged as part of the efforts being made to control the performance of low cost building materials.
References Astley RJ, Stol KA, Harrington JJ (1998) Modelling the elastic properties of softwood. Part II: the cellular microstructure. Holz Roh Werkst 56:43–50 Bejo L, Lang EM (2004) Simulation based modelling of the elastic properties of structural composite lumber. Wood Fiber Sci 36:395–410 Bodig J, Jayne BA (1982) Mechanics of wood and wood composites. Van Nostrand Reinhold Company, New York, NY Bucur V (2003) Ultrasonic imaging of wood structure. Proceedings of 5th world conference in ultrasonics, Paris, pp 299–302. http://www.sfa.asso.fr/wcu2003/procs/webside/artickes. Accessed 7 September 2004 Castro G, Paganini F (2003) Mixed glue laminated timber of poplar and eucalyptus grandis clones. Holz Roh Werkst 61:291–298 Chakraborty A, Sain M, Kortschot M (2006) Reinforcing potential of wood p[ulp – derived microfibres in a PVA matrix. Holzforschung 60:53–58 Claisse PA, Davis TJ (1998) High performance jointing systems for timber. Constr Build Mater 12:415–425 Clouston P (2007) Characterization and strength modelling of parallel strand lumber. Holzforschung 61:394–399 Clouston P, Lam F (2001) Computational modelling of strand-based wood composites. ASCE J Eng Mech 127:844–851 Clouston P, Lam F (2002) A stochastic plasticity approach to strength modelling of strand-based wood composites. Compos Sci Techn 62:1381–1395 Clouston P, Lam F, Barrett JD (1998) Incorporating size effects in the Tsai-Wu strength theory for Douglas –fir laminated veneer. Wood Sci Techn 32:215–226 Coutts RSP (2005) A review of Australian research into natural fibre cement composites. Cem Concr Compos 27:518–526 Dai C, Steiner PR (1994) Spatial structure of wood composites in relation to processing and performance characteristics. Part 3. Modelling the formation of multi-layered random flake mats. Wood Sci Techn 28:229–239 Fan M, Dinwoodie JM, Bonfield PW, Breese MC (1999) Dimensional instability of cement bonded particleboard : Behaviour of cement paste and its contribution to the composite. Wood Fiber Sci 31:306–318 Fritsch A, Hellmich Ch (2007) ‘Universal’ microstructural patterns in cortical and trabecular, extracellular and extravascular bone material: Micromechanics – base prediction of anisotropic elasticity. J Theor Biol 244:597–620 Garg CA (1988) Delamination. A damage mode in composite structures. Eng Fract Mech 29(5):557–584 Gerrard C (1987) The equivalent orthotropic elastic properties of plywood. Wood Sci Techn 21:335–348 Gibson LJ, Ashby MF (1988) Cellular Solids. Structure and properties. Pergamon, Oxford Harrington JJ, Booker R, Astley RJ (1998) Modelling the elastic properties of softwood. Part I: The cell – wall lamellae. Holz Roh Werkst 56:37–41 Hata T, Umemura K, Yamauchi H, Nakayama A, Kawai S, Sasaki H (2001) Design and pilot production of a spiral winder for the manufacture of cylindrical laminated veneer lumber. J Wood Sci 47:1105–1123
14
V. Bucur
Hofstetter K, Hellmich C, Eberhardsteiner J (2007) Micromechanical modelling of solid-type and plate-type deformation patterns within softwood material. A review and an improved approach. Holzforschung 61:343–351 Hofstetter K, Hellmich C, Eberhardsteiner J (2006) The influence of the microfibril angle on wood stiffness: A continuum micromechanics approach. Comput Assisted Mech Eng Sci 13: 523–536 Hofstetter K, Hellmich C, Eberhardsteiner J (2005) Development and experimental validation of a continuum micromechanics model for wood. Eur J Mech Solid 24:1030–1053 Hunt MO, Suddarth SK (1974) Prediction of elastic constants of particleboard. Forest Prod J 24(5):52–57 Jeronimidis G (1980) Wood, one of nature’s challenging composite. In: Vincent JFV, Currey JD (eds) “The mechanical properties of biological materials”. Cambridge University Press, London, pp 169–182 Kahle E, Woodhouse J (1994) The influence of cell geometry on the elasticity of softwood. J Mater Sci 29:1250–1259 Kamala BS, Kumar P, Rao RV, Sharma SN (1999) performance test of laminated veneer lumber (LVL) from rubber wood for different physical and mechanical properties. Holz Roh- Werkst 57:114–116 Kelly A (ed) (1989) Concise encyclopedia of composite materials. Pergamon, Oxford Konnerth J, Gindl W (2006) Mechanical characterization of wood-adhesive interphase cell walls by nanoindentation. Holzforschung 60:420–433 Lang EM, Bejo L, Divos F, Kovacs Z, Anderson RB (2003) Orthotropic strength and elasticity of hardwoods in relation to composite manufacture. Part III. Orthotropic elasticity of structural veneers. Wood Fiber Sci 35:308–320 Lee AWC, Hong Z, Phillips DR, Hse CY (1987) Effect of cement /wood ratios and wood storage conditions on hydration temperature, hydration time and compressive strength of wood – cement mixtures. Wood Fiber Sci 19:262–268 Lee JN, Wu Q (2002) In – plane dimensional stability of three-layer oriented strandboard. Wood Fiber Sci 34:77–95 Lee JN, Wu Q (2003) Continuum modelling of engineering constants of oriented strandboard. Wood Fiber Sci 35:24–40 Lenth CA, Kamke FA (1996) Investigations of flakeboard mat consolidation. Part I. Characterizing the cellular structure. Wood Fiber Sci 28:153–167 Lopez-Anido R, Gardner DJ, Hensley JL(2000) Adhesive bonding of eastern hemlock glulam panels with E-glass / vinyl ester reinforcement. Forest Prod J 50, 11/12:43–47 Mark RE (1981) Molecular and cell wall structure of wood. In: Wangaaed FF (ed) Wood: Its structure and properties. Educational Modules for Material Science and Engineering Project. Pensilvania State University, University Park, Pensylvania, USA, pp 43–100 Mark RE (1967) Cell wall mechanics of wood tracheids. Yale University Press, New Haven, Connecticut Morlier P, Valentin G (Eds) (1999) Damage in wood. COST Action E8, Bordeaux Moslemi AA, Pfister S (1987) The influence of cement/wood ration and cement type on bending strength and dimensional stability of wood-cement composite panels. Wood Fiber Sci 19: 165–175 Moulin JM, Pluvinage G, Jodin P (1990) FGRG : Fiberglass reinforced gluelam – a new composite. Wood Sci Techn 24:289–294 Murata K, Masuda M (2006) Microscopic observation of transverse swelling of latewood tracheid: Effect of macroscopic/mesoscopic structure J Wood Sci 52:283–289 Nafa Z, Araar M (2003) Applied data for modelling the behaviour in cyclic torsion of beams in glued-laminated wood: Influence of amplitude. J Wood Sci 49:36–41 Ogawa H (2000) Architectural application of carbon fibers. Development of new carbon fiber reinforced glulam. Carbon 38:211–226 Okuma M (1976) Plywood properties influenced by the glue line. Wood Sci Techn 10:57–68
1
Introduction
15
Park HM, Fushitani M (2006) Effects of component ratio of the face and core laminae on static bending strength performance of three – ply cross – laminated wood panels made with sugi (Cryptomeria japonica). Wood Fiber Sci 38:278–291 Persson K (2000) Micromechanical modelling of wood and fibre properties. Ph D thesis. University of Lund Pirvu A, Gardner DJ, Lopez-Anido R (2004) Carbon fiber – vinyl ester composite reinforcement of wood using the VARTM/SCRIMP fabrication process. Compos Part A 35:1257–1265 Price AT (1929) A mathematical discussion on the structure of wood in relation to its elastic properties. Phil Trans Royal Soc A 228:1–62 Rowlands RE, van Deweghe RP, Laufenberg TL, Krueger GP (1986) Fiber – reinforced composites. Wood Fiber Sci 18:39–57 Schneider MH (1994) Wood polymer composites. State of the Art review Paper. Wood Fiber Sci 26:142–151 Schniewind A (1981) Mechanical behavior and properties of wood. In: Wangaaed FF (ed) Wood: Its structure and properties. Educational Modules for Material Science and Engineering Project. Pennsylvania State University, University Park, Pennsylvania, USA, pp 225–270 Schniewind AP (1989) Concise encyclopedia of wood & Wood-based materials. Pergamon, Oxford Shaler SM, Blakenhorn PR (1990) Composite model prediction of elastic moduli for flakeboard. Wood Fiber 22:246–261 Smulski SJ, Ifju G (1987) Flexural behaviour of glass fiber reinforced hardboard. Wood Fiber Sci 19:313–327 Suo S, Bowyer JL (1995) Modeling of strength properties of structural particleboard. Wood Fiber Sci 27:84–94 Srinivasan AV (1996) Smart biological systems as models for engineered structures. Mater Sci Eng C 4:19–26 Steiner PR, Dai C (1993) Spatial structure of wood composites in relation to processing and performance characteristics. Part I. Rationale for model development. Wood Sci Techn 28:45–51 Tascioglu C, Goodell B, Lopez – Anido R (2003) Bond durability characterization of preservative treated wood and E – glass/phenolic composite interfaces. Compos Sci Techn 63:979–991 Triche MH, Hunt MO (1993) Modelling of parallel-alligned wood strand composites. Forest Prod J 43(11/12):33–44 Tsai M, Chou HC, Xie YM, Li YF, Lin LD (2005) Study on the accelerated aging of CFRP – wood composites. Forest Prod J 24(3):237–246 Vick CB, Okkonen EA (2000) Durability of one-part polyurethane bonds to wood improved by HMR coupling agent. Forest Prod J 50(10):69–75 Vincent JFV, Currey JD (Eds) (1980) The mechanical properties of biological materials. Cambridge University Press, London Wang K, Lam F (1998) Robot – based research on three – layer oriented flakeboards. Wood Fiber Sci 30:339–347 Watanabe U, Norimoto M (2000) Three dimensional analysis of elastic constants of the wood cell wall. Wood Research. Bull. Wood Res. Institute, Kyoto, 87:1–7 Watanabe U, Norimoto M, Morooka T (2000) Cell wall thickness and tangential Young’s modulus in coniferous early wood. J Wood Sci 46:109–114 Watanabe U, Fujita M, Norimoto M (2002) Transverse Young’s moduli and cell shapes in coniferous early wood. Holzforschung 56:1–6 Wei YM, Fujii T, Hiramatsu Y (2004) A preliminary investigation on microstructural characteristics of interfacial zone between cement and exploded wood fiber by using SEM-EDS. J Wood Sci 50:327–336 Wu Q, Lee JN, Han G (2004) The influence of voids on the engineering constants of oriented stranboard: A finite element model. Wood Fiber Sci 36:71–83 Xu J, Widyorini R, Kawai S (2005) Properties of kenaf core binderless particleboard reinforced with kenaff fiber – woven sheets. J Wood Sci 51:415–420
16
V. Bucur
Xu H (2002) Structural characterization of hybrid fiber reinforced polymer – glulam panels for bridge decks. J Comp Constr 6(3):194–203 Yamamoto H (1999) A model of the anisotropic swelling and shrinkage process of wood. Part I: Generalisation of Barber’s wood fiber model. Wood Sci Techn 33:311–325 Yamamoto H, Abe K, Arakawa Y, Okuyama T, Grill J (2005) Role of the gelatinous layer on the origin of the physical properties of tension wood of Acer sieboldianum. J Wood Sci 51:222–233 Zimmermann T, Sell J, Eckstein D (1994) SEM studies on traction – fracture surfaces of spruce samples. Holz Roh-Werkst 52:223–229
Chapter 2
Terms for Delamination in Wood Science and Technology Voichita Bucur
Contents 2.1 General Terms . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Terms for Delamination in Solid Wood . . . . . . . . . . . . . . 2.3 Terms for the Delamination in the Cell Wall . . . . . . . . . . . . 2.4 Terms for the Delamination in Laminated Wood Products . . . . . 2.5 Terms for the Delamination in Wood-Based Fibre and Particle Panels 2.6 General Classification of Delamination . . . . . . . . . . . . . . 2.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
17 18 19 19 25 26 29 29
2.1 General Terms In Material Science, delamination is defined as a sub critical damage to the interfaces between the plies in a laminate composite that causes a reduction in the load carrying capacity of composite (Morris 1992). The terms which describe delamination in wood and wood- based composites are very numerous and often confusing due to a multitude of reasons (the use of terms which were considered inappropriate in recent days, new technologies related to microscopic observation of the structure, etc). A comprehensive understanding of these terms is essential for the uses of wood products under competitive conditions of modern technology. This chapter discusses the terms that refer to delamination in solid wood, in wood cell wall, in laminated products, and in fibrous and particle board wood-based composites.
V. Bucur (B) CSIRO, Materials Science and Engineering Div. Bayview Avenue, Clayton, Victoria 3168, Australia e-mail:
[email protected] V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_2,
17
18
V. Bucur
2.2 Terms for Delamination in Solid Wood Terms which express delamination in solid wood have been defined in very well known reference textbooks (Kollmann and Cote 1968; Panshin and de Zeeuw 1980) and standards under the label of defects which develop in wood after it has been cut. In what follows we quoted the terms as referred in ASTM D9 – Check – a separation of wood along the fibre direction that usually extends across the rings of annual growth, commonly resulting from stress set up in wood during seasoning. ◦ End check – a seasoning check occurring on the end of a board or other piece of wood. ◦ Heart check – a check that extends across the growth layers in one or more directions from the pith toward, but not to, the surface of a piece of wood. A synonym is pith check ◦ Roller check – a crack in the wood structure caused by a piece of cupped lumber being flattened between machine rollers ◦ Star check – a heart check in which the separation extends in more than one direction from the pith ◦ Surface check – a check occurring on the surface of a piece of wood, usually on the tangential face not extending through the piece. ◦ Through check – a check that extends through a piece of wood, or from a surface to the opposite or to an adjoining surface. – Collapse – the flattening of single cells or rows of cells during drying or pressure treatment of wood, characterized by a caved or corrugated appearance – Cracks see shake – Cross Break – a separation of the wood cells across the grain. Such breaks may be due to the internal stress resulting from unequal longitudinal shrinkage or external forces. – Honeycombing – in lumber and other wood products, is the separation of the fibers in the interior of the piece, usually along the rays. The failures often are not visible on the surface, although they can be the extensions of surface and end checks. – Shake – a longitudinal separation of the wood. Generally two forms of shake are recognized, although variations and combinations may be used in industrial definitions ◦ Heart shake – a shake that starts out at or near the pith and extends radially. Synonyms are heart cracks, rift crack. A heart shake in which several radial cracks are presented is termed a star shake ◦ Ring shake – shake occurring in standing trees, in the plane of the growth rings in the outer position of the latewood for partial or entire encirclement
2
Terms for Delamination in Wood Science and Technology
19
of the pith, occasionally moving radially to an adjacent latewood ring. A synonym is “cup shake”. Meyer and Leney (1968) described ring shakes from standing conifer trees as compound middle lamella failures, usually in latewood, with loose fibres and deposites of extraneous material on their shake surface. ◦ Handsplit and resawn shakes – a shake having a split face and a sawn back ◦ Tapersplit shake – a shake having two split faces and a natural shingle like taper ◦ Straightsplit shake – a shake having two split faces and with no pronounced taper – Split a separation of the wood parallel to the fiber direction, due to the tearing apart of the wood cells.
2.3 Terms for the Delamination in the Cell Wall The cell wall has a typical layered structure composing three main layers – S1 , S2 , S3 – of variable thickness, in the micron (μm) range, composed of cellulosic microfibrils embedded in an amorphous matrix. Delamination can occur between layers as well as inside the same layer, and can be produced by growth related defects in living trees or can be a defect which develop in wood after it has been cut. Table 2.1 synthesises the terms related to the cell wall structure, describing wood delamination at the submicroscopic level. The spectrum of terminology that has been used in profusion in the numerous articles cited in this table need to be put in concordance with the mechanical approach proposed in Chapters 3 and 4 of this book, for the description of phenomena related to the delamination in wood and wood – based composites. On the other hand, as noted by Wilkins (1986) the future nomenclature “needs to remain flexible and include further terms derived from the development of the tools for wood structure inspection”. One can speculate about the contribution of new technologies for higher resolution microscopy in relation to wood ultrastructure which influence its mechanical behaviour.
2.4 Terms for the Delamination in Laminated Wood Products Structural laminated products include plywood, various composites of veneer and of wood based laminates such as laminated veneer lumber, glued laminated lumber, wood fibre-reinforced polymer composites, etc. Plywood as defined in ASTM D 1038 – as “usually a crossbanded assembly made of layers of veneer or veneer in combination with a lumber core or other woodbased panel material jointed with an adhesive. Plywood is generally constructed of an odd number of layers with grain of adjacent layers perpendicular to one another. Outer layers and all odd-numbered layers generally have the grain direction oriented parallel to the long dimension of the panel”.
20
V. Bucur Table 2.1 Nomenclature of cell wall deformation as referred in publications until 2007
Term
Author
Buckle
Robinson (1920)
Buckling
Scurfield et al. (1972) Kucera and Bariska (1982)
Buckling
Buckling of the cell wall
Wilkins (1986)
Cell wall crinkle
Green (1962)
Cell wall fold
Green (1962)
Occurrence
Thin walled cells and tissues
Observable at microscopic level
Common in trees and stressed wood
Compression crease Wilkins (1986) refereed as microscopic Compression crease, Wilkins (1986) refereed as macroscopic Compression wood Keith (1974) and microscopic compression failure
Observable at microscopic level Observable with naked eye
Corner crinkle
Green (1962)
Corrugation
Green (1962)
Crack arrested
Thuvander and Berglund (2000)
Crack growth and microcracks
Dill-Langer et al. (2002)
Crack propagation
Fruhmann et al. (2003) Robinson (1920)
Observable with light microscope Observable with light microscope Observable with light microscope Observable with confocal Laser Scanning Microscope Observable in ESEM Spruce wood
Crinkle
Description Buckling of cell, equivalent to Brush’s (1913) term “bending” Buckling of fibres precedes macroscopic failure Deformation characterized by coarse transverse folds and longitudinal cracks in the inner cell wall layers Reference of the level of observation must be made Structural deformation of cell wall frequently referred to as slip planes Cell wall distortion or discontinuity which is more pronounced than a cell wall crinkle. Produced by growth stress in trees or in wood by applied perpendicular stress. Horizontal rows of slip planes
Horizontal rows of slip planes
Deformation with a marked resemblance to Scurfield et al. (1972) defined as “wrinkling of the cell wall linings Minute crinkle originating in, or confined to, the thickenings at the corners of tracheids Cell deformations ranging from smooth undulations to sharp peaked folds crack propagation stopped in the latewood Progress of crack in early wood
Crack initiation and propagation in earlywood Local or horizontal bands of cell wall “crinkles”
2
Terms for Delamination in Wood Science and Technology
21
Table 2.1 (continued) Term
Author
Occurrence
Description
Crinkling
Bienfait (1926)
In earlywood during early gross failure
Permanent “telescoping deformation”
Fine microscopic crease, notes C1
Dinwodie (1968)
Fine microscopic crease, notes C2 Fold Fine and transwall fractures First step of crack propagation Fracture surface
Dinwoodie (1968)
Gross compression failures with light microscopy
Gross microscopic crease C3 Initial compression failures Interwall deformations Irregular crack profile Macroscopic buckling
Macroscopic compression crease Macroscopic compression failure
Macroscopic compression failure lines or creases
Robinson (1920) Donaldson (1995)
1 to 2 slip planes in depth, covering more than 2 cell wall in width 3 to 6 slip planes in depth Spruce earlywood Folding of cell Radiate pine Observations with SEM
Sippola and Pinus sylvestris Fruhmann (2002) Reiter et al. (2002) Different species Reiter and Sinn (2002) Bienfait (1926), Tissue with initial Dadswell and failure present, Langlands (1934) also occurs in both heart and truewood of Eucaliptus diversicolor Dinwoodie (1968) Bienfait (1926) Côté and Hanna (1983) Vasic and StanzlTschegg (2007◦ ) Côté and Hanna (1983)
Numerous in earlywood Observable with SEM observable in ESEM oak Observable with SEM
Dinwoodie (1966)
Keith (1974)
Observable with SEM
Kucera and Bariska (1982)
Observable with SEM
Observations with SEM Observations with SEM
Continuous deformation formed after initial failure development
More than 5 slip planes in depth The lining up of slip planes to form definite zones of failure Slip planes as described by Keith and Côté (1968) fracture through vessels Buckling of fibres which is preceded by cell wall deformation and related to slip plane formation Horizontal zone of dislocations, produced by failures in adjacent cell wall. Involves the development of shear planes, buckling of whole fibres and is normally preceded by slip plane development and microscopic compression creases Deformation visible to the naked eye. Multilayered accumulation of the structural deformation pattern
22
V. Bucur Table 2.1 (continued)
Term
Author
Macroscopic compression lines
Kisser and Observable with Steininger (1952) light microscopy Dinwoodie (1966)
Microscopic compression crease
Occurrence
Microscopic compression crease or line
Dinwoodie (1966)
Microscopic compression failure Microscopic compression failure lines or creases
Keith and Côté (1968)
Observable with SEM
Kucera and Bariska (1982)
Observable with SEM
Description Enlarged microscopic compression line which is visible to the naked eye Severe crinkling of the cell walls, produced by increased loading following microscopic compression crease formation Distinct rows of dislocations. The second stage in cell wall failure following slip plane formation. May develop independent to slip planes. Closely associated converging or crossing slip line
Microscopically visible changes in cell shape as buckling and/or telescopic shortening, type S or U. Type S : double bending, type U – triple bending Microscopic Kisser and Common in most Progression of slip lines compression lines Steininger (1952) wood species horizontally, from fibres to fiber Minute compression Dadswell and Common in brittle A lining up of failures in failure Langlands (1934) heart adjacent cell wall, and produced by incipient decay Minute dislocation Dinwoodie (1966) Common in wood Slip line species Multiple slip plane Scurfield et al. Synonymous with “creases” (1972) defined by Dinwoodie 1968 as areas where varying numbers of slip planes are concentrated. The next stage following slip plane formation. Severe type of cell wall fold. Node Green (1962) Common in Point of flexing in pupl pulped tracheids tracheids and stressed wood Bending of fibres from the Offset Bienfait (1926) In latewood during early original axial line gross failure Pre-crack Boatright and Macroscopic Crack propagation normal to the Garrett (1983) crack extension plane of the pre-crack in LT occurring parallel to the grain Predominant fracture Donaldson (1995) Radiate pine Differences in fracturing at S1 /S2 boundary behaviour
2
Terms for Delamination in Wood Science and Technology
23
Table 2.1 (continued) Term
Author
Radial dislocations
Chafe (1977)
Slip planes, and slip lines
Kucera and Bariska (1982)
Occurrence
Observable with SEM
Common in most Robinson (1920), Slip lines, slip wood species Bienfait (1926), planes, as defined Dadswell and between 1920 and Langlands (1934), 1952 (observed Wardrop and with light Dadswell (1947); microscope) Kisser and Steininger (1952) Slip plane in relation Dinwoodie (1968) to microfibrils
Slip plane with SEM Wilkins (1986) Slip plane, defined using SEM, after 1960
Keith and Côté (1968), Keith (1971)
Slip planes
Wilkins (1986)
Stress line and microfibrils
Scurfield et al. (1972)
Stress lines, or thrust Dinwoodie (1968) lines
Observable with SEM Observable with SEM
Observable with light microscope or with SEM
Description Radial dislocations found in the inner S2 and extending to the cell lumen. Removal of growth stress causes cell shortening and the closer packing of microfibrils changes the lumen surface from smooth to convoluted Local deformation or crinkling of cellulose fibrils in the whole of the cell wall of one or two neighbouring cells, without prominent change in shape of cell Fine crack lines in the cell wall, preceding buckling, crinkling and tension failure. A crinkle in the cell wall. Fine streaks intercrossing at a certain angle and extending through the secondary wall Dislocation or crinkling of the fibrils comprising the cell wall occurring either singly or in pairs All deformation observed as wrinkled transverse lines Non-crossing single line cell wall deformation. Possibly a stage in microscopic compression failure and not always a sectioning artefact. The scale of observation and the type of microscope used must be defined Pre-cursor of slip plane. Barely detectable cellulose microfibril deformation Precursor of slip planes. Slight dislocation virtually unobservable by polarization microscopy or staining, but observable by electron microscope
24
V. Bucur Table 2.1 (continued)
Term
Author
Occurrence
Description
Stress lines or thrust lines
Kucera and Bariska (1982)
Observable with SEM
Telescopic shortening Thrust line
Kucera and Bariska (1982) Kisser and Frenzel (1950)
Mainly thick walled cells Common in most wood species
Wrinkling in the cell wall towards the lumen, affecting “cell wall lining” A diversion of cells from their natural axial orientation Slight local thickenings of the cell walls due to small deformation of the fibrils. Pre-slip plane May be considered morphologically similar
Thrust lines, slip Wilkins (1986) planes, compression creases Thrust-line or stress Wilkins (1986) lines describe only those slip planes not observable with light microscopy Transverse fracture Sell and surface Zimmermann (1998) Wrinkling of cell wall
Scurfield et al. (1972)
Observable with light microscope or with SEM Observable only with SEM
High resolution FE -SEM
Pre-slip plans deformation, which are not distinguishable from slip planes when using SEM .
Poly-laminated concentric structure of the cell wall layers observed in transverse surface Involves only the covering lining the lumina of fibres. It is a stage after multiple slip plane formation in the sequence of events occurring during axial compression
Delamination effects in plywood, as defined in ASTM D 1038 are noted below: – Blister in plywood is an elevation of the surface of an adherend (separation between plies) somewhat resembling in shape a blister on the human skin; its boundaries may be indefinitely outlined and it may have burst or become flattened. – Broken Grain (shelling, leafing, grain separation) a separation on veneer surface between annual rings. – Closed Surface Checks – Delamination – the separation of layers in a laminate because of failure of the adhesive, either in the adhesive itself or at the interface between the adhesive and the adherent – Durability as applied to the glue bond – its resistance to deterioration related to exposure conditions – see also delamination – Gap – an open joint or split in the inner plies which results when cross band or centre veneers are broken or not tightly butted – Open Joint – failure of bond or separation of two adjacent pieces of veneer so as to leave an opening, usually applied to edge joints between venerers
2
Terms for Delamination in Wood Science and Technology
25
– Rupturing of some of wood structural elements which forms cavities of various sizes in radial plane – Skips or Voids in the glueline of plywood – Starved Joints – a glue joint that is poorly bonded because of an insufficient quantity of glue. – Sunken joint in plywood – a depression in the surface of the face ply directly above the edge joint in a lumber core or crossband. Usually the result of localized shrinkage in the edge-jointed layer. – Wood Failure ◦ as applied to plywood glueline testing – the area of wood fiber remaining at the glueline following completion of the specific shear test. Determination is by means of visual examination and expressed as a percent of the test area ◦ As applied to failure in plywood not directly associated with the adhesive, is a rupture, shelling, tearing or breaking of wood itself. The standard ASTM D 1101 refers to the integrity of glue joints in structural laminated wood products for exterior use and employs the term delamination such as: “Delamination is a term used to express separation of the wood surfaces at the glue joints. When the separation takes place in the wood member, even though very close to the glue joint, it is termed wood failure or checking”. Furthermore it is noted that since glue joints at knots and knotty areas in general are not detectable under severe exposures, development of delamination at knots should be disregarded and not included in the measurements or calculations. Quantification of the delamination effect in laminated panels is noticed in the following standards: – – – – – –
the shear through the thickness of structural panels (ASTM D 2719) the shear modulus of wood based laminated structural panels (ASTM D 3044) the toughness of wood based structural panels (ASTM D 3499) the stresses for structural glued-laminated timber (ASTM D 3737) the stresses for structural composite lumber products (ASTM D 5456) the accelerated aging test (ASTM 1037, Chapter 7) for the ability of the material to withstand severe environmental exposure conditions .
2.5 Terms for the Delamination in Wood-Based Fibre and Particle Panels ASTM D 1554 gives the terms related to wood-based fibre and particle panels defined “as a group of board materials manufactured from wood or other lignocellulosic fibres or particles to which binding agents and other materials may be added during manufacture to obtain or improve certain properties”. Under the generic name of wood-based fibre and particle panels, two types of panels are included: the fibrous – felted panels and the particleboards. Fibreboard panels – is “a board generic term encompassing sheet materials of widely varying densities
26
V. Bucur
manufactured from refined or partially refined wood fibres”(for example: medium density fibreboard (MDF) having the density between 400 and 800 kg/m3). Particle board is composed of particles such as: chips, flakes, strands, wood – wool, etc. “The particle is the aggregate component of a particleboard manufactured by mechanical means from wood or other lignocelullosic material, comparable to the aggregate in concrete”. The delamination effect in wood-based fibre and particle panels is noticed in: – ASTM D 1037 and is related to the shear test in compression, to the interlaminar shear test and to the edgewise shear test which is a shear test normal to the plane of the board – ASTM D 1038 which recommends the accelerated aging test “used to obtain a measure of the inherent ability of a material to withstand severe exposure conditions. The cycling exposure to which the material shall be subjected is a simulated condition developed to determine relatively how a material will stand up under aging conditions” of high temperature and high relative humidity. The determination of the cohesive bond strength of the fibres or particles on the surface of wood – base fibre and particle panels in the direction perpendicular to the plane of the panel is regulated by ASTM D 5651.
2.6 General Classification of Delamination The myriad of terms related to delamination in wood and wood–based composites, presented previously required a new classification, which can support a more general mechanical approach related to delamination initiation and growth. In the following we propose to follow the classification suggested by Bolotin (1996) for engineering artificial composites. The criterion of this classification is the position of delamination into the member, such as: internal delamination, near surface delamination and delamination producing multi-cracking of the member (Fig. 2.1).
Fig. 2.1 Position of delamination in layered composite materials. (Bolotin 1996, Figure 1) Legend: (a) internal delamination, situated within the bulk of the material, can be studied with conventional fracture mechanics (b) near – surface, or crack – like defect, very often accompanied by their buckling, can be studied with the theory of elastic stability (c) multiple cracking – crack like flow affecting the load carrying capacity of the member and the safe life of the structure
2
Terms for Delamination in Wood Science and Technology
27
In the context of structural health monitoring, the mechanical behaviour of composites with delamination can be studied with linear or non linear Fracture Mechanics, with nondestructive ultrasonic methods and with model dependent methods implemented by finite element analysis, which are able to provide local and global damage information. The internal delamination will be referred to as delamination observed as cracks and studied with Fracture Mechanics. Internal delamination can be detected in solid wood as well in as in wood-based composites at submicroscopic, microscopic and macroscopic scale. For example: in solid wood, between the middle lamella and the other cell wall layers or between the S1 and S2 layers or between S3 and G layers as frequently observed for compression wood or tension wood (Fig. 2.2). At macroscopic scale the delamination occurs in the annual ring between zones of different densities such as earlywood and latewood, or earlywood and medullary rays. In wood-based composites such as fibreboards, the fibre adhesive interaction during manufacturing is random and sometime the fibres remain attached in bundles, the middle lamella is degraded and large voids between the fibres can be observed (Fig. 2.3).
Fig. 2.2 Delamination in poplar wood, between the middle lamella and the other cell wall layers or between the middle lamellae LM, S1 and S2 or between S3 and G due to drying. Bar _________: 10 μm (Clair 2001, Figure 69)
28
V. Bucur
Fig. 2.3 Delamination in fibreboards is observed as large void between the fibers Murmanis et al. (1986, Figure 4) Legend: V = vessel, P = parenchyma, F = fibers, ML = middle lamella (a) Wet formed hardboards, high density with 0.5% phenol – formaldehyde. The dark granular material is scattered between cells. White zones are voids. Arrow shows the softened middle lamella (ML). Microphotograph×4760. (b) Wet formed hardboards, high density with 0.5% phenol – formaldehyde. Because of the pressure ML is in the fiber lumen. White zones are voids. Microphotograph×5300. (c) Dry-formed hardboard, high density with 0.5% phenol – formaldehyde. Parenchyma (P), vessel (V) and fibres (F) are present. White zones are voids. Microphotograph×3040.
Near surface delamination is situated just near the member surface and is always accompanied by buckling such as blisters in plywood originating from the manufacturing process. Its growth is observed as interlaminar damage. Delamination producing multiple cracking through the whole thickness of the member, without separation of the layers is typical for seasoning checks in solid
2
Terms for Delamination in Wood Science and Technology
29
wood. In fibrous composites this defect is very frequent and is induced during the manufacturing process by thermal factors. In structural layered wood-based composites multiple cracks can be observed in LVL or in glulam structural members in service, used in civil engineering (i.e. houses, bridges, sport halls, etc.). Technological instabilities in the fabrication process, shrinkage produced by thermal and chemical factors, biological degradation, etc., are sources of initiation of multiple delaminations. Local instability and crack growth in glued laminated timber may produce the global instability of large structural components which in extreme cases may even result in failure of the whole structure with the potential to cause loss of life. The implementation of damage detection strategies must be a constant object of preoccupation for engineers and wood technologists for permanent structural health monitoring of buildings and structures incorporating wood or wood-based composites. Reliable information regarding the integrity of the structure can help in the prognosis of these structures under current environmental conditions and estimate the remaining useful life of the system.
2.7 Summary In Material Science, delamination is defined as a sub critical damage to the interfaces between the plies in a laminate composite that causes a reduction in the load carrying capacity of composite (Morris 1992). The terms which describe delamination in wood and wood- based composites are numerous and often confusing for multiple reasons (the use of terms which were previously considered inappropriate, new technologies related to microscopic observation of the structure, etc). A comprehensive understanding of these terms is essential for the uses of wood products under competitive conditions of modern technology. A new classification of the delamination in wood, wood products and wood-based composites is proposed, depending on its position in the member, such as: internal delamination, near surface delamination and delamination producing multi-cracking of the member. Fracture Mechanics is an useful tool for the study of initiation of cracks and growth of delamination in wood and wood-based composites. In the context of structural health monitoring, the detection of damage induced by delamination in wood-based composites can be achieved with nondestructive ultrasonic methods and with model dependent methods implemented by finite element analysis. These methods are able to provide local and global damage information, as can be seen in Chapter 3.
References American Society for Testing and Materials (2007) Standard terminology relating to wood and wood-based products. ASTM D 9 – 05. Philadelphia, PA American Society for Testing and Materials (2007) Standard test methods for evaluating properties of wood - base fibre particle panel material. ASTM D 1037-06a. Philadelphia, PA
30
V. Bucur
American Society for Testing and Materials (2007) Standard terminology relating to veneer and plywood. ASTM D 1038- 83 (2005) Philadelphia, PA American Society for Testing and Materials (2007) Standard terminology relating to wood-based fibre and particle panel material ASTM D 1554 - 01 (2005) Philadelphia, PA American Society for Testing and Materials (2007) test methods for structural panels in shear through the thickness. ASTM D 2719 – 89 (2007) Philadelphia, PA American Society for Testing and Materials (2007) Standard test method for shear modulus of wood-based structural panels. ASTM D 3044 – 94 (2006) Philadelphia, PA American Society for Testing and Materials (2007) Standard test method for toughness wood-based structural panels. ASTM D 3499 – 94 (2005) Philadelphia, PA American Society for Testing and Materials (2007) Standard practice for establishing allowable properties of structural glued-laminated timber (glulam). ASTM D 3737- 07 Philadelphia, PA American Society for Testing and Materials (2007) Specification for evaluation of structural composite lumber. ASTM D 5456-06 Philadelphia, PA American Society for Testing and Materials (2007) Standard test method for surface bond strength of wood-based fibre and particle panel material ASTM D 5651 – 95a (2002) Philadelphia, PA American Society for Testing and Materials (2007) Standard guide for evaluating mechanical and physical properties of wood-plastic composites products ASTM D 7031 -04 (2004) Philadelphia, PA ASTM D1101 - 97a (2006) Standard Test Methods for Integrity of Adhesive Joints in Structural Laminated Wood Products for Exterior Use Bienfait JL (1926) Relation of the manner of failure to the structure of wood under compression parallel to the grain. J Agri Res 33:183–194 Boatright SWJ, Garrett GG (1983) The effect of microstructure and stress state on the fracture behaviour of wood. J Mat Sci 18:2181–2199 Bolotin VV (1996) Delaminations in composite structures: its origin, buckling, growth and stability. Composites: Part B, 27B:129–145 Brush WD (1913) A microscopic study of the mechanical failure of wood. U.S. Depart Agri Rev Forest Serv 2:33–38 Chafe SC (1977) Radial dislocations in the fiber wall of Eucalyptus regnans trees of high growth stress. Wood Sci Techn 11:69–77 Clair B (2001) Etudes des proprietes mecaniques et du retrait au sechage du bois a l`echelle de la paroi cellulaire . PhD thesis Universite de Montpellier II. France Côté WA, Hanna RB (1983) Ultrastructural characteristics of wood fracture surfaces. Wood Fiber Sci 15:135–163 Dadswell HE, Langlands I (1934) Brittle heart in Australian timbers: a preliminary study. J Couns Sci Ind Res Australia 7:190–196 Dinwoodie JM (1966) Introduction of cell wall dislocations (slip planes) during the preparation of microscopic sections of wood. Nature 212:525–527 Dinwoodie JM (1968) Failure in timber. Part I. Microscopic changes in cell wall structure associated with compression failure. J Inst Wood Sci 4:37–53 Dill-Langer G, Lutze S, Aicher S (2002) Microfracture in wood monitored by confocal laser scanning microscopy. Wood Sci Technol 36:487–499 Donaldson LA (1995) Cell wall fracture properties in relation to lignin distribution and cell dimensions among three genetic groups of radiate pine. Wood Sci Techn 29:51–63 Fruhmann K, Burgert I, Stanzl-Tschegg SE, Tschegg EK Mode I (2003) Fracture behaviour on the growth ring scale and cellular level of spruce and beech loaded in the TR crack propagation system. Holzforschung, 57:653–660 Green HV (1962) Compression caused transverse discontinuities in tracheids. Pulp Paper Mag Canada 63(3):T 155 – T 168 Jacard P (1910) Etude anatomique des bois comprimés. Mitt Schw. Centralanstalt. Forst. Versuchwessen 10:53–101
2
Terms for Delamination in Wood Science and Technology
31
Keith CT (1971) The anatomy of compression failure in relation to creep – inducing stresses. Wood Sci 4:71–82 Keith CT (1974) Longitudinal compressive creep and failure development in white spruce compression wood. Wood Sci 7:1–12 Keith CT, Côté Jr. WA (1968) Microscopic characterization of lip lines and compression failures in wood cell walls. Forest Prod J 18:67–74 Kisser J, Frenzel H (1950) Mikroskopische Veränderungen der Holzstruktur bei mechanischer Überbeansprucging von Holz in der Faserrichtung. Schr Österr. Ges. Holzforschung 2:3–27 Kisser J, Frenzel H (1952) Makroscopische und microsckopische Strukturänderungen bei der Biegebeanspruchung von Holz. Holz Roh- und Werkstoff 10:415–421 Kucera LJ, Bariska M (1982) On the fracture morphology in wood. Part I: A SEM - study of deformations in wood of spruce and aspen upon ultimate axial compression load. Wood SciTechnol 16:241–259 Meyer RV, Leney L (1968) Shake in coniferous woods – an anatomical study. Forest Prod J 18(2):51–56 Morris C (ed) (1992) Dictionary of science and technology. Academic, Sandiego, p 604 Murmanis L, Youngquist JA, Myers GC (1986) Electron microscopy study of hardboards. Wood Fiber Sci 18(3):369–375 Reiter A, Sinn G (2002) Facture behaviour of modified spruce wood: a study using linear and non linear fracture mechanics. Holzforschung 56:191–198 Reiter A, Sinn G, Stanzl-Tschegg SE (2002) Fracture characteristics of different wood species under mode I loading perpendicular to the grain. Mater Sci Eng A 332:29–36 Robinson W (1920) The microscopical features of mechanical strains in timber and the bearing of these on the structure of the cell wall in plants. Phil Trans R Soc 210 B:49–82 Scurfield G, Silva SR, Wold MB (1972) Failure of wood under load applied parallel to grain. A study using scanning electron microscopy. Micron 3:160–184 Sell J, Zimmermann T (1998) The fine structure of the cell wall of hardwoods on transverse fracture surfaces. HolzRoh Werkst 56:365–366 Thuvander F, Berglund LA (2000) In situ observations of fracture mechanisms for radial cracks in wood. J Mat Sci 35:6277–6283 Tschegg EK, Fruhmann K, Stanzl-Tschegg SE (2001) Damage and fracture mechanisms during mode I and mode III loading of wood. Holzforschung 55:525–533 Vasic S, Stanzl-Tschegg SE (2007) Experimental and numerical investigation of wood fracture mechanisms at different humidity levels. Holzforschung 61:367–374 Wardrop AB, Dadswell HE (1947) The occurrence, structure and properties of certain cell wall deformations. J Coun Sci Ind Res Aust 221(5):14–32 Wilkins AP (1986) The nomenclature of cell wall deformations. Wood Sci Technol 20:97–109
Chapter 3
Delamination Detection – A Vibration-Based Approach Voichita Bucur
Contents 3.1 3.2 3.3
Introduction . . . . . . . . . . . . . . . . . . . . Delamination Detection with an Ultrasonic Technique Delamination Detection with a Model-Based Method 3.3.1 Linear Behavior . . . . . . . . . . . . . . . 3.3.2 Nonlinear Behaviour . . . . . . . . . . . . . 3.4 Some Practical Aspects . . . . . . . . . . . . . . 3.5 Summary . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
33 33 38 38 40 44 45 46
3.1 Introduction In this chapter we shall describe the aspects related to delamination in composites revealed by a vibration-based approach and related to the local and global damage detection. The local damage detection is performed with an ultrasonic technique, with Lamb waves, while the global damage detection is based on a model – based method using low frequency vibrations and undertaking the analysis of structural models implemented by finite element analysis. In this chapter the delamination detection studies are commented in the context of structural health monitoring, which is referred as the process of implementing a damage detection strategy for mechanical engineering infrastructures or for other purposes.
3.2 Delamination Detection with an Ultrasonic Technique Interfaces play an important role in determining the performance of laminated composite materials on a wide variety of scales, from interlaminar bonds to adhesive V. Bucur (B) CSIRO, Materials Science and Engineering Div. Bayview Avenue, Clayton, Victoria 3168, Australia e-mail:
[email protected] V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_3,
33
34
V. Bucur
bonds. The defects expected to be present at the interface are cracks at the interfaces of different oriented plies, inter-ply delamination, adhesion weakness at interfaces between plies or between a ply and an adhesive layer. In all cases the basic purpose of the nondestructive evaluation methods is the determination of the integrity of bonds. The efficiency of ultrasonic methods is related to the understanding of the relationship between the measured parameters and the interface mechanical properties, which is dependent on the theoretical approach used to predict the behavior of the interface, according to the specific kind of defect expected to be present Combining the experimental data with the theoretical knowledge (Hirsekorn 2001) it is possible to gain important information about the linear or non linear interface behavior (Krohn et al. 2002; Solodov et al. 2002). In the years 1970–1990 stress was put on the damage identification and health monitoring of laminated composites through the overall mechanical characteristics of the structure by measuring the stiffness matrix, the viscoelastic parameters, etc. determined with ultrasonic waves. Theoretical models for plane wave propagation in layered anisotropic composites were developed in a very impressive amount of articles and reference books. Because space limitation only several references has been selected (Green 1985–2006; Chimenti 1981–2006; Bunsell 1988; Nayfeh and Chimenti 1988; Hosten et al. 1987; Rose et al. 1990; Alleyne and Cawley 1992; Deschamps and Hosten 1992; Rokhlin and Wang 1992; Potel and de Belleval 1993a, b; Saravanos et al. 1994; Lavrentyev and Rokhlin 1998). In that follows our attention is focused on the ultrasonic method based on Lamb waves. Lamb waves are defined as mechanical waves corresponding to vibration modes of plates having the thickness of the same order of magnitude as their wavelength. Lamb waves are suitable for the nondestructive evaluation of large structural elements, due to their prominent characteristic - the long range propagation, with low dispersion energy, even in materials with high attenuation ratio. The Lamb waves are able to put in evidence the presence of defects, as noted in a very extensive body of literature from which several references has been extracted (Rokhlin 1979, 1980; Pilarski and Rose 1987; Auld 1980, Chimenti and Martin 1991; Nagy 1992; Ogilvy 1995; Huber et al. 1997; Cawley and Alleyne 1996; Wright et al. 1996; Kazys R and Svilainis 1997; Maslov and Kundu 1997; Singer 1997; Delsanto et al.1998; Delsanto and Scalerandi 1998; Kundu et al.1998; Rokhlin and Wang 1998; Royer and Dieulesaint 2000; Hayashi and Kawashima 2002; Kessler et al. 2002a; Stoessel et al. 2002; Su et al. 2002; Sohn et al. 2004; Simonetti 2004; Shkerdin and Glorieux 2004, 2005; Toyama and Okabe 2004; Beadle et al. 2005; Fritzen and Mengelkamp 2005; Giurgiutiu et al. 2005; Hera et al. 2005; Konstantinidis et al. 2005; Lucero and Taha 2005; Nieuwenhuis et al. 2005; Raghavan and Cesnik 2005; Sundararaman et al. 2005; Terrien et al. 2007). Lamb wave characteristics such as dispersion curves, phase velocity, attenuation, reflection and transmission coefficients has been used to detect delamination, porosity, matrix cracking, and other surface defects. Interaction of Lamb wave modes with defects is an extremely valuable tool in providing quantitative information on the interface flaws and bond quality. Under different propagation modes Lamb waves generate high normal and shear stresses at different plate depth and consequently
3
Delamination Detection – A Vibration-Based Approach
35
some modes should be more sensitive to the interface defects and its stiffness variation than other modes. Terrien et al. (2007) investigated the interaction of Lamb modes with micro-defects with a simulation combining a finite element approach and a modal decomposition method. The region around the defects is described by the finite element mesh. The numerical simulation required first, the finite element modeling with an explicit algorithm for solving the transient wave propagation, second, the modal decomposition which allows to plot dispersive curves and to define the real, the evanescent and the leaky Lamb modes that exists at a given frequency and third, the analytical propagation of Lamb waves which are phase velocity and frequency dependent. The experimental setup for Lamb wave generation and detection on an aluminum plate of 1 m long, 300 mm wide and 2 mm thick, with notched of different sizes is shown in Fig. 3.1. The measured ultrasonic signals at different times and distances from the source are shown in Fig. 3.2 in which the A1 are Lamb modes transmitted by the notch, and S0 and A0 are incident modes produced by mode conversion. (Note : Si – symmetric modes and Ai antisymmetric modes). The reflections from the notch are clearly visible on Fig. 3.2a. The velocities of different Lamb modes transmitted by the notch can by identified as can be seen from Fig. 3.2b–d. All the modes which can propagate at different frequencies are shown in Fig. 3.3. (i.e green rectangle for excitation window at 2.25 MHz frequency , with a tone burst of 5 cycles at 66◦ incidence angle). In Fig. 3.4 are represented the incident waves, the transmitted waves, the reflections and the mesh used to identify the mode conversion with 2D Fourier transform technique. Figure 3.5 shows the modes A0, A1 and S0 of Lamb wave at 2.25 MHz in a 3 mm thick steel plate in a sound zone and in a zone with 1.5 mm deep notch.
Fig. 3.1 Experimental setup for Lamb wave generation in a plate with two main defects, a large notch and a defect assimilated to a crack produced by 5 thin notches. (Terrien et al. 2007, Figure 18)
36
V. Bucur
Fig. 3.2 Signals measured at different points: (a) 80 mm in front of a notch of 500 μm depth and 700 μm width, (b) 20 mm, (c) 45 mm, (d) 165 mm from the notch (Terrien et al. 2007, Figure 19)
As noted by Terrien et al. (2007), “knowing the modal expansion of the wave propagating on the right of the notch and the waveform of the displacement normal to the plate” it is possible to predict the waveform at any distance from the source. The method described here is elegant and has evident advantages such as the possibility to extract the mode conversion produced by the defects, and to predict the waveform quite far from the damaged area, if the depth of the defects is smaller then one half of the plate thickness.
3
Delamination Detection – A Vibration-Based Approach
37
Fig. 3.3 Corresponding dispersion curves of symmetric (solid lines) and antisymmetric (dashed lines) propagating Lamb modes. Legend: the excitation window is shown by the rectangle is for 2.25 MHz tone burst of 5 cycles and 66◦ incidence angle). (Terrien et al. 2007, Figure 21)
Fig. 3.4 Incident, transmitted and reflected waves and the mesh used to identify the mode conversion of Lamb waves with 2D Fourier transform technique. (Terrien et al. 2007, Figure 15)
Despite of the evident advantages of the ultrasonic method described here for the nondestructive inspection and evaluation of structural elements, drawbacks and limitations are evident, when this method is applied to real – time health monitoring. This method is local in nature, passive and labor intensive. However, it is to note that the development of the time reversal concept in modern acoustics (Fink 1992, 1997; Cassereau and Fink 1992; Wu et al. 1992) brings new prospective for the utilization of guided Lamb waves for the aerospace structures (Sohn et al. 2005) and for different civil and medical applications.
38
V. Bucur
Fig. 3.5 The modes A0, A1 and S0 of Lamb wave at 2.25 MHz in a 3 mm thick steel plate in a sound zone (a) and in a zone with 1.5 mm deep notch (b) (Terrien et al. 2007, Figure 16)
3.3 Delamination Detection with a Model-Based Method Successful application of damage detection and health monitoring of structures using the measured structural dynamic response and mathematical models has been possible with the advance in computer science and technology. Compared with the nondestructive testing and evaluation procedures, the model-based methods using low frequency vibrations have a more rigorous mathematical background, but also several limitations related to the interpretation of the physical meanings of the detected results and the precise numerical representation of the structures. The mechanical behavior of a damaged structure can be studied in two hypotheses, the linear or the non linear mechanical behavior. In that follows both aspects will be succinctly described.
3.3.1 Linear Behavior Model – based methods implemented by finite element analysis under static or dynamic loading, assume that the linear monitored structure responds can be accurately described by finite element analysis. It is assumed that the behavior of the structure is linear before and after damage. The composites are usually modeled as beams (Euler beam, Timoshenko beam) with through-width delaminations parallel to the beam surface located arbitrarily, or shells. Kim et al. (1997) proposed an analytical solution for predicting delamination buckling and growth of a thin fiber reinforced plastic layer in laminated wood beams under static bending. It was noted that the delamination growth is related to an explicit form of strain-energy release rate and the critical load can be accurately predicted. Simulation of the delamination indicated an unstable growth of the delamination after buckling of the delaminated sub-laminate, followed by arrested delamination growth. For the vibrating beams, the foundation of linear analysis is based on the concept of linear normal mode and the principle of superposition. Linear normal modes
3
Delamination Detection – A Vibration-Based Approach
39
are synchronous harmonic particular solutions of the homogeneous linear system (Vakakis 1996). The vibration phenomena of beams have been studied for the case of a single or multiple delamination through the beam thickness. Model-based methods under dynamic loading have been extensively commented and remarkably reviewed, periodically, by numerous authors (Salawu 1997; Doebling et al. 1997, 1998; Zou et al. 2000; Sohn et al. 2003; Montalvao et al. 2006). The dynamic model - based methods use changes in vibrational modal properties (i.e. modal frequencies, modal damping values and mode shapes) to infer changes in mechanical properties of the structure. The impulse or continuous excitation techniques can be used for vibrating the structure. Commonly hammer technique is used for impulse excitation (Fig. 3.6).The utilization of a non-contact scanning laser vibrometer system allows acquisition of a large number of measurement points for a better definition of the mode shapes. Continuous sine excitation can be produced by using PZT – lead-zirconate-titanate - ceramic wafers as actuator (Fig 3.7). The dynamic model – dependent methods can be subdivided into: modal analysis, frequency domain, time domain and impedance domain, according to the dynamic response parameters analyzed. Frequencies, mode shapes, curvature mode shapes and modal damping, which are function of the physical properties of the structure (mass, damping and stiffness), are the most commonly measured parameters, when the dynamic model - based methods are used. Modification of physical properties of the structure, such as for example reduction of stiffness resulting from cracks or delamination, will implicitly cause detectable changes in modal parameters. Furthermore, these changes must be used as indicators of damage, and the process of vibration - based damage detection reduced to some form of pattern recognition problem, as can be seen from the references cited below and extracted from a huge literature (Adams et al. 1978; Cawley 1990; Cawley and Adams 1979, 1987; Wang et al. 1982; Tracy and Pardoen 1989; Nagesh and Hanagud 1990;
Fig. 3.6 Experimental equipment for the excitation of flexural vibrations in a cantilever beam using a hammer. The beam response is detected by the laser vibrometer. (Berthelot and Sefrani 2004, Figure 1)
40
V. Bucur
Fig. 3.7 Experimental equipment for testing a free-free beam and measurements of modal frequencies and damping. The excitation can be induced by a shaker or by a piezoelectric actuator (PZT5) glued on the surface of the specimen, near the free end. (Chrysuchoidis and Saravanos 2004, Figure 2)
Paolozzi and Peroni 1990; Petyt 1990; Hanagud et al. 1990, 1992; Pandey et al. 1991; Tenek et al. 1993; Luo and Hanagud 1996; Messina et al. 1998; Sampaio et al. 1999; Wahl et al. 1999; Lestari and Hanagud 1999; D’Ambrogio and Fregolent 2000; Brandinelli and Massabo 2002; Kessler et al. 2002b; Lee et al. 2003; Berthelot and Sefrani 2004; Chrysochoidis and Saravanos 2004; Della and Shu 2005; Ghoshal et al. 2005; Coutellier et al. 2006; de Borst and Remmers 2006; Ladevèze et al. 2006; Lestari et al. 2007). Because of the fact that the damage is a typical local phenomenon, several difficulties can arise in its detection and location such as: – higher frequency modes are able to capture local responses, whereas lower frequency modes capture the global response of the structure – for the excitation of higher modes more energy is required than for the excitation of lower modes and loss of information can result from the reduction of time history measurements – shifting from the linear to nonlinear response. For damage identification and health monitoring of structures, many different issues are critical, such as: the excitation and measurement configurations, the selection of the type of sensors and their location, the signal processing performing such as: Fast Fourier analysis, time – frequency analysis, or wavelet analysis (Castro et al. 2007).
3.3.2 Nonlinear Behaviour Nonlinear damage is observed in the case when the initially linear-elastic structure behaves nonlinearly after the damage has been produced. Nonlinear normal modes
3
Delamination Detection – A Vibration-Based Approach
41
are defined as some synchronous periodic particular solutions of the homogeneous nonlinear system under which all degrees of freedom undergo the extreme position at the same time (Vakakis 1996). The most frequent nonlinearities in a delaminated beam are introduced by: the nonlinear geometric effects such as axial stretch effects; the deflection – dependent interactions in both longitudinal and transverse directions; the intermittent contacts between the segments during vibration, the delaminated segments constraining the movement of each other or the growth of delamination. From the existing literature in this field several articles were selected to illustrate the approach with nonparametric – models (Tseng and Dugundji 1971; Abhyankar et al. 1993; Hanagud and Luo 1994, 1997; Gammadi and Hanagud 1995; Nayfeh et al. 1995; Luo and Hanagud 1997a, b, c, 2000; Lestari and Hanagud 2001; Lu et al. 2001; Caron et al. 2006; Wang and Yu 2006; Perel 2006; Wang and Yu 2006; Friswell 2007; Wang and He 2007). The existing of the “delamination modes” was demonstrated by Hanagud and Luo (1994) and Luo and Hanagud (2000). Modeling of the delamination effects is shown in Fig. 3.8. After delamination the composite beam is represented as a combination of four beams connected at the delamination boundaries, having the characteristics denoted in the previous figure. Luo and Hanagud (2000) noted that the effect between the laminated surfaces depend on the relative position between the sublaminates during vibration. Some constraints between the upper and lower delamination surfaces still exist. Under a small amplitude vibration of the delaminated beam at a frequency corresponding to a delamination opening mode, the effect between delaminated sublaminates can be modeled as a distributed soft spring between them. When the amplitude exceeds a certain level, the spring effect becomes zero because the delamination opens beyond the small amplitude constraints. On the other hand, when the vibration mode does not tend to open
Fig. 3.8 Modeling of the delamination effects in a representative composite beam (Luo and Hanagud 2000, Figure 1). Legend: b is the beam width, H is the beam height, L is the beam length, and respectively mi , Di , Si , Ai (i = 1, 2, 3, 4) the mass density per unit length, bending stiffness, cross sectional shear stiffness and extensional stiffness of four beams. H2 and H3 are the distances between the neutral axis of delaminated beam and the neutral axis of intact beam
42
V. Bucur
Fig. 3.9 The nonlinear spring model describing the behavior of the effects between the delaminated sublaminates as shown with the dashed line (Luo and Hanagud 2000, Figure 2). Legend: do – relative displacement; w2 - displacement of the beam 2, w3 – displacement of the beam 3
the delamination, the delaminated sublaminates have the same flexural displacement and slopes. Thus, the effects between the delaminated sublaminates may be described by a nonlinear spring model as shown qualitatively in Fig. 3.9 by a dashed line. Furthermore this nonlinear model was reduced into a piecewise linear model depending on the relative displacement, expressed as w2 − w3 . Three situations can be observed: (a) w2 − w3 > 0 the delamination tends to open in vibration, the distributed contact force is zero. The spring model is represented by the solid line OA. (b) w2 − w3 = at a fix value, the delamination is completely closed during the vibration. The spring model is represented by the solid line BC (c) - do < w2 − w3 < 0 the delamination beam is vibrating in a small amount of relative displacement. The spring model is represented by the solid line OB. With the above considerations and from the solutions of the governing equations of motion of delaminated structures in different stages of vibration it was possible to synthesized the nonlinear dynamic response, through a nonlinear modal analysis technique developed by Luo and Hanagud (1997c). Figure 3.10 shows a typical mode of a transverse isotropic beam with interface 3, 3-inch delamination, and it is to note that the prediction is closed to the model. In conclusion, it is to note that the nonlinear dynamic response of the studied structure is precisely predicted with the proposed piecewise-linear model by Luo and Hanagud (2000). The reader interested in the case of multiple delaminations is invited to read the articles published by: – Gummadi and Hanagud (1995) for vibration characteristics of beams with multiple delaminations – Lestari and Hanagud (1999) for multiple delamination dynamics in composite beams, using the Euler – Bernoulli beam theory in connection with piecewise – linear springs to simulate the open and closed behaviour between the delaminated layers. – Della and Shu (2005), which studied the case of a beam with overlapping delaminations.
3
Delamination Detection – A Vibration-Based Approach
43
Fig. 3.10 Comparison between the experimental data and the prediction data based on the nonlinear model (Luo and Hanagud 2000, Figure 13). Legend: (a) vibration of the composite beam with interface 3, 3-inch delamination, experimental data provided by Shen and Grady (1992) (b) prediction based in the nonlinear mode
– Sridharan (2008) – for delamination behaviour of composites. (Please note that these recommendations reflect only my opinion when these pages have been written). Chattopadhyay et al. (1999) reported the nonlinear response of a delaminated smart composite cross – ply beams. The theory is implemented through finite element method including nonlinear induced strain effects. The numerical results indicate changes in the dynamic responses of the beam due to dilamination.
44
V. Bucur
Fig. 3.11 The nonlinear response of a smart composite cross – ply cantilever beam with delamination at the first mode of vibration (Chattopadhyay et al. 1999, Figure 9)
Figure 3.11 shows the nonlinear response of a smart composite cross – ply cantilever beam with delamination at the first mode of vibration. Another group of methods used for the implementation of nonparametric- models methods and based on the identification of the nonlinear response of the structure are the neural-network-based methods (Luo and Hanagud 1997c).These methods are not commented here. Prognosis with statistical model development for feature discrimination is also a group of methods recently developed for structural health monitoring and damage detection (Montalvao et al. 2006). These methods are not commented here.
3.4 Some Practical Aspects As noted by Sohn et al. (2003) the implementation of a structural health monitoring systems must answer questions, related to the presence of damage and to the operational evaluation, such as: – the damage detection (existence of damage in the system), the damage location (where is the damage), the type of damage (what kind of damage), the extent of damage (how severe is the damage) and the prognosis (how much useful life remains). – the operational and the environmental condition which referees to the safety and economic motivations for performing the monitoring, and on the other hand which are the limitation on acquiring data. The structural health monitoring process of big wood laminated structures, in light of normal aging and degradation resulting from operational environments, must involve the periodic inspection of the system using:
3
Delamination Detection – A Vibration-Based Approach
45
– sampled dynamic response measurements from an array of transducers, establishing their number, resolution, bandwidth, data acquisition (periodically or continuously), storage and transmittal hardware; – extraction of the damage – sensitive features, normalization of data by the measured inputs or by environmental cycles (summer, winter); – statistical analysis of data to determine the current state of the system. After catastrophic events such as earthquakes, excessive snow loading, etc, the structural health monitoring process must provide reliable information about the integrity of the structure. The review of the theoretical ideas proposed in this chapter where expressed in order to perceive and identify for the future, the research directions able to identify the damage detection induced by delamination in wood products and in wood-based composites using ultrasonic and vibration measurements, for a practical implemented technology. This imply three main aspects : the understanding of the theoretical aspects related to the physical phenomena for delamination initiation and growth , the development of models and testing procedures, and the developments and validation of specific codes.
3.5 Summary In this chapter the damage detection studies in composite materials were summarized in the context of structural health monitoring, which is referred as the process of implementing a damage detection strategy for mechanical engineering infrastructure (Allix and Blanchard 2006). The review of the theoretical aspects related to the detection of damages induced by delamination in composites was oriented in two main directions: – the nondestructive evaluation method using an ultrasonic technique with Lamb waves, which is an experimental method able to provide local damage information – the model dependent method, undertaken analysis of structural models implemented by finite element analysis and able to provide global damage information, for linear and non-linear mechanical behavior of the system The structural health monitoring process of big wood laminated structures, in light of normal aging and degradation resulting from operational environments, must involve the periodic inspection of the system using: – sampled dynamic response measurements from an array of transducers, establishing their number, resolution, bandwidth, data acquisition (periodically or continuously), storage and transmittal hardware; – extraction of the damage – sensitive features, normalization of data by the measured inputs or by environmental cycles (summer, winter); – statistical analysis of data to determine the current state of the system.
46
V. Bucur
References Abhyankar NS, Hall EK, Hanagud S (1993) Chaotic vibrations of beams: Numerical solution of partial differential equations. Trans ASME J Appl Mech 60, March:167–174 Adams RD, Cawley P, Pye CJ, Stone BJ (1978) A vibration technique for non-destructively assessing the integrity of structures. J Mech Eng Sci 20:93–100 Alleyne DN and Cawley P (1992) The interaction of Lamb waves with defects. IEEE Trans Ultrason Ferroelectr Freq Control 39:381–396 Allix O, Blanchard L (2006) Mesomodeling of delamination: towards industrial applications. Compos Sci Technol 66:731–744 Auld BA (1980) Acoustic field and waves in solids. vol 1. Krieger, Malabar, FL. Beadle BM, Hurelaus S, Jacobs LJ, Gaul L (2005) Detection and localization of small notches in plates using Lamb waves. Proceedings of the 23rd international modal analysis conference. (IMAX XXIII), Paper no 96 Berthelot JM, Sefrani Y (2004) Damping analysis of unidirectional glass and Kevlar fibre composite. Compos Sci Technol 64:1261–1278 Borst R de, Remmers JJC (2006) Computational modelling of delamination. Compos Sci Technol 66:713–722 Brandinelli L, Massabo R (2002) Free vibrations of through – thickness reinforced delaminated beams. 15th ASCE engineering mechanics conference – EM 2002, June 2–5, Columbia University:1–8 Bunsell AR (ed) (1988) Quality and damage control in composite materials. Elsevier Applied Science Publishing, London Caron JF, Diaz Diaz A, Carreira RP, Chabot A, Ehrlacher A (2006) Multi- particle modelling for the prediction of delamination in multi-layered materials. Compos Sci Technol 66: 755–765 Cassereau D, Fink M (1992) Time reversal ultrasonic field. Part III. Theory of the closed time reversal cavity. IEEE Trans Ultrason Ferroelectr Freq Control 39:579–592 Castro E, Garcia-Hernandez MT, Gallego A (2007) Defect identification in rods subject to forced vibrations using the spatial wavelet transform. Appl Acoust 68(6):699–715 Cawley P (1990) Low frequency NDT techniques for the detection of disbands and delaminations. Br J Non-Destr Test. 32:454–461 Cawley P, Adams RD (1987) Vibration techniques of NDT. In Summerscales J (ed) Nondestructive testing of fibre – reinforced plastics composites, Elsevier, London, pp 151–200. Cawley P, Adams RD (1979) The location of defects in structures from measurements of natural frequencies. J Strain Anal 14, 2:49–57 Cawley P, Alleyne D (1996) The use of Lamb waves for the long range inspection of large structures. Ultrasonics 34:287–290 Chattopadhyay A, Dragomir-Daescu D, Gu H (1999) Dynamics of delaminated smart composite cross – ply beams. Smart Mater Struct 8:92–99 Chimenti DE (ed) (1981–2006) Review of progress quantitative nondestructive evaluation. Plenum Press, New York, NY Chimenti DE, Martin RW (1991) Nondestructive evaluation of composite laminates by leaky Lamb waves. Ultrasonics 29:13–20 Chrysochoidis NA, Saravanos DA (2004) Assessing the effects of delamination on the damped dynamic response of composite beams with piezoelectric actuators and sensors. Smart Mater Struct 13:733–742 Coutellier D, Walrick JC, Geoffroy P (2006) Presentation of a methodology for delamination detection within laminated structures. Compos Sci Technol 66:837–845 D’Ambrogio W, Fregolent A (2000) The use of antiresonances for robust model updating. J Sound Vibr 236:227–243 Della C N, Shu D (2005) Free vibration analysis of composite beams with overlapping delaminations. Eur J Mech A Solids 24:491–503
3
Delamination Detection – A Vibration-Based Approach
47
Delsanto PP, Scalerandi M (1998) A spring model for the simulation of the propagation ultrasonic pulses through imperfect contact interfaces. J Acoust Soc Am 104:2584–2591 Delsanto PP, Romano A, Scalerandi M, Moldoveanu F (1998) A genetic algorithm approach to ultrasonic tomography. J Acoust Soc Am 104:1374–1381 Deschamps M, Hosten B (1992) The effects of viscoelasticity on the reflection and transmission of ultrasonic waves by an orthotropic plate. J Acoust Soc Am 91:2007–2015 Doebling SW, Farrar CR, Prime MB (1998) A summary review of vibration – based damage identification methods. Shock Vibr Dig 30, 2:91–105 Doebling SW, Hermez FM, Peterson LD, Farhat C (1997) Improved damage location accuracy using strain energy based mode selection criteria. AIAA J 35(4):639–699 Fink M (1992) Time reversal of ultrasonic field- Basic principles. Part 1. IEEE Trans Ultrason Ferroelectr Freq Control 39:555–566 Fink M (1997) Time reversed acoustics. Physics Today 20:34–42 Friswell MI (2007) Damage identification using inverse methods. Phil Trans R Soc A 365:393–410 Fritzen CP, Mengelkamp G (2005) In situ damage detection and localization in stiffened structures. Proceedings of the 23rd international modal analysis conference (IMAX XXIII), Paper no 268 Gammadi LNB, Hanagud S (1995) Vibration characteristics of beams with multiple delaminations. Proceedings of the 36 AIAA/ASME/ASCE/ASC Structures, structural dynamics and materials conference – adaptive structures forum, New Orleans , LA, pp 140–150 Ghoshal A, Kim HS, Chattopadhyay A, Prosser WH (2005) Effect of delamination on transient history of smart composite plates. Finite Elem Anal Des 41(9–10) :850–874 Giurgiutiu V, Buli X, Cuc A (2005) Dual use of travelling and standing Lamb waves for structural health monitoring. Proceedings of the 23rd international modal analysis conference (IMAX XXIII), Paper no 361 Green RE Jr (Ed) (1985–2006) Nondestructive characterization of materials. Vol. 1– Vol. IXV, Plenum Press, New York, NY; Springer, Heidelberg Gummadi LNB, Hanagud S (1995) Vibration characteristics of beams with multiple delaminations. Proceedings of the 36th AIAA/ASME/ASCE/AHS/ASC – structures, structural dynamics and materials conference. New Orleans, LA, pp 140–150 Hanagud S, Luo H (1994) Modal analysis of a delaminated beam. Proceedings of the 10th international. conference experimental. mechanics, Lisabon, June 18–22, pp 880–888 Hanagud S, Luo H (1997) Damage detection and health monitoring based on structural dynamics. Structural health monitoring: current status and perspectives proceedings of international workshop on structural health monitoring, pp 715–726. Hanagud S, Nagesh Babu GL, Roglin RL, Savanur SG (1992) Active control of delaminations in composite structures. Proceedings of .33rd AIAA/ASME/ASCE/AHS/ASC SDM conference, pp 1819–1829 Hanagud S, Nagesh Babu GL, Won CC (1990) Delamination in smart composite structures. Proceedings of the 1990 SEM spring conference on experimental mechanics, Bethel, CT, Soc Exp Mech Inc:776–781 Hayashi T, Kawashima K (2002) Multiple reflections of Lamb waves at a delamination. Ultrasonics 40:193–197 Hera A, Shinde A, Hou Z (2005) Issues in tracking instantaneous modal parameters for structural health monitoring using wavelet approach. Proceedings of the 23rd international modal analysis conference. (IMAX XXIII), Paper no 338 Hirsekorn S (2001) Nonlinear transfer of ultrasound by adhesive joints – a theoretical description. Ultrasonics 39:57–68 Hosten B, Deschamps M, Tittmann BR (1987) Inhomogeneous wave generation and propagation in lossy anisotropic solids. Application to characterization of viscoelastic composite materials. J.A.S.A.M. 82:1763–1770 Huber RD, Mignogna RB, Simmonds KE, Schechter RS, Delsanto PP (1997) Dynamic full – field visualization of uktrasound interacting with material defects : Experiments and simulation. Ultrasonics 35:7–16
48
V. Bucur
Kazys R, Svilainis L (1997) Ultrasonic detection and characterization of delaminations in thin composite plates using signal processing techniques. Ultrasonics 35:367–383 Kessler SS, Spearing SM, Atalla MJ (2002a) In situ damage detection of composite structures using Lamb waves methods. Proceedings of the1st european workshop on structural health monitoring, pp 374–381 Kessler SS, Spearing SM, Atalla MJ, Cesnik CES, Soutis C (2002b) Damage detection in composite materials using frequency response methods. Compos:Part B 33:87–95 Kim Y, Davalos JE, Barbero EJ (1997) Delamination buckling of FRP layer in laminated wood beams. Compos Struct 37(3/4):311–320 Konstantinidis G, Wilcox P, Drinkwater B (2005) Damage detection using a distributed array of guided wave sensors. Proceedings of the 23rd international modal analysis conference (IMAX XXIII), Paper no 265 Krohn N, Stoessel R, Busse G (2002) Acoustic non – linearity for defect selective imaging. Ultrasonics 40:633–637 Kundu T, Maji A, Ghosh T, Maslov K (1998) Detection of kissing bonds by Lamb waves. Ultrasonics 35:573–580 Ladevèze P, Lubineau G, Marsal D (2006) Towards a bridge between the micro – and mesomechanics of delamination for laminated composites. Compos Sci Technol 66:698–712 Lavrentyev A, Rokhlin S (1998) Ultrasonic study of environmental damage initiation and evolution in adhesive joints. RNDE-Research in Nondestructive Evaluation 10, 1, 26 pages Lee S, Park T, Voyiadjis GZ (2003) Vibration analysis of multi – delaminated beams. Compos Part B:Eng 34:647–659 Lestari W, Hanagud S (1999) Health monitoring of structures: Multiple delamination dynamics in composite beams. Proceedings of the 40th AIAA/ASME/ASCE/AHS structures, structural dynamics and materials conference Lestari W, Hanagud S (2001) Nonlinear vibration of buckled beams: Some exact solutions. Int J Solids Struct 38:4741–4757 Lestari W, Qiao P, Hanagud S (2007) Curvature mode shape-based damage assessment of carbon/epoxy composite beams. J Intell Mater Syst Struct 18(March):189–208 Lu X, Lestari W, Hanagud S (2001) Nonlinear vibrations of a delaminated beam. J Vibr Control 7:803–831 Lucero J, Taha MMR (2005) A wavelet aided fuzzy damage detection algorithm for structural health monitoring. Proceedings of the 23rd international. modal analysis conference. (IMAX XXIII), Paper no 78 Luo H, Hanagud S (1996) Delamination modes in composite plates. J Aerospace Eng 9(4):106–113 Luo H, Hanagud S (1997a) An integrated equation for changes with structural dynamics of damaged structure. Int J Solids Struct, December:4557–4579 Luo H, Hanagud S (1997b) Dynamic learning rate neural network training and composite structural damage detection. AIAA J 35:1522–1527 Luo H, Hanagud S (1997c) Delaminated beam nonlinear dynamic response calculation and visualisation. Proceedings of the 38th AIAA/ASME/ASCE/AHS SDM Conference 1: 490–499 Luo H, Hanagud S (2000) Dynamics of delaminated beams. Int J Solids Struct 37(10):1501–1519 Maslov K, Kundu T (1997) Selection of Lamb modes for detecting internal defects in composite laminates. Ultrasonics 35:141–150 Messina A, Williams EJ, Contursi T (1998) Structural damage detection by a sensitivity and statistical based method. J Sound Vibr 216:791–808 Montalvao D, Maia NMM, Ribeiro AMR (2006) A review if vibration – based structural health monitoring with special emphasis on composite materials (2006) Shocks Vib Dig 38(4):1–6 Nagesh Babu GL, Hanagud S (1990) Delamination in smart structures – A parametric study on vibration. Proceedings of the 31st AIAA/ASME/ASCE/ AHS SDM Conference, pp 2417–2426 Nagy P (1992) Ultrasonic classification of imperfect interfaces. J Nondestr Eval 11: 127–139
3
Delamination Detection – A Vibration-Based Approach
49
Nayfeh AH, Chimenti DE (1988) Ultrasonic wave reflection from liquid – coupled orthotropic plates with application to fibrous composites. J Appl Mech 55:863–870 Nayfeh AH, Chin C, Nayfeh SA (1995) Nonlinear normal modes of a cantilever beam. J Vib Acoust 177:477–481 Nieuwenhuis JH, Neumann JJ, Greve DW, Oppenheimer IJ (2005) Simulation and testing for Lamb wave generation. Proceedings of the 23rd international modal Analysis conference (IMAX XXIII), Paper no 216 Ogilvy JA (1995) A model for the ultrasonic inspection of composite plates. Ultrasonics 33:85–93 Pandey AK, Biswas M, Samman MM (1991) Damage detection from changes in curvature mode shapes. J Sound Vibr 145:321–332 Paolozzi A, Peroni I (1990) Detecting of debonding damage in composite plates through natural frequency vibrations. J Reinforced Plastics Compos 9:369–389 Perel VY (2006) A new approach for dynamic analysis of composite beam with an interplay crack. Nonlinear Dyn Syst Theory 6(2):171–186 Petyt M (1990) Introduction to finite element vibration analysis. Cambridge University Press. UK Pilarski A, Rose JL (1987) A transverse – wave ultrasonic oblique incidence technique for interfacial weakness detection in adhesive bonds. J Appl Phys 63:300–307 Potel C, de Belleval JF (1993a) Propagation in an anisotropic periodically layered medium. J Acoust Soc Am 93:2669–2677 Potel C, de Belleval JF (1993b) Acoustic propagation in anisotropic periodically multilayered media: A method to solve numerical instabilities. J Appl Phys 74:2208–2215 Raghavan A, Cesnik CES (2005) Analytical models for Lamb waves based structural health monitoring. Proceedings of the 23rd international modal analysis conference (IMAX XXIII), Paper no 289 Rokhlin S (1979) Interaction of Lamb waves with elongated dalaminations in thin sheets. Int Adv Nondestr Test 6:263–285 Rokhlin S (1980) Diffraction of Lamb waves by a finite crack in an elastic layer. JAcoust Soc Am 67:1157–1165 Rokhlin SI, Wang YJ (1992) Analysis of boundary conditions for elastic waves. J. Acoust. Soc. Am. 91:1875–1887 Rokhlin SI, Wang W (1989) Critical angle measurement of elastic constants in composite materials. Journal of Acoustical Society of America. 86:1876–1882 Rose JL, Pilarski A, Huang Y (1990) Surface wave utility in composite material characterization. Res Nondestruct Eval 1:247–265 Royer D, Dieulesaint E (2000) Elastic waves in solids. Springer, Berlin Salawu OS (1997) Detection of structural damage through changes in frequency: A review. Eng Struct 19:718–723 Sampaio RPC, Maia NMM, Silva JMM (1999) Damage detection using the frequency response function curvature method. J Sound Vibr 226:1029–1042 Saravanos DA, Birman V, Hopkins DA (1994) Detection of delaminations in composite beams using piezoelectric sensors. Proceedings of the 31th AIAA/ASME/ASCE/AHS/ASC structures, structural dynamics and materials conference, pp 181–191 Shen MMH, Grady JE (1992) Free vibrations of delaminated beams. AIAA J 30(5):1361–1370 Shkerdin G, Glorieux C (2004) Lamb mode conversion in a plate with a delamination. J Acoust Soc Am 116:2089–2100 Shkerdin G, Glorieux C (2005) Lamb mode conversion in an absorptive bi- layer with a delamination. J Acoust Soc Am 117:2253–2264 Simonetti F (2004) Lamb wave propagation in elastic plates coated with viscoelastic materials. J Acoust Soc Am 115:2041–2053 Singer L (1997) Bond strength measurements by ultrasonic guided waves. Ultrasonics 35:305–315 Sohn H, Farrar CR, Hemez FM, Shunk DD, Stinemates DW, Nadler BR (2003) A review of structural health monitoring literature : 1996–2001. Los Alamos National Laboratory Report, LA-13976 MS
50
V. Bucur
Sohn H, Park G, Wait JR, Lomback NP, Farrar CR (2004) Wavelet – based signal processing for detecting delamination in composite plates. Smart Mater Struct 13:153–160 Sohn H, Park H, Law KH, Farrar CR (2005) Instantaneous online monitoring of unmanned aerial vehicles without baseline signals Proceedings of the 23rd international modal analysis conference (IMAX XXIII), Paper no 259 Solodov IY, Krohn N, Busse G (2002) CAN: an example of nonclassical acoustic nonlinearity in solids. Ultrasonics 40:621–625 Sridharan S (Ed) (2008) Delamination behavior of composites. Woodhead Publlishing, Cambridge, England Stoessel R, Krohn N, Pfleiderer K, Busse G (2002) Air-coupled ultrasound inspection of various materials. Ultrasonics 40:159–163 Su Z, Ye L, Bu X (2002) Evaluation of delamination in laminated composites based on Lamb waves methods: FEM simulation and experimental verification. Proceedings of the 1st European workshop on structural health monitoring, pp 328–335 Sundararaman S, Adams DE, Rigas EJ (2005) Characterizing damage in plates through beamforming with sensor arrays. Proceedings of the 23rd international. modal analysis conference. (IMAX XXIII), Paper no 249 Tenek LH, Henneke EG II, Gunzburger MD (1993) Vibration of delaminated composite plates and some applications of nondestructive testing. Compos Struct 23:253–262 Terrien N, Osmont D, Royer D, Lepoutre F, Déom A (2007) A combined finite element and modal decomposition method to study the interaction of Lamb modes with micro-defects. Ultrasonics 46:74–88 Toyama N, Okabe T (2004) Effect of tensile strain and transverse cracks on Lamb wave velocity in cross – ply FRP laminates. J Mat Sci 39:7365–7367 Tracy JJ, Pardoen GC (1989) Effect of delamination on the natural frequencies of composite laminates. J Comp Mat 23:1200–1215 Tseng WY, Dugundji J (1971) Nonlinear vibrations of a buckled beam under harmonic excitation. J Appl Mech 38(6):467–476 Vakakis AF (1996) Normal modes and localization in nonlinear systems. Wiley, Chichester Wahl F, Schmidt G, Forrai L (1999) On the significance of antiresonance frequencies in experimental structural analysis. J Sound Vibr 219:379–394 Wang BS, He ZC (2007) Crack detection of arch dam using statistical neural network based on the reductions of natural frequencies. J Sound Vibr 302:1037–1047 Wang JTS, Liu YY, Gibby JA (1982) Vibration of split beams. J Sound Vibr 84(4):491–502 Wang SS, Yu TP (2006) Nonlinear mechanics of delamination in fiber – composite laminates: asymptotic solutions and computational results. Compos Sci Technol 66:766–784 Wright WMD, Hutchins DA, Hayward G, Gachagan A (1996) Ultrasonic imaging using laser generation and piezoelectric air-coupled detection. Ultrasonics 34:405–409 Wu F, Thomas JL, Fink M (1992) Time reversal of ultrasonic fields Part II Experimental results IEEE Trans Ultrason Ferroelectr Freq Control 39:567–578 Zou Y, Tong L, Steven GP (2000) Vibration – based model – dependent damage (delamination) identification and health monitoring for composite structures – a review. J Sound Vibr 230: 357–378
Chapter 4
Initiation and Growth of Delamination in Wood and Wood-Based Composites, a Fracture Mechanics Approach Voichita Bucur
Contents 4.1 4.2
Introduction . . . . . . . . . . . . . . . . . . . . . . Links with Fracture Mechanics . . . . . . . . . . . . . 4.2.1 Linear Elastic Fracture Mechanics . . . . . . . . 4.2.2 Nonlinear Fracture Mechanics . . . . . . . . . . 4.3 Micro-structural Aspects in Wood . . . . . . . . . . . 4.4 Micro-structural Aspects in Wood-Based Composites . . 4.5 Fracture Mechanics Parameters for Ecological Relevance 4.6 Summary . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
. . . . . . . . .
51 52 60 72 79 84 91 91 92
4.1 Introduction Fracture Mechanics concept has been applied to wood material as reported during more than fourthly years, by numerous references, review articles and books. Some of them are cited below. (Attack et al. 1961; Porter 1964; Boyd 1973; Schniewind and Centeno 1973; Schniewind and Lyon 1971, 1973; Schniewind and Pozniak 1971; Leicester 1971, 1973, 1974; Pearson 1974; Jeronimidis 1976, 1980; Schniewind 1977; Barrett 1976, 1981; Schniewind et al. 1982; Valentin and Morlier P 1982; Jung and Murphy 1983; Petterson and Bodig 1983; Boatright and Garrett 1983, Triboulot et al. 1982, 1984; Tschegg 1986; Patton – Mallory and Cramer 1987; Gustafsson 1985; Boström 1988, Akande and Kyanka 1990; Valentin et al. 1991; Aicher 1992; Aicher et al. 1993, 1998; Stanzl-Tschegg et al. 1994, 1995, Zink et al. 1994, 1995; Renaud et al. 1996; Gibson and Ashby 1997, Bodner et al.1997; Thuvander and Berglund 1998; Tschegg et al. 2001; Sippola and Frühmann 2002; Cotterell 2002; Reiterer and Sinn 2002; Smith et al. 2003; Vasic V. Bucur (B) CSIRO, Materials Science and Engineering Div. Bayview Avenue, Clayton, Victoria 3168, Australia e-mail:
[email protected] V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_4,
51
52
V. Bucur
and Smith 2002, 2003; Conrad et al. 2003; Nairn 2006; Vasic and Stanzl-Tschegg 2007; Nairn 2007a, b, c; Hofstetter et al. 2007). The increasing interest on the physical phenomena which lead to the onset of delamination, its development and its interaction with other damage mechanisms is determined by the expected economical benefits for wood products and various wood-based composites structures. In order to avoid over dimensioning in structural elements design, it is necessary to understand as deep as possible, the physics behind the damage mechanisms and to develop theories and tools (analytical or numerical) able to take into account the onset and growth of delamination from the earliest phases of design. The applications of wood-based laminated composites are limited by delaminations which can be introduced during the fabrication process or later in service life. The presence of delaminations degrades the stiffness, strength and fatigue characteristics of structural elements and has the potential to cause catastrophic failure of the structures. In this chapter are analysed the basic concepts related to fracture mechanics which allow the understanding of initiation and growth of delamination in wood and wood-based composites. Basic theoretical approaches and the state of the art for characterization and predicting delamination are outlined.
4.2 Links with Fracture Mechanics The initiation of delamination is due to the initiation and growth of cracks. As described by Williams (1989) the crack is “a planar discontinuity which is not capable of transmitting a load normal to its faces. When it grows, new surface area is created, which is of fundamental importance in determining behaviour”. The conditions for crack growth have been studied with Griffith theory (Griffith 1920) and with modern fracture theory using the concepts of linear or nonlinear Fracture Mechanics. Figure 4.1 gives the schematic representation of a crack, located in a plate (2D representation) and the corresponding two dimensional stress states. Any deformation of the crack can be described through a combination of three fracture modes (Fig. 4.2): – the opening mode in tension – Mode I – opening – the in plane shear mode – Mode II – sliding – the out of plane shear mode – Mode III – tearing shear
Fig. 4.1 Schematic representation of a crack in a plate and in an infinite solid (Triboulot et al. 1984, Figure 7)
4
Initiation and Growth of Delamination
53
Fig. 4.2 Three fracture modes: Mode I – tension; Mode II – in plane shear; Mode III out of plane shear
The definition of the damage zone ahead of a crack tip is crucial for the studies of wood fracture (Vasic et al. 2002). Because of the anisotropic nature of wood, defined with three principal axes L, R and T (longitudinal, radial and tangential) six different fracture system orientations can be defined such as: for crack propagation in L direction, the systems RL and TL, for crack propagation in direction R the systems LR and TR, and for crack propagation in T direction, the systems RT and LT. Note that the first letter indicates the direction normal to the crack plane and the second letter indicates the direction of crack propagation. In practice the crack’s path is very complex. The crack path for the systems RL and TL propagates always parallel to grain. However, the crack path for transverse directions TR and RT could propagates in any direction. When cracks propagation in R direction, two situations were observed, the path toward the bark or toward the pith. As reported by Attack et al. (1961) the toughness in green spruce in TR was 100 J/m2 and 180 J/m2 in RT direction. Schniewind and Centeno (1973) reported no differences between both directions in the stress intensity factor in air-dried Douglas –fir (0.35 MPam–2 ). Dill – Langer et al. (2002) noted that in softwoods crack growths in TR system in tension perpendicular to the grain is not steady and rupture of earlywood cell walls was observed. Another mechanism of rupture was observed when the crack growths in the RT system, namely the rupture between adjacent tracheids. Thuvander and Berglund (2000) observed the crack arrest in earlywood. Ashby et al. (1985) noted that in low density wood such as balsa the fracture propagates by cell wall rupture, while in high density wood species the fracture between cell walls, by peeling the middle lamellae was observed. Most studies on wood fracture mechanics rely on the concept of linear elastic fracture mechanics (LEFM), because of the simplicity of this approach. The concept of linear elastic fracture mechanics (LEFM) is based on the relationships existing between the stress in the vicinity of a crack tip and different characteristics of the structure such as: the nominal stress applied, the material mechanical and physical properties, the size, shape and orientation of existing flaws. This theory stipulates that the stress level in the vicinity of the crack tip tends toward infinity. In real materials, obviously there is a zone where the elastic solution breaks down. The size of the plastic zone dp at the crack tip in a material with σ Y yield strength, can be written such as: 1 dp = nπ
K σY
2
54
V. Bucur
where n = 1 for plane stress and n = 3 for plane strain; K is the stress intensity factor and is defined as: a √ K = σ π a.F W where: σ is the representative stress, a is the crack length and F(a/W) is a function of the geometry of the specimen LEFM can be applied if the size of the plastic zone is small compared with the dimensions of the specimen. The final scope of LEFM approach is the prediction of crack propagation conditions under the hypothesis that the material exhibits a linear elastic behaviour right up to the point where fracture occurs. Two relevant parameters for fracture phenomena studies were developed: – the stress intensity factor (K) which is based on the local stress distribution around a crack tip. Critical intensity factor (KC ) is considered a material parameter that defines the resistance to crack growth (referred also as fracture toughness of the material). – the strain energy release rate (G), which is based on the global energy balance The stress intensity factor (K) and strain energy release rate (G) will be described in that follows. The stress field around a crack tip has been documented in many reference books and we cite only the most recent (Sandford 2003; Anderson 2005). Using the notations from Fig. 4.1, the stress field in the immediate vicinity of a crack tip, for an isotropic solid, can be written such as: θ 3θ KI θ 1 − sin sin cos σx = √ 2 2 2 2π r θ 3θ KII θ 2 + cos cos −√ sin 2 2 2 2π r θ 3θ KI θ 1 + sin sin σy = √ cos 2 2 2 2π r θ 3θ KII θ cos cos −√ sin 2 2 2 2π r θ 3θ KI θ sin cos τxy = √ cos 2 2 2 2π r θ 3θ KII θ 1 − sin sin +√ cos 2 2 2 2π r
(4.2)
σz = 0 for plane stress
(4.4)
(4.1)
(4.3)
4
Initiation and Growth of Delamination
55
and σz = ν σx + σy for plane strain conditions θ KIII τyz = √ cos 2 2π r KIII θ τzx = √ sin 2 2π r
(4.5) (4.6) (4.7)
The constants KI , KII , KIII are termed stress-intensity factors for the Modes I, II or III and describe the intensity of the stress field as a measure of the severity of the crack. The stresses are singular at the crack tip for r = 0 and have a square root singularity. The displacements (u, v) under stress conditions can be written as:
θ θ r cos k − 1 + 2 sin2 2π 2 2
θ r KII 2 θ sin k + 1 + 2 cos + 2μ 2π 2 2
θ r KI 2 θ sin k − 1 + 2 cos v= 2μ 2π 2 2
θ KII θ r − cos k − 1 − 2 sin2 2μ 2π 2 2 θ r 2KIII sin w= μ 2π 2
KI u= 2μ
(4.8)
(4.9)
(4.10)
where μ is the shear modulus, ν is the Poisson’s ratio, k = (3 − v)/(1 + v) for plane stress and k = (3 − 4v) for plane strain. The strain energy release rate G is related to the work required to close a crack of length a + a to a length a, and is based on the Irwin’s crack closure concept (Irwin 1957). The total strain energy release rate G is expressed such as: G = GI + GII + GIII =
KI2 K2 K2 + II + (1 + ν) III E E E
(4.11)
where G I , G II , GIII are strain energy release for the modes I, II, III and E = E in plane stress and E = E/(1 − v2 ) in plane strain, E = Young’s modulus of the isotropic material. For orthotropic materials these parameters must be corrected with the corresponding elastic constants.
56
V. Bucur
For cracks in orthotropic materials, the stress distribution is much more complex, as described by Wu (1967), Walsh (1972), Wang (1984), Tada et al. (2000), Raju and O’Brien (2008), and Sridharan (2008) . It was generally admitted, that under short load duration the dry wood exhibits brittle fracture and linear elastic behaviour. This statement implicitly requires the theory of linear elastic fracture mechanics (LEFM) for the description of wood behavior. The main limitations of (LEFM) are: – the necessity of assuming the existence of a crack, – the effects of fracture process zone are in the vicinity of the crack tip, – the available energy goes into the creation of a single new fracture zone. However the linear elastic assumption is not suitable for examining the viscoelastic behaviour of wood, the mechano-sorptive effect, the scale effect, the mechanical behaviour under long term loading, the microstructural phenomena, etc. For these cases the quasi – brittle fracture is assumed and the phenomena are studied with the nonlinear fracture mechanics. The nonlinear fracture mechanics (NLEFM) introduces the notion of planar process zone where cohesive stresses are assumed to occur (Boström 1988, 1992; Patton – Mallory and Cramer 1987; Gustafsson 1985, 1988; Vasic and Smith 2002). In such materials, the fracture is preceded by localized phenomena in the plastic zone, the damage is assumed to occur on a surface, and a nonlinear region can be detected prior to the peak load, followed by strain softening region after the peak. The crack tip opening displacement (CTOD) can integrate these phenomena and can be used to model fracture under conditions of large plastic deformation. For fracture to occur there must be a critical crack tip opening (δ) which can be calculated as: δ=
4 K2 π EσY
(4.12)
The stability of a crack depends on the interaction of the applied loads and the material toughness. When unstable the cracks can growth with different velocities. However, it is the whole system which has the property of stability and not the crack itself. Crack initiation and crack propagation are best characterized by the fracture energy, whereas the stress intensity factor only gives information on crack initiation. In wood like in other solids, the fracturing under mechanical loading takes place in three steps namely, crack initiation, crack propagation and fracture. During crack initiation a process zone is formed in front of the crack tip, with numerous microcracks. The microcracks constitute the delamination front which profuse micro cracking ahead of the delamination front. The coalescence of existing microcracks forms macrocracks which propagate. During crack propagation, in the weak zone, behind the crack tips bridging effect takes place, which becomes gradually weaker until rupture occurs, as the complete separation of fracture surfaces. In solid wood and wood-based composites bridging process induces energy dissipation which strongly influences their fracture behaviour.
4
Initiation and Growth of Delamination
57
Smith and Vasic (2003) noted that in wood mechanically loaded, “cracks start to grow from microscopic defects in the cell walls and cell boundaries. As these small cracks accumulates, the compliance of the material increases. Prior to peak load there is a localisation process in which the damage that causes failure becomes more confined to a narrow region. By the time peak load is reached, a critical crack accompanied by a fracture process zone has been established, and strain softening can occur. The reason the fracture is not sudden, is that toughening mechanisms have been mobilised near the crack tip, causing energy to be dissipated more gradually”. The concepts developed with NLEFM are: (a) crack tip opening displacement (b) crack growth resistance curve or R-curve, the energy required for the propagation of a crack of unit area [J/m2 ] (Yoshihara 2001, 2003, 2004, 2005, 2006a, b; Morel et al. 2002, 2003, 2005; Coureau et al. 2006). Figure 4.3 explains theoretical behaviour of materials exhibiting bridging zone. Bridging zone can extend from the initial crack tip x0 to the notch root at xroot
a
b
c
Fig. 4.3 J-integral paths and softening curves. (a) J-integral analysis along the path 1 . . . 6 . The bridging zone develops from the crack tip at x0 to the notch root at x root . (Nairn 2009, Figure 1); (b) in bridging modelling, the crack opening displacements normal and tangential to the crack surface can be described by different softening functions such as : A – linear elastic, B – linear elastic brittle C – triangular with initial linear regime followed by a linear softening regime (Dourado et al. 2004), D – arbitrary traction, often approximated with a cubic function, E- linear softening, F-nonlinear softening function (Schmidt and Kallske 2007, Figure 2); (c) crack opening with microcracking and bridging components (Stanzl-Tschegg et al. 1995, Figure 2)
58
V. Bucur
and can be analysed with J-integral paths noted i (Fig. 4.3a). The J-integral [energy/unit area, or J/m2 ] is the path contour integral at crack tip which takes into consideration the stress vector acting perpendicular to the contour, the displacement vector and the strain energy density. It is to note that tractions act normal and tangential to the crack surfaces. The traction forces depend on the corresponding crack opening displacements, increasing to a peak (or cohesive stress) and decreases to zero when the tractions fails (corresponding to critical opening displacement). Materials can exhibit softening behaviour as shown in Fig. 4.3b. In case of wood, most frequently the bilinear and the polynomial functions were used. Figure 4.4c shows the bilinear softening model which explains the development of the microcracking component and the bridging component as suggested by Stanzl-Tschegg et al. (1995). Some other functions were used in finite element simulation of crack growths such as bilinear and trilinear (Douardo et al. 2004; Coureau et al. 2006a) or nonlinear (Schmidt and Kallske 2007). (c) energy release rate expressed by J integral is the energy that is extracted through the crack tip singularity.
Fig. 4.4 Theoretical behaviour of materials exhibiting bridging zone, with J integral paths (Coureau et al., 2006). zone 1 – onset of softening behaviour at GR (a0 ), the resistance GR defining the onset of the crack propagation of the equivalent elastic crack; zone II – progressive increase of the resistance to crack growth au. R-curve depends on the sharp of the softening behaviour, the ultimate load depends on the slope of the softening curve; zone III – crack propagation at constant resistance GR (a > ac ) = GRc ; zone IV – propagation at constant resistance to crack growth, when successive failures of interface element located ahead of the crack tip. (Note that the experiments were with spruce, Figure 12)
4
Initiation and Growth of Delamination
59
Figure 4.4 illustrates the influence of different parameters describing the softening behaviour on three curves, namely COD –load curve (Fig. 4.4a), Rcurve − GR (GR is energy release rate from LEFM) curve (Fig. 4.4b) and, w (relative displacement in tangential direction of the interface obtained from the upper and lower substrates) – σ curve (Fig. 4.4c). Coureau et al. (2006) described four zones: zone I – onset of softening behaviour at GR (a0 ), the resistance GR defining the onset of the crack propagation of the equivalent elastic crack zone II – progressive increase of the resistance to crack growth au . R-curve depends on the sharp of the softening behavior, the ultimate load depends on the slope of the softening curve. zone III – crack propagation at constant resistance GR (a > ac ) = GRc zone IV – propagation at constant resistance to crack growth, when successive failures of interface element located ahead of the crack tip. (Note that the experiments were with spruce) From COD –load curve (Fig. 4.4a) one can see the evolution of compliance (λ) as a function of initial crack length (a0 ) and critical crack length (ac ). Crack propagates at G = GRc ∀a. The levelling of Rcurve (Fig. 4.4b) might indicate that in wood, the toughness mechanism do not tends to infinite, where crack bridging requires sufficient deformation to produce closing forces (Smith, Landis et al. 2003). The softening behaviour of the cohesive crack is shown in (Fig. 4.4c). The normal stress transmitted by the interface decreases progressively from the interfacial normal strength (ft ) to 0, when critical opening displacement (wc ) is generated. Numerical methods can be used to evaluate the J- integral for any crack, type of loading and body configuration (Atluri 1986; Anderson 2005). Since numerical analyses are time consuming, simplified approaches for engineering calculations have been developed (Berto and Lazzarin 2007). The limitations of NLFM are related to J integral. Theoretically, the utilisation of this parameter is based on the elastic response of the material. However it is assumed that the nonlinear elastic material will not have permanent deformation. J integral is appropriate for monotonic loading conditions (where material unloading behaviour is not significant) and for small newly form process zone when crack advances due to the creation of stress free surfaces. A compromise between the LEFM and NLEFM has been proposed through the development of Damage Mechanics which is a phenomenological approach for material that do not exhibit plastic deformation and can not de characterized by brittle rupture. In such materials the formation of microcracks, defined as damages, induced stiffness decreasing which can be quantified by a damage variable which express the magnitude of this stiffness decreasing. Using Damage Mechanics (DM) approach Daudeville (1999) simulated the fracture in wood, by treating the problem of crack initiation in “originally uncraked “ structure of spruce specimens loaded in bending and by comparing the load displacement curves obtained with LEFM and DM. Both approaches correctly predicted the load-displacement curves. Moreover, the critical energy rate (parameter of LEFM) and the fracture energy (parameter of
60
V. Bucur
DM) where compared with the experimentally determined dissipated energy to fracture of the specimens. It was demonstrated that in both approaches fracture energy is the major parameter that governs crack propagation in wood. In that follows concepts related to LEFM and NLEFM will be discussed in more details in view of application to solid wood and wood-based composites.
4.2.1 Linear Elastic Fracture Mechanics The development of the concepts related to the application of LEFM required several hypotheses (Stanzl-Tschegg et al. 1995; Tschegg et al. 2001; Vasic et al. 2002; Vasic and Smith 2002; Jensen 2005a, b, c; Keunecke et al. 2007) such as: – the homogeneity of the linear elastic material in which fracture takes place – the pre-existing crack propagates always along one direction – crack-tip displacement is associated with three principal pure modes of fracture, Mode I, Mode II, Mode III. – the intensity of stress distribution in the vicinity of the single crack tip is fully characterized by the stress intensity factors by three intensity factors, KI , KII , KIII , associated with three principal pure modes of fracture – crack surfaces are traction free at all stages of loading – the crack propagates dynamically at a certain velocity once the critical fracture toughness (KC ) or strain energy (GC ), release has been reached – the inelastic process zone is limited to a small volume at crack tip. The experimental conditions that influence the fracture process in wood are: the geometry of the specimens, the loading orientation and rate, and the moisture content. Wood fracture toughness is also strongly dependent on wood species and density. The most common geometry of specimens used for the measurements of fracture toughness in Mode I and Mode II are shown in Fig. 4.5. The specimens can be tested in tension, bending, or shear. The effect of loading rate on wood fracture toughness has been studied by Conrad et al. (2003) and Vasic, Ceccotti et al. (2009). Conrad et al. (2003) noted that substantial crack growth can take place at low strain rate, whereas at high strain rates higher toughness values were measured. In this late case, the dissipation of energy is slow down because of the relatively short time of the process. Vasic, Ceccotti et al. (2008) noted that the fracture resistance curves at deformation speed between 0.05 and 200 m/min is influenced by the structural inertial effect. The twice-as-high fracture resistance at 200 m/min deformation rate proves the existence of a critical deformation rate above which the viscoelastic response of wood is suppressed. This phenomenon can characterize the ductile brittle transition limit for wood. As regards the loading orientation Table 4.1 gives some experimental values of fracture toughness in Mode I and Mode II determined for different species. As can be seen from this table, wood anisotropy is well expressed by the values of KIc . For example, for Mode I, for Douglas fir the values are such as: LR LT RL TR RT TL KIc· > KIc· > KIc· > KIc· = KIc > KIc·
4
Initiation and Growth of Delamination
b1
Fig. 4.5 (continued)
61
b2
62
V. Bucur
c1
c2
Fig. 4.5 Specimens geometry for fracture testing (a) fracture testing in mode I (Figure 4.5.a1) and mode II (Figure 4.5.a2) Yoshihara (2006, Figures 3 and 4). (b) Splitting test for macroscopic studies (Figure 4.5b) (Tschegg 1986, patent AT 390328) (c) splitting test for micromechanical studies in SEM chamber (Vasic et al. 2002, Figure 2) Figure 4.5.c1 loading device. Figure 4.5 c2 specimen
4
Initiation and Growth of Delamination
63
Table 4.1 Experimental values of fracture toughness in Mode I and Mode II determined for different species Species
Crack system Fracture toughness Plane
kNm–3/2
TL RL TR RT LT LR TL TL TL TL TL TL TL TL TL TL and RL
309 410 355 355 2417 2692 190 185 494 492 564 407 478 505 681 112
TL TL TL TL RL TL RL TL
1890 2187 1626 2143 2230 2190 1665 159
References
Mode I Douglas fir
Western white pine Western red cedar Hoop pine Hard maple Paper birch Red oak Lauan Messymate stringbark Maiden’s gum Balsa
Schniewind and Centeno (1971)
Johnson (1973) Johnson (1973) Walsh (1971) Johnson (1973) Johnson (1973) Johnson (1973) Johnson (1973) Walsh (1971) Walsh (1971) Wu (1963) cited in Wood handbook (1999)
Mode II White spruce Lodgepole pine Amabilis fir Douglas fir Red spruce Balsa
Barrett (1981) Barrett (1981) Barrett (1981) Barrett (1981) Wood handbook (1999) Wood handbook (1999) Wood handbook (1999) Wu (1963) cited in Wood handbook (1999)
The highest value of KLR Ic is explained by the fact that when crack propagates in R direction in LR plane, the transversal section of rays and tracheids is the major obstacle for crack propagation. Moreover KTL Ic has the smallest value because of the weakest split behaviour of wood in this plane; in this case it is suppose that the crack is initiated and then propagates in the middle lamella rich in amorphous lignin and poor in cellulose. As noted by Boatright and Garrett (1983) because of anisotropic and heterogeneous structure of wood, the TL system is “weak” and the LT system is “tough”. Similar remarks can be pointed out when fracture energy to failure (GC ) in tension perpendicular to grain is calculated for softwoods (Table 4.2). TL RL TL It was observed that always GRL C > GC and the ratio GC /GC is between 1.26 and 1.55. It is generally admitted that Mode I cracks propagate in a brittle manner with low energy consumption, whereas for Mode II cracking much energy is consumed in creating and breaking the hairy fragments that have been seen on the microfractographic images on the crack surfaces.
64
V. Bucur
Table 4.2 Mechanical and fracture characteristics of some species determined with splitting technique and FEA (data from Reiterer et al. 2002)
Spruce Alder Oak Ash
Young’s moduli (GPa)
Kinitial (N/m)
√ KIc (MPa m )
Gf (J/m2 )
EL
ER
ET
RL
TL
RL
TL
RL
TL
10 11.7 13 15.8
0.8 1.1 1.6 1.5
0.45 0.6 0.9 0.8
1.44 2.33 2.58 3.57
1.01 0.95 1.31 1.60
0.49 0.67 0.83 1.16
0.31 0.33 0.41 0.65
337 244 348 551
213 155 271 345
It is to note that for real structures the Mixed Mode is dominant and consequently this mode is of major interest for the studies related to health monitoring of structures. In wood very often a Mixed Mode I/II is possible because of the fact that cracks propagates along the fibres, irrespective of original crack orientation. Jernkvist (2001) proposed a theoretical model for a Mixed Mode I/II based on the fact that ” the Mixed Mode loading is supposed to displace the microcrack zone to one side of the main crack plane, and the coalescence of the microcracks with the parent crack may in this case require transverse cutting of tracheids walls. This process will create a rough crack surface which does not follow the fibre directions as can be seen in the simulation shown in Fig. 4.6. The quality of the surface observed in-situ with ESEM for spruce specimens loaded in Mode I in TR system by splitting technique is shown in Fig. 4.7. The wood structure depicted in this figure is perfectly localized on the load-displacement diagrams. The arrow at position 3a indicates the crack tip at – 20 N shortly before loading. The crack was located in the early wood zone with a razor blade. The crack front is widened, but no propagation occurred. The profile of the crack mouth opening is parabolic, wider in earlywood than in latewood. This image corresponds to the initial step – no crack propagation. At position 3b, in spruce the first propagation event occurs, the load dropped, the crack penetrated the latewood and stopped in the earlywood zone of the next ring. For beech specimens, the initial position is shown at the position 3c, corresponding to – 52 N. The profile of the crack is parabolic. For beech, the first propagation occurred at – 65 N at position 3d. In TR system and Mode I the behavior of spruce is different than that of beech. The behavior of different species (ash, oak, alder and spruce) related to the crack propagation in RL and LT, Mode I is shown in Fig. 4.8 with load displacement curves obtained by the wedge splitting test. In hardwoods a macrocrack initiation takes place at the maximum splitting force, followed by unstable crack propagation and several steps for crack arresting. The spruce specimens behaved very differently, showing a continuous load-displacement curve, with a maximum load peak related to a deviation from the linear behaviour. It was noted that “spruce displays more ductile and the hardwoods more linear elastic and brittle behaviour. Table 4.2 gives some fracture mechanics parameters deduced for ash, oak, alder and spruce with FEA. For all species the fracture RL parameters are higher then TL parameters and this is explained by the higher proportion of medulary rays. Table 4.3 gives the value
4
Initiation and Growth of Delamination
65
Fig. 4.6 Crack growths along the fibres in pure Mode I and in Mixed Mode I/II – theoretical models (Jernkvist 2001, Figure 5)
of fracture energy Gf in tension perpendicular to grain in different orientations, for spruce, fir and Sitka spruce. The ratio between the values in RL system and TL system is between 1.33 and 1.37. The fracture parameters are strongly influences by the density (Forest Products Laboratory 1999; Reiter et al. 2002). Figure 4.9a shows the variation of fracture toughness for different species with density. Fracture toughness in Mode I increases with density. The density range was between 100 kg/m3 and 800 kg/m3 and the fracture toughness was measured parallel and perpendicular to the grain. More refined studies were reported by Donaldson (1997) related to the variation of microdensity in fractured zones. Figure 4.9b, c shows the microdensity variation in a transwall fracture zone in the middle lamella region. There is a linear decrease of microdensity in the vicinity of the fractured surface. Studies related to Mode III in wood were possible with the development of a specific experimental devices as for example those proposed by Tschegg (1986), Ehart et al. (1998, 1999), Tschegg et al. (2001). It was point out by Tschegg et al. (2001) that for larch and beech, crack initiation energy in Mode III is over twice as high as Mode I in RL and TL fracture systems, because of a much larger fracture process zone in Mode III than in Mode I. Moreover the Mode III crack has
66
V. Bucur
Fig. 4.7 Load-displacement diagrams in situ and specimens of spruce and beech loaded in Mode I, inside the chamber of an environmental scanning electron microscope loaded in Mode I in the TR crack propagation system (Frühmann et al. 2003, Figures 2 and 3). The arrows show the position a, b, c, d on the load displacement diagrams when the corresponding images were taken such as: a and b for spruce and c and d for beech
ten times higher crack growth resistance compared to Mode I. “Under pure Mode III load, crack initiation takes place under Mode III in beech as well as in larch. More advanced cracks, however, propagate predominantly as Mode I. The change of the fracture mode takes place preferentially in RL orientation in beech and in TL orientation in larch” (Tschegg et al. 2001). This behaviour is related to the presence of medullary rays, much more numerous and important in size in beech than in larch. The influence of wood moisture content on fracture characteristics was thoroughly reviewed by Wang et al. (2003). The maximum fracture toughness was
4
Initiation and Growth of Delamination
67
Fig. 4.8 Mode I, Load-displacement curves obtained by the wedge splitting test in RL and TL system for ash, oak, alder and spruce (Reiterer et al. 2002, Figure 5)
reported at 17% wood moisture content. King et al. (1999) noted that the mode I fracture toughness was lower for wet wood, in all fracture directions, (Table 4.4) than for dry wood (radiata pine specimens) tested in bending (three point bending and single edge notched). In situ examination with environmental scanning Table 4.3 Fracture energy Gf in tension perpendicular to grain and in three point bending test, for specimens with constant width (b = 45 mm) [data from Daudeville 1999] Orientation Fracture energy Gc (J /m2 ) Species
Plan
Mean
Minimum
Maximum
Coeff. Variation (%)
Spruce
RL TL RL/TL RL TL RL/TL RL TL RL/TL
220 160 1.37 210 157 1.33 220 164 1.34
159 100 1.59 126 97 1.29 157 136 1.15
345 247 1.39 367 236 1.55 248 196 1.26
19 29 − 26 37 − 16 16 −
Ratio Fir Ratio Sitka spruce Ratio
68
Fig. 4.9 (continued)
V. Bucur
4
Initiation and Growth of Delamination
69
Fig. 4.9 √ Influence of density on fracture behaviour. (a) Variation of fracture toughness, Mode I KIc (MPa m) versus density (Conrad et al. 2003, Figure 5) (b) fractured zone in middle lamella region (Donaldson 1997, Figure 5) (c) Microdensity variation in a fractured zone (Donaldson 1997, Figure 6)
microscope has shown that in green wood water droplets moved away from the cell lumen around the crack tip. During drying microcracks were observed. Crack bridging is part of toughening mechanisms. Based on in-situ experiments with ESEM, Vasic and Stanzl-Tschegg (2007) have shown the influence of moisture content on fracture toughness and fracture energy (Fig. 4.10) on several European species. Three main regions of moisture content can be observed, in which the influence of wood structure is obvious. – region between 5% and 12%, in which ◦ Gf decrease for beech and oak and increased for pine and spruce ◦ KIc decreased for beech, oak spruce and pine Table 4.4 Fracture toughness in Mode I and Mode II for dry and wet Pinus radiata specimens (data from King et al. 1999) √ Fracture toughness [MPa m] Fracture
Wood
RL
RT
TL
TR
LT
LR
Mode I
Dry Wet Dry/wet Dry Wet Dry/wet
0.486 0.214 ∗∗∗ 2.826 2.328 ∗∗∗
0.351 0.236 ∗∗∗ 1.088 0.458 ∗∗∗
0.282 0.270 ns 2.664 1.905 ∗∗∗
0.195 0.235 ∗∗∗ 1.228 0.443 ∗∗∗
2.69 2.21 ∗∗∗ – – –
2.39 1.88 ∗∗∗ – – –
Test Student( t) Mode II Test Student (t)
NB: Test Student ∗∗∗ confidence level 90%
70
V. Bucur
Fig. 4.10 Influence of moisture content on some fracture parameters in RL crack propagation (Vasic and Stanzl-Tschegg 2007, Figure 7). (a) total fracture energy Gf (N/m) versus moisture content; (b) fracture toughness KIc (kNm−3/2 ) versus moisture content
– region between 12% and 18 %, in which ◦ Gf increased for all species ◦ KIc increased for beech and oak but decreased for spruce and pine
4
Initiation and Growth of Delamination
71
– region between 18% and 30%, in which ◦ Gf is relatively constant for oak and spruce and decreased for beech ◦ KIc decreased for spruce and pine From this data it seems evident that the moisture content has the ability to change wood fracture mechanism from brittle to ductile. Vasic and Stanzl-Tschegg (2007) noted that the stress gradient at the crack tip might have a significant effect on the local moisture distribution, free water flow and vapour diffusion in the vicinity of the crack. Experiments on green wood and modelling with discrete finite elements (Frühmann et al. 2003; Vasic and StanzlTschegg 2005, 2008; Sedighi – Gilani and Navi P 2007) has shown that the process zones are confined to one or only a few cell rows, and the lattice fracture model shown distributed damage in the most stressed regions between the area where a concentrated force is applied and, the notch plane where the fracture is initiated. The aspects discussed previously have proved the limitations of LEFM concepts (synthesized in Table 4.5) and the necessity to introduce new concepts.
Table 4.5 Limitation of LEFM for wood fracture studies (data from Vasic et al. (2002) Linear elastic fracture mechanics 1
2
3
4
5 6
7
Wood is a homogeneous linear elastic medium (isotropic or orthotropic)
Comments
Wood is heterogeneous, cylindrically orthotropic with discontinuities on macro and micro structural levels. Brittle fracture occurs in an elastic range The pre-existing crack always propagates The crack does not grow along the original along the original crack direction orientation. The initial crack extension is always parallel to the grain, even when starter crack lies across the grain. At microscopic level fractured surfaces are irregular and tortuous Crack tip displacements can be separated into Only displacements can be separated into three different modes (Mode I – in plane three independent modes tension, Mode II – in plane shear, Mode III – out of plane shear) The stress intensity factors KI, KII KIII fully The simplicity of K characterization with characterize the intensity of stress only one parameter for all complex fracture distribution in the vicinity of the single phenomena is no more acceptable crack tip The inelastic zone is confined to a small The inelastic zone is not small volume of crack tip Crack surface are traction free at all stages of The crack surfaces are not traction free and loading, and the crack tip is anatomically not anatomically sharp. See the sharp “ligamentary bridging “of fibres. Crack bridging is part of toughening mechanisms The crack propagates dynamically at some The stability and rate of crack velocity can be terminal velocity controlled through appropriate choice of the rate of loading and experimental configuration.
72
V. Bucur
4.2.2 Nonlinear Fracture Mechanics Nonlinear Fracture Mechanics (NLEFM) is more appropriate the LEFM to describe wood behaviour in several practical situations such as: fracture beyond initial cracking, creep rupture effects, size-effect on small clear specimens, adhesive joints, etc. To estimate the level of non linearity either fracture energy methods or stress-based concepts can be applied. Comparative studies on linear and non linear fracture mechanics on wood have been performed by Patton-Mallory and Cremer (1987), Boström (1992), StanzlTschegg et al. (1995), Tan et al. (1995), Reiter and Sinn (2002), and Vasic and Smith (2002), using non linear energy based fracture theory in order to quantify the relevance of deviations from the theoretical brittle response. It was admitted that wood has a softening behaviour (Vasic and Smith 1996a, b, 1998, 1999a, b). The apparent non linearity in the fracture response beyond the peak load is attributed to the gradual development of damage and microcraking in a fracture process zone around the crack tip. Stanzl-Tschegg et al. (1995) noted also that the fracture mechanism in wood is not purely brittle. Vasic and Smith (2002) explained the non linear behaviour of spruce in Mode I, by fibres bridging behind the crack tip, in the presence of stress concentrations. The bridging crack model propose by Vasic and Smith (2002) assumed that the sharp crack tip coexist with a bridging zone behind the tip crack (Fig. 4.11a, b). The variation of energy release rate and fracture toughness versus the crack length is shown in Figs. 4.12 and 4.13 on which “the influence of bridging stresses clearly increases with any increase in the crack length, if the maximum bridging stress is kept constant“. The fracture parameters reach a maximum at 4 mm which corresponds to the tracheids length of spruce. Vasic and Smith (2002) demonstrated that bridging of the fibres behind the crack tip is a major factor in toughening mechanism in wood. They confirmed the previous statements of Boström (1992) and Tan et al. (1995) that wood in fracture has a non linear behaviour similar to concrete. “The nonlinearity beyond the peak load was attributed to gradual development of damage in a fracture process zone around the crack tip”.
Fig. 4.11 SEM micrograph (Eastern Canadian spruce) of a crack tip (Vasic and Smith 2002, Figures 2, 4) (a) the crack tip coexists with a bridging zone behind the tip crack, towards the end of the experiment. The bar line is 100 μm. (b) Fibers bridging. The bar line is 100 μm
4
Initiation and Growth of Delamination
73
Fig. 4.12 Energy release rate Gfc versus crack size for end-tapered DCB specoimen (Vasic and Smith 2002, Figure 11)
Table 4.6 synthesizes the crack models suggested for wood fracture studies with nonlinear fracture mechanics at overall macroscopic level. As underlined by Landis and Navi (2009), these models break away from classic continuum framework, referred to the cross grain fracture of wood, represent material heterogeneity and used FEA with different stress – crack opening functions (linear, bilinear, trilinear, non linear). All these models ask for high computational expenses.
Fig. 4.13 Fracture toughness versus crack length for end-tapered DCB specimen (Vasic and Smith 2002, Figure 12)
74
V. Bucur Table 4.6 Crack models, with a crack interface element, for wood NLEFM studies Model description
1
2
3
4
5
6
Comments
– fictitious model – existing of a cohesive zone – Cracks closing stresses act – FEA as a function of crack – the model is simple and separation distance incorporates the – Stress and crack opening are nonlinearities in to the related by Gf closing stress function – the fracture energy is predicted from the area under the closing stress versus opening function – post peak softening – pine specimens parameter and fracture – FEA energy – separate the microcracking – wedge-splitting tests – FEA from crack bridging – softening curve with contribution to fracture bilinear representation of energy stress-crack opening – the bridging components in relationship RL > TL, due to rays normal to the crack plane – bridging model – Canadian spruce specimens, – bridging stress occur on the Mode I crack faces close to the – in situ ESEM crack tip – the strength of the bridging – bridging zone length = stress determine whether 4 mm = tracheids length, fracture is brittle, which is the intrinsic quasi-ductile or ductile material length scale to a – combination on FEA and continuum fracture model ESEM observations – pine and spruce specimens, – crack propagation occurs Mode I when peak tensile stress is – bilinear and trilinear reached constitutive relationship for – crack interface element crack interface element represents the closing – in trilinear model the stresses softening is broken down – whole fracture process zone into microcracking and is lumped into a crack line bridging phenomena and is characterized by the – load displacement curves stress-crack opening law and R curves which exhibits softening – FEA – cohesive crack simulations – crack face friction is – Mode II and Mixed Mode negligible – measured values for GIc and GIIc – bilinear constitutive relationship for crack interface element -
References Hillerborg et al. (1976) Hillerborg (1991) Homberg et al. (1999)
Boström (1988, 1992)
Stanzl-Tschegg et al. (1995)
Vasic and Smith (2002)
Dourado et al. (2004, 2008)
Silva et al. (2006)
4
Initiation and Growth of Delamination
75
Table 4.6 (continued) Model description 7
8
Comments
References
Coureau et al. (2006a, b) – spruce specimens, crack and RL, Mode I – stress opening relationship Lespine (2007) is initial linear elastic regime – brittle law – the peak load of a lod –crack opening displacement curve is strongly affected by the slope of softening behaviour – the roughness is related to the microstructural features and toughness – the effect of crack closing stress function, the load crack opening displacement and R curve are related quantitatively – FEA Schmidt and Kallske (2007) – 3D anisotropic constitutive – nonlinear, continuous softening stress- crack law is implemented, which opening function – realistic could loaded and unloaded incorporate damage and load behaviour history – FEA
– cohesive crack model – elastic layer model – tensile strength σt and overall fracture energy Gf , are the most important properties of the cohesive zone – the fracture energy related to the constitutive law and must correspond to the plateau value of the R-curve – crack resistance of R-curve behaviour is related to the roughness of fracture surface
Morphological based models – lattice models and material point model – has been developed to understand the structural complexity of wood and to relate micro and macro mechanical behavior. Lattice models have been developed by Landis et al. (2003), Davids et al. (2003), Wittel et al. (2005), Vasic and Stanzl-Tschegg (2007), Mishnaevsky and Qing (2008), and Landis and Navi (2009). Material point model (MPM) is a very recent and promising model that discretized the solid in an array of points, developed by Nairn (2006–2009), Guo and Nairn (2006). Figure 4.14 shows the numerical modeling of wood structure when fracture occurs in TR plane. Figure 4.15 shows a digitized image of Douglas fir specimen with a notch (a), the corresponding MPM converted image for radial direction on a scale of 0◦ –90◦ from white to black (b), the crack growth started with an initial kink (c) and the simulated crack growth (d). With MPM to each point specific properties such as stiffness and toughness can be assigned. In a numerical study of the transverse modulus of wood as a function of grain orientation and properties including heterogeneity and anisotropy Nairn (2007b, c) demonstrated in a very elegant manner the feasibility of the material point model using different degrees of complexity for the mechanical behaviour of wood, ranging from the simplest transverse isotropic hypothesis to the more
76
V. Bucur
complex heterogeneous cylindrical orthotropy (Table 4.7). The material point model requires data for reliable mechanical parameters of wood structural element that can be obtained through the development of new methodology such as acoustic microscopy (Bucur et al. 1995; Clair et al. 2000; Bucur 2003, 2005) or other ultrasonic techniques as demonstrated by Bucur et al. (1994). As a conclusion of this section, it can be noted that analytical and numerical models have been successfully developed for wood structure and fracture mechanics studies. Gibson and Ashby (1997) derived an analytical model for wood structure described as a regular array of hexagonal cells and derived results for initiation of fracture by either elastic or plastic buckling. This is a 2D model and mimics only the softwood structure. The numerical modelling of wood structure is more complex and includes the finite element analysis, the lattice method and the material point method. Finite element analysis reduces the analysis to an idealized structure. The limitations of this approach are described by Smith, Landis et al. (2003) such as: – the wood structure is very complex, difficult to discretized into an FEA mesh – the common practice of reducing analysis to a small idealized structure limits its value for numerical modelling of the details of failure mechanisms, – the number of elements required to accurately mesh realistic wood morphology is computationally expensive – the difficulty to consider the contact between cells and the large deformations
Fig. 4.14 Material point numerical modelling of wood structure (Nairn 2007a, Figure 2)
4
Initiation and Growth of Delamination
77
Fig. 4.15 Crack growths with Material Point numerical modeling (Nairn 2007c, Figure 2, Figure 3) (a) digitized image of Douglas fir specimen with a notch (b) the corresponding MPM converted image for radial direction on a scale of 0–90◦ from white to black (c) the crack growth started with an initial kink (d) the simulated crack growth
The lattice methods – for which the wood structure is replaced by a model of rod and spring elements is limited to linear elastic material properties. Variations in wood structure have been introduced by allowing strength and/or stiffness properties of the elements to be statistical quantities. Lattice models have focused on longitudinal properties of wood where the rods are wood fibres and springs represent transverse properties. In principle lattice models could be applied to transverse properties or 3D modelling, but that capability has not been demonstrated.
78
V. Bucur
Table 4.7 Some hypotheses for wood structure modelling and fracture mechanics studies at annual ring level with material point method Model 1
2
3
4
5
Characteristics
Comments
Transverse isotropic – described by 5 elastic constants, – the simplest model material – Young’s moduli are as – valid mostly for tropical trees EL >> (10 . . . 20)ER ; . ER = ET . – miss low transverse shear – Finite Element Analysis modulus GRT – data from Nairn 2007 b – miss ring curvature and structure, EW, LW properties – miss ring curvature and Rectilinear – described by 9 elastic constants structure EW, LW properties orthotropic – simplify the analysis by aligning material coordinates of the anisotropy with – mesh generation relatively simple the rectilinear natural axes L,R,T – allow a low transverse shear modulus GRT – can describe the off-axis loading – Finite Element Analysis – data from Nairn 2007 b Homogenized – accounts for growth rings curvature – complicate mesh generation, compared to rectilinear cylindrical within a specimen, orthotropic material orthotropy – simplifies the analysis by using homogenized properties in transversal plane. – one can use large elements where stress gradients are small – small elements are required throughout the specimen in order to resolve orientation of material axes along curved growth rings. – can approximate effective mechanical properties, account for differences between pith and periphery boards, and account for size effects. – Finite Element Analysis – data from Nairn 2007 b – the closest approximation of – accounts for both growth ring Heterogeneous real wood structure curvature within a specimen and cylindrical – needs new methodology for variation in material properties orthotropy reliable values of such as EW and LW – fine mesh is required to resolve the EW, LW mechanical properties (acoustic microscopy Bucur structure of wood et al. 1995; Clair et al. 2000, – is recommended for modelling Bucur 2003; Bucur 2005 or failure processes induced by other ultrasonic techniques localized stresses – EW, LW Bucur et al. 1994) – Finite Element Analysis – data from Nairn 2007 b Monoclinc symmetry – described by 21 constants (Bucur – never used until now – 2009) and Rasolofosaon 1998)
4
Initiation and Growth of Delamination
79
The material point method for numerical modelling of wood structure is capable of modelling wood anatomy in more details than the methods described previously. The advantages of this method can be such as: – the facility to discretize realistic wood structures via a process of digitization of an image into pixels. – automatically handles contact and thus can be extended to high strain without numerical difficulty – can handle the specificity of the internal configuration of wood structure, the large deformations and the large calculations in relatively short time. – the material point model requires data for reliable mechanical parameters of wood structural element that can be obtained through the development of new methodology such as acoustic microscopy
4.3 Micro-structural Aspects in Wood In this section detailed micro-structural aspects of delamination in solid wood will be discussed. We have seen that the space – time multi-scale nature of the delamination process in wood can be related to the prediction of crack nucleation, growth and arrest. The initiation of failure can be marked by the first acoustic signal (Lee et al. 1995; Dill – Langer and Aicher 2000; Dill – Langer et al. 2002; Reiter et al. 2000; Bucur 2005), or by the non-linearity point on the load/displacement curve (Tschegg 2001; Frühmann et al. 2003). Crack initiation, crack growth and crack arrest emerge as natural outcomes of the imposed load. In all these processes wood microstructure plays a very important role. Studies on wood fracture in relation to its structure using optical microscopy have been published since several decades (Mark 1967; Debaise et al. 1966, 1972; Dinwoodie 1966, 1968, 1974; Jeronimidis 1976, 1980). Gordon and Jeronimidis (1980) suggested that the cells in the vicinity of the fracture zone can absorb a great quantity of energy before breaking. The helical structure of the cellulose microfibrils in the S2 and the helically wound pattern of the microfibrils induces a specific form of buckling failure in tension, which causes a high energy absorption during fracture (Jeronimidis 1980a, b). In his previous work, Jeronimidis (1976) emphasises the essential part played by the S2 layer in the fracture process upon longitudinal tension. Keith and Côté (1968) described the layer boundary S1–S2 of the secondary wall as the place where intra-wall failure arises as a result of shear strains. Experimental studies on hollow cylindrical tubes scale models with helical fibres at different winding angles showed that the optimal trade-off between stiffness and toughness can be observed at a microfibril angle of about 15◦ (Gordon and Jeronimidis 1980). Kucera and Bariska (1982) using tube multilayered model specimens for direct observation of the formation of failure in longitudinal compression noted that “cracks do not occur until the reduction in specimen length reaches a stage where the wall folds fill the whole lumen of the tube. They always arise parallel to the axis across one or more folds in the longitudinal direction, and
80
V. Bucur
starting on the outside they traverse groups of lamellae or the entire wall”. Thus, to identify the specific behaviour of hierarchical microstructure of wood is essential for further developments of advanced models of damage and fracture. During the 1980s a new step in understanding wood behaviour and delamination was achieved with fractographic studies using scanning electron microscope, ex-situ (Borgin 1971; Kucera and Bariska 1982; Bariska 1994; Bodner et al. 1996; Zimmermann et al. 1994; Donaldson 1997; Seel and Ziemmermann 1998; Ando and Ohta 1999). Some micrographs ex-situ are shown in Fig. 4.16 for spruce and in Fig. 4.17 for beech. In spruce loaded on Mode I and impact bending, brittle fracture was observed in latewood tracheids as well as delamination between S1 and S2 . Ductile fracture was observed in fracture in long term bending with specimens at 20◦ C and 65% relative humidity. The microfibrils are pulled out of the secondary
Fig. 4.16 Fracture morphology in spruce. (Zimmermann et al. 1994, Figures 3, 6, 8, 11) fracture in impact bending, with specimens at 20◦ C and 35% relative humidity. Latewood tracheids, brittle fracture with S2 clean surface. Delamination between S1 and S2 . (a) fracture in long term bending with specimens at 20◦ C and 65% relative humidity. Latewood tracheids, ductile fracture the microfibrils are pulled out of the secondary wall (b) fracture in impact bending, with specimens at 20◦ C and 35% relative humidity. Delamination of middle lamella (matrix) and secondary wall composed from microfibrils. (c) fracture in impact bending, with specimens at 20◦ C and 35% relative humidity. Latewood tracheids. “A fast and very brittle fracture led to a partially smooth fracture surface whereas the remaining part of the fracture was more ductile and exhibits a rough surface with a certain separation of microfibrils and matrix”
4
Initiation and Growth of Delamination
81
Fig. 4.17 Fracture morphology in beech in static bending and in impact bending (Seel and Zimmermann 1998, Figures 1, 2) (a) libriform fiber after short time static bending, strong deformation in tension zone of the cell wall with radial agglomerations (arrows) on S2 (b) libriform fiber after impact bending. Brittle fracture with delamination in fibril/matrix structure. Some microfibrils are oriented radially and some others are arranged in a layered structure
wall. Delamination was observed between the middle lamella (matrix) and S2 . The ductile fracture led to a relatively rough fracture surface. In beech libriform fiber in fracture in impact bending, brittle fracture with delamination between microfibril and matrix was observed. Some microfibrils are oriented radially and some others are arranged in a layered structure. Donaldson (1997) reported the aspects of the ultrastructure of transwall fracture surfaces in Radiata pine wood using transmission electron microscopy. The fracture initiation and growth was studied under tensile stress parallel to the cell wall layers. Figure 4.18 shows a tangential fracture of two adjacent cells in Pinus radiata loaded in tension. A delamination is observed where
Fig. 4.18 Tangential fractures of two adjacent cells in Pinus radiata loaded in tension (scanning electron micrograph). A delamination is observed where intra-wall fracturing undergoes a transition between cell walls layers (Donaldson 1997, Figure 1)
82
V. Bucur
intra-wall fracturing undergoes a transition between cell walls layers. Donaldson (1997) noted that the transwall fractures are on tangential surfaces and are more common outside the juvenile wood rings, This fractures are due to changes in cell dimensions and microfibril angle with cambial age (Donaldson 1996). “Transwall fractures are favoured by large cell size and by low microfibril angles, and occur only on tangential wood fractures in Pinus radiata. Radial wood fractures are exclusively intrawall due to the favourable energy conditions provided by the alignment of tracheids in radial files”. Most of the transwall fractures observed had an irregular aspect. In contrast, the intrawall paths tend to follow the lamellate structure of the cell wall matrix, producing smooth surfaces. These aspects are described in Fig. 4.19 and in Fig. 4.20. More details related to fracture and wood anatomy are given in the outstanding contribution of Donaldson et al. (1996) – Rotorua Laboratory, New Zealand. The development of the equipment for in-situ studies with ESEM (Bodner et al. 1996; Vasic et al. 2002; Smith and Vasic 2003; Turkulin et al. 2005; Vasic and Stanzl-Tschegg 2008) has been a big step towards the understanding of crack initiation and propagation in wood during loading. The equipment allowed loading operation very precisely. Small inconvenient can be introduced by the action of electronic beam which weakened the cell wall in S3 (Hoffmeyer and Hanna 1989). Electronic beam damage induced fractures have very characteristic patterns, different from the other mechanical fractures. Bodner et al. (1996), for tension tests on Norway spruce observed that” in samples with parallel growth rings cracks propagated with jumps”, (probably the system LR was tested). In specimens with perpendicular growth rings, “the initial cracks developed into the final fracture eruptively and, without intermission”. Serrated (saw tooth) fracture pattern occurs in S2 and S3 , the microfibrils are pulled out. Thuvander and Berglund (2000) described the micromechanics of fracture in radial growth cracks in green pine (Pinus sylvestris) specimens with in-situ optical microscope. Figure 4.21 shows the morphology of the radial cracks. At the cells
Fig. 4.19 Higher magnification (transmission electron micrograph) of the cross section of the transwall fracture. S2 follows the line of least resistance while S1 and S3 layers protrude from the fracture zone. The intrawall fracture is seen along the ML surface (Donaldson 1997, Figure 2)
4
Initiation and Growth of Delamination
83
Fig. 4.20 Transwall fracture in cross section with transmission electron technique (Donaldson 1997, Figure 3). (a) the smoothness of the intrawall fracture compared to the roughness of the transwall fracture. Transwall fractures (short arrows) are separated by intervening intrawall fractures (long arrows) between S1 and S2 and S1 and ML. (b) the irregular sawtooth fracture surface in a transwall fracture
level the crack tip propagates by separating cell walls at the middle lamella in a splitting or peeling mode. At the annual ring level “stick-slip type” of crack growth was observed. Because of non-uniform stress distribution, the cracks deviate from the pure radial direction namely in earlywood zone. The latewood fracture mostly is without plane deviation. When crack propagation in earlywood approaches the latewood zone, its growth rate decreases and could be arrested in earlywood. Latewood failure occurred mostly by cell splitting because of weak middle lamella. DillLanger et al. (2002) studied the in-situ the damage mechanisms of crack propagation in tension perpendicular to the grain, in spruce micro specimens (12 mm3 ) with initial notch. For spruce at 12% moisture content two different mechanisms were identified: the rupture of earlywood cell walls when crack propagation is in tangential direction, and debonding between adjacent tracheids, when crack propagates in radial direction. The cell wall rupture is related to the meso-scale behaviour of annual ring structure while the debonding mechanism is very brittle and related to the micro-scale wood behaviour. The development of in-situ techniques will serve to the modelling approaches and for implementation of non linear and anisotropic laws in different fracture models of wood and wood-based composites.
84
V. Bucur
Fig. 4.21 Fracture morphology for radial cracks in green sapwood of Pinus sylvestris observed in situ with under tension loading (Thuvander and Berglund 2000, Figures, 4, 7, 8, 10) (a) TR crack arrested in latewood (b) Two cracks are linked and the bridging zone is torn (c) Crack alignment in R direction because of the rays (d) TR crack tip in the middle lamella of earlywood. (mode of crack growth: cell splitting or peeling)
4.4 Micro-structural Aspects in Wood-Based Composites The structure of wood-based composites is spatially much more complex than that of wood as can be seen from Figs. 4.22 and 4.23 for the fracture surfaces of woodbased composites tested in tension fracture Mode I (Niemz and Diener 1999). The failure of adhesive layers introduces new problems in old and new structures. A delamination test for structural wood adhesives used in thick joints has been proposed by Lavisci et al. (2001). While the technology to produce wood-based composites has advanced significantly in last decades, the theories for predicting the behaviour of these materials advanced less. The industry needs reliable and specific
4
Initiation and Growth of Delamination
85
Fig. 4.22 Fracture surfaces of wood-based composites on ASTM E 399 – 1994 specimens type tested in tension fracture Mode I (Niemz and Diener 1999, Figure 1, Figure 2, Figure 3) (I) Front view of: (a)- OSB split parallel to the particle orientation; (b) a- OSB split perpendicular to the particle orientation; (c) MDF ; (d) plywood – 7 layers (II ) Fracture surfaces of (a)- OSB split parallel to the particle orientation; (b) a- OSB split perpendicular to the particle orientation; (c) MDF ; (d) plywood – 7 layers (III) ASTM E 399 Specimen used for delamination testing in wood based composites
86
Fig. 4.22 (continued)
V. Bucur
4
Initiation and Growth of Delamination
87
Fig. 4.23 MDF specimens size and orientation (Matsumoto and Nairn 2008, Figure 1); (a) compact tension specimen a= 30.48 mm w= 76.2 mm = 31.75 mm. Note = 0 for the ASTM specimen (b) orientation of specimens in a panel. The first letter indicates the normal to the crack and the second one the crack propagation direction
modelling techniques to predict the influence of species, size, engineering properties of the constituents, etc., on wood-based composites properties. Under normal conditions, strengths of wood-based composites are directionally dependent on the structural features and, on the other hand, are time dependent and sensitive to moisture content. Their heterogeneity is also their main source of weakness, irrespective of the nature of the constituents. As en example, in glulam, the interface between fibers and adhesive layer is critical for damage onset and development. The damage mechanisms themselves are numerous and closely connected. In many situations the most critical damage mechanism for composites design is the delamination between adjacent layers. Whatever the cause the delaminations, they can be very dangerous and can easily lead to a premature collapse of the structures. The physical phenomena behind delamination onset in wood-based composites can have the following causes: – the residual stress induced during manufacturing – the environmental conditions, such as moisture content and temperature gradients – machining and drilling producing peel-off of the uppermost plies and heat generation; the angle of penetration, the drill geometry, the fibre orientation, lay-up sequence are factors of influence – the geometrical configuration – free edge interlaminar stress, skin debonding, joints, tapered structures – the inclusions such as bolt, holes and notches The physical phenomena behind delaminations growth are induced by the application of any type of load – compression, tensile in joints, bending or fatigue. Fracture mechanics is the best tool for the identification of a threshold level for the growth of delaminations and have a fundamental importance in understanding the real mechanical behaviour of delaminated composite structures. When designing with wood-based composites the causes associated with the delamination failure
88
V. Bucur
must be taken into account. Delamination management approaches can also strongly influence the lifecycle costs and the maintenance costs of structures. In order to efficiently exploit the potential of wood-based composite materials in structural applications the delaminations initiation and growth must be adequately predicted and controlled. Multi-scale approaches simulating delamination related physical phenomena at different levels of detail with different degrees of accuracy were suggested by Ladis et al. (2002), Smith and Vasic (2003), Moses and Prion (2004), Smith et al. (2007), and Stanzl-Tschegg and Navi (2009). Since delamination and fracture process in wood-based composites is with fibers bridging across the crack plane, the general preferred approaches are based on energy release rate and not on stress intensity factor. Crack resistance curves (Ehart et al. 1996, 1998, 1999) were determined for some wood-based composites (particleboards, MDF, Parallam, etc) with wedge splitting technique, under the assumption of linear elastic material behavior. Difficulties determined by the frontal process zone and bridging zone and the measurements of crack length required the calculation of an effective crack length by normalization and comparison with an equivalent linear elastic material with no crack tip process zone. Two models were derived, the plastic energy model and the microcracking model which relies an effective crack length. Matsumoto and Nairn (2008) developed an original new energy based method for crack growth detection in MDF. For crack growth under continuous loading, detection image correlation method has been developed with simultaneous optical detection of crack length. In Fig. 4.23 are shown the specimens for four orthogonal crack directions in a MDF panel. The increment of crack growth Δa was measured, between two successive images from the shift in the strain profile (Fig. 4.24). In the case of MDF the unloading curves after crack propagation do not return to the origin probably because of residual stresses, plasticity or crack-plane interference. “Crack-plane interference means the bridging material left in the wake of the crack cannot be unloaded back to the original specimen configuration. Instead,
Fig. 4.24 Axial strain as a function of the position along the crack line with DIC and measurement of increment crack growth a (Matsumoto and Nairn 2008, Figure 2) Curve 1 – prior to crack growth, Curves 2 . . . 7 profiles after subsequent increments in crack growth, a = the crack growth between two point in the test, the shift between curves
4
Initiation and Growth of Delamination
89
Fig. 4.25 R-curve revised method (Matsumoto and Nairn 2008, Figure 4) (a) – integral transformation of force and crack length data as a function of displacement (b) – energy area to crack length (c)– R-curve as found from the slope of energy area
the bridging material is crushed, causing the unloading compliance to be lower, resulting in a residual displacement”. To take into consideration the bridging zone, material point method and CRAMP – cracks in the material point method – (Nairn 2003, 2006; Guo and Nairn 2004, 2006) have been developed as summarized in Fig. 4.25. R-curve with fiber bridging process zone was calculated from the slope of energy area. The slop of the cumulative energy released per unit thickness is deduced by integrating force displacement data up to some displacement and subtracting the area under an assumed elastic return to the origin. Figure 4.26 shows the
Fig. 4.26 R-curve (J/m2 ) versus a (mm), the crack growth between two point in the test for LT fracture with discrete and revised analysis, for MDF specimen 12 mm and 609 kg/m3 (Matsumoto and Nairn 2008, Figure 5)
90
V. Bucur
Table 4.8 Average values for the initiation toughness (Gc ), the slope of rising R – curve for MDF specimens and σ c for MDF (Matsumoto and Nairn 2008) Crack
Panel thickness 12 mm
Panel density
Panel thickness 19 mm t
Type
Gc
Slope
σc
Gc
Slope
σc
(kg/m3 ) 609
– //
737
//
J/m2 2062 54 4153 75.3
J/m3 21700 222 59600 814
MPa 0.78 0.038 2.55 0.14
J/m2 2233 48.2 4452 48.4
J/m3 10500 296 18400 303
MPa 0.43 0.056 0.66 0.10
good agreement between discrete and revised analyses of the R-curve for LT fracture for MDF specimens, 19 mm thickness and 609 kg/m3 density. Table 4.8 gives the initiation toughness (Gc ) and the slope of rising R – curve for MDF specimens in which the effect of panel density and thickness and of the crack orientation has been demonstrated. The originality of the model “material point model” proposed by Nairn and co-workers compared with previous approaches, started with machined notch, for which the subsequent process zone does not influence the initial crack growth (Niemz et al. 1997, 1999, Morris et al 1999, Ehart et al. 1996) is related to the following points: – – – –
the introduction of an explicit crack, the crack tip energy release rate was calculated including bridging effect, the total fracture energy released was calculated at the time of crack propagation, the COD – crack opening displacement – along the crack surface was calculated, the bridging fibers failed whenever COD > δc , the calculation can be performed until crack length reached the end of the specimen, – fracture mechanics methods were used to model crack tip processes and traction law methods were used to model the bridging zone, inserted only as the crack propagated. Another experimental approach for crack propagation detection was proposed by Watanabe and Landis (2007) with 3D micro tomography. Cutting and drilling of MDF are often required for boards used in the manufacture of furniture, cabinets and flooring. Delamination is one of the major defects observed with cutting and drilling. Digital image analysis was used by Davim et al. (2007) to study the delamination in an MDF plate with coating layers induced by drilling. An empirical factor has been proposed to characterize the delamination at the entrance and exit surfaces of the hole produced by drilling. The digital image analysis shows a typically brittle fracture on the coating layer. The damage area at the drill entrance is slightly larger than at the drill exit. Higher cutting speed should be used to induce minimal delamination and to obtain greater material removal during drilling.
4
Initiation and Growth of Delamination
91
4.5 Fracture Mechanics Parameters for Ecological Relevance Fracture mechanics criteria have been developed by Mattheck (1996, 2004), Mattheck and Bethge (1991), Matthek et al. (1995), Mattheck and Kubler (1997) to describe failure modes of trees. Strong winds (hurricanes, cyclones, etc) cause extensive damage to forests, world-vide every year. Considerable research has been carried out to understand the physical processes involved in order to improve the sylvicultural practice. Spectral analysis was used to identify the dynamic behaviour of trees (Guitard and Castera 1995; Gardiner 1995). “Wind affects the growth rate of the tree and determines the occurrence of windthrow in the later years of the development of a forest. Consequently the forecast of the financial viability of any forest project is dependent upon an accurate assessment of the wind speed” and of the modelling of trees behaviour (Gardiner 1995). Fracture mechanics parameters have been used to put in evidence potential influence of air pollution on wood quality in Europe and Canada (Grosser et al. 1985; Bondietti et al. 1990; Niemz et al. 1990; Koch et al. 1996; Stanzl-Tschegg, Filion et al. 1999; Beismann et al. 2000; Beismann et al. 2002). Unfortunately no unified methodology was used and the results have been ambiguous. Only tendencies have been observed. In spruce, exposed to SO2 emission, the fracture morphology (Koch et al. (1996) has shown short fibres, and crack initiation with bent tracheids in the vicinity of rays. Stanzl-Tschegg, Filion et al. (1999) described the SO2 pollution in spruce with the notch-tensile strength via ring width and density. A pronounced influence was observed on trees grown between 1970 and 1985, and a subsequent recovery in trees that had survived this period. Beismann et al. (2002) noted the response of stems of 6 to 7 year old spruce and beech trees studied after 4 years growth in elevated atmospheric CO2 in combination with a nitrogen treatment and on two different soil types. The fracture toughness, modulus of elasticity (EL ) and wood density were strongly influenced. Smith and Chui (1994) observed differences in Mode I fracture energy of premature plantation grown red pine for crack growth in the L direction. Differences in bending properties of plantation grown white spruce have been reported by Zho and Smith (1991). Differences in fracture energy of Pinus radiata wood from different plantations were reported by King and Vincent (1998). Donaldson (1995) put in evidence cell wall fracture properties in relation to lignin distribution and cell dimensions among three genetic groups of radiate pine. As a conclusion it can be suggested that the environmental influences on wood quality require the development of specific techniques for wood microstructural studies.
4.6 Summary Reliable prediction of delamination growth is still proving to be problematic, leading to the use of large safety factors and reticence in using wood-based composites
92
V. Bucur
in safety critical applications. This has led to composite structures being perceived as expensive to fabricate and needing frequent inspection and repair. The recognise approach to study delamination has been the fracture mechanics. The theory of fracture mechanics has been successfully applied to wood, wood products and wood-based composites since more than 50 years and provided valuable concepts for evaluation of the influence of cracks, notches or other stress raisers in structural elements. The space – time multi-scale nature of the delamination process in wood can be related to the prediction of crack nucleation, growth and arrest. Crack tip displacement is related to crack growth and propagation. The definition of the damage zone ahead of a crack tip is crucial for the studies of wood fracture. If the fracture process zone is small compared to the length of the crack, linear elastic fracture mechanics (LEFM) methods yield an accurate prediction of the load level at which a crack in a structural component will grow. Any deformation of the crack can be described through a combination of three fracture pure modes: Mode I – opening mode in tension, Mode II – the in plane shear mode and, Mode III – the out of plane shear mode. However, mixed fracture modes can be recognised also. The anisotropic nature of wood allows the development of six different fracture system orientations. For the situations where the fracture process zone is not small compared with the length of a crack, the energy methods and the concepts of nonlinear fracture mechanics (NLEFM) can be used. This approach can be used to accurate prediction of wood fracture behaviour through laboratory tests and in reliable interpretation of the mechanical capacities of notched small dimension timbers, or structures with mechanical connections made with fastenings (nails, bolts, shear plates, split rings), etc. A range of failure criteria have been developed based on the physics of delamination fracture in wood and in wood-based composites. These criteria included parameters that relate to the influence of loading, material characteristics and environmental factors. Experimental investigation and predictive (analytical and numerical) modeling are linked through microfractographic studies. The fundamental knowledge on fracture behaviour of wood can have relevance for structural use of timber, in pulping industry, for wood drying technology, or in processes of machining and cutting.
References Aicher S (1992) Fracture energies and size effect law for spruce and oak in mode I. Determination of fracture properties of wood especially in mode II and mixed modes I and II. Proceedings of the RILEM TC 133 Meeting, Bordeaux Aicher S, Reinhardt HW, Klöck W (1993) Nichtlineares bruch-mechanik-massstabsgesetz für Fichte bei Zugbeanspruchungsenkrecht zur Fasserrichtung. Holz als Roh Werkst 5:385–394 Aicher S, Dill Langer G, Ranta – Maunus A (1998) Duration of load effect in tension perpendicular to the grain of glulam in different climates. Holz Roh-Werkst 56:279–284 Akande JA, Kyanka GH (1990) Evaluation of tensile fracture in aspen using fractographic and theoretical methods. Wood Fiber Sci 22:283–297
4
Initiation and Growth of Delamination
93
Anderson TL (2005) Fracture mechanics: fundamentals and applications. 3 rd edition, CRC, Boca Raton Ando K, Ohta M (1999) Variability of fracture toughness by the crack tip position in an annual ring of coniferous wood. J Wood Sci 45:275–283 Ashby MF, Eastering KE, Harryson R, Maiti SK (1985) The fracture and toughness of wood. Proc Roy Soc London A 398(1815):261–280 ASTM E 399 (1994) Standard test method for plain strain fracture toughness of metallic materials Annual Book of Standards, American Society for Testing Materials, Philadelphia, PA Atluri SN (1986) Computational methods in the mechanics of fracture. North Holland, Amstrerdam Attack D, May WD, Morris EL, Sproule RN (1961) The energy of tensile and cleavage fracture in black spruce. Tappi J 44(8):555–567 Bariska M (1994) Fracture mechanics and wood anatomy. IAWA J 15:199–200 Barrett JD (1976) Effect of crack – front width on fracture toughness of Douglas fir. Eng Fract Mech 8:711–717 Barrett JD (1981) Fracture mechanics and the design of wood structures. Phil Trans Roy Soc London 299(8):217–226 Beismann H, Schweingruber F, Speck T, Körner C (2002) Mechanical properties of spruce and beech wood grown in elevated CO2 . Trees 16:511–518 Beismann H, Wilhelmi H, Baillères H, Spatz HC, Bogenrieder A, Speck T (2000) Brittless of twig bases in the genus Salix: fracture mechanics and ecological relevance. J Exp Bot. 51: 617–633 Berto F, Lazzarin P (2007) Relationship between J – integral and the strain energy evaluated in a finite volume surrounding the tip of sharp and blunt V-notches. Int J Solid Struct 44(14/15):4621–4645 Boatright SWJ, Garrett G (1983) The effect of microstructure and stress state on the fracture behavior of wood. J Mat Sci 18:2181–2199 Bodner J, Grüll G, Schlag MG (1996) In-situ fracturing of wood in the scanning electron microscope. Holzforschung 50:487–490 Bodner J, Schlag MG, Grüll G (1997) Fracture initiation and progress in wood specimens stressed in tension. Part I: Clear wood specimens stressed parallel to the grain. Holzforschung 51:479–484; Part II: Compression wood specimens stressed parallel to grain Holzforschung 51: 571–576 Bondietti EA, Momoshima N, Shortle WC, Smith KT (1990) A historical perspective on divalent trends in red spruce stemwood and the hypothetical relationship to acid deposition. Can J For Res 20:1850–1858 Borgin K (1971) The cohesive failure of wood studied with the scanning electron microscopy. J Microsc 94(1):1–11 Boström L (1992) Method for determining of the softening behaviour of wood and the applicability of a nonlinear fracture mechanics model. PhD Thesis LUTVDG/TVBM 1012 Lund Institute of Technology, Lund, Sweden Boström L (1988) The fictitious crack model – A fracture mechanics approach applied on wood. International Conference on Timber Engineering, Seattle DC vol 2:559–565 Boyd JD (1973) Helical fissures in compression wood cells: causative factors and mechanics of development. Wood Sci Technol 7:92–111 Bucur V (2003) Imaging of microscopic structure of wood. Proceedings of the 5th world congress on ultrasonics Paris, pp 299–302. http://www.sfa.asso.fr/wcu 2003/procs/webside articles/. Accessed 20 December 2007 Bucur V (2005) Acousics of wood. Springer, Heidelberg Bucur V, Rasolofosaon PNJ (1998) Dynamic elastic anisotropy and nonlinearity in wood and rock. Ultrasonics 36:813–824 Bucur V, Saied A, Attal J (1995) Identification of wood anatomical elements by acoustic microscopy, 22nd International. Symposium on Acoustical Imaging, Florence, Italy
94
V. Bucur
Bucur V, Herbé C, Nosei G (1994) Annual ring characteristics of Pinus taeda measured by ultrasonic and X-ray techniques. IAWA J 15(2):121–132 Clair B, Gilles Despaux G, Chanson B, Thibaut B (2000) Utilisation de la microscopie acoustique pour l étude des propriétés locales du bois : étude préliminaire de paramètres expérimentaux. Ann For Sci 57:335–343 Coureau JL, Gustafsson Persson K (2006) Elastic layer model for application to crack propagation problems in timber engineering. Wood Sci Technol 40:275–290 Coureau JL, Gustafsson PJ, Persson K (2006a) Elastic layer model for application to crack propagation problems in timber engineering. Wood Sci Technol 40:275–290 Coureau JL, Morel S, Gustafsson PJ, Lespine C (2006b) Influence of the fracture softening behaviour of wood on load-cod curve and R-curve. Mater Struct 40:97–106 Conrad MPC, Smith GD, Fernlund G (2003) Fracture of solid wood: a review of structure and properties at different length scale. Wood Fiber Sci 35:570–584 Cotterell B (2002) The past, present and future of fracture mechanics. Eng Fract Mech 69:533–553 Davids WG, Landis EN, Vasic S (2003) Lattice models for the prediction of load – induced failure and damage in wood. Wood Fiber Sci 35:120–134 Davims JP, Rubio JC, Abrao AM (2007) Delamination assessment after drilling medium-density fibreboard (MDF) by digital image analysis. Holzforschung 61:294–300 Daudeville L (1999) Fracture in spruce: experiment and numerical analysis by linear and non linear fracture mechanics. Holz Roh Werkst 57:425–432 DeBaise GR (1972) Morphology of wood shear fracture. J Materials 7(4):568 –572 DeBaise GR, Porter AW, Pentoney RE (1966) Morphology and mechanics of wood fracture. Mater Res Stand 6:493–499 Dill – Langer G, Aicher S (2000) Monitoring of microfracture by microscopy and acoustic emission. In Aicher S (ed) Wood and wood fiber composites. University of Stuttgart, Otto Graff – Institute, Department of Wood and Timber Engineering, Stuttgart, pp 93 –104 Dill – Langer G, Lütze S, Aicher S (2002) Microfracture in wood monitored by confocal laser microscope. Wood Sci Technol 36:487–499 Dinwoodie JM (1966) Induction of cell wall dislocations (slip planes) during preparation of microscope sections of wood. Nature 212:525–527 Dinwoodie JM (1968) Failure in timber. Part I. Microscopic changes in cell wall structure associated with compression failure. J Inst Wood Sci 21:37–53 Dinwoodie JM (1974) Failure of timber. Part II. The angle of shear through the cell wall during longitudinal compression stressing. Wood Sci Technol 8:5–67 Donaldson LA (1995) Cell wall fracture properties in relation to lignin distribution and cell dimensions among three genetic groups of radiate pine. Wood Sci Technol 29:51–63 Donaldson LA (1996) Clonal variation in the fracture properties of radiate pine wood. In Donaldson LA, Singh AP, Butterfield BG, Whitehouse L (eds) Recent advances in wood anatomy. New Zealand Forest Research Institute Ltd, Rotoruapp 283–291 Donaldson LA (1997) Ultrastructure of transwall fracture surfaces in Radiata pine wood using transmission electron microscopy. Holzforschung 51:303–308 Donaldson LA, Singh AP, Butterfield BG, Whitehouse L (1996) Recent advances in wood anatomy. New Zealand Forest Research Institute Ltd, Rotorua Dourado N, de Moura MFSF, Morel S, Valentin G, Morais J (2004) A finite element simulation of the SENB test with crack propagation. Proceedings of the 3rd international Conferenceon European Society for Woods Mechanics, Vila Real, Portugalpp 67–76 Dourado N, Morel S, de Moura MFSF, Valentin G, Morais J (2008) Comparison of fracture properties of two wood species through cohesive crack simulations. Compos Part A: 39: 415–427 Ehart RJA, Stanzl-Tschegg SE, Tschegg EK (1996) Characterization of crack propagation in particleboard. Wood Sci Technol 30:307–321 R – Ehart RJA, Stanzl-Tschegg SE, Tschegg EK (1998) Fracture characterization of PARALAM PSL in comparison to solid wood and particleboard. Wood Sci Technol 32:43–55
4
Initiation and Growth of Delamination
95
Ehart RJA, Stanzl-Tschegg SE, Tschegg EK (1999) Mode III fracture energy of wood composites in comparison to solid wood. Wood Sci Technol 33:391–405 Forest Products Laboratory (1999) Wood handbook – wood as an engineering material. FPL GTR 113, USDA, Forest Service, Madison, WI Frühmann K, Tschegg EK, Dai C, Stanzl-Tschegg SE (2002) Fracture behaviour of laminated veneer lumber under Mode I and III loading. Wood Sci Technol 36:319–334 Frühmann K, Burgert I, Stanzl-Tschegg SE, Tschegg EK (2003) Mode I fracture behaviour on the growth ring scale and cellular level of spruce (Picea abies L. Karst.) and beech (Fagus sylvatica L.) loaded in the TR crack propagation system. Holzforschung 57:653–660 Gardiner B (1995) The interaction of wind and tree movement in forest canopies. In Coutts MP Grace J (eds) Wind and trees, Cambridge University Press, Cambridge, pp 41–59 Gibson LJ, Ashby MF (1997) Cellular solids: structure and properties. Cambridge University Press, Cambridge Gordon JE, Jeronimidis G (1980) Composites with high work of fracture. Philos Trans R Soc A 294:545–550 Griffith AA (1920) The phenomenon of rupture and flow in solids. Phil Trans R Soc A 221:163–169 Grosser D, Schulz H, Utschig H (1985) Mögliche anatomische Veränderubgen in erkrankten Nadelbäumen. Holz Roh-Werkst 43:315–323 Guitard D, Castera P (1995) Experimental and mechanical modelling of wind induced trees sways. In Coutts MP, Grace J (eds) Wind and trees, Cambridge University Press, Cambridge, pp 195–203 Guo YJ, Nairn JA (2006) Three-dimensional dynamic fracture analysis using the material point method. CMES 1(1):11–25 Guo YJ, Nairn JA (2004) Calculation of J-integral and stress intensity factors using Material Point Method. Comput Model Eng Sci 6:295–308 Gustafsson PJ (1985) Fracture mechanics studies of non-yielding materials like concrete. Technical Report TVBM – 1007 Lund Institute Technology, Sweden Gustafsson PJ (1988) A study of strength of notched beams. Proceedings of CIB W18 Meeting 21, Vancouver, September 1988 Hillerborg A (1991) Application of the fictitious crack model to different types of materials. Int J.Fracture 51:95–102 Hillerborg A, Modeer M, Petersson PE (1976) Analysis of crack formation and crack growth in concrete by means of fracture mechanics and finite elements. CemConcr Res 6:773–782 Hoffmeyer P, Hanna RB (1989) Electron beam damage during testing of wood in the SEM Wood Sci Technol 23:211–214 Hofstetter K, Hellmich C, Eberhardsteiner J (2007) Micromechanical modelling of solid-type and plate-type deformation patterns within softwood materials. A review and an improved approach. Holzforschung 61:343–351 Homberg S, Persson K, Petersson H (1999) Nonlinear mechanical behaviour and analysis of wood and fiber materials. Comp Struct 72:459–480 Irwin GR (1957) Analysis of stresses and strain near the end of a crack traversing a plate. J Appl Mech 24:361–364 Jensen JL (2005a) Quasi-non-linear fracture mechanics analysis of the double cantilever beam specimen. J Wood Sci 51:566–571 Jensen JL (2005b) Quasi-non-linear fracture mechanics analysis of the splitting failure of single dowel joints loaded perpendicularly to the grain J Wood Sci 51:559–565 Jensen JL (2005c) Quasi-non-linear fracture mechanics analysis of splitting failure in simply supported beams loaded perpendicular to grain by dowel joints. J Wood Sci 51: 577–582 Jeronimidis G (1980a) Wood, one of nature’s challenging composite. In Vincent JFV, Currey JD (eds) The mechanical properties of biological materials Cambridge University Press, Cambridge, pp 169–182 Jeronimidis G (1980b) The fracture behaviour of wood and the relations between toughness and morphology. Proc R Soc B-Biol Sci 208:447–460
96
V. Bucur
Jeronimidis G (1976) The fracture of wood in relation to its structure. Leiden botanical Series No 3: 253–265 Jernkvist LO (2001) Fracture of wood under mixed mode loading. Eng. Fract. Mech. Part I Derivation of fracture criteria 68(5):549–563; Part II Experimental investigation of Picea abies. 68(5):565–576 Johnson JA (1973) Crack initiation in wood plates. Wood Sci 6:151–158 Jung J, Murpjy JF (1983) An investigation of the fracture of butt joints in parallel – laminated veneer. Wood Fiber Sci 15:116–134 Keith CT, Côté WA Jr. (1968) Microscopic characterization of slip lines and compression failures in wood cell walls. For Prod J 18(3):67–74 Keunecke D, Stanzl-Tschegg S, Niemz P (2007) Fracture characterisation of yew (Taxus baccata L) and spruce (Picea abies L.Karst) in the radial-tangential and tangential radial crack propagation system by a micro wedge splitting test. Holzforschung 61:582–588 King MJ, Sutherland IJ, Le-Ngoc L (1999) Fracture toughness of wet and dry Pinus radiata. Holz Roh- Werkst 57:235–240 King MJ, Vincent JFV (1998) Fracture energy during clearing of Pinus radiata. Holz Roh Werkst 56:259–265 Koch G, Bauch J, Dünisch O, Seehann G, Schmitt U (1996) Sekundäre Veränderungen im Holz akut belasteten Ficten in Hochlage des Osterzgebirges. Holz Roh Werkst 54:243–249 Kucera LJ, Bariska M (1982) On the fracture morphology in wood. Part I. A SEM – study of deformation in wood of spruce and aspen upon ultimate axial compression load. Wood Sci Technol 16:241–259 Landis EN, Navi P (2009) Modelling crack propagation in wood and wood composites. A review. Holzforschung 63:150–156 Landis EN, Vasic S, Davids WG, Parrod (2002) Coupled experiments and simulations of structural damage in wood. Exp Mech 42(4):389–394 Lavisci P, Berti S, Pizzo, B, Triboulot P, Zanuttini R (2001) A delamination test for structural wood adhesives used in thick joints. Holz Roh Werkst 59:153–154 Lee SH, Quales SL, Schniewind AP (1995) Wood fracture, acoustic emission, and the drying process. Part 2. Acoustic emission pattern recognition analysis. Wood Sci Technol 30:283–292 Leicester RH (1971) Some aspects of stress fields at sharp notches in orthotropic materials. Division of Forest Products, Technical Paper No 57, CSIRO Australia Leicester RH (1973) Effect of size on the strength of structures. Division of Buildingg Research Tech Paper No 71 CSIRO Australia Leicester RH (1974) Fracture strength of wood. Paper at 1st Australian Conference Eng. Mater. University of New South Wales, Sydney Lespine C (2007) Courbe R et propriétés de fissures cohésive dans la rupture quasi-fragile. Proceedings of the « XXX Rencontres Universitaires de Génie Civil, Prix René Houpert » Loidl D, Tschegg EK, Stanzl-Tschegg (2008) Fracturing of wood under superimposed tension and torsion. Wood Sci Technol 42:61–74 Mark RE (1967) Cell wall mechanics of wood tracheids. Yale University Press, New Haven, CT Matsumoto N, Nairn JA (2008) The fracture toughness of medium density fibreboard including the effects of fiber bridging and crack-plane interference. Eng. Fracture Mech 76:2748–2797 Mattheck C (2004) The face of failure in nature and engineering. Verlag, Forschungszentrum Karlshrue Mattheck C (1996) Trees – the mechanical design. Springer, Heidelberg Mattheck C, Bethge K (1991) Failure of trees induced by delamination. Arboricultural J 15: 243–253 Mattheck C, Kubler H (1997) Wood – the internal optimization of trees. Springer, Heidelberg Mattheck C, Bethge K, Albrecht W (1995) Failure modes of trees and related failure criteria. In Coutts MP, Grace J (eds) Wind and trees, Cambridge University Press, Cambridge, pp 195–203 Mishnaevsky L, Qing H (2008) Micromechanical modelling of mechanical behaviour and strength of wood: state of the art review. Comput Mat Sci 44:363–370
4
Initiation and Growth of Delamination
97
Morel S, Mourot G, Schmittbubl J (2003) Influence of the specimen geometry on R – curve behaviour and roughening of fracture surfaces. Int J Fract 121(1/2):23–42 Morel S, Bouchard E, Schmittbuhl J, Valentin G (2002) R-curve behaviour and roughness development of fracture surfaces. Int J Fract 114:307–325 Morel S, Dourado N, Valentin G, Morais J (2005) Wood a quasi-brittle material R-curve behavior and peak load evaluation. Int J Fracture 131(4):385–400 Morris VL, Adam JM, Hunt DG, Bonfield PW (1996) Fracture mechanics in wood and woodbased panel products at different testing speeds and relative humidity. COST Action – wood mechanics. Workshop on mechanical properties of panel products, 22 – 23 March Watford, pp 336–351 Morris VL, Hunt DG, Adam JM (1999) The effect of experimental parameters on fracture energy of wood-based panels. J Inst Wood Sci 15(1): 85 Moses DM, Prion HGL (2004) Stress and failure analysis of wood composites: a new model. Compos: Part B, 35:251–261 Nairn JA (2009) Analytical and numerical modelling of R curves for cracks with bridging zones Int J Fract 155(2):167–181 Nairn JA, Matsumoto N (2009) Fracture modelling of crack propagation in wood and wood composites including crack tip process and fiber bridging mechanics. Proceedings of the 12th International Conference on Fracture, Ottawa, July 12–17 Nairn JA (2007a) Material point simulation of transverse fracture in wood with realistic morphologies. Holzforschung 61:375–381 Nairn JA (2007b) A numerical study of the transverse modulus of wood as a function of grain orientation and properties. Holzforschung 61:406–413 Nairn JA (2007c) Numerical modelling of deformation and fracture of wood including heterogeneity and anisotropy. Cost Action 35, Lausanne, May 21–23. Nairn JA (2006) Numerical simulation of the transverse compression and densification in wood. Wood Fiber Sci 38:576–591 Nairn JA (2003) Material Point Method calculations with explicit cracks. Comput Model Eng Sci 4:649–664 Niemz P, Diener M (1999) Vergleichende Untersuchungen zur Ermittlung der Bruchzähigkeit an Holzwerkstoffen. Holz Roh Werkst 57:222–224 Niemz P, Diener M, Pöhler E (1997) Untersuchungen zur Ermittlung der Bruchzähigkeit an MDF – Platten. Holz Roh Werkst 55:327–330 Niemz P, Bemmann A, Wagebführ R (1990) Untersuchungen zu den Eigenschaften von immissionsgeschädigtem Fichtenholz. Holztechnologie 30(2):70–71 O’Brien TK (1998) Interlaminar fracture toughness: the long and winding road to standardisation. Compos B 29(1):57–62 Patton – Mallory M, Cramer SM (1987) Fracture mechanics: a tool for predicting wood component strength. For Prod J 37(7/8):39–47 Pearson RG (1974) Application of Fracture mechanics to the study of tensile strength of structural lumber. Holzforschung 29(1):11–19 Petterson RW, Bodig J (1983) Prediction of fracture toughness of conifers. Wood Fiber Sci 15: 302–316 Porter AW (1964) On the mechanics of fracture in wood. For Prod J 14(8):325–331 Raju IS, O’Brien TK (2008) fracture mechanics concepts, stress fields, strain energy release rates, delamination initiation and growth criteria. In Sridharan S (ed) Delamination behaviour of composites, Woodhead Publishing, Cambridge, England Renaud M, Rueff M, Rocaboy AC (1996) Mechanical behaviour of saturated wood under compression. Part 2: Behaviour of wood at low rates of strain some effects of compression on wood structure. Wood Sci Technol 30:237–243 Reiterer A, Stanzl- Tschegg SE, Tschegg EK (2000) Mode I fracture and acoustic emission of softwood and hardwood. Wood SciTechnol 34:417–430
98
V. Bucur
Reiterer A, Sinn G (2002) Fracture behaviour of modified spruce wood: a study using linear and non linear fracture mechanics. Holzforschung 56:191–198 Reiterer A, Sinn G, Stanzl-Tschegg SE (2002) Fracture characteristics of different wood species under mode I loading perpendicular to the grain. Mater Sci Eng A 332::29–36 SAA (1975) AS 1720 – 1979 Timber engineering code. Sydney Standards Association of Australia Sandford RJ (2003) Principles of fracture mechanics. Pearson Education Inc., Upper Saddle River, NJ Sedighi – Gilani M, Navi P (2007) Micromechanical approach to wood fracture by three-dimensional mixed lattice-continnum model at fiber level. Wood Sci Technol 41: 619–634 Scheffler M, Niemz P, Diener M, Lostig V, Hardtke HJ (2004) Holz Roh Werkst 62(2): 93–100 Schmidt J, Kallske M (2007) Simulation of cracks in wood using a coupled material model for interface elements. Holzforschung 61:382–389 Schniewind AP (1977) Fracture toughness and duration of load factor. II. Duration factor for cracks propagating perpendicular to grain. Wood Fiber Sci 9:216–226 Schniewind AP, Centeno JC (1973) Fracture toughness and duration of load factor. I. Six principal systems of crack propagation and the duration factor for cracks propagating parallel to grain. Wood Fiber 5(2):152–158 Schniewind AP, Lyon DE (1971) Tensile strength of redwood dimension lumber. I. Relation to grade and working stress. Forest Prod J 21(7):45–49 Schniewind AP, Lyon DE (1973) A fracture mechanics approach to the tensile strength perpendicular to grain of dimension lumber. Wood Sci Technol 7:45–59 Schniewind AP, Ohgama T, Aoki T, Yamada T (1982) Effect of specific gravity, moisture content, and temperature on fracture toughness of wood. Wood Sci 15(2):101–109 Schniewind AP, Pozniak A (1971) On fracture toughness of Douglas – fir wood. Eng Fract Mech 2:223–233 Sedighi-Gilani M, Navi P (2007) Micromechanical approach to wood fracture by threedimensional mixed lattice continuum model at fiber level. Wood Sci Technol 41: 619–634 Seel J, Ziemmermann T (1998) The fine structure of the cell wall of hardwoods on transverse fracture surfaces. Holz Roh Werkst 56:365–366 Silva MAL, de Moura MFSF, Morais JJL (2006) Numerical analysis of the ENF test for mode II wood fracture. Compos Part A:1334–1344 Sippola M, Frühmann K (2002) In situ longitudinal tensile tests of pine wood in an environmental scanning electron microscope. Holzforschung 56:669–675 Smith I, Chui YH (1994) Factors affecting mode I fracture energy of plantation grown red pine. Wood Sci Technol 28: 147–157 Smith I, Landis E, Gong M (2003) Fracture and fatigue in wood. Wiley, New York, NY Smith I, Snow M, Asiz A, Vasic S (2007) Failure mechanism in wood based materials: a review of discrete, continuum and hybride finite-element representation. Holzforschung 61:352–359 Smith I, Vasic S (2003) Fracture behaviour of softwood. Mech Mater 35:803–815 Smith J, Kallske M (2007) Simulation of cracks in wood using a coupled material model for interface elements. Holzforschung 61:383–389 Sridharan S (2008) Delamination behaviour of composites Woodhead Publishing, Cambridge, England. Stanzl-Tschegg SE, Navi P (2009) Fracture behaviour of wood and its composites. A review. Holzforschung 63:139–149 Stanzl-Tschegg SE, Filion L, Tschegg EK, Reiter A (1999) Strength properties and density of SO2 polluted spruce wood. Holz Roh Werkst 57:121–128 Stanzl-Tschegg SE, Tschegg EK, Teischinger A (1994) Fracture energy of spruce wood after different drying procedures. Wood Fiber Sci 26:467–478 Stanzl-Tschegg SE, Tan DM, Tschegg EK (1995) New splitting method for wood fracture characterization. Wood Sci Technol 29: 31–50
4
Initiation and Growth of Delamination
99
Stanzl-Tschegg SE, Ehart RJA, Tschegg EK (1997) Fracture behaviour of glued wood laminate compounds. Proceedings of the 9th international conference on fracture, Sydney, Australia Swinehart DE, Broek D (1995) Tenacity©, fracture mechanics, and unknown coater web breaks. Tappi Journal 79(2):233–237 Tada H, Paris PC, Irwin GR (2000) The stress analysis of cracks handbook. 3rd edn. ASME, New York, NY Tan DM, Stazl-Tschegg S, Tschegg EK (1995) Models of wood fracture in Mode I and Mode II. Holz als Roh- und Werkst. 53:159–164 Tschegg EK (1986) Equipment and appropriate specimen shapes for tests to measure fracture values. (in German). Patent no 390328 Österreichisches Patentamt Tschegg EK, Frühmann K, Stanzl-Tschegg SE (2001) Damage and fracture mechanisms during mode I and III loading of wood. Holzforschung 55:525–533 Thuvander F, Berglund LA (2000) In situ observations of fracture mechanisms for radial crack in wood. J Mater Sci 35:6277–6283 Thuvander F, Berglund LA (1998) A multiple fracture test for strain to failure distribution in wood. Wood Sci Technol 32:227–235 Thuvander F, Wallström L, Berglund LA, Lindberg KAH (2001) Effects of an impregnation procedure for prevention of wood cell wall damage due to drying. Wood Sci Technol 34:473–480 Thuvander F, Berglund LA (2000) In situ observations of fracture mechanism for radial cracks in wood. J Mater Sci 35(24):6277–6283 Triboulot P, Asano I, Ohta M (1983) An application of fracture mechanics to the wood cutting process. Mokuzai Gakk 29:111–117 Triboulot P, Jodin P, Pluvinage G (1984) Validity of fracture mechanics concepts applied to wood by finite element calculation. Wood Sci Technol 18:51–58 Turkulin H, Holzer L, Richter K, Ssell J (2005) Application of the ESEM in wood research. Part II. Comparison of operational modes. Wood Fiber Sci 37:565–573 Valentin G, Boström L, Gustafsson PJ, Ranta-Maunus A, Gowda S (1991) Application of fracture mechanics to timber structures. RILEM State of the art report. Res. Note 1262, Technical Research Centre of Finland, Espoo, Finland Valentin G, Morlier P (1982) A criterion of crack propagation in timber. Mater Struct 15:88–95 Vasic S, Ceccotti A, Smith I, Sandak J (2009) Deformation rates effects in softwoods. Crack dynamics with lattice fracture modelling. Eng Fract Mech 76(9):1231–1246 Vasic S, Stanzl-Tschegg S (2008) Softwood/hardwoods fracture at different humidity levels: ESEM in-situ real time experiments. Holzforschung 62 Vasic S, Stanzl-Tschegg S (2007) Experimental and numerical investigation of wood fracture mechanisms at different humidity levels. Holzforschung 61:367–374. Vasic S, Stanzl-Tschegg S (2005) Fracture mechanisms and properties of green wood subjected to opening Mode I. In Tschegg S, Sinn G(eds) Proceedings of the COST Action E35, Rosenheim Workshop September 29– 30. Vasic S, Smith I (2002) Bridging crack model for fracture of spruce. Eng Fract Mech 69:745–760 Vasic S, Smith I, Landis E (2002) Fracture zone characterization–micro-mechanical study. Wood Fiber Sci 34:42–56 Vasic S, Smith I (2003) Contact – crack problem with friction in spruce. Holz Roh–Werkst 61(3):182–186 Vasic S, Smith I (1996a) On the influence of ultrastructure and fibres bridging in Mode I fracture of wood. Proceedings of the 2nd international conference on the deevelopment of wood science /technology and Forestry ICWSF’96, Sopron, Hungary Vasic S, Smith I (1996b) The brittleness of wood in tension perpendicular to the grain: micromechanical aspects. Proceedings of the COST 508 Wood Mechanics Conference, Stuttgart, Germany, pp 555–569 Vasic S, Smith I (1998) Bridged crack model of wood fracture: analysis and numerical modelling. Proceedings of the world timber conference Montreux, 17–20 August, Swiss Federal Institute of Technology, Lausanne, Suisse, pp 1818–1819
100
V. Bucur
Vasic S, Smith I (1999a) The effect of bridging stress on fracture toughness of wood. RILEM Symposium Timber Engineering, Stockholm Vasic S, Smith I (1999b) Failure analysis of tensile strength perpendicular to the grain. RILEM Symposium Timber Engineering, Stockholm Walsh PF (1971) Cleavage Fracture of Timber. Div. For. Prod. Tech. Pap. No. 65, CSIRO, Melbourne Walsh PF (1972) Linear fracture mechanics in orthotropic materials. Eng Fract Mech 4:533–541 Walsh PF (1973) The interaction of butt joints. J Inst Wood Sci 6(2):22–27 Wang SS (1984) Edge delamination in angle –ply laminates. AIAAA J 22(2):256–264 Wang L, Lu Z, ZhaoG (2003) Wood fracture pattern during the water adsorption process. Holzforschung 57:639–643 Watanabe K, Landis EN (2007) An acoustic emission –based study of energy dissipation in radially loaded spruce. In Proceedings of the 3rd international symposium on wood machining. Lausanne, Switzerland, pp 179–182 Williams JG (1989) The fracture mechanics of delamination tests. J Strain Anal 24(4):207–214 Wittel FK, Dill-Langer G, Kröplin BH (2005) Modelling of damage evolution in softwood perpendicular to grain by means of a discrete element approach. Comput Mater Sci 32:594–603 Wu EM (1967) Application of fracture mechanics to anisotropic plate. J Appl Mech 34:967–974 Yoshihara H (2001) Influence of span/depth rate on the measurement of mode II fracture toughness of wood by end – notched flexure test. J Wood Sci 47(1):8–12 Yoshihara H (2003) Resistance curve for the mode II fracture toughness of wood obtained by the end – notched flexure test under the constant loading point displacement condition. J Wood Sci 49(3):210–215 Yoshihara H (2004) Mode II R-curve of wood measured by 4-ENF test. Eng Fract Mech 71: 2065–2077 Yoshihara H (2005) Mode II initiation fracture toughness analysis for wood obtained by 3 ENF test. Compos Sci Technol 65:2198–2207 Yoshihara H (2006a) Estimation of the 4–ENF test for measuring the mode III R -curve of wood. Eng Fract Mech 73(1):42–63 Yoshihara H (2006b) Characterization of fracturing properties of wood and wood based materials on fracture mechanics. Mokuzai Gakk 52:185–195 Zimmermann T, Sell J, Eckstein D (1994) SEM studies on tension – fracture surfaces of spruce samples. Holz als – Roh und Werkst 52:223–229 Zho H, Smith I (1991) Influence of drying treatment on bending properties of plantation – grown white spruce. For Prod. J. 41(3):8–14 Zink AG, Pellicane PJ, Shuler CE (1994) Ultrastructural analysis of softwood fracture surfaces. Wood Sci Technol 28:329–338 Zink AG, Pellicane PJ, Anthony RW (1995) A stress transformation approach to predicting the failure mode of wood. Wood Sci Technol 30:21–30
Chapter 5
A Theoretical Model of Collapse Recovery Philip Blakemore
Contents 5.1 5.2
Introduction . . . . . . . . . . . . . . . . . . . Repeating Cell Unit Model with Cyclical Constrains 5.2.1 Cell Wall Layer Properties . . . . . . . . . 5.2.2 Circular Based Cell Model . . . . . . . . . 5.2.3 Squared Based Cell Model . . . . . . . . . 5.3 Summary . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
. . . . . . .
101 102 102 104 106 118 119
5.1 Introduction The theory that is thought to best explain the recovery of collapse reconditioning supposes that the S1 and S3 layers are largely responsible for providing restoring the cells to the un-collapsed shape. This is because these two layers are particularly important in providing circumferential stiffness to each individual cell lumen. Hence, it is the potential energy stored in these layers that principally provides the force to restore the cell shape. In contrast, the S2 layer is considered to be the most important for providing the inelastic material properties required to hold the cell in the collapsed or deformed state. While moisture content is important for its effect on the cell wall material properties (i.e. stiffness, creep, mechano-sorptive creep), the uptake or movement of moisture within the cell walls is not thought to be critical for collapse recovery. In this sense, the recovery phenomenon can largely be attributed to a thermal effect (Blakemore and Langrish, 2008), and hence it is the relationships with temperature for the various material properties which are critical for this modelling work. The effect of heat then is to soften the S2 layer, which is holding the cell in the deformed shape, allowing the stored mechanical energy in the S1 and S3 layers to restore the cell shape. P. Blakemore (B) Department of Materials Science and Engineering, CSIRO, Clayton South, VIC 3169, Australia e-mail:
[email protected] V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_5,
101
102
P. Blakemore
The intention of this chapter is to outline the development of a numerical model to assess the importance of the different secondary cell wall layers for both collapse and recovery, to demonstrate if the proposed collapse recovery mechanism is plausible. The model presented was developed using a model that Innes (1995) originally created. As the name implies, collapse is essentially a structural or mechanical phenomenon. Currently, the most appropriate tool for numerically modelling this type of problem is to use Finite Element Analysis (FEA) or Modelling (FEM). A repeating cell unit model with cylindrical constraints is proposed, firstly with a circular based cell model and secondly with a square based cell model, and is discussed.
5.2 Repeating Cell Unit Model with Cyclical Constrains 5.2.1 Cell Wall Layer Properties The three dimensional ultrastructure of the secondary and primary layers in the cell wall, and the corresponding microfibril orientations, are shown in Fig. 5.1 The FEM developed tries to incorporate as many of the basic features of this structure as possible. One of the primary limiations of the Innes (1995) was that it was only for a single three layered cell in isolation. For the model developed here it was important to incorporate a double cell wall (S3 , S2 , S1 , CML, S1 , S2 , S3 ). As introduced above, the important characteristics for collapse recovery are the effect of temperature on the stiffness properties of the S1 and S3 layers, and the effect of heat on the plastic properties of the S2 layer. The orthotropic orientation of
Fig. 5.1 Three-dimensional representation of microfibril orientation in the primary cell wall (P) and the secondary cell wall (S) of a typical fibre or tracheid (Wardrop and Bland, 1959)
5
A Theoretical Model of Collapse Recovery
103
these properties within the layers is also crucial. For these reasons, an orthotropic viscoelastic material model is used for the S1 and S3 layers, while an isotropic elastic–plastic material model is used for the S2 layer and the CML. Given that in the S2 layer the radial and tangential properties are thought to be very similar, this simplification is thought to be acceptable. In that follows we discuss the viscoelastic material model (S1 , and S3 ) and the non-linear (elastic plastic) material properties (S2 , and CML). 5.2.1.1 Viscoelastic Material Model (S1 , and S3 ) The FEM package used for this modelling was MSC.Marc, which is a good solver for non-linear problems. The viscoelastic material model it uses is based on a de Prony-series that relates E, K or G against time at a certain temperature. For example, the de Prony-Series for E is shown in Eq. (5.2) The basic effect of temperature on E∞ was based on Eq. (5.1) (Innes 1996). A shift function is then used to adjust the creep curves for other temperatures. One of the more common forms of the shift function for polymers is the Williams-Landel-Ferry (WLF) equation (Williams et al. 1955) (Eq. (5.3)). In this instance, the shift factor and estimates of C1 and C2 were fitted such that the creep curves produced in a simple uniaxial tension model matched the creep model developed by Oliver (1991). Full details of the Oliver (1991) model, and how the appropriate MSC. Marc parameters were fitted to provide a match, are outlined in Chapter 6 of Blakemore (2008). Egreen = exp(4.206 + 0.003265BD − 0.03029T) E(t) = E∞ +
N
En exp
n=1
−t λn
(5.1)
(5.2)
where En = Modulus constant for series n λn = Relaxation time constant for series n – h t = Time – h log [αT ] =
−C1 (T − Ts ) C2 + (T − Ts )
(5.3)
where αT = Shift Factor C1 = Constant (Specific to Ts ) C2 = Constant (Specific to Ts ) T = Measurement temperature – K Ts = Reference temperature – K αT =
λnT λnTs
(5.4)
104
P. Blakemore
where λnT = Relaxation time constant for series n at the temperature T λnTs = Relaxation time constant for series n at the reference temperature Ts 5.2.1.2 The Non-linear (Elastic plastic) Material Properties (S2 , and CML) Innes (1995, 1996) used a non-linear elastic stress-strain relationship in his three layer, orthotropic, single-cell collapse model. The relationship was essentially broken into two parts: an initial linear relationship, and then a non-linear relationship. While MSC.Nastran can handle this form of non-linear relationship reasonably easily, MSC.Marc had no such capacity. However, given the large strains involved, in reality some of that strain is likely to be plastic strain and so using an elasticplastic relationship, which MSC.Marc can employ readily, seemed appropriate. This is despite the fact that this material type is mainly used to model smaller strain behaviour in metals. The MSC.Marc elastic-plastic relationship is by default isotropic. So unfortunately no orthotropic behaviour could be modelled. The match between the MSC.Marc elastic-plastic relationship and Innes (1996) used a nonlinear elastic stress-strain relationship can be found in Chapter 6 of Blakemore (2008).
5.2.2 Circular Based Cell Model For any cell scale model to be representative of macroscopic behaviour, it needs to be based on a geometrically representative unit with cyclically repeating boundary conditions. This ensures that many such models could be joined together in both directions and the behaviour would still be consistent at a macro scale. Based on the perfectly cylindrical nature of the single cell model initially used by Innes (1995, 1996), the simplest repeating unit, based on this, is a hexagon (Fig. 5.2). Given the circular simplicity of the secondary cell wall layers, it is geometrically implausible that the cell would collapse flat on its own under a uniform hydrostatic tension pressure alone. Hence, a lateral displacement was applied to the model to force the cell into a non-circular shape. This was done by displacing (3 μm) the left and right edges inwards towards each other. Node displacements, and not edge forces, were used as otherwise the edge forces would have to be specified so as to maintain the straight edges required for the model to be cyclically repeating. There is some basis for this compressive force in reality as drying stresses can, and do, contribute to the occurrence of collapse, and for many cells these stresses will occur primarily in one axis only. To maintain cyclical constraints in the y-axis direction, it was necessary that the horizontal lines above and below the central whole cell should remain as a straight edge and horizontal (to prevent free body rotation). The simplest way of doing this was to tie the Y degree of freedom (DOF) for all the nodes on each line to the Y DOF to that of the central node (coloured red in Fig. 5.4). The horizontal compression was applied with a load case whereby all the nodes on the left and right edges were incrementally and linearly displaced towards each other over a
5
A Theoretical Model of Collapse Recovery
105
Fig. 5.2 Multi–cellular collapse model in MSC.Mentat showing areas of different material properties as developed in previous sections. (Blakemore 2008)
given time period. The increments had to be small enough to allow the solver to converge on the intermittent solutions, particularly to be able to calculate the non-linear plastic strain in the S2 layer and the CML. Given the temperature dependence of the properties in both the viscoelastic and elastic-plastic material models, a coupled solution was undertaken where the node temperatures were fixed at a given temperature for each increment. The internal negative pressure was also loaded linearly in the same loadcase. Once the full load had been applied, a second loadcase, which lasted for 16 hrs in total, was undertaken to observe the stress relaxation that occurs because of the viscoelastic behaviour of the S1 and S3 layers. In the first instance, a small-strain solver solution was obtained (Fig. 5.3) highlighting the first problem with this model, which is, the large amount of plastic shear strain occurring in the CML between the central cell and the four outer quarter cells. Using a large strain (Total Lagrange) solver, which is a more realistic method, the problem becomes even more evident (Fig. 5.4) since there is a large amount of shear occurring along the axis joining the centre of the four outer quarter cells and in the centre of the central whole cell, related to these geometric weak points (Fig. 5.2). Even if the geometry was changed so that the S1 layers shared at least one node along the centre joining lines, there would still be a large area of CML filling in the
106
P. Blakemore
Fig. 5.3 Cylindrical cell model obtained with the small strain solver. Highlighted in the close up section is the point between the cells where a large amount of plastic strain is predicted to occur in the CML. (Blakemore 2008)
gaps between any three adjoining cells. It should be noted that this sharing of nodes in the above model was not done as it would produce a long thin element of CML on either sides of the shared node. As a general rule, elements are more likely to provide a good solution if they are composed of approximately even-length sides.
5.2.3 Squared Based Cell Model The next improvement to the model then was to base the cells more on a square shape, as shown in Fig. 5.5. This square shape is also possibly more realistic of the type of lumen shape that occurs in the collapse prone group of eucalypts (Fig. 5.6). One of the reasons that a cylindrical model was attempted first was that the orthotropic orientations for the viscoelastic model are most easily assigned in terms of a cylindrical co-ordinate system. Fortunately, in meshing the central cell shown in Fig. 5.5, the quad elements were generated in a cylindrical pattern such that a given edge was always on the inside. The orientations of the orthotropic properties were then transformed to be relative to that edge. To apply a simple compressive displacement to the edge nodes in a similar manner to the previous model would still produce a similar problem, possibly even worse, of shear planes in the four corners,
5
A Theoretical Model of Collapse Recovery
107
Fig. 5.4 Cylindrical cell model obtained with large strain (Total Lagrange option) solver. (Blakemore 2008)
and a non-realistic collapsed shape. Instead of this then, a shear displacement in the y-axis was forced onto the model. A shear force in the x-axis would result in similar problems to that of applying a compressive force in this direction; except it would only occur in two opposing corners and not in all four. Given the longer unsupported edges at the top and bottom of the central cell, these edges are inherently more likely to buckle or collapse and hence the y-axis shearing is the more realistic for this base model. The shear displacements were applied to the highlighted nodes (black circles – Fig. 5.5) on the y-axis mid-plane. A shear displacement, instead of a shear force, was applied for similar reasons as in the previous stage of the model development (Fig. 5.4), where a compressive displacement was used instead of a compressive force. The compressive displacement keeps periodic symmetry, such that all cells at the boundaries move inward by the same amount. In this instance, while there was no requirement to maintain a straight edge, there were still problems with how realistic the resultant stress distribution from the applied shear forces would be. The displacements of the two outer nodes were constrained to be twice those of their respective neighbouring inner nodes. The two black circled nodes on the x–axis mid–plane also had their x-displacement fixed to zero to prevent any rotation of the model caused by the shear displacements.
108
P. Blakemore
Fig. 5.5 Grid layout, regions of the cell wall and forces for the square-cell based model showing a much reduced area of CML. The arrows indicate the negative hydrostatic pressure on the inner S3 layer. The other red lines between nodes show the tie and servo links used to enforce the symmetry considerations for this model. Black dots indicate important nodes for shear displacements and related symmetry conditions. (Blakemore 2008)
To maintain the cyclical constraints requirement in the y-axis direction, the nodes along the straight edge, at the top and bottom of the square central unit, were tied in the y-axis degree of freedom. To ensure the cyclical constraints were met in both the x and y directions, servo links were used on the nodes on the mid–planes of the four quarter cells, so that their displacements matched the pattern for the equivalent plane line on the central cell. A limitation in the software being used meant that the y-axis displacement of the two side edges could not be enforced, but this was not pursued further as the x-axis servo links on these nodes did a reasonable job of ensuring that the collapsed shape in the four quarter cells matched the equivalent section of the central whole cell. The locations of the servo and tie links are indicated by the red lines on Fig. 5.5. From the initial attempts to run this model, Fig. 5.7 shows the final increment of the model shown in Fig. 5.5 before the solver failed to converge on the next increment. This problem was largely a meshing issue in the corner areas, one of which is highlighted in Fig. 5.7 which was experiencing a high level of stress and
5
A Theoretical Model of Collapse Recovery
109
Fig. 5.6 Scanning electron micrograph of E. regnans cross section. C = collapsed fibres, U = uncollapsed fibres, V = vessel, R = ray cells. (Chafe et al. 1992)
Fig. 5.7 Square-cell based model in collapsed state that highlights a convergence issue in large strain corners of the S3 layer. (Blakemore 2008)
110
P. Blakemore
Fig. 5.8 Refinement of mesh in the high stress corners of the square based multiple cell model. (Blakemore 2008)
became too distorted for the next incremental solution to be found. Figure 5.8 shows the mesh refinement in this region that was used to overcome this difficulty. The collapsing period of this model was run with all of the nodal temperatures fixed at 25◦ C. The internal negative hydrostatic tension and shear (y-axis) displacements were chosen iteratively to just initiate contact on the internal walls (Fig. 5.9).
Fig. 5.9 Final shape of the collapsed square-cell based model with the stress distribution shown after the 16 hrs of stress relaxation were allowed to occur. (Blakemore 2008)
5
A Theoretical Model of Collapse Recovery
111
More extensive contact started to cause convergence problems for the solver. These problems may have been related, at least partially, to the artificial shear boundary conditions that were being applied in this model. The internal pressure applied in this instance was –4.9 MPa. The shear displacements applied to the four circled nodes on the horizontal mid-plane (Fig. 5.5) were –3.5, –1.75, 1.75 and 3.5 μm from left to right respectively. The internal pressure value of –4.9 MPa is not too dissimilar to the –5.33 MPa (at 25◦ C) that Innes (1996) obtained to strain the inner edge of his model to 95% of the value assumed to result in collapse. However, both values are high compared with the pressures that Kauman (1964) estimated were likely in collapsing cells which he estimated to be in the range of 1–2.35 MPa, based on the liquid meniscus having a radius in the range of 600–1000 Å. The high negative pressures in both this model and the Innes (1996) model could at least partially be explained by the artificially regular geometry and uniformity of material properties used in both of these models. In reality, shape and material irregularities are likely to act as weak points where collapse is initiated at lower negative pressures than required in the models here. Figure 5.10 shows the shape of the solutions after the negative internal
Fig. 5.10 Final shape of the collapsed square-cell based cell model after the negative hydrostatic pressure and the shear displacements have been removed. (Blakemore 2008)
112
P. Blakemore
Fig. 5.11 Heat up profile of temperature against time used to simulate reconditioning in the FEM. (Blakemore 2008)
pressure and the shear displacement boundary conditions have been removed. This figure shows that the model, despite a little relaxation, essentially maintains the deformed shape. Figure 5.11 shows the approximation to the measurements of the central internal board temperature shown in Blakemore and Langrish (2008) that was used as the basis for the application of heat to simulate reconditioning in this model. This was applied by using a table of these temperatures to change all of the nodal temperatures uniformly as a function of time. Figure 5.12 shows the shape and stress distribution of the model after the nodal temperatures have been increased to 100◦ C to simulate steaming. The recovery of cell shape is barely discernible from the shape shown in Fig. 5.13. The next main reason for the lack of collapse recovery in the model relates to the material properties being used. As discussed earlier, Eq. (5.1) was central to many of the material properties, and up to this point a density of 673 kg m–3 was used to be consistent with Innes (1996). This density is very high compared with most of the experimental material that has been used here. For this reason, the model was run again using a more moderate density value of 500 kg m−3 . The viscoelastic parameters used for this model are shown in Table 5.1 and the elastic-plastic model was also reconfigured for this density value. Figure 5.14 shows that the change of basic density to 500 kg m−3 had little effect on the recovery of the cell shape. The main differences were that a negative internal pressure of only 3 MPa was required to just initiate internal contact, and the highest residual stress in the model at the end of the steaming was predicted to be reduced from ∼55 to ∼36 MPa. Again, this might just highlight how important the effect of temperature is in this model, and how poorly it is currently understood at the cell wall scale. All of the material properties used in this model were also almost entirely obtained by analogy with measured material properties on small samples or boards. This is largely because of the difficulties in measuring
5
A Theoretical Model of Collapse Recovery
113
Fig. 5.12 Shape and stress distribution after nodal temperatures were increased to 100◦ C. (Blakemore 2008)
the mechanical properties of secondary cell wall layers directly. An alternative to measurement though, as outlined in (Blakemore 2008, Chapter 1) is the significant progress that has been achieved in modelling the mechanical properties of timber based on knowledge of the chemical microstructure and the mechanical properties of extracted chemical constituents. For example, Harrington et al. (1998) calculated the following cell wall elastic constants for Pinus radiata at 12% moisture content (Table 5.2 ). Many of the values shown in this table are an order of magnitude greater than the equivalent values shown in Table 5.3. Although, it should be noted that in this table the longitudinal direction (l) is parallel to the microfibril direction, and not parallel to the longitudinal orientation of the cell as is the “z” direction as indicated in Table 5.3. This is because in the model that used these data (Astley et al. 1998) the mean microfibril angle and known random variation around the mean, used in the different cell wall layers, could be varied to analyse its effect on board scale properties. To some extent, if the values used here are low it may have to some degree be compensated for the inherent stability of the simple and regular geometry of the model being used. While an attempt at generating similar values as shown in Table 5.2 might produce much more realistic values for an ash-type eucalypt
114
P. Blakemore
Fig. 5.13 Shape and stress distribution after nodal temperatures were increased to 100◦ C. In this instance the temperature effect on the elastic moduli in the S1 and S3 layers has been removed and the values remain fixed. (Blakemore 2008)
than those shown in Table 5.3 it is unlikely to lead to a significant improvement in the model at this stage. This is because this approach is still not able to provide essential information on the non-linear large deformation behaviour, the time dependent behaviour, or the temperature-dependent behaviour of the cell wall properties. All of which are critical for a significant improvement in the collapse and recovery behaviour of the model being attempted here. The other most obvious reason for explaining the differences in the two tables (Tables 5.2 and 5.3) is the stated moisture content for the two tables; respectively 12% moisture content and green. This highlights another problem with attempting to replicating (Table 5.2) for an ash-type eucalypt species, and that is the need to generate a table for green moisture Table 5.1 Single term de Prony-series values fitted to E(overall) as a function of time for a density value of 500 kg m–3 T (◦ C)
Ei (MPa)
λ1 (h)
E1 (MPa)
Eα (MPa)
21.5
358
2.45535
12.1385
345.583
5
A Theoretical Model of Collapse Recovery
115
Fig. 5.14 Shape and stress distribution after nodal temperatures were increased to 100◦ C. Model run assuming a basic density of 500 kg m–3 . (Blakemore 2008)
content. As Harrington et al. (1998; Harrington et al. 1998) acknowledged, “while it is expected that at the lower moisture contents – such as at 12% – the properties of extracted constituents will be much the same as those in situ, this may not hold true at higher moisture contents”. Hence, this approach may not be as useful for predicting green cell wall properties as it is for dried material properties. Related to this discussion about the effect of moisture content, the next major limitation of Table 5.2 Calculated elastic constants for the cell-wall layers at 12% moisture content in Pinus radiata (Harrington et al. 1998) Wall Layer
Et (GPa)
Er (GPa)
El (GPa)
υrt
υlt
υlr
Grl (GPa)
Glt (GPa)
Gtr (GPa)
S3 S2 S1 CML
8.43 9.85 8.54 5.07
7.98 9.16 8.02 5.12
50.36 63.96 53.10 18.43
0.39 0.39 0.38 0.38
0.33 0.33 0.33 0.31
0.32 0.33 0.32 0.31
2.65 3.02 2.66 1.78
3.00 3.38 3.02 2.11
2.68 2.96 2.66 1.88
NB: “l” refers to a direction parallel to the length of the microfibres, and not the longitudinal direction of the cell
116 Table 5.3 Material properties for different cell wall layers
P. Blakemore
Er (E11 ) Eθ (E22 ) Ez (E33 ) νrθ (ν12 ) νθ z (ν23 ) νzr (ν31 )
S1 and S3
S2
620 6,200 620 0.05 0.5 0.38
620 620 6200 0.38 0.05 0.5
the recovery model is that it was attempted with green material properties maintained throughout. It is possible that the change of mechanical properties occurring upon drying, with the cell in the collapsed and stressed state, could be critical for collapse recovery. Based on analogy again with the mechanical properties of whole boards, it could be assumed that the modulus of elasticity and shear strength in the dried state, at around 12% moisture content, are approximately 1.5 times the green values. However, this board analogy is for the board at two different moisture contents, in which the board is essentially drying stress-free in both states. Clearly this is not the case at the cellular level in the collapsed state, and hence it is likely that this analogy would be even more tenuous than the similar assumptions used up till this point. Nevertheless, given that the viscoelastic properties appear to have minimal effect on the model attempted here, it may be possible to formulate a different time-dependent behaviour to mimic the change in elastic moduli as moisture is removed. For this to be successful, it would be necessary to include some component of mechanosorptive strain, which is dependent on some form of drying model for the change in moisture content. Another initial response to this might be to try and incorporate a simple drying model. After all, the finite element method is very suited to analysing this sort of diffusion problem. Unfortunately, most of the commonly used standard finite element modelling packages typically only include a diffusion-based heat transfer capability, and have no capability for modelling moisture content in its own right. The main reason that only a heat transfer capability is included is that these type of software packages are mostly used to design and test metallic, laminate and elastomeric parts where heat transfer properties are often very important. The underlying governing equations for heat and mass diffusion are very similar and it is possible to obtain solutions for simple moisture diffusion problems by reformulating and analysing them as a heat conduction problem. However, even if a combined heat transfer and moisture diffusion model could be easily implemented, the collapse and recovery model will not be progressed significantly until a reliable relationship between the mechanical properties, at the cell wall scale, and (changing) moisture content can be better established. At the moment there are no good data, apart from at a board scale level again, for these relationships (Table 5.4).
−1 node
−0.00969 −0.47345 −11.85242 −0.77253
5.2
−0.00981 −0.12648 −12.02427 −0.64927
r (μm)
U (μm) σr (MPa) σt (MPa) σz (MPa)
S3
−0.00967 −0.54445 −6.57728 −0.87351
5.37 −0.00966 −0.55488 −1.32599 −0.94043
+1 node
S2
−0.00982 −0.61094 −1.14803 −0.87948
−1 node
−0.00986 −0.61525 −5.11821 −0.78025
7.2
Table 5.4 Results from FEM solution. Three layer orthotropic case (Blakemore 2008)
−0.00984 −0.68686 −9.06328 −0.71417
+1 node
S1
−0.00987 −0.98577 −9.03168 −0.82618
7.5
5 A Theoretical Model of Collapse Recovery 117
118
P. Blakemore
5.3 Summary A finite element model was attempted here to demonstrate the theoretical basis for collapse recovery. The model built on a simple single-cell based model of collapse that was developed by Innes (1995, 1996). The theoretical mechanism for recovery of collapse by steam reconditioning has essentially been that proposed by Ilic (1987, personal communication) (cited by Oliver, 1991), which assumes that the S1 and S3 layers are largely responsible for providing an elastic component of the cell walls and that the S2 layer is considered the most important for providing the inelastic material properties required to hold the cell in the collapsed or deformed state. To this end, a viscoelastic material model was developed for the S1 and S3 layers, while an elastic-plastic model was developed for the S2 layer. The model based on these material properties was not able to clearly demonstrate collapse recovery. This was largely attributed to the similarity of the dependence of the elastic moduli as a function of temperature in all cell wall layers. By removing the temperature dependence in the S1 and S3 layers, a much more significant, although still incomplete, recovery of cell shape was demonstrated. The lack of realistic behaviour for the model predictions has highlighted the paucity of knowledge about mechanical properties at the cell wall scale. Obviously, direct measurements at this scale are extremely difficult, if not impossible. The most successful approaches so far to estimate these properties has been to use a range of homogenisation and finite element modelling techniques based on the generalised knowledge of the cell wall ultrastructure and the properties of extracted chemical constituents. While not directly comparable, published values (Harrington et al. 1998) for Pinus radiata at 12% moisture content suggest that values used in this model may have been rather low. To some extent this may have compensated for the stiffening effect of the simple, but inherently stable, geometry used in this model. Even though the method used by Harrington et al. (1998) could have been used to determine better elastic moduli for the different cell wall layers than those used, it was not attempted here because there are still several critical limitations with this approach. These include that the method is possibly less reliable at high moisture content states and that it provides no additional information on critical behaviours such as non-linear large deformation stress-strain relationships, time or temperature-dependent behaviour, or moisture content (including moisture change or mechanosorptive strain) dependent behaviour. All of which may be critical for accurately modelling the deformation and stress distribution in the cell wall layers prior to steam reconditioning. Even if alternate attempts to simplify the moisture related behaviours were pursued, but that still accounted for the significant reduction in collapse recovery below 15% moisture content, the lack of good temperature-dependent data in the different secondary cell wall layers is currently a major impediment for developing the current model further. The other major improvement that could be made to the model developed here would be to include multiple cells with more realistic geometries and arrangements. Such an approach was attempted in the models by Astley et al. (1998), where real cross-sections of Pinus radiata tracheids were scanned and skeletonised to form the
5
A Theoretical Model of Collapse Recovery
119
geometrical basis of a finite element model used to predict the macroscale elastic properties. Unfortunately, while the skeletonisation process makes the realistic cell geometries much simpler to implement in a finite element model, it is not possible here. Mostly this is because it is the two-dimensional spatial arrangement of the different secondary cell wall layers that is considered critical to the collapse and collapse recovery behaviour. Scanning in real cross-sections is still a possibility, but, it would require much more complex image analysis programming to approximate where the cell wall layer boundaries occurred for the relevant material properties to be applied to the appropriate elements. The number of elements required in this type of approach would also make the finite element model considerably more computationally intensive. Nevertheless, as computer processing continues to become faster and cheaper, even in the near future this is unlikely to be a significant restraint for a model with up to 100 cells. This approach would also largely avoid the need for artificial constraints in the current model, such as the shear displacements required to achieve a more realistic flattening of the cell lumen.
References Astley RJ, Stol KA, Harrington JJ (1998) Modelling the elastic properties of softwood. Part II: The cellular microstructure. Holz Roh Werkst 56:43–50 Blakemore P (2008) Optimisation of steam reconditioning for regrowth-ash and plantation grown eucalypt species. PhD Thesis, The University of Sydney. 327 pp. http://hdl. handle.net/2123/2343. Accessed 3 August 2010 Blakemore PA, Langrish TAG (2008) Effect of pre-drying schedule ramping on collapse recovery and internal checking with Victorian Ash eucalypts. Wood Sci Technol 42(6):473–492 Chafe SC, Barnacle JE, Hunter AJ, Ilic J, Northway RL, Rozsa AN (1992) Collapse: an introduction. CSIRO Division of Forest Products, Melbourne, 9 pp Harrington JJ, Booker R, Astley RJ (1998) Modelling the elastic properties of softwood. Part 1: The cell wall lamellae. Holz Roh Werkst. 56:37–41 Innes TC (1995) Stress model of a wood fibre in relation to collapse. Wood Sci Technol 29:363–376 Innes TC (1996) Improving seasoned hardwood timber quality with particular reference to collapse. PhD Thesis, Faculty of Engineering, University of Tasmania, Tasmania, 207 pp Kauman WG (1964) Cell collapse in wood. CSIRO Division of Forest Products, Melbourne, 59 pp Oliver AR (1991) A model of the behaviour of wood as it dries (with special reference to Eucalypt materials). Civil and Mechanical Engineering Department, University of Tasmania, Tasmania, 107 pp Wardrop AB, Bland DE (1959) The process of lignification on woody plants. In: Proceedings of the 4th international congress of biochemestry. Pergamon Press, New York, NY, pp 76–81 Williams M.L, Landel R.F, Ferry John D (1955) The temperature dependence of relaxation mechanisms in amorphous polymers and other glass-forming liquids. J Am Chem Soc 77:3701–3707
Part II
Methodology for Delamination Detection and Factors Inducing and Affecting Delamination
Chapter 6
Delamination of Wood at the Microscopic Scale: Current Knowledge and Methods Lloyd Donaldson
Contents 6.1
Anatomical Features of Wood Delamination . . . . . . . . . 6.1.1 Weathering and Decay . . . . . . . . . . . . . . . . 6.1.2 Internal and Intra-Ring Checking . . . . . . . . . . . 6.1.3 Resin Pockets . . . . . . . . . . . . . . . . . . . . 6.1.4 Shelling . . . . . . . . . . . . . . . . . . . . . . . 6.1.5 Reaction Wood . . . . . . . . . . . . . . . . . . . 6.1.6 Induced Delamination . . . . . . . . . . . . . . . . 6.2 Ultrastructural Features of Cell Wall Delamination . . . . . 6.2.1 Ultrastructure of Wood Cell Walls . . . . . . . . . . 6.2.2 Location of Cell Wall Delamination . . . . . . . . . . 6.2.3 Mechanism of Delamination . . . . . . . . . . . . . 6.2.4 Influence of Microfibril Angle . . . . . . . . . . . . 6.2.5 Influence of Delignification and Pulp Refining . . . . . 6.2.6 Influence of Species . . . . . . . . . . . . . . . . . 6.2.7 Influence of Moisture Content . . . . . . . . . . . . 6.2.8 Influence of Temperature . . . . . . . . . . . . . . . 6.3 Microscopic Methods for Evaluation of Delamination in Wood 6.3.1 Light Microscopy . . . . . . . . . . . . . . . . . . 6.3.2 Confocal Microscopy . . . . . . . . . . . . . . . . 6.3.3 Electron Microscopy . . . . . . . . . . . . . . . . . 6.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . .
124 126 128 128 128 129 130 130 130 131 134 134 135 135 136 137 137 137 138 138 139 139
L. Donaldson (B) Bioproduct Development, Scion - Next Generation Biomaterials, 49 Sala St. Rotorua, Private Bag 3020, Rotorua 3046, New Zealand e-mail:
[email protected] V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_6,
123
124
L. Donaldson
6.1 Anatomical Features of Wood Delamination At the microscopic level, wood delamination can be defined as the separation or disintegration of fibres as a result of physical or chemical processes resulting in fracturing. Wood shows complex anisotropic behavior related to its microscopic structure, and this is also reflected in its fracture behaviour (Figs. 6.1 and 6.2). Delamination of wood can occur by intrawall fracture between adjacent tracheids
Fig. 6.1 Diagram of cell wall delamination for both radial and tangential planes, showing transwall and intrawall fracture types. Interwall fracture is a special case of intrawall fracture directly through the middle lamella or directly between individual tracheids or fibres
Fig. 6.2 Diagram illustrating the phenomenon of fibre bridging where single fibres or short rows of fibres span the developing fracture. This feature is more common in some species than others and is thought to prevent abrupt crack propagation resulting in a stepwise failure process
6
Delamination of Wood at the Microscopic Scale
125
or fibres, or in association with rays, or less commonly by transwall fracture, where the cell lumen is exposed (Koran 1967, 1968; Jeronimidis 1976; Kucera and Bariska 1982; Boatright and Garrett 1983; Côté and Hanna 1983; Zink et al. 1994; Donaldson 1997) (Figs. 6.2 and 6.3). Interwall fracture is a special case of intrawall fracture that occurs within the middle lamella (Côté and Hanna 1983; Zink et al. 1994). Delamination can occur naturally from weathering or decay, growth stress or drying stress (including brittleheart, surface, intra-ring and internal checking), physical damage to the living tree (resin pockets), and physiological stress (resin pockets, traumatic resin canals, shelling). These phenomena can be divided into radial delamination (intra-ring checking, internal checking, weathering, decay) and tangential delamination (resin pockets, traumatic resin canals, shelling). Brittleheart is a special case of transverse delamination. These cases of natural delamination often occur as a result of changes to the anatomical or chemical properties of the wood. Because wood tracheids and fibres are aligned in radial files, delamination may occur preferentially in the radial longitudinal plane between the rows of cells. Thus, wood splits easily in the radial longitudinal plane by crack propagation along the radial files of tracheids, and/or along the rays (Thuvander and Berglund 2000; Thuvander et al. 2000). There are differences between radial cracks that grow radially from a tangential surface, and those that grow longitudinally from a transverse surface. The former often shows deflections of the propagation when encountering a latewood boundary, with a stick-slip method of propagation due to the stress distribution produced by the alternating layers of soft earlywood and stiff latewood (Thuvander and Berglund 2000; Thuvander et al. 2000). Delamination by intrawall fracture is more difficult in the tangential plane because the tracheids or fibres are randomly arranged in this direction. Crack propagation must therefore follow an irregular course, which requires more energy, and may propagate more slowly. There are also fewer potential failure points, such as bordered pits, on the tangential
Fig. 6.3 Light micrographs of radiata pine showing: (a) Tangential delamination in thin walled mild compression wood with examples of both transwall (t) and intrawall (i) fracture. (b) Tangential delamination in thick walled tracheids showing exclusively intrawall fracture. (c) Radial delamination associated with internal checking showing intrawall fracture at the middle lamella (ml) region. Scale bar = 30 μm
126
L. Donaldson
walls. Delamination in the transverse plane is difficult because there is usually no defined path for crack propagation. Thus transverse delamination occurs mainly by transwall fracture but with some intrawall stepping (Zimmermann and Sell 1997). This type of fracture may be associated with brittleheart in some hardwood species, where compression failures in the cell wall provide the pathway for crack propagation (Dadswell and Langlands 1934; Green 1962; Wilkins 1986a, b). Of course, delamination can also occur in wood-based products as a result of applied stresses during service or manufacture and may be associated with adhesion problems when failure occurs at glue lines, for example. Wood structure can also influence these types of delamination. For example, penetration of adhesive into the wood structure, or even into the cell wall, will result in stronger bonds. The presence of any delamination at the glue line as a result of inadequate surface preparation will weaken the adhesion (Jokerst and Stewart 1976; Murmanis et al. 1983; Singh et al. 2002). Various other types of delamination occur on wood surfaces as a result of sawing or planing and these micro-cracks can be penetrated by various surface coatings resulting in improved adhesion of the coating (Singh and Dawson 2004, 2006; Singh et al. 2007).
6.1.1 Weathering and Decay Delamination associated with weathering and/or decay occurs as a result of chemical modification of the wood cell walls, primarily the breakdown and removal of lignin by the action of UV radiation in sunlight, and the stresses caused by wetting/drying cycles (Borgin 1971a; Sell and Leukens 1971; Bamber and Summerville 1981; Voulgaridis and Banks 1981; Feist and Hon 1984; Singh et al. 1995; Evans et al. 2000; Turkulin et al. 2001; Singh and Dawson 2003; Kim et al. 2008), or by the activities of microorganisms (Sandberg 1999). In addition to wetting and drying cycles, under some conditions freezing and thawing cycles also contribute. Delamination as a result of weathering usually occurs by simple cell separation at the middle lamella (Voulgaridis and Banks 1981; Evans et al. 2000), as a result of breakdown of lignin (Bamber and Summerville 1981). The wood breaks up into individual fibres or fibre bundles (Singh and Dawson 2003), which may also show thinning of the cell wall as a result of erosion of the secondary wall from exposed lumen surfaces. Bundles of microfibrils are peeled away due to degradation of the cell wall resulting from chemical and physical processes (Kim et al. 2008). In some species such as pine (Pinus spp), ray cells are mostly unlignified, and hence may break down well before tracheids and fibres, resulting in radial delamination. When tissues other than rays are involved, the separation of cells may result in mixed radial and tangential delamination (Bamber and Summerville 1981). In cases where mainly polysaccharides are degraded, leaving an intact lignin residue (e.g., brown rot), delamination may occur also in the transverse direction because of the brittleness of dry lignin, allowing transwall fracture to easily take place (Irbe et al. 2006). The presence of cellulose microfibrils tends to resist transwall failure in the transverse plane in undegraded tracheids or fibres.
6
Delamination of Wood at the Microscopic Scale
127
Radial and tangential surfaces may show different responses to weathering (Sandberg 1999; Sandberg and Söderström 2006). Sandberg and Söderström (2006) found that tangential surfaces show a greater number and width of cracks in pine and spruce (Picea spp), compared to radial surfaces. Cracks on the tangential surface occur in both earlywood and latewood, but on radial surfaces, cracks are more common at the ring boundaries (Sandberg and Söderström 2006). Delamination of the middle lamella occurs frequently in the latewood on tangential surfaces (Sandberg and Söderström 2006). The tracheid or fibre cell walls exposed on weathered or decayed surfaces may show checks that follow the alignment of the microfibrils in the S2 layer. These checks may be related to softrot cavities, or may result from breakdown of the cell wall matrix and the influence of wetting and drying cycles (Borgin 1971a; Nilsson and Daniel 1990; Blanchette et al. 1994). Delamination of pit membranes as a result of weathering has been described by Turkulin and Sell (1997). Checks in the secondary wall induced by weathering, show bridging involving macrofibrils (bundles of microfibrils), a behavior that mirrors the bridging of whole tracheids in radial delamination at a larger scale (Stanzl-Tschegg 2006; Keunecke et al. 2007) (Fig. 6.2). It seems likely that such bridging behavior will also influence fracture toughness (resistance to fracture) at the nanostructural level, as it does at the anatomical level. Borgin et al. (1975) examined a range of ancient wood samples with ages ranging from 900 to 4400 years. Delaminations were found at the middle lamella/S1 interface with cracks and fissures also present in other parts of the cell wall. Blanchette et al. (1994) found intrawall cracks and fissures within the secondary walls of archaeological wood from ancient Egypt. This form of physical degradation was associated with exposure to limestone, gypsum, sodium chloride and moisture. Donaldson (1993) found similar delamination associated with sodium chloride deposits in the cell walls of Podocarpus tracheids from wood that had been buried on the sea floor. In ancient wood samples, cracks and intrawall delaminations are often observed, even in apparently sound wood. Borgin (1971a) found intrawall delamination within both the secondary wall and the middle lamella in the absence of microbial degradation. Similar delaminations were observed in wood from ancient tombs by Nilsson and Daniel (1990), and in at least one case, this was associated with softrot attack. Daniel et al. (2004) used cryo-FESEM to study delamination in white-rot decayed birch wood (Betula verrucosa Ehr.). The decay progressively removed matrix material between macrofibrils, resulting in concentric delamination of layers of macrofibrils within the secondary wall. Ando et al. (2006) compared fracturing properties of old Japanese red pine (Pinus densiflora Sieb. Et Zucc.) wood from a Buddhist temple and new wood within 3 years of felling. They found more uneven and complicated surfaces dominated by transwall fracture in the old wood (270 years old), suggesting that there was a prolonged formation of microcracks before fracture in these samples. Many of these transwall fractures were initiated from bordered pits, suggesting that these structures have a role in concentrating stress. Microcracks may have accumulated at the
128
L. Donaldson
bordered pits during the 270 years in service. Fracture in new wood (120 mm) cracks than the internal boards, because in those boards there is a predominant influence of the tangential anisotropic direction. Aspen got a high number of relatively short cracks rather spruce got fewer, but longer and injurious. For both species, the boards near the bark have more injuries than those near the pith. The boards near the pith contain more juvenile wood than those near the bark. On the other hand, the juvenile wood has an important number of medulary rays which prevent the generation and development of cracks. (b) Microscopic Aspects The technological advancement with scanning electronic microscopy allowed the developments of studies related to the fine structure of wood. Numerous micrographs related to the photodegradation and weathering of wood have been published in the last three decades (Sell and von Luekens 1971; von Luekens and Sell 1972; Raczkowski 1980; Kucera and Sell 1987; Kuo and Hu 1991). In that follows we selected several to illustrate the crack formation and delaminations in fine structure of wood induced by artificial exposure to weathering for species from temperate zone (European) and tropical. The delamination effect in the transversal section of Southern pine cross section after exposure to UV (λ > 200 nm) during 1000 h can be observed in Fig. 9.9. The middle lamella was completely eroded, UV radiation producing delamination between the tracheids. The deterioration of the pits after 1000 h UV exposure generated delamination at the border of pits which extends in alignment with the microfibril orientation (Fig. 9.10). The high energy protons degraded the lignin and the cohesion between wood anatomical elements. Hon and Feist (1986) reported the effect of UV (λ > 220 nm) irradiation on yellow poplar after 500 h, 1000 h and 2000 h. Figure 9.11 shows the corresponding
a
b
Fig. 9.9 Delamination observed on the transversal section of Southern pine cross section after exposure to UV (λ > 200 nm) during 1000 h. (Williams 2005, Figure 7.26, Figure 7.27) (a) Cross section before the exposure (b) cross section after 1000 h exposure to UV (λ > 200 nm). The middle lamella was completely eroded producing delamination between the tracheids
188
V. Bucur
Fig. 9.10 Delamination observed on the radial section of Southern pine section after exposure to UV (λ > 200 nm) during 1000 h. Deterioration of pits generated cracks and delamination between tracheids (Williams 2005, Figure 7.28)
Fig. 9.11 Delamination observed on the transversal section of yellow poplar after exposure to UV (Hon and Feist 1986, Figure 7, Figure 8, Figure 9) (a) cross section after 500 h irradiation with λ > 200 nm; (b) cross section after 1000 h irradiation with λ > 220 nm (c) cross section after 2000 h irradiation with λ > 220 nm
9
Delaminations Induced by Weathering in Wood
189
Fig. 9.12 Delamination observed on the tangential section of yellow poplar after exposure to UV (Hon and Feist 1986, Figure 10, Figure 11) (a) roughening of the tangential section after irradiation with λ > 220 nm during 2000 h (b) delamination into the cell wall induced by the deterioration of pits observed on tangential section after irradiation with λ > 220 nm during 2000 h after 1000 h irradiation
deterioration of transversal microscopic structure (middle lamellae, checking and roughening of the cell wall). After 2000 h of exposure the delamination is evident through the separation of cells. It was reported that the tangential section the development of the very rough surface is due to the degradation of lignin and consequently the microfibrils emerged to the surface. The pits were severely damaged generating the delamination into the adjacent cell wall (Fig. 9.12). Kishino and Nakano (2004) reported the development of delamination produced by artificial weathering of eight tropical species (auri – Acacia auriculiformis; bangkai – Shorea spp, cumaru – Amnurana acreana, ipe – Tabebula spp, jahhra – Eucalyptus marginate, keruing – Dipterocarpus spp, robusta – Eucalyptus robusta). The development of cracks has been studied in relationship with the wettability from the prospective of chemical and structural modifications of wood surface. Each 120 min weathering cycle was composed from 120 min light irradiation (300 nm 30◦ ). The m coefficient decreases quite linearly with increasing grain angle in Picea jezoensis. Figure 15.5 depicts the variation of the acoustic events number, the signal amplitude, at various levels of loading up to the proportional limits. Two groups of different signals have been observed, at small angle (0◦ and 15◦ ) only signals of low amplitude 40–50 dB were generated in the early stage. The increasing of load determined the gradual increasing of higher amplitudes. For large angles, signals of higher amplitudes (65–75 dB) were generated from the early stage of fracture. The ex-situ inspection of fractured specimens showed that at small angles the crack propagated along the grain, with transwall typical fractures in earlywood. At large angles the cracks propagate along the grain, as interwalls fractures. The acoustic emission signals generated at small angle, before crack propagations correspond probably to the microcracks at the tip zone induced by delaminations between cell wall layers. Cyra and Tanaka (2000) studied fracture phenomena and acoustic emission in relation to routing cutting process. The acoustics events were related to the grain angle, the state of cutting and the surface roughness. The acoustic emission technique is promising for routing monitoring.
15
Acoustic Emission Activity Induced by Delamination and Fracture
a
b
c
d
293
Fig. 15.4 Relationship between the grain angle and fracture parameters and acoustic emission parameters (Ando et al. 1992, Figures 3, 4, 6, and 8), (a) COD and grain angle, (b) KIc and grain angle, (c) AE cumulative events (counts) and grain angle up to proportional limit of Load –COD curves, (d) m and grain angle Mokuzai Gakkaishi (1992, Figures 4, 5, 6, and 8)
The literature is very scarce in data on acoustic activity on specimens subjected to static torsion and fatigue. The shear mode of rupture under torsional loading is one of the most complex possible modes. Chen et al. (2006) investigated the behavior of Red lauan (called also Philipine mahogany, Shorea spp.) and Sitka spruce under torsional fatigue experiments. Table 15.2 shows the complexity of the fracture modes developed under torsional loading. In static torsional testing microcrack initiation was observed through the acoustic activity prior to maximum loading. Acoustic emission activity (events number) increases as the grain angle increases from 45◦ to 90◦ . The red lauan produced more counts than Sitka spruce. Keiser effect before cracking, in fatigue testing was reported for both species. Specimens under fatigue testing produced more events than under static tests.
294
V. Bucur
Fig. 15.5 Amplitude, acoustic events number and relative load, up to the proportional limit, on notched specimens in tension. (a) for small angle (0◦ and 15◦ ) and (b) for other angles between 30◦ and 90◦ . Mokuzai Gakkaishi (1992, Figure 7)
Table 15.2 Fracture modes at different grain angle orientation of specimens under torsional loading (data from Chen et al. 2006) Fracture modes at different grain angle orientation under torsion Species Red lauan Sitka Spruce
0◦
45◦
90◦
II RL, III RT
I RL, I RT
II RL, III RT
II TL, III TR
I TL, I TR
II TL, III TR
and III TR; II TL, III TR and III RT and III RT; II RL, III RT and III TR
15.3.3 Annual Ring Structure Ansell (1982) was probably the first to demonstrate the influence of the earlywoodlatewood ratios under tensile loading on acoustic emission activity, expressed by the shape of AE strain curve. Dill – Langer and Aicher (2000) observed micro fracture nucleation of spruce under tensile loading; it was an on-set of AE prior to the first visible crack growth step. Ando et al. (1991) studied the effect of the location of the crack tip in an annual ring of sugi (Cryptomeria japonica) in single edge notched specimens of the TR crack propagation system. The critical stress intensity factor KIc varied according to the location of the crack tip, from the pith to the bark or from the bark to the pith. The crack tip was located in earlywood or in latewood. When located in the
15
Acoustic Emission Activity Induced by Delamination and Fracture
295
earlywood, KIc increased with decreasing distance between the crack tip and the latewood located forward of the tip. The patterns of the AE amplitude distribution at various levels of loading up to the proportional limit in load-COD diagrams were studied. Two populations of peaks amplitudes were observed, at 45 dB and 70 dB, corresponding with two microcracks at different energy levels and recognized as transwall failure and intrawall failures (Ando and Ohta 1995). It was supposed that the variability of KIc by the crack tip position in the annual ring is due either to the difference in cell shape or cell wall thickness around the crack tip, and to the difference in stress concentration induced by crack location and the direction of crack propagation. Ando and Ohta (1999) extended the previous studies to Sitka spruce (Picea sitchensis), sugi (Cryptomeria japonica) and akamatsu (Pinus densiflora) taking into account the anisotropy and the heterogeneity of different zones in the annual ring with FEA. Figure 15.6 shows the variation of KIc as a function of the location of the crack tip in the annual ring for sugi (Cryptomeria japonica), spruce (Picea sitchensis) and akamatsu (Pinus densiflora), for pith side notched and bark side notched specimens. The stress at the tip is expressed by σtip = α σ and α = σtip /σ where: σ tip is the stress at the crack tip, in the tangential direction obtained with FEA α is the stress concentration factor σ is the nominal stress Figure 15.7 shows the variation of stress concentration factor α as a function of the location of the crack tip in the annual ring for sugi (Cryptomeria japonica), spruce (Picea sitchensis) and akamatsu (Pinus densiflora) for pith side notched (filled circles) and bark side notched (open circles) specimens. When the degree of stress concentration was small an acoustic emission signal was generated and an intrawall failure was observed, before the crack initiation. When the degree of stress concentration was large, a signal of large amplitude was generated and transwall failure was observed. The annual ring scale was also studied by Dill –Langer and Aicher (2000). They monitored simultaneously the crack propagation and the acoustic emission activity of notched spruce specimens in tension load, in RT and TR systems. A third configuration was studied for specimens at 45◦ between radial growth direction and load axis, with notch at 45◦ versus R. The damage mechanism was studied at micro (tracheids diameter 50 μm) and mezzo scales (annual ring width 3. . .5 mm). Confocal laser scanning microscopy was used for in-situ observation of crack growth. Two characteristic damage phenomena have been observed. When crack propagation is in TR system, or T propagation, the rupture of earlywood cell walls was observed (intrawall failure). When crack propagation in RT system, or R propagation, the debonding of the interface middle lamella between two adjacent tracheids was observed (interwall failure). The crack path was in zigzag for the third configuration with the specimen oriented at 45◦ . “The more the crack
296
V. Bucur
Fig. 15.6 Variation of KIc in the annual ring as a function of the relative position of the crack for sugi, spruce and akamatsu Permission J Wood Sci 45:275–283 Figure 9
approaches the earlywood/latewood transition the more it deviates from the initial direction until propagation coincides with 45◦ At this stage the crack surface consists predominantly of ruptured cell walls comparable to the rupture in RT system. Having reached half the specimen with the crack turns 90◦ anticlockwise propagating through late and transition wood. Thereby the crack surface is smooth as
15
Acoustic Emission Activity Induced by Delamination and Fracture
297
Fig. 15.7 Variation of stress concentration factor versus the relative position of the crack for sugi, spruce and akamatsu Ando and Ohta (1999, Figure 10)
in case of pure TR configuration, caused by the debonding of adjacent fibres with hardly any cell wall damage”. The 45◦ specimen was used to put in evidence the relationship between structural damage and acoustic emission activity. The onset of acoustic emission was prior to the first in-situ microscopic visible crack. More that 95% of the acoustic emission events and visible crack propagation were observed during a plateau of load displacement curve. The acoustic emission plot exhibited peaks in the same time as the load deformation plot. Dill –Langer and Aicher (2000) suggested using these observations as support for further theoretical modeling of damage in wood.
15.3.4 Tension Wood The presence of reaction wood in general (compression or tension wood) in lumber submitted to drying induce quality degradation. Cunderlik et al. (1996) monitored the drying cracks in the tension and opposite wood by acoustic emission and SEM. The opposite wood generates higher numbers of acoustic events than tension wood (Fig. 15.8). Tension wood activity decreases with increase of gelationous fibres proportion in wood tissue. In opposite wood dominated the cracks of high acoustic activity in the middle lamella (intercell), while in tension wood the delamination of gelatinous layer (G-layer) from the secondary wall (S2-layer) is characterized by low acoustic activity.
15.3.5 Moisture Content Acoustic emission technique from the early 1980 was related to checking detection and wood drying (Kawamoto and Williams 2002). Acoustic emission signals during drying are related to events produced by checking and water movement and it is difficult to distinguish the source of emission. Wave pattern recognition using cluster analysis was limited in applications for monitoring and controlling kilns (Schniewind et al. 1996; Lee et al. 1996). The major AE sources are the surface tensile stress induced by the water movement below the FSP and the thermal stress related to temperature variation. Wood microscopic structure shows slips in the crystalline segments of cellulose (Booker 1994). The checking occurs when the rate of
298
V. Bucur
Fig. 15.8 Acoustic emission count rate during drying of tension and opposite wood in beech (Cunderlik et al. 1996, Figure 1)
slips exceeded a critical value. The slip lines generate slow AE events. Brittle microcracks in the cell wall and delaminations generate rapid AE events (Sato et al. 1984; Schniewind 1989). Acoustic emission energy has been used to identify the damage during drying by Kowalski et al. (2004). Roughly three groups of AE signals were identified. The first group, observed at the beginning of drying process – with small amount of energy and an important number of events, inducing microcracks. The second group, identified during drying, with increasing acoustic emission energy and diminishing the number of events, when the surface of the specimen shrinks and the moisture content is around the fiber saturation point, inducing macrocraks. The final detected stage of drying corresponds to specimen core drying, and has relatively low energy AE signals. For better understanding of fracture phenomena during drying and the related acoustic emission activity, problems related to the attenuation of signals and transducers sensitivity must be solved.
15
Acoustic Emission Activity Induced by Delamination and Fracture
299
Fig. 15.9 Acoustic emission signals during oak drying (Kim et al. 2005, Figures 2 and 3). (a) cumulative counts versus drying time (b) AE waveform during surface check (c) AE waveform during water movement
Pattern classification of acoustic emission signals during wood drying by principal component analysis and artificial neural network for oak (Quercus variabilis) was proposed by Kim et al. (2005). The acoustic emission parameters selected for this study were: peak amplitude, ring-down count, event duration, frequency, energy, rise time and peak amplitude/rise time. The sources of acoustic emission events during drying were due to water movement and to surface check. The cumulative AE hits versus drying time and the waveforms of AE signals corresponding to surface check and water movement are shown in Fig. 15.9. AE signal produced by water movement shows lower in peak amplitude, longer in rise time and lower in peak frequency, then the signals caused by surface check. 96% of the variability of AE signals was accounted in the principal components plane 12. Kim et al. (2005) noted that when the value of the first principal component is greater than 1, the number of AE events cause by water movement are higher then AE events caused by surface checking. The classification of AE signals with artificial neural network (ANN) was performed in two hypotheses. The first ANN classifier has six input nodes for six AE parameters. The second ANN classifier has only two inputs for two principal components. All classifiers have two output nodes, the surface check pattern and the water movement pattern. Eight nodes (hidden layer) were chosen for optimum pre processing of ANN classifier. The neurons between layers were activated with tangent sigmoid function. The recognition rate was used to evaluate the performance of AE and ANN classifiers (Table 15.3). Surface checking was recognised at 83% with ANN classifier and AE parameter inputs and by 85% with ANN classifier and principal components inputs. This technique seemed to be promising for improving wood drying technology.
15.3.6 The Reused Old Wood Recycling old wood salvaged from old structures, is a real problem for the modern and green sustainable society (Ando et al. 2006, 2007). Acoustic emission technique
300
V. Bucur
Table 15.3 Recognition rate with ANN classifiers using AE parameters and principal components inputs (data from Kim et al. 2005) Water Inputs
Pattern
Learning (number)
AE parameters
Water Check Total Water Check Total
458 66 − 471 44 −
Principal components
Check Validation (number)
Learning (number)
396 83 − 448 74 −
42 434 − 29 456 −
Recognition Validation (number) 104 417 − 52 426 −
Rate
%
91.6 86.8 89.2 94.2 91.2 92.7
72.9 83.4 81.3 89.6 85.3 87.4
is appropriated for examining the differences between new and old wood in shearing fracture (Japanese standard JIS Z 2101-1994). Specimens of Japanese red pine (Pinus densiflora) from 270 old structural members were compared with specimens of new wood, lumbered within 3 years before testing. The cumulative AE event counts versus shearing stress is presented in Fig. 15.10. In the initial stage of loading
a
c
b Fig. 15.10 Behavior of old and new wood (Ando et al. 2006, Figure 3, 5) (a) cumulative AE events versus shearing stress in new wood; (b) cumulative AE events versus shearing stress in old wood; (c) m value versus relative stress (ratio of stress to maximum stress)
15
Acoustic Emission Activity Induced by Delamination and Fracture
301
the behaviour of new wood is very different from that of old wood. The difference in behaviour of old and new wood is well put in evidence in Fig. 15.10c with the variation of m value versus the relative stress. The increase in m value expresses the acoustic emission events of small amplitudes with stable crack propagation. The decrease in m value signifies the frequent occurrence of high-amplitudes AE events, with predominantly unstable fractures. The fractographic analysis revelled in new wood a smooth flat surface with intrawall failure. The fracture surface of old wood was rough and irregular of trans-wall type, initiated from the bordered pits. Under shearing test the old wood underwent stable crack propagation before the final fracture.
15.4 Some Aspects Related to the Energy of the Acoustic Emission Signals The acoustic emission energy, monitored with a high speed waveform acquisition system was reported by Landis and Whittaker (2001). The progressive crack growths along the grain in direction in notched eastern hemlock (Tsuga Canadensis) specimens has been studied in Mode I (Fig. 15.11). Crack length data and loadCMOD data were used to calculate the fracture energy. The release of the acoustic emission energy was calculated by integrating the instantaneous power of an elastic wave over all frequencies, and by multiply the result by the length of the AE waveform. From data plotted in Fig. 15.12 (which gives the variation of the measured
Fig. 15.11 Load (N), strain energy release (GIc ) versus fracture time (seconds) of a notched specimen of eastern hemlock (Tsuga Canadensis) (Landis and Whittaker 2001, Figure 2)
Fig. 15.12 The measured load and the cumulative acoustic emission energy release during fracture. (Landis and Whittaker 2001, Figure 4)
302
V. Bucur
Fig. 15.13 The linear relationship Between the acoustic emission energy rate and fracture energy release (Landis and Whittaker 2001, Figure 5)
load and cumulative acoustic emission energy release during fracture) it can be noted that: – an internal damage is produced before the growth of the crack – a constant rate at which the AE energy is released after the starting of crack growth – a constant rate of fracture energy release Figure 15.13 shows the linear relationship between the acoustic emission energy rate and the fracture energy release. A more refined approach to examine the mechanisms of energy dissipation during fracture perpendicular to grain and crack propagation in radial direction, in spruce was proposed by Watanabe and Landis (2007). It was hypothesized that the total dissipation energy during fracture is composed from two main components, the first one, corresponding to the dissipation of energy induced by short bursts and unstable crack growth, reflected by the strong acoustic emission activity and the second one, corresponding to additional energy, dissipated in the form of more gradual processes that include creep deformation, and slow crack growth, as can be seen from Fig. 15.14. Towards the end of the experiment the acoustic energy rises again, but before that, the fracture was slowed probably because of bridging effect combined with the action of other cohesive forces. More research is needed to understand the rapid rise of energy at the end of the test. It was advanced that this behavior is related to cohesive forces at the crack tip.
15.5 Summary Delamination and fracture phenomena in wood can be monitored non-destructively, continuously and in real time with acoustic emission technique. The conventional classical parameters of acoustic emission are: hit, counting/ring-down count/emission count, amplitude, duration, rise time, energy, average frequency, initial frequency, reverberation frequency, frequency centroid, peak frequency, rise time divided by amplitude called RA value, RMS (root mean squared values), the threshold voltage (ISO 12716). The approaches in analysing AE signals
15
Acoustic Emission Activity Induced by Delamination and Fracture
303
Fig. 15.14 The variation of AE energy, the consumed energy and load versus time, for a specimen of Canadian Eastern spruce tested in mode I (Watanabe and Landis 2007, Figure 5)
are: parameter-based techniques and signal-based quantitative techniques. The parameter-based techniques are the most popular for wood material studies. Acoustic emission technique operates using resonant or wideband transducers in ultrasonic frequency range (100 kHz–1 MHz). The most appropriated technique for wood is with piezoelectric transducers between 100 and 200 kHz. Acoustic emission is highly sensitive to the initiation and growth of delamination in wood and has advantages over conventional ultrasonic and radiographic methods. Transverse failure is one of the most important damage mechanisms controlling the loss of stiffness in wood which may be lifetime limiting in for structural members. Factors such as species, grain angle orientation, annual ring structure, moisture content, tension wood, etc effects the acoustic emission activity related to crack propagation, delamination and fracture phenomena. Acoustic emission technique provides a sensitive approach for real time detection of cracking, and also an unique view into the micromechanics of crack initiation and growth of delamination. The damage processes in the material under test can be observed during the entire load history without any deterioration of the specimen. The final objective of monitoring acoustic emission phenomena in wood is to provide beneficial information to prevent deterioration during processing (drying, etc) and catastrophic failure of the material. For the future, it could be suggested the development of new signal-based procedures, for wood and wood based composites, with a more quantitative analysis of the acoustic emission signals based on a 3D localization of AE sources and the recordings obtained from a sensor network. Another research field to be developed is related to the acoustic emission activity produced by micro-cracks with a two or
304
V. Bucur
tri-dimensional disordered lattice model of dynamic fracture, which can relate the acoustic response to the internal damage of the sample.
References Aicher S, Höfflin L, Dill-Langer G (2001) Damage evolution and acoustic emission of wood at tension perpendicular to fiber. Holz als Roh Werkst 59:104–116 Ando K, Takita A, Hirashima Y, Sasaki Y (2007) Fractography of old wood. Nagoya Univ Forest Sci 26:1–7 Ando K, Hirashima Y, Sugihara M, Hirao S, Sasaki Y (2006) Microscopic processes of shearing fracture of old wood, with acoustic emission technique. J Wood Sci 52, 6:483–489 Ando K, Ohta M (1999) Variability of fracture toughness by the crack tip position in an annual ring of coniferous wood. J Wood Sci 45:275–283 Ando K, Ohta M (1995) Relationship between the morphology of micro-fractures of wood and the acoustic emission characteristics. Mokuzai Gakk 41:640–646 Ando K (1993) Direct observation of micro-fracture process of wood by SEM and its acoustic emission characteristics under tension test. Proceedings 9th conference on acoustic emission, Osaka, Japan, pp 85–90 Ando K, Sato K, Fushitani M (1991) Fracture toughness and acoustic emission characteristics of wood. I. Effect of the location of a crack tip in an annual ring. Mokuzai Gakk 37:1129–1134 Ando K, Sato K, Fushitani M (1992) Fracture toughness and acoustic emission characteristics of wood II: effects of grain angle. Mokuzai Gakkaishi, 38(4):342–349 Ansell MP (1982) Acoustic emission from softwoods in tension. Wood Sci Technol 16:35–38 American Society for Nondestructive Testing – ASNT (2005) Acoustic emission testing. In Nondestructive Testing Handbook, 3rd edition, vol 6, Published by ASNT, Columbus OH ASTM E750-98 Standard practice for characterizing acoustic emission instrumentation Ballad EM, Vezirov SY, Pfleider K, Solodov IY, Busse G (2004) Nonlinear modulation technique for NDE with air-coupled ultrasound. Ultrasonics 42:1031–1036 Beall FC (2002) Overview of the use of ultrasonic technologies in research on wood properties. Wood Sci Technol 36(3):197–212 Booker JD (1994) Acoustic emission and surface checking in Eucalyptus Regnans boards during drying. Holz als Roh Werkst 52:383–388 Bucur V (2005) Acoustics of wood. Springer, Heidelberg Chen Z, Gabbitas B, Hunt D (2006) Monitoring of fracture of wood in torsion using acoustic emission. J Mater Sci 41(12):3645–3655 Chui CK (1992) Introduction to wavelets. San Diego, Academic Cunderlik I, Molinski W, Raczkowski J (1996) The monitoring of drying cracks in the tension and opposite wood by acoustic emission and SEM. Holzforschung 50:258–262 Cyra G, Tanaka C (2000) The effects of wood-fiber directions on acoustic emission in routing. Wood Sci Technol 34(3):237–252 Dill –Langer G, Aicher S (2000) Monitoring of microfracture by microscopy and acoustic emission. Proceedings internation conference Wood and wood fiber composites, Stuttgart, pp 93–104 Drouillard TF (1990) Anecdotal history of acoustic emission from wood. J Acoust Emission 9(3):155–176, 1990. Fausett LV (1994) Fundamentals of neural networks: architecture, algorithms and applications. Prentice Hall, Englewood Cliffs Grabec I, Sachse W (1997) Synergetics of measurements, prediction and control. Springer, Berlin Green RE Jr (2004) Non-contact ultrasonic techniques. Ultrasonics 42:9–16 Grosse C, Ohtsu M (2008) Acoustic emission testing basics for research – applications in civil engineering. Springer, Heidelberg
15
Acoustic Emission Activity Induced by Delamination and Fracture
305
Grosse CU, Finck F (2006) Quantitative evaluation of fracture processes in concrete using signalbased acoustic emission techniques. Cem Concr Compos 28:330–336 Grosse CU, Reinhardt HW, Finck F (2003) Signal-based acoustic emission techniques in civil engineering. J Mat Civ Eng. 15(3):274–279 Hill R, Brooks R, Kaloedes D (1998) Characterization of transverse failure in composites using acoustic emission. Ultrasonics 36:517–523 ISO 12716 – 2001 Non-destructive testing – Acoustic emission inspection JIS Z 2101 – 1994 Methods of test for woods Kawamoto S, Williams RS (2002) Acoustic emission and acousto-ultrasonic techniques for wood and wood-based composites: a review – General Technical Report FPL-GTR-134. Madison, WI Kim KB, Kang HY, Yoon DJ, Choi MY (2005) Pattern classification of acoustic emission signals during wood drying by principal component analysis and artificial neural network. Key Eng Materials 297– 300:1962–1967 Kowalski SJ, MolinskiW, Musielak G (2004) The identification of fracture in dried wood based on theoretical modelling and acoustic emission; Wood Sci Technol 38(1):35–52 Landis E N (2008) Acoustic emission in wood. In: Grosse C, Ohtsu M (eds) Acoustic emission testing basics for research – applications in civil engineering. Springer, Heidelberg, pp 311–322 Landis E N, Whittaker DB (2001) Acoustic emission as a measure of fracture energy. Proceedings of the 1st conference of the European society for wood mechanics, Vila Real, Portugal Lee SH, Quales SL, Schniewind AP (1996) Wood fracture, acoustic emission and the drying process. Part II. Acoustic emission pattern recognition analysis. Wood Sci Technol 30:283–292 Minozzi M, Caldarelli G, Pietronero L, Zapperi S (2003) Dynamic fracture model for acoustic emission. Eur Phys J B 36:203–207 Muravin B (2009) Acoustic emission, science and technology. J of Building and Infrastructure Engineering of the Israeli Assoc of Engineers and Architects (in press). www.muravin.com. Accessed 22 July 2010 Murphy JC, Majerowicz S, Green RE Jr, Glass JT (1990) Laser interferometric probe for detection of acoustic emission. Mater Eval 48:714–720 Ogawa M, Sobue N (1999) Effect of loading speed on fracture of timber with a crack. Mokuzai Gakk 45(6):461–470 Okoroafor EU, Hill R (1995) Investigation of complex failure modes in fibre bundles during dynamic mechanical testing using acoustic emission and Weibull statistics. J Mater Sci 30:4233–4243 Ono K (1997) Acoustic emission. In: Crocker MJ (ed) Encyclopedia of acoustics, Wiley, New York, NY Chapter 68:797–809 Persson K (1997) Modeling of wood properties by a micro-mechanical approach. Ph D Thesis, Lund University Report TV SM – 3020 Petri A (1996) Acoustic emission and microcrack correlation. Phil Mag B 77(2):491–498 Reiter A, Stanzl-Tschegg SE, Tschegg EK (2000) Mode I fracture and acoustic emission of softwood and hardwood. Wood Sci Technol 34(5):417–430 Reiter A, Stanzl-Tschegg SE, Tschegg EK (2002) Fracture characteristics of different wood species under Mode I loading perpendicular to the grain. Mater Sci Eng A 332:29–36 Ringger T, Höfflin L, Dill-Langer G, Aicher S (2003) Measurement of the acoustic anisotropy of soft and hardwood; effect of source location. Otto-Graff J 14:231–253 Sachse W, Kim KY (1987) Quantitative acoustic emission and failure mechanics of composite materials. Ultrasonics 25:195–203 Sasikumar T, Rajendraboopathy S, Usha KM, Vasudev ES (2008) Artificial Neural Network Prediction of Ultimate Strength of Unidirectional T-300/914 Tensile Specimens Using Acoustic Emission Response J. Nondestructive Eval. 27(4):127–133 Sato KN, Kamei M, Fushitani M, Noguchi M (1984) Acoustic emission generated upon mechanosorptive creep of wood. Mokuzai Gakk 30(8):653–659 Schniewind AP (1989) Concise encyclopedia of wood and wood-based material. Pergamon Press, Oxford
306
V. Bucur
Schniewind AP, Quales SL, Lee SH (1996) Wood fracture, acoustic emission and the drying process. Part I Acoustic emission associated with fracture. Wood Sci Technol 30:273–282 Scott IG (1991) Basic Acoustic Emission. Gordon and Breach Science Publishers, New York Serrano EP, Fabjo M (1996) Application of wavelet transform to acoustic emission signal processing. IEEE Trans Signal Proc 44(5):1270–1275 Solodov I Y (1998) Ultrasonics of non-linear contacts: propagation, reflection and NDEapplications. Ultrasonics 36:383–390 Stephens RWB, Leventhal HG (1974) Acoustic and vibration. Chapman and Hall, London Stoessel R, Predak S, Solodov I, Busse G (2003) In:Green RE Jr, Djordjevic BB, Hentschel MP (eds) Nondestructive materials characterization, Springer, Berlin, XI:117 Watanabe K, Landis EN (2007) An acoustic emission based study of energy dissipation in radially loaded spruce. In: Navi P, Guidon A (eds) Proceedings of the 3rd international symposium on wood machining, Lausanne, Switzerland, pp 179–182
Chapter 16
Delamination Detection in Wood – Based Composites Panel Products Using Ultrasonic Techniques Voichita Bucur and Saeed Kazemi-Najafi
Contents 16.1 16.2
Introduction . . . . . . . . . . . . . . . . . . . . . . Basic Aspects . . . . . . . . . . . . . . . . . . . . . 16.2.1 Waves Propagation Paths . . . . . . . . . . . . 16.2.2 Linear Ultrasonic Inspection Techniques . . . . . 16.2.3 Ultrasonic Transducers and Scanning Procedures . 16.2.4 Non-linear Ultrasonic Inspection Techniques . . . 16.3 Delamination Detection in Wood-Based Composite Panels 16.3.1 Through Transmission Technique . . . . . . . . 16.3.2 Plate Wave Technique . . . . . . . . . . . . . . 16.3.3 Industrial Applications of Non-contact Technique . 16.4 Summary . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
307 309 309 311 320 327 329 329 339 343 344 345
16.1 Introduction Wood – based composite panel products (WBCP) are manufactured from veneer, wood particles, strands or fibres bind together with different types of adhesives such as urea-formaldehyde resin, phenol-formaldehyde resin, melamine formaldehyde resin, methylene diphenyl diisocyanate or polyurethane resins. The nature and the quality of the raw material and of the adhesives determine the characteristics of the products (mechanical properties, water resistance, dimensional stability, surface quality and machinability). The products existing on the market can be classified such as: – glued laminated timber – glulam, crosslam, glulam slabs, – veneer based panels – plywood, laminated veneer lumber, V. Bucur (B) CSIRO, Materials Science and Engineering Div. Bayview Avenue, Clayton, Victoria 3168, Australia e-mail:
[email protected] V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_16,
307
308
V. Bucur and S. Kazemi-Najafi
– particle based boards – oriented strandboard, particleboard, fibre board, – mineral bonded particleboard and fibreboard, – wood – plastic composites. Wood-based panel products have become increasingly specialized in recent years and are used in a wide range of applications. The demand for panels is forecast to increase in the next decades as quality logs for traditional products become increasingly scarce and as designers and consumers gain experience with positive product attributes and new applications. Wood based composite panels will probably in the future, substitute some metallic structural elements and plastics. The strength and durability of wood based materials are determined by their homogeneity and flaw presence. Similar to other composite materials, flaws and damage in WBCPs are: – inherent processing flaws (voids, debonding, fibre breakage, non-uniform fibre and matrix distribution, fibre misalignment, foreign inclusions, ply gaps, delaminations, and matrix cracking). These defects tend to degrade certain structural design properties and remain in the material throughout its life cycle – in-service damage, induced by the exposition to various environmental and mechanical loading conditions (impact events, hygrothermal cycles and fatigue in very variable environmental conditions of temperature and humidity, etc). The main defects observed in service are: matrix cracking and crazing, ply gaps and delamination, fibre pullout, fibre fracture, degradations by aging and environmental aggressive conditions. Delamination is one of the most important defect which occurs in WBCP. In order to design and use WBCP with confidence, it is important to assess their integrity and evaluate their tolerance to flaws. Unfortunately, some flows cannot be detected until failure occurs. Therefore, the structural integrity of WBCP should be assessed by means of nondestructive evaluation methods at the earliest possible stage of damage existence. Nondestructive evaluation techniques have been developed to minimize the effects of defects and to insure the quality control of the final product. Non-destructive inspection techniques (Bucur 2003a, b) with different sensitivity levels can be used for non-destructive evaluation of wood-based composites, such as: X-ray and γ-ray radiography, thermography, ultrasonic techniques, acoustic emission and acousto-ultrasonics. Among these techniques, ultrasonic techniques are the most popular, due to the ease of integration into the production line, of the relatively low cost and inherent safety. The advantages or the disadvantages of different available techniques depend on the type of damage to be detected and on the testing conditions. Sophisticated laboratory techniques can give highly accurate results, but may not be able to assess the state of the structure under in-service conditions. The aim of this chapter is to review the existing ultrasonic techniques for delamination detection in WBCP.
16
Delamination Detection in Wood – Based Composites Panel Products
309
16.2 Basic Aspects Theoretical aspects related to the propagation of elastic waves in anisotropic solids have been discussed in numerous reference books and articles (Kolsky 1963; Viktorov 1967; Musgrave 1970; Auld 1973; Green 1973; Graff 1975; Krautkrämer and Krautkrämer 1990; Birks and Green 1991; Nayeh 1995; Schmerr 1998; Rose 1999). It was stated that the real part of the elastic constants of the materials can be determined by measuring two main parameters, the velocity of the elastic waves and the density of the material. The presence of delamination induces modification of materials mechanical and elastical properties which can be detected with ultrasonic techniques. Ultrasonic inspection involves the utilization of stress waves having a frequency higher than 20 kHz. Linear and non linear ultrasonic techniques have been developed for ultrasonic inspection of wood and wood – based composite materials.
16.2.1 Waves Propagation Paths Bulk waves or surface waves can be used for the characterization of the mechanical behaviour of wood-based composites panel products. The waves characteristics related to the propagation in an infinite solid of bulk waves – longitudinal and shear waves, (named also P and S waves) are shown in Fig. 16.1. The longitudinal waves are characterized by the fact that the direction of wave propagation is parallel to the direction of particle motion (polarization). In the case of shear waves the particle motion is perpendicular to the direction of wave propagation. For the inspection of plate type specimens, as wood based composites panels, it is also interesting to use Lamb waves. The Lamb waves are elastic wave modes
Fig. 16.1 Schematic representation of propagation of bulk longitudinal and shear waves in solids (Olympus, Panametrics, Figure 3, page 41)). http://www.olympus-ims.com/data/File/ panametrics-UT.en.pdf (visited 16 june 2009)
310
V. Bucur and S. Kazemi-Najafi
Fig. 16.2 Plate modes (Tucker 2001, Figure 1, page 7). https://research.wsulibs. wsu.edu:8443/dspace/bitstream/ 2376/53/1/b_tucker_052101.pdf. (a) Schematic representation of the displacement with plate wave modes; (b) plate mode signal composed from extensional and flexural waves
propagating in solid plates with free boundaries and representing a combination of both compression waves and shear waves. There are two distinct modes of Lamb waves, discernible by their particle displacement patterns and velocities (Fig. 16.2), – extensional mode which is symmetric versus the axis of the plate; compression and tension effects can be observed; the displacement is due to Poisson’s effect. Each symmetric mode has infinite number modes (s0 , s1 , s2 , . . . , sn ). The particle motion is parallel with the direction of wave propagation. This mode is relatively nondispersive (not dependent on frequency f ) if hf < 0.5 and where h is the plate thickness. – flexural mode which is antisymmetric versus the axis of the plate. Each antisymmetric mode have an infinite number modes (a0 , a1 , a2 , . . . an ) The particle motion is perpendicular to the direction of wave propagation. This mode is highly dispersive The lowest order Lamb wave modes (s0 and a0 ) which are the two fundamentals are commonly termed “plate waves” and are mostly used for non-destructive testing of wood – based composite panels. Lamb wave propagation occurs when the wavelength λ is as 0.1h < λ < 10h, where h is the plate thickness. For experimental reason λ > 10h was recommended by Bray and Stanley (1997) while λ > 5h and λ > 3h were proposed by Huang
16
Delamination Detection in Wood – Based Composites Panel Products
311
(1999) for experiments with composites and with extensional and flexural waves, respectively. It is understood that for a type plate specimen, the length and width of the plate must be much greater than the wavelength used for the inspection. The layered solid in which plate waves propagate can be understand as a homogeneous solid because of the long wavelength. The techniques based on Lamb wave propagation are able to measure the flexural and transverse shear rigidity of laminated fibre composites. Defect detection using Lamb waves can be performed by measuring wave velocity or attenuation. Numerous reference books and articles (Bland 1988; Birks and Green 1991; Nayeh 1995; Rose et al. 1987, 1992; Schmerr 1998; Rose 1999; BarCohen and Chimenti 1986; Tang 1988; Tang and Henneke 1989a; Huang et al. 1998) reported the utilization of velocity and attenuation measurements for the evaluation of elastic constants of fibre composites and for the detection of defects such as delaminations and disbondings. The measurements of attenuation of ultrasonic waves are more complicated than of velocity. The coupling medium could be a high source of error for attenuation measurements, but is not as critical in velocity measurements. Non contact transducers are recommended to avoid all these problems. Tucker (2001), Tucker and Bender (2003), and Tucker et al. (2003) successfully utilised the Lamb waves for continuous nondestructive inspection of wood-based composite panels and of wood-plastic composites.
16.2.2 Linear Ultrasonic Inspection Techniques The linear inspection techniques have been developed in the hypothesis that the acoustic wave amplitude is infinitesimally small and the response of the material is assumed to be linear to the excitation signal, obeying Hook’s law. Under the label– linear ultrasonic techniques- three main groups of techniques are recognized: – reflection technique, or pulse – echo technique, for which only one transducer is used. The ultrasonic wave is directed into the specimen and after propagating twice through its thickness is recorded by the same transducer. This technique works with continuous waves or with pulses – through –transmission technique for which two transducers are used, the transmitter and the receiver. The energy injected by the transmitter travel through the specimen and is recorded by the receiver. This technique works with continuous waves or with pulses – emission technique, known also as acoustic – emission technique, uses only one transducer which is a receiver and which collects waves emitted by the specimen under mechanical stress. This technique is not described in this chapter. For all these techniques the contact transducers are coupled to the specimen with a coupling medium or with a delay line. Non contact transducers called also air coupled transducers can be used for the generation of bulk waves or surface waves without contact between the specimen and the transducers. In this case the coupling medium is the air (or an other gas).
312
V. Bucur and S. Kazemi-Najafi
16.2.2.1 Contact Techniques A typical ultrasonic inspection system consists, very roughly, of three main units: the signal generator system with a corresponding preamplifier and amplifier, the transducers and the data acquisition system (PC and oscilloscope). The signal generating system or the pulser, generates short, large amplitude electric pulses of controlled energy, which are converted into short ultrasonic pulses when applied to an ultrasonic transducer. The transducer is the core of all non-destructive ultrasonic inspection procedures. The basic requirements for ultrasonic transducers are: good sensitivity and resolution, controlled beam pattern, reproducible performance under various testing condition and high signal to noise ratio. In that follows pulse –echo technique and through transmission technique will be described. Pulse Echo Technique The typical pulse echo inspection configuration is shown in Fig. 16.3a.The operation principle of the pulse echo technique is to excite the test sample into a mechanical vibration with an ultrasonic transducer driven by a pulse, and to measure the time of wave propagation through the material using the same transducer acting
Fig. 16.3 Pulse echo technique with one transducer. (a) transducer in contact with the specimen (Olympus Panametrics 2009); (b) schematics of the pulse echo technique and signal display in a sound and defective zone ( Stoessel 2004, Figure 19). http://elib.uni-stuttgart.de/ opus/volltexte/2004/1622/pdf/ Dis_Stoessel.pdf
16
Delamination Detection in Wood – Based Composites Panel Products
313
as a receiver. Driven by the pulser, the transducer generates energy converted into mechanical vibration which travels twice through the material to be tested. When the signal travels through a structural discontinuity of the material, a flaw (i.e. delamination), part of the energy will be reflected back from the flaw surface. The reflected signal is captured by the receiver and is displayed on the screen of the oscilloscope. The ratio between the distance and the travelling time of the signal gives the velocity of propagation of the ultrasonic wave through the material under test. A more sophisticated signal treatment gives information about the location, the size, the orientation and other features of the defect. In pulse echo methods, a single-sided access is needed and this is an advantage for testing panels. The transducers used for this purpose must generate pulses of broad band frequency and short rising time. For high attenuator materials such as wood – based composites this technique has some limitations, especially in high frequency range. The echo technique allows the direct localization of a reflector as for example the back wall of the specimen or flaw. A clear echo from the back of a specimen means that the specimen is free of defects. The presence of a defect in a specimen can be identified when echoes with short time of flight and echoes with long time of flight are measured (Fig. 16.3b). If the velocity in the sound material is known as well as the size of the specimen, the defect can be located. Figure 16.4 shows a structural element with important cracks detected with pulse echo technique. The measurements were performed with longitudinal waves and 100 kHz. A very strong echo was observed at 6.2 cm, which corresponds to the big crack observed on the photograph. The echo from the back wall is at 12.5 cm for which the velocity was 1630 m/s. Through Transmission Technique The operation principle of the through transmission technique is to excite the test sample into a mechanical vibration with an ultrasonic transducer driven by a short pulse, and to measure the time of propagation with a second transducer acting as a receiver and disposed to the opposite side of the sample (Fig. 16.5). In this case the bulk wave energy is transmitted through the panel thickness and is received by a second transducer on the opposite side of the specimen This technique is much easier in application than the pulse echo technique because the signal travels only once through the thickness. Changes observed in received signal amplitude or other signal characteristics are induced by the internal structure of the material. This response is then used to measure the velocity of propagation of the ultrasonic wave and furthermore for the characterization of the mechanical behaviour of the specimen. By measuring the time-of-flight, the amplitude of the ultrasonic signal or other parameters of the signal (i.e the RMS voltage), the location and the size of the defects can be estimated. The technique using the normal incidence of the transducers to the surface of the specimen is most sensitive to flaws parallel to the surface (delaminations) (Smith et al. 1989; Wooh and Daniel 1994; Wooh and Wei 1999; Hosur et al. 1998; Žukauskas et al. 2005) while defects lying perpendicular to the surface (cracks in
314
V. Bucur and S. Kazemi-Najafi
Fig. 16.4 A structural element inspected with pulse echo technique (Hasenstab et al. 2006). http://www.ndt.net/article/ecndt2006/doc/Th.2.4.1.pdf (a) view of the structural element in pine (15.5 × 12.5 × 75 cm) (original Figure 3, page 4); (b) pulse echo measurements along the sound zone in A-scan mode and in B-scan mode, with P wave transducers of 100 kHz (original Figure 2, page 4); (c) ultrasonic image of the specimen scanned in B-mode, the crack is clearly identified at 6.2 cm (original Figure 5, page 5)
the matrix, fractured fibres, etc) are detectable with transducers oriented at an angle to the surface of the specimen (Moran et al. 1985; Wooh and Daniel 1990; Gorman 1991; Steiner et al. 1995). It was demonstrated that a combination of normal and oblique incidence with pulse-echo ultrasonic techniques can be used to produce a highly detailed volumetric image of complex damage states dominated by transverse matrix cracks and delamination. A comparison between the pulse echo method and the through transmission method in B-scan mode is shown in Fig. 16.6. The echo from the defect can be seen between the echoes from the front and back wall of the specimen. In through transmission technique the defect is represented by a lower amplitude signal.
16
Delamination Detection in Wood – Based Composites Panel Products
Fig. 16.5 Through transmission method. (a) Schematic representation of the principle of the through transmission method; (b) transmission of acoustic waves in two zone, a sound zone and a zone with defects (Stoessel 2004, Figure 20, page 75). http://elib.unistuttgart.de/opus/volltexte/2004/ 1622/pdf/Dis_Stoessel.pdf
315
a
b
Fig. 16.6 Comparison between the pulse echo method and through transmission method in B – scan mode (Stoessel 2004, Figure 21a, b). (a) reflection technique in B scan mode; (b) through transmission technique in B scan mode. http://elib.uni-stuttgart.de/opus/volltexte/ 2004/1622/pdf/Dis_Stoessel.pdf
316
V. Bucur and S. Kazemi-Najafi
16.2.2.2 Non-contact Techniques The non contact ultrasonic technique and the utilisation of air-coupled ultrasound transducers have gained considerable attention in the last 20 years. The elimination of the physical contact (with a gel) between the transducers and the material under inspection has been an enormous step towards the successful implementation of ultrasonic techniques in numerous applications and industrial processes (Hayward 1997; Castaings et al. 1998; Buckley 1999, 2000; Bharadwaj 2002; Bharadwaj et al. 2000; Blomme et al. 2002; Stoessel et al. 2001; Vun et al. 2003; Kleinschmidt 2003; Döring et al. 2006; Solodov et al. 2006a, b; Lionetto et al. 2007). Technological advancements with air-coupled transducers have made possible the study of the non linear acoustical behaviour of many materials and on the other hand permitted the development of automated nondestructive techniques for materials quality assessment. The non contact technique is a sensitive tool for testing materials’ structure and strength, for monitoring the degradations produced by the environmental conditions, and for quantifying the damage of structural elements. The air-coupled transducers can be configured to work in through transmission or to generate guided plate waves (Fig. 16.7). In conventional through transmission configuration of normal incidence, a beam of air coupled ultrasound excites longitudinal waves which can detect defects in materials. The flexural wave velocity can be detected using focused slanted transmission of air-coupled ultrasound. This transmission mode is used to generate and detect locally the flexural waves in wood and to measure their velocities. Figure 16.8 shows the methodology proposed by Solodov et al. (2004b). To excite the flexural waves in plate specimen with air coupled ultrasound, the “resonance” values of the angle
Fig. 16.7 Possible configurations of air-coupled transducers (Airstar 2001). http://airstar1.com/ air-coupled%20us.htm – air star coupled ultrasound, 2009-06-22. (a) through transmission, longitudinal wave; (b) shear wave; (c) plate wave two sided inspection; (d) plate wave one-sided inspection
16
Delamination Detection in Wood – Based Composites Panel Products
317
Fig. 16.8 Focused slanted transmission of air coupled ultrasound (Solodov et al. 2004b, page 506). (a) experimental configuration θ angle of incidence, θ 0 “resonance” angle α azimuthal angle; (b) plate wave excitation with air coupled ultrasound, the ultrasonic wave in air and the plate wave in the sample move in phase along the surface (Döring et al. 2006, Figure 2, page 2). http://www.ndt.net/article/ecndt2006/doc/P123.pdf
of incidence need to be determined. This is obtained by sample rotation. The “resonance” is manifested by important rise of the amplitude of the transmitted plane wave. In this case, the angle of incidence θ must satisfy the following condition sin θ =
Vair Vplate
(16.1)
where V are the velocities in air and in plate. In this case, the ultrasonic wave in air and the propagating plane wave in the material move in phase along the surface (Fig. 16.8b). This phase coincidence determines the “resonance”. The “resonance” is understood as the synchronous excitation of the plate wave and its re-radiation into the air from the opposite side of the plate. The acoustic coupling between the plate surface, its thickness t is obtained if t ≤ λs , where λs is the wavelength of the shear wave of the plate material. By measuring θ 0 the velocity of the flexural mode of the plate wave can be measured, along an arbitrary direction of propagation. The wavelength in air is small, about 0.7–0.35 mm in ultrasonic frequency range between 500 kHz and 1 MHz. this provides a localized quasi-plane wave spot of 2–4 mm in the focused area. The advantage of this technique is to provide a single point excitation of the flexural plate waves for use in the experiments for remote monitoring of in-plane local anisotropy in plate type specimens. As noted by Haberger et al. (1979), for Lamb waves propagating in symmetrical directions of thin orthotropic plates, the following relation can be used Vplatei =
√
Ei ωt 3ρ(1 − νik .υki )
1/4 (16.2)
where Vplatei is the velocity in the plate, ω is the frequency, t is the thickness of the plate, Ei is the Young’s modulus, ρ the density and ν the Poisson ratios. Their product is 120 mm) and in OSB, OSL, LVL, LVP and plywood having the thickness up to 200 mm. The system detects blows, delaminations and air pockets using a resonance technique. Papadakis and Kovacs (1980) described in detail this technique used firstly for quality assurance of iron parts. The principle of this method is based on the fact that a resonance occurs at the frequency at which an integral number of half-wavelengths fit into some dimensions of the workpiece. The ultrasonic energy is imparted to the workpieces which vibrates at its natural frequencies. In a zone with delaminations, the natural frequencies are altered. The analysis of the resonance frequencies permits the detection of the defective zone. The main advantage of the resonance technique is the capability to sample almost the entire workpiece (except the areas around nodes that are not stressed). “The question of corrective action in the presence of a reflection from a potential flaw that may or may be not of a detrimental size or shape is still a management decision rather then a purely scientific one” (Papadakis 1999). The patent developed by Electronic Wood System (2006) used non contact ultrasonic transducers installed across the panel width, in transmission technique configuration (Fig. 16.30). The minimum size of detectable delamination is 1 cm (Kleinschmidt 2003). Several thresholds allow display of a multicoloured ultrasound picture of the panel. The picture recognizes variations of moisture content, density, thickness and temperature using ultrasound, microwave and X ray techniques. The air coupled ultrasonic transducers are encapsulated and the system is protected against noise produced by sander, saws, compressed air, heat and dust. A similar system exists in Australia at Heyfield, Victoria (Hurley 2008) for the detection of the internal checking within dried eucalyptus boards. The non contact scanning system is equipped with 16 emitters and 16 receivers. The transducers are manufactured by Airstar Inc. US (Loertscher et al. 1996). The detection is performed on boards travelling at 1800 m/min in production line. A real time coloured picture is displayed on each board and is used for further operations (trimming, etc). The non contact ultrasonic technology is very robust, inherent safety and cost effective; the transducers are protected against external noise, heat and dust, humidity and high temperature; the maintenance time is significantly reduced.
344
a
V. Bucur and S. Kazemi-Najafi
b
Fig. 16.30 Industrial equipment for delamination detection on LVL (courtesy of Airstar) (https://airstar1.com/BLOW%20DETECTOR/.htm)(a) transversal view of the equipment for testing of LVL and air coupled ultrasonic transducers; (b) frontal view of the same equipment with Power Sonic Resonance Technology (https://ews-usa.com/images/ewsproducys/eval.ofwood.pdf)
16.4 Summary Wood – based composite panel products are manufactured from veneer, wood particle, strands or fibres bind together with different type of adhesives. Delamination is one of the most important defect which occurs in these products. The aim of this chapter is to review the ultrasonic techniques used for delamination detection. Delamination induces modifications of the mechanical and elastical properties of the materials, which can be observed with ultrasonic techniques. Ultrasonic inspection involves the utilisation of stress waves having a frequency higher then 20 kHz. Linear and non linear ultrasonic techniques have been developed for ultrasonic inspection of wood based composites. Linear techniques have been developed in the hypothesis that the acoustic wave amplitude is infinitesimally small and the response of the material is assumed to be linear tp the excitation signal. Hook law is valid. Under the label of –linear ultrasonic techniques- three main groups of techniques are recognized: the reflexion technique or the pulse echo technique, the transmission technique, and the emission technique. This last technique is not described in this chapter. For ultrasonic signal transmission to the specimens, contact and non contact transducers called also air coupled transducers, can be used. Bulk waves and Lamb waves (plate waves) are used for the mechanical and elastical characterization. Technological advancements with non contact transducers have made possible the development of studies related to the non linear behaviour of materials. For wood-based composites testing the contact transducers are piezoelectric, rather the air coupled transducers can be either piezoelectric or capacitive transducers. Images of panels’ internal structure can be obtained with different scanning procedure. The most common modes are: A-scan, B-scan and C scan. The non linear behaviour of anisotropic materials can be observed by the increasing number of higher harmonics having increasing amplitude with the distance. In these materials the non linear effect is enhanced by the presence of defects, because of non linear motion of their boundary. Delamination detection in wood-based composites was studied
16
Delamination Detection in Wood – Based Composites Panel Products
345
with through transmission technique and with plate wave technique. In both cases the transducers can be in direct contact with the specimen or can be air coupled. Laboratory experiments have been reported for the delamination detection in clear specimens or in specimens of structural size. An industrial non contact technique was patented for the detection of delamination in MDF OSB OSL LVL LVP, etc. The principle of this technique is based on the fact that a resonance occurs at the frequency at which an integral number of half-waveslengths fit into some dimensions of the workpiece, which vibrates at its natural frequencies. The presence of a delamination modifies the frequency. A real time picture is displayed on each workpiece and is used for further operations. The main advantage of this technique is its capability to sample almost the entire workpiece. The air coupled transducers are encapsulated and protected against dust, noise, etc. the non contact technology for delamination detection is very robust, inherent safety and cost effective.
References Aicher S, Dill-Langer G, Ringger T (2002) Non-destructive detection of longitudinal cracks in glulam beams. Otto-Graf-J 13:165–182 Alleyene DN, Cawley P (1992) The interaction of Lamb waves with defects. IEEE Trans Ultrasonics Feroelectric Freq Control 39:381–397 Auld BA (1973) Acoustic fields and waves in solids, vol I and II. Wiley, New York Aymerich F, Meili S (2000) Ultrasonic evaluation of matrix damage in impacted composite laminates. Composites: Part B 31:1–6 Bar-Cohen Y, Chimenti DE (1986) Nondestructive evaluation of composite laminates by leaky Lamb waves. In: Thompson DO, Chimenti DE (eds) Review of progress in quantitative NDE, vol l. 5B. Plenum Press, New York, pp 1199–206 Beall FC (1987a) Fundamentals of acoustic emission and acousto-ultrasonics. In: Proceedings of 6th International Symposium Nondestructive Testing of Wood, Washington State University, U. S Pullman, WA. 1987 September 14–16, pp 3–22 Beall FC (1987b) Acousto – ultrasonic monitoring of glueline curing. Wood Fiber Sci 19:204–214 Bekhta PA, Niemz P, Kucera, LJ (2000) The study of sound propagation in the wood-based composite material. In: Proceeding of 12th International Symposium on Nondestructive Testing of Wood, Sopron, Hungary, Sopron, 13–15 September 2000, pp 33–41 Berndt H, Schniewind AP, Johnson GC (1999) High resolution ultrasonic imaging of wood. Wood Sci Techn 33:185–189 Bharadwaj MC, Neeson I, Stead G (2000) Introduction to contact free ultrasonic characterization and analysis of consolidated materials. In: Proceedings of Application of NDE in powder metals. Invited paper Iowa State University, NDT net-june 2000, Vol 5, No. 06, http://www.ndt.net/article/v05n06/bhardwaj/bhardwaj.htm Bharadwaj MC (2002) Nondestructive evaluation: introduction of non-contact ultrasound. In: Schwartz M (ed) Encyclopedia of smart materials. Wiley, New York, pp 690–714 Birks AS, Green RE Jr.(1991). Ultrasonic testing. vol 7, 2nd edn. In: McIntire P (ed) Nondestructive testing handbook. American Society for Nondestructive Testing, Columbus, OH Bjorno L (2001). Forthy years of non-linear ultrasound. Ultrasonics 40:11–17 Bland DR (1988) Wave theory and applications Oxford University Press, New York Blomme E, Bulcaen D, Declerc NF (2002) Recent observations with air coupled NDE in the frequency range of 650 kHz to 1.2 MHz. Ultrasonics 40:153–157 Bray D E, Stanley RK (1997) Nondestructive evaluation: a tool in design, manufacturing and service. CRC Press, New York, NY
346
V. Bucur and S. Kazemi-Najafi
Buckley J (1999) Principles and applications of air-coupled ultrasonics. INSIGHT 40, 11:755–759 Buckley J (2000) Air-coupled Ultrasound – A millennial review. In: Proceedings of the 15th World Conference of Non-Destructive Testing, Roma, Italy, 15–21 October 2000, Available on http://www.ndt.net/article/wcndt00/papers/idn507/idn507.htm Buckley J, Loertscher H (1999) Frequency considerations in air-coupled ultrasonic inspection. In: British Institute of Non-Destructive Testing (BINT), September 1999, UK Cheshire, Plenum Press, 41(11):696–699, Available on: http://joe.buckley.net/papers/BINDT99%20Airscan.PDF Bucur V (1992) Anisotropy characterization of structural flackboard with ultrasonic methods, Wood Fiber Sci 24(3):337–346 Bucur V (1995) Acoustics of wood. CRC Press, Boca Raton, FL, pp 266 Bucur V (2003a) Nondestructive characterization and imaging of wood, Springer, Heidelberg, p 354 Bucur V (2003b) Techniques for high resolution imaging of wood structure: a review. Meas Sci Technol 14:R 91–R98 Bucur V, Ansell M P, Barlow C Y, Prithard J, Garros S, Deglise X (1998) Physical methods for characterization wood composites panel products, Holzfurchung 52(5):553–561 Burmester A (1968) On the correlation between sound velocity and density, tensile strength perpendicular to the surface and bending strength of particleboard. Holz als Roh werkstoff 113–117 Castaings M, Cawley P (1996) The generation, propagation, and detection of Lamb waves in plates using air-coupled ultrasonic transducers. J Acoust Soc Am 100(5):3070–3077 Castaings M, Cawley P, Fralow R, Hayward G (1998) Single side inspection of composite materials using air-coupled ultrasound. J Nondestructive Eval 17:37–45 Chimenti DE (1997) Guided waves in plates and their use in materials characterization. Appl Mech Rev 50(5):247–284 Dang C, Schmerr LW, Sedov A (2002) Ultrasonic transducer sensitivity and model-based transducer characterization. Research. Nondestructive Eval 14(4):203–228 Declercq NF, Degrieck J, Leroy O (2005) Bounded beam interaction with plate-edge at Lamb angle. Acta Acustica united with Acustica 91(2):326–332 Declercq NF, Degrieck J, Leroy O (2006) Ultrasonic polar scans: numerical simulation on generally anisotropic media. Ultrasonics 45(1–4):32–39 Delsanto PP (2007) Universality of nonclassical nonlinearity. Applications to non-destructive evaluations and ultrasonics. Springer, New York Dill-Langer G, Bernauer W, Aicher S (2005) Inspection of glue-lines of glued-laminated timber by means of ultrasonic testing. In: 14th International Symposium on Nondestructive Testing on Wood, Hanover, Germany, pp 49–60 Dill-Langer G, Bernauer W, Aicher A (2006) Inspection of glue-lines of glued laminated timber by means of ultrasonic testing. NDT.net 11(4) http://www.ndt.net/article/v11n04/dill-langer1/dilllanger1.htm Dimanche M, Capretti S, Del Senno M, Facaoaru I (1994) Validation of a theoretical approach for the detection of delamination in glued laminated beams. In: proceedings of 1st European symposium on nondestructive evaluation of wood, Sopron, pp 250–260 DIN EN 1330 – 4 (2000) Non-destructive testing – terminology Part 4: terms used in ultrasonic testing. Trilingual version DIN – Deutsches Institut für Normung Beuth Verlag GmbH Berlin, Berlin Döring D, Pfleiderer K, Solodov I, Busse G (2006) Automated NDT of stiffness anisotropy with air-coupled ultrasound. ECNDT 2006, p 8. http://www.ndt.net/article/ecndt2006/doc/P123.pdf. Accessed 7 November 2009 Electronic Wood Systems North America (2006) Evaluation of wood-based panel properties and delamination detection by means of ultrasound, based on current technology on online sonic measurements. See www.ews-usa.com and Electronic Wood Systems GmbH, Germany www.electronic-wood-systems.de and http://ews-usa.com/index.php?option=com_ content&task=view&id. Accessed 15 March 2006
16
Delamination Detection in Wood – Based Composites Panel Products
347
Gan TH, Hutchins DA, Green RJ, Andrews MK, Harris PD (2005) Noncontact, high-resolution ultrasonic imaging of wood samples using coded chirp waveforms. IEEE Trans Ultrason Ferroelect Freq Control 52(2):280–288 Gorman MR (1991) Ultrasonic polar backscatter imaging of transverse matrix cracks, J Compos Mater 25:1499–514 Graff K (1975) Wave motion in elastic solids. Ohio State University Press, Columbus, OH Grandia WA, Fortunko CM (1995) NDE applications of air-coupled ultrasonic transducers. In: Proceedings of IEEE International Ultrasonics Symposium, Seattle, WA, USA, pp 697–709 Green RE Jr. (1973) Ultrasonic investigation of mechanical properties, vol 3 of treatise on materials science and technology. Academic, New York Green RE Jr. (1999) Application of non-contact ultrasonics to nondestructive characterization of materials. In: Green RE Jr (ed) Nondestructive characterization of materials, vol 10, Kluver Academic, Plenum Press, New York, pp 419–420 Green RE Jr. (2004) Non-contact ultrasonic techniques. Ultrasonics 42:9–16 Grimberg R, Savin A, Lupu A, Iancu L, Rotundu C (2000) A method to determine the debonding zones in multilayer wood materials. In: Proceedings of the 15th world conference of non-destructive testing, Roma, Italy, 2000, IDN 522 Grimberg R, Savin A, Steigmann R, Bruma A (2005) Ultrasound and visual examination of wood based products, In: The 8th international conference of the slovenian society for non-destructive testing -application of contemporary non-destructive testing in engineering, September 1–3, Portorož, Slovenia, pp 109–115 Guo N, Cawley P (1993) The interaction of Lamb waves with delaminations in composite laminates. J Acoust Soc Am 94:2240–2246 Haberger CC, Mann RW, Baum GA (1979) Ultrasonic plate waves in paper. Ultrasonics 17: 57–62 Han Jun-Bo, Cheng JC, Wang TH, Berthelot Y (1999) Mode analyses of laser- generated transient ultrasonic Lamb waveforms in a composite plate by wavelet transform. Mater Eval 57(8): 837–840 Hasenstab A, Krause M (2005) Defect localisation in wood with low frequency ultrasound echo technique. In: proceedings of 14th international symposium on nondestructive testing of wood, Hannover, pp 329–336 Hasenstab A, Krause M, Osterloh K (2006) Testing of wooden construction elements with ultrasonic echo technique and X –ray. ECNDT 2006 –Th.2.4.1:1–8 http://www.ultrasonic.de/article/ecndt2006/doc/Th.2.4.1.pdf Hayward G, Gachagan A (1996) An evaluation of 1–3 connectivity composite transducers for air-coupled ultrasonic applications. JASA 99(4):2148–2157 Hayward G (1997) Air-coupled NDE: constraints and solutions for industrial implementations. In: Proceeedings of IEEE Ultrasonics Symposium, Toronto, Ontario, Canada, pp 665–673 He Y, Manful D, Bardossy A, Dill-Langer G, Ringger T, Aicher S (2004) Fuzzy logic based denoising of ultarsound signals from non-destructive testing, Otto-Graf-J 15:103–120 Hillger W (1992) Ultrasonic imaging of damages in CFRP-laminates. Acoust Imaging 19:575–579 Hillger W (1997) Ultrasonic imaging of internal defects in composites. NDTnet 2(5). Paper at UTonline application workshop May ’97. http://www.ultrasonic.de/article/ wsho0597/hillger/hillger.htm and http://www.ultrasonic.de/wshop/wshop_ap/wshop_ap.phtml. Accessed 12 June 2007 Hosten B (1991) Reflection and transmission of acoustic plane waves on an immersed orthotropic and visoelastic solid layer. JASA 89:2745–2752 Hosur MV, Murthy C RL, Ramamurthy TS, Shet A (1998) Estimation of impact-induced damage in CFRP laminates through ultrasonic imaging. NDT E Int 31(5):359–374 Huang W (1998) Application of Midlin plate theory to analysis of acoustic emission waveform in finite plates. In: Thompson DO, Chimenti DE (eds) Proceedings 24th Annual Symposium on Quantitative NDE, July 27–August 1, 1997, San Diego, CA, Plenum Press
348
V. Bucur and S. Kazemi-Najafi
Huang W, Ziola SM, Dorighi JF, Gorman MR (1998) Stiffness measurement and defect detection in laminated composites by dry-coupled plate waves. In: Bossi RH, Pepper DM (eds) Proceedings of SPIE,. March 31–April 2, San Antonio, TX, USA Hurley V (2008) A successful case study of R&D, commercialization and adoption of new technology by a wood products company. In: Wood Innovation 2008 Proceedings 18–19 September, Melbourne: 182–183 (www.fridayoffcuts.com ) Hutchins DA, Schindel DW (1994) Advances in non-contact and air – coupled transducers. In: Proceedings IEEE Ultrasonic Symposium 1994, Electronic Identifier 10.1109/ULTSYM. 1994.40181,1–4 November 1994, pp 1245–1254 Kabir M F, Araman P A (2002) Nondestructive evaluation of defects in wood pallet parts by ultrasonic scanning. In: Proceedings of 13th international symposium on nondestructive testing of wood, California, Berkley, USA Kabir MF, Schomoldt DL Shafer ME (2001) Roller-transducer scanning of wooden pallet parts for defect detection. In: Thompson DO, Chimenti DE (eds) Review of Progress in Quantitative NDE, vol 20. pp 1218–1225 Karim MR, Mal AK, Bar-Cohen Y (1990) Determination of the elastic constants of composites through the inversion of leaky Lamb wave data. In: Thompson DO, Chimenti DE (eds) Review of Progress in Quantitative NDE, vol 9A, Plenum Press, New York, pp 109–116 Kazemi Najafi S, Bucur V, Ebrahimi G (2005) Elastic constants of particleboard with ultrasonic technique. Mater Lett 59:2039–2042 Kazemi Najafi S, Abbasi Marasht A, Ebrahimi G (2007) Prediction of ultrasonic wave velocity in particleboard and fiberboard. J Mat Sci 42:789–793 Kazemi Najafi S, Bucur V (2002) Nondestructive characterization of particleboard with acoustic methods. In: Proceedings of 6th conference of French acoustical Soceity, April 2002, Lille, France Kleinschmidt H (2003) Ultra-Scan delamination detection with new power sonic resonance technology increases panel board production. In: Proceedings of 7th European panel products symposium, North Wales Conference Center, Llandudno UK - October 2003, pp 111–115 Kolsky H (1963) Stress waves in solids. Dover, New York, NY, USA Krautkrämer J, Krautkrämer H (1990). Ultrasonic testing of materials, 4th edn. Springer, Berlin Kundu T, Potel C, de Belleval JF (2001) Importance of the near Lamb mode imaging of multilayered composite plates. Ultrasonics 39:283–290 Lionetto F ; Tarzia A ; Maffezzoli A (2007) Air-coupled ultrasound :a novel technique for monitoring the curing of thermosetting matrices. IEEE Trans Ultrason Ferroelect Freq Control 54:1437–1444 Loertscher H, Grandia B, Strycek J, Grandia W (1996) Airscan transducers, techniques and applications. NDTnet 1(9):4. http://www.ndt.net/article/qmi/qmi.htm Lynworth LC (1989) Ultrasonic measurement for process control. Theory, techniques, applications. Academic, Boston, MA Martin RW, Chimenti DE (1987) Signal processing of leaky Lamb wave data for defect imaging in composite laminates. In: Thompson DO, Chimenti DE (eds) Review of Progress in Quantitative NDE, vol 6A, Plenum Press, New York, pp 815–824 Moran TJ, Crane RL, Andrews RJ (1985) High-resolution imaging of microcracks in composites, Mater Eval 43:536–540 Musgrave MJP (1970) Crystal acoustics. Holden-Day, San Francisco, CA Nayeh AH (1995) Wave propagation in layered anisotropic media: with applications to composites. Elsevier, New York Niemz P, Kucera L J, Poblete H, Baradit E(1996) On the application sound velocity to the determination of bending strength of particleboard. In: Proceedings of 10th international symposium on nondestructive testing of wood, Lausanne, Switzerland Niemz P, Kucera L J, Schob M, Scheffler M (1999) Possibility of defect detection in wood with ultrasound. (Experimentelle Untersuchungen zur Erkennung von Defekten in Holz mittels Ultraschal). Holz als Roh und Werk 57:96–102
16
Delamination Detection in Wood – Based Composites Panel Products
349
Neuenschwander J, Niemz P, Kucera LJ (1997) Orientierende Untersuchungen zur Anwendung der bildgebenden Ultraschallprüfug zur Fehlererkennung in Holz. Holz als Roh-und Werkst 55:339–340 Olympus, Panametrics (2009) http://www.olympus-ims.com/data/File/panametrics-UT.en.pdf. Accessed 16 June 2009 Papadakis EP (1999) Ultrasonic instruments and devices. Reference for modern instrumentation, techniques and technology. Academic, San Diego, CA Papadakis EP, Kovacs BV (1980) Theoretical model for comparison of sonic-resonance and ultrasonic velocity for assuring quality in instant nodular iron parts. Mater Eval 38(6):25–30 Potel C, de Belleval KF (1993) Propagation in a periodically anisotropic multilayered media. J Acoust Soc Am 93(5):2669–2677 Potel C, de Belleval JF, Genay E, Gatignol Ph (1996) Behavior of Lamb waves and multilayered Rayleigh waves in an anisotropic periodically multilayered medium. Application to the longwave length domain, Acustica-Acta Acustica 82(5):738–748 Potel C, Leduc D, Morvan B et al. (2008) Lamb wave attenuation in a rough plate. I. Analytical and experimental results in an anisotropic plate. J Appl Phys 104:074908–074908–10 Rogers WP (1995) Elastic property measurement using Rayleigh-Lamb waves. Res Nondestruct Eval 6:185–208 Rodgers JM, Green AT, Borup SW (1991) Acousto-ultrasonic measurement of internal bond strength in composite wood products. Mater Eval 49(5):566–571 Rose JL (1999) Ultrasonic waves in solid media. Cambridge University Press, Cambridge Rose JL, Zhu W, Cho Y (1992) Boundary element modelling for guided wave reflection and transmission factor analyses in defect classification. IEEE Ultrason Proc Symp 1:885–888 Rose WR, Rokhlin SI, Alder L (1987) Evaluation of anisotropic properties of graphite – epoxy composites using Lamb waves. In: Review of Progress in Quantitative Nondestructive Evaluation, vol 6B, Plenum Press, New York, pp 1111–1118 Schindel DW, Hutchins DA, Zou L, Sayer M (1995) The design and characterization of micromachined air-coupled capacitive transducers. IEEE Trans Ultrason Ferroelect Freq Contr 42:42–50 Schmerr LW Jr (1998) Fundamentals of ultrasonic nondestructive evaluation, a modelling approach. Plenum Press, New York Schmoldt D L, Ross RJ, Nelson R M (1996) Ultrasonic defect detection in wooden pallet parts for quality sorting. SPIE Proc Series 2944:285–295 Solodov IY (1998) Ultrasonics of non-linear contacts: propagation, reflection and NDE applications. Ultrasonics 36:383–390 Solodov IY (2001) CAN: an example of nonclassical acoustic nonlinearity in solids. Ultrasonics 40:621–624 Solodov I, Strössel R, Busse G (2004a) Material characterization and NDT using focused slanted transmission mode of air-coupled ultrasound. Res NonDestruct Eval 15:1–21 Solodov I, Pfleiderer K, Busse G (2004b) Nondestructive characterization of wood by monitoring of locall elastic anisotropy and dynamic nonlinearity. Holzforschung 58:504–510 Solodov I, Busse G (2006) New advances in air-coupled ultrasonic NDT using acoustic mode conversion. ECNDT 2006 Berlin – We.2.4.2 http://www.ultrasonic.de/article/ ecndt2006/doc/We.2.4.2.pdf. Accessed 7 January 2007 Solodov I Y, Doering D, Pfleiderer K, Busse G (2006a) Linear and nonlinear NDE using aircoupled Lamb waves. AIP Conf Proc 820:1492 Solodov I Y, Pfleiderer K, Gerhard H, Predak S, Busse G (2006b) New opportunities for NDE with air-coupled ultrasound NDT Int 39:176–183 Stiffler RC (1986) Wave propagation in composite plates. Ph.D. Dissertation, College of Engineering, Virginia Polytechnic Institute and State University, Blacksburg, VA Smith BT, Heyman JS, Buoncristiani AM, Blodgett ED, Miller JG, Freeman SM (1989) Correlation of the deply technique with ultrasonic imaging of impact damage in graphite-epoxy composites. Mater Eval 47(12):1408–1416
350
V. Bucur and S. Kazemi-Najafi
Steiner KV, Eduljee RF, Huang X, Gillespie JW Jr (1995) Ultrasonic NDE techniques for the evaluation of matrix cracking in composite laminates. Compos Sci Technol 53:193–198 Stössel R, Krohn N, Pfleider K, Busse G (2001) Air-coupled ultrasound inspection of various materials. Ultrasonics 40:159–163 Stössel R, Krohn N, Busse G (2000) Measurements with air-coupled ultrasound. In: Proceedings of 7th Internat. Congres on Sound and Vibration, July 4–7, Garmish – Partenkirchen vol II: 795–802 and Acoust Phys J (2002) 48(3):159–163 Stössel R (2004) Air-couples ultrasound inspection as a new non-destructive testing tool for quality assurance. PhD Thesis, University of Stuttgart, Germany Strycek JO, Loertscher H (1999) Ultrasonic air-coupled inspection in advanced material. NDT.net 4:46–50 Strycek JO, Grandia WA, Loertscher H (1997) Wave modes produced by air coupled ultrasound. NDTnet May 1997, 2(5). http://www.ndt.net/article/wsho0597/qmi2/qmi2.htm. Accessed 10 September 2008 Tang B, Henneke EG II (1989a) Lamb wave monitoring of axial stiffness reduction of laminated composite plates. Mater Eval 47(8):928–932 Tang B, Henneke EG II (1989b) Long wavelength approximation for Lamb wave characterization of composites laminates. Res Nondestructive Eval 1:51–64 Tang B, Henneke EG II, Stiffler RC (1988) Low frequency flexural wave propagation in laminated composite plates. In: Duke JC Jr (ed) Acousto – ultrasonics: theory and application. Plenum Press, New York, NY, pp 45–65 Tucker BJ, Bender DA, Pollock DG, Wolcott MP (2003a) Ultrasonic plate waves evaluation of natural fiber composite panels. Wood Fiber Sci 35:266–281 Tucker BJ, Bender DA (2003) Continuous ultrasonic inspection of extruded wood-plastic composites. Forest Products J 53(6):27–32 Tucker BJ (2001) Ultrasonic plate waves in wood-based composite panels. Ph. D Dissertation, Department of Civil and Environmental Engineering, Washington State University, p 112 Ty Ch (1989) Modulus of elasticity of particleboard determined by nondestructive testing methods. J Agric For 38(2):151–164 Viktorov I A (1967) Rayleigh and Lamb waves: physical theory and applications. Plenum Press, New York, NY Vun RY (2003) Ultrasonic characterization of engineering performance of oriented strandboard. Louisiana State University – etd (electronic thesis and dissertations), http://etd.lsu.edu/ docs/available/etd-0708103-163628/ Vun RY, Wu Q, Bhardwaj MC, Stead G (2003) Ultrasonic characterization of structural properties of oriented strandboard: a comparison of direct-contact and non-contact methods. Wood Fiber Sci 35(3):381–396 Vun RY, Wu Q, Monlezun CJ (2003) Ultrasonic characterization of horizontal density variations in oriented strandboard. Wood Fiber Sci 35(3):482–498 Vun RY, Hoop C, Beall FC (2005) Monitoring critical defects of creep rupture in oriented strandboard using acoustic emission: incorporation of EN300 standard. Wood Sci Techn 39(3):199–214 Vun RY, Hoover K, Janowiak J, Bhardwaj M (2008) Calibration of non-contact ultrasound as an online sensor for wood characterization: effects of temperature, moisture, and scanning direction. Appl Physics A Mat Sci 90(1):191–196 Vun YR, WuQ, Bhardwaj M, Stead G (2000) Through – thickness ultrasonic transmission properties of oriented strandboard. In: Proceedings of 12th international symposium on nondestructive testing of wood, Sopron, pp 76–68 Wang CS, Wu F, Chang FK (2001) Structural health monitoring from fiber reinforced composites to steel reinforced concrete. Smart Mat Struct 10:548–552 Wooh SC, Daniel IM (1990) Enhancement techniques for ultrasonic nondestructive evaluation of composite materials. J Eng Mat Tech 112:175–182
16
Delamination Detection in Wood – Based Composites Panel Products
351
Wooh SC, Daniel IM (1994) Three-dimensional ultrasonic imaging of defects and damage in composite materials. Mater Eval 52(10):1199–1206 Wooh SC, Wei C (1999) A high-fidelity ultrasonic pulse-echo scheme for detecting delaminations in composite laminates. Composites: Part B 30:433–441 Žukauskas E, Cic˙enas V, Kažys R (2005) Application of air–coupled ultrasonic technique for sizing of delamination type defect in multilayered materials. Ultragarsas 54(1):7–11
Chapter 17
Delamination Evaluation of in-Service Glulam Beams and other Structural Members Via Ultrasonics Ferenc Divos
Contents 17.1 17.2
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Crack Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.2.1 Crack Between Lamellae . . . . . . . . . . . . . . . . . . . . 17.2.2 Crack Inside the Lamella Material . . . . . . . . . . . . . . . . 17.3 Crack Depth Detetrmination with Ultrasonics . . . . . . . . . . . . . . 17.4 Strength Prediction of Lamellas In situ . . . . . . . . . . . . . . . . . 17.5 Detection of Other Internal Defects in Glulam . . . . . . . . . . . . . . 17.6 Shear Strength Determination Between the Lamellae . . . . . . . . . . . 17.7 Delamination and Other Defects in Structural Element of Historic Buildings 17.8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
353 355 355 355 356 358 358 358 360 363 364
17.1 Introduction The importance of nondestructive techniques for assessing the internal conditions of structural members is generally accepted for old or more recent buildings (Ross and Pellerin 1994; Pellerin and Ross 2002, Tanaka et al. 1998; Brashaw et al. 2005; Wacker et al. 2007). Efforts to preserve buildings that have nationally significant importance have been based on the development of nondestructive techniques able to assess their mechanical performances (Wang X and Wacker 2006; Sandoz and Benoit 2007; Lee et al. 2007; Divos et al. 2007). The overall goal of these efforts is to preserve to the maximum possible extent of the historical materials used for the monuments. On the other hand, in recent years, the stability of existing glulam structures became an important issue after the collapse of some of the early structures. For example, in Germany on Jan.2, 2006, 15 people, most of them children
F. Divos (B) Faculty of Wood Science, University of West Hungary, Sopron, Hungary e-mail:
[email protected];
[email protected] V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_17,
353
354
F. Divos
Fig. 17.1 Crack depth determination with 0.1 mm thick feller gauge
on a school holiday, died when the roof of a skating rink fell in the Bavarian town of Bad Reichenhall, after heavy wet snow falls (Associated Press 2006). Since the tests should be performed in the field, simple devices must be used and the experimental processes modified to improve the workability and efficiency of tests. A preliminary inspection phase must be followed by an advance phase of inspection. In the preliminary phase, a comprehensive in situ assessment of the structure in its current condition must be made via a visual inspection. This visual inspection defines damage severity and extends of deterioration, having in mind that the engineering challenge is to assess member integrity. The first symptoms that can be detected are surface cracks. (The biological degradations are not discussed in this chapter). In the advance phase, the methods and procedures to be used for inspection are defined, and the experimental results analyzed. This phase must be ended with practical recommendations. One of the most frequent observed defects are the cracks. The primary characteristic of a crack is the depth. The common test for crack depth determination is the penetration depth of 0.1 mm thick feller gauge (Fig. 17.1). This simple technique provides useful information about the crack, but has its limitation, especially when the crack path is not straight (Fig. 17.2). Due to the above limitation in testing,
Fig. 17.2 Crack depth determination by feller gauge penetration is limited by the crack path, which is not straight
17
Delamination Evaluation of in-Service Glulam Beams
355
ultrasonic technique can be an alternative when applied to crack depth determination. Typically the depth measured by ultrasonic technique is a bit higher, than feller gauge plate penetration depth.
17.2 Crack Types In glulam structures, two main crack types are recognised, the crack between the lamellae and the cracks inside the lamellae material.
17.2.1 Crack Between Lamellae Adhesive failure or other technological failures cause partial separation between lamellas. It is to note that after relatively few years of service (ex 6 years), the glulam beams exposed to sun and rain suffer from delamination, as shown in Fig. 17.3. Fig. 17.3 Cracks in glulam beam in service for 6 years. Crack depth is 24 mm
Figure 17.4 shows a typical example of delamination in a glulam structure, long time in service, on Robinia pseudoacacia. This sample has been taken from a demolished structure after 34 years of service. The structure has been demolished due to severe delamination and shape change.
17.2.2 Crack Inside the Lamella Material In the past, glulam elements containing pith have often been used by some manufacturers, typically at the central zone of the beam on which the mechanical loading corresponds to the neutral axis. However an extreme case of utilization of a lamellae is shown in Fig. 17.5a. Where the outer laminates are pith – containing. Cracks are
356
F. Divos
Fig. 17.4 Severe delamination in a glulam beam made of Robinia pseudoacacia after 34 years of service. Lamella thickness is 20 mm
Fig. 17.5 Cracks in lamellae with piths zones. (a) two lamellae with pith in transversal section; (b) cracks due to the pith at beam surface
also clearly visible on the surface (Fig. 17.5b). Note that the crack propagates to the pith. The beam ends are particularly vulnerable to initiation and growth of delamination because the dowels or screws of the fastening can be a crack initiator. A crack having such origin is shown in Figs. 17.5 and 17.6.
17.3 Crack Depth Detetrmination with Ultrasonics Cracks form a material discontinuity and as such are a vibration propagation barrier. Ultrasonic or stress wave tools are applicable for the crack depth determination.
17
Delamination Evaluation of in-Service Glulam Beams
357
Fig. 17.6 Cracks at the glulam beam end due to the fastening screws
Placing a vibration source close to the crack line and a receiver sensor on the opposite side of the crack, the transit time between the emitter and receiver give us information about the crack depth. The recommended distance between sensors is 2–5 cm. To compute the crack depth it is also necessary to know the transit time in the intact material (see: material without crack). The following equation gives us the crack depth (L) L = 0.5∗ sqrt(V 2 (Tc − Ti )2 + 2V(Tc − Ti )D)
(17.1)
where: Tc transit time across crack, Ti transit time in case of intact material, D Distance between sensors, V P-wave velocity perpendicular to fibers in intact material. Figure 17.7 shows two examples for the measurement of the time of flight of the ultrasonic pulse. The first one uses ultrasonic surface waves. In this case a pair of ultrasonic probes (1 MHz) is displayed at 45◦ and pressed to the surface
a)
b)
Fig. 17.7 Crack depth determination via acoustic technology. (a) with ultrasonic transducers at 45◦ (b) with stress wave sensors
358
F. Divos
of the beam. The second example shows the probes used for stress wave technique (40 kHz), which uses a hammer and an amplification system to excite vibration into the tested element. It is easy to understand that with ultrasonic techniques, which are more precise, the value corresponding to crack depth measurement is greater than that determined using feeler gauge penetration. The ultrasonic techniques work very well for normal cracks. Including all error sources the relative error of crack depth determination is around 10%. The ultrasonic techniques are strongly limited if the crack is filled by paint.
17.4 Strength Prediction of Lamellas In situ Re-evaluation after many years of service of an existing glulam structure needs strength and stiffness data. In the case of old beams, in function for more than 30 years, often no strength grading was applied to produce the beams. Hence wave velocity determination in the fiber direction can provides the lamella modulus of elasticity (MOE) according to the following equation: MOE = ρV 2
(17.2)
where: ρ density V P-wave velocity In-situ determination of density is however rather difficult but possible using gamma or X rays. In practice the density values cited in the literature, for given species is suitable for computation of MOE. The velocities can be measured relatively easily and it is recommended to limit the distance between the two probes at 20 times the lamella thickness. Figure 17.8 shows the P-wave velocity measurement using an ultrasonic device. A plastic – PVC – bar connects the sensors, helping to keep the distance constant between during measurements.
17.5 Detection of Other Internal Defects in Glulam Detecting internal decay, holes or other defects in glulam structures is possible by measuring the transit time of P or S-waves between the two faces of the beam. This type of defect was simulated on the sample used for this experiment. The test set up using P waves is shown in Fig. 17.9. The defective zone showing a longer transit time is located clearely by the red color on the graph.
17.6 Shear Strength Determination Between the Lamellae The shear strength between the lamellae provides critical information about the glulam beam mechanical capacities. Predicting the shear strength of the glue layer via nondestructive techniques is possible but quite difficult as can be seen in Fig. 17.10.
17
Delamination Evaluation of in-Service Glulam Beams
359
Fig. 17.8 Velocity determination in a lamella along the fibers, using a special device for keeping a constant distance between the transducers
a)
b)
Fig. 17.9 P-wave velocity determination perpendicular to the grain. (a) the device (b) defect location on the time of flight map. Dark (red) spot shows the location of the defective region
Shear sensors equipped with a knife type wave guide are applied to the opposite faces of adjacent elements The plane of the knife determines the shearing plane. This plane is prallel to the glue layer. In this experimental situation there were no glue in between the lamellae. It was suggested that (Fig. 17.11) the attenuation measurements via shear wave amplitude can be a potential tool for the
360
a)
F. Divos
b)
Fig. 17.10 Shear transducers. (a) The position of shear transducers on the transversal section of the laminar structure; (b) the probes
Fig. 17.11 Relationship between the measured shear strength and signal amplitude
evaluation of delamination between glued layers in beams. More research and technical development is necessary for in situ applications.
17.7 Delamination and Other Defects in Structural Element of Historic Buildings Hungary abounds in historical buildings, several of which have wooden roof and ceiling structures. Most of these buildings are in a run-down state, needing renovation. Wood experts evaluate the bio-degradation of wood, identifying fungi and insect attack by visual inspection, by touching the material or using a simple screw driver. In this chapter the effectiveness of the ultrasonic technique, allowing the time of flight measurements is demonstrated on structural elements and on the ceiling structure of an old baroque castle – Esterhazy Castle – in Pápa in Hungary (Fig. 17.12).
17
a)
Delamination Evaluation of in-Service Glulam Beams
361
b)
Fig. 17.12 Esterhazy Casel in Papa, Hungary. (a) genral view of the restored casel, (b) plane view of the first flor
The first floor of the two storied, U-shaped building is more than 500 years old, and had been used as a fortress until 1752. From the original structure, the builders have kept the walls of the ground floor. The length of the building is 165 m. The roof and the top ceiling structure (a system of dowelled, closely aligned beams) is made on wood (larch, oak, spruce and lime). The first floor and the walls of the second floor are made on stone and bricks, and have been built probably during the 18th century. The ceiling structure of the first flor is different from that of thesecond one, the first being vaulted, and the second composed of closely fitted, doweled beams. The inspection of the structure started with the visual examination of the sound and decayed zones of all elements. The second phase of the inspection with ultrasonic device and micro drilling device are commented in this chapter and are related to the physical and mechanical tests of the Sections 16 and 38. Figure 17.13 shows some aspects during in situ measurements For the ultrasonic test, the piezoelectic transducers of the device-FAKOPP timer, have been equipped with 60 mm long nails to facilitate the inspections of wooden beams. Using a long cable of 11 m lenhght, all beams have been measured without difficulty. The test for velocity measurements is fast, two experienced person can carry out the test within 30 s. To obtain more data about the physical state of the inner layers of the beams, micro drilling technique was used. The screw withdrawal force measurements (Fscrew ) were performed for 5 mm diameter drill to 120 mm depth. The consistency and the odor of wood particles falling out from the hole were also analyzed to state about the wood quality of the inspected beams. For the inspected beams, the strength predictor parameters are stress wave velocity V and screw withdrawal resistance. The predictor coefficient was calculated as Fscrew · V2 . It was demonstrated previously (Divos and Tanaka 1997, Divos et al. 1998, 1999) that screw withdrawal resistance is well correlated shear modulus, withdrawal force and with density. This suggested to use the empirical relationship σ = FCS v2 (similar to E = ρv2 ). Moreover, Kollmann (1965) noted a strong relationship between modulus of elasticity and modulus of rupture in bending of full size beams. Using the screw withdrawal force and the velocity of stress wave the following empirical strength predictor equation applied for coniferous wood species was derived. The
362
F. Divos
a)
b)
Fig. 17.13 In situ measurements
applied units in the equations are: MORest [MPa], Fscrew [kN] and v, the velocity [km/s]: MORest = 0.809Fscrew · v2 + 26.8 A similar MOR predictor formula applies for hardwoods: MORest = 1.258Fscrew · v2 + 36.9 The correlation coefficients between the bending strength and MORest is 0.74. Figure 17.14 shows the correlation between modulus of rupture in bending of fill
Fig. 17.14 Relationship between the modulus of rupture and the predictor parameters. represents coniferous, ∇ represents hardwood specimens
MOR [Mpa]
140 120 100 80 60 40 40
60
80 100 Predictor [MPa]
120
140
17
Delamination Evaluation of in-Service Glulam Beams
363
Fig. 17.15 Strength distribution over the 214 years old ceiling in larch – a system of doweled, closely aligned beams. The numbers represent the residual stress bending strength in MPa
size specimens and the predictor parameters.. The residual bending strength of individual wooden beams have been estimated with ± 9 MPa accuracy. Figure 17.15 shows the strength distribution in the larch ceiling.
17.8 Summary In this chapter are discussed several procedures for delamination evaluation of in-service glulam beams and other structural members via ultrasonics Efforts to preserve buildings that have nationally significant importance have been based on the development of nondestructive techniques able to assess their mechanical performances. The stability of existing glulam structures became an important issue after the collapse of some of the early structures. One of the most frequent observed defects are the cracks. The primary characteristic of a crack is the depth. The common test for crack depth determination is the penetration depth of 0.1 mm thick feller gauge. Ultrasonic technique can be an alternative when applied to crack depth determination. Typically the depth measured by ultrasonic technique is a bit higher, than feller gauge plate penetration depth. In glulam structures, two main crack types are recognized, the crack between the lamellae and the cracks inside the lamellae material. Cracks form a material discontinuity and as such are a vibration propagation barrier. Re-evaluation after many years of service of an existing glulam structure needs strength and stiffness data. In the case of old beams, in function for more than 30 years, often no strength grading was applied to produce those beams. Wave velocity determination with P waves, in the fiber direction can provides the modulus of elasticity of the lamellae. The shear strength between the lamellae provides critical information about the glulam beam mechanical capacities. Predicting
364
F. Divos
the shear strength of the glue layer via nondestructive techniques is possible using shear probes equipped with a knife type wave guide. The plane of the knife determines the shearing plane. This plane is parallel to the glue layer. It was suggested that the attenuation measurements via shear wave amplitude can be a potential tool for the evaluation of delamination between glued layers for in –situ beams.
References Associated Press (2006) German ice rink collapse. AP Image January 3, 2006 by Diether Endlicher, www.highbeam.com. Accessed 3 August 2010 Brashaw BK, Vatalaro RJ, Wacker JP, Ross RJ (2005) Condition assessment of timber bridges. 1. Evaluation of a micro-driling resitance tool. FPL – GTR 159 USDA Forest Service Forest Products Laboratory. Madison WI, USA Divos F, Divos P, Divos G (2007) Acoustic techniques: from seedling to wood structures. Proceedings of the 15th international symposium on nondestructuctive testing of wood. Duluth, MN, pp 3–12 Divos F, Nemeth L, Bejo L (1999) Evaluation of the wooden structure of a baroque place in Papa, Hungary. Proceedings of the 11th international symposium on nondestructive testing of wood. Lausanne, Suisse, pp 153–160 Divos F, Tanaka T (1997) Lumber strength estimation by multiple regression, Holzforshung 51:467–471 Divos F, Tanaka T, Nagao H, Kato H (1998) Determination of shear modulus on construction size timber. Wood Sci Technol 32:393–402 Kollmann, F. 1965 Relationship between elasticity and bending strength of wood, Proceedings of the 2nd symposium on nondestructive testing of wood. Spokane, WA Lee SJ, Oh JK, Yeo H, Lee JJ, Kim KB, Kim KM (2007) Field application on nondestructive testing for detecting deterioration in Korean historic wood buildings. Proceedings of the 15th international symposium on nondestructuctive testing of wood. Duluth, MN, pp 227–232 Pellerin RF, Ross RJ (2002) Nondestructive evaluation of wood. Forest Products Society, Madison, WI Ross RJ, Pellerin RF (1994) Nondestructive testing of assessing wood members in structures, USDA, Forest Products Laboratory, FPL-GTR-70, Madison, WI Sandoz JL, Benoit Y (2007) Acousto-ultrasonic nondestructive evaluation of historic wood structures. Proceedings of the 15th international symposium nondestructuctive testing of wood. Duluth, MN, p 245 Tanaka T, Divos F, Fazan, T (1998) Nondestructive evaluation of residual bending strength of wood with artificial defects by stress wave, Proceedings of the 11th international symposium on nondestructive testing of wood, Madison, WI Wacker JP, Wang X, Ross RJ, Brashaw BK (2007) Condition assessment of historic vessels. Proceedings of the 15th international symposium nondestructuctive testing of wood. Duluth, MN, pp 223–226 Wang X, Wacker JP (2006) Condition assessment of main structural members of US Brig Niagara. Final Report Project no 187–2419. Erie Maritime Museum, Erie, PA
Chapter 18
Moisture Induced Stresses and Deformations in Parquet Floors Samuel Blumer, Erick Serrano, Per Johan Gustafsson, and Peter Niemz
Contents 18.1 18.2
Introduction . . . . . . . . . . . . . . . . . . . . . . . . Material and Methods . . . . . . . . . . . . . . . . . . . 18.2.1 Tests on the Basic Material . . . . . . . . . . . . . 18.2.2 Test on Parquet Planks . . . . . . . . . . . . . . . 18.2.3 Analytical Model A: Calibration Model . . . . . . . 18.2.4 Analytical Model B: Distortional Effects . . . . . . . 18.2.5 Analytical Model C: Gap Opening . . . . . . . . . . 18.3 Deformations in Parquet Floors . . . . . . . . . . . . . . . 18.3.1 Model A: Calibration and Comparison to Test Series 2 18.3.2 Model B: Cupping of the Parquet . . . . . . . . . . 18.3.3 Model C: Stresses in the Glue Line and Gap Opening . 18.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . 18.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
365 366 366 367 368 370 370 371 371 373 374 376 377 378
18.1 Introduction During the last decade the use of wood flooring systems in Europe has increased dramatically. In Sweden for example, the proportion of wood flooring systems rose steadily from 30% in the seventies to its current proportion which is 80%. This rapid growth has fostered the development of new products, enabling the industry to maintain and increase its market share. The main objective of this paper is to improve understanding of the behaviour of parquet floors exposed to different climates by applying numerical analysis techniques using the commercial finite element program ABAQUS. In addition, S. Blumer (B) b-h-e GmbH, Holzinnovationszentrum 1a, 8740 Zeltweg, Austria e-mail:
[email protected] V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_18,
365
366
S. Blumer et al.
Bernoulli’s beam theory applied in two dimensions to the square specimens gives a simple and additional validation instrument for determining the deformational behaviour of parquet flooring (Bodig and Jayne 1982). This approach is complemented by a parametric study of the long time behaviour of parquet planks that emphasises the influence of creeping on the aforesaid deformations and failure modes.
18.2 Material and Methods Parquet floor product from Sweden was tested. The parquet specimen as seen in Fig. 18.1, has three main layers: the surface layer – denoted SL – is 3.6 mm thick, the core layer – denoted CL – is 8.6 mm thick and backing layer – denoted BL – is 2 mm thick. The layers are glued together crosswise with urea formaldehyde resin. The geometry of the specimen, a parquet plank, is a plate of size 188 × 14.2 × 2500 mm. 188 mm 3.6 mm
Half element = 94 mm
X Y
click joint
local coordinate system
2 mm
14.2 mm
Z
Surface layer: Oak
(Quercus robur L.)
Core layer: Pine
(Pinus sylvestris L.)
8.6 mm
Half element = 94 mm
Backing layer: Veneer pine
Fig. 18.1 Geometry and consistency of the parquet floor
Laboratory tests on the materials which compose the parquet planks (referred to as basic material) were performed to determine mechanical characterization and for providing data for calibration and validation of the finite element method calculations.
18.2.1 Tests on the Basic Material Wood species used in these experiments are: pine, oak, beech and ash. The adsorption behaviour, density, static modulus of elasticity in the longitudinal direction and the hygroexpansion factors of pine (Pinus sylvestris L.) and oak (Querqus robur L.) have been determined. Data for beech and ash were obtained from the literature. The static modulus of elasticity in the longitudinal direction was determined using a three point-bending test. The measured values were in the same range and showed the same variation as noted in the literature (Kollmann 1982, Wood Handbook 1999). For determining the adsorption behaviour, 20 samples from each species were conditioned in a climatic chamber at 20◦ C and 25% relative humidity (RH) until
18
Moisture Induced Stresses and Deformations in Parquet Floors
367
they reached equilibrium moisture content (EMC). Thereafter, the climate was changed from 25% to over 50% and then up to 85% relative humidity, all at a temperature of 20◦ C. The shrinkage and swelling coefficients in the longitudinal, radial and tangential directions were measured by changing the climate from 60% RH down to 25% RH (shrinking, 20 specimens) and then/from 60% RH up to 85% RH (swelling, 20 specimens) respectively, all at a temperature of 20◦ C.
18.2.2 Test on Parquet Planks The influences of different parameters such as material properties, material orientation, properties of the glue line and the geometry of the product on the stresses and deformations were tested. 30 square samples with a side length of 150 mm were cut out of parquet planks and conditioned in a standard climate of 20◦ C/65% relative humidity until reaching equilibrium moisture content. Ten specimens had a lacquered surface layer (denoted test series 2, A1–A10) and ten specimens were non-lacquered (denotes test series 1, and A11–A20). In addition, the transport of moisture in one dimension only was enforced on 5 lacquered and 5 non-lacquered specimens by applying moisture insulation on the edges. The samples were dried in a climate of 20◦ C for a period of 28 days. The bending deformation of parquet specimens was measured at four points on the surface layer of the specimens (Fig. 18.2). The vertical deformation was measured along two orthogonal directions of the plate. In position A the grain direction of the core layer and thus x axis was parallel to the primary axis of the global
Position A H
G D A
2 (Z) 1 (X)
3 (Y)
I
E C
B
measurement points for horizontal deformation
measurement points for horizontal deformation
F
direction of x axis
Position B G' = I
A' = G
2 (Z)
I' = C
H'
F' = B
E' = E
D' = H B' = D
C' = A direction of x axis
3 (X)
1 (Y)
Fig. 18.2 Measurement of the plate’s vertical deformation in two directions and evaluation with the beam theory of Bernoulli (t: thickness change, κ: curvature in x and y direction respectively)
368
S. Blumer et al.
coordinate system, thereafter 90◦ rotation counter clockwise of the plate was done for the second measurements. The plate was supported at the downside of point B, G and I., B’, G’ and I’ respectively. Three variables (thickness of the plate, curvature in the x and y directions respectively), describing the vertical deformation of the specimen, have been evaluated from the eight data measurements using a least square fit and Bernoulli beam theory. Several finite element models were created to simulate the behaviour of the parquet planks under different climatic conditions, such as a calibration model, a model for the study of distortional effects and a model for gap opening. These models are described in following sections. The influence of different parameters such as the properties of the material, material orientation, glue line properties and the geometry of the product on the stress and deformations will be tested using these models.
18.2.3 Analytical Model A: Calibration Model The calibration model has been constructed for calibration and verification of the numerical analysis applied to specimens with moisture-isolated edges. Some assumptions have been made in estimating the effective coefficient of diffusion, These are: • The diffusion coefficient of pine and oak wood in the radial and tangential directions are equal; • The estimated diffusion coefficient is assumed constant below 15% MC, (Jönsson 2005) • The model does not explicitly consider any interfacial layer between the wood layers and the glue layer. The interfacial layers have been reduced to a continuative 0.l mm thick composite layer. The glue lines between the backing and core and core and surface layer respectively, were modelled as a 0.1 mm thick layer (UF resin) with material data taken from Hagstrand (1999). The relationships between the diffusion coefficient, the specific heat, the density and the thermal conductivity (Carslaw and Jaeger (1959) and Eriksson (2005) are given in the Eqs. (18.1) and (18.2)) ∂ λ(T) ∂T ∂T = ∂t ∂x cρ ∂x ∂u ∂ ∂u = DW (u) ∂t ∂x ∂x Where the parameters are expressed in the following units: – diffusion coefficient [m2 s−1 ] – temperature T – specific heat c [Jkg−1 K −1 ]
(18.1) (18.2)
18
Moisture Induced Stresses and Deformations in Parquet Floors
369
– density ρ [kgm−3 ] – thermal conductivity λ [Js−1 m−1 K −1 ] – time t [s] Thermally coupled and quadratic interpolated brick elements have been used for the calculation. To simplify the transport model, the moisture content at the surface was set in equilibrium with the moisture content corresponding to the relative humidity of circulating air, μsurf = μair . This imposed boundary condition is called the boundary condition of Drichlet (Koc and Houska 2002). An effective coefficient of diffusion for the entire parquet plank was determined and compared to experiments on specimens with moisture-isolated edges (test series 2). The bending deformations of the plate in plane xz and yz have been calculated and compared to experiments on test series 2. The static boundary conditions were included consistent with the test set up of test series 2. The degrees of freedom u1 , u2 and u3 were restrained on the lower edge of point B and in the vertical direction u2 on the lower edge of point G and I respectively, see Fig. 18.4. The gaps between the pine strips in the core layer were modelled with ABAQUS. The layer consisted of three strips (left, middle and right) each with different direction of the growth rings (longitudinal L, radial R and tangential T). The material orientation of the surface layer has been varied from 0◦ (Tangential direction parallel to u1 or x direction according to Fig. 18.3) to 90◦ (Tangential direction perpendicular u1 or x direction). The angle of the growth rings has been set similar to the test specimens of test series 2. Transforming the stiffness matrix of the oak layer’s different strips simulated the influence of the growth ring’s direction. The transformation was done for both the stiffness matrix and the hygroexpansion factors. A coordinate
150
R (variable) surface layer L
425
65
425
SL: left strip
SL: middle strip
SL: right strip
T (variable) H
G D A
2 (Z) 3 (Y)
I
E
F
B
C
1 (X) core layer
T L R
R backing layer
T L
Fig. 18.3 Geometry and material of the parquet plank. The angle of the growth rings in the surface layer differs between the strips
370
S. Blumer et al.
transformation of the orthogonal coordinate system around the longitudinal axis (L) has been performed.
18.2.4 Analytical Model B: Distortional Effects A finite element model (named the analytical model B and shown in Fig. 18.4) was applied to predict the distortional behaviour of the parquet plank’s central component. Parametric study on the influence of geometry, material and creeping of the surface layer were performed. The model corresponded to half the width of the strip and a depth of 26 mm. This depth included 2 half width core sticks (2×12.5 mm) and 1 mm spacing between the sticks. The depth of the model was relatively small compared to the length of the parquet planks, which is 2500 mm. The parquet strip was 10 mm wide, 14.4 mm thick in a three-layer structure (3.6 mm surface layer (SL), 8.6 mm core layer (CL) and 2 mm backing layer (BL)) glued together with two 0.1 mm thick UF resin layers. The vertical deformation of the parquet planks has been calculated between point A and point B. These points were located on nodes, point B on the boundary edge whereas point A was located 10 mm from the boundary to minimize the local deformation shape of the unconstrained face in plane yz at x = 0. The surface in plane yz at the value x = 94 mm was constraint in u1 or x direction according to the coordinate system shown in Fig. 18.2. The surface could not be blocked in u2 or z direction in order to allow free movement of the surface layer in the vertical direction. The edge below was also constraint in u2 or z direction for stability reasons. The surfaces in plane xz were constraint in u3 or y direction. This boundary condition simulated an infinite depth of the parquet plank. Coupled temperature-displacement and quadratic interpolated elements have been chosen for the model.
A (0/14.2/13)
Moisture exchange limited to top surface
B (94/14.2/13) u1 = 0
2 (z) u3 = 0
125 + 1 + 125
1(x) 94
3 (y)
u2 = 0
A v
B
Fig. 18.4 Analytical model A: Geometry, static system and boundary conditions used for the modelling of the distortional effects
18.2.5 Analytical Model C: Gap Opening The geometry, the static system and the boundary conditions of the model C are shown in Fig. 18.5. The proposed model was applied to predict the deformation
18
Moisture Induced Stresses and Deformations in Parquet Floors
371
2 (Z) 94.000
94.000
2 8.6 3.6
1 (X)
A
B B.3
A.3
uA.1
uB.1
Fig. 18.5 Geometry, static system and boundary conditions of model C
behaviour of the click joint and the gaps of the surface layer. The model corresponded to half the width of the parquet plank on the right and left side of the click joint. Because stresses mainly occurred in xz, the model was reduced to two dimensions. The side edges of the model were coupled to the reference points A and B in u1 or x direction. The symmetrical behaviour has been introduced to the model by constraint equations. The rotation of the edge at point A is the same as that of the corresponding surface at point B (ϕA.3 = ϕB.3 ). The horizontal deformation in u1 or x direction had similar values but opposite signs (uA.1 = −u1.B ). The model was based on an elastic layer with a very small E-modulus, to give the stability in u2 or z direction. This was done in order to simplify the model, such that no algorithm for modelling the contact with the foundation had to be used. The contact of the model in the click joint was modelled by a contact algorithm triggering reaction forces in the case where the elements of the tongue come in contact with the element of the groove in the joint region. Seams have been introduced for simulating the gaps in the surface layer. Coupled temperature-displacement and quadratic interpolated triangular and quadratic elements were chosen for this model.
18.3 Deformations in Parquet Floors In the following the moisture induced stresses and deformation in parquet floors determined with the models A, B and C will be discussed.
18.3.1 Model A: Calibration and Comparison to Test Series 2 The estimation of the moisture transport using different effective diffusion coefficients resulted in Deff = 1.8e−11 [m2 s−1 ]. The value obtained was about 1.5−2 times smaller than that given by Simpson (1993). This difference may be caused by the influence of the two glue lines in the parquet element that acts as a moisture barrier.
372
S. Blumer et al.
L Z Y
L
X v
L L
w
Fig. 18.6 Bending v in xz plane
The specimens without edge isolation (denoted A11–A20 for non lacquered specimens) reached equilibrium moisture content in a climate of 20◦ C/25% RH after 28 days conditioning. Thus, the numerical analyses were performed as steady state calculations. The comparison of the vertical deformation v (Fig. 18.6) between the test specimen and numerical analysis according are shown in Figs. 18.7 and18.8 for lacquered specimens A1 . . .. A10, and for non lacquered specimens were denoted A11. . ..A20. The bending of the plank in xz plane was strongly dominating (v >> w). Figure 18.9 gives the comparison between the experimental and the numerical results with the Analytical model A, for lacquered and non lacquered
0
Fig. 18.7 Test results versus numerical FEM calculations, moisture content change from 10.25 to 6.85%. Bending deformation of lacquered specimens (A1–A10 are the specifications of the samples of test series 2)
Bending v in xz plane [mm]
–0.05
Numerical simulation Measurements
–0.1 –0.15 –0.2 –0.25 –0.3
Mean values
–0.35 A1
A2 A3 A4 A5 A6 A7 A8 A9 A10 Specification of lacquered specimens
0
Fig. 18.8 Test results versus numerical FEM calculations, moisture content change from 10.25 to 6.85%. Bending deformation of non lacquered specimens (A11–A12 are the specifications of the samples of test series 2)
Bending v in xz plane [mm]
–0.05
Numerical simulation Measurements
–0.1 –0.15 –0.2 –0.25 –0.3 Mean values –0.35 A11 A12 A13 A14 A15 A16 A17 A18 A19 A20 Specification of non lacquered specimens
Moisture Induced Stresses and Deformations in Parquet Floors
Fig. 18.9 Overall comparison of Analytical model A: Test results versus numerical FEM calculations, moisture content change from 10.25 to 6.85%
373
0 Bending v, measurements [mm]
18
Laquered specimens Non lacquered specimens
–0.05 –0.1 –0.15 –0.2 –0.25 –0.3 –0.35
–0.4 –0.4 –0.35 –0.3 –0.25 –0.2 –0.15 –0.1 –0.05 Bending v, numerical analyse [mm]
0
specimens, for moisture content decreasing between 10.25 and 6.85%. Further comments regarding this model are: – the introduction of different angles of the growth rings for each of the tree strips of the surface layer shown in Fig. 18.3 had an important influence on the plate deformation behaviour; – the surface treatment (lacquered or non lacquered) did only slightly influence the bending behaviour of the square samples.
18.3.2 Model B: Cupping of the Parquet The cupping effect under different drying conditions using different materials, angle of the growth rings and geometry has been calculated. At the start of the calculation, the boundary condition of the surface layer was determined for moisture content decreasing from 7.5 down to 5% and for different angles of growth rings (Fig. 18.10). The cupping minimum occurs under 45◦ in both strips of the surface layer. The influences of the surface thickness on the cupping of the parquet are
0.8 Angle = 30 deg Angle = 45 deg Angle = 90 deg Angle = 0 deg
0.7
Fig. 18.10 Influence of the surface layer’s angle of growth rings after reduction of the moisture content from 7.5 down to 5%. (Local tangential direction parallel to horizontal plane at α = 0◦ )
Cupping v [mm]
0.6 0.5 0.4 0.3 Beech Beech SL 3.6mm Angle ==Beech 30deg Angle 30deg Angle =SL 45 deg Angle Oak Oak =Oak 45 3.6mm deg Angle =SL 90 deg Angle =Ash 03.6mm deg Ash Ash Angle deg Angle==090 deg
0.2 0.1 0
0
20
40
60 80 Time [days]
Beech Oak Ash
100
120
374 0.8 0.7 0.6
Cupping v [mm]
Fig. 18.11 Influence of the surface layer’s thickness after reduction of the moisture content from 7.5 down to 5%
S. Blumer et al.
0.5 0.4 0.3 4.8 mm Surfacelayer layer2.6mm 2.6mm Surface 3.6 mm Surfacelayer layer3.6mm 3.6mm Surface 2.6 mm Surfacelayer layer4.8mm 4.8mm Surface
0.2 0.1 0
20
40
60 80 Time [days]
100
120
100
120
0.8 0.7 0.6
Cupping v [mm]
Fig. 18.12 Influence of the surface layer’s material after reduction of the moisture content from 7.5 down to 5%
0
0.5 0.4 0.3 Beech SL 3.6mm Beech Oak Oak SL 3.6mm Ash SL Ash3.6mm
0.2 0.1 0
0
20
40
60 80 Time [days]
shown in Fig. 18.11, for moisture content decreasing from 7.5 to 5%. The influences of the surface layer’s material – beech, oak and ash – on the cupping of the parquet are shown in Fig. 18.12, for the same moisture content decreasing range. The significant influence of the geometry (layers thickness) and of the species can be observed. Beech had the highest cupping , a maximum of 0.78 mm after 20 days while ash had the lowest cupping , 0.44 mm after 20 days.
18.3.3 Model C: Stresses in the Glue Line and Gap Opening Figure 18.13 shows the model and the cuts for stress calculation for the model C where S11 is the horizontal stress, S22 the vertical stress and S12 the shear stress. The influence of the materials used for the surface and core on the vertical stress as a function of the horizontal distance to the middle of the gag is shown in Fig. 18.14. Figure 18.15 shows the variation of the horizontal stress through the cut of the parquet plank for different geometries of the surface layer and core. It is notable that the absolute values of the stresses are extremely mesh-size dependent. Maximum
18
Moisture Induced Stresses and Deformations in Parquet Floors
375
Fig. 18.13 Model and cuts for stress calculation. S11 : Horizontal stress, S22 Vertical stress and S12 Shear stress
values of the stresses cannot be evaluated in this model. For this reason the curves have to be compared on the basis of their gradients. The main target of the parametric study was to minimize the gradient of the vertical, horizontal and shear stresses. The vertical stress S22 and the shear stresses S12 in the glue line can lead to delamination. A steeper curve close to the gap indicates an increased risk for crack formation and propagation of delamination. Here, the highest gradient of vertical stresses can be observed for the beech wood. The creeping is stress depending as demonstrated by Jönsson (2005) and Hanhijärvi (1995). Thus, higher horizontal stresses can lead to higher creeping effect in the wood. A stress gradient in the surface layer may result in stronger creeping of the surface layer at the bottom as on the top. The effect of a creeping gradient in the surface layer has been modelled and the variation of the vertical stresses S22 after creeping gradient is shown in Fig. 18.16. The minimum is observed at about 6 mm distance to the gap.
Fig. 18.14 Vertical stresses S22 in CUT A: Different materials in the surface and core layer (HDF) respectively
376
S. Blumer et al.
Cut through the parquet plank [mm]
15 SL 4.8 mm/CL 8.6 mm SL 3.6 mm/CL 8.6 mm SL 2.6 mm/CL 8.6 mm
Surface layer (SL)
10
Groove Core layer (CL)
5
Backing layer (BL) 0
–30
–20
–10 0 10 20 Horizontal stresses S11 [MPa]
30
Fig. 18.15 Horizontal stresses S11 in CUT B Influence of the geometry of the surface layer
3
φR,T = 0.8 φR,T = 1.6 φR,T = 3 hoak hpine
Vertical stresses S22 [MPa]
2.5 2 1.5
ΦR,T = 0
1
ΦR,T = 0.8..1.6..3 (gradient)
0.5 0 –0.5 –1 –1.5 –2
0
5 10 15 Horizontal distance to the gap [mm]
20
Fig. 18.16 Influence of creeping gradient in the surface layer
18.4 Conclusion As concluding remarks it is noted that a model that includes the whole parquet system helps find optimal solutions as a function of stresses in the glue line and gap opening of the surface layer. In the previous discussion, it was demonstrated that the finite element method brings several advantages compared to traditional testing in laboratory conditions. The time needed for simulating changing climatic cycles is much smaller compared to laboratory tests. In the future, the design process for
18
Moisture Induced Stresses and Deformations in Parquet Floors
377
wood flooring systems should include basic material testing, finite element analyses and, finally, testing of the developed product. Other advantages with modelling techniques include the possibility to optimize the geometry of the joint and the lay-up of the planks in a rather straightforward manner. The material and the angle of growth ring in the surface layer have a considerable influence on the deformation and stress distribution of the parquet planks. An angle of 45◦ (between tangential direction and horizontal plane) in the surface layer minimized the cupping deformation. From the design perspective, results based on calculations with elastic properties of the glue line without introducing creeping factors are conservative; bigger deformations than experienced in practice are predicted. The material properties of the glue line and lacquer are difficult to determine, although the finite element method can be used for parameter estimation. The long time behaviour of the glue line did not significantly influence the deformation and stress distribution. It seems to be a good approach in terms of modelling to assign the UF resin layer properties making it less hygroscopic than wood and acting as a linear elastic layer. A hygroscopic material model may make more sense for the wood material than for the glue line. Periodic loading can increases the creeping effect, delamination may also occur after several summer – winter cycles.
18.5 Summary The indoor climate in buildings has changed in the last decade due to more efficient climatic systems, floor heating systems and larger open floor areas with more natural light. All these factors have induced increasing ranges of relative humidity between different seasons. Also with decreasing relative humidity (in the winter 30–50% RH, in the summer 70–90% RH), floor-heating systems increase the temperature in wooden parquet planks for example. Such variations can result in troublesome deformations, delamination of the surface layer and development of cracks in the parquet flooring boards. Sometimes there is only deterioration of the appearance but the durability of the flooring system can also be reduced. Many laboratory tests have to be done before reaching an optimal design of the parquet elements. Due to the high costs and time constraints of experiments, other supplementary research methods should be tested and evaluated. The articles’ main objective was to improve understanding of the behaviour of parquet floors exposed to different climatic conditions by using numerical calculation. The use of the finite element models provides options for design purposes of wood flooring systems. Several finite element models to aid adequate design have been created, tested and applied. After calibration and validation of the calculation method, parametric studies on the influence of material properties, geometry of the parquet floors and the long-term behaviour of the wood and glue line were performed. The results show a strong relation between material and geometry choice on the deformation, for example the gap opening and the stress distribution in the glue line, which can induce delamination of the surface layer and distortional effects of the parquet boards.
378
S. Blumer et al.
References Bodig J, Jane A (1982) Mechanics of wood and wood composites. – Van Nostrand Reinhold, New York, NY Carslaw HS, Jaeger JC (1958) Conduction of heat in solids. Oxford University Press, London Eriksson J (2005) Moisture transport and moisture induced distortions in timber. – Doctoral Thesis, Department of Applied Mechanics, Chalmers University of Technology, Göteborg Hagstrand PO (1999) Mechanical analysis of melamine-formaldehyde composites – Doctoral Thesis, Department of Polymeric Materials, Chalmers University of Technology, Göteborg Hanhijärvi A (1995) Modelling of creep deformation mechanisms in wood. – Dissertation, Technical Research Centre of Finland, Espoo Jönsson J (2005) Moisture induced stresses in timber structures. -Doctoral Thesis, Report TVBK1031, Division of Structural Engineering, Lund University Koc P, Houska M (2002) Characterisation of the sorptive properties of spruce wood by the inverse identification method. Holz als Roh und Werk 60:265 – 270 Kollmann F (1982) Technologie des Holzes und der Holzwerkstoffe 2. Auflage- Springer, Berlin, Heidelberg, New York, NY Simpson WT (1993) Determination and use of moisture diffusion coefficient to characterize drying of northern red oak (Quercus rubra). Wood Sci Technol 27:409–420 Wood Handbook (1999) Wood as an engineering material. -USDA Forest Products Laboratory, Forest Laboratory, Madison, WI
Chapter 19
Glue Line Nondestructive Assessment in Timber Laminates with an Air-Coupled Ultrasonic Technique Sergio J. Sanabria, Christian Müller, Jürg Neuenschwander, Peter Niemz, and Urs Sennhauser
Contents 19.1 19.2 19.3
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . Theoretical Considerations . . . . . . . . . . . . . . . . . . . . Material and Methods . . . . . . . . . . . . . . . . . . . . . . 19.3.1 Sample Preparation . . . . . . . . . . . . . . . . . . . . 19.3.2 Experimental Setup . . . . . . . . . . . . . . . . . . . . 19.4 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . 19.4.1 Imaging of Glue Presence and Repeatability of Measurements 19.4.2 Influence of Natural Variability and Anisotropy of Wood . . 19.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . 19.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
. . . . . . . . . . .
379 380 381 381 382 383 383 385 385 386 386
19.1 Introduction Glued solid wood products have gained much importance during the last years, as they allow an efficient and versatile use of the renewable timber material. Current standardized methods for bonding quality assessment consist of tests of small specimens cut from the structure during production or visual in-service inspection. Ultrasonic diagnostics are traditionally based on discrete point measurements using contact techniques. Transducers are generally pressed onto the timber surface with a coupling gel, liquid, or membrane couplant. Large glue line defects in glued timber constructions have been detected with this method (Dill-Langer et al. 2005). The disadvantages are a low precision in signal level measurements, which are highly dependent on the coupling pressure, and that the coupling agent may deteriorate the object. Better repeatability and one-dimensional continuous scanning is achieved S.J. Sanabria (B) Electronics/Metrology/Reliability Laboratory, Swiss Federal Laboratories for Materials Science and Technology, Empa, Überlandstrasse 129, CH-8600, Dübendorf, Switzerland e-mail:
[email protected] V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_19,
379
380
S.J. Sanabria et al.
with roller transducers, which have been applied to defect inspection in hardwood lumber (Kabir et al. 2002). Non-linear effects have been used to assess delaminations between veneer lamina and particle board (Solodov et al. 2004); a welding piezoelectric stack couples high power ultrasound into the sample and its surface is scanned with a laser vibrometer. Air-coupled ultrasonics (ACU) provides much more flexibility than traditional techniques since the transducer can be moved at a certain distance from the surface of the object, so fine and reproducible scanning in any direction is possible. A high-power low-frequency ACU system for split detection in wood composites is currently used in production lines (Niemz and Sander 1990). Ultrasonic imaging has been performed in solid wood using through-transmission mode for inspection of density, knots, microcracks and drilled holes (Gan et al. 2005; Hasenstab 2006). Delaminations in wood panel paintings between solid wood and a thin plaster layer have been assessed with both through-transmission and single-sided inspection (Siddiolo et al. 2007). In this work we present preliminary results of the application of ACU to assess disbonding in glued solid wood objects. A specific measurement set-up and data evaluation based on voltage level measurements of recorded A-scans allows precise imaging of areas with and without adhesive. Advantages and limitations of this method are discussed.
19.2 Theoretical Considerations The interpretation of the measurements is based on the theory of plane waves in homogeneous isotropic layered media (Brekhovskikh 1980). The sample is modeled as a three layers system, i.e. wood/glue/wood for glued material and wood/air/wood in the case of non-glued material. Due to the high acoustic impedance mismatch between air and solids the pressure level of an ACU signal which propagates through non-glued material is significantly lower than the level for glued material. Only a single echo of a longitudinal wave propagating through the three layers is considered. The acoustic attenuation in the glue line is neglected. It is assumed that the voltage level measured with an ACU transducer is proportional to the force exerted on its surface by ultrasonic waves (Schmerr and Song 2007). A simplified expression for the level ratio is given in Eq. (19.1): Lglued/non glued = 20 · log10
Vglued Vnon glued
= Twood→glue→wood − Twood→air→wood
1 ·Z2 Zi = ρi · ci T1→2→1 = 20 · log10 (Z4·Z+Z 2 ) 1 2 (19.1) Where: Lglued/non glued (dB) is the amplitude level ratio between ACU signals propagating through glued and non-glued material; Vglued (V) and Vnon glued (V) are corresponding amplitude measurements in the recorded A-scans. T1→2→1 (dB) is
19
Glue Line Nondestructive Assessment in Timber Laminates
381
the transmission coefficient for a single echo propagating through a layer of material 2 between two semi-infinite media of material 1. Zi (Pa·s/m), ρi (kg/m3 ) and ci (m/s) are the acoustic impedance, density and speed of sound in the propagation direction for medium i. From available data cspruce = 1300 m/s (measured in T orthotropic direction from ACU data following a similar method to (Vun et al. 2003) and ρspruce = 409 kg/m3 (gravimetrical determination), therefore Zwood = 0.532·106 Pa·s/m . From literature data (Deutsch et al. 1997) Zglue = 2.2·106 Pa·s/m and Zair = 0.000427 · 106 Pa · s/m (dry air T = 20◦ C). From Eq. (19.1) it follows that Twood→glue→wood = −4.1 dB and Twood→air→wood = −49.9 dB.
19.3 Material and Methods 19.3.1 Sample Preparation A total of 46 samples of common spruce (Picea abies Karst.) were manufactured in the Wood Physics Laboratory of ETH Zurich; each consisting of two 5 mm thick solid wood lamellas glued together except for some defined areas (Table 19.1). The R HB 110) applied adhesive is a one-component polyurethane resin (PURBOND 2 to one side of the boards with an amount of 200 g/m . The boards were pressed together hydraulically during 3 h with a stress of 0.8 N/mm2 . Before the gluing the wood was conditioned to normalized climatic conditions (T = 20◦ C and RH = 65%), which were afterwards also used for storage. Only solid wood lamellas with a small percentage of knots, resin pockets, grain distortion, etc. were used in order to analyze the variability of ultrasonic signals propagating through defect free glued timber. The cross-section of the samples is approximately in the orthotropic R-T plane and the curvature of the year rings is negligible (Fig. 19.1). After ultrasonic measurement, samples of type C and D were broken up and the profile of the transition between glued and non-glued areas was recorded with optical means.
Table 19.1 Geometry of glued timber samples manufactured for ACU measurement Type
Description
A B
Single solid wood lamella of dimensions 500×100×10 mm3 Two lamellas of dimensions 500×100×5 mm3 glued together to form a glued timber object of 500×100×10 mm3 Same geometry as B. No adhesive applied in the left half area (250×100 mm2 ) of the lamellas Same sample type as B. Adhesive only applied in two small areas (about 30×100 mm2 ) on the left and right edges of the lamellas
C D
382
S.J. Sanabria et al.
Fig. 19.1 Photographs of external surfaces of a typical C sample. An aluminum spacer of 1 mm thickness on the edge of the non-glued part allows control of the gap thickness. The year ring angle varies between 90º (propagation in orthotropic T direction) and 45º. Small defects like a 26 mm long resin pocket were allowed
19.3.2 Experimental Setup The measurement setup is shown in Fig. 19.2. Two ACU broadband planar transducers (model NCG100-D50, The Ultran Group) with a central frequency of 120 kHz and 50 mm active diameter were aligned perpendicularly to the surfaces of the sample, one transmitting an ultrasonic signal and the other one receiving it. The distance between the transmitter and the sample (210 mm) was chosen to minimize the diamR moves eter of the sound field penetrating the latter. A three-axis system from ISEL the two transducers together as a fixed unit; scanning the surface of the samples with steps of 1 mm in the fast axis and 4 mm in the slow axis. A sinusoidal pulse of 115 Vpp amplitude and 33 μs length windowed with a Gaussian function was applied to the transmitter. Received waveforms were amplified with a gain of 52 dB and digitized with a sampling frequency of 2.5 MHz and 14 bits resolution, the generated A-scans being stored for each scanned position. No averaging of A-scans was performed. C-scans were generated from a peak or root mean square (RMS) voltage measurement for each A-scan and a defined time gate [t1 , t2 ]:
VPEAK = max V(t) [t1, t2]
VRMS
=
1 t2 − t1
t2 V(t)2 dt
(19.2)
t1
The distance between receiver and sample (80 mm) allows separating in time multiple reflections between their surfaces (3) from measured waves (1) and (2). A-scans received through bonded material (1) present a signal-to-noise ratio of 55 dB, which allows for enough dynamic range to record waveforms from glued and non-glued areas in a single scan. Waves diffracted at the edges of the sample (4) are blocked by a frame built around the inspected object. The frame is made from wood (Norway spruce) covered by several layers of paper with small gaps of air
19
Glue Line Nondestructive Assessment in Timber Laminates
383
Fig. 19.2 Experimental set-up. Top image: Inspection principal and main propagation paths. 1 and 2 are waves propagating through the sample for glued and non-glued areas respectively. 3 are multiple reflections between the receiver and the surface of the sample. 4 are waves diffracted at the edges of the sample, which are blocked by a frame built around the object. The noise level in the A-scans is 1.3 mVRMS . Bottom image: Profile of voltage level along the fast axis normalized with respect to the glued area. The gap thickness decreases linearly between fast axis 0 and 250 mm from 1 mm down to the glue line thickness
in between. Waves propagating through the frame are highly attenuated due to the accumulation of impedance mismatch losses.
19.4 Results and Discussion 19.4.1 Imaging of Glue Presence and Repeatability of Measurements Figure 19.3 demonstrates successful ultrasonic imaging of absence and presence of adhesive of a typical glued timber sample of type C, which corresponds to the object photographed in Fig. 19.1. As expected, there is a strong voltage reduction in the left area of the board surface, corresponding to the non-glued region. The transition between glued and non-glued areas could be imaged accurately. Figure 19.4 shows an ultrasonic image of a sample type D, in this case the two glued areas on the left
384
S.J. Sanabria et al.
Fig. 19.3 ACU imaging of glue presence for the sample in Fig. 19.1. Top image: Photograph of open board; the transition between glued and non-glued area is highlighted. Bottom image: ACU C-scan of the sample. Feature 1 of the transition between glue/no glue and the resin pocket (2) can be visualized
and right sides of the surface of the image can be clearly distinguished; the amplitude values being higher than the ones measured in non-glued areas. In both images, details of the transition between glued and non-glued regions can be resolved. The spatial resolution of the images is limited by the sound field diameter (about 35 mm); features smaller than 20 mm cannot be resolved. Preliminary tests applying spatial deconvolution algorithms to the ultrasonic images showed an improvement of the resolution limit down to 10 mm.
Fig. 19.4 ACU imaging of glue presence and absence of a sample type D. Top image: Photograph of open board; in this case there was a small drop of glue (feature 1) joining the two lamellas, separated from the glued area on the right side of the sample. Bottom image: ACU C-scan. The presence of the drop of glue can be clearly recognized; however, the non-glued area between feature 1 and the glued region on the right side is smeared by the finite diameter of the sound field
19
Glue Line Nondestructive Assessment in Timber Laminates
385
Figure 19.2 shows in logarithmic scale the voltage level variations between glued and non-glued area for a typical fast axis amplitude profile of a sample type C, and for specific air gap thicknesses between the two lamellas. The voltage level in the non-glued area shows a minimum of –50 dB with respect to the glued area, in good agreement with the estimation for Twood→air→wood . As the gap thickness between boards decreases, this level rises due to multiple reflections of the ultrasonic wave adding constructively in the gap. About 20 mm from the boundary to the glued area propagation through bonded material becomes dominant, owing to the finite diameter of the sound field. An amplitude rise is observed in the delaminated area from fast axis 0 to 50 mm. It probably corresponds to residual ultrasonic energy diffracting through air at the edges of the sample, which is not blocked by the frame. Best measurement performance was achieved by limiting secondary ultrasonic propagation paths by evaluating a reduced number of cycles at the beginning of the received waveform. RMS voltage and peak voltage give similar results with sufficiently short temporal gates (less than 40 μs). Repeated ACU measurements of the same object showed variations of less than 0.1 dB (error < 1%). A homogenous amplitude level was observed in the ACU images of samples type A and B. The average voltage level measured for type B glued samples is –1 dB with respect to the value for type A solid wood samples; a smaller difference than predicted by Twood→glue→wood , which further enhances the contrast of ACU images. A probable reason is the constructive interference of multiple reflections of the ultrasonic wave in the glue line.
19.4.2 Influence of Natural Variability and Anisotropy of Wood Wood inhomogeneity introduces variations of up to 8 dB in voltage measurements of glued material without compromising the detectability of non-glued areas. Due to the small uncertainty of ACU measurements specific wood structure features can be visualized in the C-scans. Regions with highest latewood concentration show lowest voltage levels. A possible reason is the fact that latewood has higher acoustic impedance than earlywood and therefore larger impedance mismatch with air. Small defects in the material decrease the measured voltage, since they scatter partially the ultrasonic field; for instance, a resin pocket can be visualized in Fig. 19.3. Variations of the year ring angle could not be correlated to voltage amplitude changes in a clear fashion, an indication that the influence of anisotropy is not large for the inspected objects.
19.5 Conclusion We have demonstrated that air-coupled ultrasound is well-suited for glued timber inspection; combining the high sensitivity to disbonded interfaces of traditional ultrasonic methods with a phenomenal reproducibility in amplitude measurements,
386
S.J. Sanabria et al.
and precise spatial data acquisition. Moreover it is a fully non-invasive method, since no couplant is required between transducers and sample. Current state-of-theart transducers plus moderate pulser voltage and receiver gain allow transmission through 10 mm thick glued timber samples with a signal-to-noise ratio of 55 dB; therefore inspection of thicker objects is promising. The repeatability error is smaller than 1%. A through-transmission measurement set-up achieves level variations of up to 50 dB between glued and non-glued material, which ensures a reliable glue line assessment despite amplitude variations of up to 8 dB in bonded regions, due to the heterogeneity of the wood structure. Future research work is planned to inspect thicker (over 10 cm) multiple laminated glued timber. The main challenge is to resolve small amplitude level variations between bonded and disbonded areas from larger level variations within bonded material (higher influence of natural variability and anisotropy).
19.6 Summary Wood is a sustainable construction material. Glued timber products make efficient use of the strength properties of solid wood; moreover, structural members of expanded dimensional and geometrical properties can be produced. The integrity of the glue lines of timber laminates needs to be assessed during the full life cycle of the product; therefore, a non-destructive reproducible inspection method is required. As part of an ongoing project, we performed air-coupled ultrasound (ACU) measurements in glued timber laminates. A normal transmission setup with 120 kHz commercial transducers was used to analyze samples consisting of two spruce solid wood lamellas glued together with polyurethane adhesive introducing defined delaminated areas. Ultrasonic scanning with high resolution was performed to successfully image the presence or absence of glue. The geometry of the delaminated regions and features of the wood structure could also be visualized. We have demonstrated that ACU is a sensitive, accurate, reproducible and non-invasive inspection alternative with respect to conventional contact techniques; therefore, it is wellsuited for glued timber inspection. Future work is planned for the inspection of more complex glued timber structures. Acknowledgements This research has been supported by the Swiss National Science Foundation under contract 200021-115920. The authors acknowledge the work of Oliver Tolar and Fabian Binkert in the analysis of optical images and ultrasonic data.
References Brekhovskikh LM (1980) Waves in layered media. New York, NY, Academic Deutsch V, Platte M, Vogt M, Verein Deutscher Ingenieure (1997) Ultraschallprüfung Grundlagen und industrielle Anwendungen. Springer, Berlin Dill-Langer G, Bernauer W, Aicher S (2005) Inspection of glue-lines of glued-laminated timber by means of ultrasonic testing. In: Proceedings of the 14th international symposium on nondestructive testing of wood. Eberswalde, pp 49–60
19
Glue Line Nondestructive Assessment in Timber Laminates
387
Gan TH, Hutchins DA, Green RJ, Andrews MK, Harris PD (2005) Noncontact, high-resolution ultrasonic imaging of wood samples using coded chirp waveforms. IEEE Trans Ultrason Ferroelectr Freq Control 52(2):280–288 Hasenstab A (2006) Integritaetspruefung von Holz mit dem zerstoerungsfreien Ultraschallechoverfahren.Technische Universitaet Berlin. PhD Thesis Kabir MF, Schmoldt DL, Schafer ME (2002) Time domain ultrasonic signal characterization for defects in thin unsurfaced hardwood lumber. Wood Fiber Sci 34:165–182 Niemz P, Sander D (1990) Prozessmesstechnik in der Holzindustrie. VEB Fachbuchverlag, Leipzig Schmerr LW, Song SJ (2007) Ultrasonic nondestructive evaluation systems models and measurements. Springer, New York, NY Siddiolo AM, D’Acquisto L, Maeva AR, Maev RG (2007) Wooden panel paintings investigation: an air-coupled ultrasonic imaging approach. IEEE Trans Ultrason Ferroelectr Freq Control 54(4):836-846 Solodov I, Pfleiderer K, Busse G (2004) Nondestructive characterization of wood by monitoring of local elastic anisotropy and dynamic nonlinearity. Holzforschung 58:504–510 Vun RY, Wu QL, Bhardwaj MC, Stead G (2003) Ultrasonic characterization of structural properties of oriented strandboard: a comparison of direct-contact and non-contact methods. Wood Fiber Sci 35(3):381–396
Chapter 20
From Present Researches to Future Developments Voichita Bucur
Content References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
391
Delamination phenomena in manmade composites (Tay 2003, Sridharan 2008) as well as in wood and wood-based composites have received much attention from scientists and practitioners due to serious technological implications and obvious scientific curiosity for this subject. The long term objectives for the application of current research on delamination in wood and wood-based composites reported in this book are to develop robust design/modelling tools for minimizing the potential failures of both conventional and new products. Topics such as the theoretical aspects, the methodology for delamination detection and the factors inducing and affecting delamination were reviewed. The industry prospective of delamination in different products was also presented. The theoretical aspects have been related to physical understanding of phenomena for delamination initiation and growth. For structural health monitoring and damage detection techniques two approaches were used: the vibration – based monitoring and the fracture mechanics. The vibration – based approach involves model-based methods using low frequency vibrations, the fracture mechanics approach requires linear elastic and nonlinear concepts. In the vibration-based approach, the specimens are assumed to be free of defects: however, in numerical approaches, the stress concentration near to a notch or a flaw leads to mesh dependency. Stress criteria are needed in order to evaluate the occurrence of failure. Stress criteria require the definition of a critical crack dimension which depends on the material and stacking sequence. The main purpose of the model is to predict the deviation in materials properties if damage occurs (cracks, voids etc). The availability of a reliable model has many benefits such as the design and optimization of efficient testing configuration, the correct interpretation V. Bucur (B) CSIRO, Materials Science and Engineering Div. Bayview Avenue, Clayton, Victoria 3168, Australia e-mail:
[email protected] V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3_20,
389
390
V. Bucur
of experimental data, the development of an inverse technique based on quantitative data and the generation of training set for neural network. Fracture mechanics approach is based on the concept of strain energy release rate and assumes the presence of an inherent defect in the specimen (a notch). Crack propagation occurs when the strain energy at the crack front is equal to the critical strain energy rate, which is a material property. Fracture mechanics approach has demonstrated satisfactory accuracy in modelling the propagation of delamination; however, in many structural applications the locus of damage initiation is not obvious. In order to overcome the referred drawbacks cohesive damage models combined with continuum damage mechanics emerge as a suitable option which is not necessary to take into account an initial defect and in the same time, mesh dependency problems are minimized. In addition to these two approaches, the analysis of cumulative damage is fundamental in life prediction of components and structure under loading (Lemaitre and Desmorat 2005). Test procedures related to delamination of wood and wood based composites were commented. The behaviour of wood and wood-based composites at different length scales was discussed. The mechanism of delamination under static loading is well understood. Models were developed to explain the delamination growth and propagation under static loading or induced by other stresses such as microwave, drying and weathering. The development of ultrasonic techniques for non-destructive inspection of structural members was emphssized. Due to the great potential of incorporating novel biomaterials or integrating nanofibers/nanoparticles advanced wood-based composites are very attractive for new structural designs and applications. However the heterogeneity of wood – based composites is the main challenge irrespective to the unique nature of the constituents in the advance-composite design. In order to avoid over – dimensioning in wood or wood-based composites design it is necessary to develop theories and analytical/numerical tools that can take into account the initiation and growth of delamination in these new advanced wood-based composites. The delamination resistant design concept applied to wood-based composites can strongly influence their performance and cost. However more research is needed in order to achieve a fully mature methodology for use in design and certification of such wood-based composites structures. If delamination onset has been successfully predicted in laboratory samples using different codes (finite elements, etc), delamination predictions using these codes need to be validated on full size structural elements by comparing field data with experimental data. Changes in temperature and hygrometry can result in significant properties variations, so laboratory simulated testing is essential both to check fabrication quality of wood-based composites and to validate design data. The anisotropic nature of wood and wood-based composites as well as their multiple failure modes, have caused major difficulties in testing procedures for products strength. With respect to initiation of delamination, their stable propagation or unstable growth, the effect of load type (tensile, compressive, biaxial, etc) the rate of load application (monotonic, quasi static, dynamic, combined, etc) and of environmental conditions (temperature, hygrometry and pressure) should be investigated.
20
From Present Researches to Future Developments
391
Damage involves mechanisms at all scales, from the molecular level revealed by sophisticated instruments to massive scale during physical inspections. The damages can be induced by mechanical or thermal loads, environmental influences, or result from residual stress or combinations of them, such as thermo-mechanical stress. The ranges of different mechanisms and scales in damage accumulation in wood or wood-based composites raise permanent problems for a detailed experimental characterization as well as for modelling of delaminations. Implementation of more sophisticated approaches for mechanical behaviour of wood-based composites will require standardization of all three pure fracture modes and mixed fracture modes characterization test methods for delamination onset threshold and growth. There are new research areas to be suggested, such as: testing of delamination growth under dynamic loading, testing of multidirectional laminates, adaptation of tests to new wood-based composites with through- thickness reinforcements, or determination of in-situ interlaminar shear strength as a controlling factor for the initiation of delamination. Although the delamination induced by dynamic fracture toughness is of fundamental importance for wood machining and for pulp and paper industry, the dynamic delamination test is not easy to perform because it is experimentally difficult to induce high speed delamination growth in a simple and controlled manner (Ravi-Chandar 2004; Freund 1998). It must be borne in mind that traditional current, as well as new wood or wood-based composite structures can be highly vulnerable to damages, in particular delaminations that might have been introduced during manufacturing, tooling, processing or in service. Delaminations are difficult to detect by visual inspection, thus reliable and if possible inexpensive detection methods and technologies (active or passive) must be developed to improve safety and reliability of new wood – based laminated composites structures in service. It is of vital importance to identify the delaminations in new wood-based composites structures at the early stage, so as to prevent any potential failure. For further development in structural health monitoring of wood-based composites, the key issues are the prediction of delamination in different products and the improvement of the design of new advanced wood - based composites and structures, to prevent and minimize the products from delaminations. Without doubt, these fascinating challenges will be solved in the future by scientists and engineers with new perspectives involving in this highly interdisciplinary field, which has enormous potential for practical applications.
References Freund LB (1998) Dynamic fracture mechanics. Cambridge University Press, Cambridge Lemaitre J, Desmorat R (2005) Engineering damage mechanics. Springer, Berlin, Heidelberg Ravi-Chandar K (2004) Dynamic fracture. Elsevier, Amsterdam Sridharan S (ed) (2008) Delamination behaviour of composites. Woodhead Publishing, Cambridge Tay TE (2003) characterization and analysis of delamination fracture in composites: an overview of development from 1991 to 2001. Appl Mech Rev 56:1–31
Index
A Accuracy, 88, 225, 323, 363, 390 Acoustic coupling, 317 emission, 287–303, 308 imaging, 261–265 impedance, 380–381, 385 tomography, 255–266 waves, 259, 266, 311, 315, 318–320, 344 Active control, 391 Adhesion, 9, 34, 126, 145, 149, 152, 181, 329, 334 Adhesive joints, 72 Advanced materials, 3 Aging, 25–26, 44, 192 Air circulation rate, 200 Air-coupled transducers, 316, 324 American society for testing and materials (ASTM), 18–19, 24–26, 85, 87, 175, 192, 230, 289, 337 Amplitude, 41, 259, 288–289, 291–292, 294–295, 299, 301, 311–314, 317, 319–320, 324–331, 334, 336, 341, 359–360, 380, 382, 384–386 distribution, 295 Anatomic features, 124–130 Anisotropic, 3, 5–7, 9, 34, 53, 63, 75, 83, 92, 124, 131–132, 139, 187, 205–206, 211, 231, 233, 240, 242, 256, 258, 309, 318, 328, 339, 344, 390 Anisotropy, 60, 75, 78, 175, 179, 184, 193, 203–204, 206, 242, 292, 295, 317–320, 339, 385–386 Annual rings, 6, 24, 27, 78, 83, 167, 179, 181–184, 186, 193, 233, 271–272, 276, 279, 287, 290, 294–297, 303, 318, 337 Array, 6, 45, 75–76, 324 Attenuation, 34, 289, 298, 311, 323, 331, 333–339, 359, 364, 380
Automatic, 79, 176, 193, 323–324 Average frequency, 288, 302 values, 90 Axial strains, 88 Axial tension, 103 B Beam, 11, 38–44, 82, 154, 232, 244, 312, 316, 321, 325–326, 331, 334–335, 340, 342, 353–364, 366–368 Bending stress, 128, 226 Bending test, 67 Bernoulli theory, 42, 366–368 Biomaterials, 5, 145, 390 Biomechanical, 236 Boundary conditions, 104, 111–112, 199, 369–371, 373 Brittle fracture, 56, 71, 80–81, 90 Buckling, 20–23, 26, 28, 38, 76, 79, 166, 222, 235 C Calibrating, 176, 366, 368–371, 377 Cantilever, 39, 44 Capacitive sensors, transducers, 320–321, 323, 337, 344 Cellulose, 5–8, 11, 23, 63, 79, 126, 130–132, 183, 218, 233, 237, 240, 297 Chemical, 7, 9, 11, 29, 113, 118, 124–126, 138, 145, 169, 173, 175, 189, 191, 198–200, 233, 242 Chipboard, 223 Clear wood, 221, 225, 231 Climate, 6, 174–175, 178–179, 181–182, 186, 234, 270, 272–274, 277, 282, 365, 367, 372, 377 climatic cycling, 376 Coefficient of diffusion, 368–369
V. Bucur (ed.), Delamination in Wood, Wood Products and Wood-Based Composites, C Springer Science+Business Media B.V. 2011 DOI 10.1007/978-90-481-9550-3,
393
394 Cohesive strength, 9, 240 Collapse, 18, 87, 101–118, 159, 165, 204–205, 208, 234, 237–249, 333, 353, 363 Components, 3, 6, 29, 57, 74, 153, 161, 163, 166, 179, 182, 185, 198, 217–218, 242, 246, 299–300, 302, 321–322, 326, 390 Composites fiber reinforced, 9 wood-based, 3–5, 9–12, 17, 26–27, 29, 45, 51–92, 173–193, 288, 308–309, 311, 323, 329, 339–340, 343–344, 389–391 Compression, 5, 7, 10, 22–24, 26–27, 79, 87, 104, 125–126, 129–131, 133–134, 146, 166, 210, 219–220, 222–223, 225, 256, 261, 266, 297, 310, 332, 337 Concept, 37–38, 51, 53, 55, 152, 169, 208, 390 Constitutive law, 76 Constraints fracture toughness, 54, 60, 63, 65–67, 69–70, 72–73, 91, 136, 292, 391 mechanical, 102 Continuous emission, 288, 291–292, 302, 311 Contrast of the image, 150, 153–154, 385 Correlations, 88, 147, 149–150, 152, 155, 178, 362 Costs, 88, 199, 215–216, 222, 288, 377 Count, 288, 298–299, 302 Coupling, 176, 275, 277, 311, 317, 321, 331–332, 379 coupling media, 331–332 Crack arrest, 20, 53, 60, 64–65, 79 propagation, 7, 21–22, 54, 59–60, 63, 66, 70, 74, 87–88, 90, 124, 126, 134–136, 232, 292, 294, 297, 301–303 size, 73 Crack growth rate applied stress, 126 average crack length, 90, 179–180 crack propagation, 83, 87 damage, 83, 91 environmental effects, 233 fracture control, 88 fracture initiation, 81, 136 initiation, 91 length, 88 orientation, 85 stress intensity factor, 53–56, 60, 71, 88, 135–136, 291, 294 stress intensity rate, 88
Index Crack mouth opening displacements (CMOD), 301 Crack opening displacements (COD), 57–58, 75, 90, 292 Creeping, 366, 370, 375–377 Curing, 12, 227–232 cure, 226–229, 231 Curvature, 19, 39, 48–49, 78, 170, 218, 276, 367 Curve dispersion, 34, 37 R-curve, 57–59, 75, 89–90 D Damage evolution, 287 mechanics, 59, 390 Damping, 39–40, 321 Debonding, 4–5, 83, 87, 295, 297, 308, 337, 339–340 Decay, 22, 125, 126–128, 138, 145, 175, 178, 198, 223, 225, 232–234, 343, 358, 361 Decibels (dB), 292, 295, 328–329, 335, 342, 380–382, 385–386 Defect, 18–19, 26, 29, 34–36, 57, 90, 159, 165–166, 168, 197, 204–206, 217, 221, 223–228, 231, 259, 308, 311, 313–316, 320, 323, 326–328, 332–334, 336, 339–341, 344, 354, 358–363, 379, 380–382, 385, 389–390 size, 259, 334 Density, 9–11, 25–27, 28, 41, 53, 58, 60, 65, 69, 90–91, 112, 114–115, 133, 135–136, 160, 162–163, 179–180, 183, 205, 208, 218, 221–223, 227, 229, 237–238, 242, 246–247, 250–252, 276, 278, 309, 317, 323, 336–338, 343, 358, 361, 366, 368–369, 380–381 Dielectric properties, 160, 162–163, 169 Disbonds, 311, 328, 380, 385–386 Dislocation, 21–23, 151–152, 155, 256–257, 261, 288 Distortional effect, 368, 370, 377 Double cantilever beam (DCB), 73 Drying air, 200–203, 208 kiln, 200–203 Ductile fracture, 80–81 Duration, 56, 165, 175, 177, 217, 232, 288–289, 299 Dynamic fiber reinforced composites, 9 fracture, 391 toughness, 391
Index E Ecological relevance, 91 Effective crack length, 88 Elastic constants, 8–9, 55, 78, 113, 115, 309, 311 Elasticity, 10, 91, 116, 270, 276, 283, 291, 337, 339, 358, 361, 366 Elliptical, 242 Energy balance, 54 Energy release rate, 38, 54–55, 58–59, 72–73, 88, 90, 390 Engineered wood products (EWP), 216–217, 226–232, 235 Engineering materials, 26, 45 Errors, 4, 8 Event, 24, 45, 64, 288–289, 291–294, 297–301, 308 Experimental determination of fracture toughness, 391 stress intensity factor, 391 Exposure artificial, 178, 186–192 outdoor, 149, 174–175, 178–186, 191–192 F Failure modes, 91, 288, 366, 390 stress, 211, 389 Fatigue tests, 293 Fiber board, 229, 308 pullouts, 308 Finite ABAQUS, 365 element method, 43, 116, 366, 376 finite element analysis (FEM), 27, 33, 38, 45, 76, 78, 102–103, 112, 117, 372–373 Flaws, 34, 53, 308, 313, 320 Flexural, 39, 42, 310–311, 316–319 Flexural vibration, 39 Focused air-coupled transducers, 324 Fourier Transform Technique, 35, 37 Four-point-bending test, 367 Fractography, 80, 301 Fracture process zone, 56–57, 65, 72, 74, 92 Frequency natural, 343 spectrum, 327 Fundamental frequency, 327
395 G Gain, 34, 201, 217, 308, 320, 382, 386 Gap opening, 368, 370–371, 374–376 Geometry, 11, 54, 60, 62, 87, 105, 111, 113, 163, 167, 244, 329, 333–334, 337, 339, 341, 366–371, 373–374, 376–377, 381 Glue, 9–10, 24–25, 126, 134, 138, 192, 199, 227, 277, 283, 331, 340, 358–359, 367–368, 371, 374–377, 379–386 Glued laminated timber – glulam, crosslam, glulam slabs, 307 Grading, 221, 224–225, 358, 363 Grain angle effect, 89 slope, 166 Griffith, 52 Griffith theory, 52 Growth rings, 18, 78, 82, 128, 206, 222, 225, 242, 252, 272, 278, 283, 369, 373, 377 Guided waves, 320 H Hardness, 138, 178, 252, 297 Hardwood, 6, 11, 126, 130, 161–162, 164, 176, 185, 207–208, 211, 220, 222–223, 230, 235, 291–292, 362, 380 Harmonic, 39, 290, 328–329, 341–342 Health monitoring, 3, 12, 27, 29, 33–34, 37–38, 40, 44–45, 64, 389, 391 Heat transport, 160 Hemicelluloses, 7 Hidden specimens, 220–221, 225, 299–300 High resolution, 24, 138–139, 145–155, 386 Hook’s Law, 311, 344 Humidity, 4, 12, 26, 80, 136, 173, 177–178, 189, 198–203, 206, 233, 239, 244, 270, 273, 275, 308, 343, 366–367, 369, 377 I Impact testing, 223, 308 Impregnation, 159, 163–166, 169, 199 Infrared spectroscopy, 332 Inhomogeneities, 385 Initiation, 5, 20, 26, 29, 51–91, 135–136, 183–185, 218, 232, 235, 242, 291, 293, 295, 303, 356, 389–391 In-plane, 317, 320, 336 In-service damage, 308 In-situ, 64, 69, 82–83, 134, 288, 295, 297, 358, 391 Inspection continuous, 45 one side, 316 two side, 316
396 Integrity, 3, 25, 29, 34, 45, 137–138, 232, 308, 354, 386 Interface, 4, 9, 11–12, 24, 34–35, 42–43, 58–59, 74, 87, 127, 132–133, 145–156, 165, 295, 318, 321 Interface wood-cement, 12 Interfacial cracking, 59 Interfacial layers, 368 Interlaminar fracture toughness, 391 Inverse Fourier transform, 35, 37 Irwin lamellar tearing, 55 Iterative, 110 J J-integral, 57–59 Joints, 24–25, 72, 84, 87, 192, 226 K Knots, 25, 166, 204, 217, 221, 223, 225–226, 231, 380–381 L Lamb wave, 33–35, 37–38, 45, 309–311, 317, 339–341, 344 Laminated veneer lumber (LVL), 3, 9–10, 12, 19, 29, 227, 229, 307, 343–344 Laminates, 4, 19, 355, 379–391 Laser scanning microscopy, 145, 147–152, 295 Laser vibromerty, 39, 319, 380 Lateral, 104, 166, 220, 260, 262, 266, 323, 327 Leaky surface acoustic waves (SAW), 318–320 Levels, release rate, 38, 54–55, 58–59, 72–73, 88, 90, 390 Life prediction, 390 Lignin, 6–7, 63, 91, 126, 130, 135, 173, 181, 185, 187, 189, 193, 218–220, 233, 235 Linear elastic fracture mechanics (LEFM), 53, 56, 60, 74, 92 Load displacement curves, 59, 64, 67, 74–75, 79, 134, 297 modes, 64 rate, 38, 60, 71–72 Local, 20, 23–24, 27, 29, 33, 37, 40, 54, 71, 130, 136, 185, 241, 256, 276, 288, 317, 328, 341, 366, 370, 373 residual stress, 3, 87–88, 112 Localization, 287, 289, 292, 303, 313 Longitudinal, 6–7, 18, 20, 41, 53, 77, 79, 113, 115, 125, 129, 131–137, 159, 161–162, 164–165, 169, 186, 205–206, 217–220,
Index 233, 256, 258, 261, 269, 282–283, 309, 313, 316, 322, 366–367, 369–370, 380 Low frequency vibrations, 33, 38, 389 Low temperature, 137, 208–210 Lumber, 3–4, 9–10, 18–19, 25, 191, 215–217, 221–227, 231, 233–234, 255–256, 297, 300, 307, 324, 380 M Machining, 87, 149, 252, 391 Macrocracks, 56, 64, 291 Main, 19, 35, 45, 56, 64, 69, 87, 112, 116, 139, 147, 152–153, 162–163, 179, 192, 206, 220–221, 228, 240, 242, 246, 251, 256, 269, 289, 302, 308–309, 311–312, 322, 333, 337, 343, 355, 365–366, 375, 383, 386, 389–390 amplifier, 312 Manufacturing defects, 223–224 process, 28–29, 223–230 Mass production, 3 Material properties, 77–78, 101, 103–105, 111–112, 115–116, 251, 367, 377 Maximum, 64, 66–67, 72, 134, 136–137, 168, 178, 185, 198, 202, 206–208, 210–211, 216, 221, 239, 246, 279, 291, 293, 300, 325, 329, 353, 374 Mechanical performance, 353 properties, 3, 6–11, 34, 39, 78, 113, 116, 166, 175, 193, 205, 225, 233, 240, 292, 307, 329 Mechanical properties of cell wall, 7, 113, 240 Medium density fiberboard (MDF), 26, 85, 87–90, 229, 340–341, 343 Microcracks, 20, 56, 59, 64, 69, 127, 133, 136, 185, 291–293, 295, 298, 337, 380 Microdefects, 35 Microfibril angle, 6, 79, 82, 113, 129–131, 134, 219, 240, 242, 260 Microscopy confocal, 134, 138 electron, 81, 131, 136, 138–139, 146, 149–152 light, 9, 21–22, 24, 131, 137–138, 147, 149–152 Microstructural, 7, 56, 75, 91 Micro-voids, 149, 159, 165 Microwaves, 159–170, 200, 227, 229, 235, 343, 390 Mineral bonded particleboard and fiberboard, 308
Index Modal analysis, 39, 42 Mode I, 52–53, 60, 62–67, 69, 71–72, 74–75, 80, 84–85, 91, 290, 301, 303 Mode II, 52–53, 60, 62–65, 69, 71, 74 Model analysis of structural, 33 circular based, 102, 104–106 collapse recovery, 101–119 linear behaviour, 38–40 local and global information, 27 mathematical, 38, 205–211, 240 model-based methods, 38–44, 389 nonlinear behaviour, 40–44 squared based, 106–117 Modeling, 8–9, 35, 41, 75, 77, 92, 168–169, 205–206, 297 computational modelling, 5, 9 Modulus of elasticity, 91, 116, 291, 337, 358, 361, 366 Moisture, 136–137, 174, 178, 207, 209, 270–272, 274–278, 297–299, 365–377 Monotonic loading, 59 Monte Carlo simulation, 10 Morphology, 76, 80–82, 84, 91, 134 Multi-layered materials, 6 Multiple delaminations, 29, 39, 42 N Natural defects of wood knots, 204, 225 slop of grain, 225 Neural network, 44, 289, 299, 390 Nominal stress, 53, 295 Non-contact, 39, 311, 316–320, 323–324, 329, 334–344 ultrasonic transducers, 320, 323, 329, 334–343 Nonlinear acoustic modulation, 290 nonlinear behaviour, 40–44, 72, 328, 344 Nonlinearity, 50, 349 Nonparametric models methods, 44 Notch effect, 129 Nuclear magnetic resonance (NMR), 9 Numerical analysis, 365, 368, 372 calculation, 377 O Opening, 24–25, 41, 52, 56–59, 64, 73–75, 90, 92, 134, 228, 235, 239, 292, 368, 370–371, 374–377 mode stress intensity, 56
397 Operational, 44–45, 153, 215 Orientation, 5, 11, 24, 53, 60, 64–67, 71, 75, 78, 85, 87, 90, 92, 102, 106, 113, 128, 130–131, 136, 152, 167, 184, 186–187, 218–219, 223, 230–231, 233, 243, 256, 266, 272, 276, 278, 283, 289, 294, 303, 313, 318, 367–368 crack, 64, 90 OSB oriented strandboards, 336 Out of plane, 52–53, 71, 92 Overall mechanical characterization, 34 P Parameters, 3, 5, 34, 39, 54, 55, 59, 64–65, 70, 72, 76, 79, 91–92, 103, 112, 162–163, 167, 169, 189, 191, 275–277, 280, 282–283, 288–293, 299–300, 302, 309, 313, 321, 324, 327, 329–330, 332, 336–337, 339, 361–363, 367–368 linear fracture mechanics, 72, 92 Parquet cupping, 373–374 deformation, 377 floors, 365–377 gap opening, 370, 374, 376–377 geometry, 366–371, 373–374, 376, 377 parquet lacquered specimens, 367, 373 parquet planks, 366–368, 370, 376–377 simulate the behaviour of parquet planks, 368 Particle based boards – oriented strandboards, particle board, fiber board, 308 Pattern recognition, 39, 297 Peak amplitude, 289, 299, 329, 336 Periodical, 39, 45, 175 Phase velocity, 34–35, 340 Physical methods, 45, 52, 87–88 Piezoelectric actuators, 40 sensors, 40 Pin contact forces, 42 Planar, 52, 56, 382 Plane stress, 54–55 Plane wave propagation, 34 Plastic deformation, 56, 59, 131–132 strains, 104–106, 137, 337 zone, 53–54, 56 Plasticity, 88 Plate wave technique, 329, 339–342, 345 Plywood, 9–10, 12, 19, 24–25, 28, 85, 149–155, 174, 191–192, 216, 227, 229, 339, 343
398 Point source, 36, 201 Poisson’s ratios, 8, 271, 276, 280 Polymers, 103, 119, 217–218 Poor cure, 227 Porosity, 34 Prediction of life, 390 Pressure, 18, 28, 104–105, 108, 111–112, 159–163, 167, 192, 200–201, 215, 218, 227–228, 230, 233, 235, 238–239, 272, 379, 380, 390 sound, 380 Principle of superposition, 38 Processing, 12, 40, 119, 150, 160, 162, 167, 169, 198, 206, 220–221, 224, 228, 237–253, 256, 288–289, 299, 303, 308, 343, 391 internal checking, 220, 237–253, 343 Production process, 11, 169, 217, 227, 229, 236, 308, 323, 343 Propagation, 5, 7, 20–22, 34–35, 53–54, 56–60, 63, 64, 66, 70, 74, 82–83, 87–88, 90, 92, 124–126, 134–136, 175, 183–184, 217–218, 221, 232, 235, 255–256, 289, 291–292, 294–297, 301–303, 309–313, 317, 319, 322, 327, 336, 339, 356, 363, 375, 381–383, 385, 390 Pulp, 11, 22, 92, 133, 135, 216–217, 233, 246, 256, 391 P wave, 314 Q Q factor, 193 Quality assessment, 3, 256, 316, 379 Quantitative, 3, 34, 75, 148, 192, 246, 288, 303, 390 R Radial, 6–7, 18–19, 23, 25, 53, 75, 77, 81–84, 103, 124–131, 134–138, 159, 161–162, 164–165, 167, 169, 179–180, 183–185, 188, 205–206, 216–262, 217–221, 233–234, 242, 244, 256–259, 265–266, 269–270, 275, 279, 282–283, 295, 302, 367–369 Radiations, 126, 145, 160, 167–168, 173, 177–178, 185–189, 233, 290, 317, 321 Radiographic, 303 Ratios, 8, 180, 271, 276, 279–280, 294, 317 Raw material, 9, 11, 215–217, 220–224, 227, 236, 307 Rayleigh, 320 R-curve, 57–59, 75, 89–90 Reaction wood, 129–130, 179, 193, 220, 256, 261, 266, 297
Index Reconditioning, 101, 112, 118, 165, 227–228, 247–250, 252 Reference, 11, 18, 20, 34, 39, 51, 54, 63, 74–75, 103–104, 175–176, 191, 240, 255, 261, 281–282, 309, 311, 320, 325, 329, 371 stress, 18, 20, 34, 54, 74–75, 92, 104, 175, 240, 261, 282, 309, 311, 320, 371 Reflection, 34, 176, 193, 311, 315, 318, 326, 343 elastic waves, 34, 311 Refraction, elastic waves, 311 Relative humidity, 26, 80, 173, 189, 199–200, 202, 206, 239, 244–245, 270, 273, 275, 366–367, 369, 377 Reproducibility, 323, 332, 385–386 Residual strength, 363 stress, 3, 87–88, 112, 363, 391 Resin pockets, 125, 128, 166, 381–382, 384–385 Resistance, 3, 11, 24, 54, 57–60, 66, 75, 82, 88, 127–128, 132–133, 135, 191, 232, 234, 244, 270, 275, 307, 361 Resonance, 316–317, 343–345 Restraint cracking, 220 Review, 39, 45, 51, 66, 155, 160, 169, 175, 191, 217, 242, 290, 308, 344, 389 Rise time, 288, 299, 302 Rock, 219 Rods, 77, 323, 328 Root mean square (RMS) voltage, 288, 302, 382 Roughness, 75, 83, 146, 153, 292 Round wood, 234 Rupture, 7, 25, 53, 56, 59, 72, 83, 160–161, 166, 168–169, 204, 231, 293, 295–296, 337, 361–362 S Safe design, safety, safe performance, 5, 9, 44, 91–92, 229, 232, 235, 308, 343, 345, 391 Safety factors, 91–92 Samples, 82, 112, 127, 133–139, 165, 168, 198, 225, 240–241, 246–248, 251, 260, 271–273, 275, 278, 366–367, 372–373, 381–382, 385–386, 390 Sampling rate, 382 Scanning electron microscopy (SEM), 21–24, 62, 72, 131, 134, 136–139, 146, 149–155, 219, 297 Scanning modes, 325
Index Scattering, 261 Serviceability of structures, 3 Shape change, 355 Shear bands, 129 modulus of elasticity, 116, 358, 361, 363 waves, 309–310, 316–317, 359, 364 strains, 79, 105 Shrinkage, 18, 25, 29, 175, 179–180, 183, 185, 192, 198, 203–207, 219–220, 231–233, 237, 241–244, 246–252, 271, 367 Simulation, simulated, 7, 9–10, 13, 26, 35, 38, 42, 58–59, 64, 74–75, 77, 88, 112, 160, 162, 168–169, 175, 206, 270, 278–280, 282–283, 290, 331, 337, 339–340, 342, 358, 368–372, 376, 390 Size, 10, 53–54, 66, 72–73, 78, 82, 87, 146, 165–169, 176, 218, 225–226, 230–231, 239, 247–248, 256, 259–260, 313, 323, 332–335, 343, 345, 361, 363, 366, 374, 390 Smart composites, 43–44 Softwood, 6–7, 11, 53, 63, 76, 128–129, 132–133, 136, 161–162, 164, 176, 181, 192–193, 198, 202, 208, 211, 218, 220, 230, 234, 279, 291–292 Solid state adhesive layer, 12 Species, 6, 21–24, 53, 60, 63–65, 67, 69–70, 87, 114, 124, 126, 130, 133, 135–136, 159, 161–162, 164, 166–167, 174–175, 178–180, 184–185, 187, 189–191, 198, 203, 205, 207, 216, 218, 221–223, 234–235, 237–240, 242, 246–248, 290–294, 303, 320, 358, 361, 366, 374 Spectral analysis, 91 Spectroscopy, 332 Split, 19, 24, 62–64, 67, 74, 83–85, 88, 92, 125, 159, 165–166, 168, 174, 179, 191, 197, 203, 204, 208, 211, 217–218, 220–226, 231, 233–234, 236, 272, 290, 380 Spring, 6, 41–42, 77, 204, 206, 219–220, 235 Stability, 26, 56, 71, 113, 137, 149, 198, 226, 241, 269, 279, 307, 353, 363, 370–371 Static curves, 293, 337 loading, 390 tests, 293 Statistical based methods, 44 Stiffness, 5, 7, 10, 34–35, 39, 41, 52, 59, 75, 77, 79, 101–102, 135, 225–226, 228, 241–242, 252, 278–279, 283, 292, 303, 358, 363, 369
399 Strain basic concept, 211 curves, 88, 134, 232, 294 energy, 38, 54–55, 58, 60, 301, 390 energy release rate, 38, 54, 55, 390 field, 54–55, 233, 390 hardening, 337 Strength, 5–7, 9–10, 26, 52–53, 59, 74–75, 77, 87, 91, 116, 129, 134, 136, 149, 152, 192, 199, 205, 210–211, 217, 222, 225–226, 228, 231, 233–234, 238, 240–241, 261, 288, 290–291, 308, 316, 358–364 Stress concentration, 72, 225–226, 295, 297, 389 distribution, 54, 56, 60, 71, 83, 107, 110, 112–115, 118, 125, 232, 282–283, 377 distribution in glue line, 377 field, 54–55, 183 stress intensity factor, 53–56, 60, 71, 88, 135–136, 291–294 Structural integrity, 137–138, 232, 308 Structures, 5–7, 9, 12, 18–19, 24, 26, 29, 34, 38–40, 42, 44–45, 53, 59, 63–64, 69, 75–79, 81–84, 102, 124, 126, 131, 133, 135–137, 139, 159–161, 163–164, 166–170, 179, 181, 183, 185, 187, 189, 193, 207, 217–220, 230–232, 234–235, 255–256, 261–262, 284, 287–303, 308, 313, 316, 326, 337, 344, 354–355, 358, 360–361, 363, 370, 379, 385–386, 390 Substrate, 9, 59, 152 Surface, 18, 24, 149–152, 179, 209, 211, 238, 241–246, 250, 262, 299, 366, 374, 376 T Tangential, 6–7, 18, 53, 57–59, 81–83, 103, 124–129, 131–132, 134–135, 137–138, 166–167, 169, 179–180, 183–187, 189, 205–206, 219–220, 233, 242, 244, 258, 260, 262–266, 269, 271–272, 275, 279, 292, 295, 328, 367–369, 373, 377 Temperature, 4, 12, 26, 87, 101–105, 110, 112–115, 118, 137, 152, 160–163, 174, 177–178, 181, 198–203, 206–211, 218, 223–224, 227–230, 232–233, 235, 238–241, 246, 248–252, 297, 308, 343, 367–368, 370–371, 377, 390 Tensile loads, 131, 134, 294 stress, 74, 81, 175, 182, 205, 210, 297, 328
400 Tension, 5, 18, 22–23, 27, 41, 52–53, 63, 65, 67, 71, 79, 81–85, 87, 103–104, 110, 130–131, 134, 136, 165–166, 204, 210, 219–220, 225–226, 231, 238–241, 244–245, 255–266, 290, 292, 294–295, 297–298, 303, 310–311 Tests, 25–26, 62, 64, 67, 69, 84, 88–89, 192, 198, 208, 223, 240, 261, 271–275, 279, 288, 290, 301–303, 312–313, 321, 322–323, 325, 332, 340, 354 testing configuration, 329–331, 389–390 Theoretical, 3, 7, 34, 45, 52, 57–58, 64–65, 72, 92, 101–119, 219, 289, 297, 309, 320, 334, 380–381, 389 Thermal stress, 297 waves, 161, 297 Thermo mechanical analysis, 135, 391 Thickness, 19, 25, 28–29, 34, 36, 39, 89–90, 150–152, 174, 176, 193, 203, 206–207, 227–228, 244, 248, 250–252, 271–274, 276, 279–280, 284, 295, 310–311, 313, 317–318, 320–321, 323, 326, 331–334, 337, 340–343, 356, 358, 367–368, 373–374, 382–383, 385, 391 Three dimensional systems, 270 three or Tri or 3D non-linear stochastic finite element model, 10 three point bending test method, 67, 366 Threshold, 87, 208, 211, 241, 288–289, 302, 343, 391 Timber, 9, 12, 25, 29, 92, 113, 128, 159–170, 186, 192, 197–211, 215–216, 221, 235, 237, 241, 246–248, 252, 266, 269–284, 307, 331–332, 379–386 Time reversal concept, 37 Tomography, 90, 255–266 Toughness, 5, 25, 53–54, 56, 59–60, 63, 65–67, 69–70, 72–73, 75, 79, 90–91, 127, 136, 222, 232, 292, 391 Transform Fourier transform, 35, 37 Transient, 35, 288 effect, 288 Transition, 60, 81–82, 129–130, 132, 137, 229, 235, 296, 381, 383–384 Transmission electron microscopy (TEM), 81, 128, 131–134, 138–140 Travel time, 326
Index U Ultrasonic nondestructive evaluation - local damage information, 33, 45, 288, 328 path, 260, 333 Ultrasonic techniques contact techniques, 334–340, 343, 379, 386 non-contact, 329, 334–342 pulse echo, 312–314, 344 through transmission, 312–314, 324, 329, 380, 386 Ultrasound, 165, 229, 316–319, 323, 337, 340–343, 380, 385–386 Ultrastructural features, 130–137 Uncertainties, 260, 385 Uniaxial loading, 103 Uniform stress field, 83 V Validity tests, 336 Velocity plate wave, 317–318 stress wave, 258, 262–264, 266, 361 surface wave, 320 ultrasonic, 256, 258, 260–262, 266, 332 Veneer, 3, 9–10, 12, 19, 24, 192, 221, 227–229, 256, 307, 318–320, 337, 339–342, 344, 366, 380 Veneer based panels – plywood, laminated veneer lumber, 307 Verification tests, 368 Vibration vibration based damage identification, 34 Viscoelasticity, 339 Viscoelastic properties of wood, 56, 60, 116 Voids, 25, 27–28, 147–149, 151–153, 155–156, 159, 161, 165, 167, 226, 229, 308, 336–337, 389 Volume, 60, 71, 163, 169, 179, 206–207, 220–221, 270, 274–276, 278, 314, 325 Volumetric strain, 206 W Warping, 167, 174, 191, 204, 206, 269–284 Waves modes, 34, 309–310, 323 plane, 34, 261, 317, 380 surface, 260, 309, 311, 318–320, 357 ultrasonic, 34, 176, 256–257, 261, 311, 313, 317, 336, 380, 385 Weak bond, 34, 229 Weathering, 125–128, 145, 173–193, 233, 390 Wedge loads, 64, 67, 74, 88
Index Wood based composite panels, 192, 308, 311, 329–344 Wood coating interface, 145–156 Wood flooring systems, 365, 377 Wood-plastic composites, 311 Wood species, 6, 22–24, 53, 60, 126, 130, 159, 161–162, 164, 166, 167, 175, 178, 185, 205, 223, 234–235, 292, 361, 366
401 Y Yield, 53, 92, 146, 153, 206 strength, 53 Young’s modulus, 7–8, 10, 55, 260, 317, 337 structural modifications, 7 Z Zone plastic deformation, 56, 59
