This page intentionally left blank
Carbon Dioxide Sequestration and Related Technologies
Scrivener Publishing 3 Winter Street, Suite 3 Salem, MA 01970 Scrivener Publishing Collections Editors James E. R. Couper Richard Erdlac Pradip Khaladkar Norman Lieberman W. Kent Muhlbauer S. A. Sherif
Ken Dragoon Rafiq Islam Vitthal Kulkarni Peter Martin Andrew Y. C. Nee James G. Speight
Publishers at Scrivener Martin Scrivener (
[email protected]) Phillip Carmical (
[email protected])
Carbon Dioxide Sequestration and Related Technologies Edited by
Ying (Alice) Wu Sphere Technology Connection
John J. Carroll Gas Liquids Engineering, Ltd. and
Zhimin Du Southwest Petroleum University
Copyright © 2011 by Scrivener Publishing LLC. All rights reserved. Co-published by John Wiley & Sons, Inc. Hoboken, New Jersey, and Scrivener Publishing LLC, Salem, Massachusetts. Published simultaneously in Canada. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., Ill River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http://www.wiley.com/go/permission. Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com. For more information about Scrivener products please visit www.scrivenerpublishing.com. Cover design by Kris Hackerott. Library of Congress Cataloging-in-Publication ISBN 978-0-470-93876-8
Printed in the United States of America 10 9 8 7 6 5 4 3 2 1
Data:
Contents Introduction The Three Sisters - CCS, AGI, and EOR Ying Wu, John J. Carroll and Zhimin Du
xix
S e c t i o n 1: D a t a a n d C o r r e l a t i o n 1.
2.
3.
Prediction of Acid Gas Dew Points in the Presence of Water and Volatile Organic Compounds Ray. A. Tomcej 1.1 Introduction 1.2 Previous Studies 1.3 Thermodynamic Model 1.4 Calculation Results 1.5 Discussion References Phase Behavior of China Reservoir Oil at Different C 0 2 Injected Concentrations Fengguang Li, Xin Yang, Changyu Sun, and Guangjin Chen 2.1 Introduction 2.2 Preparation of Reservoir Fluid 2.3 PVT Phase Behavior for the C 0 2 Injected Crude Oil 2.4 Viscosity of the C 0 2 Injected Crude Oil 2.5 Interfacial Tension for C 0 2 Injected Crude Oil/Strata Water 2.6 Conclusions Literature Cited Viscosity and Density Measurements for Sour Gas Fluids at High Temperatures and Pressures B.R. Giri, P. Biais and R.A. Marriott 3.1 Introduction 3.2 Experimental
3 3 4 5 6 10 11
13
14 14 15 17 19 20 21
23 24 25 v
vi
CONTENTS
3.2.1 Density Measurement 3.2.2 Viscosity Measurement 3.2.3 Charging and Temperature Control 3.3 Results 3.4 Conclusions References 4.
5.
6.
Acid Gas Viscosity Modeling with the Expanded Fluid Viscosity Correlation H. Motahhari, M.A. Satyro, H.W. Yarranton 4.1 Introduction 4.2 Expanded Fluid Viscosity Correlation 4.2.1 Mixing Rules 4.2.2 Modification for Non-Hydrocarbons 4.3 Results and Discussion 4.3.1 Pure Components 4.3.2 Acid Gas Mixtures 4.4 Conclusions 4.5 Acknowledgements References Evaluation and Improvement of Sour Property Packages in Unisim Design Jianyong Yang, Ensheng Zhao, Laurie Wang, and Sanjoy Saha 5.1 Introduction 5.2 Model Description 5.3 Phase Equilibrium Calculation 5.4 Conclusions 5.5 Future Work Reference Compressibility Factor of High C0 2 -Content Natural Gases: Measurement and Correlation Xiaoqiang Bian, Zhimin Du, Yong Tang, and Jianfen Du 6.1 Introduction 6.2 Experiment 6.2.1 Measured Principles 6.2.2 Experimental Apparatus and Procedure 6.2.3 Experimental Results
25 27 30 31 37 37
41 41 42 44 45 47 47 48 52 52 52
55
55 56 58 62 62 63 65
65 67 67 67 68
CONTENTS
6.3
Methods 6.3.1 Existing Methods 6.3.2 Proposed Method 6.5 Comparison of the Proposed Method and Other Methods 6.6 Conclusions 6.7 Acknowledgements 6.8 Nomenclature References
68 68 74 78 83 84 84 85
Section 2: Process Engineering 7.
8.
Analysis of Acid Gas Injection Variables Edward Wiehert and James van der Lee 7.1 Introduction 7.2 Discussion 7.3 Program Design 7.4 Results 7.5 Discussion of Results 7.5.1 General Comments 7.5.2 Overall Heat Transfer Coefficient, U 7.5.3 Viscosity 7.6 Conclusion References Glycol Dehydration as a Mass Transfer Rate Process Nathan A. Hatcher, Jaime L. Nava and Ralph H. Weiland 8.1 Phase Equilibrium 8.2 Process Simulation 8.3 Dehydration Column Performance 8.4 Stahl Columns and Stripping Gas 8.5 Interesting Observations from a Mass Transfer Rate Model 8.6 Factors That Affect Dehydration of Sweet Gases 8.7 Dehydration of Acid Gases 8.8 Conclusions Literature Cited
89 89 90 93 94 96 96 101 104 105 105
107
108 110 111 114 115 118 119 119
CONTENTS
Carbon Capture Using Amine-Based Technology Ben Spooner and David Engel 9.1 Amine Applications 9.2 Amine Technology 9.3 Reaction Chemistry 9.3.1 Nucleophilic Pathway 9.3.2 Acid-Base Pathway (Primary Secondary and Tertiary Amines) 9.4 Types of Amine 9.5 Challenges of Carbon Capture 9.5.1 Prevention 9.5.2 Reclaimers 9.5.3 Purging and Replacing Amine 9.5.4 High Energy Consumption 9.5.5 Size of the Amine Facility 9.5.6 Captured C 0 2 9.6 Conclusion Dehydration-through-Compression (DTC): Is It Adequate? A Tale of Three Gases Wes H. Wright 10.1 Background 10.2 Water Saturation 10.3 Is It Adequate? 10.4 The Gases 10.5 Results 10.6 Discussion References Diaphragm Pumps Improve Efficiency of Compressing Acid Gas and C 0 2 Josef Jarosch, Anke-Dorothee Braun 11.1 Diaphragm Pumps 11.2 Acid Gas Compression 11.3 C 0 2 Compression for Sequestration 11.4 Conclusion Literature
121 121 122 124 124 125 126 128 128 129 129 129 130 130 131
133 133 138 138 141 147 151 152
155 162 164 167 171 172
CONTENTS
ix
Section 3: Reservoir Engineering 12. Acid Gas Injection in the Permian and San Juan Basins: Recent Case Studies from New Mexico David T. Lescinsky; Alberto A. Gutierrez, RG; James C. Hunter, RG; Julie W. Gutierrez; and Russell E. Bentley 12.1 Background 12.2 AGI Project Planning and Implementation 12.2.1 Project Planning and Feasibility Study 12.2.2 Reservoir/Cap Rock Identification and Regulatory Permitting 12.2.3 Well Drilling and Testing 12.2.4 Well Completion and Construction 12.2.5 Reservoir and Seal Evaluation 12.2.6 Documentation, System Start-up and Reporting 12.3 AGI Projects in New Mexico 12.3.1 Permian Basin 12.3.1.1 LinamAGI#l 12.3.1.2 Jal 3 AGI #1 12.3.2 San Juan Basin 12.3.2.1 Pathfinder AGI #1 12.4 AGI and the Potential for Carbon Credits 12.5 Conclusions References 13. C 0 2 and Acid Gas Storage in Geological Formations as Gas Hydrate Farhad Qanbari, Olga Ye Zatsepina, S. Hamed Tabatabaie, Mehran Pooladi-Darvish 13.1 Introduction 13.2 Geological Settings 13.2.1 Depleted Gas Reservoirs 13.2.1.1 Mixed Hydrate Phase Equilibrium 13.2.1.2 Assumptions
175
175 178 178 181 183 186 186 188 190 190 193 196 199 200 204 207 208
209
210 211 211 211 213
x
CONTENTS
13.2.2
Ocean Sediments 13.2.2.1 Negative Buoyancy Zone (NBZ) 13.2.2.2 Hydrate Formation Zone (HFZ) 13.3 Model Parameters 13.3.1 Depleted Gas Reservoir 13.3.2 Ocean Sediment 13.4 Results 13.4.1 Depleted Gas Reservoir 13.4.2 Ocean Sediment 13.5 Discussion 13.6 Conclusions 13.7 Acknowledgment References 14. Complex Flow Mathematical Model of Gas Pool with Sulfur Deposition W. Zhu, Y. Long, Q. Liu, Y. Ju, and X. Huang 14.1 Introduction 14.2 The Mathematical Model of Multiphase Complex Flow 14.2.1 Basic Supposition 14.2.2 The Mathematical Model of Gas-liquid-solid Complex Flow in Porous Media 14.2.2.1 Flow Differential Equations 14.2.2.2 Unstable Differential Equations of Gas-liquid-solid Complex Flow 14.2.2.3 Relationship between Saturation and Pressure of Liquid Phase 14.2.2.4 Auxiliary Equations 14.2.2.5 Definite Conditions 14.3 Mathematical Models of Flow Mechanisms 14.3.1 Mathematical Model of Sulfur Deposition 14.3.2 Thermodynamics Model of Three-phase Equilibrium 14.3.3 State Equations
213 213 214 216 216 217 218 218 221 221 223 224 224
227 227 228 228
229 229
230 231 232 232 232 232 234 236
CONTENTS
14.3.4 14.3.5
Solubility Calculation Model Influence Mathematical Model of Sulfur Deposition Migration to Reservoir Characteristics 14.4 Solution of the Mathematical Model Equations 14.4.1 Definite Output Solutions 14.4.2 Productivity Equation 14.5 Example 14.5.1 Simulation Parameter Selection 14.5.2 Oil-gas Flow Characteristics near Borehole Zones of Gas-well 14.5.3 Productivity Calculation 14.6 Conclusions 14.7 Acknowledgement References
xi
236
237 238 238 239 240 240 240 240 242 242 242
Section 4: Enhanced Oil Recovery (EOR) 15. Enhanced Oil Recovery Project: Dunvegan C Pool Darryl Burns 15.1 Introduction 15.2 Pool Data Collection 15.3 Pool Event Log 15.4 Reservoir Fluid Characterization 15.4.1 Fluid Characterization Program Design Questions 15.4.2 Fluid Characterization Program 15.4.3 Solubility of Acid Gas Mixtures in the Dunvegan C Oil 15.5 Material Balance 15.6 Geological Model 15.7 Geological Uncertainty 15.7.1 Formation Bulk Volume 15.7.2 Porosity 15.7.3 Permeability 15.7.4 Residual (Immobile) Fluid Saturations 15.7.5 Relative Permeability Curve Parameters 15.7.6 Fluid Contacts 15.8 History Match 15.9 Black Oil to Compositional Model Conversion
247 248 249 252 255 255 257 263 263 264 269 269 269 269 270 270 272 272 282
CONTENTS
Recovery Alternatives Economics Economic Uncertainty Discussion and Learning 15.13.1 Reservoir Fluid Characterization 15.13.2 Material Balance 15.13.3 Geological Model 15.13.4 History Match 15.13.5 Black Oil to Compositional Model Conversion 15.13.6 Recovery Alternatives 15.13.7 Economics 15.14 End Note References 15.10 15.11 15.12 15.13
C 0 2 Flooding as an EOR Method for Low Permeability Reservoirs Yongle Hu, Yunpeng Hu, Qin Li, Lei Huang, Mingqiang Hao, and Siyu Yang 16.1 Introduction 16.2 Field Experiment of C 0 2 Flooding in China 16.3 Mechanism of C 0 2 Flooding Displacement 16.4 Perspective 16.5 Conclusion References Pilot Test Research on C 0 2 Drive in Very Low Permeability Oil Field of in Daqing Changyuan Weiyao Zhu, Jiecheng Cheng, Xiaohe Huang, Yunqian Long, and Y. Lou 17.1 Introduction 17.2 Laboratory Test Study on C 0 2 Flooding in Oil Reservoirs with Very Low Permeability 17.2.1 Research on Phase Behavior and Swelling Experiments 17.2.2 Tubule Flow Experiments 17.2.3 Long Core Test Experiments 17.3 Field Testing Research 17.3.1 Geological Characteristics of Pilot 17.3.1.1 Structural Characteristics 17.3.1.2 Characteristics of Reservoir
290 307 312 312 312 315 315 316 317 317 317 317 318 319
319 320 321 324 326 326 329
329 330 330 331 332 333 333 334 334
CONTENTS
Reservoir Properties and Lithology Characteristics 17.3.2 Distribution and Features of Fluid 17.3.3 Designed Testing Scheme 17.3.4 Field Test Results and Analysis 17.3.4.1 Low Gas Injection Pressure and Large Gas Inspiration Capacity 17.3.4.2 Production Rate and Reservoir Pressure Increase after Gas Injection 17.3.4.3 Reservoir Heterogeneity Is the Key to Control Gas Breakthrough 17.3.4.4 C 0 2 Throughput as the Supplementary Means of Fuyu Reservoir's Effective Deployment 17.3.4.5 Numerical Result Shows that Carrying Out Water Flooding after Injecting Certain Amount of C 0 2 Slug is Better 17.4 Conclusion 17.5 Acknowledgement References
Xlll
17.3.1.3
18. Operation Control of C0 2 -Driving in Field Site. Site Test in Wellblock Shu 101, Yushulin Oil Field, Daqing Xinde Wan, Tao Sun, Yingzhi Zhang, Tiejun Yang, and Changhe Mu 18.1 Test Area Description 18.1.1 Characteristics of the Reservoir Bed in the Test Area 18.1.2 Test Scheme Design 18.2 Test Effect and Cognition 18.2.1 Test Results 18.2.2 The Stratum Pressure Status 18.2.3 Air Suction Capability of the Oil Layer 18.2.4 The Different Flow Pressure Control 18.2.5 Oil Well with Poor Response 18.3 Conclusions References
336 339 339 340 340
341
342
343
344 346 349 349 351
352 352 352 353 353 354 356 356 358 359 359
xiv
CONTENTS
19. Application of Heteropolysaccharide in Acid Gas Injection Jie Zhang, Gang Guo and Shugang Li 19.1 Introduction 19.2 Application of Heteropolysaccharide in C 0 2 Reinjection Miscible Phase Recovery 19.2.1 Test of Clay Polar Expansion Rate 19.2.1.1 Test Method 19.2.1.2 Testing results as the Figure 2 and Table 1 shows 19.2.2 Test of Water Absorption of Mud Ball in Heteropolysaccharide Collosol 19.3 Application of Heteropolysaccharide in H2S Reinjection formation 19.3.1 Experiment Process, Method and Instruction 19.3.1.1 Experiment Process 19.3.1.2 Experiment Method 19.3.1.2 Experiment Results 19.4 Conclusions References
361 361 363 364 364 366 367 370 370 370 370 372 373 373
Section 5: Geology and Geochemistry 20. Impact of S 0 2 and NO on Carbonated Rocks Submitted to a Geological Storage of C0 2 : An Experimental Study Stéphane Renard, Jérôme Sterpenich, Jacques Pironon, Aurélien Randi, Pierre Chiquet and Marc Lescanne 20.1 Introduction 20.2 Apparatus and Methods 20.2.1 Solids and Aqueous Solution 20.2.2 Gases 20.3 Results and Discussion 20.3.1 Reactivity of the Blank Experiments 20.3.2 Reactivity with pure S0 2 20.3.3 Reactivity with pure NO 20.4 Conclusion Acknowledgments References
377
377 378 379 380 381 381 384 387 391 392 392
CONTENTS
21. Geochemical Modeling of Huff 'N' Puff Oil Recovery With C 0 2 at the Northwest Mcgregor Oil Field Yevhen I. Holubnyak, Blaise A.F. Mibeck, Jordan M. Bremer, Steven A. Smith, James A. Sorensen, Charles D. Gorecki, Edward N. Steadman, and John A. Harju 21.1 Introduction 21.2 Northwest McGregor Location and Geological Setting 21.3 The Northwest McGregor Field, E. Goetz #1 Well Operational History 21.4 Reservoir Mineralogy 21.5 Preinjection and Postinjection Reservoir Fluid Analysis 21.6 Major Observations and the Analysis of the Reservoir Fluid Sampling 21.7 Laboratory Experimentations 21.8 2-D Reservoir Geochemical Modeling with GEM 21.9 Summary and Conclusions 21.10 Acknowledgments 21.11 Disclaimer References
22. Comparison of C 0 2 and Acid Gas Interactions with reservoir fluid and Rocks at Williston Basin Conditions Yevhen I. Holubnyak, Steven B. Hawthorne, Blaise A. Mibeck, David J. Miller, Jordan M. Bremer, Steven A. Smith, James A. Sorensen, Edward N. Steadman, and John A. Harju 22.1 Introduction 22.2 Rock Unit Selection 22.3 C 0 2 Chamber Experiments 22.4 Mineralogical Analysis 22.5 Numerical Modeling 22.6 Results 22.7 Carbonate Minerals Dissolution 22.8 Mobilization of Fe
XV
393
393 395 395 397 398 400 401 402 403 404 404 405
407
407 409 411 412 413 413 414 416
xvi
CONTENTS
22.9
Summary and Suggestions for Future Developments 22.10 Acknowledgments 22.11 Disclaimer References
418 418 418 419
Section 6: Well Technology 23 Well Cement Aging in Various H 2 S-C0 2 Flui(is at High Pressure and High Temperature: Experiments and Modelling Nicolas Jacquemet, Jacques Pironon, Vincent Lagneau, Jérémie Saint-Marc 23.1 Introduction 23.2 Experimental equipment 23.3 Materials, Experimental Conditions and Analysis 23.3.1 Cement 23.3.2 Casing 23.3.3 Environment 23.3.4 Exposures (Figure 3): 23.3.5 Analyses 23.4 Results and Discussion 23.4.1 Cement 23.4.2 Steel 23.5 Reactive Transport Modelling 23.6 Conclusion Acknowledgments References 24. Casing Selection and Correlation Technology for Ultra-Deep, Ultra- High Pressure, High H2S Gas Wells Yongxing Sun, Yuanhua Lin, Taihe Shi, Zhongsheng Wang, Dajiang Zhu, Liping Chen, Sujun Liu, and Dezhi Zeng 24.1 Introduction 24.2 Material Selection Recommended Practice 24.3 Casing Selection and Correlation Technology
423
424 425 426 426 427 427 427 427 428 428 430 430 432 433 434
437
438 438 441
CONTENTS
Casing Selection and match Technology Below 90°C 24.3.2 Casing Selection and Match Technology Above 90°C 24.4 Field Applications 24.4 Conclusions 24.5 Acknowledgments References
xvii
24.3.1
25. Coupled Mathematical Model of Gas Migration in Cemented Annulus with Mud Column in Acid Gas Well Hongjun Zhu, Yuanhua Lin, Yongxing Sun, Dezhi Zeng, Zhi Zhang, and Taihe Shi 25.1 Introduction 25.2 Coupled Mathematical Model 25.2.1 Gas Migration in Cement 25.2.2 Gas Migration in Stagnant Mud 25.2.3 Gas Unloading and Accumulation at Wellhead 25.2.4 Coupled Gas Flows in Cement and Mud 25.3 Illustration 25.4 Conclusions 25.5 Nomenclature 25.6 Acknowledgment References
442 443 443 445 447 447
449
449 450 451 452 454 456 458 459 460 461 461
S e c t i o n 7: C o r r o s i o n 26. Study on Corrosion Resistance of L245/825 Lined Steel Pipe Welding Gap in H 2 S+C0 2 Environment Dezhi Zeng, Yuanhua Lin, Liming Huang, Daijiang Zhu, Tan Gu, Taihe Shi, and Yongxing Sun 26.1 Introduction 26.2 Welding Process of Lined Steel Pipe 26.3 Corrosion Test Method of Straight and Ring Welding Gaps of L245/825 Lined Steel Pipe 26.4 Corrosion Test Results of Straight and Ring Welding Gaps of 1245/825 Lined Steel Pipe
465
466 466 467 472
xviii
CONTENTS
26.4.1 Atmospheric Corrosion Test Results 26.4.2 Corrosion Test Results at High Pressure 26.4.3 Field Corrosion Test Results 26.5 Conclusions 26.6 Acknowledgments References Index
472 472 474 477 477 477 479
Introduction The Three Sisters - CCS, AGI, and EOR Ying Wu1, John J. Carroll2 and Zhimin Du3 1
Sphere Technology Connection, Calgary, AB, Canada 2 Gas Liquids Engineering, Calgary, AB, Canada 3 Southwest Petroleum University, Chengdu, People's Republic of China
Although there remains some debate about whether or not man is changing the global climate and, if so, whether or not carbon dioxide is the cause of it, there is a significant capital, both political and financial, to reduce carbon emissions. It is not the purpose of this introduction, or this volume for that matter, to enter this debate. The purpose is to review the technology to achieve this and the inter-relations within available technologies. One of the main foci for reducing carbon emission is the so-called process, carbon capture and storage (CCS), removing carbon dioxide from combustion gases and storing them in subsurface formations. The main source of these combustion gases is coal-fired power plants, but other sources are targeted as well. In the petroleum and natural gas business there are two other mature technologies for injecting gas streams. The first of these is acid gas injection (AGI), and the other is injecting carbon dioxide for enhanced oil recovery (EOR). This makes CCS, AGI and EOR three sisters, of sorts. Whereas AGI and EOR are relatively mature processes, CCS is not and there is much those working in the CCS world can learn from both AGI and CCS. Table 1 summarizes the main components for the three technologies. Each of these will be discuss here. Whereas the impetus for acid gas injection is to eliminate sulfurous emissions, and there is little doubt about the effect of these emissions, they also sequester C0 2 . On the other hand, the purpose of injecting C 0 2 for EOR is to produce more oil. Burns [1], in a chapter in this volume discusses, the economics of an EOR xix
xx
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Table 1. The three sisters: CCS, AGI, and EOR. CCS
AGI
EOR i. virgin ii. recovered
Source of fluid
capture from flue gas
sweetening of natural gas
Pressure
compression
compression
compression
Pipeline
probably a network
commonly a single pipeline
pipeline network
commonly a single well
multiple, injection pattern (5-spot, for example)
disposal
oil recovery
reduced C 0 2 emissions
co2
Well
Purpose By-product
i. probably multiple wells ii. probably deviated wells to achieve high injectivity storage
sequestration
project. Nonetheless, sequestration of C 0 2 is a by-product of these EOR schemes. For CCS the purpose is simply to eliminate carbon emission into the atmosphere. However, C 0 2 captured from flue gas may have value as a source of virgin C 0 2 for EOR projects.
Capture The flue gas stream from a combustion process produces a flue gas that is from 5% to 15% carbon dioxide. The rest of this stream contains mostly nitrogen but also some oxygen and smaller amount of sulfur oxides and nitrogen oxides. The volume of the raw flue gas is too large to make compression and injection feasible. Thus the first step is to "capture" the C 0 2 from the flue gas. In the natural gas business the removal of carbon dioxide (and hydrogen sulfide for that matter) is called sweetening. Much of the technology developed over 75 years in the natural gas business can be transferred to the capture of C0 2 . However there are many
INTRODUCTION
xxi
problems associated with capturing C 0 2 that are not as common in the natural gas business. These include the low pressure of the flue gas stream (near atmospheric pressure versus tens of bars for natural gas) and the contaminants. Oxygen is poison to the common solvents used in the natural gas business. The chapter by Spooner and Engel [2] in this volume discusses the use of amine technology for capturing C 0 2 from flue gas. Among the problems Spooner and Engel address are the high oxygen content of the flue gas and the low pressure. In EOR there must be a source of carbon dioxide when the project begins. This is the so-called "virgin" C0 2 . Once the project starts, some of the C 0 2 will be produced with the oil. This C 0 2 is recovered from the oil and used for re-injection. Initially the recycled C 0 2 will be small but as the project matures this may become as large as 80% or 90% of the carbon dioxide injected.
Compression The next step for each of the three processes is to compress the stream to sufficient pressure such that it can be injected into a subsurface reservoir. In EOR the virgin C 0 2 is usually delivered at such a pressure that little or no compression is required. However the recycled C 0 2 is at low pressure and must be compressed for injection. In AGI the acid gas stream is at low pressure and in comes the sweetening process, where low pressure is used to regenerate the solvent. In acid gas injection and the compression of C 0 2 for EOR it is common to use compression and cooling alone to reduce the water content of an acid gas stream. The water holding capacity of acid gas was discussed in the previous volume in this series by Marriott et al. [3] and also by Satyro and van der Lee [4]. In a chapter in this volume Wright [5] discusses the use of compression and cooling in order to dehydrate an acid gas stream. In particular Wright addresses when dehydration is required and when it is not based on the composition of the gas and its water holding capacity. In some cases, compression alone cannot achieve sufficiently high pressures to inject the stream. In these cases, the stream can be liquefied (using a combination of high pressure and low temperature) and then pumped to higher pressure. Later in this
xxii
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
book Janusch and Braun [6] discuss the pumping of acid gas with diaphragm pumps.
Pipeline For all of the three sisters the compressed gas is transported via pipeline to the injection well(s). In an EOR project the compressed C 0 2 must be distributed through the oil filed such that the optimum oil recovery can be achieved. This requires a network of pipes. For small AGI projects usually only a single injection well is used and thus a single pipeline. However, for very large projects, AGI may require a network of line similar to an EOR project. The volumes injected in a typical CCS project will be very large and thus a single well is probably not an option.
Injection Again in each of the three sisters, the compressed fluid enters a well and travels downward to the target formation. In EOR it is common to have multiple wells arranged in a pattern, some for injecting C 0 2 and some for producing oil. It is also possible to use C 0 2 for huff 'n puff. This involves injecting C 0 2 for a period of time and then allowing the C 0 2 to soak (the "huff"). The same well is the used for producing the oil (the "puff"). Because of the properties of the gas injected and the phase behavior encountered, some unusual behavior can be observed in acid gas injection wells. Mirreault et al. [7] in the previous volume in this series, describe some seeming unusual behaviour in an injection well that have some relatively simple explanation.
Geochemistry The effect of the acid gas, and perhaps more specifically C0 2 , on the reservoir rock is an important consideration in the design of an injection scheme. How does the injected fluid affect the native rock? A case study related to the geochemical interactions is presented in this volume by Holubnyak et al. [8].
INTRODUCTION
xxiii
Summary The three sisters: CCS, AGI, and EOR share many common components. Many lessons can be shared especially between the more mature technologies of AGI and EOR and the newer one, CCS. These commonalities demonstrate that carbon capture and storage is a feasible technology. The remaining chapters in this volume discuss specific aspects of these three sisters and the reader should keep in mind the common aspects of these seemingly different technologies.
References 1. Burns, D. "Enhanced Oil Recovery Project: Dunvegan C Pool", Carbon Dioxide Sequestration and Related Technologies, Scrivener Publishing, Salem, MA. (2011). 2. Spooner, B. and D. Engel, "Carbon Capture Using Amine-Based Technology", Carbon Dioxide Sequestration and Related Technologies, Scrivener Publishing, Salem, MA (2011). 3. Marriott, R.A., E. Fitzpatrick, E Bernard, H. H. Wan, K. L. Lesage, P. M. Davis, and P. D. Clarke, "Equilibrium Water Content Measurements For Acid Gas Mixtures" Acid Gas Injection and Related Technologies, Scrivener Publishing, Salem, MA, (2011). 4. Satyro, M. and J. van der Lee, "The Performance of State of the Art Industrial Thermodynamic Models for the Correlation and Prediction of Acid Gas Solubility in Water", Acid Gas Injection and Related Technologies, Scrivener Publishing, Salem, MA, (2011). 5. Wright, W. "Dehydration-through-Compression: Is it Adequate? A Tale of Three Gases", Carbon Dioxide Sequestration and Related Technologies, Scrivener Publishing, Salem, MA, (2011). 6. Janusch, J. and A.-D. Braun, "Diaphragm Pumps improve Efficiency of Compressing Acid Gas and C0 2 ", Carbon Dioxide Sequestration and Related Technologies, Scrivener Publishing, Salem, MA, (2011). 7. Mireault, R., R. Stocker, D. Dunn, and M. Pooladi-Darvish, "Dynamics of Acid Gas Injection Well Operation", Acid Gas Injection and Related Technologies, Scrivener Publishing, Salem, MA, (2011). 8. Holubnyak, Y.I., S.B. Hawthorne, B.A. Mibeck, D.J. Miller, J.M. Bremer, S.A. Smith, J.A. Sorensen, E.N. Steadman, and J.A. Harju, "Comparison of C 0 2 and Acid Gas Interactions with Reservoir Fluid and Rocks at Williston Basin Conditions ", Carbon Dioxide Sequestration and Related Technologies, Scrivener Publishing, Salem, MA, (2011). 9. Taiman, S.J. and E.H. Perkins, "Concentration Gradients Associated With Acid Gas Injection", Acid Gas Injection and Related Technologies, Scrivener Publishing, Salem, MA, (2011).
SECTION 1 DATA AND CORRELATION
1 Prediction of Acid Gas Dew Points in the Presence of Water and Volatile Organic Compounds Ray. A. Tomcej Tomcej Engineering Inc. Edmonton, AB, Canada
Abstract Aromatic hydrocarbons which are present in sour natural gas streams can be absorbed into the amine treating solution at the bottom of the contactor and exit in the rich amine stream. Depending on the process configuration, these dissolved hydrocarbons can end up in the acid gas leaving the amine regenerator. In acid gas injection facilities, trace amounts of heavy hydrocarbons in the acid gas may lead to the formation of a sour hydrocarbon liquid phase in the compressor interstage scrubbers. In this exploratory work, a cubic equation-of-state (EOS) model was used to make predictions of non-aqueous (Lj) dew points in acid gas systems. The objective was to develop a better understanding of the conditions under which this phenomenon can occur, and to reinforce the need for accurate experimental vapor-liquid equilibrium data to support cost effective design and model development.
1.1
Introduction
Benzene, toluene, ethyl benzene and xylene isomers are commonly referred to collectively as BTEX compounds. These compounds are known to be toxic to humans and their containment and disposal are of special interest to the hydrocarbon industry. BTEX environmental contamination is often linked to leakage from underground gasoline storage tanks or accidental spills. Awareness of this toxicity led to regulated clean air emission standards that directly impact Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (3-12) © Scrivener Publishing LLC
3
4
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
the natural gas processing industry as trace amounts of BTEX compounds are associated with produced fluids such as natural gas. Sour gas production generally involves a subsequent processing step in which the hydrogen sulphide (H2S) and carbon dioxide (C0 2 ) are removed to produce an acid gas stream that may be a candidate for acid gas injection. Liquid solvents that are used to remove the H2S and C 0 2 from the gas stream are often aqueous solutions of organic chemicals that have a high affinity for the BTEX compounds. Distribution of the BTEX compounds within the various streams of a natural gas processing plant is a complex phenomenon involving many interrelated process variables such as operating pressures and temperatures, amine composition, amine circulation rates, and others. Of particular interest in acid gas injection, is the amount of BTEX compounds that end up in the acid gas product leaving the amine regenerator. The presence of trace quantities of BTEX compounds in the acid gas, if unaccounted for at the design stage, may lead to the unexpected formation of a sour non-aqueous liquid phase in the compressor train, and considerable operational difficulties. The objective of this work was to develop a better understanding of the conditions under which this phenomenon can occur, and to reinforce the need for accurate experimental vapor-liquid equilibrium data to support cost effective design and model development.
1.2
Previous Studies
In order to estimate the levels of BTEX compounds that will be present in the acid gas, there is a need for accurate vapor-liquid equilibria (VLE) a n d / o r vapor-liquid-liquid equilibria (VLLE) data for BTEX and similar hydrocarbons in amine treating solutions under rich amine conditions. Operating plant data are also useful to verify the predictions of any thermodynamic model. Ng et al. (1999) provided an overview of specific phase equilibria data and physical properties that are required for reliable design of acid gas injection facilities. Hegarty and Hawthorne (1999) presented valuable operating data for a Canadian gas plant using MDEA in which measured BTEX compositions were reported. Mclntyre et al. (2001) and Bullin and Brown (2004) tabulated the experimental data available for hydrocarbon and BTEX solubility in amine treating solutions and demonstrated general trends
PREDICTION OF ACID GAS DEW POINTS
5
in amine plant BTEX absorption using computer simulation. Valtz et al. (2002) presented a comprehensive set of fundamental solubility data for aromatic hydrocarbons in aqueous amine solutions. Miller and Hawthorne (2000) and Jou and Mather (2003) measured the solubility of BTEX compounds in water. Clark et al. (2002) measured bubble and dew points for a nominal 10 mol% H 2 S/90 mol% C 0 2 mixture and regressed an equation of state to match the phase envelope. Satyro and van der Lee (2009) demonstrated that with suitable modification to interaction parameters, a cubic equation of state can provide reliable predictions of phase behavior in sour gas mixtures.
1.3
Thermodynamic Model
A rigorous treatment of the complex phase behavior in the H 2 S-C0 2 water-BTEX system was beyond the scope of this work, which was intended to be exploratory in nature. The Peng-Robinson equationof-state with classical van der Waals mixing rules was used in this study. The interaction parameter for the H 2 S-C0 2 binary was set to 0.1 and all others were set to zero. Table 1 contains the critical properties used for the system components. Table 1. Component critical properties. Critical P, kPa
Critical T, °C
Hydrogen Sulphide
9007.8
100.45
Carbon Dioxide
7386.6
31.05
Benzene
4898.0
Toluene
Component
Acentric Factor
Molecular Weight
0.1
34.076
0.225
44.01
289.0
0.2092
78.112
4105.8
318.7
0.2637
92.138
Ethyl Benzene
3605.9
344.1
0.3026
106.165
o-Xylene
3734.2
357.2
0.3118
106.165
m-Xylene
3536.3
343.9
0.3255
106.165
p-Xylene
3510.8
343.1
0.3211
106.165
6
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
The performance of the Peng-Robinson equation of state has been well documented in the literature. The model reproduced the dew point locus of Clark et al. (2002) to within 2.5%.
1.4
Calculation Results
The conditions of the calculations were chosen to encompass those normally found in acid gas injection compression: pressures from 150 kPa to 10 MPa, and temperatures above the hydrate formation curve from 0° to 100°C. Three different nominal acid gas compositions were considered: 20/80, 50/50, and 80 mole% H 2 S/20 mole% C O r Hydrocarbon components studied included: benzene, toluene, ethyl benzene and dimethyl benzenes (xylenes). The model was used to generate the phase envelope for each of the three nominal acid gas compositions. The influence of associated water on the location of the bubble and dew-point loci was not considered in this work. A typical injection profile was generated for each nominal composition using a starting pressure of 150 kPa and constant compression ratio. Temperatures in the compression process were restricted to remain under 150°C. Cooling temperature was set to 50°C. The final pressure was selected to be under 10 MPa but above the mixture critical point. Initial calculations indicated that the phase behavior of the acid gas mixtures in the presence of each of the three xylene isomers was similar. For simplicity only o-xylene was considered in this study. To establish a reasonable range of BTEX compositions, a sensitivity study was undertaken using pure H2S. The model was used to determine the L^ dew point temperature at 4000 kPa using various compositions of benzene and o-xylene ranging from 0 to 5000 ppmv. The results are shown in Figure l. 1 Below concentrations of 100 ppmv, the aromatic compounds increase the dew point temperature by less than 1°C. Hegarty and Hawthorne (1999) reported BTEX content of up to 2500 ppmv in the acid gas of an operating MDEA plant. Using this as a guideline, non-aqueous liquid (L,) dew points were calculated for each of the three nominal acid gas compositions with 500-, 2000- and 5000 ppmv of each of the four aromatic compounds.
1
Figures 1 through 4 appear at the end of this paper.
PREDICTION OF ACID GAS DEW POINTS
7
Figure 1. Effect of BTEX compounds on L, dew point in pure H2S.
Figure 2. Effect of BTEX compounds in 80% H2S - 20% C O r
Clearly this range of calculated points generated a significant amount of data. The results for the 2000 ppmv cases are presented in Figures 2 through 4 and provide an adequate representation of the general trends that were observed. Note that curves labeled as organic compounds represent the dew point loci for the acid gas mixture with 2000 ppmv of only that organic compound.
8
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 3. Effect of BTEX compounds in 50% H2S - 50% C0 2 .
Figure 4. Effect of BTEX compounds in 20% H2S - 80% C0 2 .
Using data from Mclntyre et al. (2001) for BTEX component distribution in the acid gas from an MDEA plant as a guideline, flash calculations were performed at 50°C for the mixture given in Table 2. Identical calculations were performed for a mixture containing 80 mol% H2S and 20 mol% C O r The results are shown in Table 3.
PREDICTION OF ACID GAS DEW POINTS
Table 2. Composition of mixture used for condensation study. Composition, mol %
Component Hydrogen Sulphide
79.82
Carbon Dioxide
19.955
Benzene
1000 ppmv
Toluene
750 ppmv
Ethyl Benzene
250 ppmv
o-Xylene
250 ppmv
Table 3. Condensation study results at 50°C. Pressure, kPa 3268.3
Volume% Lj, BTEX Mixture
Volume %Lj, 80/20 H2S/
co 2
Dew point P
3400
0.009
0
3600
0.031
0
3800
0.074
0
4000
0.169
0
4200
0.420
0
4400
1.29
0 Dew point P
4466.6 4600
3.90
2.46
4800
8.55
7.29
5000
15.6
14.2
5200
26.6
24.8
5400
45.4
42.5
5600
82.8
77.2
5654.1 5674.5
Bubble point P Bubble point P
9
10
1.5
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Discussion
In the absence of experimental data for dew point conditions in acid gases with contaminants, there can be no absolute conclusions drawn on the accuracy of the predictions. This exploratory study clearly emphasizes the importance of experimental research to provide fundamental information for process design and advanced model development. The results in Figures 2 through 4 illustrate that with conservative cooling temperatures and with BTEX contaminant levels in the range of those already measured in an operating MDEA plant, it is possible to enter the three-phase region in the higher pressure interstage coolers and separators in acid gas injection facilities. More aggressive cooling escalates the potential for three-phase conditions. The formation of a second liquid phase in the compression interstage cooling system, in itself is not a problem, provided that the phase behavior phenomenon is understood at design time. The Lj phase is less dense than water, contains u p to 20 mol% BTEX and, if formed, will accumulate in the interstage separators. As pointed out by Hegarty and Hawthorne (1999), it is extremely important to obtain an accurate inlet gas composition, including an extended analysis of the C6+ fraction to determine the aromatic content. Once the BTEX content, if any, is identified it can be accounted for in any process design, modeling, or operational troubleshooting of downstream processes such as acid gas injection. In spite of the purely predictive nature of the calculated results, the following general observations can be made by analyzing Figures 2 through 4. The same behavior is observed in the 500 ppmv and 5000 ppmv calculated results. • At a given pressure, the presence of BTEX compounds in acid gas widens the phase envelope, with this effect being more pronounced in acid gases with higher C 0 2 content. • At a given pressure, the presence of BTEX compounds in acid gas increases the L^ dew point temperature, with this effect being more pronounced in acid gases with higher H2S content. This is, in part, a result of the shift of the acid gas phase envelope to higher temperatures in high H2S mixtures.
PREDICTION OF ACID GAS DEW POINTS
11
• At equal concentration in the acid gas and at equal pressure, BTEX compounds increase the L] dew point temperature in the order: benzene, toluene, ethyl benzene and o-xylene with o-xylene having the most pronounced effect. • In all cases, the possibility of non-aqueous Lj formation is highest in the separator before the final stage of compression. • If compressed acid gas is cooled to lower temperatures (e.g. 30°C) in the compressor facility, this increases the possibility of Lj formation. • If BTEX compounds are present in the acid gas at levels less than 100 ppmv, the acid gas dew point locus is relatively unaffected. The dew point loci shown in Figures 2 through 4 indicate where the first droplet of L^ forms. Table 3 contains an example of the condensation behavior inside the phase envelope at constant temperature. Note that the condensation behavior of the BTEX mixture is similar to the BTEX-free system except for the deep depression of the dew point pressure. Lines of constant liquid volume % are widely spaced in this region of the phase envelope. This behavior is similar to the condensation behavior of rich gas systems. The location of the bubble point is relatively unaffected by the organic compounds.
References Bullin, Jerry A. and William G. Brown, "Hydrocarbons and BTEX Pickup and Control from Amine Systems", Proceedings of the 83rd Gas Processors Association Annual Convention, New Orleans, March 14-17,2004. Clark, M.A., W.Y. Svrcek, W.D. Monnery, A.K.M. Jamaluddin and E. Wiehert, "Acid Gas Water Content and Physical properties: Previously Unavailable Experimental Data for the Design of Cost Effective Acid gas Disposal Facilities, and Emission Free Alternative to Sulfur Recovery Plants", Hycal Energy Research Laboratories, 2002. Hegarty, Mike and Dean Hawthorne, "Application of BTEX/Amine VLE Data at Hanlan Robb Gas Plant", Proceedings of the 78th Gas Processors Association Annual Convention, Nashville, March 1-3,1999. Jou, Fang-Yuan and Alan E. Mather, "Liquid-Liquid Equilibria for Binary Mixtures of Water+Benzene, Water+Toluene and Water+p-Xylene from 273K to 458K", /. Chem. Eng. Data, 48, 750-752(2003)
12
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Mclntyre, G.D., V.N. Hernandez-Valencia and K.M. Lunsford, "Recent GPA Data Improves BTEX Predictions for Amine Sweetening Facilities", Proceedings of the 80th Gas Processors Association Annual Convention, San Antonio, March 12-14,2001. Miller, David J. and Steven B. Hawthorne, "Solubility of Liquid Organics of Environmental Interest in Subcritical (Hot/Liquid) Water from 298K to 473K", /. Chem. Eng. Data, 45, 78-81(2000). Ng, Heng-Joo, John J. Carroll and James Maddocks, "Impact of Thermophysical Properties Research on Acid Gas Injection Process Design", Proceedings of the 78th Gas Processors Annual Convention, Nashville, March 1-3,1999. Satyro, Marco A. and James van der Lee, "The Performance of State of the Art Industrial Thermodynamic Models for the Correlation and Prediction of Acid Gas Solubility in Water", Proceedings of the First International Acid Gas Injection Symposium, Calgary, Alberta, Canada, October 5-6, 2009. Valtz, A., P. Guilbot and D. Richon, "Amine BTEX Solubility", Gas Processors Association Research Report RR-180, 2002.
2
Phase Behavior of China Reservoir Oil at Different COJnjected Concentrations Fengguang Li, Xin Yang, Changyu Sun, and Guangjin Chen State Key Laboratory of Heavy Oil Processing, China University of Petroleum Beijing, People's Republic of China
Abstract The phase behavior of China reservoir oil at different C 0 2 injected concentrations has been studied at the temperature of 339.2 K using a high-pressure PVT unit. Seven groups of reservoir fluids with C 0 2 molar contents of 0, 10.0, 34.1, 44.7, 48.9, 57.8, and 65.0 mol% have been prepared. Saturation pressure of reservoir fluids at seven C 0 2 injected contents were measured. The reservoir oil density and viscosity at different pressures under reservoir temperature were also obtained. The influence of C 0 2 molar contents on the interfacial tension of C 0 2 injected reservoir oil under stratum conditions was determined using a pendant drop method. The experimental data indicated that when C 0 2 content is lower than 45 mol%, the increase of bubble point pressure is slow. After that, the bubble point pressure value increases more sharply with the increase of C 0 2 molar concentrations. The reservoir viscosities decrease sharply with the increase of C 0 2 concentration when the system pressure is above the bubble point for different injection contents. The experimental results of interfacial tension for C 0 2 injected crude oil/stratum water show that it decreases with the increase of C 0 2 injected concentrations. The pressure has a slight effect on the interfacial tension value. These phase behavior data will be helpful for evaluating the effect of C 0 2 injected method to enhance oil recovery.
Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (13-22) © Scrivener Publishing LLC
13
14
2.1
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Introduction
The fluid phase behavior study is used as an important basis for miscible-slug process and predominant displacement mechanism, which is of critical importance during the miscible displacement process (1). The conventional fluid phase behavior test is usually conducted using PVT (Pressure-Volume-Temperature) unit. It is of great concern in many high-pressure technologies, such as fluid extraction process, exploration of near-critical gas condensate/ volatile oil reservoir, and gas-injected enhanced oil recovery processes. C 0 2 displacement technology is recognized as a significant and well-established means for oil and gas enhanced recovery both at home and abroad. Miscible gas injection could minimize the trapping effect of capillary forces and is recognized as an economic enhanced oil recovery process. Although some PVT fluid phase behavior data are available in the published papers, they are still insufficient because of the complexity of multi-component reservoir fluid. In this work, the phase behavior of China reservoir fluids collected from Jilin oil field were analyzed at different C 0 2 injected concentrations and pressures using a high-pressure PVT device. The density, bubble point pressure, viscosity, and interfacial tension properties of reservoir fluid at different C 0 2 injected mole percents and pressures under the stratum temperature were systematically measured.
2.2
Preparation of Reservoir Fluid
The reservoir fluid sample was collected from China Jilin oil field at reservoir conditions. The stratum temperature was 339.2K. The reservoir fluid arriving from the well was separated and flashed to standard condition. The molar composition of reservoir fluids was then obtained from analysis of the gas and oil samples. The gas phase was analyzed by HP6890 gas Chromatograph. The liquid phase was analyzed by simulating distillation process using HP5890A. Afterwards, the reservoir fluid composition was obtained by combining the gas and liquid phase compositions using the gasoil ratio (GOR). The measured composition for reservoir fluid was shown in Table 1. Molecular weights of the oil phase were determined by vapor pressure osmometer (VPO) and the determined molecular weight was 420 g/mol.
PHASE BEHAVIOR OF CHINA RESERVOIR O I L
15
Table 1. The composition of reservoir fluid. Gas phase (mol%)
Oil phase (mol%)
Reservoir fluid (mol%)
N2
2.491
0.968
co2
0.190
0.074
CH 4
61.921
24.059
C2H6
9.585
3.724
C3H8
11.226
4.361
i"C 4 H 10
1.721
0.669
n-C 4 H 1 0
6.983
2.713
i"C 5 H 12
1.301
0.505
n-C 5 H 1 2
2.721
1.057
C
6H14
1.861
0.723
C
7H16
0.884
0.540
C
8H18
2.998
1.833
C
9H20
2.178
1.332
C
10H22
2.980
1.823
90.960
55.619
C1I+
Seven groups of C 0 2 injected concentration (including 0% C0 2 ) were chosen to study the reservoir fluid behavior under gas injection process. The C 0 2 injected crude oil was prepared using RUSKA PVT device.
2.3
PVT Phase Behavior for the C 0 2 Injected Crude Oil
Phase behavior of China reservoir oil was systematically investigated using a RUSKA high-pressure PVT system which was described in our previous papers (2,3). The PVT data at different
16
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
C 0 2 injected molar components was measured to build the relationship between the volume and pressure of reservoir oil. The bubble point pressure and density of reservoir fluid at different pressures could then be determined according to the measured PVT data, which is useful to calculate the phase behavior properties such as the relatively volume, solubility of injected C 0 2 in oil, and so on. The density of the C 0 2 injected reservoir fluid at different pressure under the strata temperature was plotted in Figure 1. From Figure 1, it can be found that there exists an inflexion for the curve of reservoir fluid density and pressure, showing the process of phase transition. When the C 0 2 concentration achieves 65.0 mol%, there is no significant difference between the gas phase and liquid phase, showing that there may exhibit first contact miscibility condition when the C 0 2 injected contents is 65.0 mol%. The bubble point pressure at seven C 0 2 molar compositions determined from PVT measurement was shown in Figure 2. According to Figure 2, it shows that bubble point pressure increases with the increase of C 0 2 injected concentrations. When C 0 2 content is lower than 45mol%, the increase of bubble point pressure is slow. However, when C 0 2 content is higher than 45mol%, the bubble point pressure value increases more sharply with the increase of C 0 2 molar concentrations. The bubble point pressure data is also used to choose the suitable C 0 2 injected concentration.
Figure 1. Variation of reservoir oil density for C 0 2 injected crude oil at different C 0 2 mole percents and pressures.
PHASE BEHAVIOR OF CHINA RESERVOIR O I L
17
Figure 2. Bubble point pressure at different C0 2 injected concentrations for China reservoir crude oil.
2.4 Viscosity of the C0 2 Injected Crude Oil Viscosity is an important transport property in petroleum production and processing. RUSKA falling ball viscometer connected with RUSKA high-pressure PVT device was used in this work to investigate the viscosity of China Jilin oil samples after different C 0 2 content was injected under stratum conditions. The basic principle of falling ball viscometer is based on Stokes law. The fluid viscosity could be exactly calculated by Stokes law according to the time of the ball travels through internal pipe from the top to the bottom. If the falling ball behaves to be laminar flow, the following equation was used: p = kt(pB-pF)
(1)
where pB and pF are the density of the ball and fluid, respectively. t is the travel time. A: is a constant value related to the diameter of the falling ball and the angel of the apparatus. Before the experiment, a falling ball was selected to measure the constant value k in Eqn. (1) using standard silicon oil for the viscometer. Thereafter, the reservoir crude oil viscosities were systematically measured with the same calibrated ball at different C 0 2 injected molar concentrations and pressures. The reservoir fluid viscosity was tested from higher pressure under single phase conditions until close to the
18
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
saturation pressure. After the pressure was lower than the bubble point pressure, a gas exhaust valve was open to slowly reduce to the experimental pressure and the stable time was prolonged to 4-5 h. The measured viscosity for C 0 2 injected crude oil at different C 0 2 mole percents and pressures were plotted in Figure 3. As shown in Figure 3, the viscosity for C 0 2 injected crude oil decreased apparently with increasing of C 0 2 content. When the C 0 2 injected amount changed from 0 to 65.0 mol%, the reservoir oil viscosity value decreased greatly. At about 30 MPa, the viscosity value can decrease from 10.6 cP to 1.1 cP when 65 mol% C 0 2 was injected. It can be found that when the experimental pressure is higher than the saturated value, the reservoir oil viscosity increases with the increase of pressure; When it is lower than the saturated pressure, the reservoir oil viscosity increases with the decrease of pressure. With the decrease of pressure, more C 0 2 was released from the reservoir oil and induced the increase of viscosity of the residual oil. From Figure 3, it can be concluded that C 0 2 injecting is significant in favor of the decrease of viscosity of Jilin reservoir crude oil. However, after C 0 2 content was higher than 44.7 mol%, the reservoir oil viscosity at single phase condition does not decrease significantly with the further increase of C 0 2 injecting concentration. Meanwhile, During the C 0 2 injecting concentration increases from 0 to 44.7 mol%, the bubble point pressure only increases from 11.28 MPa to 14.14 MPa. However, when the C 0 2 injected concentration
Figure 3. Variation of viscosity for C 0 2 injected crude oil at different C 0 2 mole percents and pressures.
PHASE BEHAVIOR OF CHINA RESERVOIR O I L
19
increases from 44.7 mol% to 65.0 mol%, the bubble point pressure increases from 14.14 MPa to 25.0 MPa. Therefore, from the view of decrease of viscosity and bubble point pressure, there exists a suitable C 0 2 injecting concentration and high C 0 2 concentration is not needed.
2.5 Interfacial Tension for C0 2 Injected Crude Oil/Strata Water A great amount of reservoir water exists in the stratum after water displacement process of oil field. There is a special need for accurate interfacial tension estimation because the movement of reservoir fluids is influenced to a great extent by capillary forces. The C 0 2 injected concentration also plays an important role on the interfacial phenomena. In this work, the influence of C 0 2 molar contents on the interfacial tension of injected crude oil/water was systematically investigated using the JEFRI pendant drop high-pressure interfacial tension apparatus manufactured by D.B.Robinson (Canada), which the maximum working pressure is 34.5 MPa (5,000 psi) and the operating temperature range is 253-473 K. The experimental apparatus and procedures were detailed described in our previous papers (4,5). The interfacial tension measurement is based on the following principle: If the drop is in equilibrium with its surroundings gas, the interfacial tension (y) values can be calculated directly from an analysis of the stresses in the static, pendant drop, using the following equations developed by Andreas et al. (6): = ApDe2g/H
(2)
l/H = f(ds/de)
(3)
7
where Ap is the density difference between the two phases, De is the unmagnified equatorial diameter of the drop, g is the gravitational constant, ds is the diameter of the drop at a selected horizontal plane at height equal to the maximum diameter de. Andreas et al. have prepared a detailed table of 1/H as a function (djd). The difference in density between reservoir oil and water could be calculated from the measured density data. The interfacial tension of C 0 2 injected crude oil /reservoir water were all measured
20
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
under single-phase conditions at the stratum temperature. The measured interfacial tension data for C 0 2 injected reservoir oil/ water at different C 0 2 injected molar concentrations and pressures are plotted in Figure 4. As shown in Figure 4, the interfacial tension for C 0 2 injected oil/reservoir water decreased apparently with the increase of C 0 2 injected molar concentration when C0 2 content varies from 0 to 65.0 mol%. The dissolvability of C 0 2 in oil has a significant influence on the interfacial tension value. The interfacial tension decreased by about one-third as the C 0 2 injected amount changed from 0 to 65.0 mol%. It also shows that the interfacial tension of the C 0 2 injected crude oil/water increased with increasing pressure. During the experiment process, the experimental pressure was always higher than the bubble point pressure at the corresponding C 0 2 injected condition. Compared with the effect of C 0 2 injected amounts, the pressure has only a slightly effect. When the C 0 2 composition was 65.0 mol%, the C 0 2 injected oil system approached complete irascibility and the interfacial tension data of C 0 2 injected crude oil/ reservoir water changed a little with an increase in pressure.
2.6
Conclusions
The phase behavior of reservoir oil collected from China Jilin oil field was systematically investigated by using a high-pressure RUSKA PVT device at different C 0 2 injected concentrations and
Figure 4. Variation of interfacial tension for C 0 2 injected oil/reservoir water at different C 0 2 mole percents and pressures.
PHASE BEHAVIOR OF CHINA RESERVOIR O I L
21
pressures under strata temperature. Seven groups of C 0 2 injected concentrations varying from 0 to 65.0 mol% were prepared. The bubble point pressure increases from 11.28 MPa to 25.0 MPa when C 0 2 content increases from 0 to 65.0 mol%. When the C 0 2 concentration achieves 65.0 mol%, there is no significant difference between the gas phase and liquid phase, showing that there may exhibit first contact miscibility condition under the corresponding C 0 2 injected content. The viscosity for C 0 2 injected crude oil decreased apparently with increasing of C 0 2 content. C 0 2 injecting is significant in favor of the decrease of viscosity of Jilin reservoir crude oil. However, after C 0 2 content was higher than 44.7 mol%, the reservoir oil viscosity under single phase condition does not decrease significantly with the further increase of C 0 2 injecting concentration. The interfacial tension for C 0 2 injected oil/reservoir water decreased apparently with the increase of C 0 2 injected molar concentration when C 0 2 content varies from 0 to 65.0 mol%. When the C 0 2 composition was 65.0 mol%, the C 0 2 injected oil system approached complete miscibility and the interfacial tension data of C 0 2 injected crude oil/reservoir water changed a little with an increase in pressure.
Literature Cited 1. W. Yan, L.K. Wang, L.Y. Yang, T.M. Guo, Fluid Phase Equilibria, Vol. 190, p. 159-178, 2001. 2. M.X. Gu, Q. Li, X.Y Zhou, W.D. Chen, T.M. Guo, Fluid Phase Equilibria, Vol. 82, p. 173-182,1993. 3. H.Q. Pan, T. Yang, T.M. Guo, Fluid Phase Equilibria, Vol. 105, P. 259-271,1995. 4. C.Y. Sun, G.J. Chen, L.Y. Yang, /. Chem. Eng. Data, Vol. 49, p. 1023-1025,2004. 5. C.Y. Sun, G.J. Chen, /. Chem. Eng. Data,, Vol. 50, p. 936-938, 2005. 6. J.H. Andreas, E.A. Hauser, W.B. Tucker, /. phys. Chem., Vol. 42, p. 1001-1019, 1938.
This page intentionally left blank
3
Viscosity and Density Measurements for Sour Gas Fluids at High Temperatures and Pressures B.R. Giri, P. Biais and R.A. Marriott* Alberta Sulphur Research Ltd. Department of Chemistry University of Calgary Calgary, AB, Canada
Abstract Designing an acid gas injection scheme requires an accurate knowledge of the density and viscosity of the injected fluid as these properties are used to optimize compression, monitor transportation and model gas mobility in the reservoir. Fit-for-purpose models are developed based on the available literature data, which in some instances are either inaccurate or studied at industrially irrelevant temperatures and pressures. Moreover, the errors for predicted data at high pressures and temperatures can be as large as 20-50%. An extensive literature search by Schmidt et al. [1] revealed that there are limited data for H2S and its mixtures available in the literature; most of which are limited to gaseous H2S and saturated liquids. The only existing data that extend to higher pressures (p = 10 to 50 MPa) and temperatures (T = 115 to 140°C) are from Monteil et al. [2] which were reported in the late 60's, after which no measurements appeared to have been carried out. Expansion of the literature data to fill the void temperature and pressure regions, especially at relevant conditions for acid gas injection schemes (T = 0 to 150°C and p = 0.1 to 75 MPa) are desired so that the discrepancies of existing data sets can be resolved and reference viscosity models can be further tested and parameterised. It is worthwhile to note that during the recent development of the H2S viscosity model of Schmidt et al., [1] the data set from Monteil et al. [2] was excluded due to inconsistency. This further demonstrates the importance of additional experimental studies for the determination of H2S viscosity and density at elevated pressures and temperatures. Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (23-40) © Scrivener Publishing LLC
23
24
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
We have recently begun an experimental program aimed at measuring the high-pressure densities and viscosities of H2S and other acid gas mixtures using an Anton Parr vibrating tube densimeter and a Cambridge oscillating piston viscometer at p = 1 to 100 MPa and T = 0 to 150°C. This paper discusses how these instruments were commissioned, calibrated and operated. Interim C0 2 , CH 4 and H2S results show the accuracy and reproducibility of the high-pressure measurements.
3.1
Introduction
Design of an acid gas injection (AGI), sour gas injection or C 0 2 injection scheme requires that the density and viscosity properties of the fluid be well known [1,3,4]. From pre-compression to the reservoir, the viscosity is required to assess frictional pressure drops and the density is required to calculate pressure gains due to static head. Expansion of the literature data to fill the applicable temperature and pressure regions, especially at relevant conditions for AGI schemes are desired so that the discrepancies within existing data sets can be resolved and models can be further parameterised. While density and viscosity properties have been well studied for pure C 0 2 and methane, the data for H2S are sparse at industrially relevant conditions, particularly H2S viscosities at higher pressures [1]. A notable exception is Monteil et al. [2] who have reported some H2S viscosities at high pressures (p = 10 to 50 MPa; T = 115 to 140°C). However, it should be noted that, recently Schmidt et ah, [1] have excluded the data set from Monteil et al. [2] due to inconsistency. In order to determine the range of conditions which would be applicable to industry we considered that acid gas streams moving through traditional compression cycles involve a broad range of temperatures from T = 0 to 150°C. For examples of applicable pressures, Mireault et al. [5] have used pressures of 30 MPa for a target aquifer and 3 MPa for a targeted depleted reservoir. If the fluid is being used for reservoir pressure maintenance, one can expect even larger target reservoir pressures (p > 700 MPa). Thus the overall temperature and pressure ranges targeted by our research includes T = 0 to 150°C and p = 0.1 to 100 MPa. Within this range of conditions we intend to measure a variety of pure acid gas components and acid gas mixtures, beginning with C0 2 (calibrant), CH 4
SOUR GAS FLUIDS AT H I G H TEMPERATURES AND PRESSURES
25
and H2S. The large temperature and pressure ranges outlined is experimentally challenging in large part because of fluids involved. Pure C 0 2 and H2S are gaseous, liquid and supercritical within these conditions; therefore, measurements must cover a range from p « 10 to 1200 kg m 3 and 77 « 10 to 300 uPa s (0.01 to 0.30 cP). So far we have gained enough data to estimate the accuracy of our new high pressure instruments, vibration tube densimeter (VTD) and oscillating cylinder viscometer (Cambridge). This paper discusses our experimental methods, some preliminary data for C 0 2 (calibrant), CH 4 and H2S; and provides some evaluation of the instruments capabilities.
3.2
Experimental
3.2.1 Density Measurement There are several methods for accurately measuring densities at high pressure. Providing the mass of the fluid can be measured with high confidence, isochoric vessels with good pressure measurement and stable temperature control are simple and have yielded high quality results in all fluid regions [6-8]. With the isochoric method the vessel can be heated to desired temperatures and the resulting pressure measured. A second vessel can be used for controlled isothermal fluid expansion (Burnett Expansion) [9,10]. Another accurate method includes measuring the buoyancy of a sinker, or better yet two sinkers, which are completely immersed in a high-pressure fluid [11,12]. Vibrating Tube Densimeters, VTDs, have the advantage of a small volume, applicability over a wide range of densities, typically p = 1 to 2000 kg m 3 , and they can be used to measure densities for static or flowing fluids. VTDs have long been used by the brewing and distillation industry to quantify alcohol content [13]. The precision of the VTD technique was improved in 1974 by Picker et al. [14] and extended to high pressure in 1984 by Albert and Wood [15]. Provided they are coupled with good temperature and pressure control, they can yield accurate results up to very high pressures. High pressure VTDs have been commercialized by Anton Paar (DMA-HDT and DMA-HPM). In this study densities were measured using an Anton Paar DMA HPM vibrating tube densimeter, VTD. The densimeter
26
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
has Hastelloy C-276 wetted parts, a stated temperature range of T = -10 to 200°C and a pressure limit of p = 140 MPa. A vibrating tube densimeter can be theoretically described using the undamped resonance frequency of a simple harmonic oscillator, co:
(1) where k is the spring constant and m is the mass of the vibrating system consisting of the tube, mt, and the fluid inside the tube, ma or mb. At a specific temperature and pressure, changing the internal fluid from a to b results in a system mass change of ma - my which can be related to the difference in density, pa - pb:
k 2
2
= k'tf-Tt),
(2)
where T.=2jt/COK is the time period of oscillation for the tube containing fluid / and k' is the calibration constant for the instrument. The calibration constant can be determined by measuring the time period for two fluids of well known density. Due to thermal expansion and compressibility of the vibrating tube, the calibration constant, k', of Equation 2 is both temperature and pressure dependent. Because the temperature is reproducible to within 0.01 °C, isothermal calibrations have been determined at T = 0,50,100 and 150°C and from p = 0.09 to 100 MPa. The isothermal expression used for calculating the density was Pr,a = K (P) ■ {rf,a
- 4,b ) + PT,b
O)
For Equation 3, pTa is the density of fluid a, x\A is the oscillation time period of the tube filled with fluid a and x\h is the oscillation time period for the tube filled with air at laboratory pressure. For each temperature, a simple linear expression, kT(p) = c + dp, was fit by least square regression using the time period of oscillation for a = C 0 2 [16] (p = 1, 2, 5,10, 20, 50 and 100 MPa) and air at 0.09 MPa (atmospheric pressure in Calgary).
SOUR GAS FLUIDS AT H I G H TEMPERATURES AND PRESSURES
3.2.2
27
Viscosity Measurement
There are many potential high pressure viscosity techniques to choose from, such as a capillary viscometry, falling body viscometers (e.g., Stokes, rolling ball, falling piston, etc.) or oscillating viscometers (oscillating disc, vibrating wire, etc.). Unlike several of the densimeter techniques, most viscometers are built to measure liquid or gaseous viscosity and are rarely designed for a large range of viscosity. Wakeham et al. [17] have recently published a review on the development for some of these techniques. Some common instruments will be briefly discussed here. The high pressure capillary viscometer is similar to the commonly used Ostwald viscometers (u-tube) which are often used for liquids under gravity flow and normally at atmospheric pressure. Rather than gravity flow, most high-pressure capillary viscometers use pistons to drive fluids through a capillary tube either at constant flow (measuring the difference in pressure) or at constant pressure difference (measuring the flow). Through Poiseuille's law for steady state fluid flow, the viscosity can be calculated. Capillary viscometers can be adapted for both liquid and gaseous fluids by changing the size of the capillary line (length a n d / o r internal diameter). A common experimental issue is the low tolerance for small particles which can obstruct flow. Falling body or sinker type viscometers can include falling ball, falling piston and rolling ball viscometers. In general they all involve some object falling through a static fluid under constant gravitational force with an opposing drag. The falling ball and/or falling piston viscometer measurement was originally conceived by Stokes [18] and applied within the work of Flowers [19]. The accuracy of the viscometer depends on the accuracy of the velocity measurement, i.e., the travel time measurement for the object to traverse some known distance. In order to optimize the elapse time the falling object's density (buoyancy) can be changed, the object's shape (drag and tolerance) can be changed, or additional friction can be added by allowing the object to roll/slide on the surface of an inclined tube. Finally several techniques for have been used to better measure elapse time, e.g., optical [20] and electromagnetic [21,22]. Falling body viscometers are well suited for high-pressure applications, because the fluid is static; however, they are commonly used for liquid phase conditions versus the gas phase where the viscosity is very low (77 < 20 uPa-s). Other high pressure instruments adapted
28
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
to low viscosity measurements include oscillating disks, [23-27] vibrating wire [28,29] and torsionally vibrating piezoelectric quartz crystal [30,31]. Finally, a modification of the falling cylinder viscometer has been commercialized by Cambridge Viscosity Inc. This type of viscometer, henceforth referred to as a Cambridge Viscometer, was designed for high viscosity fluids; however, by using a hollow cylinder with less tolerance between the cylinder and vessel wall, it has been possible to reduce minimum measurable viscosity. The primary advantage of this viscometer is the small volume and broad viscosity range, eg., other cylinders can be purchased to accommodate very viscous fluids. The ASRL Cambridge Viscometer is operated with a low-mass silco-coated magnetic piston of diameter 0.312" and an advertised viscosity range of 20 to 200 uPa-s. The piston resides in a cylindrical SS-316 chamber with an internal diameter of 0.314" and operating conditions of T=190°C and p = 140 MPa. For this work the viscometer was held horizontal. The piston is moved a predetermined distance (0.2") at a constant force determined using two magnetic coils outside the SS-316 stainless steel chamber. By alternating the power to the coils, the round trip travel time is measured and translated into absolute viscosity. The measurement is completed for the motion in both directions. The optimal travel time for each viscometer piston is ca. 26 seconds at full scale; therefore, for a 20 - 200 uPa-s viscosity range, a total cycle time of 26 seconds should correspond to a viscosity of 20 uPa-s. To our knowledge, there are some research groups using this instrument; however, no viscosity data from this instrument at these low viscosities have been published in the open literature. Therefore, we have undertaken extensive testing of the instruments performance over a wide range of experimental conditions. Our early testing of the viscometer resulted in the conclusion that the factory calibration settings were inadequate, especially at high-pressures and low-temperatures. We have explored our own calibration procedure using pentane, hexane and pure C0 2 . The Calibration Drive Level (CDL) is the primary parameter which determines the magnitude of current flowing into the magnetic coils to drive the piston at a constant force. To begin our calibration, all other adjustable instrument parameters were initially set equal to zero. After cleaning the internals with isopropyl alcohol, the viscometer was evacuated for several hours, flushed and charged with
SOUR GAS FLUIDS AT H I G H TEMPERATURES AND PRESSURES
29
n-pentane (99.5%¿ Fischer Scientific) at a pressure slightly above 1 atm and at T = 25°C. This small pressure ensured that no bubbles were formed inside the chamber. Pentane was chosen to determine CDL because its viscosity is 217.9 uPa-s at p = 1 atm. and T = 25°C, [32] which is close to the high viscosity limit of the piston used. The next step in the calibration procedure was to optimize the high end correction factor (CHC) using hexane as high-viscosity fluid (77 = 296.3 uPa-s at p = 1 atm. and T - 25°C). Through an interative calibration and re-checking the calibration over time, the CDL and CHC were found to be 420 mA and -0.15, respectively. These values are significantly different from the factory calibration settings (CDL = 452 mA and CHC = 1.3). The low-end viscosity correction factor (CLC) was checked by measuring laser grade C 0 2 (PRAXAIR, 99.9995%) at P = 3 bar and T = 25°C. The CLC value was found to be insignificant and was set to zero. The parameters determined above worked well for the several fluids tested during this procedure as long as the measurements were carried out at low pressure. When the pressure or the temperature is changed significantly, the tolerance between the piston and the viscometer chamber also increases thus decreasing the resistance to motion. To compensate for these effects, a corrected viscosity, r¡a, is calculated from raw viscosity, 77^ using an isothermal correction factor which is linear in pressure: Va=^n+dvPynr
(4)
Note that the form of equation is equivalent to those outlined by Cambridge, [33] where (5)
I « = M VA c,=Jfc r =
1 + TPC
(T-25°C)^
(6)
100°C
and j
_ (,/Cp — L)kj
1-
p
PRC 20,O0Opsia
kT
(7)
In this case, cn and dn of Equation 4 were determined at each temperature by least square regression of the raw viscosities for C 0 2 (p = 1, 2, 5, 10, 20, 50 and 100 MPa) and the calculated viscosities
30
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
from Fenghour et al. [34] Again, these correction factors were found to be significantly different from the factory recommended values.
3.2.3
Charging and Temperature Control
A schematic of the experimental charging, control and logging system is shown in Figure 1. H2S was charged at pressure through a custom built SS-316 hydraulic floating piston (ca. 250 cm3). Ethylene glycol hydraulic fluid was delivered using a Waters High Performance Liquid Chromatography, HPLC, pump. Laser grade C 0 2 (PRAXAIR 99.9995%) was charged using a liquid C 0 2 pump (SFT-10, supercritical fluid technology) and methane (PRAXAIR 99.999%) was delivered using an air operated diaphragm gas compressor (pmax = 75 MPa; Supperpressure Inc. 46-14025-1). Pressure was measured via a Hastelloy Honeywell Sensotec TJE pressure transducer with a maximum calibrated pressure of p = 140 MPa. All valves and tubing were SS-316 (pmax = 210 MPa). Extra valves were included for fine adjustment of pressures, i.e., by displacement of the fluid by adjusting the valve stem position. All measurements were completed for static fluids.
Figure 1. A schematic of the vibrating tube densimeter and oscillating cylinder viscometer system. Component details can be found in the text.
SOUR GAS FLUIDS AT H I G H TEMPERATURES AND PRESSURES
31
The temperature of the VTD unit was controlled using a NESLAB RTE740 circulating bath which can control to within ± 0.01 from T = -40.00 to 200.00 °C. Temperature was measured at the VTD using an internal platinum resistance thermometer, PRT, and a second PRT (100 Q, 3 wire) inserted into the face plate and between the unit inlet and outlet. This second PRT was previously calibrated using the triple point of pure water and melting point of pure indium (99.9999%) according to ITS-90 (T mn = 0.01 °C; T = 156.5985 °C) [35]. The calibrations for both t,H20
'
m,ln
/ L
J
PRTs were checked by slowly melting distilled water which had been frozen inside the VTD. The inflection in density/time period upon melting was within ± 0.02°C for both PRTs. The temperature of the viscometer was controlled using a Julabo F12 with a range of -20 to 190°C and a stability of ± 0.03°C.
3.3
Results
Figure 2(a) shows the final correlation plot for the C 0 2 VTD calibration data at p = 1, 2, 5, 10, 20, 50 and 100 MPa and all four isothermal temperatures (T = 0, 50, 100 and 150°C). Figure 2(b) shows the differences between the experimental values and the Span and Wagner [16] reference equation used for calibration. Figure 2(b) also show a similar comparison of some literature data. The comparison of the calibrated experimental data shows a pooled standard deviation of 1.2 kg m~3. This accuracy is less than much of the literature data; however, we have found that this densimeter can produce slightly better results if applied to a narrower range of densities. Also the instantaneous time period has been used; whereas, some averaging may improve future results. Previous work with benzene showed an estimated error of 0.4 kg m~3. Figures 3(a) and 3(b) show the similar plots for the C 0 2 viscosity data; experimental viscosity versus those calculated using Fenghour et al. [34] and the relative difference between the experimental and calculated values. With the exception of the three largest pressures at T - 100°C, the pooled standard error based on the correlation plot is ca. 2% which is similar to the stated accuracy of the reference equation. The overall estimated relative error for each measurement has been calculated using 877/77 = [0.0004 + (2 a/77)2]05., where 0.0004 is the square of the calibration confidence (2%) and a is the standard deviation for the averaged measurement. Note
32
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 2. (a) Correlation plot of the isothermal VTD calibration using C 0 2 at T = 0,50,100 and 150°C (p = 1 to 100 MPa); (b) difference between experimental C 0 2 densities and those calculated using the Span and Wagner EOS;(16) ♦, this work (VTD); o, pooled isochoric densimeter literature;'6"9' A , pooled float/ sinker densimeter literature;" 112 ' x, pooled piezometer data;13738' O , Ihmels and Gmehling (VTD).09»
that each measurement represents an averaged reading of 20 data points. The estimated relative errors for the individual C 0 2 measurements ranged from Ô77/77 = 2 to 6 %. The relative difference plot in Figure 3(b) shows that these estimated relative errors are consistent with the overall differences and the differences shown with the literature data, which is a little sparse at the high pressures and temperatures (larger densities). Those literature values
SOUR GAS FLUIDS AT H I G H TEMPERATURES AND PRESSURES
33
Figure 3. (a) Correlation plot of the isothermal viscometer calibration using C 0 2 at T = 0,50,100 and 150°C (p = 1 to 100 MPa); (b) difference plot for the experimental C 0 2 viscosities from this work and the literature, p is calculated from the Span and Wagner EOS;r + T-a 1=0
(1) n
where p*w is the vapor pressure of water, T is the temperature and a. is the coefficient with the values listed in Table 1. In the liquid-liquid or vapor-liquid-liquid equilibrium, algorithm considers the heaviest phase as the aqueous phase only when mole fraction of water, xlw, is greater than 0.5. To ensure the model continuous at x'w - 0.5, a transition range is created where 0.4 <xw< 0.6. If x'w < 0.4, the fugacities of all components are calculated by EOS, and if xlw > 0.6, the API model is used. In this transition range, the fugacities of all the components are corrected by / = has + (fAPI -ÍEOS )• «
-0.4)/(0.6-0.4)
(2)
where / is the fugacity of a compound and subscripts EOS and API represent for the fugacity of the compound calculated by the equation of state and the API model, respectively. The default binary interaction parameters (BIP) in UniSim Design for the EOS part are shown in Table 2. The BIPs between H 2 0 - C 0 2 and H 2 0-H 2 S are the same as those used by Oellrich et al. [3]. The other constants such as Henry constants and chemical equilibrium constants of all components are taken from Wilson's work [1]. Although some of these parameters are user-tunable, the default values are used in this study. Table 1. Parameters in vapor pressure equation of water. i
ai
0
10.4592
6
9.0367E-16
1
-4.0490E-03
7
-1.9969E-18
2
-4.1752E-05
8
7.7929E-22
9
1.9148E-25
3
3.6851 E-07
i
«
,
■
4
-1.0152E-09
10
-3.9681E+03
5
8.6531E-13
11
39.574
58
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Table 2. Default interaction parameter in unisim design for the EOS part. H20-C02
-0.5572 + 0.001879 x T-1.274 x 10' 6 P
H 2 0-H 2 S
-0.3897 + 0.001565 x T - 1.142 xl0' 6 P
H 2 0-NH 3
-0.2533
NH3-CO3
0.0
NH 3 -H 2 S
0.0
C0 2 -H 2 S
0.1
5.3
Phase Equilibrium Calculation
To evaluate the applicability of the sour property packages, the results from UniSim Design are compared with experimental vapor-liquid equilibrium data from literature for different systems. Since the Sour PR and Sour SRK provide similar results under many conditions, only the results from the Sour PR property package are reported here. Experimental values of partial pressures of solutes in NH 3 water [1], NH 3 -C0 2 -water [4] and NH 3 -H 2 S-water [5] solutions are compared against the values predicted by the Sour PR model from UniSim Design in Figures 2, 3, and 4, respectively. The data points below 5 mmHg in Figure 2 and 5 kPa in Figures 3 and 4 are omitted because of the possible measurement error. It can be see that the prediction from Sour PR model fits the trend of experiment data very well, especially in the dilute region. With the molality of sour gases increasing, as illustrated in Figures 3 and 4, the total vapor pressure decreases first and reaches a minimum. At this minimum point, almost all of the NH 3 is converted into ammonium ions and the ions from the sour gases are saturated. Any excess sour gases injected would go into the vapor phase from this point on, resulting in a sharp increase in total pressure. This behavior of NH 3 solution can be utilized as a controllable operational variable for sour gas treating processes. Figure 5 shows the predicted molality ratios of C 0 2 and NH 3 at the vapor pressure inflection point under different concentrations and temperatures. These ratios are the measure of the captured
SOUR PROPERTY PACKAGES I N U N I S I M DESIGNES
59
Figure 2. Partial pressure of NH 3 in NH 3 -water solution. Experimental data: □, 293.15 K; A, 333.15 K; UniSim Design: —
Figure 3. Partial pressure in NH 3 -C0 2 -water solution at 313 K and mNH3=6 mol/kg. Experimental data: G, C0 2 ; A, NH 3 ; 0, total vapor pressure; UniSim Design: —
C 0 2 per mole of NH 3 . Figure 5 indicates that the amount of C 0 2 captured decreases as the temperature increases. Based on this phenomenon, it is desirable to establish a NH 3 recycle process to absorb C 0 2 at low temperature and to be regenerated at high temperature. This finding has been patented by Eli Gal [6]. Figure 5 also shows that high molality of NH 3 will increase the amount of C 0 2 captured due to higher rate of chemical equilibrium. This behavior has been confirmed by the experiments [4], [5]. At high temperature, impact
60
CO
SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 4. Partial pressure in NH 3 -H 2 S-water solution at 313 K and mNH3=3 mol/kg. Experimental data: □, H2S; A, NH 3 ; 0, total vapor pressure; UniSim Design: —
Figure 5. Molality ratio of C 0 2 and NH 3 at the inflection point. The curves from bottom to top: mNH3=3 mol/kg, wNH3=6 m o l / k g and mNH3=12 mol/kg.
of NH 3 concentration on C 0 2 absorption is more prominent and this phenomenon can help an engineer to select suitable process conditions to remove sour gases. Figure 6 shows the relative deviations of partial pressures between the calculated results and experimental data [1], [,4], [5] for ternary and quaternary sour water systems, where wt% is the total weight percentage of the solutes. This figure suggests that
SOUR PROPERTY PACKAGES IN U N I S I M DESIGNES
61
Figure 6. Relative deviation of partial pressure in ternary and quaternary solutions. D, C0 2 ; A, NH 3 ; 0, H2S.
Figure 7. Predicted vapor compositions of C 0 2 and H2S under different temperatures and pressures by Sour PR property package.
the Sour PR property package models dilute sour gas solutions more accurately than that of high concentration solutions. A possible reason for poor prediction at high solute concentrations is that the parameters used in the API model may not have been fitted with experimental data in these regions. Figure 7 shows the predicted compositions of C 0 2 and H2S in vapor phase with the temperature range from 30 to 140°C and
62
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
pressure range from 100 to 260 kPa for a C0 2 -H 2 S-NH 3 -H 2 0 system. The overall compositions in mole fraction are xNH =0.06, x co2 = 0-03 a n c [ xHS= 0.03. It can be seen that C 0 2 has higher solubility than H2S at the low temperature region. As temperature increases, less C 0 2 is absorbed in aqueous phase. Between 50~70°C, the composition of C 0 2 reaches a maximum. Beyond this temperature, NH 3 and water starts to vaporize and therefore absorption of C 0 2 and H2S in aqueous decrease sharply.
5.4
Conclusions
The results obtained by using UniSim Design Sour PR property package in UniSim Design for sour water system are compared with the experimental data. The evaluation indicates: 1. Sour PR property package can predict the phase behavior quite well at low solute concentration, which is more common in Oil and Gas processing where acid gases are removed 2. Although the deviation at the higher solute concentration region is little high, it can be further reduced by tuning the model parameters. 3. The solubility of C 0 2 and H2S in NH 3 solution under different operating conditions is studied. It provides solid rationale for the NH3 recycle process of the sour gas capture.
5.5
Future Work
The natural extension of this study will lead to further enhancements in UniSim Design through 1. Updated model parameters from experimental data, especially in the high concentration region, and 2. Enabled more user-tunable parameters to better represent the real plant conditions.
S O U R P R O P E R T Y PACKAGES I N U N I S I M D E S I G N E S
63
Reference 1. G.M. Wilson, "A New Correlation for NH3, C02, H2S Volatility Data from Aqueous Sour Water System", API Pub. No. 955, American Petroleum Institute, Washington, DC, 1978. 2. T. Irvine, and P. Liley, "Steam and Gas Tables with Computer Equations", New York, Academic Press, p. 21,1984. 3. L, Oellrich, U. PlOcker, J. M. Prausnitz, and H. Knapp, "Equation-of-State Methods for Computing Phase Equilibria and Enthalpies", Internat. Chem. Eng., Vol. 22, P. 1,1981. 4. F. Kurz, B. Rumpf, and G. Maurer, "Vapor-liquid-solid Equilibria in the System NH 3 -C0 2 -H 2 0 from Around 310 to 470 K: New Experimental Data and Modeling", Fluid Phase Equilibria, Vol. 204, P. 261,1995. 5. B. Rumpf, A. Pérez-Salado Kamps, R. Sing, and G. Mauer, "Simultaneous Solubility of Ammonia and Hydrogen Sulfide in Water at Temperatures from 313 K to 393 K", Fluid Phase Equilibria, Vol. 158-160, P. 293,1999. 6. E. Gal, Ultra, "Cleaning combustion gas including the removal of C0 2 ", World Intellectual Property, Patent WO 2006022885,2006.
This page intentionally left blank
6 Compressibility Factor of High C0 2 Content Natural Gases: Measurement and Correlation Xiaoqiang Bian, Zhimin Du, Yong Tang, and Jianfen Du The State Key Laboratory of Oil & Gas Reservoir Geology and Exploitation Engineering, Southwest Petroleum University, Chengdu, People's Republic of China
Abstract The JEFRI-PVT apparatus made in Canada by Schlumberger was used to obtain accurate compressibility factor measurements for high C0 2 content natural gases to study the effect of different C 0 2 content on gas compressibility factors (range covered: temperature, 263.15K to 313.15K; pressure, 3 MPa to 15 MPa). The results showed that gas compressibility factors reduce with increasing CO z content in natural gases and increase with increasing temperature. In addition, a non-integral power polynomial correlation was proposed without an iterative procedure whose coefficients were determined by fitting experimental data. The mixing rules used include: Kay's mixing rule combined with Wiehert-Aziz and Casey correlations (Kay) and Stewart-Burkhardt-Voo mixing rule with Wichert-Aziz and Casey (SBV). Comparison of the DAK-SBV, DAK-Kay, and proposed correlations showed that the presented model yielded the most accurate predictions with the lowest average absolute deviation (0.42%) among them.
6.1
Introduction
In recent years, m o r e a n d m o r e h i g h C 0 2 - c o n t e n t gas reservoirs in the w o r l d h a v e been discovered (Jokhio et al., 2001). Different C 0 2 content leads to different gas compressibility factor w h i c h is
Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (65-86) © Scrivener Publishing LLC
65
66
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
involved in calculating gas properties such as formation volume factor, density, compressibility, and viscosity (Shokir, 2008). All these properties are necessary in the natural gas industry for evaluating newly discovered gas reservoirs, calculating initial and gas reserves, and predicting future gas production. Prediction of gas compressibility factor contains four methods: experiment (Adisoemarta et al, 2004), equation of state (Li and Guo, 1991; Elsharkawy, 2002), corresponding state correlations (Riazi, 2005), and empirical formula (Li et al., 2001). Experimental measurement of gas compressibility factor is the most accuracy among all methods, but it is time-consuming and costly. Several correlations fitting the Standing and Katz chart (Standing and Katz, 1942) can be used to calculate the gas compressibility factor (Papay, 1968; Hankinson et al., 1969; Hall and Yarborough, 1973; Yarborough and Hall, 1974; Dranchuk et al., 1974; Brill and Beggs, 1974; Dranchuk and Abou-Kassem, 1975). But DranchukAbou-Kassem (DAK) correlation is the most accurate representation of Standing and Katz chart. In addition, among the mixing rules (Sutton, 1985; Elsharkawy and Elkamel, 2001; Bahadori et al., 2007), Kay's (1936) mixing rule and Stewart-Burkhardt-Voo (1959) are the most widely used. Kay's mixing rule is simple. StewartBurkhardt-Voo (SBV) rule provided the most satisfactory results (Satter and Campbell, 1963). Since the presence of C 0 2 gas, prediction of the compressibility factor is much more difficult than that of sweet gases. Wiehert and Aziz (1972) presented corrections for the presence of H2S a n d / o r C 0 2 for determining compressibility factor of sour gases. Casey (1990) proposed correlations for the presence of nitrogen (N2) and water vapor (H 2 0) to correct the pseudo-critical properties. Fortunately, in this study, due to the absence of C7+ fraction, the critical properties of the C7+ fraction are not calculated from correlations (Keseler and Lee, 1976; Pedersen et al., 1989). This study has two objectives. The first objective is to measure gas compressibility factor for different C0 2 -content natural gases (C0 2 content is about 10%, 30% and 50%, respectively) by applying JEFRI-PVT experimental equipments which were made in Canada by the Schlumberger company. The second objective is to develop a non-iterative empirical correlation to estimate gas compressibility factor based on the experimental results, and to make a comparison among the proposed model, DAK-SBV and DAK-Kay correlation.
MEASUREMENT AND CORRELATION
6.2
67
Experiment
6.2.1 Measured Principles For dry gases, compressibility factors can be calculated using the following equation:
z=PAVT-
(i)
P.-V.-T where P is experimental pressure, P s ambient pressure, T experimental temperature, Ts ambient temperature, AV volume of gas bled from the PVT vessel, and Vs is the volume of gas released at ambient pressure and temperature.
6.2.2 Experimental Apparatus and Procedure In this study, mercury free DBR-PVT vessels made in Canada were used to measure gas compressibility factors. A schematic diagram of the apparatus is shown in Figure 1. The experimental apparatus used consisted of PVT vessel of approximately 135 ml capacity, automatic pump, gas chromatography, dry gasometer, constant temperature air bath, flash separator and ground separator.
Figure 1. Schematic diagram of the experimental apparatus.
68
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
The experimental procedure is as follows: 1. Clean PVT vessel and cells, then connect the PVT vessel to the cells and evacuate the cells; 2. Control and maintain the desired temperature using the constant-temperature air bath; 3. Introduce a test gas (about 100 ml) into PVT vessel at the specified temperature and pressures in the oven and keep 5 h, then hold up half an hour and measure the gas volumes of the PVT vessel; 4. Slightly open the valve between the cells and bleed gas from the PVT vessel into the flash separators, at the same time, keep the pressures of the PVT vessel constant using automatic pump; 5. Record the bled gas volumes and remaining gas volumes of PVT vessel; 6. Use Eq.(l) to determine Z-factors; Repeat the procedure (4) ~ (6), and ensure tested Z-factors are about the same at least three times. More detailed descriptions are given by Liu et al. (2002), SY/T 6434-2000, and Varotsis and Pasadakis (1996).
6.2.3 Experimental Results The gas samples used were analyzed using a HP-6890 Gas Chromatograph. The results of analysis were shown in Table 1. The critical pressure and temperature for the pure components normally present in natural gases are also provided in Table 1 (Lu, 1982; Reid et al., 1987). Experimental compressibility factors for different C0 2 -content natural gases were listed in Table 2. It was shown in Table 2 that compressibility factors reduce with increasing C 0 2 content in natural gases and pressures, but increase with increasing temperature.
6.3
Methods
6.3.1 Existing Methods When gas composition is available, pseudo critical properties are calculated using a given mixing rule. In this study, Kay's mixing
MEASUREMENT AND CORRELATION
69
Table 1. Composition of gas mixtures and critical properties of defined component. Component
Sample 1
Sample 2
Sample P/MPa C 3
T /K c
Mw /g-mol 1
Mole fraction C02
0.0984
0.2886
0.5099
7.384
304.21
44.01
N2
0.0205
0.0124
0.0083
3.400
126.20
28.013
Cl
0.8602
0.6773
0.4654
4.595
190.56
16.043
C2
0.0167
0.0159
0.0116
4.871
305.33
30.07
C3
0.0031
0.0042
0.0033
4.247
369.85
44.097
iC4
0.0004
0.0006
0.0005
3.640
407.85
58.123
nC4
0.0005
0.0007
0.0006
3.796
425.16
58.123
iC5
0.0001
0.0003
0.0002
3.381
460.43
72.15
nC5
-
0.0001
0.0001
3.369
469.71
72.15
C6
0.0001
-
-
3.012
507.37
86.177
Table 2. Experimental compressibility factors for high C0 2 natural gases. Pressure (MPa)
Sample 1 313.15K
303.15K
293.15K
283.15K
273.15K
263.15K
3.00
0.9423
0.9307
0.9196
0.9028
0.8959
0.8833
5.00
0.9128
0.8992
0.8803
0.8608
0.8483
0.8337
7.00
0.8901
0.8721
0.8518
0.8252
0.8037
0.7829
9.00
0.8711
0.8563
0.8280
0.7981
0.7719
0.7410
11.00
0.8607
0.8369
0.8079
0.7809
0.7495
0.7054
13.00
0.8500
0.8309
0.8013
0.7714
0.7324
0.6901
15.00
0.8452
0.8286
0.8001
0.7691
0.7303
0.6860
(Continued)
70
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Table 2. Experimental compressibility factors for high C 0 2 natural gases. (Continued) Pressure (MPa)
Sample 2 313.15K
303.15K
293.15K
283.15K
4.21
-
-
-
0.8892
-
-
4.50
-
-
0.8891
-
0.8505
-
5.00
0.9025
0.8941
0.8718
0.8595
0.8297
0.8066
6.00
0.8745
0.8605
0.8452
0.8344
0.7998
0.7600
7.00
0.8576
0.8358
0.8161
0.8087
0.7710
0.7202
9.00
0.8277
0.8010
0.7809
0.7524
0.7109
0.6589
11.00
0.7957
0.7661
0.7406
0.7075
0.6719
0.6137
13.00
0.7644
0.7429
0.7136
0.6743
0.6395
0.5854
15.00
0.7513
0.7290
0.6882
0.6594
0.6431
0.5877
Pressure (MPa)
273.15K 265.65K
Sample 3 313.15K
303.15K
293.15K
283.15K
275.85K
4.00
-
-
0.8859
0.8574
0.8386
4.50
0.9118
0.8927
-
-
-
5.00
0.9002
0.8710
0.8353
0.8099
0.7674
6.00
0.8627
0.8247
0.7961
0.7710
0.6992
7.00
0.8246
0.7893
0.7683
0.7335
0.6470
9.00
0.7630
0.7423
0.7017
0.6402
0.5576
11.00
0.7158
0.7052
0.6483
0.5653
0.4915
13.00
0.6787
0.6735
0.6117
0.5211
0.4771
15.00
0.6565
0.6507
0.5890
0.5246
0.4937
MEASUREMENT AND CORRELATION
71
rule, Stewart-Burkhardt-Voo (SBV) mixing rule are considered. Kay's (1936) mixing rule, based on molar weighted average critical properties, has the following form:
^ = £yA
(2)
r
(3)
¿=i
P c=Xy.
T
¿=i
d
Stewart-Burkhardt-Voo (1959) (SBV) proposed the following mixing rule for high molecular weight gases. ( *.T
yTc
O ¡=l
\
Í
2
+—
v^ *
«=£ i=i
3
1=1
rzr\
y*
(5)
y/Ve
T*pcVC=K2/J
(6)
Ppc = T pc ' / i PC
(4)
VC '
(7)
>
Eqs. (2) and (3) or (4) through (7) provide critical properties for sweet natural gas systems. For high C 0 2 natural gases, these equations must be corrected for the presence of non-hydrocarbon components. The method proposed by Wiehert and Aziz (1972) (WA) is used to correct the pseudo critical properties of natural gases to H2S a n d / o r C 0 2 components. The WA correlation is given as follows: £ = - [ l 2 0 ( A a 9 - A L 6 ) + 15(Ba5-B4)]
(8)
Where the coefficient A is the sum of the mole fractions of H2S and C 0 2 in the gas mixture and B is the mole fraction of H2S in
72
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
the gas mixture. The corrected pseudo critical properties P' and r/ Tpc
a r e :
T' =TM -E * pc
p/
pc
(9) 7
fVV Tpc+B(l-B)Ç
pc
(10)
Correlations proposed by Casey (1990) are available for correcting pseudo-critical pressure and temperature for the presence of N 2 and H 2 0. The correlations for nitrogen and water vapor are: T p c c 0 r =-136.72y N 2 +222.22y H 2 O
(11)
ppCrC0r=-l.U7yN2+8.756yH2O
(12)
r;c-126.22yN2-647.22yH20 pc
1
1
r
= pc
_
pcœr
yN2
yH,o
P;-3-40yN2-22.06yH2O -• i —
pc,cor
3/N, ~VH7O
where Tl and P' c are the pseudo-critical temperature and pressure correlated for H2S and C 0 2 with the WA correlation. Reduced pressure (P r) and reduced temperature (T r) are calculated from pressure (P) and temperature (T) of interest and critical properties of natural gases ( T"c, P"c ) by the following relationship:
p
''!
p P" pc T PC
(15)
(16)
MEASUREMENT AND CORRELATION
73
The DAK correlation is extensively adopted to calculate the gas compressibility factor (Z) using reduced pressure (Pr) and reduced temperature (T) as follows: A,
Z=l+
A,
A¿
A + A ! A2 tf
Ac Í
T
V
r. r
v
(17)
2 \
-A
A ! A P?+Ao(l + AlPr2) Ä T T L
eX
2
P(-A!Pr2)
r 7
where (18)
Pr
ZTL
The constants Ax through An in Eq. (17) are listed in Table 3: Because the gas compressibility factor appears on both sides of DAK's correlation, Eq. (17), an iteration solution is necessary. Newton-Raphson method is used as follows: :(n)
in) _ yi.n+1) _ y{n)
(19)
dF M where ( . Ar, A-, A, Ac ? Pr + A +— - + ^ r + ^ - + —f- P,* +
0. Then differentiation of Eq.(24) with respect to x and letting x=l yield:
g'H) = bax + 2ba2 + 3ba3 + ... + nban g"il) = 1Kb - Da, + 2b(2b - \)a2 + ... + nb(nb - \)an (25) g{n\l)
= b(b-l)...[b-(n-l)]al
+
+ ... + nbiyib -l)...[nb-(n-
2b(2b-l)...[2b-(n-l)]a2 Y)]a„
The coefficients a.(i = l,2,---,n) can be determined through Eq.(25). In common, when fitting experimental data, let n=3, which can be sufficiently to satisfy an engineering requirement. Therefore, Eq.(24) can be rewritten as follows:
g(x) = g(0) + ^ 3 - [(6b2 - 5b + l)g'U) - (5b - 3)£"(1) + g'"(l)] xb + -^r-(3& 2 -4b + l)g'(l) + (4b-3)g"(l)-g'"(l)]x2h 2b
(26)
+ -L[(2b 2 -3b + l)g'(l) - 3(b - l)g"(l) + g'"0)] x3
When the conditions of the same number of terms between Eq.(23) and Eq.(24) apply, Eq.(23) becomes: f(x) = /(0) + /'(0)x + £ ^ x
2
+^-x3
(27)
In order to indicate the accuracy of Eq.(26) and Eq.(27), a random function y = (l +x)Z5 is taken for example (see Table 4). The average absolute deviations (AAD) and absolute relative deviations CARD) in the subsequent tables are, respectively, defined as:
0.8
4.3469
4.3600
4.3635
4.3457
4.3305
0.4
2.3191
2.3200
2.3365
2.3194
2.3046
y
f
g(b=0.92)
g(b=0.96)
g(b=\. 00)
X
0
5.6404
5.6556
5.6734
5.6875
5.6569
1
7.1624
7.1777
7.1954
7.2400
7.1789
1.2
10.8854
10.9004
10.9177
11.0800
10.9002
1.6
15.5841
15.5976
15.6121
16.0000
15.5885
2
Table 4. Comparison of the precision of Eq.(26) and Eq.(27) approximating to y=(\+x)2
21.3435
21.3523
21.3585
22.1200
21.3156
2.4
32.1569
32.1466
32.1195
33.8125
32.0000
3
0.256
0.086
0.282
1.718
/
AAD/%
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
MEASUREMENT AND CORRELATION
AAD(%)
=—£ ¡=i
ARD(%) ■■
Cal. - Exp. xlOO Exp.
Cal. - Exp. xlOO Exp.
77
(28)
(29)
It was shown in Table 4 that Eq.(26), whose AAD = 0.086% and b = 0.96, significantly improved the accuracy of the prediction of y = (1+x)25 as compared with Eq.(27) (AAD=1.718%). The proposed correlation of Eq.(26) was extended to apply to bi-variant one as follows: g(x, t) = a0 (t) + a, (t)xb + a2 (t)x2b + a3 (t)x3b a0 \t) = c00 + c10t + C2QI
+ c30t
flj (t) = C0/1 + C u f A + C2lt2A + C3At301
(30)
a2(t) = c02 + cX2tßl + c22t2ß2 + c22t3ßl a3(t) = c03+cl3t
+c23t
+c33t
Finally, a new correlation of gas compressibility factors for high C02-content natural gases can be obtained:
Z(Pr,Tr) = a0(Tr) + a,(Tr)Pra +a2(Tr)Pr2a +a3(Tr)P3a a0(Tr) =
Aa+BaTrß°+CXß°+DaTr3ßo
al(Tr) = Ab +
BhT^+CbT^+DbT3A
a2 (Tr ) = Ac + BcTf2 +CcT2ß2 +
D
Xßl
(31)
a3(Tr) = Ad + BdTrß>+CdT^+DdTr3ß>
where a ß. (i = 0,1,2,3) and A.,B¡,CjrD.{i-a,b,c,d) are undetermined coefficients which can be obtained by fitting experimental data.
78
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
6.5 Comparison of the Proposed Method and Other Methods The tuned coefficients of Eq.(31) can be obtained by fitting experimental data using MATLAB 7.0 (Table 5). But the tuned coefficients of the correlation needed refitting for different samples. Table 6 showed the absolute relative deviations of calculated compressibility factors of sample 2 and 3 by different methods. Figure 2(a-c) showed the predicted gas compressibility factors from the new model compared to the results of the DAK-SBV and DAK-Kay compressibility factor correlations versus the measured gas compressibility of the Sample 3. As can be seen from Table 6 and Figure 2, the proposed method was the most accurate of the three methods tested, giving an overall average absolute deviation of 0.42%. In the order of accuracy Table 5. The tuned coefficients of the proposed model for different samples. A
-8.816390
Ba
24.405726
a
\ Sample 1
A
12.520235
Bb
-31.429384
-5.306294
B
13.338221 c
0.747085
a
0.8
Aa
28.574773
\ Sample 2
A
ß«ßvß, B
65.179604
20.274849
B
-30.885941 C
1.1
A
79.235341
A
Bd
0.813848
c
c* 2.1
C
1.266354
Dd
20.748488
cb c
B
D
-29.974701
Dt
5.006821
14.296895
D
-2.482498 c
-166.38253
C
88.823846
-153.47652 C
28.118034
ßvßvßyßs
cd 2.5
0.4058196
2.3 D
-13.588054
-149.32077
Db
22.973990
83.105283
D
-12.845033
C
c
v*
Dd
a
a
B
-3.328087
a
ßvß,
cb c
0.0965256
2.5
A
2.2
72.278392
1.2
D
/VA 277.43880
a
-9.046675
-2.249317
B>
-13.225356
-2.280668
cd
4.837914
-130.68757
K
Db
c
a
C
21.474317
a
Bb
a
\
-45.647110
-41.125073
a
Sample 3
-1.874226
a
-3.301408
cb c
2.132880 a
c
B
then the saturation equation is:
+s„
dp
f
R,
pg(l-yc-zc)
Po _ PgVc X
B„ v
-R,
B„
B„
Pgtt-Vc-^
B,
PsVc + S„ B„ _ PogK
B„
j
In the equation, Rz is the defined intermediate variable.
(13)
232
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
14.2.2.4
Auxiliary
Equations
Capillary force and relative permeability equations: Kro=KJSg,Nc,oog)
(14)
Krg=Krg(Sg,Nc,aog)
(15)
P^Po
+ Pœg&g'K'Oog)
(16)
In the equations above, Nc is the capillary number, a o is the interfacial tension between gas and oil, po is the oil phase pressure, p is the gaseous phase pressure, and pco is the gas-oil capillary force. Constraint condition: Sg+Sl+Ss=l
(17)
Here, S1 is the liquid phase saturation. 14.2.2.5
Definite
Conditions
1. Gas pool original state „ |f=o __ B o
Po I n + T
PoSi+T '
T|f=o _ T o l I Q+T - L Ci+T°
2. Boundary state The closed outer boundary VO|route). = 0 , the constant pressure outer boundary p\router - Po > the constant bottomhole flow pressure p\nnner = const and the bottomhole constant f[dp/dn~\\rinner = const. Here, Q is the study domain, T is the temperature, V is the potential gradient function, Touter is the outer boundary and Tinner is the inner boundary.
14.3
Mathematical Models of Flow Mechanisms
14.3.1 Mathematical Model of Sulfur Deposition Sulfur deposition mechanisms include molecular diffusion, shear diffusion, Brownian diffusion, gravity subsidence, etc., among
MODEL OF GAS POOL WITH SULFUR DEPOSITION
233
which molecular diffusion and shear diffusion are considered to be major types of sulfur deposition according to some research. So the sulfur deposition model is: dW_=dWL+dWL dt dt dt
(18)
Here, W is the total deposition amount, Wd is the molecular diffusion deposition amount, Ws is the shear diffusion deposition amount. 1. Sulfur molecular diffusion deposition model According to the Fick Diffusion Law, the molecular diffusion deposition speed of sulfur can be expressed as: dWL dt
= c c P d a
A fi
idÇ| l dTl
dT dr
(19)
In the above formula, dW d /df is the mass of dissolved sulfur deposit from molecular diffusion in unit time, Cd is the deposit constant (generally 1500), C : is the liquid phase concentration, A is the surface area for sulfur deposition, ¡x is the liquid viscosity, C is the volume fraction of sulfur to the crude oil, dC/df is the volume fraction gradient of sulfur in the liquid and dT/dr is the radial thermal gradient. 2. Sulfur shear diffusion deposition model Sulfur particles behave two ways of horizontal migration, i.e. Brownian movement and shear diffusion, but the influence of Brownian movement is relatively small. Because of the porous flow speed-gradient field, the sulfur particles suspended in the oil flow will rotate in an angular velocity, contributing to their horizontal movement and shear diffusion. The sulfur shear deposition gradient caused by speed gradient with laminar flows can be expressed as the following: ^ - = Cäk'C'rA dt
(20)
In the equation, dW s /d£ is the mass of dissolved sulfur deposition from shear diffusion in unit time, k* is the shear deposition rate constant, C* is the volume fraction of sulfur particles on the surface and / i s the shear velocity.
234
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
14.3.2 Thermodynamics Model of Three-phase Equilibrium 1. Fugacity equilibrium equations (21)
fi=fi=n
(22)
f!=Mp
(23)
f;=4fr=xtffr
(24)
In the equations above, /,- , f¡ and ff are fugacity of the component i in the gaseous, liquid and solid phases respectively,
M
o
O O
n M Z
M
O H
M
r
M
> o
% O
H
O c¡ m
M
e/5
n
CTs O
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
261
2. Molecular weight and specific gravity of the crude oil sample were then entered into Winprop. Winprop then generated pseudo-components and the composition of the crude oil based on these pseudo-components. 3. The separator gas analysis from the 07-20 well was then entered into Winprop as a secondary fluid. 4. Using the Winprop Saturation Pressure calculation block, the crude oil and gas were then recombined, and the molar fraction of crude oil (primary fluid) in the mixture adjusted until the saturation pressure matched the initial reservoir pressure. Winprop predicts that with a mix of .68 mol fraction oil and .32 gas, the reservoir liquid will be saturated at a pressure of 8211 kPaa. If there is a target GOR that needs to be achieved, the molar fraction of oil and gas can be calculated by hand (44.5 M3 / m 3 * 1 kmol / 23.7 M3 = 1.88 kmol gas/m3 oil, 840 kg / m 3 oil * 1 kmol / 206 kg = 4.08 kmol oil / m 3 oil). At standard conditions, Winprop reports about 1% light ends remaining in the oil. This is consistent with the D86 reported 1% loss of sample during the distillation. It was likely boiled up, but the lab could not condense and recover it. 5. Since differential liberation data for the reservoir fluid was not available, 'live' reservoir fluid estimates of viscosity, gas-oil ratio, and specific gravity (3) were entered into the Winprop Differential Liberation calculation block. The Winprop Regression utility was then used to adjust properties of the pseudo-components to fit the differential liberation data. 6. Finally, for the IMEX reservoir simulation a Black Oil PVT Data block was added to export the fluid properties to IMEX. For the GEM compositional reservoir simulation, the CMG GEM EOS Model block was added. Properties of the reservoir fluid characterized in Winprop, compared to estimates from Slider, are provided in Figures 4-5. The estimates from Slider are shown as "Experiment" discrete data points. For the purpose of this project, the match on GOR, viscosity, and density is acceptable, especially considering that the match is to estimated properties in the literature. If laboratory differential
262
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
liberation results were available, additional work may be warranted to get a closer match. A complete program for laboratory analysis of reservoir fluids must be conducted as part of the detailed enhanced recovery process design.
Figure 4. Dunvegan C reservoir oil gor and bo vs estimates from slider.
Figure 5. Dunvegan C reservoir oil viscosity vs estimates from slider.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
263
15.4.3 Solubility of Acid Gas Mixtures in the Dunvegan C Oil Solubility of acid gas mixtures with compositions ranging from 100 mol% C 0 2 to 80 mol% H2S, 20 mol% C 0 2 in the live Dunvegan C Oil were studied. The study is not a textbook case for two reasons; 1. The pool is not pressure depleted. As a result, the oil contains significant methane and other light solution gas components. When acid gas comes into contact with the oil, the acid gas components replace methane in the oil, resulting in an acid gas rich oil phase, and a methane gas phase. This acts to increase minimum miscibility pressures. The study also indicates that in this case there is an increase in the oil phase density, rather than the C 0 2 swelling effect normally expected. 2. The average conditions of the pool are right around the critical point for C0 2 . Pool conditions are -7600 kPag at 34 °C. Critical point for C 0 2 is 7374 kPaa @ 31 °C. Critical point for H2S is 8963 kPaa @ 100 °C. In all cases, core flood studies are recommended to confirm solubility and miscibility predictions, and understand their effects on phase mobility and residual saturations. The somewhat unusual conditions under which enhanced oil recovery is being considered for this pool, while presenting opportunity, also further underline the need to conduct laboratory work to confirm predictions.
15.5 Material Balance Material balance calculations were completed prior to reservoir simulation to estimate fluids in place, initial condition of the reservoir oil (under-saturated or saturated), and drive mechanisms for the pool (gas cap, depletion, and water drive). According to information obtained from the ERCB, geological interpretation of the pool and the drive mechanism were among many subjects of discussion during the application process for Good Production Practice (GPP). Understanding the behavior of water in the pool proved to be a challenge throughout the project. Pool pressure data, acquired by acoustic well sounder survey,
264
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
indicate pressure depletion in the pool, leading to the conclusion that water drive is not significant. Well logs indicate high water saturation through the producing interval, but no water-oil contact, leading to the conclusion that the water present is immobile. This conflicts with reported water production, although the water-oil ratio in the down dip well 03-20 is trending downwards, with no clear trends in the other wells. For the purpose of material balance, the pool was assumed to be stratigraphically trapped with a down dip water leg providing insignificant pressure support. Pre-simulation material balance calculations are provided in Table 6. The calculations are in reasonable agreement with pool data, with the exception of pool pressure. Due to the quality of the pool pressure measurements, and the GPP information discussed above, the initial material balance calculations placed more weight on log data and geological interpretation of the pool, and less on the pool pressure data. The pre-simulation material balance indicated that the pool was initially saturated with a gas phase present. The material balance was revisited after the history match. Post simulation material balance calculations are also provided in Table 6. The history match indicated 56%+ water saturation versus 35% reported on the Reserves Summary and 46% average from well logs. The history match also indicated 423.5 E3M3 Original Oil in Place versus 789 E3M3 reported on the Reserves Summary. The post simulation material balance calculations are in better agreement with pool pressure data measurements, and are also in agreement with the history match results for the 56% water saturation case. The conclusion from this work is that the pool contains lower oil reserves than what is on record with the ERCB.
15.6 Geological Model Data available for the geological model consisted of simple geological isopach (thickness) and formation top maps of the pool, and the wireline logs used to generate the geological maps. A summary of the data sources used for this project is given in Table 7. Geological isopach and formation top maps for the Dunvegan C pool were digitized and imported into CMG Builder. Porosity and connate water saturation from logs were entered into CMG Builder and CMOST as initial estimates.
44.5 0.0114 1.11 1
M3 M3 M3/M3 m 3 /M3 m 3 /M3 m 3 /M3
Np
Wp
Rso
Bg
Bo
Bw
Swi
We
cf
oil produced
water produced
gas oil ratio
gas formation volume factor
oil formation volume factor
water formation volume factor
initial water saturation
water influx (reservoir bbl)
formation compressibility
0 5.80E-04
m3 MPa"1
0.46
0
0
0
E3M3
Gp
gas produced
8.2
MPa
Pr
Initial Condition
reservoir pressure
Parameter/ Case
Material Balance Input
0
.35?
-
-
-
-
4,181
15,489
1,722
7.0
Pool Data on Record
Grande Prairie Dunvegan C Material Balance Calculations
Table 6. Material balance calculations.
0
0.46
1
1.1
0.0126
40.7
4,181
15,489
1,722
7.6
Pre Simulation
0
(Continued)
0.56
1
1.09
0.0138
36.9
4,181
15,489
1,722
7.0
Post Simulation
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
30,177
-
-
M3/M3 m3 m 3 /m 3 m3 m3
Bt
Rp
F
m
Eo
Eg
E.
oil + solution gas formation volume factor
cumulative gas oil ratio
net total production (includes (Wp-We))
gas to oil in place ratio
oil expansion
gas expansion
water and formation expansion f,w
0 042
-
M3
N
initial oil in place
m3
m 3 /M3 -
Calculations
7.89E+05
1.3 38E+06 7.89E+05
138E+06
886
3,905
34,966
111
1.148
7.89E+05
m3
Pore Volume and OOIP Estimates
Vp 1.69E+06
Pre Simulation
pore volume
4.79E-04
Pool Data on Record
Cw
MPa"1
Initial Condition
formation water compressibility
Parameter/ Case
Table 6. Material balance calculations. (Continued)
1,052
437
35,459
0.005
36,949
111
1.196
4.13E+05
1.05E+06
Post Simulation
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Sgi
Soi
Swi
Sg
So
Sw
initial oil saturation
initial water saturation
current gas saturation
current oil saturation
currentwater satu ration -
-
fraction fraction
-
fraction
-
-
fraction fraction
-
fraction
Err ->adjust Vp or N until Error=0
initial gas saturation
F- (Eo+Eg+Efw)=0 solver
0.458
0.504
0.039
0.460
0.518
0.022
0
0.556
0.414
0.029
0.560
0.438
0.002
0
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
268
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Omitted from Table 7 are data sources such as correlations from the literature, and generic field or basin data that can be used in the absence of primary or secondary sources. It is strongly recommended that data be collected from primary or secondary sources before conducting a reservoir simulation.
Table 7. Geological model data sources. Geological Parameter
Primary Data Source
Secondary Data Source
Source for this Project
Formation Bulk Volume ( top, thickness, closure)
Well Log Interpretation
Seismic interpretation Well test interpretation Geological interpretation
Geological interpretation
Porosity
Routine Core Analysis
Well Log Interpretation
CMOST optimized with initial estimate from logs
Permeability
Well Test Interpretation
Routine and Special Core Analysis
CMOST optimized
Residual Fluid Saturations (water, oil, liquid, gas)
Special Core Analysis
Well Log Interpretation
CMOST optimized with initial Swc estimate from logs
Relative Permeability curve exponents and end points
Special Core Analysis
Well Skin or near-wellbore permeability
Well Test Interpretation
—
CMOST optimized
Huid Contacts
Well Log Interpretation
Material Balance
CMOST optimized
CMOST optimized
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
15.7
269
Geological Uncertainty
Given the data sources available for this project, there is considerable geological uncertainty associated with the Dunvegan C Pool. This is not unusual for pools of this size and vintage. Computer Modeling Group's CMOST optimization software was used to complete a computer assisted history match using the IMEX black oil reservoir simulator to complete individual simulations. Geological uncertainties in the pool were represented by thirty individual parameters input into CMOST. CMOST cases were completed with connate water saturation at 36%, 46%, and 56%, and the results compared. The history match operation was successful with global objective function error of less than 5% in all optimized cases.
15.7.1 Formation Bulk Volume Uncertainty in formation geometry was not directly accounted for in this project because only one geological interpretation of the pool geometry was available. In many cases, including this one, formation geometry is a significant uncertainty. The history match cases resulted in lower porosity than indicated by well log interpretation. One possible conclusion from this is that the reservoir simulation is responding to a 'pool compressibility', and that the pool is geometrically smaller than indicated by geological mapping.
15.7.2 Porosity Porosity from well log interpretation is 18%. This is based on sandstone matrix density of 2.65 g/cm 3 . A porosity range of 10-24% was used in the history match optimization. The history match returned optimized porosities between 12-14%. Density-neutron well log cross plot interpretation indicates a shaley matrix, with matrix density is between 2.54-2.57 g/cm 3 . Routine and special core analysis is recommended to correct well log data and confirm interpretation.
15.7.3
Permeability
No permeability data was available for this pool. Estimates of permeability in the i, j , and k directions is important data for
270
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
designing optimum enhanced oil recovery programs, and is normally obtained from well test interpretation and routine or special core analysis. For this project, a 0.01-100 md range of permeability values in each of the i, j , and k directions was input into the optimization software. The optimized history match cases returned permeability in the range of 5-50 md in each of the directions, with k (vertical) permeability being unusually high.
15.7.4
Residual (Immobile) Fluid Saturations
Residual fluid saturations are obtained from special core analysis relative permeability results for water-oil and liquid-gas fluid regimes, with attention paid to match the fluid movement history in the core with that anticipated in the field to capture hysteresis effects. For this project, the only residual fluid saturation data available for this project was connate water saturation from well log interpretation and geological interpretation. Pool average connate water saturation in the range of 35-52.5% has been reported by various sources. Connate water saturation calculated from well logs depends on the Archie equation constants used, and therefore is subject to interpretation. Using standard Archie constants of a=l, m=2, and n=2, calculated connate water saturations in the range of 65-68% were calculated. Given the above information, optimized history matches were completed for connate water saturations in the range of 36-56%.
15.7.5
Relative Permeability Curve Parameters
Relative permeability curve exponents and end points are best obtained from special core analysis of water-oil and liquid-gas fluid regimes, with attention paid to match the fluid movement history in the core with that anticipated in the field to capture hysteresis effects. No relative permeability data from core analysis were available for this project. Initial estimates of relative permeability parameters were made by using rule of thumb parameters available in the literature, and production data for each well in the pool. Ranges of relative permeability end points and exponents were input into the optimization software. CMOST optimized end points and exponents are given in Table 8.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
271
Table 8. Fractional flow study inputs versus CMOST results. Dunvegan C Fractional Flow Study Inputs Input Variable
CMOST Results for Comparison
Description
Input
Sw36
Sw46
Sw56
krw*
end point water rel perm
0.5
0.7
0.8
0.7
kro*
end point oil rel perm
1
0.8
0.8
0.6
Swi
irreducible water
0.46
0.36
0.46
0.56
Sor
residual oil
0.2
0.2
0.2
0.2
eo
oil exponent
3.5
4
4
2
ew
water exponent
2
1.5
1.5
1.5
krl*
end point liquid rel perm
1
0.8
0.8
0.6
krg*
end point gas rel perm
1
0.4
0.5
0.4
Sli
irreducible liquid
0.6
0.46
0.56
0.66
Sgr
residual gas
0.04
0.01
0.01
0.01
eg
gas exponent
3
1.5
1
1
el
liquid exponent
2
2
4
4
A study of three phase fractional flow was completed to better understand the expected range of relative permeability end points and exponents, and the fluid saturations required to allow for produced water flow. Stone's method with Aziz and Settari corrections, along with rule of thumb exponent and end point data were used to predict three phase fractional flow (4) (5) (6). Historical fractional flow data from the pool, was obtained by converting production data to fractional flow data at low pressure (near wellbore) and high pressure (reservoir) conditions. The predicted and historical fractional flow data were then compared, and water saturation modified until a match obtained.
272
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
A match at 55% water saturation was obtained for historical fractional water production, with relative permeability exponents in the 2-3 range, end points in the 0.5 to 1 range, Swc (immobile) =0.46, Sor=.2, Srg=.04. Water saturations above 55% were not investigated. Water saturation below 55% resulted in insufficient water flow. The procedure used for this study merits additional review, but the concept of translating relative permeability relationships and actual production into fractional flow data, and comparing the results, is valuable for understanding fluid flow behavior in the reservoir, and predicting the impact of various operations on oil flow and recovery from the pool. Comparison of the study results with the optimized history match results is provided in Table 8. The only value from the study was in knowing the historical fractional flow of each phase at reservoir conditions. Three phase flow through porous media is poorly understood, difficult to study and difficult to predict, but important to consider when evaluating enhanced oil recovery projects. The same can be said for the impact of fluids present on residual saturations.
15.7.6 Fluid Contacts An initial water-oil contact of 335 mSS, from geological interpretation of well logs, was used in the reservoir simulations. Uncertainty in the water-oil contact was not investigated in this project. Geological interpretation of the pool, pressure data, and water production trends from the pool do not indicate an active aquifer. Water production from the pool is not significant in terms of volume of reservoir fluids removed. Although there was no direct evidence of a gas-oil contact from well log interpretation, production data and pool material balance calculations indicated that a small gas cap may be present in the pool. Based on pool geometry and material balance calculations, an initial estimated gas-oil contact of 310 mSS was used, and range between 308-314 mSS allowed in the history match optimization. This translated to initial gas saturations of 7-11 % in the optimized history match cases, at optimized contacts of 310-312 mSS.
15.8
History Match
Computer Modeling Group's CMOST optimization software was used to complete the history match.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
273
Geological uncertainties in the pool were represented by thirty individual parameters input into CMOST. For ease of execution, connate water saturation was held constant in each of the CMOST optimization cases. CMOST cases were completed with connate water saturation at 36%, 46%, and 56%, and the results compared. In each of the cases, CMOST determined the optimum value of 29 parameters that would produce a reservoir simulation that minimized the error between the reservoir simulation and the production history. In CMOST, the production history is represented by a user defined 'global objective function'. For this history match, the global objective function was defined as a weighted average of cumulative gas, oil, and water production for each of the wells, and estimated bottom hole pressures for 03-20, 10-20, and 13-20 wells on artificial lift. Note that IMEX primarily matched oil production in all cases. A summary of the results is provided in Table 9. Charts summarizing pool production history data with the optimized history match cases are provided in Figures 6-8. Observed trends and indications from the history match are; • Global error in the history match decreased with increasing water saturation. This was due to decreasing errors in the cumulative gas and cumulative water production. Match to oil production in all cases was generally excellent. There was no trend amongst the cases in error between the simulation and the constructed bottom hole pressure data. • Porosity is much lower than reported from well log interpretation. • Original Oil in Place in the Sw46 and Sw56 cases is significantly lower than the OOIP on public record for the pool. This is supported by pressure data on file for the pool. • All history match cases indicate the presence of a gas cap, with gas oil contact at 310-312 mSS, and initial gas saturations ranging from 11% in the Sw36 case to 7% in the Sw56 case. • Permeability results in the Sw36 and Sw46 cases were unusual, with k direction (vertical) permeability being higher than i and j direction permeability. Permeability in the Sw56 case met conventional expectations,
2.14E-05 1.20E-05
0.096769
2.14E-05
1.20E-05
21.669
12.618
4.012
20.055
%
%
%
%
%
%
%
W07_20_Oil
W10_20_oil
W13_20_Oil
W03_20_Gas
W07_20_Gas
W10_20_Gas
W13_20_Gas 19.958
7.2683
23.932
8.9173
6.93E-05
1.66E-05
1.66E-05
%
W03_20_Oil
4.09
4.93
1925
Sw_46
%
1280
Sw_36
GlobalObj.
Selected Run
Units Sw_56
4.8637
4.2855
15.444
14.344
1.20E05
2.14E-05
6.93E-05
1.66E-05
3.12
1889
Mar29SW46_PASS01 .cmr
Comparison of optimized base case reservoir simulations
Parameter
Mar29SW36_PASSl 3.cmr
DNVG C Pool EOR Study
2 weighting
4 weighting
2 weighting
4 weighting
9 weighting
18 weighting
22 weighting
27 weighting
Volumetric weighting applied to production
from above .cmr files
Notes
Mar29SW56JPASS02.cmr
CMOST Results Files:
ENCH-699
Table 9. History match results. C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
%
%
%
%
%
fraction
-
W10_20_Water
W13_20_Water
W03_20JPress
W10_20_Press
W13_20_Press
Swc
AQ03RR
GasOilC
AQ13RR
AQ10RR
mSS
-
-
-
%
W07_20_Water
AQ07RR
%
W03^20_Water
14.45 13.455 43.29
14.218
8.0897
46.057
7
7
6 2 310
5
2
312
5
0.45
0.36
5
54.093
37.94
23.913
311
2.5
5
7
7
0.56
30.887
42.636
7.0923
55.289
43.529
4.9649
9.3174
13.409 9.7239
2.7389
3.4641
16.151
gas-oil contact.
(Continued)
aquifer strength indicator for well 13 (located mid-updip edge fmn)
aquifer strength indicator for well 10 (located mid-updip center fmn)
aquifer strength indicator for well 07 (located mid fmn)
aquifer strength indicator for well 03 (located low in fmn)
Reported Sw=.46from logs
1 weighting
1 weighting
1 weighting
2 weighting
4 weighting
5 weighting
7 weighting ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POO
md
fraction
fraction
fraction
permk
porosity
Sgcon
Slcon
Soirw
fraction
-
md
permj
SLT
md
-
-
-
Units
permi
krwiro
krocw
krgcl
Parameter
0.1
0.05
g=l, o=4
g=1.5,o=2
0.02
0.05
0.25
g=L o=4
0.1
0.12
0.12
0.05
10
50
0.1
0.02
0.14
50
10
50
5
10
5 10
end point relative water permeability in w-0 system
0.7
0.8
0.7
residual oil saturation in w-o system
relative permeability curve sets for 1-g system (exponents)
residual oil saturation in 1-g system [total Slr=Swc+Slcon)
residual gas saturation in 1-g system
reported porosity =0.18 from logs, .1-.24 range in CMOST
reservoir permeability in k direction
reservoir permeability in J direction
reservoir permeability in direction
end point relative oil permeability in w-o system
0.6
0.8
0.4
end point relative gas permeability in 1-g system
Notes
0.8
Sw_56 0.4
Sw_46 0.5
Sw_36
Table 9. History match results. (Continued) C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
md
md
md
md
md
md
md
md
w03fracpermk
w03fracpermj
w07fracpermi
w07fracpermk
w07fracpermj
wlOfracpermi
wlOfracpermk
-
w03fracpermi
SWT
0.01
100
1.00E+05
1.00E-02
1E5
1E5
1.00E+02
1.00E-01
w=l .5,o=4
0.01
1
(Continued)
simulating near wellbore stimulation or damage
simulating near wellbore stimulation or damage
1E2
1E6
simulating near wellbore stimulation or damage
simulating near wellbore stimulation or damage
simulating near wellbore stimulation or damage
simulating near wellbore stimulation or damage
simulating near wellbore stimulation or damage
relative permeability curve sets for o-w system (exponents)
simulating near wellbore stimulation or damage
0.001
1000
1E5
1E4
1E-2
w= 1.5,o=2
1E3
1000
0.01
1E6
1E3
1.00E+05
1E0
w=1.5,o=4
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POO
1E-6
1E9
700.5
45.9
529.2
md
md
E3M3
E6M3
E3M3
wl3fracpermk
wl3fracpermj
OOIP
OGIP
OWIP
E3m3
E3m3
Initial Oil
Solrn + Free Gas
523.26
777.555
0.1
md
wl3fracpermi
Reservoir Volume Calculations
100
Sw_36
md
Units
wlOfracpermj
Parameter
Table 9. History match results. (Continued)
705.5
579.6
379.62
302.1
470.085
26.5
33.3
560.217
423.5
504.7
Calculated at Reservoir conditions, Boi=l.ll, Bg=.0114, Bw=l
simulating near wellbore stimulation or damage
simulating near wellbore stimulation or damage
0.001
IE-6 1.00E+09
simulating near wellbore stimulation or damage
1
0.01
1.00E+09
simulating near wellbore stimulation or damage
Notes
1E1
Sw_56
1EO
Sw_46
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
OGIP-solution gas. Rsoi=44.5 Calculated, OIP/(OIP+WIP+Gas Cap Gas) Calculated, WIP/(OIP+WIP+Gas Cap Gas) Calculated, GAS Cap Gas/ (OIP+WIP+Gas Cap Gas)
87.3 0.37 0.56 0.07
123.6 0.44 0.46 0.10
167.9
0.53
0.36
0.11
So
Sw
Sg
E3m3
Gas Cap Gas
705.5
579.6
529.2
E3m3
Initial Water
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POO
280
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 6. History match pool cumulative oil.
Figure 7. History match pool cumulative gas.
with vertical permeability being 1/10 of the maximum horizontal permeability. Horizontal directional permeability was predicted in the Sw36 and Sw56 cases, important for designing enhanced oil recovery programs.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
281
Figure 8. History match pool cumulative water.
• Near wellbore permeability parameters were included in the history match to simulate stimulation or damage from drilling and completion operations. There was extreme variation in the order of magnitude of near wellbore permeability in the optimized cases. This method of modeling near wellbore reservoir properties appears to be flawed. A better approach may be to use skin factor from well test interpretation if available, or estimate skin based on the completion program. • The strength of the aquifers added to each well in order to provide free water production generally was highest with wells located down dip in the pool, with strength decreasing in the u p dip wells, and thinner pay wells. This method for accounting for water production is also flawed. Water production is likely from water mobilized during completion operations. A manual history match exercise was completed prior to using CMOST to complete the optimized history match. Observations from this exercise were; • Gas production estimates from IMEX were relatively easily tuned to the field history by adjusting gas oil
282
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
contact, gas relative permeability exponent, oil relative permeability end point, and i direction permeability in perforated grid blocks. • Matching the water production (WOR=0.2 to 0.3 for up dip and down dip wells) was particularly difficult. Well logs and geological interpretation do not indicate a water oil contact in the pool, nor do they indicate any high permeability connection to a water source. No mobile water bearing zone adjacent to D N V G C . Matrix expansion is insufficient to mobilize the produced water volume. Cumulative Water-Oil ratio is declining for each well. No indication of water bank arrival at up dip wells. Attaching a Carter-Tracy limited extent aquifer to each well provides the closest match to field history. ( Fetkovitch, Carter Tracy Infinite, and "Old" aquifer models were also investigated) Strength of the aquifer for all wells is set at 10, with the exception of 13-20 set at 1.5. Note that pay height at 13-20 is 1 m compared to the other wells at 5-9 m. Note Completions are generally 20 tonne, 45-65 m 3 gelled oil fracs, water production on all wells is present immediately, and 13-20 (one of the up dip wells) was re-completed and had highest water production immediately after recompletion. No production evidence of a 'flood front' moving into the wells. If a lower layer is being swept out (waterflooded), it would be indicated by a distinctive flood front arrival.
15.9 Black Oil to Compositional Model Conversion The optimized history match cases were converted from IMEX to GEM in order to simulate an acid gas flood on the pool. Comparison of field history with history match cases in IMEX, and the cases converted to GEM, is provided in Figures 9-17. A comparison of IMEX and GEM parameters for the various cases is provided in Table 10. The GEM simulations produced an unacceptable match to IMEX and field historical gas and water production. Predicted well bottom hole pressures in the GEM simulations were also in poor agreement with the IMEX simulations. Although the conversion process
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
Figure 9. Sw36 IMEX-GEM comparison pool cumulative oil.
Figure 10. Sw36 IMEX-GEM comparison pool cumulative gas.
283
284
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 11. Sw36 IMEX-GEM comparison pool cumulative water.
Figure 12. Sw46 IMEX-GEM comparison pool cumulative oil.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
Figure 13. Sw46 IMEX-GEM comparison pool cumulative gas.
Figure 14. Sw46 IMEX-GEM comparison pool cumulative water.
285
286
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 15. Sw56 IMEX-GEM comparison pool cumulative oil.
Figure 16. Sw56 IMEX-GEM comparison pool cumulative gas.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
287
Figure 17. Sw56 IMEX-GEM comparison pool cumulative water.
was unsuccessful here, the GEM models were used to simulate various acid gas floods for proof of concept purposes. It is also worth noting that the acid gas floods were simulated with horizontal injection and production wells, and all existing wells shut in to control flood the flood front location and defer breakthrough of injected acid gas to the producing wells. The following factors likely contributed to the unsuccessful conversion; 1. There was some difficulty in getting the GEM simulations to run successfully. Through experimentation with the simulations it was found that the simulations would run if variation of permeability in the reservoir was restricted. Near wellbore permeability for each of the wells was restricted to 1 to 100 md. Refer to Table 10. 2. Use of a large number of components in the GEM fluid model likely contributed to the instability of the simulations. Some components, such as iso-butane and normal-butane, have very similar properties and would be difficult for the simulator to resolve. One important lesson from this exercise is that lumping of such components will give better results. The
-
md
md
md
fraction
fraction
krwiro
permi
permj
permk
porosity
Sgcon
mss
GasOilC
-
-
AQ13RR
krocw
-
AQ10RR
-
-
AQ07RR
krgcl
-
fraction
Swc
AQ03RR
Units
Parameter 0.46 7
0.46 7
0.36 7
0.8
0.8 0.8 10
0.4 0.8 0.7 5
0.01
0.14
50
10
5
0.7
0.8
0.02
0.14
1
10
0.5
0.5
312
312
0.4
310
310
2
2
0.05
0.12
0.12 0.01
1
10
10
0.8
2
50
10
2
6
5
5 6
5 5
Sw_46 GEM
Sw_46 IMEX
Sw_36 GEM
5
5
7
0.36
Sw_36 IMEX
Table 10. Comparison of IMEX and GEM parameters.
0.01
0.12
10
50
5
0.7
0.6
0.4
311
2.5
5
7
7
0.56
Sw56 IMEX
0.02
0.12
1
50
5
0.7
0.6
0.4
311
2.5
5
7
7
0.56
Sw_56 GEM
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
fraction
-
fraction
-
md
md
md
md
md
md
md
md
md
md
md
md
SI con
SLT
Soirw
SWT
w03fracpermi
w03fracpermk
w03fracpermj
w07fracpermi
w07fracpermk
w07fracpermj
wlOfracpermi
wlOfracpermk
wlOfracpermj
wBfracpermi
wl3fracpermk
wl3fracpermj
1E9
IE-6
0.1
100
0.01
100
1.00E+05
1.00E-02
1E5
1E5
1.00E+02
l.OOE-01
w=1.5, o=4
0.2
g=1.5,o=2
0.46
10
1
100
10
1
100
1.00E+09
IE-6
0.01
1E0
1
1E6
1000
0.01
1 10
IE6
1E3
1.00E+05
1E0
10
1
100
10
1
100
10
1
100
10
1
100
w=1.5,o=4
0.05
0.2 w=1.5, o=4
g=l, o=4
0.51
g=l,o=4
0.56
100
10
1
100
w=1.5, o=4
0.1
g=1.5,o=2
0.46
1.00E+09
0.001
1
1E1
0.01
1E2
1E3
0.001
1000
50
1
100
50
1
100
50
1
100
50
1
1E4 1E5
100
w=1.5, o=2
0.25
g=l, o=4
0.66
1E-2
w=1.5,o=2
0.2
g=L o=4
0.66
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
290
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
geological model arrived at through IMEX simulations had to be compromised due to the onerous fluid system that the user imposed on the GEM simulations. Well deliverability, and near wellbore reservoir properties need to be more accurately represented in the simulations. An approach of modeling wellbore deliverability by skin, rather than permeability variations, may have resulted in simple stable GEM models. Extreme values and order of magnitude variations in near wellbore permeability likely contributed to GEM model instability.
15.10
Recovery Alternatives
A number of recovery alternatives were studied in this project, but the alternative of interest is a miscible acid gas flood. This alternative is conceptually attractive for a number of reasons; 1. Theoretical residual oil saturation in a miscible flood is 0, resulting in complete oil displacement and recovery from the reservoir volume swept by the injected fluid. The challenge is achieving acceptable sweep efficiency. 2. Environmental benefits of sequestering acid gas, including pure C0 2 , can be coupled with additional oil recovery. This means that the cost of acid gas or C 0 2 capture and storage would partially or fully paid for by the additional revenue from recovering additional oil from the reservoir. 3. There is significant potential for additional recovery from conventional oil pools. In Alberta, an estimated 7.7 Billion m 3 of oil remains unrecovered in conventional pools, 74% of the Original Oil in Place (7). A generic program was designed, consisting of one horizontal injection well placed up dip in the pool, and two horizontal production wells placed down dip in the pool. The program was then operated under the following recovery alternatives; 1. Acid gas flood, miscible at injection pressure down to 8 MPa, producing wells at 4 MPa. Identical operating
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
2.
3.
4.
5.
291
conditions were used and applied to the 36%, 46%, and 56% water saturation cases. While not comprehensive, this illustrates how the impact geological uncertainty on the project could be evaluated by optimizing a program based on one geological interpretation, then applying that program to other possible geological realities to determine what the technical and economic impact of implementing a non ideal program would be. These discrete cases, with probabilities attached to them, can then be used in a Decision Analysis process to determine the Expected Monetary Value of the project. Production from the horizontal wells only, with no flood implemented. This scenario was run to understand the incremental difference between simply investing in additional horizontal wells versus implementing a flood using the wells. Identical operating conditions were used and applied to the 36%, 46%, and 56% water saturation cases. Production from existing wells only, under 'pump off operation with wells at 1000 kPa bottom hole pressure. This scenario was run to understand the incremental difference between simply lowering the pressure in the existing wells, and producing without Maximum Rate Limitation (MRL) versus investing in additional horizontal wells, also producing without the MRL. Identical operating conditions were used and applied to the 36%, 46%, and 56% water saturation cases. Production from existing wells only, under the initial ERCB Maximum Rate Limitation (MRL) of 8 m 3 / d and corresponding Gas Oil Ratio of 100 m 3 /m 3 . C 0 2 flood, miscibility developed by increasing pool pressure above 13 MPa, producing wells at 4 MPa. A stand alone single case, using 46% water saturation and 18% porosity, was run to demonstrate the potential of coupling C 0 2 disposal with Enhanced Oil Recovery.
Event logs, describing pool operation under each of the above recovery alternatives, are provided in Tables 11-15. Screen shots of the pool are provided in Figures 18-20. Production and injection forecasts for the pool, and individual wells, are provided in Figures 21-42.
292
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Table 11. Acid gas flood event log. Program: Acid Gas Flood Event Description
Well
Date l-Oct-10
Hz Injector-1
begin drill, complete, tie in operations
l-Oct-10
Hz Producer-2
begin drill, complete, tie in operations
1-Jan-ll
Hz Producer-2
Produce with Constraint Min BHP=4000 kPa
1-Jan-ll
13-20
Shut In
1-Jan-ll
10-20
Shut In
1-Jan-ll
07-20
Shut In
1-Jan-ll
03-20
Shut In
1-Jan-ll
Hz Injector-1
Inject with Constraints Max Rate=30 E3M3/D, Max BHP=15 MPa
l-Oct-22
Hz Producer-3
begin drill, complete, tie in operations
l-Oct-22
Hz Producer-2
Prepare to convert the well form a Producer to an Injector
l-Jan-23
Hz Producer-3
Produce with Constraint Min BHP=4000 kPa
l-Jan-23
Hz Injector-2
Inject with Constraints Max Rate=30 E3M3/D, Max BHP=15 MPa
Table 12. Horizontal wells event log. Program: Horizontal Producers Date
Well
Event Description
l-Sep-09
13-20
Produce with Constraint Min BHP=1000 kPa
l-Sep-09
10-20
Produce with Constraint Min BHP=1000 kPa
l-Sep-09
07-20
Produce with Constraint Min BHP=1000 kPa
l-Sep-09
03-20
Produce with Constraint Min BHP=1000 kPa
l-Oct-10
Hz Producer-2
begin drill, complete, tie in operations
1-Jan-ll
Hz Producer-2
Produce with Constraint Min BHP=4000 kPa, Max oil rate=40 m 3 /d
l-Oct-22
Hz Producer-3
begin drill, complete, tie in operations
l-Jan-23
Hz Producer-3
Produce with Constraint Min BHP=4000 kPa, Max oil rate=40 m 3 /d
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
293
Table 13. Pump off event log. Program : Pump Off Date
Well
Event Description
l-Sep-09
13-20
Produce with Constraint Min BHP=1000 kPa
l-Sep-09
10-20
Produce with Constraint Min BHP=1000 kPa
l-Sep-09
07-20
Produce with Constraint Min BHP=1000 kPa
l-Sep-09
03-20
Produce with Constraint Min BHP=1000 kPa
Table 14. Maximum Rate Limitation (MRL) event log. Program : Status Quo Date
Well
Event Description
l-Sep-09
13-20
Produce with Constraint Max oil Rate=8 m 3 / d , Penalty GOR 100 m 3 / m 3
l-Sep-09
10-20
Produce with Constraint Max oil Rate=8 m 3 / d , Penalty GOR 100 m 3 / m 3
l-Sep-09
07-20
Produce with Constraint Max oil Rate=8 m 3 / d , Penalty GOR 100 m 3 / m 3 , Min BHP=7000 kPa (no p u m p on this well)
l-Sep-09
03-20
Produce with Constraint Max oil Rate=8 m 3 / d , Penalty GOR 100 m 3 / m 3
Table 15. C 0 2 Flood event log. Program : C 0 2 Flood Date
Well
Event Description
l-Oct-10
Hz Injector-1
begin drill, complete, tie in operations
l-Oct-10
Hz Producer-2
begin drill, complete, tie in operations
1-Jan-ll
Hz Producer-2
Produce with Constraint Min BHP=5000 kPa (Continued)
294
C 0 2 SEQUESTRATION A N D RELATED TECHNOLOGIES
Table 15. C 0 2 Flood event log. (Continued) Date
Well
Event Description
1-Jan-ll
13-20
Shut In
1-Jan-ll
10-20
Shut In
1-Jan-ll
07-20
Shut In
1-Jan-ll
03-20
Shut In
1-Jan-ll
Hz Injector-1
Inject with Constraints Max Rate=40 E3M3/D, Max BHP=15 MPa
l-Oct-19
Hz Producer-3
begin drill, complete, tie in operations
l-Jan-20
Hz Producer-2
Shut In
l-Jan-20
Hz Producer-3
Produce with Constraint Min BHP=5000 kPa
Figure 18. Screen shot of the dunvegan C pool in CMG results.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
295
Figure 19. North-South cross section of the pool under C 0 2 flood.
Figure 20. East-West cross section indicating trajectory of horizontal producer 002.
296
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 21. Enhanced oil recovery pool cumulative oil.
Figure 22. Enhanced oil recovery pool cumulative gas.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
Figure 23. Enhanced oil recovery pool cumulative water.
Figure 24. Enhanced oil recovery cumulative gross tonnes acid gas injected.
297
298
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 25. Enhanced oil recovery pool net tonnes acid gas sequestered.
Figure 26. Enhanced oil recovery Mol% C 0 2 in producer 2 gas.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
Figure 27. Enhanced oil recovery Mol% H2S in producer 2 gas.
Figure 28. Enhanced oil recovery k g / d C 0 2 in producer 2 oil.
299
300
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 29. Enhanced oil recovery k g / d H2S in producer 2 oil.
Figure 30. Enhanced oil recovery Mol% C 0 2 in producer 3 gas.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
Figure 31. Enhanced oil recovery Mol% H2S in producer 3 gas.
Figure 32. Enhanced oil recovery kg/d C0 2 in producer 3 oil.
301
302
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 33. Enhanced oil recovery k g / d H2S in producer 3 oil.
Figure 34. Sw36 Primary production pool cumulative oil production.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
Figure 35. Sw36 Primary production pool cumulative gas production.
Figure 36. Sw36 Primary production pool cumulative water production.
303
304
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 37. Sw46 Primary production pool cumulative oil production.
Figure 38. Sw46 Primary production pool cumulative gas production.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
Figure 39. Sw46 Primary production pool cumulative water production.
Figure 40 Sw56 Primary production pool cumulative oil production.
305
306
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 41. Sw56 Primary production pool cumulative water gas production.
Figure 42. Sw56 Primary production pool cumulative water production.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
307
Recovery alternatives that may be attractive, but were not investigated in this project, include surfactant flood and gas flood. In the case of surfactant flood, incremental oil recovery would be expected due to driving the residual oil saturation down (reduction in interfacial tension) along with high sweep efficiency. Oil viscosity is very attractive here, resulting in high sweep efficiency. In the case of gas flood, incremental oil recovery would be expected due to pressure maintenance in combination with low residual oil saturation in a gas-liquid system. A summary of the technical performance of recovery alternatives is provided in Table 16. The highest oil volumes recovered (248-305 E3M3), and recovery factors (35-74%), were realized by the Acid Gas and C 0 2 Flood programs, followed by horizontal wells (79-160E3M3, 19-23%), pump off without horizontal wells (58-102E3M3, 14-15%), and continued rate restriction operation (58-63E3M3, 9-14%). Refer to the Economics section of this report for economic performance of the recovery alternatives. Original Oil in Place for the C 0 2 flood is larger than the other Sw46 cases because it was run as a separate stand alone case using porosity and water saturation from well log interpretation. In the flood programs, Recovered oil volumes generally do not 'plateau' in the 2011-2030 time periods, indicating that further optimization could be done to accelerate recovery.
15.11
Economics
The economic summary of development cases is provided in Table 17. Cumulative cash flows are shown in Figures 43-45. Calculations are on a full project, go forward basis. The Pump Off cases represents the base case for each geological realization, against which the various development cases can be compared. Results are not incremental to the base case. The pump off cases generally realized the highest Net Present Value ($ 2.2-5.3 MM) of the pool, with the minimum capital investment. The exception is the 36% water saturation (Sw36) case, where drilling horizontal wells resulted in the highest NPV. However, the probability that the Sw36 case is reality is low given this case had the highest error in history match, and is not supported by the current geological evidence.
248
35%
285
E3M3
%
E3 Tonnes
Oil Recovered at 2030
Recovery Factor
Acid Gas Sequestered at 2030
700
E3M3
Sw36
336
74%
374
505
Sw46
285
72%
0
23%
160
700
424
305
Sw36
0
20%
102
505
Sw46
0
19%
0
15%
102
700
424
79
Sw36
Sw56
Horizontal Wells
Sw56
Acid Gas Flood
Original Oil In Place
Case
Table 16. Technical performance of recovery alternatives.
0
0
14%
58
71
14%
424
505
0
9%
63
700
0
12%
60
505
Sw46
Sw36
Sw46 Sw56
MRL
Pump Off
0
14%
58
424
Sw56
co 2
231
36%
255
716
Sw46
Flood
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
>40%
1
1.83
N/A
0
N/A
N/A
10,432
Rate of Return
Payout, years
Profit Investment Ratio (0%)
Profit Investment Ratio (10%)
Royalty Revenue to Alberta, M$ CAD
19,684
1.90
7,528
5,331
Net Present Value (10%), M$CAD
Sw36 Hz Wells
Sw36 Pump Off
Profitability Indicators
35,402
-0.06
0.68
12
8.5%
-1,094
Sw36 AG Flood
5,263
N/A
N/A
0
N/A
3,019
Sw46 Pump Off
Table 17. Economic summary of recovery alternatives.
10,225
0.76
0.60
1 year
>40%
3,011
Sw46 Hz Wells
55,971
0.04
1.22
3,766
N/A
N/A
0
11
7,003
0.24
0.18
2
>40%
N/A
10.6%
Sw56 Hz Wells 945
Sw56 Pump Off 2,173
735
Sw46 AG Flood
44,259
-0.05
0.98
11
9.1%
35,192
0.09
0.70
6
12.6%
1,519
-944
co 2 Flood
Sw56 AG Flood
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
310
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 43. Cumulative cash flow after tax, acid gas and C 0 2 floods.
Figure 44. Cumulative cash flow after tax, horizontal wells.
All acid gas flood economics are marginal. A positive change in oil price, capital cost, royalty framework, or other government policy are required in order for an acid gas flood development to proceed. Comparing royalty received from the base 'Pump Off cases versus potential royalty from the acid gas floods, government has significant negotiation room in the royalty framework as shown by Figure 46.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
Figure 45. Cumulative cash flow after tax, pump off.
Figure 46. Pool net present value and royalty revenue comparison.
Calculations were completed using the following parameters; • 2011 project start. • 10% Discount Rate. • No escalation.
311
312
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
• • • • •
$377/M3 ($60/STB) Oil. 40% Royalty. $63/M3 ($10/STB) operating cost for oil production. $ 3/tonne operating cost for acid gas disposal (8). $30/tonne capital cost for acid gas disposal, spend in 2011 (8). This capital cost classified as Class 41 with 25% declining balance deduction with Vi year rule. • $0 revenue / $0 cost for the acid gas or C 0 2 itself (operating and capital costs carried per above). • $3 MM per well drill, complete, and tie in capital cost. This capital cost classified as CDE with 30% declining balance deduction. • 28% Income Tax.
15.12
Economic Uncertainty
The impact of +/-50% changes to oil price, capital cost (Capex), royalty, and operating cost (Opex) on the acid gas flood economics were calculated, and the results reported in Table 18. The price of oil had the largest impact on project NPV, closely followed by capital cost and royalties respectively. Operating cost had the lowest impact on project economics, although this cost is significant to the project.
15.13
Discussion and Learning
15.13.1 Reservoir Fluid Characterization Regarding the oil characterization a case was initially completed where mole fraction, MW, S.G., and Boiling point of the D86 cuts were entered directly into plus fraction splitting, but the result appeared to be a sample with the light and heavy components missing from the distribution. This may have had more to do with how the cuts were generated rather than Winprop. A constant K assumption was used because it was quick and easy, and the D86 analysis did give a characterization factor = 12. This warrants additional study. It would be good to go back and re-visit the characterization method, and do a comparison on a volume basis. A comparison
2.59
0.62
1.22
0.04
55,971
Profit Investment Ratio (0%)
Profit Investment Ratio (10%)
Royalty Revenue to Alberta, M$ CAD
83,957
7
11
Payout, years
19.5%
10.6%
12,519
735
Net Present Value (10%), M$CAD
Rate of Return
$90 Oil
Sw46 AG Flood
Profitability Indicators
0.42
-0.54
27,986
2.13
-0.15
27,986
8
20
16.6%
8,591
-11,049
-2.0%
20% Royalty
$30 Oil
Table 18. Economic sensitivity of acid gas flood.
83,957
-0.35
0.30
17
3.3%
-7,121
60% Royalty
55,971
0.21
1.62
9
13.3%
4,201
$5/STB $1.5/T Opex
55,971
-0.13
0.81
13
7.6%
-2,731
$15/STB $4.5/T Opex
55,971
0.86
3.15
7
23.0%
8,685
$11 MM Capex
55,971
-0.24
0.57
7
5.7%
-7,216
$33 MM Capex
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
314
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 47. Economic sensitivity of acid gas flood.
Figure 48. Mol fraction oil distilled versus temperature.
of experimental vs. characterized data on a molar basis was done, and it appeared that the very light ends and very heavy ends were truncated.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
315
Complete component lumping before any simulation to avoid inconsistencies. Reservoir temperature of 34 °C is approaching hydrate formation temperature of approximately 30 °C for 80% H2S:20% C 0 2 acid gas mixture. Reservoir temperature should be confirmed, and hydrate formation conditions verified by laboratory experiment.
15.13.2
Material Balance
It was very useful in this case to use material balance to understand the relative influence of the various drive mechanisms for this pool. In future studies it would be useful to expand on this and complete calculations for various geological interpretations covering the range of geological uncertainty.
15.13.3
Geological Model
It is strongly recommended that data be collected from primary or secondary sources before conducting a reservoir simulation. It is important to discuss results of calculations with the geologist and geophysicist on the team, acknowledge that there can be more than one depositional interpretation, and get a number of possible geological pay maps to include in the history match exercise. The reason for water production from the wells is not well understood, and attempting to match it by adding aquifers may be leading CMOST astray. Generally, non equilibrium conditions as a result of well completion operations are not well understood or simulated. Consider using skin next time instead of manipulating near wellbore perms. Wellbore performance in general needs to be better understood. Regarding use of relative permeability tables in IMEX, the end points that are specified below the relative permeability tables in the .dat file generally over-ride the tables. The tables are re-scaled to match the end points. In the case of SLCON, IMEX seems to have ignored the SLCON value specified and run with the values in the table. To be sure of what end points IMEX is using, run one time step and have IMEX output the end points that it is using. Note that
316
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
the tables in the .out file may only be régurgitation of input, rather than what IMEX has actually used.
15.13.4
History Match
Next time adhere to a 'formula' approach on objective term weightings for the history match. Stick with production weighting based upon cumulative volume of reservoir fluid withdrawn at initial reservoir conditions until a more logical method is revealed. Adjusting weightings on terms, and including 'artificial' terms without data to get a perceived 'better match' only serve to put the basis of the simulation in question. Recommend against using any artificial objective function terms. In this project, bottom hole pressure data for the three wells on artificial lift was constructed in an effort to produce simulations with lower bottom hole pressures. This likely was not helpful. The reservoir simulation is working with monthly production volume data. Given that the pool is in initial production stages, and was until September 2009 on rate restriction, there were daily variations in well operating conditions that the reservoir simulation cannot capture. This may be an explanation for gas production history match discrepancies. Including well cumulative water production terms in the history match objective function was likely not beneficial in this case because the method for accounting for water production did not reflect actual reservoir conditions. The reservoir simulator needs to be improved to reflect 'non equilibrium' conditions present due to well completion operations. One possible explanation for the water production in this pool is that the gelled oil completions mobilized connate water and created a water bank that was produced back once the wells were put on production. Complete a sensitivity analysis in CMOST prior to the history match. Eliminating variables that are found to have minor influence allows CMOST to produce results more quickly, and produce optimized simulations that may be more stable in GEM. Rather than completing a manual history match exercise to 'get a feel' for simulation behavior, similar knowledge can be gained quickly from the sensitivity analysis. The method of modeling near wellbore reservoir properties appears to be flawed. A better approach may be to use skin factor from well test interpretation if available, or estimate skin based on the completion program.
ENHANCED O I L RECOVERY PROJECT: DUNVEGAN C POOL
317
15.13.5 Black Oil to Compositional Model Conversion One important lesson from this exercise is that lumping of such components will give better results. The geological model arrived at through IMEX simulations had to be compromised due to the onerous fluid system that the user imposed on the GEM simulations. Well deliverability, and near wellbore reservoir properties need to be more accurately represented in the simulations. An approach of modeling wellbore deliverability by skin, rather than permeability variations, may have resulted in simple stable GEM models. Extreme values and order of magnitude variations in near wellbore permeability likely contributed to GEM model instability. 15.13.6
Recovery Alternatives
Additional optimization can be done, this is conceptual study. The evaluation method used here was to optimize an enhanced recovery program for one particular geological realization, and then apply that identical program to the other geological realizations to understand the impact of making the wrong decision. This was done to demonstrate the core of what would be done in a full Decision Analysis type study. The method of looking at incremental capital investment steps (pump off->horizontal wells->acid gas flood) was also done to demonstrate the core of what would be done in a full Decision Analysis type study. 15.13.7
Economics
Economic uncertainty was evaluated for one particular geological realization for demonstration purposes. In a full Decision Analysis study, this uncertainty would be evaluated for all geological realizations.
15.14
End Note
This report has been condensed International Acid Gas Injection 2010, Calgary, AB. The author and not give any warranty express or
for presentation at the Second Symposium, September 28-29, his employer and affiliations do implied, and shall not be liable
318
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
for any loss, claims, costs, damages or any other action caused by direct or indirect use of this material. Application of the information contained in this material is entirely at the risk of the user
References 1. Fekete Associates Inc. Application for Good Production Practice - Primary Depletion Pool Dunvegan C Pool, Grande Prairie Field. Calgary : s.n., 2009. 2. IHS. Oil Reserves Summary, Dunvegan C Pool, Grande Prairie Field. Calgary : s.n., 2010. 3. Slider, H.C. Worldwide Practical Petroleum Reservoir Engineering Methods. Tulsa, Oklahoma : PennWell Publishing Company, 1983. 4. Aziz, K., Settari, T. Petroleum Reservoir Simulation. London : Applied Science Publishers, 1979. 5. Evaluation of Normalized Stone's Methods for Estimating Three Phase Relative Permeabilities. Fayers, F.J., Matthews, J.D. 1984, Society of Petroleum Engineers Journal, pp. 224-232. 6. An improved model for estimating three phase oil-water-gas relative permeabilities from two phase oil-water and oil-gas data. Maini, B.B., Kokal, S.L.,. 1990, The Journal of Canadian Petroleum Technology, pp. 105-114. 7. Energy Resources Conservation Board. ST98: Alberta's Energy Reserves and Supply/Demand Outlook. Calgary : s.n., 2006. 8. Study shows 'huge' C 0 2 storage potential in Alberta. Carbon Capture Journal. March/April, 2010,14. 9. Alberta Research Council. Hydrochemistry of the Peace River Arch Area, Alberta and British Columbia, Open File Report 1990-18.1990. 10. Energy Resources Conservation Board. Directive 65: Resources Applications. Calgary : s.n., 2009. 11. —. Directive 007-1: Allowables Handbook-Guidelines for Calculation of Monthly Production Allowables. Calgary : s.n., 2007. 12. A New Method for Petroleum Fractions and Crude Oil Characterization. Castells, F., Hernandez, J., Miquel, J. 1992, SPE Reservoir Engineering, pp. 265-270. 13. Alberta Geological Survey, Energy Resources Conservation Board. ERCB/ AGS Special Report 094: Stress Regime at Acid Gas Injection Operations in Western Canada. Edmonton : s.n., 2008.
16 C0 2 Flooding as an EOR Method for Low Permeability Reservoirs Yongle Hu1, Yunpeng Hu2, Qin Li2, Lei Huang1, Mingqiang Hao1, and Siyu Yang1 ^hina Petroleum Exploration and Development Research Institute Beijing, People's Republic of China 2 China University of Geosciences Beijing, People's Republic of China
Abstract
Carbon dioxide flooding is an efficient enhanced oil recovery (EOR) method for low permeability reservoirs. C0 2 swelling oil, reducing oil viscosity significantly, and obtaining miscibility at specified temperature and pressure can decrease the surface tension considerably. Simultaneously, injecting C0 2 into reservoirs is an important way for C0 2 sequestration. The C0 2 flooding technique has not been widely implemented in China. Technology suitable for low permeability reservoirs in China should be developed further.
16.1
Introduction
Carbon dioxide injection can effectively make up the voidage of low permeability reservoirs. Because of the injection difficulties and poor pressure transmission of low permeability reservoirs, it is difficult for water flooding to build up an effective displacing system to maintain reservoir pressure. In practical, the pressure level of low permeability reservoir developed by water only maintains at 70% of the primary pressure, which seriously affects the oilfield development effects. The viscosity of C 0 2 is far less than that of water, so C 0 2 can be injected into low permeability formation more easily and pressure can be recovered efficiently.
Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (319-328) © Scrivener Publishing LLC
319
320
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Carbon dioxide injection can effectively decline the lower limit of the reservoir put to use to improve oil use rate. C 0 2 injection can effectively decrease lower limit of used extent of formations, and increase the oil producing degree. C 0 2 can flow into tiny pores that water cannot, to mobile the reserves and improve the injection profile. For water sensitive low permeability formation, C 0 2 is a favorable alternative. Carbon dioxide flooding can increase displacement efficiency and oil recovery factor. The mobile oil saturation of low permeability reservoir is low, the efficiency of water flooding and water flooding recovery are low, too. For those main oil fields in China, in terms of development stage under high water content, it has been proved that C 0 2 flooding can further improve the recovery of old fields with high water content. Flooding experiments in long cores showed that: compared with water flooding, C 0 2 flooding can dramatically increase oil displacement efficiency. While C 0 2 is implemented in a field with water content 95%, oil recovery can increase 10% [1].
16.2
Field Experiment of C 0 2 Flooding in China
In China C 0 2 flooding was focus in the early 1960s and some experiments and pioneering tests were carried out. In 1963, Daqing Oilfield first researched on C 0 2 flooding and designed some pilotscale experiments in the field. These experiments showed C 0 2 flooding technology can improve the recovery by about 10%. From 1990 to 1995, experiments of water alternating C 0 2 gas were implemented in the Well Zone 45, 3-3 C, located in eastern transition zone of Sanan area, and the water content in field was up to 98%. Oil recovery was enhanced by 6% and C 0 2 utilization efficiency was 0.23 t / t C0 2 . In 1999, C 0 2 flooding experiments are tried in Xinli Oilfield, Jilin Province. There was 5200 ton more oil extracted while 1500 ton C 0 2 was injected into subsurface. Also, experiment of water alternating C 0 2 gas to form miscible displacement was taken in block 14, Jiangsu Oilfield in 1998 with water content in field above 95%. The oil recovery was increased by 4% while the C 0 2 utilization efficiency is 0.4t/tCO 2 . Since 2006, experiments of Water and gas synchronizing injection have commenced in Caoshe Oilfield. Although the experiment is still in progress, it has obtained miscible displacement and promoted well production
C 0 2 FLOODING AS AN EOR METHOD
321
from the data available. In 2008, C 0 2 flooding was used as primary oil recovery in the Tree 101 block and Songfangtun block, Daqing Oil field, and preliminary effect of enhancing oil recovery has been realized. In 2008, C 0 2 flooding experiments started in the Black 59 block, and the effect of gas injection is remarkable. Well production was substantially promoted compared with its initial production. At present, oil fields in China such as Daqing, Jilin, Shengli, Liaohe, Jiangsu have fulfilled some significant work of C 0 2 flooding on research and implementation in field. However, C 0 2 flooding in China is still immature, since the related research has just advanced in a short period and the fields in which tests were implemented are small. More attention should be paid to the future research, such as C 0 2 flooding tests, integrated technique of C 0 2 flooding and resolution of the key technology of C 0 2 flooding.
16.3 Mechanism of C0 2 Flooding Displacement C 0 2 is a kind of gas with high solubility in both water and oil. A large amount of C 0 2 dissolving in crude oil can result in crude oil's volume inflation, viscosity decrease and interfacial tension decline. In addition, the Carbonic acid generated after the C 0 2 dissolves in the water could play a role of acidification. If the compositions of the crude oil are favorable, C 0 2 could be mixed with oil at certain pressure and the recovery efficiency would be significantly increased. It has been proved that C 0 2 is an efficient medium for enhancing oil recovery through a large number of laboratory and field experiments. C 0 2 flooding falls into three categories: miscible phase displacement (semi-miscible phase included), non- miscible phase displacement and carbonated water displacement. The high efficiency of C 0 2 displacing oil in porous medium mainly attribute to following merits: 1. Distention C 0 2 can significantly dissolve into crude oil. The full dissolution can give rise the crude oil to a high volume expansion which is commonly about 10% - 40% [2]. The volume expansion can play an important role in oil displacement. Firstly, the residual oil in-situ after water flooding is reciprocal to the expansion coefficient, i.e., the higher the expansion coefficient, the less the residual oil in-situ.
322
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Secondly, the dissolved oil drops could extrude the water out of the porous space to form a water wet system with water drainage rather than water suction. This can lead to higher oil relative permeability curve of oil drainage than that of oil suction. Therefore, a beneficial oil flowing environment is imposed in any given saturation conditions. Thirdly, the oil volume expansion, on one hand, can increase stratum's elastic energy significantly; on the other hand, the residual oil after expansion could completely or partly escape from the bound water and become mobile oil [3]. 2. Viscosity reduction effect When the crude oil is saturated with C 0 2 gas, its viscosity of crude oil can be greatly reduced. Under the subsurface condition, the higher the pressure is, the more the C 0 2 dissolves in the crude oil, and the reduction of the crude oil viscosity will be more significant [4]. The crude oil viscosity may reduce by 1.5-2.5 times after C 0 2 dissolve in it. In general, the viscosity reduction ratio is proportional to the viscosity of crude oil, i.e., the viscosity reduction ratio in heavy crude oil is much greater than that in light crude oil after C 0 2 dissolution. Therefore, it is suggested that C 0 2 should be used to develop the heavy crude oil, since the viscosity of heavy crude oil with saturated C 0 2 can decline remarkably. The mobility ratio improves and oil relative permeability will be correspondingly promoted, too. 3. Improvement on the mobility ratio and reduction on the interfacial tension When C 0 2 dissolves in water, the water's viscosity can increase 20%-30% and its mobility increase by 2 to 3 times. In the meanwhile, with the decreasing of oil mobility, oil /water mobility ratio and their interfacial tension will be further reduced, so that the oil could flow more easily. 4. Improvement on injection capacity and acidification [5] C0 2 -water mixture is slightly acidic and it can react with the formation matrix as follows: C02+H2O^H2C03 H2C03 + CaCOs -> Ca(HC03 ) 2
(1) (2)
C 0 2 FLOODING AS AN EOR METHOD
H2C03 + MgC03
-> Mg(HC03
)2
323
(3)
The generated bicarbonate can easily dissolve in water and increase the permeability of reservoir, particularly those formations whose vicinity around bore hole a great amount of water and C 0 2 pass by. In addition, due to acidification, C0 2 -water mixture can relieve inorganic scale blocking, dredge the oil flowing pathway and recover single well production to a certain extent. 5. The role of dissolved gas drive [6] The solubility of C 0 2 in crude oil is very high. With gas injecting, part of the C 0 2 will dissolve in crude oil and the amount of C 0 2 dissolution will increase with increasing injection pressure. After C 0 2 injection into reservoir, the reservoir pressure will reduce with oil extraction. As a result, the C 0 2 dissolving into the crude oil will be separated from the crude oil, which can play a role as gas drive similar to the natural type of solution gas drive. 6. Extraction and vaporization of the light components of crude oil There is a good miscibility between light hydrocarbons and C0 2 . When pressure exceeds a certain value which depends on the oil properties and temperature), C 0 2 can make the light components extraction and vaporization, which is more prominent for the light crude oil. Extraction and vaporization of light hydrocarbons in crude oil is one of the main mechanisms of enhancing oil recovery through C 0 2 injection. 7. Miscibility [7] Under the reservoir temperature, the pressure at which C 0 2 and oil reach miscible phase is called minimum miscibility pressure (MMP), which depends on the pureness of C0 2 , oil component and reservoir temperature. When the reservoir temperature goes up, the MMP increases; in addition, it also increases while the molecular weight of component above C5 in crude oil increased. MMP can be influenced by pureness (impurity) of C0 2 . MMP will decrease while the critical temperature of impurity is lower than that of C0 2 , and vice versa. The mixture of C 0 2 and primary oil can not only extract and vaporize light hydrocarbon, but also can realize an oil zone mixed with C 0 2 and light hydrocarbon, which is the most effective oil displacement process when the oil zone is mobile.
324
16.4
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Perspective
Different from marine sediment reservoir, most of the oil fields found in China belong to continental sedimentary reservoir feathered with a complex tectonic geological features, serious heterogeneity, high content of heavy component in crude oil and larger viscosity. In order to implement C 0 2 flooding successfully, we need to resolve the following issues: 1. Determine the screening criteria to implement C 0 2 flooding and evaluation methods of enhancing oil recovery of C 0 2 flooding based on the geological characteristics in China. 2. Currently, there are a number of oversea screening criteria with respect to C 0 2 enhancing oil recovery; however, it is still uncertain that those standards are suitable for continental sedimentary reservoirs in China, especially for the low permeability reservoir. Therefore, the domestic geological features should be considered while we determine the range of application of C 0 2 miscible flooding, immiscible flooding and throughput, formulate the reservoir screening criteria of C 0 2 enhancing oil recovery and C 0 2 sequestration and the evaluation methods of EOR under low permeability reservoirs conditions. All standards should base on the geological features in china: under continental sedimentary geology, synthesizing the complex tectonic geological features (including basin characteristics, geological structure, sedimentary faces characteristics, fault characteristics, cap sealing characteristics, etc.), reservoir characteristics factors (including factors affecting the ability of the reservoir injection, such as the reservoir permeability boundaries, heterogeneity parameters, etc.), and the quality of crude oil factors (including the influence of crude oil components, and C 0 2 purity, etc.). 3. Develop phase evaluation and characterization technology for the Conformation fluid mixing system Components exchange will take place between C 0 2 and crude oil during the process of C 0 2 flooding, which may give rise to a complex process of phase change. Therefore, the principle to compile scientific
C 0 2 FLOODING AS AN EOR METHOD
325
scheme for C 0 2 flooding should base on phase evaluation of Conformation fluid mixing system under formation conditions. The current Conformation fluid mixing system phase evaluation is mainly accounted for through PVT experiments, including sampling, mixing with samples, testing and data calculations and analysis. Inaccuracy in each step could lead to incorrect results. How to make an experiment more closely reproduce C0 2 -crude oil system phase under formation conditions should be focused. With respect to the phase characterization, in order to build a basic principle for the following C 0 2 flooding simulation, integrated methods should be proposed on the evaluation of thermal stability of well flow properties, the division and combination of pseudocomponents, the solid precipitation characterization, the adjustment of phase equation, the calculation methods of the minimum miscibility pressure (MMP), and so on. 4. Develop the applicable C 0 2 flooding fine reservoir characterization technology and the numerical simulation technology Most oil fields in China belong to continental deposit featured with strong heterogeneity, small sand body distribution, and more interbeds. Investigations on the distribution characteristics of the sand body, the development characteristics of intercalation and the connectivity between injection wells and production wells should be based on the reservoir characteristics with thin bed and narrow channel, so that a reliable geological recognition should be provided for implementing C 0 2 flooding. To the numerical simulation technology, we should further develop the multiphase and multi -component simulation considering advection and diffusion effect among different phases, since there are multi-liquid flow and solid-Phase Precipitation in the process of C 0 2 flooding. Therefore, applicable C 0 2 simulation methods can be built as a good technical storage for the implementation of C 0 2 to the large quantities of complex type of reservoir in China. 5. Develop dynamic monitoring technology for displacement front of C 0 2 gas drive It's very important to monitor the displacement front of C 0 2 gas drive in the process of C 0 2 flooding. Development of the monitoring
326
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
technology for the displacement front of C 0 2 gas drive based on the seismic data, well logging data and production performance data is appealing very urgently. 6. Improve the drilling and the surface engineering for the C 0 2 flooding As research on the drilling and the surface engineering for the C 0 2 flooding in China starts late, the application effect of process equipment related to C 0 2 flooding needs to be tested and the C 0 2 anticorrosion technology as well as C 0 2 separation technology needs to be further developed, too. 7. Develop new technology of C 0 2 flooding for enhancing the oil recovery In view of the shortcomings of the conventional technology, the new generation of C 0 2 flooding technology for enhancing the oil recovery has the following improvements: a. using horizontal wells to adjust the well pattern and the displacement methods, and improve the sweep extent to the remaining oil and the displacement efficiency, b. increasing the mobility ratio and controlling the viscous fingering of C 0 2 to expand swept volume, c. reducing the minimum miscibility pressure (MMP) by adding miscible agents, d. paying attention to integrating all technologies.
16.5
Conclusion
In conclusion, the implementation of the C 0 2 enhancing oil recovery in China is still in initial stage. We need to further our research urgently and try our best to provide the technical support for the large-scale industrial implementation of C 0 2 flooding in the low permeability oil field.
References 1. A.T.F.S. Gaspar, S.B. Suslick, D.F. Ferreira, and G.A.C. Lima, "Economic Evaluation of Oil Production Project with EOR: C 0 2 Sequestration in Depleted Oil Field," SPE94922.
CO
FLOODING AS AN EOR METHOD
327
2
2.
3.
4.
5.
6. 7. 8.
Cai Xiulin, "The Mechanism and Application of Single Well C 0 2 cyclic injection technology for production improvement," Petroleum Drill Technology, vol. 24(24), pp. 45-46,2002 (in Chinese). Wang Shouling, et a l , "Research on the mechanism of production improvement and application of C 0 2 cyclic injection technology," Drill Technology, vol. 1, pp. 91-94,2004 (in Chinese). Yu Yunxia, "The Application of Single Well C 0 2 cyclic injection technology for production improvement in oil field," Drill Technology, Vol. 27, pp. 89-90, 2004 (in Chinese). Liang Fuyuan, "The Application of C 0 2 cyclic injection technology in Fault Block Hydrocarbon Reservoir," Producing Test Technology, Vol. 22(3), pp. 31-33, 2001 (in Chinese). Chen Tielong, "The Tertiary Oil Recovery Introduce," Petroleum Industry Press, 2000 (in Chinese). F. Stalkup, "Field Developing by Miscible Displacement," Petroleum Industry Press, Beijing, 1989. J.H. Goodrich, "Target reservoir for C 0 2 Miscible Flooding," Report DOE/ MC/08341-17, U.S.DOE, Washington,DC, 1984.
This page intentionally left blank
17 Pilot Test Research on C 0 2 Drive in Very Low Permeability Oil Field of in Daqing Changyuan Weiyao Zhu1, Jiecheng Cheng2, Xiaohe Huang1, Yunqian Long1, and Y. Lou1 1
Civil and Environmental Engineering School, University of Science Technology Beijing, People's Republic of China 2 Daqing Petroleum Administration Bur of Petroleum Daqing People's Republic of China
Abstract
The oil reserves about 3.7xl08tonne do not obtain economic development by water flooding in Daqing Changyuan. For obtaining an availability development method to fit a very low permeability oil field, according to the research results of some experiments and reservoir engineering, some testing schemes are designed and numerical simulations are investigated. Based on the testing results of C0 2 injection, some injection gas feasibility and immiscible displacement condition for C0 2 drives are presented. The technology ambit and product change curve is given. The appropriate technology and C0 2 injection condition is gained. Thus, in a very low permeability oil field the C0 2 drive has succeeded in enhancing well production and oil recovery.
17.1 Introduction According to the statistical results of more than 70 foreign oil field, for the very low-permeability oil field, gas injection (especially C 0 2 flooding) is the main technical measure to improve development effectiveness and to establish effective deployment system [1-8]. Utility of gas-water alternating injection and miscible injection could enhance oil recovery by 7 to 15 percent. Proportion of C 0 2 Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (329-350) © Scrivener Publishing LLC
329
330
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
flooding in low permeability oil reservoirs is high, but decreases as permeability increases. Proportion of nitrogen injection is close to that of hydrocarbon gas injection, and decreases as permeability increases. C 0 2 flooding projects usually have the following common conditions: depth less than 2000m; crude oil density between 0.8 and 0.9 g/cm 3 . But some nitrogen and hydrocarbon gas injection projects are used when crude oil density is smaller or larger. The number of C 0 2 flooding projects increases as viscosity increases and the fact shows that CÖ 2 flooding is capable of exploiting highviscosity oil. Gas injection in domestic low-permeability oil fields was blocked because of gas supply shortage in the last few years. Now more and more field tests for gas injection projects are carried out in low-permeability reservoirs. Laboratory test results show that gas injection in very low permeability reservoirs differs significantly from that in common permeability and low permeability reservoirs: gas flow has significant non-Darcy flow characteristics; oil and water have obvious threshold pressures. The characteristics above are also revealed in field tests. Formation conditions and fluid characteristics of lowpermeability oil reservoirs in the periphery of Changyuan Daqing satisfy the selection criteria for C 0 2 flooding, which has better adaptability than hydrocarbon gas flooding. Consequently it is necessary to carry out C 0 2 flooding theory research and field tests in order to summarize experience and lay a solid foundation for further development.
17.2
Laboratory Test Study on C 0 2 Flooding in Oil Reservoirs with Very Low Permeability
Study of phase behavior and experiments of swell, tubule flow and long core flow were carried out on the natural rocks of the Fuyu oil layer and the oil/gas samples collected from the object regions.
17.2.1 Research on Phase Behavior and Swelling Experiments Experiments of simple degassing, P-V relationship, multi-stage degassing and swelling were performed on the object samples.
PILOT TEST RESEARCH ON C 0 2 DRIVE
331
The simple degassing experiment aims to obtain the main parameters such as gas/oil ratio, volume factor, density of initial oil in place and so on; the P-V experiment aims to measure the parameters such as the saturated pressure of the fluid, the fluid density and volume factor under the saturated pressure and so on; the multistage degassing experiment aims to measure the dissolved gas/ oil ratio, volume factor, density, viscosity and the change of liquid volume under the conditions of multi-stage degassing. Simple degassing experiments: Under the formation temperature 86 °C and the formation pressure 22.64 MPa, a simple degassing experiment produced a gas/oil ratio 18 m 3 /m 3 , a crude oil volume factor 1.088 m 3 /m 3 , a crude oil shrinkage factor 8.06%, an average solubility of gas 3.51 m 3 /(m 3 .MPa), and a crude oil viscosity 3.314 mPa.s. These data indicate that the Fuyu oil layer is a reservoir with high oil density and low volume factor, swell, shrinkage and solubility. P-V experiments: According to the measured data, the crude oil volume factor changed a little (1.0566-1.0673) with varied pressures, which suggests a small amount of energy for the volume swelling. Multi-stage degassing experiments: The measured parameters for crude oil under varied pressures show that the viscosity and density of the crude oil increase with decreased pressures while the gas/oil ratio and the volume factor decrease with decreased pressures. That is, the crude oil possesses such characteristics as intermediate density, high viscosity, small expansibility and shrinkage, and low density for the displaced gas. Swelling experiments: For Shengqi Well 1-4 and Fangshen Well 6, the experiments under C 0 2 flooding presented swell factors 1.10, 1.15 and 1.26, respectively. Generally, the parameters for the raw oil hardly changed with varied pressures.
17.2.2
Tubule Flow Experiments
The tubule flow experiments were designed to determine whether the injected gas is miscible with the crude oil. According to such experiments, the lowest miscible pressure at the formation temperature (86 °C) was 47 MPa, with an oil-displacement efficiency 92% (see the pressure dependent oil-displacement efficiencies plotted in Figure 1). Therefore, the oil-displacement experiment under C 0 2 flooding in the Fuyu oil layer was immiscible displacement.
332
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 1. The variation of oil-displacement and pressure.
17.2.3
Long Core Test Experiments
The conditions in long rock test experiments are usually much more close to the actual situations in the formation. The used rocks that were 28.85 cm in the length with an average air-permeability 2.694xl0~3 urn2. Five series of oil-displacement experiments were carried out on the long rocks, which used gas flooding for Fangshen Well 6, gas flooding, gas/water alternate flooding, pure water flooding and pure C 0 2 flooding for Shengqi Well 1-4, respectively. The following conclusions can be made based on the experiments: 1. the displacement under gas flooding is more facile than that under water flooding, which shows that the threshold pressure difference under gas flooding is smaller than that under water flooding (2.06-2.19 MPa vs 5.45-5.77 MPa). 2. The displacement efficiency under the gas/water alternate flooding is not ideal. For Shengqi Well 1-4 under the water/gas alternate flooding with a threshold pressure difference of 5.77 MPa, the injection pressure kept increasing until that is close to the formation fracturing pressure, which did not lead to both water and gas to break through, and the final recovery factor was only 25.96%. 3. Before the breakthrough point, the recovery factor under gas flooding is higher than that under water flooding (27.4-29.081% vs 23.28% at the breakthrough point). 4. The recovery efficiencies under C 0 2 flooding increase with increased injection pressure. For example, with the injection pressure going up from 6.0 MPa to 35 MPa, the recovery factor at the breakthrough point increased from 32.61% to 44.76% and the final efficiency increased from 39.06% to 56.27%.
PILOT TEST RESEARCH ON CO. DRIVE
333
Table 1. Data comparison of different injected medium. Projects
Recovery Factor at the Breakthrough Point (%)
Starting Pressure (MPa)
Stating Pressure Gadient (MPa/m)
Fangshen 6 C 0 2 flooding
2.19
7.59
29.08
34.32
Shengqi 1-4 C 0 2 flooding
2.06
7.14
27.41
32.20
Shengqi 1-4 gas/water alternate flooding
5.77
20.00
Not breakthrough
25.96
Water flooding
5.45
18.89
23.28
Final Recovery Factor (%)
/
Table 2. Data comparison of different injection pressure. Starting Pressure (MPa)
Stating Pressure Gradient (MPa/m)
Recovery Factor at the Breakthrough Point (%)
Final Recovery Factor (%)
6.0
2.43
8.42
32.61
39.06
22.64
2.29
7.94
41.80
48.15
35.0
2.26
7.83
44.76
56.27
Injection Pressure (MPa)
5. C 0 2 flooding can strongly improve the oil recovery factor. For example, the value at the breakthrough point under C 0 2 flooding was 41.08%, 12.72% higher than that for Fangshen Well 6; the final value was 48.15%, 13.83% higher than that for Fangshen Well. 6. All experiment results are listed in table 1 and table 2.
17.3
Field Testing Research
17.3.1 Geological Characteristics of Pilot Fang 48 fault block is located on the southeast of Songfangtun oilfield, and on the Zhaozhou nose structure, which is in the Sanzhao
334
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
depression of central northern Songliao basin. The three exploration well, Fang 48, Zhaoshen 4 and Zhao 401, which drilled through Fuyu oil layer, were successfully finished drilling from 1989 to 1990 in Fang 48 fault block. 3D high-resolution seismic exploration with a bin of 20x20m, was finished to explore deep gas in 1996, and the developmental condition of the structure and fault was identified in Fang 48. In 1998, for Fang 48 fault block, with other six wells, such as Zhou 7, have been submitted and proved reserve was 812xl0 4 t, oil bearing area was 23km 2 (unit coefficient is 5.5xl0 4 t/km 2 .m). In 1999, for designing Putaohua oil layer development wells, giving consideration to Fuyu oil layer, 5 development control wells were drilled, and oil test was conducted in the 5 wells. At present, there are 8 exploration and development control wells, of which 5 wells test oil yield is more than 1.5t/d. 17.3.1.1
Structural
Characteristics
Fang 48 fault block is located on nose structure of east Songfangtun oilfield. Songfangtun nose structure uplifted extent is small, actually a moderate slope, so Fang 48 fault block is flat air streamed structure. The nose structure has the maximum uplifted extent at -1600m contour line in the structure map of Fuyu oil layer. Fractures are developed around Fang 48 fault block. There are two near northsouth faults, MF13 and MF16, which consist of Fang 48 horst block, in the test area from Tl-1 and T2 reflection layer structure map. The extension of the two faults is about 5 km, and the fault throw is the big end up mold. There are up to 34 minor faults in Fang 48 horst block. The fault throw is about 50m of the top surface fault in Fuyu oil layer, and the horst block scale is larger than that of PI group. Though the faults around test area developed well, that of gas injection test area didn't develop well. 17.3.12
Characteristics of Reservoir
Fuyu oil layer in Sanzhao area is cretaceous for Quan 4 and upper Quan 3 segments, and the distribution is relatively stable. Fuyu oil layer in Fang 48 fault block is in the sandstone enrichment zone, which is effected by northern Songfangtun and southern Zhaoyuan water systems. Fuyu oil layer is river-lake flood plain faciès deposition, which is formed in the HST ancient lake, and the lithology is dark purple, purple mixed green and gray mudstone and gray green, green, gray muddy siltstone, siltstone and gray-brown,
PILOT TEST RESEARCH ON C 0 2 DRIVE
335
oily brown powder, and fine sandstone. The formation thickness of Quan 4 segment (Fu 1 group) is about 100m, and the Quan 4 segment is divided into 7 small layers. The formation thickness of upper Quan 3 segment (Fu 2 and 3 groups) is about 100m, and the segment is divided into 5 small layers. The top surface depth of Fuyu oil layer of Fang 48 fault block is about 1742m. Only Fu 1 group developed oil layer. 10 wells were drilled, and the average drilled sandstone thickness is 13.9m and the effective thickness is 9.3m. Vertical evolution sequences: fluvial facies-Lake flood plain facies-delta facies. The main type of oil layer sand body is channel sand, and the shape of sand body is short strip and intermittent banding. The drilling ratio is 90% of the main oil layers, F14 and F17, and the drilled thickness is 24.5% and 64.4% of total effective thickness respectively. The layer FI4 belongs to the branched channel sand of lake flood facies, the micro-gradient curve is bell-shaped, the lithology is positive cycle, and from bottom to top is calcareous siltstone - sandstone oil powder - muddy siltstone, the bedding is parallel bedding, small oblique bedding and wavy bedding, the sandstone thickness is l~5.8m, the effective thickness is 0.6~2.6m, the average drilled sandstone thickness is 3.5m and effective thickness is 2.1m, and the width of sand body is about 500m based on the well drilling. The layer F17 belongs to the branched channel sand of fluvial facies, the micro-gradient curve is box-shaped, the lithology is dual structure, the bedding is parallel and small oblique bedding, the sandstone thickness is 5.8~10.0m and the effective thickness is 4.8~10.0m, the average drilled sandstone thickness is 7.3m and the effective thickness is 6.0m, the width of sand body is 600m based on well drilling. The development wells (well pattern:300x300m) in the Zhou 16 Pu-Fu commingled test area, which is 9km far away from Fang 48 fault block, is proved that the width of Fu 1 group is about 600m too (See Figure 2). The buried depth of F17 sand body is -1696- -1703m, and the sand body tilted from east to west. The effective sandstone thickness is the largest in the vicinity of Fang 190-138, and is 10.0m. The minimum thickness is in the vicinity of 187-138 (6.0m and 5.6m respectively). The sand body is getting thicker from north to south. Based on the distribution of porosity and permeability of layer F17, in general, the reservoir properties of layer F17 in Fang 48 block is not good, and it belongs to low porosity, ultra-low permeability reservoir. The porosity ranges from 5.8%~17.4%, and the average
336
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 2. Test well location map of gas injection. is 14.5%; the range of permeability is 0.02~3.66xl0-3um2, and the average is 1.4xl0-3um2. Based on the surface distribution, there is little change in porosity and permeability from north to south (See Figure 3 and Figure 4). 27.3.2.3
Reservoir Properties and Lithology
Characteristics
The typical lithological features of Fuyu oil formation from the Fang 48 well is argillaceous siltstone and fine sandstone containing mud with secondary quartz developing well in the pore space. The pores are mainly narrowing intergranular pores, most of which are not connected, and the rock core analysis indicated that the porosity is 9.0-17.6% with an average value of 14.5% while the average air permeability is 1.4xl0"3um2 with the maximum value is 5.22xl0"3|i,m2 and the minimum is 0.1x10 3 |im 2 (the permeability ratio is 279.5 and mutation coefficient is 5.7). Besides, the sandstone is mainly composed of quartz (21-26.7%), feldspar (29.2-36.2%) and the average rock debris is 33.8% while median grain diameter is 0.068-0.111 and sorting coefficient is 5.7-10.08. Slice analysis suggested that the dominant cementation types of the reservoir are shale cementation and calcareous mixing cementation with average 9.7% shale content. And calcic cementation accumulated locally while shale cementation is mainly recrystallization and distributed in clusters and films. The quartz and feldspar have
PILOT TEST RESEARCH ON C 0 2 DRIVE
337
Figure 3. Distribution map of FI7 porosity.
Figure 4. Distribution map of FI7 permeability.
characters of secondary enlargement and regenerated cementation, which are mainly types of pore-film, regeneration-pore-film. The clay minerals of the Fuyu oil formation are mainly illite (31%) and chlorite (39%), 70% of which is enriched in iron and mixing layers of montmorillonite/illite and montmorillonite /chlorite also composed the clay minerals. The primary attitudes are listed in table 3.
3.5
7.3
FI4
FI7
14.5
14.7
31 39
0.9 1.4
30.8 36.0 41.6 36.1
0.865
0.874
0.872
0.869
Fang 190-138
Zhao 401
Fang 48
Average
125
157
148
80
33
30
35
35
32
124
38.1
0.870
Fang 184-136
35
115
33.8
0.866
Freezing Point (°C)
48
0
IBP (°C)
Oil Viscosity (mPa.s)
Fang 190-140
^~~~-~-^Pro j ects Oil Density Well N o / " ~ \ ^ ^ (t/m3)
Clay Mineral Components (%)
17.0
18.1
19.2
14.0
17.8
16.0
Gel Content (%)
4
4
25.1
28.6
26.2
30.8
20.7
19.4
Wax Content (%)
7
65
Porosity Permeability Illite Chlorite Montmorillonite/ Montmorillonite/ Illite Chlorite 3 2 (10 iim ) (%)
Physical Properties
Table 4. Character table of oil properties from fault block Fang 48.
6
2.1
Sandstone
Thickness (m)
Effectiveness
Level No.
Table 3. Reservoir character of FI4 and FI7 from fault block Fang 48.
M
h-1
n
O1 r O
n
a
25
M
z > a *>
O
I—I
H
CD
M
O ci
M
n
p
oo
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
PILOT TEST RESEARCH ON C 0 2 DRIVE
339
Capillary pressure curve of Fuyu oil formation in Sanzhao didn't show obvious platform and the pore distributed in single peak shape with complicated pore throat structure and well developed micro pore, which indicated that the original pore and secondary pore are all well developed with the maximum pore throat radius is 2.14um and the average is 0.257um. Core observation suggested that fractures didn't develop well and sixteen years' of water flooding experiences from Shengnan testing field proved that the fractures didn't cause bad influences.
17.3.2 Distribution and Features of Fluid Oil reservoir in Fuyu area was mainly formed by controls of lithology and distribution of oil was also controlled by fault block. Generally, oil column are higher in horst block and no united oil-water interface exists. The types of the reservoirs were horstlithology reservoirs. Oil-water distribution in Fang 48 fault block are characterized by pure oil layer in Fuyi formation and dry layer or water layer in F2 and F3 formation. Statistical analysis of crude oil property from five wells in Fuyu oil layer shows that averaging density of crude oil, crude oil viscosity; freezing point, glue content and wax content are 0.869t/m 3 , 36.1mPa.s, 33.0 °C, 17.0% and 25.1% respectively. Analysis of high pressure property of the samples from well Fang 48 and well Zhou 7 shows that averaging density of crude oil, crude oil viscosity, saturation pressure; volume factor and original gas oil ratio are 0.815t/ m3, 6.6mPa.s, 5.3MPa, 1.089 and 17.5 m 3 /t respectively. See table 4. Averaging CLcontent in formation water in Fuyu oil layer is 3067.6mg/L. Total salinity is 7158.0mg/L. Water type is NaHC0 3 . Original strata pressure is between 17.06 and 24.19MPa (average 20.4MPa); pressure gradient ranges from 0.9426 to 1.3151MPa /100m, with an averaging value of 1.1212 MPa/100m. Reservoir temperature ranges from 81.1 to 87.8°C, with the mean value of 85.9°C. Geothermal gradient are 4.51-4.85°C/100m (average 4.72°C/100m), which belongs to normal geo thermal field.
17.3.3 Designed Testing Scheme According to experimental results, recovery factor increases significantly when the injected carbon dioxide slug is lower than 0.3PV and recovery factor increases little when it is more than 0.3PV.
340
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
So, it is determined to inject carbon dioxide slug before 0.3PV and then transform it to water drive after that. Specific schemes are as follows: 1. Injected medium: liquid carbon dioxide (1-6 year); water (after 6 year); 2. Injection-production ratio: 1.5 (early stage); 1.0 (stable stage); 3. Daily oil production designed in early stage: 2.0-2.5t; 4. Daily gas injection of a single well: liquid C 0 2 of 8m 3 /d in early stage; 6m 3 /d in stable stage; About 1.5xl04m3 of liquid carbon dioxide will be injected in above process in six years, then, it is transformed to water drive. Gas injection rate will be investigated and adjusted according to dynamic variations of production wells in implementation process of designed schemes.
17.3.4 Field Test Results and Analysis In 2003, a pilot area for C 0 2 flooding was pioneered in the Fuyu reservoir in the southern Songfangtun oil field. The oil-bearing area was 0.43 km 2 , average air permeability was 1.4xl0~3um2, and effective porosity was 14.5%. Currently the pilot area has one gas injection well and five production wells. The average sandstone thickness of the target layer (FI7) is 8.2 m and the effective thickness is 6.6 m. Fong 188-138 gas injection well started testing in March 2003, only penetrating FI7 layer. The sandstone thickness is 10.3 m; effective thickness is 6.0m; air permeability is 0.79~1.35xl0"3um2. Gas was injected without fracturing. Injection pressure currently is 12.5-13.0MPa, cumulative volume of injected liquid C 0 2 is 16500m3 (0.33PV) and the cumulative injection production ratio is 2.5 (See Figure 5). 173.4.1
Low Gas Injection Pressure and Large Gas Inspiration Capacity
From July to November in 2004, the average bottom-hole pressure was 30.2 MPa at the average daily liquid C 0 2 injection rate of 68m3 in Fong 188-138 gas-injection-well. From August to December 2005
PILOT TEST RESEARCH ON C 0 2 DRIVE
341
Figure 5. The variation of recovery percent and injection pore volume.
the bottom-hole pressure was 30 MPa and the average apparent gas inspiration index was 0.57m 3 / (d.MPa.m) at the average daily liquid C 0 2 injection rate of 50 m3. In Zhou 2 pilot area, the geologic characteristics are similar to Fong 188-138, it has two water injection wells with well spacing 212 m. The wells started fracture injection in December 1999. At the beginning the average oil pressure per well was 13.3 MPa, average water injection rate per day was 16 m3, and the apparent water-intake index per effective thickness calculated according to bottom hole pressure was 0.05 m 3 / (d.MPa.m). After two years the apparent water-intake index per effective thickness was 0.079m 3 / (d.MPa.m). Compared with the two water injection wells in Zhou 2 pilot area, the apparent gas inspiration index per effective thickness of gas injection wells without fracturing was 7.2 times more than that of water injection wells fractured. It shows that gas injection pressure is lower and the gas inspiration capacity is larger in Fuyu layer. 17.3.4.2
Production Rate and Reservoir Pressure Increase after Gas Injection
At the beginning, of the five oil wells in the pilot area, average production rate per day was 2.8t, intensity of oil recovery was 0.28t/d.m. Currently average production rate per day was 1.5t, and the intensity of oil recovery was 0.15t/d.m. Cumulative oil recovery was 7751t, recovery percent to OOIP was 3.37%, oil recovery rate was 0.92%, and total water-cut was 5.2%.
342
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
The production change of the well group shows that, from the commissioning date to July 2004 production followed the elastic recovery law. Production per day of well group increased steadily after gas injection test quickening in July 2004; decreased slightly after gas breakthrough of Fong 190-136, 190-140 in March 2005; and stayed above 7t because of response of Fong 188-137,190-138, and increase of production of Fong 190-136, 190-140. It was found that reservoir pressure increased from 8.6 MPa to 12.2 MPa after response through monitoring Fong 187-138 well. By analysis of production change since commissioning date in Zhou 2 pilot area, gas injection took effect in Fong 48 well group in August 2004 and cumulative incremental oil production was 1524 tons. Currently daily incremental oil production of the test well group was about 4 tons. 173.4.3
Reservoir Heterogeneity Is the Key to Control Gas Breakthrough
Breakthrough of wells Fong 190-136 and 190-140 occurred in March 2005. The present quantity of C 0 2 contained in casing pipe were 90.3% and 91.8%. The earlier breakthrough of the two wells was mainly due to the reservoir heterogeneity. From the horizontal distribution, the porosity and permeability of layer FI7 increased gradually from north to south. The permeability of well Fong 190-140 was the highest (about 2.6xl0-3um 2 ), and that of the other five wells were about 1.6-1.8x10 3 um 2 . In view of the vertical rhythmic profile of layer FI7, thickness with the relatively high permeability of well Fong 190-136 and 190-140 were apparently greater than other two wells. Due to reservoir heterogeneity, C 0 2 breakthrough of wells Fong 190-136,190-140 occurred earlier. Production well after gas breakthrough has following characteristics: 1. Production rate increased steadily. Seen from the curve of Fong 190-136,190-140 well's production change, oil production per day increased steadily from November 2004. The production slightly decreased early after gas breakthrough in March 2005, but increased steadily afterwards.
PILOT TEST RESEARCH ON C 0 2 DRIVE
343
2. C 0 2 content in the gas production increased gradually as well with gas-oil ratio and casing head pressure. Current gas-oil ratio of Fong 190-136 andl90-140 calculated according to Molar composition of gas production are 186 and 218m 3 /m 3 respectively; daily liquid C 0 2 production is 0.48~0.75m3. Besides, casing head pressure after gas breakthrough in production wells is gradually increased. At present the casing head pressures of the two gas breakthrough wells are between 3.7 and 4.6MPa, and that of wells without breakthrough of gas are below 0.5MPa. 3. Reservoir pressure is relatively high. Integral well test was carried out in the injection well group from May 2005 to June 2005. The reservoir pressure of Fong 190-136 and 190-140 well were 13.4 and 14.8MPa respectively, which were significantly higher than the other 3 wells (between 3.6 and 10.6MPa). 173.4.4
C02 Throughput as the Supplementary Means Reservoir's Effective Deployment
ofFuyu
Test well Fong 188-137 was put into production with 80m well spacing in August 2004. The complete well was only perforated in layer FI17. Sandstone thickness is 8.4m. Effective thickness is 5.7m. Besides, the well was put into production without fracturing. Early daily oil production was only 0.02t on in the test, and the daily oil production was between 0.2 and 0.3t from January to May 2005. C 0 2 throughput test was carried out in well Fong 188-137, and overall of liquid C 0 2 injected was 120m3. Early after throughput daily oil production was 2.3t, and oil recovery rate was 0.4t/d.m. Afterwards daily oil production was between 0.6t and 1.3t. According to the well's production change, daily oil production gradually increased from late August, which shows that C 0 2 throughput plays an early role (See Figure 6). Besides, Fong 190-138 well, which had a low oil recovery rate since it had been put into production, carried out C 0 2 throughput. Early incremental production was relatively high. However, due to the influence of the pump operating duty and project, the validity only lasted 50 days, and cumulative incremental oil production was 61ton (See Figure 7).
344
C0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 6. Fong 188-137 well daily oil production curve and water cut curve.
Figure 7. Fong 190-138 well daily oil production curve and water cut curve.
17.3.4.5
Numerical Result Shows that Carrying Out Water Flooding after Injecting Certain Amount of C02 Slug Is Better
In order to simulate the technical measures improving effect of gas injection, nine numerical schemes of four types were designed (See Table 5).
PILOT TEST RESEARCH ON C 0 2 DRIVE
345
Table 5. Numerical simulation program. No.
Description
Basic
1
Close gas injection well and keep production well producing
Gas injection
2
Maintain injecting Inject liquid C 0 2 at rate of gas in injection well 4, 9,14,18,27m3/d and producing oil in production well
3
Impulse gas injection
Carry out impulse gas injection (close wells with breakthrough of gas. After completing gas injection, open all the wells) and keep continuous gas injection after completion of impulse gas injection. Simulate the effect of impulse gas injection with different cycles
4
Inject a water slug first and then carry out gas flooding
Study the effect of various water slug sizes and velocities of follow-up gas injection on displacement efficiency
5
Carry out water flooding directly
Three water injection velocities: 10,15,20m3/d
6
Continue injecting certain amount of gas and then carry out water flooding
Continue injecting 4000, 6000, 8000,10000,15000, 25000, 30000m3 liquid C02 and then carry out water flooding. Study the effect of gas injection with various injection velocities
7
Continue impulse gas injection and then carry out water flooding
Carry out impulse gas injection at first and then change to water flooding after gas injection is completed
Category
Continue gas injection and change to water flooding afterwards
Details
(Continued)
346
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Table 5. Numerical simulation program. (Continued) Category Gas-water alternating flooding
No.
Description
Details
8
Inject gas and water alternatively and then change to water flooding
After injecting water and gas alternatively, carry out water flooding. Simulate the effect of gas injection adopting different proportion of gas alternating water
9
Carry out tapered gas-water alternating injection and then change to water flooding
Change gas-water ratio. Increase water and decrease gas, or contrarily
In summary, due to low permeability and high underground crude oil viscosity (6mPa.s), if basic scheme is adopted, productivity and the ratio of total oil produced to OOIP would be few. Concerning the three gas injection schemes; gas oil ratio increased so fast that after 6 to 8 years it would be greater than 1000 m 3 /m 3 and the well had to be closed. The ratio of total oil produced to OOIP was low. In regard to the water alternating gas displacement, persistent increase in the injecting pressure made the scheme hard to carry out. Consequently, the preferred scheme is to carry out water flooding after injecting certain amount of C 0 2 slug. However, numerical simulation did not take non-Darcy flow into account, and its conclusion needs to be further studied (See Table 6).
17.4
Conclusion
1. Both laboratory research and field test results proved that gas injection could reduce interfacial tension and enhance oil recovery, having unique advantages of developing very low-permeability oil reservoirs similar to Fuyu oil layer. 2. The pressure of miscible phase was 47 MPa in the laboratory research on gas injection in Fuyu layer. However, field tests could only adopt immiscible
Gas-water alternating flooding
Continue gas injection and change to water flooding afterwards
Gas injection
Basic
Category Prediction 10 Years Later
42000
37000
9470
15000
14500
29800
25500
4
5
6
7
8
9
39000
2
3
9470
19765
26440
29545
27982
29372
0
0
0
0
23495
30527
23855
24577
20653
21963
18700
18741
14861
8634
15665
8994
9716
5792
7982
5183
5707
0
(Continued)
9.08
11.79
9.22
9.49
7.98
8.48
7.23
7.24
5.74
Cumulative Liquid Cumulative Water Cumulative Oil Cumulative Oil Ratio of total C0 2 (m 3 ) Injection (m3) Production (t) Production (t) Oil Produced to OOIP (%)
1
No.
Table 6. Results of numerical simulation.
PILOT TEST RESEARCH ON C 0 2 DRIV
Gas-water alternating flooding
Continue gas injection and change to water flooding afterwards
Gas injection
Basic
Category
0
18107
53139 53914 58047 58800 36551
9470
15470
14500
29800
25500
5
6
7
8
9
28262
37516
31422
31924
27458
10155
19409
13315
13817
9352
10.92
14.50
12.14
12.33
10.60
Continue injecting gas for eight years and injecting 5700 m 3 water and then close oil wells.
7.0
4
0
Ratio of total Oil Produced to OOIP (%)
Continue injecting gas for seven years and then close oil wells.
Continue injecting gas for six years and then close oil wells.
9470
Cumulative Liquid Cumulative Water Cumulative Oil Cumulative C0 2 (m 3 ) Injection (m3) Production (t) Incremental Oil production (t)
Prediction 20 Years Later
3
2
1
No.
Table 6. Results of numerical simulation. (Continued)
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
PILOT TEST RESEARCH ON C 0 2 DRIVE
349
flooding. In long-core test, threshold pressure of gas injection was lower than that of water injection by 11 MPa/m, and breakthrough recovery of gas injection was higher than that of water injection by 4-6 percentages. Consequently it is feasible to test C 0 2 flooding in Fuyu layer. 3. Compared with hydrocarbon gas injection, oil recovery of C 0 2 flooding increased by 15 percentage points, and therefore achieved better results. 4. C 0 2 flooding test in Fuyang reservoir shows that injection pressure is lower and gas inspiration capacity is larger. The advantage proves that C 0 2 flooding could build u p effective deployment system in very low-permeability Fuyu reservoir without grown fractures. 5. Gas injection could form breakthrough hard to control and adjust, which could cause imbalance of effect in the horizontal after breakthrough of gas in some wells. Balance of gas drive in the horizontal is the key to improve sweep efficiency.
17.5 Acknowledgement This research was supported by the National Natural Science Foundation of China (10772023) and the National Key foundation of China (50934003).
References 1. Bentsen R G. "Effect of Momentum Transfer Between Fluid Phases on Effective Mobility." / Pet Sei Eng, 1998, 21 (1-2), 27 2. Morrow N R. Interfacial Phenomena in Petroleum Recovery. Monticello: USA, Mercel Dekker Inc, 1991 3. Zhu Weiyao. Liu Xuewei. Luo Kai. "Dynamic Model of Gas-Liquid-Solid Porous Flow with Phase Change of Condensate Reservoirs." Natural Gas Geoscience. 2005,16 (3): 363 4. Zhu Weiyao. "Theoretical Studies on the Gas-Liquid Two-phase Flow." Petroleum Expoloration and Development, 1988,15 (3): 63 5. Zhu Weiyao. "Theoretical Studies on the Gas-Liquid Two-phase Flow (Including a Phase Change)Through Porous Media." Acta Petrolei Sínica, 1990, 9(1): 15
350
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
6. Yu Mingzhou, Lin Jianzhong. The Dynamics of Nanoparitcle-Laden Multiphase Flow and Its Applications. 7. T.P Fishlock,C.J Probert. "Waterflooding of gas-condensate reservoirs." SPERE, 1996,11(3) 8. Prieditis J,Brugman R J."Effects of Recent relative Permeability data on C 0 2 flood modeling (A)." In: the 68 Annual Technical Conferences and Exhibition of SPE (C). SPE26650, Huston, Texas, 1993, 467-481.
18 Operation Control of C02-Driving in Field Site. Site Test in Wellblock Shu 101, Yushulin Oil Field, Daqing Xinde Wan, Tao Sun, Yingzhi Zhang, Tie j un Yang, and Changhe Mu CNPC Daqing Branch, Daqing, People's Republic of China
Abstract Based on C0 2 -driving test of extra-low permeable Fuyang oil layer in Well block ShulOl, Yushulin Oil Field, Daqing, the relationship of pore volume and injection mode, intensity of gas injection and injection rate was found in stratum that has an air permeability of around 1 millidarcy. Methods of adjusting injection profiles and production profiles were initially formed, according to the injection situation and dynamical property. Liquid state C 0 2 can be injected into oil layer as required in order to complement producing energy and oil production. Injecting gas in advance based on the stratum pressure and then bringing in oil wells can guarantee that oil wells take affect earlier and achieve economic yield without taking other measures. The strong ability of absorbing gas in stratum keeps the reservoir pressure high, for a longer time, and creates conditions for miscible displacement. Taking the different flow pressure control and systematic management can control one-way gas onrush and postpone gas channeling in line with oil yield, production intension and beneficial situation. For oil wells that are not responding, we can improve the benefit rate by C 0 2 throughput lead to well connected oil wells. Yushulin Oil Field is regarded as the typical large-scale oil deposit with extra low permeability, low fluidity, and low yield; with extra low permeability Fuyang oil layer as the interest bed under the main development, with the average air permeability of 2.71 xl0~3 urn2 and porosity of 10.8%. In addition, it has a bad water drive development effect, which is featured by low daily oil yield per well (0.7 t / d ) , low oil production speed (0.56% currently), low recovery degree, and low geologic reserve recovery degree (8.5%); thus, it is difficult to adopt the water drive for the oil layer with the Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (351-360) © Scrivener Publishing LLC
351
352
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
permeability lower than 1.5xl0"3um2. Accordingly, by the end of 2007, the site test of C02-driving was done in the Well block Shu 101 of Yushulin Oil Field, which was intended to explore the effective development approach of the reservoir bed with the permeability of 1.0xl0"3um2 and difficult exploration.
18.1 Test Area Description 18.1.1 Characteristics of the Reservoir Bed in the Test Area The test area is with the oil-bearing area of 2.36 km 2 and geologic reserve of 217.8xl0 4 t, which is mainly used to explore Fuyang oil layer. Besides, Fuyu oil layer is with the average porosity of 10.0% and air permeability of 1.16x10" 3 um 2 ; Yangdachengzi oil layer is with the porosity of 10.8% and air permeability of 0.96xl0~3um2. Additionally, the oil deposit is with the buried depth of 1806-2283 m; and the average original saturated pressure is 4.94 MPa, average initial gas-oil ratio is 22.8 m 3 /t, crude oil viscosity of the stratum is 3.6 MPa-s, original stratum pressure is 22.05 MPa, and the stratum temperature is 108°C. Through the slime-tube test, we find out that the min. miscible-phase pressure is 32.2 MPa, and the test area is the C0 2 - immiscible driving.
18.1.2 Test Scheme Design The test area is applied with the well pattern of 300x250 m rectangle five-spot area, which is featured by 23 wells in 7 rows, 7 injection wells and 16 exploratory wells, well array direction of NE77.5 0 and consistent with the max main stress direction. Firstly, three main oil layers including YI 6, YII 41, and YII 42 will be perforated, with the reserve of 118.7xl0 4 t, accounting for 54.5% of the total reserve; in the later stage, the upper part of Fuyu oil layer will be perforated, which will cost the total reserve of 148.5xl0 4 t. The scheme predicts that, the recovery ratio will be 20.1%. Based on the advance gas injection of six months and normal injection of three months and rest period of one month, it is designed that the well head injection pressure < 25.5 MPa and oil well production flow pressure a 5 MPa.
OPERATION CONTROL OF C 0 2 - D R I V I N G IN FIELD SITE
18.2
353
Test Effect and Cognition
Based on the advance gas injection principle, injection wells were successfully put into operation in December 2007, and all oil wells were put into operation till April 2009. By the end of June 2010, the cumulatively-injected liquid C 0 2 w a s 5.86xl0 4 t, cumulative oil production was 2.12xl0 4 t, recovery degree was 1.79%, oil production speed was 1.18%, cumulatively-injected HCPV amount was 0.063, annual injection-production ratio was 2.08, cumulative injection-production ratio was 3.38, annual oil replacement ratio was 0.65 t / m 3 , and cumulative oil replacement ratio was 0.39 t/m 3 .
18.2.1 Test Results The test shows that the liquid C 0 2 can be injected into the oil layer as specified to timely supply the oil layer energy and keep the stable production of the oil well. With the same gas injection amount, the gas-injection wells in the Well block Shu 101 are with the initial gas injection pressure of 18.5 MPa, and the current gas injection pressure is 17.8 MPa. With the reduction of the injection allocation amount and extension of the shut-in period of partial wells, the gas injection pressure is stable with a slight decline. In addition, the water injection pressure of Yushulin Oil Field is with typical increase, and the water injection pressure and daily water injection amount of the adjacent Well block Shu 8 increases by 4.4 MPa and reduces by 50% in the same period (see Figure 1). The first 2 wells are with the initial air suction index of 115.2 t / d . MPa and air suction pressure of 17.4 MPa. Currently, seven gasinjection wells are with the air suction index of about 42.0t/d.MPa and air suction pressure of about 17.2 MPa (see table 1). Viewing from the yield variation, it is always kept stable. On average, the daily oil yield per well keeps at 2.7 t; however, the index of the water drive yield is in a diminishing law, which is of larger reduction amplitude. By the end of the next year, the daily oil yield per well is only 33.3% of that of the initial period and the yield reduces by 66.7%, which is largely different from the law of the gas drive yield (see Figure 2).
354
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 1. Variation of injection pressure of well block Shu 101 and well block Shu 8.
Table 1. Index curve test result of well block ShulOl. Test Date
Air Suction Pressure (MPa)
Air Suction Index (t/d. MPa)
December 2007
17.4
115.2
April 2008
18.0
79.0
November 2008
17.4
42.3
March 2009
17.5
41.5
March 2010
17.2
42.4
18.2.2 The Stratum Pressure Status Based on the stratum pressure status, if the advance gas injection is properly done, the oil well will become effective earlier and have higher natural productivity. The gas-injection well is injected with the liquid C 0 2 of 2531 tons based on the advance gas injection of 6 months, which is with the averagely-injected HCPV times of 0.021 per well. In case the oil
Figure 2. Comparison between production status of well block Shu 101 and well block Shu 8.
I
H M W Oí
^ *-(
o
I-
w1
l—i
»Tí
2
O
FeS + H 2
(1)
The presence of H 2 has been confirmed by the analysis of synthetic inclusions having trapped experimental fluid in gold capsule.
23.5
Reactive Transport Modelling
The purpose of the reactive transport modelling was to reproduce the alteration front observed after aging in the aqueous liquid phase; no attempts were made so far for the diphasic and / o r dry supercritical phase experiments.
WELL CEMENT AGING IN VARIOUS H 2 S-C0 2 FLUIDS
431
Figure 5. Binocular (left), optical microscope on polished sections (centre) and SEM (right) observations of sulphidation crust (pyrrhotite) developed on steel micro-cylinder.
Figure 6. ID model showing the four different simplified hydro-geochemical zones.
The numerical simulation of the cement alteration needs a reactive transport code as both chemistry and transport processes (diffusion) are involved in the cement alteration. The reactive transport code HYTEC (13), based on the geochemical code CHESS, was used to simulate the reactions between the cement and the acidic brine. The CTDP (Common Thermodynamic Data Project) thermodynamic database (17) was selected for the study and enriched with additional data about H2S and C 0 2 solubility from the literature (12, 15, 18, 19). A ID model of a cross section of the cement bar was constructed for the calculations (Figure 6). It contains four initial hydro-geochemical zones (film of brine in contact with the gas, brine, crust and cement) defined by their initial diffusion properties, porosities and chemical compositions. The results of the simulations are shown in the Figure 7. The model reproduced the successive layers observed experimentally:
432
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
• a calcite crust formation at the surface of the cement bar, • a decalcified zone marked by the presence of silica and gyrolite (CSH with Ca/Si=0.66) / • The unaltered cement mimicked by a CSH association of tobermorite 11Â (Ca/Si=0.83) and foshagite (Ca/ Si=1.33). The formation of the calcite crust efficiently reduces the diffusion at the cement-solution interface. The flux of calcium and carbonates through the interface decreases dramatically so that the progression of the front stops after 5 to 10 days. This numerical result confirmed that the diffusion blockage (caused by the calcite deposit formation) is the process responsible for the front progression stop. However the thickness of the decalcified zone predicted by the model is two to three times larger than the thickness experimentally observed. It should be improved by a better fit of the model parameters.
23.6
Conclusion
Figure 7. Numerical simulation results obtained with HYTEC at 15 days duration.
WELL CEMENT AGING IN VARIOUS
H 2 S-C0 2 FLUIDS
433
We designed a specific experimental device to study well cement aging with H 2 S-C0 2 in geological-relevant conditions (brine, high pressure and temperature). The aim of this study was the evaluation of well cement durability in acid-gas storage hydrocarbon operations and by extension in C 0 2 geological storage in DBR (Deeply Buried Reservoirs) conditions. SEM characterization and water porosimetry figured out the carbonation of cement beneficial decrease of porosity (and by extension improvement of isolation properties). While immersed in aqueous liquid with dissolved H 2 S-C0 2 , a calcite deposition occurred at cement surface acting as a passivation layer that stops further carbonation. The passivation character of the deposit was evidenced by reactive transport modelling. We suppose those beneficial carbonate deposition to be related to low W / R (Water/ Rock) ratio (~1) occurring in the reactors. In presence of supercritical fluid, the initially dried cement is completely carbonated with a thin crust of aragonite and calcite. When pores are initially brine saturated, alteration fronts are detected showing digitations. These results show how the fluid state is important to predict the behaviour of solid phases. C 0 2 is the main actor of solid transformation in cement, whereas H2S is responsible for steel sulphidation in the extreme P-T conditions applied in this work. Future experiments with refreshing of surrounding fluid e.g. plug-flow, or large W / R ratio would be necessary to assess the long-term persistence of carbonates deposits within of at surface of cement.
Acknowledgments This study has been supported by TOTAL in the form of a Ph.D. thesis of Nancy Université, France, as part of a R&D project "Residual Gas Management" dedicated to C 0 2 capture and storage as well as acid gas injection. The authors thank TOTAL for the authorization of publishing these results. The authors thank also Alain Köhler from Nancy Université for the SEM analyses. Any opinions, findings, conclusions, or recommendations expressed herein are those of the authors and do not necessarily reflect the views of the sponsors.
434
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
References 1. Chakma, A. "Acid gas re-injection - a practical way to eliminate C02 emissions from gas processing plants." Energy Convers. and Manage. 1997, 38 (Supplt), S 205-S 209. 2. Connock, L. "Acid gas re-injection reduces sulphur burden." Sulphur 2001, 272,35-41. 3. Bachu, S., Watson, T.L., "Review of Failures for Wells used for C 0 2 and Acid Gas Injection in Alberta," Canada Energy Procedía 2009,1,3531-3537. 4. Watson, T.L. and Bachu, S. 2009. "Evaluation of the Potential for Gas and C 0 2 Leakage Along Wellbores." SPE Drill & Compl 24 (1): 115-126. SPE-106817-PA. 5. Carey, J. W; Wigand, M.; Chipera, S. J.; WoldeGabriel, G.; Pawar, R.; Lichtner, P. C ; Wehner, S. C ; Raines, M. A.; Guthrie, G. D. "Analysis and performance of oil well cement with 30 years of C 0 2 exposure from the SACROC unit, West Texas, USA." International ]. of Greenhouse Gas Control 2007,1 (1), 75-85. 6. Savage D, Maul P R, Benbow S J and Stenhouse, M. (2003) "The assessment of the long-term fate of carbon dioxide in geological systems." In Coping with Climate Change, 25-27 March 2003. Geological Society of London Online Extended Abstracts. 7. Celia, M. A. and Bachu, S. (2002) "Geological sequestration of C0 2 : is leakage unavoidable and acceptable ?" GHGT-6 (6th Int. conf. on GreenHouse Gas control Technologies). 6 pages, http://www.princeton.edu/~cmi/research/ kyoto02/celia&bachu.kyoto%2002.pdf 8. Kutchko, B.G., Strazisar, B.R., Dzombak, D.A., Lowry, G.V., Thaulow, N. "Degradation of Well Cement by C 0 2 under Geologic Sequestration Conditions." Environ. Sei. Technol. 2007,41:12,4787-4792. 9. Kutchko, B., Strazisar, B., Dzombak, D., Lowry, G. "Rate of C0 2 Attack on Hydrated Class H Well Cement under Geologic Sequestration Conditions." Environ. Sei. and Technol. 2008, 42:16, 6237-6242. 10. Fabbri, A., Corvisier, J., Schubnel, A., Brunet, E, Goffé, B., Rimmele, G., Barlet-Gouédard, V. "Effect of carbonation on the hydro-mechanical properties of Portland cements," Cement and Concrete Research, Volume 39, Issue 12, December 2009, Pages 1156-1163. 11. Jacquemet N, Pironon J, Saint-Marc J (2008). « Mineralogical changes of a well cement in various H 2 S-C0 2 (-brine) fluids at high pressure and temperature." Environ. Sei. Technol. 42, 282-288. 12. Jacquemet, N., 2006, "Well materials durability in a context of carbon dioxide and hydrogen sulfide geological storage," Ph.D. Thesis, Université Henri Poincaré, Nancy, France. 13. van der Lee, J., De Windt, L., Lagneau, V. and Goblet, P, 2003, "Module-oriented of reactive transport with HYTEC." Computers and Geosciences, 29,265-275. 14. Jacquemet, N., Pironon, J. and Caroli, E., 2005, "A new experimental procedure for simulation of H 2 S+C0 2 geological storage-Application to well cement aging." Oil & Gas Science and Technology-Rev. IFP, Vol. 60, No. 1, pp. 193-206. 15. Pironon, J., Jacquemet, N., Lhomme, T., Teinturier, S. 2007 "Fluid inclusions as micro-samplers in batch experiments: a study of the system C-O-H-S-cement with application to the geological storage of industrial acid gas." Chemical Geology 237,264-273.
WELL CEMENT AGING IN VARIOUS H 2 S-C0 2 FLUIDS
435
16. Fernandez Bertos, M., Simons, S. J. R., Hills, C. D., Carey, P. J., 2004, "A review of accelerated carbonation technology in the treatment of cement-based materials and sequestration of CO,." Journal of hazardous materials, B112,193-205. 17. CTDP v. 1.0.0, available at http://ctdp.ensmp.fr/ 18. Duan, Z. et Sun, R. (2003) "An improved model calculating C 0 2 solubility in pure water and aqueous NaCl solutions from 273 to 533 k and from 0 to 2000 bar." Chemical geology, 193, 257-271. 19. Bakker, R.J. 2009 "Package FLUIDS. Part 3: correlations between equations of state, thermodynamics and fluid inclusions," Geofluids 9, 63-74.
This page intentionally left blank
24
Casing Selection and Correlation Technology for Ultra-Deep, Ultra- High Pressure, High H2S Gas Wells Yongxing Sun, Yuanhua Lin, Taihe Shi, Zhongsheng Wang, Dajiang Zhu, Liping Chen, Sujun Liu, and Dezhi Zeng CCDC Drilling & Production Technology Research Institute Guanghan People's Republic of China State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation (SWPU) Chengdu, People's Republic of China CNPC Key Laboratory for Tubular Goods Chengdu, People's Republic of China
Abstract With ultra-deep (more than 5000 m), ultra-high pressure (more than 100 MPa), ultra-high temperature (more than 150°C), high H2S (2-70%) gas wells increasing, the casing service environment tends to be critical, sour and complex in the Northeast Sichuan province. In these wells, when well temperature is below 90°C, failures due to sulfide stress cracking (SSC) of High Strength Sour Service grades for downhole applications have been reported in recent years. Furthermore, it is well known that the plastic creep formation made of rock salt, gypse, and clay shale will give rise to much higher external collapse pressure on casing, which needs much higher collapse strength OCTG grades (such as non-sulfur resistance casing 140ksi, 150ksi grades) to meet test and production criteria. All of the complex cases result in catastrophic consequences that require the sulfur resistance casing (such as C110) not only to meet the ultra-high pressure criteria (more than lOOMPa) but also to meet the high sulfur resistance criteria (H2S partial pressure is 0.3-2MPa in Sichuan). Though rules for the selection of proper materials to avoid the catastrophic consequences of SSC are gathered in the latest edition of the NACE MR0175/ISO151516 standard "Materials for use in H2S containing environments" and "Sumitomo Metal sulfur resistance casing selection Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (437-448) © Scrivener Publishing LLC
437
438
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
recommended practice", both selections' recommended practices cannot present suitable casing. So this paper discusses an economical and suitable method of string design combined with sulfur resistance packer completion test technology, and this method successfully deals with the current difficult problems of sulfur resistance casing selection and casing program design in current ultra-deep, ultra-high pressure, high H2S gas wells in Northeast Sichuan Gas Field China.
24.1
Introduction
Currently, more ultra-deep, ultra -high pressure, ultra-high temperature oil & gas wells are being drilled and completed in the presence of H2S and C0 2 , the need for high strength (>=110ksi) sour service-rated tubular is ever increasing. In these wells, when well temperature is below 90°C, downhole High Strength Sour Service grades failures accidents due to sulfide stress cracking (SSC) have been reported in recent years. East Sichuan gas fields contain 7.12% to 17.03% of H2S, and 3.29% to 10.41% of C0 2 . Chloride ion content is about 20429 mg/1 in the L well, as determined in the water analysis report. The corrosion problems of oil country tubular goods (OCTG) and equipments often take place due to H2S and C 0 2 [1, 2, 3, and 4]. Though rules for the selection of proper materials to avoid the catastrophic consequences of SSC are gathered in the latest edition of the NACE MR0175/ISO 151516standard15-61 "Materials for use in H2S containing environments" and "Sumitomo Metal sulfur resistance casing selection recommended practice", both selection recommended practices cannot present suitable casing. So this paper discusses an economical and suitable method of string design combined with sulfur resistance packer completion test technology, and this method successfully deals with current difficult problem of sulfur resistance casing selection and casing program design in current ultra-deep, ultra-high pressure, high H2S gas wells in Northeast Sichuan Gas Field China.
24.2
Material Selection Recommended Practice
The severity of the sour environment, determined in accordance with ISO 15156-1 (2009), with respect to the SSC of a carbon or low-alloy steel shall be assessed using Figure 1 [6]. In defining the
CASING SELECTION AND CORRELATION TECHNOLOGY
439
severity of the H 2 S-containing environment, the possibility of exposure to unbuffered, condensed aqueous phases of low pH during upset operating conditions or downtime, or to acids used for well stimulation and / o r the backflow of stimulation acid after reaction should be considered. When partial pressure is higher than 0.7 bar, Cr-Ni materials should be used. Carbon and low-alloy steels selection should be tested by casing and tubing serve conditions from regions 1, 2, and 3 (shown in Figure 1) [6]. If the recommended casing and tubing such as C110 has been applied downhole, whose serve environments can be referred as casing and tubing selection recommended practice. The discontinuities in the figure below 0.3 kPa (0.05 psi) and above 1 MPa (150 psi) partial pressure H 2 S reflect uncertainty with respect to the measurement of H2S partial pressure (low H2S) and the steel's performance outside these limits (for both low and high H2S). In this case, it must use fit for purpose (FFT)[7] method to evaluate the OCTG materials according actual well conditions, then, it can be determined whether or not to use the recommended OCTG materials.
Figure 1. Regions of environmental severity with respect to the SSC of carbon and low-alloy steels Key: X-H2S partial pressure, expressed in kilopascals; Y-in situ pH; 0-region 0; l-SSC region 1; 2-SSC region 2; 3-SSC region 3. Note 1 Guidance on the calculation of H2S partial pressure is given in ISOl 5156:2009, Annex C. Note 2 Guidance on the calculation of pH is given in ISO15156:2009, Annex D.
440
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
Figure 2 provides a simplified guideline to choose from the material applications based on C 0 2 content, H2S content and temperature by Sumitomo Metals [8]. Reliable and optimized material selection depends upon a large set of parameters, from fluid characteristics to well conditions. The chart provides a quick way to make a pre-selection of material type. It should be used as a guideline. For a more detailed assessment, please refer to each Material Data Sheet [8]. This material selection chart features Sumitomo Metals proprietary grades. Sumitomo Metals also does manufacture API 5CT/ISO 11960 grades. It has been recognized that there are still much inadequacies to design casing program by Figures 1 and 2 for oil and gas wells with H 2 S[6,8,and9]. 1. The low grade anti-sulfur casing such as J55, T95 can just be applied in shallow oil and gas well for their low strength. 2. The low grade casing strength (burst and collapse) can be increased by adding the wall thickness, which can be done within limited conditions, because of the increase in cost and change of casing program that has been wide applied in oil and gas fields. In addition, a significant increase in wall thickness will make the casing cannot meet the tensile strength safe criteria. So it cannot still meet high collapse resistance and internal pressure strength criteria in HTHP sour gas well through adding wall thickness. 3. Stainless steel can meet anti-sulfide criteria, but it cannot meet strength design criteria in ultra-HTHP sour gas well, as well as its high-cost. 4. Though lots of anti-sulfide casing such as 110SS, 125SS have been applied in oil and gas fields at abroad and home, but all of them must be evaluated through FFT method before they are applied in oil and gas fields, otherwise, whose strength cannot still meet strength design criteria in ultra-HTHP sour gas well in east Sichuan oil fields, and 140 ksi or 150 ksi grade OCTG may be just meet the high collapse criteria [10].
CASING SELECTION AND CORRELATION TECHNOLOGY
441
Figure 2. Material Selection According to Gas (C0 2 and H2S) Partial Pressure (Note: Cl" content should be less than 30,000ppm for SMCR and SM 13CR).
24.3 Casing Selection and Correlation Technology All above discussed cannot deal with casing strength design problem for ultra-HTHP, high H2S well in east Sichuan oil fields. Based on the further study, the contradiction between the anticorrosion and high collapse criteria has been perfectly solved in following technology: 1. When well temperature is above 90°C, sulfide-stress cracking (SSC) can be neglected, but the plastic creep formation made of rock salt, gypse, clay shale will
442
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
give rise to much higher external collapse pressure on casing, for this case, it can use higher collapse strength OCTG grades such as non-sulfur resistance casing 140ksi, 150ksi grades to meet test and production criteria. 2. During completion operations, especially in the stimulation operations, the well temperature will decrease to 60°C -90°C, which results in the high strength casing such as 140 or 150 ksi grade stress SSC. So the two methods below have been used to prevent the high strength OCTG such as 140 or 150 ksi casing from SSC, which can be described as follows. a. Completion strings with an anti-sulfide packer can separate the high strength casing (140 or 150 ksi) from natural gas containing H 2 S, which can decrease the possibility of SSC and meet the strength criteria. b. For test well, it is necessary to prevent the bottom temperature from decrease and make the high strength casing (140 or 150 ksi) immune to SSC. Because the high collapse strength and anticorrosion capability of OCTG cannot be simultaneously presented by Figures 1 and 2. So this paper presents an economical and suitable method of string design combined with sulfur resistance packer completion test technology. By anti-sulfide packer, the low temperature (below 90°C) natural gas containing H2S is isolated from the casing which prevent the high strength casing from SSC. In this case, the high strength casing such as 140 ksi etc. can be applied in the plastic creep formation containing H2S to meet strength design criteria.
24.3.1
Casing Selection and Match Technology Below 90°C
During completion operations, when well temperature is below 90°C, H2S corrosion cannot be ignored. The low-temperature wells containing H2S (below 90°C) should select the low grade casing. If the casing strength cannot meet the design criteria, the suitable casing can be obtained through reasonable increase in the wall thickness.
CASING SELECTION AND CORRELATION TECHNOLOGY
443
24.3.2 Casing Selection and Match Technology Above 90°C When well temperature is above 90°C, H2S corrosion can be ignored. In this case, the high strength casing such as 140 ksi etc. can be applied in high temperature well. 1. This approach is feasible when the well temperature is constant 2. During acidizing treatment, well temperature will be below 90°C, in this case, the SSC must be considered. In order to prevent the high strength casing from corrosion due to well temperature decreasing, the following match technology measures are used: a. By sulfur resistance packer completion test technology, the gas containing H2S is isolated from low temperature casing above sulfur resistance packer. b. Because of extending acid reaction time can make the acid temperature above 100°C in formation, and then the casing below sulfur resistance packer can be protected. c. After controlling the higher wellhead pressure, because of gas volume expansion, it is necessary to keep the gas from the information in constant temperature. When the high strength casing such as 140 ksi contacts the gas with H 2 S, the well temperature remains above 90°C, the casing below sulfur resistance packer can be effectively protected.
24.4
Field Applications
This new method has been successfully applied in most of gas wells containing H2S in M gas fields in China, and the typical well programs are shown in Figures 3 and 4. Figure 3 is the original casing program of L well in M gas fields, and the OD 177.8 mmxll.51 mm wall thickness VMllOSS casing is selected for its good anti-sulfide, but whose strength is not immune to L well with formation pressure 127.79 MPa (mid-depth 5942.9 m) and the wellhead pressure 107 MPa. The OD 177.8 mmxll.51 mm wall thickness VMllOSS casing collapse and anti-pressure safe factor is respectively 0.76 and
444
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
0.68 (shown in Table 1) in air. All the safety factors cannot meet the "Drilling technical operation" [11] (collapse safety factor is 1.125, intermediate casing and production's is not less than 0.80), as well as the SY/T5322-2000[12] (collapse safety factor is from 1.00 to 1.125, anti-pressure safety factor is from 1.05 to 1.15). It is worth noting how to avoid the catastrophic consequences of casing collapsing of M gas fields, because the casing strength cannot meet the strength criteria. So the original casing program of L well has been partly changed. On the basis of guaranteeing completion and cementing quality, the OD 177.8 mmxll.51 mm wall thickness VM110SS casing is replaced by OD 193.68 mmxl9.05 mm wall thickness TP110SS casing, and the collapse strength is 15% higher than that of VM110SS, which not only meet the strength criteria but meet anti-corrosion requirement, in the meantime, all the safety factors is higher than 1.0 The high strength casing such as VM140HC (collapse strength is 117.62 MPa) is applied in the plastic creep formation zone, and the
Figure 3. The original casing program L well.
CASING SELECTION AND CORRELATION TECHNOLOGY
445
Figure 4. The changed casing program L well.
changed casing program is shown in Figure 4. This method successfully deals with current difficult problem of sulfur resistance casing selection and casing program design in current ultra-deep, ultra-high pressure, high H2S gas wells in Northeast Sichuan Gas Field China.
24.4
Conclusions 1. The method of string design combined with sulfur resistance packer completion test technology can successfully deal with current difficult problem of sulfur resistance casing selection and casing strength design in current ultra-deep, ultra-high pressure, high H2S gas wells in Northeast Sichuan Gas Field China.
VM140HC
180
1500
177.8
177.8
139.7
Production Tieback
Production suspend
WM110SS
4568
244.5
Intermediate 2 hang
3300
1448
244.5
Intermediate 2 Tieback
KO-140V
TP-95TS
TP-95ss
P-110
1698
339.7
Intermediate 1
J-55
Grade
98
Setting Depth(m)
508
(mm)
OD
Surface
String
Table 1. The original casing program of L well.
12.7
12.65
11.51
11.99
11.99
598
689
2175
2308
1012
1818
155
12.7 13.06
Weight (kN)
Wall (mm)
159.5
117.62
74
45.5
35.09
19.89
5.3
Collapse Strength (MPa)
1.19
1.16
0.76
0.70
1.77
1.86
5.25
Safety Factor
Collapse
108.5
120.18
86
56.19
56.19
34
16
Burst Strength (MPa)
1.4
1.51
0.68
0.83
0.74
1.02
15.94
Safety Factor
Burst
3866
6334
4559
5846
5846
10160
7495
Tensile (kN)
6.46
9.19
2.09
2.53
5.77
5.58
48.35
Safety Factor
Tensile
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
CASING SELECTION AND CORRELATION TECHNOLOGY
447
2. For critical corrosion environments, when well bottom temperature is below 90°C, the lower grade anti-sulfide casing such as T95, J55 etc. can be applied. If the lower grade casing anti-pressure cannot meet strength criterion, adding the pipe wall thickness.
24.5 Acknowledgments The authors are grateful to the support from Program for New Century Excellent Talents in University (NCET-08-0907) and China National Natural Science foundation and Shanghai Baosteel Group Corporation (Grant No.: 51074135).
References 1. Li Luguang, Huang Liming, Gu Tan, Li Feng. "Corrosion Characters and Inhibition Methods for Sichuan Gas Fields." Chemical Engineering of Oil & Gas, 2007,36(l):46-54 (in Chinese with English Abstr.). 2. Jiang Fang. "In Lab. Evaluation Methods of Metal Materials for High Sour Gas Fields." Natural Gas Industry, 2004, 24(10): 105-107 (in Chinese with English Abstr.). 3. Zeng Dezhi, Huang Liming, Lin Yuanhua. "Material Evaluation and Selection of OCTG and Gathering Lines for High Sour Gas Fields." SPE 131943, 2010, 6 (in Chinese with English Abstr.). 4. Liu Zhide, Huang Liming, Yang Zhongxi, et al. "Material Corrosion Factors of Ground Gathering Line in High Sulfurous Environment." Natural Gas Industry, 2004, 24(12):122-123(in Chinese with English Abstr.). 5. NACE TM0177-2005. "Laboratory Testing of Metals for Resistance to Specific Forms of Environmental Cracking." NACE International, Houston, TX, 2005. 6. IS015156/NACE MR0175. "Petroleum and Natural Gas Industries-Materials for Use in H2S Containing Environments in Oil and Gas Production[S]." Switzerland: The International Organization for Standardization, 2009. 7. Fitness-For-Services. API 579-1 /ASME FFS-1, JUNE 5, 2007(SECOND EDITION). 8. Sumitomo Metal, http://www.sumitomo-tubulars.com/materials/index. htm. 9. "Petroleum and natural gas industries-Steel pipes for use as casing or tubing for wells." ISO/FIDS 11960 (E):2009. 10. Sun, Y.X., Lin, Y.H. 2010. "Study on Casing and Tubing Design for Sour Oil & Gas Field." Corrosion Science and protection technology, (in Chinese, with English Abtsr.). 11. Q/SYCQZ 001-2008 Drilling technical operation.2008. 12. SY/T5322-2000. Design method of casing string strength[S]. Beijing: China Machine Press, 2001.
This page intentionally left blank
25
Coupled Mathematical Model of Gas Migration in Cemented Annulus with Mud Column in Acid Gas Well Hongjun Zhu, Yuanhua Lin, Yongxing Sun, Dezhi Zeng, Zhi Zhang, and Taihe Shi State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation Southwest Petroleum University, Chengdu, People's Republic of China.
Abstract Sustained casing pressure (SCP) in an acid gas well brings serious threat to worker safety and environmental protection. In this paper, the mathematical model of gas flow through porous media in cement sheath was obtained based on the generalized Darcy percolation theory. And the establishment of buoyancy model for gas migration in mud columns was based on multi-phase fluid dynamics theory. On this basis, the coupled mathematical model of gas migration in cemented annulus with mud column has been improved. Then the value of SCP changing with time in an acid gas well in a field of China has been calculated by this model. Calculation results coincided well with the actual field data, which provide some reference for the following security evaluation and solution measures of SCP.
25.1
Introduction
Well cement problems such as small cracks or channels can result in gas migration and lead to sustained casing pressure (SCP) at casing heads. Several other reasons for casing pressure buildup are tubular corrosion or mechanical failures, packer failures, and connection leaks. In some cases, the pressure can reach dangerously Wu/Carroll/Du (ed.) Carbon Dioxide Sequestration and Related Technologies, (449-462) © Scrivener Publishing LLC
449
450
C 0 2 SEQUESTRATION AND RELATED TECHNOLOGIES
high values. Moreover, SCP in acid gas wells brings serious threat to worker safety and environmental protection. SCP is the universal existing problem in gas wells in China. After the 12 May 2008 Wenchuan earthquake, the possibility of forming SCP in gas wells in northeast of Sichuan is much greater than previous. While acid gas reservoirs widely locate in Sichuan-Chongqing region in China, which contain high content of acidic components such as carbon dioxide and hydrogen sulfide. The material of current intermediate casing and surface casing in gas well is usually the common carbon steel. Acid gas migrating into annulus of protection casing and production casing or annulus of protection casing and surface casing may cause corrosion on the strings and endanger the safety of wellbore. The gas escaping from high-sulfur gas well will specially give rise to significant personnel casualty and property damage. Therefore, we are in urgent need of an explicit reason for acid gas migration and a model to calculate sustained casing pressure. The work presented here focuses on the coupled mathematical model of gas migration in cemented annulus with mud column in acid gas well, which provide some reference for the following security evaluation and solution measures of SCP.
25.2 Coupled Mathematical Model As shown in Figure 1, two possible configurations of the cement column in the annulus are commonfl]: cemented to the surface or a mud column above the cement. And gas column or gas-liquid column may present in the cement if the cementing job is poor. In wells cemented to the surface, gas migration can be considered as a one dimensional flow through a medium have some conductivity [2]. After bleed-down, the buildup behavior is controlled by cement properties, such as permeability and porosity, and by gas formation pressure. While in wells with a mud column above the cement, gas migration occurs in two stages. Firstly, gas flow follows Darcy's Law in the cement column. Then gas rises as bubbles through stagnant non-Newtonian drilling fluids, in which the gas migration is affected by the characteristics of mud and the status of the top gas cap. We focus our attention on the coupled mathematical model of gas migration in cemented annulus with mud column, which includes gas migration in cement column. Considering acid gas migration in cement and stagnant mud, the coupled model was obtained by improving Xu's SCP model[3,4].
COUPLED MATHEMATICAL MODEL
451
Figure 1. Configurations of the cement column in the annulus (a) Annulus cemented to the surface (b) Annulus with a mud column above cement.
25.2.1
Gas Migration in Cement
Gas migration in cement column can be considered as a one-dimensional flow through a medium having some conductivity, which related the cement properties, interface pressure, interface flow rate, gas formation pressure and elapsed time. The following assumptions were made for establishing the mathematical model[5]. Firstly, the gas formation pressure is constant because permeability of gas zone is much higher than that of cement. Secondly, the pseudo gas pressure concept is used. Finally, gas is vented out from the well at a small constant rate at the end of bleed-down. Then with a constant flow rate qn during the n-th period, the pressure in cement can be obtained as:
^
+
dz2
i i ^ T 4 ^
p{dz)
if
p< 15 MPa
and
v
Çdt
zl5MPa
and
z0 ^
25.2.2 Gas Migration in Stagnant Mud Gas migration in stagnant mud can be modeled as dispersed twophase flow that can be described by a drift-flux model[6]. Basic assumptions are concentric annulus, equal phase pressure, uniform phase densities normal to the flow direction, constant temperature profile and thermodynamic equilibrium. Due to slow phase segregation after the bleed-down, it is assumed that the relative velocity term is negligible. Under the assumptions, the one-dimensional two-equation drift-flux model is summarized in the foliowing[3]: Continuity equation for dispersed phase (gas) is: d(apg)
^(ccpgvg)
dt
=Q
(3)
dz
And mixture momentum equation is:
-^+pmg+jj-pym=o
(4)
where: V
v
g=
ü
s+C0vm
(5)
m=(^+