Detection and Diagnosis of Stiction in Control Loops (eBook)

Mohieddine Jelali is Project Manager and Vice-Head of Department of Plant and System Technology at VDEh Betriebsforschungsinstitut (BFI), the Applied Research Institute of the German Steel Industry. He received his Dipl. Eng. and Dr. Eng. degrees in Mechanical Engineering from the University of Duisburg, Germany, in 1993 and 1997, respectively. Before joining BFI in 1999, he worked for three years with Mannesmann Demag Metallurgy as an R and D Engineer on the design and application of advanced process control systems in rolling mills. For more than a decade Doctor Jelali has worked in the areas of advanced (model-predictive) control and control performance monitoring as the initiator and coordinator of multinational research projects in the field of metal processing. He has filed patents related to this. M. Jelali has received the Medal of the Werner-von-Siemens-Stiftung for his excellent research in the field of control in the rolling. He has published about 40 papers on process control and related topics in different international journals and conference proceedings. He is the principal author of Hydraulic Servo-systems: Modelling, Identification and Control (978-1-85233-692-9) published by Springer in 2002. He is a member of VDEh, VDI, GMA, and DUG. Biao Huang is a Professor and researcher in system identification and control performance monitoring. He has received a number of awards for his contributions in these areas including Germany’s Alexander von Humboldt Research Fellowship award, Canadian Chemical Engineeriing Society’s Syncrude Canada Innovation Award, University of Alberta’s McCalla Professorship award, and Petro-Canada Young Innovator Award, and has been the recipient of best paper award for Journal of Process Control. Biao Huang has worked specifically in systems and control over the last 20 years. He is one of the leading experts in control loop performance monitoring, and has made contributions in this area including a book published by Springer. Biao Huang applied his expertise extensively in industrial practice particularly in oil sands industry. His research contributions in control performance monitoring have enjoyed widely applications in chemical, petrochemical, oil and gas, mineral processing, and pulp and paper industries throughout the world. He has served on a number of national and international engineering and science communities including as Chair of CSChE’s System and Control Division, Associate Editor of Control Engineering Practice, Associate Editor of CJChE. Since 1997 Biao Huang has published over 100 refereed papers in international journals and conference proceedings. He has been invited to speak in a number of institutions as well as workshops worldwide. He is the author of Performance Assessment of Control Loops (978-1-85233-639-4) and Dynamic Modeling, Predictive Control and Performance Monitoring (978-1-84800-232-6).

Preface 11
Copyright Acknowledgements 15
Contents 16
List of Contributors 24
Abbreviations and Acronyms 27
1 Introduction 29
1.1 Motivation 29
1.2 Typical Valve-controlled Loop 30
1.3 Stiction Phenomenon and Related Effects 32
1.4 Input–Output Relation of Valves Under Stiction 34
1.5 Limit Cycles due to Stiction 37
1.6 Typical Observations in Control Loops with Sticky Valves 40
1.7 Industrial Examples of Loops with Stiction 43
1.8 Summary and Conclusions 46
Part I: Stiction Modelling and Oscillation Detection 47
2 Stiction Modelling 48
2.1 Introduction 49
2.2 Physics-based Stiction Modelling 49
2.3 Data-driven Stiction Modelling 51
2.4 Comparison Between Choudhury’s and Kano’s Stiction Models 59
2.5 Summary and Conclusions 62
3 An Alternative Stiction-modelling Approach and Comparison of Different Stiction Models 63
3.1 Introduction 63
3.2 He’s Two-parameter Model 64
3.3 Three Data-driven Models 66
3.4 Further Investigation of Valve Stiction 71
3.5 He’s Three-parameter Model 75
3.6 Simulation Results 77
3.7 An Industrial Example 82
3.8 Summary and Conclusions 83
3.9 Appendix: Proof of the Equivalence Between He’s Two- parameter and Three- parameter Model 84
4 Detection of Oscillating Control Loops 86
4.1 Introduction 87
4.2 Root-causes for Oscillatory Control Loops 87
4.3 Characterisation of Oscillations 89
4.4 Techniques for Detection of Oscillations in Control Loops 92
4.5 Critical Evaluation of Oscillation-detection Methods 105
4.6 Comprehensive Oscillation Characterisation 118
4.7 Industrial Case Studies 119
4.8 Summary and Conclusions 125
Part II Advances in Stiction Detection and Quantification 126
5 Shape-based Stiction Detection 127
5.1 Introduction 127
5.2 Method Description 128
5.3 Key Issues 131
5.4 Simulation Results 132
5.5 Application to Industrial Loops 135
5.6 Summary and Conclusions 137
6 Stiction Detection Based on Cross-correlation and Signal Shape 138
6.1 Introduction 138
6.2 The Cross-correlation Function 140
6.3 Industrial Examples 143
6.4 Theoretical Explanation 149
6.5 Conclusions (Cross-correlation Method) 154
6.6 Stiction Detection for Integrating Processes 155
6.7 Detection in Integrating Loops – Basic Idea 155
6.8 Examples 165
6.9 Self-regulating Processes 166
6.10 Summary and Conclusions 170
7 Curve Fitting for Detecting Valve Stiction 171
7.1 Introduction 171
7.2 Method Description 173
7.3 Key Issues 176
7.4 Simulation Results 176
7.5 Application to Industrial Loops 181
7.6 Summary and Conclusions 183
8 A Relay-based Technique for Detection of Stiction 186
8.1 Introduction 187
8.2 Trends of Different Variables 189
8.3 Method Description 191
8.4 Simulation Results 196
8.5 Application to Plant Data 199
8.6 Summary and Conclusions 201
9 Shape-based Stiction Detection Using Area Calculations 203
9.1 Introduction 203
9.2 Method Description 205
9.3 Key Issues 217
9.4 Simulation Results 218
9.5 Application to Industrial Loops 220
9.6 Summary and Conclusions 223
10 Estimation of Valve Stiction Using Separable Least-squares and Global Search Algorithms 225
10.1 Introduction 225
10.2 Basic Approach 227
10.3 Identification Approach 230
10.4 Key Issues 235
10.5 Simulation Studies 239
10.6 Industrial Case Studies 241
10.7 Summary and Conclusions 247
11 Stiction Estimation Using Constrained Optimisation and Contour Map 249
11.1 Introduction 249
11.2 Stiction Model of Control Valve 251
11.3 Existing Stiction-detection Methods 253
11.4 Closed-loop Stiction Detection and Quantification 255
11.5 Stiction Detection: Identifiability Analysis 261
11.6 Simulations 265
11.7 Industrial Applications 271
11.8 Graphical User Interface 280
11.9 Summary and Conclusions 285
12 Oscillation Root-cause Detection and Quantification Under Multiple Faults 287
12.1 Introduction 287
12.2 Preliminaries and Brief Review of Model-based Oscillation Diagnosis 288
12.3 Overview of the Root-cause Detection and Quantification Methodology 290
12.4 Process-model Identification Under Non-stationary Disturbances 291
12.5 Root-cause Detection and Quantification 298
12.6 Illustrative Example: Oscillation Diagnosis Under Various Faulty Situations 301
12.7 Industrial Case Studies 310
12.8 Summary and Conclusions 313
13 Comparative Study of Valve-stiction-detection Methods 314
13.1 Introduction 314
13.2 Selected Methods 315
13.3 Industrial Control Loops Involved in the Study 317
13.4 Application Results and Discussion 321
13.5 Suggestions 340
13.6 Summary and Conclusions 342
13.7 Appendix: Tables of Results of the Comparative Study 343
14 Conclusions and Future Research Challenges 378
14.1 Summary of the Book 378
14.2 Future Research Challenges 381
Appendix A: Evaluated Industrial Control Loops 386
Appendix B: Review of Some Non-linearity and Stiction-detection Techniques 389
B.1 Bicoherence Method 389
B.2 Surrogates Analysis 392
References 395
Contributor Biographies 401
Index 407