Condition Monitoring and Assessment of Power Transformers Using Computational Intelligence (eBook)

Dr. W.H. Tang received his BEng and MSc(Eng) degrees in Electrical Engineering from Huazhong University of Science and Technology, China, in 1996 and 2000, respectively. He obtained a PhD from The University of Liverpool, Liverpool, UK, in 2004. From 2004 to 2006 he worked as a postdoctoral research associate and a university teacher in The University of Liverpool. Since 2006, he has held a lectureship in power engineering in the Department of Electrical Engineering and Electronics in The University of Liverpool. He has published 21 refereed journal papers and presented 24 international conference papers since 2000. His research interests include transformer modelling, substation condition monitoring, computational intelligence and multiple criteria decision analysis. He has also worked on multi-agent systems, renewable energy and power systems. His research has been funded by the Engineering Physical Science Research Council, UK, and industrial companies. Professor Q.H. Wu is the Chair of Electrical Engineering at The University of Liverpool, UK. He obtained an MSc(Eng) in Electrical Engineering from Huazhong University of Science and Technology (HUST), China, in 1981 and a PhD in Electrical Engineering from The Queen's University of Belfast (QUB) in 1987. Before joining The University of Liverpool in 1995, Professor Wu worked as senior research fellow, lecturer and senior lecturer at QUB and Loughborough University, UK, respectively. He has published 3 monographs, 150 journal papers, 20 book chapters and 180 refereed conference papers. In 1994 he was awarded the Donald Julius Groen Prize for the best paper published in the Journal of Systems and Control Engineering, Institution of Mechanical Engineers. He is a Chartered Engineer, Fellow of IET and Senior Member of IEEE. Professor Wu’s research interests include systems modelling and control, mathematical morphology, computational intelligence, multi-agent systems and their applications for power system operation and control.

Preface 6
Contents 8
Acronym 13
Introduction 16
1.1 Background 16
1.2 Main Aspects of Transformer Condition Monitoring and Assessment 18
1.3 Drawbacks of Conventional Techniques 21
1.4 Modelling Transformer and Processing Uncertainty Using CI 23
1.5 Contents of this Book 24
1.6 Summary 26
References 27
Evolutionary Computation 29
2.1 The Evolutionary Algorithms of Computational Intelligence 29
2.2 Genetic Algorithm 32
2.3 Genetic Programming 38
2.4 Particle Swarm Optimisation 43
2.5 Summary 48
References 48
Methodologies Dealing with Uncertainty 51
3.1 The Logical Approach of Computational Intelligence 51
3.2 Evidential Reasoning 52
3.3 Fuzzy Logic 62
3.4 Bayesian Networks 64
3.5 Summary 67
References 67
Thermoelectric Analogy Thermal Models of Power Transformers 69
4.1 Introduction 69
4.2 Conventional Thermal Models in IEC and IEEE Regulations 70
4.3 The Thermoelectric Analogy Theory 74
4.4 A Comprehensive Thermoelectric Analogy Thermal Model 75
4.5 Parameter Estimation of a Thermoelectric Analogy Model 80
4.6 Identification of Thermal Model Parameters 82
4.7 A Simplified Thermoelectric Analogy Thermal Model 82
4.8 Summary 84
References 85
Thermal Model Parameter Identification and Verification Using Genetic Algorithm 86
5.1 Introduction 86
5.2 Unit Conversion for Heat Equivalent Circuit Parameters 87
5.3 Fitness Function for Genetic Algorithm Optimisation 88
5.4 Parameter Identification and Verification for the Comprehensive Thermal Model 89
5.5 Parameter Identification and Verification for the Simplified Thermal Model 98
5.6 Summary 106
References 107
Transformer Condition Assessment Using Dissolved Gas Analysis 108
6.1 Introduction 108
6.2 Fundamental of Dissolved Gas Analysis 109
6.3 Combined Criteria for Dissolved Gas Analysis 114
6.4 Intelligent Diagnostic Methods for Dissolve Gas Analysis 115
6.5 Summary 116
References 117
Fault Classification for Dissolved Gas Analysis Using Genetic Programming 118
7.1 Introduction 118
7.2 Bootstrap 120
7.3 The Cybernetic Techniques of Computational Intelligence 121
7.4 Results and Discussion 123
7.5 Summary 135
References 136
Dealing with Uncertainty for Dissolved Gas Analysis 138
8.1 Introduction 138
8.2 Dissolved Gas Analysis Using Evidential Reasoning 139
8.3 A Hybrid Diagnostic Approach Combining Fuzzy Logic and Evidential Reasoning 151
8.4 Probabilistic Inference Using Bayesian Networks 165
8.5 Summary 174
References 175
Winding Frequency Response Analysis for Power Transformers 176
9.1 Introduction 176
9.2 Transformer Transfer Function 178
9.3 Frequency Response Analysis Methods 179
9.4 Winding Models Used for Frequency Response Analysis 181
9.5 Transformer Winding Deformation Diagnosis 181
9.6 Summary 187
References 187
Winding Parameter Identification Using an Improved Particle Swarm Optimiser 189
10.1 Introduction 189
10.2 A Ladder Network Model for Frequency Response Analysis 190
10.3 Model-Based Approach to Parameter Identification and Its Verification 191
10.4 Simulations and Discussions 192
10.5 Summary 195
References 195
Evidence-Based Winding Condition Assessment 196
11.1 Knowledge Transformation with Revised Evidential Reasoning Algorithm 196
11.2 A Basic Evaluation Analysis Model 197
11.3 A General Evaluation Analysis Model 198
11.4 Results and Discussions 199
11.5 Summary 204
References 205
Index 208