Intelligent Systems and Signal Processing in Power Engineering ppt

Power Systems Abhisek Ukil Intelligent Systems and Signal Processing in Power Engineering Abhisek Ukil Author Intelligent Systems and Signal Processing in Power Engineering With 239 Figures and 36 Tables Author Dr. Abhisek Ukil ABB Corporate Research Segelhofstrasse 1K CH-5405 Baden-Daetwill Switzerland abhisek.ukil@ch.abb.com abhiukil@yahoo.com ISSN 1612-1287 ISBN 978-3-540-73169-6 Springer Berlin Heidelberg New York Library of Congress Control Number: 2007929722 Matlab  is a registered trademark of Mathworks Inc. In short, no guaraentees, whatsoever, are given for the example computer programs provided in this book. They are intended for demonstration purpose only. The author or the publisher would not be responsible for any consequences or damages of any sort from the usage of the programs or any other relevant ideas from the book. This work is subject to copyright. All rights are reserved, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilm or in any other way, and storage in data banks. Duplication of this publication or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its current version, and permission for use must always be obtained from Springer. Violations are liable for prosecution under the German Copyright Law. Springer is a part of Springer Science+Business Media springer.com c  Springer-Verlag Berlin Heidelberg 2007 The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. Typesetting: Integra Software Services Pvt. Ltd., India Cover Design: deblik, Berlin Printed on acid-free paper SPIN: 11884910 60/3180/Integra 5 4 3 2 1 0 Dedicated to all my teachers who enabled me to write this book, and my family & friends for supporting me all along. Preface Power engineering is truly one of the main pillars of the electricity-driven modern civilization. And over the years, power engineering has also been a multidisciplinary field in terms of numerous applications of different subjects. This ranges from lin- ear algebra, electronics, signal processing to artificial intelligence including recent trends like bio-inspired computation, lateral computing and the like. Considering the reasons behind this, in one hand, we have vast variety of application sub-domains in power engineering itself; on the other hand, problems in these sub-domains are complex and nonlinear, requiring other complementary techniques/fields to solve them. Therefore, there is always the need of bridging these different fields and the power engineering. We often encounter the problem of distributed and scattered nature of these different disciplines while working in various sub-domains of power engineering, trying to apply some other useful techniques for some specific problem. Oftentimes, these are not direct field of work for many power engineers/researchers, but we got to use them. This book is urged by that practical need. As the name suggests, the book looks into two major fields (without undermining others!) used in modern power systems. These are the intelligent systems and the signal processing. These broad fields include many topics. Some of the common and useful topics are addressed in this book. In this book, the intelligent systems section comprises of fuzzy logic, neural network and support vector machine. Fuzzy logic, driven by practical humanoid knowledge incorporation, has been a powerful technique in solving many nonlinear, complex problems, particularly in the field of control engineering. Neural network, on the other hand, is inspired by the biological neuronal assemblies that enable the animal kingdom (including us!) to perform complex tasks in everyday life. Sup- port vector machine is a relatively newer field in machine learning (neural network also falls in this category) domain. It augments the robustness of machine learning scenario with some new concepts and techniques. Although there are many more extensions of the concept of machine learning and intelligent systems, we confine ourselves to these three topics in this book. We look at some theories on them without assuming much particular background. Following the theoretical basics, we study their applications in various problems in power engineering, like, load forecasting, phase balancing, disturbance analysis and so on. Purpose of these, so called, application studies in power engineering is to demonstrate how we can utilize vii viii Preface the theoretical concepts. Finally, some research information are included, showing utilizations of these fields in various power systems domains as a starting point for further futuristic studies/research. In the second part, we look into the signal processing which is another universal field. Whenever and oftentimes we encounter signals, we need to process them! Power engineering and its enormous sub-fields are no exceptions, providing us with ample voltage, current, active/reactive power signals, and so forth. Therefore, we look in this section about the basics of the system theory, followed by fundamentals of different signal processing transforms with examples. After that, we look into the digital signal processing basics including the sampling technique and the digital filters which are the ultimate (signal) processing tools. Similar to the intelligent systems part, here also the theoretical basics are substantiated by some of the appli- cations in power engineering. These applications are of two types: full application studies explained like in-depth case-studies, and semi-developed application ideas with scope for further extension. This also ends up with pointers to further research information. As a whole, the book looks into the fields of intelligent systems and signal pro- cessing from theoretical background and their application examples in power sys- tems altogether. It has been kind of hard to balance the theoretical aspects as each of these fields are vast in itself. However, efforts have been made to cover the essential topics. Specific in-depth further studies are pointed to the dedicated subject intensive resources for interested readers. Application studies are chosen with as much real implications as possible. Finally, the book is a small effort to bridge and put together three great fields as a composite resource: intelligent systems, signal processing and power engineering. I hope this book will be helpful to undergraduate/graduate students, researchers and engineers, trying to solve power engineering problems using intelligent systems and signal processing, or seeking applications of intelligent systems and signal process- ing in power engineering. April, 2007 Abhisek Ukil Contents 1 Introduction 1 1.1 About the Book . . . 1 1.2 ProspectiveAudience 1 1.3 OrganizationoftheBook 2 1.3.1 BookChapters 2 1.3.2 ChapterStructure 3 2 Fuzzy Logic 5 2.1 Introduction . 5 2.1.1 History and Background . . . 5 2.1.2 Applications 6 2.1.3 ProsandCons 7 2.2 FuzzyLogic 8 2.2.1 Linguistic Approach . 8 2.2.2 SetTheory 9 2.2.3 FuzzySetTheory 12 2.2.4 ClassicalSetTheoryvs.FuzzySetTheory 24 2.2.5 Example 27 2.3 FuzzySystemDesign 28 2.3.1 Fuzzification 29 2.3.2 FuzzyInference 29 2.3.3 Defuzzification 34 2.4 ApplicationExample 35 2.4.1 BrakeTestApplication 36 2.4.2 Fuzzification 37 2.4.3 FuzzyInference 40 2.4.4 Defuzzification 42 2.4.5 Conclusion 45 References 46 2.5 LoadBalancing 46 2.5.1 FeederRepresentation 47 2.5.2 Proposed Technique. . 48 2.5.3 DesigningFuzzyController 49 2.5.4 Results 51 References 55 ix x Contents 2.6 EnergyEfficientOperation 56 2.6.1 ProjectOverview 56 2.6.2 DesigningFuzzyController 56 2.6.3 FinalOutput 57 References 58 2.7 Stability Analysis . . 58 2.7.1 UseofFuzzyLogic 58 References 59 2.8 DemandSideManagement 59 2.8.1 LoadProfiling 59 2.8.2 EnergyConsumptionModeling 60 Reference 61 2.9 PowerFlowController 61 2.9.1 SystemOverview 61 Reference 62 2.10 ResearchInformation 62 2.10.1 GeneralFuzzyLogic 62 2.10.2 FuzzyLogicandPowerEngineering 63 2.10.3 Electrical Load Forecasting 63 2.10.4 FaultAnalysis 64 2.10.5 PowerSystemsProtection 65 2.10.6 DistanceProtection 65 2.10.7 Relay 65 2.10.8 PowerFlowAnalysis 66 2.10.9 PowerSystemsEquipments&Control 67 2.10.10 FrequencyControl 67 2.10.11 HarmonicAnalysis 68 2.10.12 PowerSystemsOperation 68 2.10.13 PowerSystemsSecurity 69 2.10.14 Power Systems Reliability . 69 2.10.15 PowerSystemsStabilizer 70 2.10.16 Power Quality . . 71 2.10.17 RenewableEnergy 71 2.10.18 Transformers 71 2.10.19 RotatingMachines 72 2.10.20 Energy Economy, Market & Management . . . 73 2.10.21 UnitCommitment 73 2.10.22 Scheduling 73 2.10.23 PowerElectronics 74 3 Neural Network 75 3.1 Introduction . 75 3.1.1 History and Background . . . 76 3.1.2 Applications 76 3.1.3 ProsandCons 78 Contents xi 3.2 Artificial Neural Networks (ANN). . . 78 3.2.1 BasicStructureoftheArtificialNeuralNetworks 78 3.2.2 StructureofaNeuron 79 3.2.3 TransferFunction 81 3.2.4 ArchitectureoftheANN 84 3.2.5 StepstoConstructaNeuralNetwork 85 3.3 LearningAlgorithm 85 3.3.1 TheDeltaRule 86 3.3.2 GradientDescent 87 3.3.3 EnergyEquivalence 88 3.3.4 TheBackpropagationAlgorithm 88 3.3.5 TheHebbRule 92 3.4 DifferentNetworks 93 3.4.1 Perceptron 93 3.4.2 Multilayer Perceptrons (MLP) . . . 94 3.4.3 Backpropagation(BP)Network 95 3.4.4 RadialBasisFunction(RBF)Network 96 3.4.5 HopfieldNetwork 104 3.4.6 Adaline 105 3.4.7 Kohonen Network . . . 105 3.4.8 SpecialNetworks 108 3.4.9 SpecialIssuesinNNTraining 111 3.5 Examples 114 3.5.1 LinearNetwork:BooleanLogicOperation 115 3.5.2 Pattern Recognition . . 117 3.5.3 Incomplete Pattern Recognition . . 120 References 126 3.6 Load Forecasting . . 127 3.6.1 DatasetfortheApplicationStudy 128 3.6.2 UseofNeuralNetworks 129 3.6.3 LinearNetwork 129 3.6.4 BackpropagationNetwork 131 3.6.5 RadialBasisFunctionNetwork 134 References 137 3.7 FeederLoadBalancing 138 3.7.1 PhaseBalancingProblem 139 3.7.2 FeederReconfigurationTechnique 139 3.7.3 NeuralNetwork-basedSolution 140 3.7.4 NetworkTraining 141 3.7.5 Results 142 References 143 3.8 FaultClassification 143 3.8.1 Simple Ground Fault Classifier . . 144 3.8.2 AdvancedFaultClassifier 144 Reference 145 xii Contents 3.9 Advanced Load Forecasting . . . 145 References 146 3.10 Stability Analysis . . 146 References 147 3.11 ResearchInformation 148 3.11.1 GeneralNeuralNetworks 148 3.11.2 NeuralNetworkandPowerEngineering 148 3.11.3 Electrical Load Forecasting 149 3.11.4 FaultLocator&Analysis 150 3.11.5 PowerSystemsProtection 151 3.11.6 HarmonicAnalysis 152 3.11.7 TransientAnalysis 152 3.11.8 PowerFlowAnalysis 153 3.11.9 PowerSystemsEquipments&Control 153 3.11.10 PowerSystemsOperation 154 3.11.11 PowerSystemsSecurity 154 3.11.12 Power Systems Reliability . 155 3.11.13 Stability Analysis 155 3.11.14 RenewableEnergy 156 3.11.15 Transformers 156 3.11.16 RotatingMachines 157 3.11.17 Power Quality . . 158 3.11.18 StateEstimation 158 3.11.19 EnergyMarket 158 3.11.20 PowerElectronics 159 4 Support Vector Machine 161 4.1 Introduction . 161 4.1.1 History and Background . . . 161 4.1.2 Applications 162 4.1.3 ProsandCons 163 4.2 Basics about Statistical Learning Theory . . 164 4.2.1 MachineLearning&AssociatedProblem 164 4.2.2 StatisticalLearningTheory 165 4.2.3 VapnikChervonenkis(VC)Dimension 167 4.2.4 StructuralRiskMinimization 169 4.3 Support Vector Machine . 171 4.3.1 LinearClassification 171 4.3.2 OptimalSeparatingHyperplane 174 4.3.3 Support Vectors 179 4.3.4 ConvexOptimizationProblem 181 4.3.5 OverlappingClasses 183 4.3.6 NonlinearClassifier 185 4.3.7 KernelMethod 186 4.3.8 Support Vector Regression . 193 [...]... of the intelligent systems, signal processing techniques, power engineering has truly become a multi-disciplinary field Modern applications of intelligent systems, e.g., fuzzy logic, neural network, support vector machines, etc and signal processing, like, the Fourier transform-based digital filters, etc are being applied more and more in various sub-domains of the vast power engineering These include... intelligent systems and signal processing techniques in power engineering applications However, ample power engineering specific and application-oriented references are provided in order to follow up particular further research objective Also, power engineering researchers and industry people would be interested at the dispositions of the different multi-domain present and future trends of research in power engineering. .. research in power engineering Basic theoretical discussions are followed up with ample pointers to the up to date specialized references, which should be a starting/supporting point for multi-disciplinary projects involving the intelligent systems, signal processing and power engineering Basic understanding of power engineering concepts, matrix computation, complex numbers are in general assumed Nevertheless,... depicted in Fig 1.1 1.2 Prospective Audience The book is primarily for the graduate and the undergraduate students of electrical engineering, power engineering and the related fields This book is not a textbook 1 2 1 Introduction Fig 1.1 Subject overview of the book for power engineering, rather it is mainly oriented towards inter-disciplinary applications in power engineering, demonstrating how to apply intelligent. .. load-forecasting, load-balancing, load-profiling, disturbance analysis, fault classification, energy management, energy efficient operation and so on However, it is often difficult to find a resource off the shelf which can altogether provide basic understanding of the various important intelligent systems and signal processing technologies along with possible applications in the power engineering Power engineering. .. vector machine share some common grounds on machine learning The book starts with this introductory chapter explaining the scope and the layout of the book This is followed by part I: intelligent systems and part II: signal processing Part I contains three chapters: Chap 2 on fuzzy logic, Chap 3 on neural network, Chap 4 on support vector machine Part II contains Chap 5 on signal processing The chapters... under one cover in a modular fashion Exclusive subject-intensive references are provided for further specialized studies Modern, up to date applications as thorough case studies alongside many a prospective project idea would nurture the current research trends and inspire future exhaustive, inter-disciplinary research works involving intelligent systems, signal processing and power engineering The subject... Rotating Machines 366 5.13.24 Power Electronics 367 Index 369 Chapter 1 Introduction 1.1 About the Book Power engineering is an ever-growing, important, multi-dimensional field for electrical engineering students and the associated industry people And with the increasing... dealing with the intelligent systems and part II, the signal processing Both part I and II would be modular in nature, treating each specific topics with in depth theory, ample practical applications, future directions and references Here, ‘modular’ means mostly mutually exclusive, self-contained chapters However, some chapters share some points, like the neural 1.3 Organization of the Book 3 network and. .. 5.12.2 Inrush Currents 352 5.12.3 Analyzing Lightning Strike 352 References 352 5.13 Research Information 353 5.13.1 General Signal Processing 353 5.13.2 Signal Processing and Power Engineering . Power Systems Abhisek Ukil Intelligent Systems and Signal Processing in Power Engineering Abhisek Ukil Author Intelligent Systems and Signal Processing in. problems using intelligent systems and signal processing, or seeking applications of intelligent systems and signal process- ing in power engineering. April,