Distributed Energy Management of Electrical Power Systems
Description
Distributed Energy Management of Electrical Power Systems provides the most complete analysis of fully distributed control approaches and their applications for electric power systems available today. Authored by four respected leaders in the field, the book covers the technical aspects of control, operation management, and optimization of electric power systems.
In each chapter, the book covers the foundations and fundamentals of the topic under discussion. It then moves on to more advanced applications. Topics reviewed in the book include:
* System-level coordinated control
* Optimization of active and reactive power in power grids
* The coordinated control of distributed generation, elastic load and energy storage systems
Distributed Energy Management incorporates discussions of emerging and future technologies and their potential effects on electrical power systems. The increased impact of renewable energy sources is also covered.
Perfect for industry practitioners and graduate students in the field of power systems, Distributed Energy Management remains the leading reference for anyone with an interest in its fascinating subject matter.
More details
Other editions
Additional editions
Persons
YINLIANG XU, PHD, is now an Associate Professor with Tsinghua-Berkeley Shenzhen Institute, Tsinghua Shenzhen International Graduate School, Tsinghua University, P. R. China.
WEI ZHANG, PHD, is a Postdoc Resercher Associate with Department of Civil, Environmental, and Construction Engineering of College of Engineering & Computer Science, University of Central Florida, Orlando, Florida, USA.
WENXIN LIU, PHD, is an Associate Professor with the Department of Electrical and Computer Engineering, Lehigh University, Bethlehem, PA, USA.
WEN YU, PHD, is a Professor with the Departamento de Control Automatico with the Centro de Investigación y de Estudios Avanzados, Instituto Politécnico Nacional (CINVESTAV-IPN), Mexico City, Mexico.
Content
About the Authors xiii
Preface xv
Acknowledgment xix
List of Figures xxi
List of Tables xxxi
1 Background 1
1.1 Power Management 1
1.2 Traditional Centralized vs. Distributed Solutions to Power Management 4
1.3 Existing Distributed Control Approaches 5
2 Algorithm Evaluation 9
2.1 Communication Network Topology Configuration 9
2.1.1 Communication Network Design for Distributed Applications 9
2.1.2 N -1 Rule for Communication Network Design 11
2.1.3 Convergence of Distributed Algorithms with Variant Communication Network Typologies 13
2.2 Real-Time Digital Simulation 16
2.2.1 Develop MAS Platform Using JADE 16
2.2.2 Test-Distributed Algorithms Using MAS 18
2.2.2.1 Three-Agent System on the Same Platform 18
2.2.2.2 Two-Agent System with Different Platforms 19
2.2.3 MAS-Based Real-Time Simulation Platform 20
References 22
3 Distributed Active Power Control 23
3.1 Subgradient-Based Active Power Sharing 23
3.1.1 Introduction 24
3.1.2 Preliminaries - Conventional Droop Control Approach 26
3.1.3 Proposed Subgradient-Based Control Approach 27
3.1.3.1 Introduction of Utilization Level-Based Coordination 27
3.1.3.2 Fully Distributed Subgradient-Based Generation Coordination Algorithm 28
3.1.3.3 Application of the Proposed Algorithm 31
3.1.4 Control of Multiple Distributed Generators 33
3.1.4.1 DFIG Control Approach 33
3.1.4.2 Converter Control Approach 34
3.1.4.3 Pitch Angle Control Approach 35
3.1.4.4 PV Generation Control Approach 36
3.1.4.5 Synchronous Generator Control Approach 36
3.1.5 Simulation Analyses 37
3.1.5.1 Case 1 - Constant Maximum Available Renewable Generation and Load 38
3.1.5.2 Case 2 - Variable Maximum Available Renewable Generation and Load 41
3.1.6 Conclusion 45
3.2 Distributed Dynamic Programming-Based Approach for Economic Dispatch in Smart Grids 46
3.2.1 Introduction 46
3.2.2 Preliminary 49
3.2.3 Graph Theory 49
3.2.4 Dynamic Programming 49
3.2.5 Problem Formulation 49
3.2.6 Economic Dispatch Problem 50
3.2.7 Discrete Economic Dispatch Problem 50
3.2.8 Proposed Distributed Dynamic Programming Algorithm 51
3.2.9 Distributed Dynamic Programming Algorithm 52
3.2.10 Algorithm Implementation 53
3.2.11 Simulation Studies 54
3.2.12 Four-generator System: Synchronous Iteration 54
3.2.12.1 Minimum Generation Adjustment ¿pi = 2.5MW 54
3.2.12.2 Minimum Generation Adjustment ¿pi = 1.25MW 57
3.2.13 Four-Generator System: Asynchronous Iteration 59
3.2.13.1 Missing Communication with Probability 59
3.2.13.2 Gossip Communication 60
3.2.14 IEEE 162-Bus System 61
3.2.15 Hardware Implementation 63
3.2.16 Conclusion 64
3.3 Constrained Distributed Optimal Active Power Dispatch 65
3.3.1 Introduction 65
3.3.2 Problem Formulation 67
3.3.3 Distributed Gradient Algorithm 68
3.3.4 Distributed Gradient Algorithm 68
3.3.5 Inequality Constraint Handling 70
3.3.6 Numerical Example 72
3.3.6.1 Case 1 72
3.3.6.2 Case 2 74
3.3.7 Control Implementation 75
3.3.8 Communication Network Design 76
3.3.9 Generator Control Implementation 76
3.3.10 Simulation Studies 77
3.3.11 Real-Time Simulation Platform 78
3.3.12 IEEE 30-Bus System 78
3.3.12.1 Constant Loading Conditions 80
3.3.12.2 Variable Loading Conditions 82
3.3.12.3 With Communication Channel Loss 84
3.3.13 Conclusion and Discussion 86
3.A Appendix 86
References 87
4 Distributed Reactive Power Control 97
4.1 Q-Learning-Based Reactive Power Control 97
4.1.1 Introduction 98
4.1.2 Background 99
4.1.3 Algorithm Used to Collect Global Information 99
4.1.4 Reinforcement Learning 101
4.1.5 MAS-Based RL Algorithm for ORPD 101
4.1.6 RL Reward Function Definition 102
4.1.7 Distributed Q-Learning for ORPD 103
4.1.8 MASRL Implementation for ORPD 104
4.1.9 Simulation Results 106
4.1.10 Ward-Hale 6-Bus System 106
4.1.10.1 Learning from Scratch 108
4.1.10.2 Experience-Based Learning 110
4.1.10.3 IEEE 30-Bus System 112
4.1.10.4 IEEE 162-Bus System 114
4.1.11 Conclusion 115
4.2 Sub-gradient-Based Reactive Power Control 116
4.2.1 Introduction 116
4.2.2 Problem Formulation 119
4.2.3 Distributed Sub-gradient Algorithm 120
4.2.4 Sub-gradient Distribution Calculation 122
4.2.4.1 Calculation of ¿¿¿¿f /¿¿¿¿Qci for Capacitor Banks 122
4.2.4.2 Calculation of ¿¿¿¿f /¿¿¿¿Vgi for a Generator 124
4.2.4.3 Calculation of ¿¿¿¿f /¿¿¿¿tti for a Transformer 124
4.2.5 Realization of Mas-Based Solution 126
4.2.5.1 Computation of Voltage Phase Angle Difference 127
4.2.5.2 Generation Control for ORPC 128
4.2.6 Simulation and Tests 129
4.2.6.1 Test of the 6-BusWard-Hale System 129
4.2.6.2 Test of IEEE 30-Bus System 134
4.2.7 Conclusion 141
References 141
5 Distributed Demand-Side Management 147
5.1 Distributed Dynamic Programming-Based Solution for Load Management in Smart Grids 148
5.1.1 System Description and Problem Formulation 150
5.1.2 Problem Formulation 151
5.1.3 Distributed Dynamic Programming 153
5.1.3.1 Abstract Framework of Dynamic Programming (DP) 153
5.1.3.2 Distributed Solution for Dynamic Programming Problem 154
5.1.4 Numerical Example 157
5.1.5 Implementation of the LM System 158
5.1.6 Simulation Studies 160
5.1.6.1 Test with IEEE 14-bus System 160
5.1.6.2 Large Test Systems 166
5.1.6.3 Variable Renewable Generation 168
5.1.6.4 With Time Delay/Packet Loss 170
5.1.7 Conclusion and Discussion 171
5.2 Optimal Distributed Charging Rate Control of Plug-in Electric Vehicles for Demand Management 172
5.2.1 Background 175
5.2.2 Problem Formulation of the Proposed Control Strategy 175
5.2.3 Proposed Cooperative Control Algorithm 180
5.2.3.1 MAS Framework 180
5.2.3.2 Design and Analysis of Distributed Algorithm 180
5.2.3.3 Algorithm Implementation 181
5.2.3.4 Simulation Studies 183
5.3 Conclusion 190
References 191
6 Distributed Social Welfare Optimization 197
6.1 Introduction 197
6.2 Formulation of OEM Problem 200
6.2.1 SocialWelfare Maximization Model 200
6.2.2 Market-Based Self-interest Motivation Model 203
6.2.3 Relationship Between Two Models 204
6.3 Fully Distributed MAS-Based OEM Solution 207
6.3.1 Distributed Price Updating Algorithm 207
6.3.2 Distributed Supply-Demand Mismatch Discovery Algorithm 209
6.3.3 Implementation of MAS-Based OEM Solution 210
6.4 Simulation Studies 212
6.4.1 Tests with a 6-bus System 212
6.4.1.1 Test Under the Constant Renewable Generation 214
6.4.1.2 Test Under Variable Renewable Generation 217
6.4.2 Test with IEEE 30-bus System 218
6.5 Conclusion 221
References 221
7 Distributed State Estimation 225
7.1 Distributed Approach for Multi-area State Estimation Based on Consensus Algorithm 225
7.1.1 Problem Formulation of Multi-area Power System State Estimation 227
7.1.2 Distributed State Estimation Algorithm 228
7.1.3 Approximate Static State Estimation Model 231
7.1.4 Regarding Implementation of Distributed State Estimation 233
7.1.5 Case Studies 234
7.1.5.1 With the Accurate Model 235
7.1.5.2 Comparisons Between Accurate Model and Approximate Model 238
7.1.5.3 With Variable Loading Conditions 239
7.1.6 Conclusion and Discussion 241
7.2 Multi-agent System-Based Integrated Solution for Topology Identification and State Estimation 242
7.2.1 Measurement Model of the Multi-area Power System 244
7.2.2 Distributed Subgradient Algorithm for MAS-Based Optimization 245
7.2.3 Distributed Topology Identification 248
7.2.3.1 Measurement Modeling 248
7.2.3.2 Distributed Topology Identification 251
7.2.3.3 Statistical Test for Topology Error Identification 252
7.2.4 Distributed State Estimation 253
7.2.5 Implementation of the Integrated MAS-Based Solution for TI and SE 254
7.2.6 Simulation Studies 255
7.2.6.1 IEEE 14-bus System 255
7.2.6.2 Large Test Systems 263
7.3 Conclusion and Discussion 266
References 267
8 Hardware-Based Algorithms Evaluation 271
8.1 Steps of Algorithm Evaluation 271
8.2 Controller Hardware-In-the-Loop Simulation 273
8.2.1 PC-Based C-HIL Simulation 274
8.2.2 DSP-Based C-HIL Simulation 277
8.3 Power Hardware-In-the-Loop Simulation 279
8.4 Hardware Experimentation 281
8.4.1 Test-bed Development 281
8.4.2 Algorithm Implementation 284
8.5 Future Work 288
9 Discussion and Future Work 291
References 296
Index 297
Preface
The ever-growing demand, the rising penetration level of renewable generation, and the increasing complexity of electric power systems pose new challenges to control, operation, management, and optimization of power grids. Conventional centralized control structure requires a complex communication network with two-way communication links and a powerful central controller to process a large amount of data, which reduces overall system reliability and increases its sensitivity to failures; thus, it may not be able to operate under the increased number of distributed renewable generation units. Distributed control strategy enables easier scalability, simpler communication network, and faster distributed data processing, which can facilitate highly efficient information sharing and decision making. The distributed approach is a promising candidate to address the features of modern power grids by providing fast, flexible, reliable, and cost-effective solutions.
Considering the foreseeable large deployment of distributed renewable generation in electric power systems, a comprehensive technical source book is needed for both academic and industrial fields. This book will be providing in-depth analysis and discussion of fully distributed control approaches and their applications for electric power systems. The book will cover a wide range of the topics, including both large- (hours) and short- (seconds) time-scale-level coordination control and optimization. The technical aspects in terms of control, operation management, and optimization of electric power systems will be elaborated with the fundamental knowledge and advanced techniques.
This book focuses on fully distributed approaches for control, operation management, and optimization of electrical power systems with high distributed renewable generation penetration level. With the integration of more and more controllable and dispatchable aggregator of distributed generation, energy storage systems, elastic and inelastic loads, not only the power supply-demand should be balanced in a timely manner but also the frequency and voltage must be properly regulated to ensure efficient, safe, and reliable operation of power systems. Conventionally, power systems are commonly managed via centralized control approaches because of the existing structure of the traditional bulk power system. However, with the increasing number of highly intermittent distributed renewable generation with a vast geographical span, distributed control approaches have been proposed and thereafter been deployed in some pilot projects. It is foreseeable that distributed control approaches for power systems will be a promising and effective solution to advance the existing control and energy management technologies for grid integration of renewable energy sources.
In Chapter 1, an introduction on centralized and distributed solutions to power management and several distributed algorithms are provided. Chapter 2 introduces the communication network topology configuration, the multi-agent framework-based hardware-in-the-loop controller and setup of the real-time digital simulation test-bed.
Chapter 3 discusses distributed active power control and optimization of power systems. For a microgrid with a high level of renewable energy penetration to operate autonomously, it must maintain the instantaneous supply-demand balance of active power. A distributed sub-gradient-based solution to coordinate the operations of different types of distributed renewable generators in a microgrid is developed for the primary and secondary frequency control. An effective distributed tertiary control strategy based on distributed dynamic programming algorithm is then proposed to optimally allocate the total power demand among different generation units using asynchronous communication, which can lead to simpler implementation and faster convergence speed. Finally, an improved distributed gradient algorithmic-based approach, which can address both equality and inequality constraints, is proposed to realize online power system control and optimization.
In Chapter 4, to accommodate the increasing penetration level of distributed generators in the power system, appropriate voltage and reactive power control and optimization approaches are investigated to improve voltage profiles and reduce the power loss and the opportunity cost of reactive power generation. Distributed model-free Q-learning-based algorithm, which is easy to carry out and is adaptable in various operating conditions, is developed, and a sub-gradient-based reactive power control algorithm is proposed to reduce the possibility of random control updates and increase the rate of convergence.
In Chapter 5, a distributed demand response solution to maximize the overall utility of users while meeting the requirement for load reduction requested by the system operator is proposed in emerging smart grids is presented. The proposed distributed dynamic programming algorithm is implemented via a distributed framework, which can reduce the computation and communication burden and protect users' private information. As an important load in future smart grids, the flexibility of PEV's charging behavior can be beneficial to the power systems by offering various ancillary services, such as load leveling, reserve provision, frequency regulation, etc. It is important to design proper charging strategies so as to satisfy the consumers in terms of the financial charging cost, charging time, and state of charge, without violating operational constraints of the power system.
In Chapter 6, a distributed energy management approach that considers both generation suppliers and load users is proposed to maximize the total social welfare of the whole power system while satisfying various system constraints. Furthermore, an online solution based on a multi-agent system framework is developed to achieve better consumer participation and efficient supply-demand balance with faster response than traditional centralized methods.
In Chapter 7, two approaches for distributed estimation approaches are presented. The consensus algorithm-based distributed estimation is a two-loop iteration architecture, which is developed for multi-area power systems. The inner loop is used to discover the information of gain matrix and gradient vector by applying the consensus confusion technique, and the outer loop is used to update the estimation based on the second-order Newton's method. The distributed sub-gradient algorithm-based state estimation is a fully distributed integrated solution for multi-area TI and SE solution, which can be implemented with MAS framework. Numerical studies for variable scales of systems demonstrate that it is adaptable, robust and is a promising option for the state estimation of large interconnected power systems.
In Chapter 8, the steps to evaluate the algorithm are discussed. We start with controller-hardware-in-the-loop simulation, then the power hardware-in-the-loop simulation, and finally the hardware experimentation. Hardware experimentation is necessary, which can tell us whether an algorithm will work in reality or not and how to make it work and work better. In addition, it can also tell us the implementation requirement.
In Chapter 9, more issues and constraints in power system such as coupling variables, non-convexity, and communication topology change are investigated. Towards to more mature implementation of distributed control solution in real world, many facets and technology details such as investment, reliability, optimality, feasibility, etc. need to be taken into consideration, and research efforts and contributions from all around the world are more than welcome.
This book is written for power system engineers who are moving into the area of renewable energy and distributed generation, smart grids, economic dispatch, demand response, electric vehicle charging management, virtual power plant, state estimation, etc., and researchers and practitioners working in these areas who are enthusiastic to see what benefits and advantages distributed control and optimization algorithms can bring. This book follows an evolutionary procedure. Particularly, the founding technologies and background knowledge are systematically introduced first for each topic, and then the technical advances and the specific and emerging technologies will be elaborated. Therefore, the readers can get a clear picture of all the knowledge following a relatively fast learning curve. Meanwhile, only the proven and feasible solutions will be introduced in this book so that the applicability of the technical solutions can be guaranteed. Most of the advanced control and optimization strategies presented in this book are accompanied by extensive simulation and experimental results. Therefore, this book is also very useful for practitioners in this area to see how distributed control and optimization strategies could improve system performance and work in practice. This book also provides a precious opportunity for graduate students and researchers who work in the area to become familiar with the up-to-date techniques. It can be adopted as a textbook for graduate programs on power system engineering, microgrids, renewable energy integration, and smart grid.
Tsinghua-Berkeley Shenzhen Institute (TBSI)
Tsinghua Shenzhen International Graduate School
(Tsinghua SIGS), Tsinghua University, Shenzhen,...
