This book aims to work out the distributed economic operation in smart grids in a systematic way, which ranges from model-based to model-free perspectives. The main contributions of this book can be summarized into three folds. First, we investigate the fundamental economic operation problems in smart grids from model-based perspective. Specifically, these problems can be modeled as deterministic optimization models, and we propose some distributed optimization algorithms by integrating the multi-agent consensus theory and optimization techniques to achieve the distributed coordination of various generation units and loads. Second, due to the randomness of the large-scale renewable energies and the flexibility of the loads, we further address these economic operation problems from a model-free perspective, and we propose learning-based approaches to address the uncertainty and randomness. At last, we extend the idea of model-based and model-free algorithms to plug-in electric vehicles (PEVs) charging/discharging scheduling problem, the key challenge of which involves multiple objectives simultaneously while the behavior of PEVs and the electricity price are intrinsically random. This book presents several recent theoretical findings on distributed economic operation in smart grids from model-based and model-free perspectives. By systematically integrating novel ideas, fresh insights, and rigorous results, this book provides a base for further theoretical research on distributed economic operation in smart grids. It can be a reference for graduates and researchers to study the operation and management in smart grids. Some prerequisites for reading this book include optimization theory, matrix theory, game theory, reinforcement learning, etc.
Author(s): Jiahu Qin, Yanni Wan, Fangyuan Li, Yu Kang, Weiming Fu
Series: Studies in Systems, Decision and Control, 455
Publisher: Springer
Year: 2023
Language: English
Pages: 245
City: Singapore
Preface
Contents
1 Introduction
1.1 Background and Motivation
1.2 Economic Operation in Smart Grid
1.2.1 Economic Dispatch
1.2.2 Demand Response
1.2.3 Unit Commitment
1.2.4 Optimal Power Flow
1.2.5 PEV Scheduling
1.3 Organization of the Book
References
2 Preliminaries
2.1 Graph Theory
2.2 Multi-Agent Consensus Algorithm
2.3 Related Optimization Algorithms
2.3.1 Classical Alternating Direction Method of Multiplier
2.3.2 Centralized Newton Method
2.3.3 Pattern Search Algorithm
2.4 Reinforcement Learning
2.5 Game Theory
References
3 Model-Based Distributed Optimization
3.1 Consensus-Based Coordination Between Economic Dispatch and Demand Response
3.1.1 Introduction
3.1.2 Problem Formulation
3.1.3 Consensus-Based Coordination Algorithm
3.1.4 Case Studies
3.1.5 Conclusion
3.2 Hierarchical Operation for Optimal Unit Commitment and Economic Dispatch
3.2.1 Introduction
3.2.2 Problem Formulation
3.2.3 Proposed Algorithm
3.2.4 Simulation Cases
3.2.5 Conclusion
3.3 Newton Method-Based Coordination for Multi-area Economic Dispatch
3.3.1 Introduction
3.3.2 Problem Formulation
3.3.3 Development of the Proposed NMDA
3.3.4 Case Studies
3.3.5 Conclusion
3.4 Negotiation-Based Cooperative Power Flow of Multiple Interconnected Microgrids
3.4.1 Introduction
3.4.2 Standalone Microgrid Model
3.4.3 Coupled Microgrid Model
3.4.4 Solution Methodology
3.4.5 Simulation Results
3.4.6 Conclusion
References
4 Model-Free Distributed Optimization
4.1 Distributed Pattern Search-Based Optimization for Economic Dispatch
4.1.1 Introduction
4.1.2 Problem Formulation
4.1.3 DPSA Algorithm for ED
4.1.4 Simulations and Discussions
4.1.5 Conclusion
4.2 Reinforcement Learning-Based Management of Residential Demand Response
4.2.1 Introduction
4.2.2 Problem Statement
4.2.3 Reinforcement Learning-Based DR Algorithm
4.2.4 Case Studies
4.2.5 Conclusion
4.3 Distributed Q-Learning-Based Optimization for Unit Commitment and Dispatch
4.3.1 Introduction
4.3.2 Problem Formulation
4.3.3 Q-Learning Based Online Optimization Algorithms
4.3.4 Theoretical Analysis
4.3.5 Case Studies
4.3.6 Conclusion
References
5 Extensions to PEVs Charging/Discharging Scheduling
5.1 Deep Reinforcement Learning-Based Coordinated PEVs Charging
5.1.1 Introduction
5.1.2 Problem Formulation and System Model
5.1.3 Distributed Deep Reinforcement Learning Algorithm
5.1.4 Evaluation Results
5.1.5 Conclusion
5.1.6 Appendix: Derivations of the Gradient of Actor Network
5.2 Multi-Agent Based Deep Reinforcement Learning for Charging Scheduling
5.2.1 Introduction
5.2.2 System Model and Problem Formulation
5.2.3 Markov Game Formulation
5.2.4 Multi-agent DRL-Based Charging Scheduling Algorithm
5.2.5 Simulation Results
5.2.6 Conclusion
5.3 Game Theoretic-Based Distributed Charging Scheduling Strategy
5.3.1 Introduction
5.3.2 System Model and Problem Formulation
5.3.3 Non-Cooperative PEV Charging Game
5.3.4 Distributed GNE-Seeking Algorithm
5.3.5 Simulation Results
5.3.6 Conclusion
References
