Artificial Intelligence Enabled Computational Methods for Smart Grid Forecast and Dispatch

Description

With the increasing penetration of renewable energy and distributed energy resources, smart grid is facing great challenges, which could be divided into two categories. On the one hand, the endogenous uncertainties of renewable energy and electricity load lead to great difficulties in smart grid forecast. On the other hand, massive electric devices as well as their complex constraint relationships bring about significant difficulties in smart grid dispatch.  Owe to the rapid development of artificial intelligence in recent years, several artificial intelligence enabled computational methods have been successfully applied in the smart grid and achieved good performances. Therefore, this book is concerned with the research on the key issues of artificial intelligence enabled computational methods for smart grid forecast and dispatch, which consist of three main parts.   (1) Introduction for smart grid forecast and dispatch, in inclusion of reviewing previous contribution of various research methods as well as their drawbacks to analyze characteristics of smart grid forecast and dispatch. (2) Artificial intelligence enabled computational methods for smart grid forecast problems, which are devoted to present the recent approaches of deep learning and machine learning as well as their successful applications in smart grid forecast. (3) Artificial intelligence enabled computational methods for smart grid dispatch problems, consisting of edge-cutting intelligent decision-making approaches, which help determine the optimal solution of smart grid dispatch.              The book is useful for university researchers, engineers, and graduate students in electrical engineering and computer science who wish to learn the core principles, methods, algorithms, and applications of artificial intelligence enabled computational methods. Foreword Preface Acknowledgments Contents 1 Introduction for Smart Grid Forecast and Dispatch 1.1 Smart Grid Forecast 1.2 Smart Grid Dispatch 1.2.1 Problem Statement 1.2.2 Problem Properties References 2 Review for Smart Grid Forecast 2.1 Introduction 2.2 The Load and Netload Forecasting 2.2.1 The Representative Patterns of Load Forecasting 2.2.2 The Statistical Model of Load/Net Load Forecasting 2.2.3 The Machine Learning Model of Load and Netload Forecasting 2.3 The Electrical Price Forecasting 2.3.1 The Mathematical Method for Electrical Price Forecasting 2.3.2 The Learning Method for Electrical Price Forecasting 2.4 The Electrical Vehicle Charging Station Charging Power Forecasting 2.4.1 Model-Based Electrical Vehicle Charging Station Charging Power Forecasting Method 2.4.2 Data-Driven Electrical Vehicle Charging Station Charging Power Forecasting Method References 3 Review for Smart Grid Dispatch 3.1 Introduction 3.2 Real-World Applications 3.2.1 Distribution Network 3.2.2 Microgrid Network 3.2.3 Electric Vehicles 3.2.4 Integrated Energy System 3.3 The Methods for Smart Grid Dispatch 3.3.1 Mathematical Programming 3.3.2 Evolutionary Algorithms 3.3.3 AI-Enabled Methods References 4 Deep Learning-Based Densely Connected Network for Load Forecast 4.1 Introduction 4.2 Residual Architecture 4.3 Unshared Convolution 4.4 Densely Connected Network 4.4.1 Overall Framework 4.4.2 Densely Connected Block 4.4.3 Clipped L2-norm 4.4.4 Smooth Loss 4.4.5 Smooth Quantile Regression 4.5 Case Study 4.5.1 Data Description 4.5.2 Case 1: Methods Validation 4.5.3 Case 2: Deterministic Forecasting 4.5.4 Case 3: Probabilistic Forecasting 4.6 Conclusion References 5 Reinforcement Learning Assisted Deep Learning for Probabilistic Charging Power Forecasting of EVCS 5.1 Introduction 5.2 Framework 5.2.1 Problem Formulation 5.2.2 The Probabilistic Forecast Framework of EVCS Charging Power 5.3 Data Transformer Method 5.4 Reinforcement Learning Assisted Deep Learning Algorithm 5.4.1 Long Short-Term Memory 5.4.2 The Modeling of LSTM Cell State Variation 5.4.3 Proximal Policy Optimization 5.5 Adaptive Exploration Proximal Policy Optimization 5.6 Case Study 5.6.1 Data Description and Experiential Initialization 5.6.2 The Performance of Probabilistic Forecasting Obtained by LSTM-AePPO 5.6.3 Metrics Comparison Among Different Algorithms 5.6.4 The Effectiveness of AePPO 5.7 Conclusion References 6 Dense Skip Attention-Based Deep Learning for Day-Ahead Electricity Price Forecasting with a Drop-Connected Structure 6.1 Introduction 6.2 Structure of the Proposed Framework 6.2.1 Data Preprocessing 6.2.2 Feature Extraction 6.2.3 Autoweighting of Features 6.2.4 Target Regression 6.3 Drop-Connected UCNN-GRU 6.3.1 Advanced Residual UCNN Block 6.3.2 Drop-Connected Structure 6.4 Dense Skip Attention Mechanism 6.4.1 Dense Skip Connection 6.4.2 Feature-Wise Attention Block 6.5 Case Study 6.5.1 Data Description 6.5.2 Implementation Details 6.5.3 Case 1: Model Effectiveness Evaluation 6.5.4 Case 2: Comparison with Statistical Techniques 6.5.5 Case 3: Comparison with Conventional DL Techniques 6.6 Conclusion 6.7 Quantile Regression 6.8 Formulation of the Evaluation Index 6.9 PReLU: A Solution to the Neuron Inactivation References 7 Uncertainty Characterization of Power Grid Net Load of Dirichlet Process Mixture Model Based on Relevant Data 7.1 Introduction 7.2 A Bayesian Framework Based on the Dirichlet Mixture Model of Data Association 7.2.1 Net Load Time-Series Correlation 7.2.2 Bayesian Framework Based on the Dirichlet Mixture Model of Data Association 7.2.3 Dirichlet Process and Folded Stick Construction Representation 7.2.4 Nonparametric Dirichlet Mixture Model 7.3 The Dirichlet Mixture Model Based on VBI for Data Association 7.3.1 Nonparametric Dirichlet Mixture Model 7.3.2 Variational Posterior Distribution Considering Data Association 7.4 Example Analysis 7.4.1 Description of the Algorithm 7.4.2 DDPMM Convergence Analysis 7.4.3 Analysis of DDPMM Fitting Effect 7.4.4 DDPMM Interval Indicator Analysis 7.5 Conclusion References 8 Extreme Learning Machine for Economic Dispatch with High Penetration of Wind Power 8.1 Introduction 8.1.1 Background and Motivation 8.1.2 Literature Review 8.1.3 Contribution of This Paper 8.2 Multi-objective Economic Dispatch Model 8.2.1 Formulations of Economic Dispatch 8.2.2 Multi-objective Economic Dispatch Model 8.3 Extreme Learning Machine Assisted Group Search Optimizer with Multiple Producers 8.3.1 Group Search Optimizer with Multiple Producers 8.3.2 ELM Assisted GSOMP 8.4 Simulation Studies 8.4.1 Simulation Settings 8.4.2 Simulation Results 8.5 Conclusion References 9 Multi-objective Optimization Approach for Coordinated Scheduling of Electric Vehicles-Wind Integrated Power Systems 9.1 Introduction 9.2 Operation Models of EV and Wind Power 9.2.1 Operational Model of EV Charging Station 9.2.2 Model of Uncertain Wind Power 9.2.3 Wind Power Curtailment Based on Probability Model 9.3 Coordinated Scheduling Model Integarated EV and Wind Power 9.3.1 Objective Functions 9.3.2 Decision Variables 9.3.3 Constraints 9.4 Solution of Coordinated Stochastic Scheduling Model 9.4.1 The Parameter Adaptive DE Algorithm 9.4.2 Decision-Making Method 9.4.3 Solution Procedure 9.5 Case Study 9.5.1 Case Description 9.5.2 Result Analysis 9.5.3 Algorithm Performance Analysis 9.6 Conclusion References 10 Many-Objective Distribution Network Reconfiguration Using Deep Reinforcement Learning-Assisted Optimization Algorithm 10.1 Introduction 10.2 Many-Objective Distribution Network Reconfiguration Model 10.2.1 Problem Formulations 10.2.2 Objectives 10.3 Deep Reinforcement Learning-Assisted Multi-objective Bacterial Foraging Optimization Algorithm 10.3.1 Multi-objective Bacterial Foraging Optimization Algorithm 10.3.2 Deep Reinforcement Learning 10.3.3 Multi-objective Material Foraging Optimization Algorithm Based on Deep Reinforcement Learning 10.4 Simulation Studies 10.4.1 Simulation Settings 10.4.2 Simulation Result and Analysis 10.4.3 Comparison with Other Algorithms 10.5 Conclusion References 11 Federated Multi-agent Deep Reinforcement Learning for Multi-microgrid Energy Management 11.1 Introduction 11.2 Theoretical Basis of Reinforcement Learning 11.3 Decentralized Multi-microgrid Energy Management Model 11.3.1 Isolated Microgrid Energy Management Model 11.3.2 Isolated MG Energy Management Model via MDP 11.3.3 Decentralized Multi-microgrid Energy Management Model 11.4 Federated Multi-agent Deep Reinforcement Learning Algorithm 11.4.1 Proximal Policy Optimization 11.4.2 Federated Learning 11.4.3 Federated Multi-agent Deep Reinforcement Learning Algorithm 11.5 Case Study 11.5.1 Experiment Setup 11.5.2 Analysis of the F-MADRL Algorithm 11.5.3 Performance Comparison 11.6 Conclusion References 12 Prospects of Future Research Issues 12.1 Smart Grid Forecast Issues 12.1.1 Challenges 12.1.2 Future Research Directions 12.2 Smart Grid Dispatch Issues 12.2.1 Challenges 12.2.2 Future Research Directions