New Technologies for Power System Operation and Analysis

Description

New Technologies for Power System Operation and Analysis considers the very latest developments in renewable energy integration and system operation, including electricity markets and wide-area monitoring systems and forecasting. Helping readers quickly grasp the essential information needed to address renewable energy integration challenges, this new book looks at basic power system mathematical models, advanced renewable integration and system optimizations from transmission and distribution system sides. Sections cover wind, solar, gas and petroleum, making this a useful reference for all engineers interested in power system operation. New Technologies for Power System Operation and Analysis Copyright Contents List of contributors one Introduction 1.1 Overview of power systems 1.2 The development history of power systems in the United States 1.3 Distributed energy resource units 1.4 Steady-state conditions 1.5 AI and machine learning 1.6 Network dynamic operation 1.7 Multisector coupling 1.8 Structure of this book References two Decoupled linear AC power flow models with accurate estimation of voltage magnitude in transmission and distribution sy... Nomenclature 2.1 Introduction 2.2 Linear power flow models for the meshed transmission systems 2.2.1 Decoupling of voltage magnitude and phase angle 2.2.2 Matrix formulation of the decoupled linearized power flow model 2.2.3 Transformers and phase shifters 2.2.3.1 Contributions to the admittance matrix 2.2.3.2 Influence on the branch MW flow 2.2.4 Derivation and justification of the fast decoupled linearized power flow 2.2.4.1 An illustrative example 2.2.4.2 Theoretical derivation 2.2.4.3 A numerical example 2.3 Linear three-phase power flow models of the unbalanced distribution systems 2.4 Case study 2.4.1 Meshed transmission systems 2.4.2 Balanced distribution systems 2.4.3 Unbalanced distribution systems 2.5 Conclusion References three Renewable energy integration and system operation challenge: control and optimization of millions of devices 3.1 Introduction 3.2 Distribution system model with high penetration of renewables 3.2.1 Distribution network model 3.2.2 An explicit branch model of distribution network 3.2.3 Dynamic distributed generation model 3.3 Autonomous distributed voltage control 3.3.1 Distributed subgradient algorithm 3.3.2 Distributed subgradient voltage control 3.3.3 Reactive power control and power factor control 3.4 Hierarchical multiagent control of large-scale distribution system 3.4.1 Virtual leader design 3.4.2 Case study 3.5 Islanded microgrid with high penetration of distributed generations 3.6 Grid-edge situational awareness: enhanced observability by voltage inference 3.6.1 Voltage inference method 3.6.2 Network sensitivity 3.6.3 Implementation 3.7 Control-enabled dynamic hosting allowance: P and Q control capacity and impact analysis 3.7.1 Traditional hosting analysis 3.7.2 Dynamic hosting allowance analysis 3.8 Cosimulation of integrated transmission and distribution systems 3.8.1 The framework of cosimulation 3.8.1.1 Integrated T&D system simulation 3.8.1.2 Parallel cosimulation of integrated T&D systems 3.8.2 Simulation results 3.8.2.1 Power flow of integrated T&D systems 3.8.2.2 Integrated T&D systems with PVs and control 3.9 Conclusion References four Advances of wholesale and retail electricity market development in the context of distributed energy resources 4.1 Introduction 4.2 Modern wholesale electricity market 4.2.1 Driving forces of the electricity market development 4.2.2 Operation process of wholesale electricity markets 4.2.3 New products and designs in wholesale electricity markets 4.2.3.1 Considering flexibility in resource adequacy 4.2.3.2 Flexible ramping products 4.2.3.3 Energy storage market modeling 4.2.3.4 Demand response resources market modeling 4.2.4 Running the electricity market with 100% renewables 4.2.4.1 Motivations 4.2.4.2 Bulk electric power grid dispatch in the all inverter-based resources systems 4.2.4.3 Increased uncertainty and variability 4.2.4.4 Increased complexity for the controls of utility-scale inverter-based resources 4.2.4.5 All zero marginal cost resources 4.2.4.6 Unit commitment and economic dispatch for up to 100% renewables 4.2.4.7 Determination of reserve requirements 4.2.4.8 New frequency response mechanism design 4.3 Modern retail electricity market 4.3.1 State-of-the-art for retail electricity market development 4.3.2 Design of retail electricity market operation framework 4.3.3 Implementation of blockchain technology in electric power systems 4.4 Conclusion References five Wide-area monitoring and anomaly analysis based on synchrophasor measurement 5.1 Synchrophasor measurement technology introduction 5.1.1 Situational awareness 5.1.2 Advanced control 5.2 Wide-area measurement system example—FNET/GridEye 5.3 FNET/GridEye wide-area measurement system applications overview 5.3.1 Visualization 5.3.2 Disturbance detection and location 5.3.3 Interarea oscillation detection and event-data-based oscillation modal analysis 5.3.4 Online ambient-data-based oscillation modal analysis 5.3.5 Islanding detection 5.3.6 Event replay and postevent analysis 5.3.7 Statistical analysis of historical data 5.3.8 Model validation and parameter verification 5.3.9 Machine learning–based inertia estimation References Further reading six Advanced grid operational tools based on state estimation 6.1 Introduction 6.2 Model validation 6.2.1 Largest Normalized Lagrange Multiplier test 6.2.1.1 Extraction of Lagrange multipliers from state estimation problem 6.2.1.2 Normalized Lagrange multipliers and hypothesis testing 6.2.1.3 Detection, identification, and correction of model parameter errors 6.2.2 Computationally efficient implementation of the Largest Normalized Lagrange Multiplier test 6.2.3 Detectability and identifiability of parameter and measurement errors 6.3 System monitoring 6.3.1 Motivations for dynamic state estimation 6.3.2 Problem formulation of dynamic state estimation 6.3.3 Unified framework for Bayesian dynamic state estimation through nonlinear regression 6.3.3.1 Bayesian state estimators for nonlinear dynamic system models 6.3.3.2 Proposed unified framework for nonlinear Bayesian dynamic state estimation 6.3.3.3 Proposed framework for robustifying the Bayesian dynamic state estimation 6.3.3.4 Decentralized versus centralized dynamic state estimation for power system 6.4 Protective relaying 6.4.1 Theoretical basis of dynamic state estimation–based protection 6.4.1.1 Dynamic model of the protection zone 6.4.1.2 Quantification of consistency through dynamic state estimation algorithm 6.4.1.3 Trip decision 6.4.2 Numerical experiments 6.4.2.1 Example test system: series compensated transmission line 6.4.2.2 Dynamic model of the series compensated transmission line 6.4.2.2.1 Dynamic model of section k in the multisection line 6.4.2.2.2 Dynamic model of the three-phase series capacitors 6.4.2.2.3 Formulation of the overall dynamic model 6.4.2.3 Legacy protection functions for comparison and corresponding settings 6.4.2.4 Event study 6.5 Conclusion remarks References Further reading seven Advanced machine learning applications to modern power systems 7.1 Introduction 7.2 Modern forecasting technology 7.2.1 Prior research work 7.2.1.1 Deterministic forecasting methods 7.2.1.2 Probabilistic forecasting methods 7.2.1.3 Ensemble learning 7.2.2 Ensemble learning forecasting methodologies 7.2.2.1 Single-machine learning algorithm models 7.2.2.2 Competitive ensemble learning 7.2.2.3 Cooperative ensemble learning 7.2.3 Forecasting results 7.2.3.1 Case study I: wind speed forecasting based on competitive ensemble learning 7.2.3.2 Case study II: wind power forecasting based on cooperative ensemble learning 7.3 Machine learning–based control and optimization 7.3.1 Prior research work 7.3.1.1 Machine learning–based control 7.3.1.2 Machine learning–based optimization 7.3.2 A Machine learning–based network reconfiguration methodology 7.3.2.1 Literature review on network reconfiguration 7.3.3 Network reconfiguration results 7.4 Advanced artificial intelligence and machine learning applications to building occupancy detection 7.4.1 Prior research work 7.4.2 The convolutional neural network–long short-term memory deep learning architecture 7.4.2.1 Occupancy detection problem formulation 7.4.2.2 Convolutional neural network 7.4.2.3 Long short-term memory network 7.4.2.4 The developed convolutional neural network–long short-term memory architecture 7.4.3 Experiments 7.4.4 Results 7.5 Conclusion References eight Power system operation with power electronic inverter–dominated microgrids Nomenclature 8.1 Power system evolution toward modernization 8.2 Networked microgrids with parallel inverters 8.2.1 Advanced microgrid structures 8.2.2 Concept of dynamic microgrids 8.3 Parallel inverter operation in microgrids 8.3.1 Parallel inverter operation in the context of dynamic microgrids—steady-state operation 8.3.2 Parallel inverters operation in the context of dynamic microgrids—transient-state operation 8.3.2.1 Inverter dynamic stability during network reconfiguration 8.3.2.2 Network reconfiguration with improved inverter operation performance 8.3.2.3 Seamless network reconfiguration using advanced inverter control 8.4 Conclusion References nine Automated optimal control in energy systems: the reinforcement learning approach 9.1 Introduction 9.1.1 Background 9.1.2 Markov decision process and Bellman equations 9.1.2.1 Policy 9.1.2.2 Value function 9.1.2.3 Bellman equations 9.1.2.4 Bellman expectation equations 9.1.2.5 Bellman optimal equations 9.1.3 Solving Markov decision process problems 9.1.4 Value-based reinforcement learning 9.1.4.1 Monte Carlo reinforcement learning 9.1.4.2 Temporal difference reinforcement learning 9.1.4.3 Value function approximation 9.1.5 Policy-based reinforcement learning 9.1.6 Actor–critic reinforcement learning 9.1.7 Summary 9.2 Deep reinforcement learning 9.2.1 What is deep reinforcement learning 9.2.2 Introduction to three deep reinforcement learning algorithms 9.2.2.1 Deep Q-network: a value-based approach 9.2.2.2 Asynchronous advantage actor–critic: an actor–critic approach 9.2.2.3 Evolution strategies–based reinforcement learning 9.2.3 Scalable reinforcement learning frameworks 9.2.4 Curriculum learning 9.2.5 Meta learning 9.2.6 Multiagent system 9.2.7 Summary 9.3 Reinforcement learning in energy systems 9.3.1 Advantages of applying reinforcement learning in engineering problems 9.3.2 Training a reinforcement learning controller 9.3.2.1 Typical workflow 9.3.2.2 Building a reinforcement learning environment 9.3.2.3 Selecting the right algorithm 9.3.3 Reinforcement learning applications in energy systems 9.3.3.1 Smart buildings 9.3.3.2 Demand response 9.3.3.3 Grid operation 9.3.3.4 Renewable generation and battery control 9.3.4 Some interesting research topics 9.3.5 Limitations and challenges 9.3.6 Summary References Ten Power, buildings, and other critical networks: Integrated multisystem operation 10.1 Introduction 10.1.1 Aging infrastructure and climate-related impacts 10.1.2 Increasing electrification in the built environment 10.1.3 Increasing connectivity in power, water, and gas networks 10.2 Grid-interactive buildings 10.2.1 Distributed energy resources 10.2.2 Demand response 10.2.3 Emerging considerations for demand response 10.2.4 Climate and environment 10.2.5 Cybersecurity and privacy 10.3 Interdependent critical networks 10.3.1 Water and energy 10.3.2 Power and gas 10.3.3 Combined heat and power 10.4 Electrification of the transportation sector 10.4.1 Consumer vehicles 10.4.2 Public transportation 10.4.3 Rideshare services and emerging methods of transportation 10.4.4 Vehicle-to-grid 10.5 Considerations for future power systems 10.5.1 Physical considerations 10.5.2 Market and organizational considerations 10.5.3 Cyber considerations Acknowledgments References Index