Enhancing resilience in power distribution systems / Fangxing Fran Li, Department of Electrical Engineering and Computer Science & CURENT, University of Tennessee, Knoxville TN, United States, Qingxin Shi, North China Electric Power University, Beijing, P.R. China, Jin Zhao, Trinity College Dublin, Dublin, Ireland.

Format:

Book

Author/Creator:

Li, Fangxing, author.

Shi, Qingxin (Researcher in electrical engineering), author.

Zhao, Jin, no linkage, author.

Language:

English

Subjects (All):

Electric power distribution--Reliability.

Electric power distribution.

Physical Description:

1 online resource (viii, 224 pages) : illustrations (chiefly color)

Place of Publication:

Amsterdam, Netherlands : Elsevier, [2025].

Biography/History:

"Fangxing ‘Fran’ Li is the James W. McConnell Professor in Electrical Engineering and the Campus Director of CURENT at the University of Tennessee at Knoxville, USA. His current research interests include resilience, artificial intelligence in power, demand response, distributed generation and microgrid, and energy markets. From 2020 to 2021, he served as the Chair of the IEEE PES Power System Operation, Planning and Economics (PSOPE) Committee. He has been the Chair of IEEE WG on Machine Learning for Power Systems since 2019 and the Editor-In-Chief of IEEE Open Access Journal of Power and Energy (OAJPE) since 2020. Prof. Li has received numerous awards and honours including R&D 100 Award in 2020, IEEE PES Technical Committee Prize Paper award in 2019, 5 best or prize paper awards at international journals, and 6 best papers/posters at international conferences."--Provided by publisher.

"Qingxin Shi is an Assistant Professor in the School of Electrical and Electronic Engineering, North China Electric Power University, Beijing, China. His research interests include demand response, resilient urban power systems, and hydrogen-electric integrated energy systems. He serves as the Associate Editor of Protection and Control of Modern Power Systems and the IEEE Open Access Journal of Power and Energy (OAJPE)."--Provided by publisher.

"Jin Zhao is an Assistant Professor in the Department of Electronic & Electrical Engineering, Trinity College Dublin, Ireland. Her research interests include power system resilience, climate adaptive energy systems, microgrids and machine learning. She currently serves as Senior Editor for IET Generation, Transmission & Distribution, Associate Editor for the IEEE Trans. on Smart Grid, Chair of the IEEE Task Force AISR, and as a member of the Steering Committee and PES representative for IEEE DataPort."--Provided by publisher.

Summary:

"Enhancing Resilience in Power Distribution Systems presents practical guidance for readers on the challenges and potential solutions for resilience in modern power systems. The book begins by explaining the risks and problems for resilience presented by renewable-based power systems. It goes on to clarify the current state of research and propose several novel methodologies and technologies for analysis and improvement of power system resilience. These methods include deep learning, linear programming, and generative adversarial networks.Packed with practical steps and tools for implementing the latest technologies, this book provides researchers and industry professionals with guidance on the resilient systems of the future. Key features: Breaks down novel methodologies and tools from deep learning to generative adversarial networks. Supports readers in implementing practical steps towards resilient renewable energy. Presents practical guidance for readers on the challenges and potential solutions for resilience in modern power systems"--Provided by publisher.

Contents:

1. Resilience in Modern Distribution Systems

2. Solutions, Current Issues, and Future Challenges

3. Components in Distribution Systems

4. Resilience-Oriented Long-term Planning in Distribution systems

5. Resilience-Oriented Short-term Planning in Urban-Level Power Networks

6. Optimal Operation to Enhance Distribution Resilience

7. Machine Learning for Pre-Event Preparation

8. Machine Learning for During-Event Mitigation

9. Machine learning for post-event restoration

10. Conclusions.

1 Resilience in modern distribution systems

1.1 Background

1.2 Fragility model of distribution components under extreme weather events

1.2.1 Hurricane or typhoon

1.2.1.1 Fault probability of high-voltage overhead lines

1.2.1.2 Fault probability of medium-voltage overhead lines

1.2.1.3 Fault probability of substations

1.2.2 Cryogenic, frozen rain, and snow disaster

1.2.3 Flood

1.2.4 Wildfire

1.3 Summary

References

2 Solutions, current issues, and future challenges

2.1 Resilient distribution system operation

2.1.1 Type I: Predisaster operation

2.1.2 Type II: Postdisaster restoration

2.1.3 Type III: Pre- and postdisaster actions

2.2 Resilience-oriented distribution system long-term planning

2.2.1 Trilevel robust optimization

2.2.2 Two-stage stochastic optimization

2.3 Discussion for future work

2.3.1 Predisaster resource allocation

2.3.2 Postdisaster resource dispatch

2.3.3 Cooperative restoration of urban electricity-water-gas system

2.4 Content in the remaining chapters

3 Components in distribution systems

3.1 Basic components of distribution systems

3.1.1 Distribution networks and voltage levels

3.1.2 Distribution transformers

3.1.3 Switchable components

3.1.4 Fault indicators

3.2 Emergency resources

3.2.1 Repair crews and supporting infrastrures

3.2.2 Distributed energy resources

3.2.3 Mobile power sources

3.3 Microgrid

4 Resilience-oriented long-term planning in distribution systems

4.1 Introduction

4.2 Problem formulation

4.2.1 The first-stage problem

4.2.2 The second-stage problem

4.2.3 The compact notation

4.3 Probabilistic scenario sampling and reduction.

4.3.1 Probabilistic model of line faults

4.3.2 Stochastic scenario generation

4.3.3 Scenario reduction based on K-means clustering

4.3.4 Resilience index

4.4 Solution algorithm

4.5 Case studies

4.5.1 IEEE 33-bus system

4.5.2 IEEE 123-bus system

4.6 Conclusion

5 Resilience-oriented short-term planning in urban-level power networks

5.1 Introduction

5.2 Topology feature and fragility model of urban power networks

5.3 Probabilistic modeling of load loss in urban power network

5.3.1 Load loss of medium-voltage distribution(MVD) network

5.3.1.1 Topology reduction

5.3.1.2 Power supply path and influence node matrix

5.3.1.3 Power supply mode a

5.3.1.4 Power supply mode b

5.3.2 Load loss of high-voltage network

5.3.2.1 Supply path search algorithm

5.3.2.2 Expected load loss calculation

5.3.2.3 Summary

5.4 Predisaster mobile power source allocation

5.5 Numerical study

5.5.1 Description of the test case

5.5.2 Simulation result

5.5.2.1 Expected load loss in medium-voltage distribution network

5.5.2.2 High-voltage distribution expected load loss

5.5.2.3 Discussion

5.6 Conclusion

6 Optimal operation to enhance distribution resilience

Nomenclature

Operator

Indices and sets

Parameters

Variables

6.1 Introduction

6.2 Framework of the outage management strategy

6.2.1 Operating feature of distribution networks and mobile resources

6.2.2 Framework of two-stage outage management strategy

6.3 Preventive allocation stage

6.3.1 The first stage

6.3.2 The second stage

6.3.3 Compact notation of two-stage stochastic optimization

6.3.4 Stochastic scenario generation based on Monte Carlo simulation

6.4 Postdisaster restoration stage

6.4.1 Repairing crew dispatch

6.4.2 Mobile power source dispatch.

6.4.3 Microgrid connectivity in dynamic formation

6.4.4 System operation

6.5 Solution algorithm of preventive allocationStage

6.6 Case study

6.6.1 IEEE 123-bus system

6.6.1.1 System description

6.6.1.2 Result analysis

6.6.1.2.1 Scenario 1

6.6.1.2.2 Scenario 2

6.6.2 Real-world 118-bus system

6.6.2.1 System description

6.6.2.2 Result analysis

6.6.3 Discussion

6.7 Conclusion

7 Machine learning for preparation before events

7.1 Introduction

7.2 Multimicrogrid formation using deep reinforcement learning framework

7.2.1 Formulate dynamic multimicrogrid formation as an Markov decision process

7.2.2 Dynamic multimicrogrid formation using deep Q-learning

7.3 Action generation of dynamic multimicrogrid formation process

7.3.1 Search space reduction of spinning forest

7.3.2 Convolutional neural network-based action-decoupling Q-value

7.4 Learning and application of CM-DDQN

7.4.1 Techniques for better learning

7.4.1.1 Experience replay and multiple buffers

7.4.1.2 Epsilon-greedy-based exploration

7.4.1.3 Fixed Q network (double deep Q-network)

7.4.2 CM-DDQN learning process

7.4.3 Deep reinforcement learning-based dynamic model-free multimicrogrid formation scheme

7.5 Case study

7.5.1 DRL-based dynamic multimicrogrid formation process using 7-bus system

7.5.1.1 Pretraining of deep reinforcement learning

7.5.1.2 Comparison of performances of different schemes

7.5.2 Performance of CM-DQN in IEEE 123-bus system

7.5.2.1 Comparison of learning abilities of deep reinforcement learning methods

7.5.2.2 Computation performance of the CM-DDQN

7.6 Conclusions

8 Machine learning for during-event mitigation

8.1 Introduction.

8.2 Generative adversarial network-based dataset processing for resilient networked microgrids under sequential extreme events

8.2.1 Review of Wasserstein generative adversarial network

8.2.2 Proposed self-attention generative adversarial network

8.2.3 Training process of self-attention generative adversarial network for data generation

8.3 Self-attention generative adversarial network integrated deep reinforcement learning for survival of critical loads

8.3.1 Double deep Q-network-based reconfiguration strategy for survival of critical loads

8.3.2 Design reward and environment

8.3.3 Generative adversarial network-enhanced deep Q-network

8.4 Case study

8.4.1 Self-attention generative adversarial network-based data processing

8.4.2 Self-attention generative adversarial network deep reinforcement learning for survival of critical loads in the 7-bus and 123-bus systems

8.5 Conclusion

9 Machine learning for postevent restoration

9.1 Introduction

9.2 Optimization models of robust load restoration

9.2.1 Objective function

9.2.2 Constraints

9.3 Deep learning-based uncertainty handling method

9.3.1 Find worst-case condition using deep neural network

9.3.1.1 Deep neural network data processing

9.3.1.2 Deep neural network structure

9.3.1.3 Training deep neural network

9.3.1.4 Accuracy improvement

9.3.2 Security check of worst-case condition using convolutional neural network

9.3.2.1 Convolutional neural network structure

9.3.2.2 Convolutional neural network data processing

9.3.2.3 Training convolutional neural network

9.4 Optimal robust load restoration strategy

9.4.1 Network topology list generation and load pickup list generation algorithms

9.4.1.1 Network topology list generation algorithm

9.4.1.2 Load pickup list generation algorithm.

9.4.2 Online optimal robust strategy acquisition

9.4.3 Case study

9.4.3.1 Data off-line training results

9.4.4 Load restoration performance of model-free method

9.5 Conclusion

10 Conclusions

Index.

Notes:

Includes bibliographical references and index.

Description based on online resource, publisher supplied metadata and other sources.

Other Format:

Print version:

ISBN:

9780443236396

0443236399

OCLC:

1526232229