Applications of deep machine learning in future energy systems /

This book explores the applications of deep machine learning in the realm of future energy systems, providing insights into the integration of artificial intelligence in energy management and control systems. The book is edited by Khooban and features contributions from experts in the field. It delv...

Full description

Bibliographic Details
Corporate Author:	ScienceDirect (Online service)
Format:	eBook
Language:	English
Published:	Amsterdam : Elsevier, 2024.
Subjects:	Deep learning (Machine learning) Electric power systems. Apprentissage profond. Réseaux électriques (Énergie) Electronic books.
Online Access:	Connect to the full text of this electronic book

Table of Contents:

Front Cover
Applications of Deep Machine Learning in Future Energy Systems
Copyright Page
Contents
List of contributors
Preface
1 Introduction
2 Artificial intelligence and machine learning in future energy systems (state-of-the-art, future development)
2.1 Introduction
2.2 Machine learning basics
2.2.1 Supervised learning
2.2.2 Unsupervised learning
2.2.3 Reinforcement learning
2.3 Machine learning in future energy systems
2.3.1 Forecasting
2.3.2 Fault and anomaly detection
2.3.3 Control and operation
2.4 General observation and future trends
2.5 Conclusion
References
3 Digital twins−assisted design of next-generation DC microgrid
3.1 Introduction
3.2 Modeling of DCMG
3.2.1 PV system
3.2.2 Battery ESS
3.2.3 DCMG's general state space model
3.3 Proposed control
3.3.1 Nonlinear droop control
3.3.2 The ULM controller
3.3.3 Basic Q-learning
3.4 Digital twin−based control strategy for DCMG
3.4.1 The ISSA-based digital twin system parameter estimate
3.5 Results
3.5.1 Scenario 1
3.5.2 Scenario 2
3.5.3 Scenario 3
3.5.4 Scenario 4
3.6 Conclusion
Acknowledgement
References
4 Intelligent charging station recommendations for electric vehicles in the charging market: a fuzzy−deep learning approach
4.1 Introduction
4.2 Methodology
4.2.1 System model
4.2.2 SoC estimation
4.3 FDCR model in the charging market
4.3.1 EVCSs model
4.3.2 EV model
4.3.2.1 Charging demand
4.3.2.2 Arrival time
4.3.2.3 Staying time
4.3.2.4 EV driver classification
4.3.3 Decision-making criteria
4.4 Supervised learning
4.4.1 Fuzzy-supervised learning algorithm
4.5 FDCR's decision-making strategy
4.5.1 Basic concept of LSTM NN
4.6 Score prediction of each EVCS
4.7 Simulation results
4.7.1 Simulation setup.
4.8 Conclusion
Acknowledgment
References
5 Deep frequency control of power grids under cyber attacks
5.1 Introduction
5.2 Basics of frequency control in power grids
5.3 LFC system vulnerability to cyber attacks
5.4 Modeling of the LFC system under cyber attacks
5.5 Category of cyber attacks on the LFC system
5.6 DoS attack
5.7 Time delay attack
5.8 FDI attack
5.9 Replay attack
5.10 Covert attack
5.11 Zero dynamic attack
5.12 Deep learning−based methods
5.13 Conclusion
References
6 Application of Q-learning in stabilization of multicarrier energy systems
6.1 Introduction
6.2 Methodologies for modeling
6.2.1 CHP model
6.2.2 PV model
6.2.3 Battery model
6.2.4 System dynamic model
6.3 Q-Learning-based ULM controller
6.3.1 The ULM controller
6.3.2 Basic Q-Learning
6.3.3 Multiagent fuzzy Q-learning
6.4 Real-time results
6.4.1 Scenario A: the absorption chiller meets the thermal load
6.4.2 Scenario B: the absorption chiller unmeets the thermal load
6.4.3 Scenario C: the robustness analysis
6.5 Conclusion
Acknowledgment
References
7 Design of next-generation of 5G data center power supply based on artificial intelligence
7.1 Introduction
7.1.1 Model description of dc/dc full bridge dc/dc converter
7.1.1.1 Adaptive model-free sliding mode control based on deep reinforcement learning
7.1.1.1.1 Structure of nonlinear model-free sliding mode controller
7.1.1.1.2 Deep reinforcement learning−based controller design
7.1.1.1.3 DDPG-based actor-critic framework
7.1.1.1.4 Design of DNNs of actor-critic
7.2 Real-time simulation verifications
7.3 Conclusion
Appendix A
Appendix B
References
8 Smart EV battery charger based on deep machine learning
Nomenclature
8.1 Introduction
8.2 Bus, path, and power requirements.
10.2.2 Calculation of system's error
10.3 Sensorless backstepping controller
10.3.1 Observer design
10.3.2 Desired magnitude for current
10.3.3 Controller design
10.3.4 Stability analysis
10.3.4.1 Stability of the equilibrium point
10.3.4.2 Robustness of the closed-loop system
10.4 SAC-based parameter tuner for sensorless controller design
10.4.1 Notation
10.4.2 SAC strategy
10.4.3 The SAC-DRL based on SMC controller
10.5 Real-time simulation verifications
10.6 Conclusion
References
11 Multilevel energy management and optimal control system in smart cities based on deep machine learning
11.1 Introduction
11.1.1 Problem statement
11.1.2 Literature review
11.1.3 Highlights and main contribution
11.2 IEMOCS topology and three-level model in INMG structure
11.2.1 The primary level of IEMOCS operation in the IMG structure
11.2.2 The secondary level of IEMOCS operation in the IMG structure
11.2.3 The tertiary level of IEMOCS operation in the IMG structure
11.2.4 DR programs structure with load shift format
11.2.5 Uncertainty of RER and load
11.3 IEMOCS structure based on hybrid TSKFS&amp
MADRL model
11.3.1 TSKFS modeling based on uncertainties in MADRL structure
11.3.1.1 Critic and actor network learning model based on DDPG method
11.3.1.2 Optimal coordination between critic and actor network based on ADMM method
11.3.1.3 Offline training and online operation
11.3.2 Optimal power and energy distribution approach based on triple objectives
11.4 Simulation results
11.4.1 Simulation requirements and parameters
11.4.2 Results of hybrid TSKFS&amp
MADRL algorithm execution and control of voltage and frequency
11.4.3 Optimal distribution of power and energy in INMG and optimal results of primary, secondary, and tertiary
11.5 Conclusion
Appendix 1.