AI Based Design Methodology for Power Converters
Published on 20. August 2022
Book
Hardback
978-981-19-1403-4 (ISBN)
Description
This book presents new techniques for engineers (and researchers) to design power converters using artificial intelligent (AI)-based methods. The book first reviews existing AI technologies in power converters followed by an introduction to the proposed special AI algorithms for power converters considering their unique features. Based on the proposed AI-based design methods, the book discusses suitable applications in the design of power converters such as, power devices, DC/DC converters, resonant DC/DC converter, bidirectional DC/DC converter, DC/AC inverter, etc. This book is a useful guide for both the academic and professional audiences interested in the area of power converters.
More details
Series
Edition
1st ed. 2022
Language
English
Place of publication
Singapore
Singapore
Target group
Professional and scholarly
Illustrations
50 farbige Abbildungen
50 Illustrations, color; Approx. 180 p. 50 illus. in color.
Dimensions
Height: 235 mm
Width: 155 mm
ISBN-13
978-981-19-1403-4 (9789811914034)
Copyright in bibliographic data is held by Nielsen Book Services Limited or its licensors: all rights reserved.
Schweitzer Classification
Persons
Xin Zhang received his Ph.D. in Automatic Control and Systems Engineering from the University of Sheffield, UK, in 2016 and the Ph.D. degree in Electronic and Electrical Engineering from Nanjing University of Aeronautics & Astronautics, China, in 2014. Currently, he is a Professor at the College of Electrical Engineering, Zhejiang University, China. From 2017 to 2020, he was an Assistant Professor at the School of Electrical and Electronic Engineering of Nanyang Technological University. He is the Senior Member of IEEE society. He services as the Associate Editor of IEEE Transactions on Industrial Electronics, IEEE Journal of Emerging and Selected Topics in Power Electronics, IET Power Electronics, IEEE open journal of power electronics, IEEE Access, Journal of Artificial Intelligent. He is also the TPC member in IEEE IA/PELS Singapore joint Chapter. His research interests are in the area of power electronics and advanced control theory, green building, together with their applications in various sectors.
Xinze Li received his bachelor's degree in power engineering from Shandong University, China, in 2018, and is currently working towards his Ph.D. in Electrical and Electronic Engineering from Nanyang Technological University, Singapore. He is fully committed to applying advanced AI algorithms in power electronics. His current research interests mainly include the applications of AI in power electronics, power converter design, control and modulation design, artificial neural network, computational intelligence, and machine learning algorithms.
Hao Ma received his Ph.D. in Electrical Engineering from Zhejiang University, Hangzhou, China, in 1997. Currently, he is a Professor in the College of Electrical Engineering and serves as the Vice Dean of ZJU-UIUC Institute, Zhejiang University. His research interests include advanced control in power electronics, wireless power transfer, fault diagnosis of power electronic circuits and systems, and application of power electronics.
Bin Zhao received his Ph.D. from the Institute of Electronics, Chinese Academy of Sciences, Beijing, China, in 2017. From September 2015 to December 2016, he was Visiting Scholar with the National University of Ireland, Galway, Ireland. From January 2017 to January 2018, he was a Postdoc Researcher with the Department of Electrical Engineering, Technical University of Denmark. From February 2018 to January 2019, he was a Research Fellow with Energy Research Institute at NTU, Nanyang Technological University. Since January 2019, he has been with the Space Travelling-wave Tube Research & Development Center, Institute of Electronics, Chinese Academy of Sciences, as a Professor. His current research interests include high-frequency magnetic simulation, design and integration in power electronics and resonant converters.
Zheng Zeng received his Ph.D. in electrical engineering from Zhejiang University, Hangzhou, China, in 2014. He joined the School of Electrical Engineering, Chongqing University, Chongqing, China, in July 2014, where he was promoted to Associate Professor in August 2017. From July 2018 to July 2019, he was a Research Fellow with the School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore. His research interests include grid-connected inverter for renewable energy integration and customized package for wideband gap power devices.
Xinze Li received his bachelor's degree in power engineering from Shandong University, China, in 2018, and is currently working towards his Ph.D. in Electrical and Electronic Engineering from Nanyang Technological University, Singapore. He is fully committed to applying advanced AI algorithms in power electronics. His current research interests mainly include the applications of AI in power electronics, power converter design, control and modulation design, artificial neural network, computational intelligence, and machine learning algorithms.
Hao Ma received his Ph.D. in Electrical Engineering from Zhejiang University, Hangzhou, China, in 1997. Currently, he is a Professor in the College of Electrical Engineering and serves as the Vice Dean of ZJU-UIUC Institute, Zhejiang University. His research interests include advanced control in power electronics, wireless power transfer, fault diagnosis of power electronic circuits and systems, and application of power electronics.
Bin Zhao received his Ph.D. from the Institute of Electronics, Chinese Academy of Sciences, Beijing, China, in 2017. From September 2015 to December 2016, he was Visiting Scholar with the National University of Ireland, Galway, Ireland. From January 2017 to January 2018, he was a Postdoc Researcher with the Department of Electrical Engineering, Technical University of Denmark. From February 2018 to January 2019, he was a Research Fellow with Energy Research Institute at NTU, Nanyang Technological University. Since January 2019, he has been with the Space Travelling-wave Tube Research & Development Center, Institute of Electronics, Chinese Academy of Sciences, as a Professor. His current research interests include high-frequency magnetic simulation, design and integration in power electronics and resonant converters.
Zheng Zeng received his Ph.D. in electrical engineering from Zhejiang University, Hangzhou, China, in 2014. He joined the School of Electrical Engineering, Chongqing University, Chongqing, China, in July 2014, where he was promoted to Associate Professor in August 2017. From July 2018 to July 2019, he was a Research Fellow with the School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore. His research interests include grid-connected inverter for renewable energy integration and customized package for wideband gap power devices.
Content
1. Introduction1.1 Overiew of power converters
1.1.1 Power converters in power electronics systems: converter, rectifier and inverter
1.1.2 Applications of artificial intelligent (AI) algorithms in power converters
1.2 Evolutionary algorithms in power converters
1.2.1 Parameter design of power converters
1.2.2 Controller design of power converters
1.2.3 Offline parameter estimation of power converters
1.3 Fuzzy logic in power converters
1.3.1 Fuzzy-based controller design
1.3.2 Fuzzy-based modulation schematics
1.3.3 Fuzzy-based fault diagnosis
1.4 Neural networks (NN) in power converters
1.4.1 NN-based controller design
1.4.2 NN-based modulation schematics
1.4.3 Online parameter estimation
1.4.4 NN-based fault diagnosis
1.5 Conclusions
2 The Proposed Adaptive Velocity Boundary Particle Swarm Optimization Algorithm with the Moving Average of Evolutionary State Estimation for Power Converter Design
2.1 Introduction
2.2 Review for particle swarm algorithms
2.2.1 PSO Variants with Hyper-Parameter Adjusting Methods
2.2.2 PSO Variants with Different Learning Strategies
2.2.3 PSO Variants with Auxiliary Strategies or Tools
2.3 Main strategies in the proposed MA-ESE based AVB-PSO
2.3.1 The proposed moving average of evolutionary state estimation
2.3.2 The proposed batch-wise velocity boundary update strategy
2.3.3 The proposed dimension-wise velocity boundary update strategy
2.3.4 The proposed adaptive boundary for the moving average of evolutionary state estimation
2.3.5 Collision and boundary restriction strategies
2.4 Framework of the proposed AVB-PSO with MA-ESE
2.4.1 Main process of the proposed algorithm
2.4.2 Charateristics of the proposed algorithm
2.5 Experimental results for performance verification
2.5.1 Test functions
2.5.2 Experiemental setups
2.5.3 Results analysis
2.6 Hyper-parameters and strategies analysis
2.6.1 Effects of the length of collision radius
2.6.2 Effects of the length of moving average window
2.6.3 Effects of the moving average strategies
2.6.4 Effects of the adaptive boundary strategy for the moving average of evolutionary state estimation
2.6.5 Effects of the silence interval and the velocity boundary partitions
2.6.6 Effects of the position and velocity restriction strategies
2.7 Conclusions
3 The Proposed AI-based Design for the Power modules: A Thermo-Mechanical-Coordinated and Multi-Objective-Oriented Optimization Methodology
3.1 Introduction
3.2 Present status and under-optimization problem of DSC power module
3.2.1 State-of-the-art of DSC power module
3.2.2 Under-optimization problem of DSC power module
3.3 Modeling of the proposed thermo-mechanical multi-objective co-design of DSC power module
3.3.1 Thermal resistance modeling of the DSC power module
3.3.2 Mechanical stress modeling of the DSC power module
3.3.3 The propsoed multi-objective optimization model
3.4 Solution of Proposed Thermo-Mechanical-Coordinated Multi-Objective Optimization Design
3.4.1 Case study of multi-objective optimization design
3.4.2 Influence of material properties on optimization
3.5 Experimental results
3.6 Conclusions
4 The Proposed AI-based Multi-objective Optimal Design for the Traditional DC/DC converter with Improved AMOSA Algorithm
4.1 Nomenclature in the proposed coevolving-AMOSA algorithm
4.2 Introduction
4.3 Concept of the proposed multi-objective design method of the LC output filter for DC-DC Buck converter
4.4 Stage 1: three optimal objectives for the proposed multi-objective LC output filter design of Buck converter: efficiency, power-density and cut-off frequency
4.4.1 Objective-I: optimal power loss for high efficiency Buck converter
4.4.2 Objective-II: optimal volume for high-power-density filter
4.4.3 Objective-III: optimal cut-off frequency for high-filtering-performance filter
4.5 Stage 2: the proposed improved AMOSA algorithm for multi-objective design: coevolving archived multi-objective simulated annealing
4.5.1 Challenges and requirements of the multi-objective output LC filter design methodology for Buck converter
4.5.2 The main process of the proposed improved AMOSA: coevolving-AMOSA
4.5.3 Comparisons with the existing popular multi-objective evolutionary algorithms
4.6 Stage 3: optimal design cases with the proposed coevolving-AMOSA algorithm: design examples
4.6.1 Design specifications of the design example
4.6.2 Design examples
4.6.3 Comparisons with the existing popular multi-objective evolutionary algorithms
4.7 Experimental verification
4.7.1 Experimental results of the designed buck converters under different cases
4.7.2 Evaluation of the four design cases
4.8 Conclusions
5 The Proposed AI-based Efficiency-oriented Two-stage Optimal Design Methodology for the Resonant DC/DC Converters
5.1 Introduction
5.2 Preliminary of the proposed efficiency-oriented two-stage optimal design method: review of the LCLC resonant converter and the calculation of the total power loss
5.2.1 Review of the calculations of the main parameters in the LCLC resonant converter
5.2.2 Power loss analysis of the LCLC resonant converter in the space TWTA applications
5.3 The proposed efficiency-oriented two-stage optimal design method of the LCLC resonant converter in the space TWTA applications
5.3.1 The proposed efficiency-oriented two-stage optimal design methodology
5.3.2 Stage-I: extraction of the optimal parameters based on the proposed GA+PSO algorithm
5.3.3 Stage-II: realization of the optimal parameters based on the proposed single-layer partially-interleaved transformer structure
5.3.4 Modified inductor current references incorporated with adaptive terms of the three-phase stand-alone inverter
5.4 Experiemental validations
5.4.1 Verifications of the ZVS and ZCS characteristics of the optimal LCLC resonant converter
5.4.2 Verifications of the proposed efficiency-oriented two-stage optimal design method of the LCLC resonant converter
5.5 Conclusions
6 The Proposed AI-based Two-stage Optimal Design Methodology for the High-efficiency Bidirectional DC/DC Converters
6.1 Introduction
6.2 Working principles and the circuit analysis of the CLLC bidirectional converter in hybrid AC/DC microgrid applications
6.2.1 Working principles of the CLLC bidirectional converter as a DC transformer
6.2.2 The circuit analysis of the CLLC bidirectional converter
6.2.3 Conditions of ZVS and ZCS
6.3 The proposed AI based high efficiency oriented two-stage optimal design method for CLLC bidirectional power converters in the hybrid AC/DC microgrid
6.3.1 Preliminary of the proposed AI based two-stage optimal design method: total power loss equation of the CLLC bidirectional converter
6.3.2 The proposed AI based (GA+PSO) two-stage optimal design methodology for the CLLC bidirectional converter
6.4 Simulation and experimental validations of the proposed AI based two-stage optimal design method
6.4.1 Design of a planar transformer with the desired parasitic parameters for a CLLC bidirectional converter
6.4.2 Experimental validation of the proposed AI based two-stage optimal design methodology
6.5 Conclusions
7 The Proposed AI-based Optimal Design and Heterogeneous Integration of the DC/AC Inverter
7.1 Introduction
7.2 State-of-the-art and main barriers of air-cooling SiC inverter
7.2.1 State-of-the-art of EV inverter
7.2.2 Challenges of light and compact EV inverter
7.3 The proposed design methodology for air-cooling inverter
7.4 The proposed optimal design for key components of air-cooling SiC inverter
7.4.1 The proposed multi-physics-based design of SiC power module
7.4.2 The proposed optimal selection of DC-link capacitance
7.4.3 The proposed thermal design of heat sink
7.5 Experimental results
7.5.1 Prototypes of power module and air-cooling inverter
7.5.2 Experimental results of SiC power module
7.5.3 Experimental results of SiC inverter
7.6 Conclusions
8 The Proposed AI-based Space Vector Modulation Control Strategy for AC/DC Rectifiers via Neural-Network Approach
8.1 Introduction
8.2 Space-vector PWM in under-modulation and over-modulation regions for three-phase rectifier
8.2.1 Under-modulation (0 < m < 0.907)
8.2.2 Over-modulation mode-1 (0.907 < m < 0.952)
8.2.3 Over-modulation mode-2 (0.952 < m < 1)
8.3 The proposed neural-network-based space-vector PWM for three-phase rectifier
8.3.1 Under-modulation region
8.3.2 Over-modulation region
8.3.3 The proposed angle neural network
8.3.4 The proposed amplitude neural network
8.4 Performance evaluation
8.5 Conclusions
1.1.1 Power converters in power electronics systems: converter, rectifier and inverter
1.1.2 Applications of artificial intelligent (AI) algorithms in power converters
1.2 Evolutionary algorithms in power converters
1.2.1 Parameter design of power converters
1.2.2 Controller design of power converters
1.2.3 Offline parameter estimation of power converters
1.3 Fuzzy logic in power converters
1.3.1 Fuzzy-based controller design
1.3.2 Fuzzy-based modulation schematics
1.3.3 Fuzzy-based fault diagnosis
1.4 Neural networks (NN) in power converters
1.4.1 NN-based controller design
1.4.2 NN-based modulation schematics
1.4.3 Online parameter estimation
1.4.4 NN-based fault diagnosis
1.5 Conclusions
2 The Proposed Adaptive Velocity Boundary Particle Swarm Optimization Algorithm with the Moving Average of Evolutionary State Estimation for Power Converter Design
2.1 Introduction
2.2 Review for particle swarm algorithms
2.2.1 PSO Variants with Hyper-Parameter Adjusting Methods
2.2.2 PSO Variants with Different Learning Strategies
2.2.3 PSO Variants with Auxiliary Strategies or Tools
2.3 Main strategies in the proposed MA-ESE based AVB-PSO
2.3.1 The proposed moving average of evolutionary state estimation
2.3.2 The proposed batch-wise velocity boundary update strategy
2.3.3 The proposed dimension-wise velocity boundary update strategy
2.3.4 The proposed adaptive boundary for the moving average of evolutionary state estimation
2.3.5 Collision and boundary restriction strategies
2.4 Framework of the proposed AVB-PSO with MA-ESE
2.4.1 Main process of the proposed algorithm
2.4.2 Charateristics of the proposed algorithm
2.5 Experimental results for performance verification
2.5.1 Test functions
2.5.2 Experiemental setups
2.5.3 Results analysis
2.6 Hyper-parameters and strategies analysis
2.6.1 Effects of the length of collision radius
2.6.2 Effects of the length of moving average window
2.6.3 Effects of the moving average strategies
2.6.4 Effects of the adaptive boundary strategy for the moving average of evolutionary state estimation
2.6.5 Effects of the silence interval and the velocity boundary partitions
2.6.6 Effects of the position and velocity restriction strategies
2.7 Conclusions
3 The Proposed AI-based Design for the Power modules: A Thermo-Mechanical-Coordinated and Multi-Objective-Oriented Optimization Methodology
3.1 Introduction
3.2 Present status and under-optimization problem of DSC power module
3.2.1 State-of-the-art of DSC power module
3.2.2 Under-optimization problem of DSC power module
3.3 Modeling of the proposed thermo-mechanical multi-objective co-design of DSC power module
3.3.1 Thermal resistance modeling of the DSC power module
3.3.2 Mechanical stress modeling of the DSC power module
3.3.3 The propsoed multi-objective optimization model
3.4 Solution of Proposed Thermo-Mechanical-Coordinated Multi-Objective Optimization Design
3.4.1 Case study of multi-objective optimization design
3.4.2 Influence of material properties on optimization
3.5 Experimental results
3.6 Conclusions
4 The Proposed AI-based Multi-objective Optimal Design for the Traditional DC/DC converter with Improved AMOSA Algorithm
4.1 Nomenclature in the proposed coevolving-AMOSA algorithm
4.2 Introduction
4.3 Concept of the proposed multi-objective design method of the LC output filter for DC-DC Buck converter
4.4 Stage 1: three optimal objectives for the proposed multi-objective LC output filter design of Buck converter: efficiency, power-density and cut-off frequency
4.4.1 Objective-I: optimal power loss for high efficiency Buck converter
4.4.2 Objective-II: optimal volume for high-power-density filter
4.4.3 Objective-III: optimal cut-off frequency for high-filtering-performance filter
4.5 Stage 2: the proposed improved AMOSA algorithm for multi-objective design: coevolving archived multi-objective simulated annealing
4.5.1 Challenges and requirements of the multi-objective output LC filter design methodology for Buck converter
4.5.2 The main process of the proposed improved AMOSA: coevolving-AMOSA
4.5.3 Comparisons with the existing popular multi-objective evolutionary algorithms
4.6 Stage 3: optimal design cases with the proposed coevolving-AMOSA algorithm: design examples
4.6.1 Design specifications of the design example
4.6.2 Design examples
4.6.3 Comparisons with the existing popular multi-objective evolutionary algorithms
4.7 Experimental verification
4.7.1 Experimental results of the designed buck converters under different cases
4.7.2 Evaluation of the four design cases
4.8 Conclusions
5 The Proposed AI-based Efficiency-oriented Two-stage Optimal Design Methodology for the Resonant DC/DC Converters
5.1 Introduction
5.2 Preliminary of the proposed efficiency-oriented two-stage optimal design method: review of the LCLC resonant converter and the calculation of the total power loss
5.2.1 Review of the calculations of the main parameters in the LCLC resonant converter
5.2.2 Power loss analysis of the LCLC resonant converter in the space TWTA applications
5.3 The proposed efficiency-oriented two-stage optimal design method of the LCLC resonant converter in the space TWTA applications
5.3.1 The proposed efficiency-oriented two-stage optimal design methodology
5.3.2 Stage-I: extraction of the optimal parameters based on the proposed GA+PSO algorithm
5.3.3 Stage-II: realization of the optimal parameters based on the proposed single-layer partially-interleaved transformer structure
5.3.4 Modified inductor current references incorporated with adaptive terms of the three-phase stand-alone inverter
5.4 Experiemental validations
5.4.1 Verifications of the ZVS and ZCS characteristics of the optimal LCLC resonant converter
5.4.2 Verifications of the proposed efficiency-oriented two-stage optimal design method of the LCLC resonant converter
5.5 Conclusions
6 The Proposed AI-based Two-stage Optimal Design Methodology for the High-efficiency Bidirectional DC/DC Converters
6.1 Introduction
6.2 Working principles and the circuit analysis of the CLLC bidirectional converter in hybrid AC/DC microgrid applications
6.2.1 Working principles of the CLLC bidirectional converter as a DC transformer
6.2.2 The circuit analysis of the CLLC bidirectional converter
6.2.3 Conditions of ZVS and ZCS
6.3 The proposed AI based high efficiency oriented two-stage optimal design method for CLLC bidirectional power converters in the hybrid AC/DC microgrid
6.3.1 Preliminary of the proposed AI based two-stage optimal design method: total power loss equation of the CLLC bidirectional converter
6.3.2 The proposed AI based (GA+PSO) two-stage optimal design methodology for the CLLC bidirectional converter
6.4 Simulation and experimental validations of the proposed AI based two-stage optimal design method
6.4.1 Design of a planar transformer with the desired parasitic parameters for a CLLC bidirectional converter
6.4.2 Experimental validation of the proposed AI based two-stage optimal design methodology
6.5 Conclusions
7 The Proposed AI-based Optimal Design and Heterogeneous Integration of the DC/AC Inverter
7.1 Introduction
7.2 State-of-the-art and main barriers of air-cooling SiC inverter
7.2.1 State-of-the-art of EV inverter
7.2.2 Challenges of light and compact EV inverter
7.3 The proposed design methodology for air-cooling inverter
7.4 The proposed optimal design for key components of air-cooling SiC inverter
7.4.1 The proposed multi-physics-based design of SiC power module
7.4.2 The proposed optimal selection of DC-link capacitance
7.4.3 The proposed thermal design of heat sink
7.5 Experimental results
7.5.1 Prototypes of power module and air-cooling inverter
7.5.2 Experimental results of SiC power module
7.5.3 Experimental results of SiC inverter
7.6 Conclusions
8 The Proposed AI-based Space Vector Modulation Control Strategy for AC/DC Rectifiers via Neural-Network Approach
8.1 Introduction
8.2 Space-vector PWM in under-modulation and over-modulation regions for three-phase rectifier
8.2.1 Under-modulation (0 < m < 0.907)
8.2.2 Over-modulation mode-1 (0.907 < m < 0.952)
8.2.3 Over-modulation mode-2 (0.952 < m < 1)
8.3 The proposed neural-network-based space-vector PWM for three-phase rectifier
8.3.1 Under-modulation region
8.3.2 Over-modulation region
8.3.3 The proposed angle neural network
8.3.4 The proposed amplitude neural network
8.4 Performance evaluation
8.5 Conclusions