Series-parallel conversion systems, in which multiple standardized converter modules are connected in series or parallel at the input and output sides, to meet the demands of various applications. This book focuses on the control strategies for the series-parallel conversion systems with DC-DC converters and DC-AC inverters as the basic modules, respectively, to achieve input voltage/current sharing and output voltage/current sharing among the constituent modules. The detailed theoretical analysis with design examples and experimental validations are presented. This book is essential and valuable reference for graduate students and academics majoring in power electronics and engineers engaged in developing DC-DC converters, DC-AC inverters and power electronics transformers.

Autor
Xinbo Ruan | Wu Chen | Tianzhi Fang | Kai Zhuang | Tao Zhang | Hong Yan | Mohammad Dulal Miah | Yasushi Suzuki | David Parsons | Kathryn MacCallum

108 downloads 4K Views 13MB Size

Empty story

CPSS Power Electronics Series

Xinbo Ruan · Wu Chen Tianzhi Fang · Kai Zhuang Tao Zhang · Hong Yan

Control of Series-Parallel Conversion Systems

CPSS Power Electronics Series Series editors Wei Chen, Fuzhou University, Fuzhou, Fujian, China Yongzheng Chen, Liaoning University of Technology, Jinzhou, Liaoning, China Xiangning He, Zhejiang University, Hangzhou, Zhejiang, China Yongdong Li, Tsinghua University, Beijing, China Jingjun Liu, Xi’an Jiaotong University, Xi’an, Shaanxi, China An Luo, Hunan University, Changsha, Hunan, China Xikui Ma, Xi’an Jiaotong University, Xi’an, Shaanxi, China Xinbo Ruan, Nanjing University of Aeronautics and Astronautics, Nanjing, Jiangsu, China Kuang Shen, Zhejiang University, Hangzhou, Zhejiang, China Dianguo Xu, Harbin Institute of Technology, Harbin, Heilongjiang, China Jianping Xu, Xinan Jiaotong University, Chengdu, Sichuan, China Mark Dehong Xu, Zhejiang University, Hangzhou, Zhejiang, China Xiaoming Zha, Wuhan University, Wuhan, Hubei, China Bo Zhang, South China University of Technology, Guangzhou, Guangdong, China Lei Zhang, China Power Supply Society, Tianjin, China Xin Zhang, Hefei University of Technology, Hefei, Anhui, China Zhengming Zhao, Tsinghua University, Beijing, China Qionglin Zheng, Beijing Jiaotong University, Beijing, China Luowei Zhou, Chongqing University, Chongqing, China

This series comprises advanced textbooks, research monographs, professional books, and reference works covering different aspects of power electronics, such as Variable Frequency Power Supply, DC Power Supply, Magnetic Technology, New Energy Power Conversion, Electromagnetic Compatibility as well as Wireless Power Transfer Technology and Equipment. The series features leading Chinese scholars and researchers and publishes authored books as well as edited compilations. It aims to provide critical reviews of important subjects in the ﬁeld, publish new discoveries and signiﬁcant progress that has been made in development of applications and the advancement of principles, theories and designs, and report cutting-edge research and relevant technologies. The CPSS Power Electronics series has an editorial board with members from the China Power Supply Society and a consulting editor from Springer. Readership: Research scientists in universities, research institutions and the industry, graduate students, and senior undergraduates.

More information about this series at http://www.springer.com/series/15422

Xinbo Ruan Wu Chen Tianzhi Fang Kai Zhuang Tao Zhang Hong Yan •

•

•

Control of Series-Parallel Conversion Systems

123

Xinbo Ruan College of Automation Engineering Nanjing University of Aeronautics and Astronautics Nanjing, Jiangsu, China Wu Chen School of Electrical Engineering Southeast University Nanjing, Jiangsu, China Tianzhi Fang College of Automation Engineering Nanjing University of Aeronautics and Astronautics Nanjing, Jiangsu, China

Kai Zhuang SDAAC Automotive Air-Conditioning System Co., Ltd. Shanghai, China Tao Zhang Flextronics R&D (ShenZhen) Co., Ltd. Shenzhen, Guangdong, China Hong Yan School of Mechanical and Electronic Engineering Bengbu University Bengbu, Anhui, China

ISSN 2520-8853 ISSN 2520-8861 (electronic) CPSS Power Electronics Series ISBN 978-981-13-2759-9 ISBN 978-981-13-2760-5 (eBook) https://doi.org/10.1007/978-981-13-2760-5 Jointly published with Science Press, Beijing, China The print edition is not for sale in China Mainland. Customers from China Mainland please order the print book from: Science Press. Library of Congress Control Number: 2018957246 © Springer Nature Singapore Pte Ltd. and Science Press, Beijing, China 2019 This work is subject to copyright. All rights are reserved by the Publishers, whether the whole or part of the material is concerned, speciﬁcally the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microﬁlms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a speciﬁc statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publishers, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publishers nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publishers remains neutral with regard to jurisdictional claims in published maps and institutional afﬁliations. This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd. The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721, Singapore

Preface

In the past three decades, power electronics technology has been rapidly developing and has found wide applications in almost all kinds of ﬁelds of industrial and agricultural production and daily life. Recently, the development and utilization of renewable energy for addressing the traditional fossil energy shortage and environmental pollution has been received more and more attention, which provides new opportunities for power electronics technology. Power electronics technology has been playing an increasingly important role in renewable energy generation, flexible AC/DC transmission, smart grid, and electric vehicle. At present, most of power electronics products are composed of various discrete devices. Inspired by the success of the development of microelectronics technology, the concept of power electronics system integration is emerging for the purpose of reducing the manufacture cost and improving the reliability. Generally, the power electronics system integration can be divided into three levels. The ﬁrst one is component-level integration including active devices integration and passive components integration. The second one is module-level integration, in which a variety of integrated active devices and/or passive components are integrated together to form highly versatile standardized converter modules. The last one is system-level integration, in which a number of converter modules are put together to form different kinds of power conversion systems. In the system-level integration, series–parallel power conversion system, in which multiple standardized converter modules are connected in series or parallel at the input and output sides, is one of the typical forms. Series–parallel power conversion systems have the following advantages, including reduced design difﬁculties, ease of capacity expansion, redundancy, and improved reliability. Series– parallel power conversion systems can be categorized into four possible architectures on the basis of the connection forms, namely, input-parallel-output-parallel (IPOP), input-parallel-output-series (IPOS), input-series-output-parallel (ISOP), and input-series-output-series (ISOS), and each type of series–parallel power conversion system has its own speciﬁc application areas. The IPOP system is suitable for the applications with high output current, the IPOS is suitable for the applications with low input voltage and high output voltage, the ISOP system is suitable for the v

vi

Preface

applications with high input voltage and high output current, and the ISOS system is attractive for the applications with high input voltage and high output voltage. This book focuses on the control techniques for the series–parallel power conversion systems with DC–DC converters and DC–AC inverters as the basic modules, respectively, to achieve equilibrium among the constituent modules. The detailed theoretical analysis with design examples and experimental validations are presented. This book includes nine chapters. In Chap. 1, the classiﬁcation and characteristics of the series–parallel power conversion systems are presented, and the control strategies for the series–parallel power conversion systems with DC–DC converters and DC–AC inverters as the basic modules, respectively, are briefly reviewed. Chapters 2–4 focus on the control strategies for the DC–DC series–parallel power conversion systems. In Chap. 2, the inherent relationship between the sharing of input voltages/ currents and the sharing of output voltages/currents among the constituent modules in the four types of DC–DC series–parallel power conversion systems is analyzed systematically from the view of power conservation, and the stability of various possible control strategies is discussed. Based on the analysis, a general control strategy is proposed for the DC–DC series–parallel power conversion systems, which not only achieve equilibrium among the constituent modules, but also decouples the output voltage control loop and the sharing control loop. In Chap. 3, the small-signal model of ISOP system consisting of multiple phase-shifted full-bridge converters is derived to prove the decoupling of the system output voltage control loop and input voltage sharing (IVS) control loop, and the parameters design of the two control loops is given. Furthermore, in order to improve the system dynamic response and realize output current short-circuit protection, the inner current loop is incorporated into the general control strategy, and the parameters design of the IVS control loop is analyzed. In Chap. 4, the wireless IVS control strategies for the input-series-connected power conversion systems based on positive output voltage gradient method are presented, which increases the output voltage of each module with the increase of module input voltage. With this control strategy, there is no control interconnection among the constituent modules leading to a truly modular design and high system reliability. Chapters 5–8 are dedicated to the control strategies for the DC–AC series– parallel power conversion systems. In Chap. 5, the inherent relationship between the input voltage/current sharing (IVS/ICS) and output voltage/current sharing (OVS/OCS) of the four kinds of DC– AC series–parallel power conversion systems is analyzed systematically, and the control strategies to achieve IVS/ICS and OVS/OCS and the system stability are discussed. A general control strategy for the four kinds of DC–AC series–parallel power conversion system is proposed.

Preface

vii

Chapters 6 and 7 present the implementation of the compound balanced control strategies for the DC–AC ISOP system and DC–AC ISOS system, respectively, which control the module input voltages to be equal for achieving IVS and control the magnitudes or phases of modular output currents (for ISOP system) or output voltages (for ISOS system) to be identical for achieving OCS/OVS. In Chap. 8, the load current feed-forward scheme is incorporated for the inverter module in the DC–AC IPOP system, and the load current feed-forward node is intentionally located before the current-limiting unit, for not only making the output voltage of the inverter independent of load current but also limiting the load current. In Chap. 9, the control strategy with changeable weighted values of the IVS loops and the IVS auxiliary circuit are presented for the input-series-connected power conversion system for achieving IVS under extreme load conditions, including light-load, no-load, and short-circuit current-limiting. This book is essential and valuable reference for graduate students and academics majoring in power electronics and engineers engaged in developing DC–DC converters, DC–AC inverters, series–parallel power conversion systems and power electronics transformers. Senior undergraduate students majoring in electrical engineering and automation engineering would also ﬁnd this book useful. Nanjing, China Nanjing, China Nanjing, China Shanghai, China Zhuhai, China Bengbu, China

Xinbo Ruan Wu Chen Tianzhi Fang Kai Zhuang Tao Zhang Hong Yan

Acknowledgements

This research monograph systematically summarizes the research work on the control techniques for series–parallel power conversion systems since 2003. The work in this book was partially supported by the Delta Power Electronics Science and Education Development Foundation under Award DREO2006006. We would like to express our sincere thanks to the Executive Committee of Delta Power Electronics Science and Education Development Foundation. It has been a great pleasure to work with the colleagues of Springer, Science Press, China, and China Power Supply Society (CPSS). The support and help from Mr. Wayne Hu (the Project Editor) are greatly appreciated. June 2018

ix

Contents

. . . . . . . . . . . . . . . . . . .

1 1 4 5 5 5 7 7 9 11 13 15 15 16 20 21 22 25 26

........

29

. . . . .

30 32 32 32 33

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1 Development Trend of Power Electronics . . . . . . . . . . . . . . . 1.2 Power Electronics System Integration . . . . . . . . . . . . . . . . . . 1.2.1 Component-Level Integration . . . . . . . . . . . . . . . . . . . 1.2.2 Module-Level Integration . . . . . . . . . . . . . . . . . . . . . . 1.2.3 System-Level Integration . . . . . . . . . . . . . . . . . . . . . . 1.3 DC–DC Series–Parallel Power Conversion Systems . . . . . . . . 1.3.1 IPOP System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3.2 IPOS System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3.3 ISOP System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3.4 ISOS System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4 DC–AC Inverter Series–Parallel Power Conversion Systems . . . . 1.4.1 DC–AC Inverter . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4.2 IPOP DC–AC Inverter System . . . . . . . . . . . . . . . . . . 1.4.3 IPOS DC–AC Inverter System . . . . . . . . . . . . . . . . . . 1.4.4 ISOP DC–AC Inverter System . . . . . . . . . . . . . . . . . . 1.4.5 ISOS DC–AC Inverter System . . . . . . . . . . . . . . . . . . 1.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 A General Control Strategy for DC–DC Series–Parallel Power Conversion Systems . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Possible Control Strategies for DC–DC Series–Parallel Power Conversion Systems . . . . . . . . . . . . . . . . . . . . . 2.1.1 IPOP Systems . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.2 IPOS Systems . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.3 ISOP Systems . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.4 ISOS Systems . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . . . . . . . . . . . . . . . .

. . . . .

. . . . . . . . . . . . . . . . . . .

. . . . .

. . . . .

xi

xii

Contents

2.2 Stability of the Control Strategies . . . . . . . . . . . . . . . . . . . . . . 2.2.1 Controlling at Input Side . . . . . . . . . . . . . . . . . . . . . . . 2.2.2 Controlling at Output Side . . . . . . . . . . . . . . . . . . . . . . 2.3 General Control Strategy for DC–DC Series–Parallel Power Conversion Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4 Modularization Architecture for DC–DC Series–Parallel Power Conversion Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5 Experimental Veriﬁcation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Mathematical Model and Closed-Loop Parameters Design for DC–DC ISOP System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 Small-Signal Model of PSFB-Based ISOP System . . . . . . . . . 3.1.1 Small-Signal Model of PSFB Converter . . . . . . . . . . . 3.1.2 Small-Signal Model of the PSFB-Based ISOP System . 3.2 Decoupling the Control Loops of the PSFB-Based ISOP System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3 Closed-Loop Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1 Output Voltage Regulator . . . . . . . . . . . . . . . . . . . . . . 3.3.2 IVS Regulator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4 Experimental Veriﬁcation of the General Control Strategy . . . 3.5 Three-Loop Control Strategy for ISOP System . . . . . . . . . . . . 3.5.1 Positive Resistance . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5.2 IVS Regulator Design . . . . . . . . . . . . . . . . . . . . . . . . 3.6 Experimental Veriﬁcation of the Three-Loop Control Strategy 3.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

.. .. ..

34 35 36

..

40

. . . .

. . . .

43 46 51 53

. . . .

. . . .

. . . .

55 56 56 61

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

67 69 70 72 74 74 76 80 82 83 84

..

85

..

86

4 Wireless IVS Control Strategies for Input-Series-Connected Systems Based on Positive Output Voltage Gradient Method . . . . 4.1 Genesis of the Wireless IVS Control Strategies Based on Positive Output Voltage Gradient Method . . . . . . . . . . . . . . . . . . . . . . . 4.2 Basic Features of the Input-Series-Connected Systems with Wireless IVS Control Strategies . . . . . . . . . . . . . . . . . . . . . . . 4.2.1 ISOP System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.2 ISOS System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3 System Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4 Experimental Veriﬁcation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4.1 Experimental Results of ISOP System . . . . . . . . . . . . . . 4.4.2 Experimental Results of ISOS System . . . . . . . . . . . . . . 4.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . .

. 90 . 90 . 92 . 94 . 97 . 97 . 102 . 105 . 105

Contents

xiii

5 A General Control Strategy for DC–AC Series–Parallel Power Conversion Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1 Relationship Between Input Voltage/Current Sharing and Output Voltage/Current Sharing . . . . . . . . . . . . . . . . . . . 5.1.1 IPOP Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1.2 IPOS Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1.3 ISOP Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1.4 ISOS Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2 Stability of the Control Strategies for DC–AC Series–Parallel Power Conversion Systems . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2.1 Compound Balanced Control Strategy . . . . . . . . . . . . . 5.2.2 Output Voltage/Current Sharing (OVS/OCS) Control Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3 General Voltage/Current Sharing Control Strategy for DC–AC Series–Parallel Power Conversion Systems . . . . . . . . . . . . . . . 5.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . 107 . . . . .

. . . . .

. . . . .

108 110 111 111 112

. . . 113 . . . 114 . . . 115 . . . 115 . . . 118 . . . 119

6 Compound Balanced Control Strategy for DC–AC Input-Series Output-Parallel Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1 DC–AC Inverter Topology for the ISOP Systems . . . . . . . . . . . 6.2 Implementation of the Compound Balanced Control Strategy . . 6.3 Closed-Loops Design for DC–AC ISOP Systems . . . . . . . . . . . 6.3.1 Decoupling Relationship Between IVS Loop and Output Voltage Loop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3.2 Design of Control Loop of the First-Stage Full-Bridge Converter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3.3 Design of IVS Loop . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3.4 Design of Output Voltage Loop . . . . . . . . . . . . . . . . . . 6.4 Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.5 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.5.1 Steady-State Experiment . . . . . . . . . . . . . . . . . . . . . . . 6.5.2 Dynamic Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . 6.5.3 Comparison of the Output Current Feedback and Output Filter Inductor Current Feedback . . . . . . . . . . . . . . . . . 6.6 Distributed IVS/OCS Control Strategy . . . . . . . . . . . . . . . . . . . 6.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . .

. . . . . . .

130 133 137 138 141 141 141

. . . .

. . . .

145 148 151 151

7 Compound Balanced Control Strategy for DC–AC Input-Series-Output-Series Systems . . . . . . . . . . . . . . . . . . . . . 7.1 Implementation of the Compound Balanced Control Strategy 7.2 Decoupling the Control Loops . . . . . . . . . . . . . . . . . . . . . . . 7.3 Design of the Control Loops for the DC–AC ISOS Systems .

. . . .

. . . .

153 154 158 161

. . . .

. . . .

. . . .

. . . .

121 122 124 127

. . 127

xiv

Contents

7.3.1 Design of the Output Voltage Loop . . . . . . . . . . . . . . . 7.3.2 Design of the Control Loop of the First-Stage Converter and IVS Loop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.4 Simulation and Experimental Veriﬁcation . . . . . . . . . . . . . . . . 7.4.1 Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.4.2 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . 7.5 Distributed Control Strategy for DC–AC ISOS Systems . . . . . . 7.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 An Improved Average Current Control Strategy for DC–AC Input-Parallel-Output-Parallel Systems . . . . . . . . . . . . . . . . . . . . 8.1 Current Three-State Hysteresis Modulation . . . . . . . . . . . . . . . 8.2 Improvement of the Output Characteristics of Inverter . . . . . . 8.3 Conﬁguration and Control Scheme for DC–AC IPOP Systems . . 8.3.1 Conﬁguration of the DC–AC IPOP Systems . . . . . . . . 8.3.2 Control Scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.4 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . .

9 Input Voltage Sharing Control Strategy for ISOP Systems Under Extreme Load Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.1 Reason of Input Voltages Imbalance of ISOP Systems Under Extreme Load Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2 Input Voltage Sharing Control Strategy for ISOP Systems Under Extreme Load Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2.1 Changing the Weighted Values of the IVS Loops . . . . . 9.2.2 Input Voltage Sharing Auxiliary Circuit . . . . . . . . . . . . 9.3 Simulation and Experimental Veriﬁcation . . . . . . . . . . . . . . . . 9.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . 161 . . . . . . .

. . . . . . .

163 163 164 167 167 172 172

. . . . . . . . .

. . . . . . . . .

175 176 178 180 180 183 183 187 188

. . 191 . . 192 . . . . . .

. . . . . .

194 194 197 198 208 208

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209

About the Authors

Xinbo Ruan was born in Hubei Province, China, in 1970. He received B.S. and Ph.D. in electrical engineering from Nanjing University of Aeronautics and Astronautics (NUAA), Nanjing, China, in 1991 and 1996, respectively. In 1996, he joined the Faculty of Electrical Engineering Teaching and Research Division, NUAA, where he became a Professor in the College of Automation Engineering in 2002 and has been engaged in teaching and research in the ﬁeld of power electronics. From August to October 2007, he was a Research Fellow in the Department of Electronic and Information Engineering, Hong Kong Polytechnic University, Hong Kong, China. From March 2008 to August 2011, he was also with the School of Electrical and Electronic Engineering, Huazhong University of Science and Technology, China. He is a Guest Professor with Beijing Jiaotong University, Beijing, China, Hefei University of Technology, Hefei, China, and Wuhan University, Wuhan, China. He is the author or co-author of nine books and more than 300 technical papers published in journals and conferences. His main research interests include soft-switching DC–DC converters, soft-switching inverters, power factor correction converters, modeling the converters, power electronics system integration, and renewable energy generation system. Dr. Ruan was a recipient of the Delta Scholarship by the Delta Environment and Education Fund in 2003 and was a recipient of the Special Appointed Professor of the Chang Jiang Scholars Program by the Ministry of Education, China, in 2007. From 2005 to 2013, again from 2017, he served as Vice President of the China Power Supply Society. From 2014 to 2016, he served as Vice Chair of the Technical Committee on Renewable Energy Systems within the IEEE Industrial Electronics Society. Currently, he is an Associate Editor for the IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, IEEE TRANSACTIONS ON POWER ELECTRONICS, IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS-II, and the IEEE JOURNAL OF EMERGING AND SELECTED TOPICS ON POWER ELECTRONICS. He was elevated to IEEE Fellow in 2015. Wu Chen was born in Jiangsu Province, China, in 1981. He received B.S., M.S., and Ph.D. in electrical engineering from Nanjing University of Aeronautics and Astronautics (NUAA), Nanjing, China, in 2003, 2006, and 2009, respectively. xv

xvi

About the Authors

From 2009 to 2010, he was a Senior Research Assistant in the Department of Electronic Engineering, City University of Hong Kong, Kowloon, Hong Kong, China. From 2010 to 2011, he was a Postdoctoral Researcher in Future Electric Energy Delivery and Management Systems Center, North Carolina State University, Raleigh. Since September 2011, he has been an Associate Research Fellow with the School of Electrical Engineering, Southeast University, Nanjing, China, where he has been a Professor since 2016. His main research interests include soft-switching converters, power delivery, and power electronic system integration. Tianzhi Fang was born in Jiangsu Province, China, in 1977. He received B.S. in electrical engineering and automation from Nanjing University of Aeronautics and Astronautics (NUAA), Nanjing, China, in 2000, M.S. in control theory and engineering from Jiangsu University, Zhenjiang, China, in 2003, and Ph.D. in electrical engineering from NUAA, Nanjing, China, in 2009. In 2009, he joined the Faculty of the College of Automation Engineering, NUAA, where he is currently an Associate Professor. His main research interests include series–parallel inverters, grid-connected inverter, and power electronics system integration. Kai Zhuang was born in Shandong Province, China, in 1975. He received B.S in electrical engineering from Jinan University, Jinan, China, in 1998, M.S. in electrical engineering from Anhui University of Technology, Ma’anshan, China, in 2001, and Ph.D. in electrical engineering from Nanjing University of Aeronautics and Astronautics (NUAA), Nanjing, China, in 2008. He had worked in GE (China) research center, LG Electronics (Shanghai) Research Center, Shanghai Delphi Automotive Air-Conditioning System Co., Ltd. In Nov. 2017, he joined Valeo Powertrain (Shanghai) Co., Ltd, Shanghai, China, as hardware department manager. His current research interests include iBSG, DC– DC converter, onboard charger, and wireless power transfer system for electric vehicles. Tao Zhang was born in Hubei Province, China, in 1977. He received B.S. in industrial automation from Daqing Petroleum Institute, Daqing, China, in 2000, and M.S. in electrical engineering from Nanjing University of Aeronautics and Astronautics (NUAA), Nanjing, China, in 2005. In 2005, he joined Astec Power Co., Ltd. and in 2008 joined TDI Power Co., Ltd. for design and development of power products. At the beginning of 2011, he joined the Flex R&D (Shenzhen) Co., Ltd. for the research and development of new power technology. He joined Innoscience (Zhuhai) Technology Co., Ltd. in 2018, as AE director responsible for the application development of wide band gap semiconductor products GaN. He holds one US patent and two pended US patents.

About the Authors

xvii

Hong Yan was born in Hunan Province, China, in 1979. She received B.S. in automation from Hefei University of Technology, Hefei, China, in 2001, and M.S. in electrical engineering from Nanjing University of Aeronautics and Astronautics (NUAA), Nanjing, China, in 2009. Since April 2009, she joined the College of Electronic and Electrical Engineering at Bengbu University, where she is currently a Lecturer. Her main research interests include power electronics system integration and DC–DC converters.

Abbreviations

AC BJT CPES DC DSP GaN GTO HVICs ICS IGBT IPEM IPM IPOP IPOS ISOP ISOS IVS IVSR LSI MOSFET OCS OVR OVS PEBB PFC PFM PI PLL PSFB PWM

Alternating Current Bipolar Junction Transistor Center for Power Electronics Systems Direct Current Digital Signal Processor Gallium Nitride Gate Turn-off Thyristor High-Voltage Integrated Circuits Input Current Sharing Insulated Gate Bipolar Transistor Integrated Power Electronics Modules Intelligent Power Module Input-Parallel-Output-Parallel Input-Parallel-Output-Series Input-Series-Output-Parallel Input-Series-Output-Series Input Voltage Sharing Input Voltages Sharing Regulator Large-Scale Integrated Metal-Oxide-Semiconductor Field-Effect Transistor Output Current Sharing Output Voltage Regulator Output Voltage Sharing Power Electronics Building Block Power Factor Correction Pulse Frequency Modulation Proportional-Integral Phase-Locked Loop Phase-Shifted Full-Bridge Pulse-Width Modulation

xix

xx

RMS SCR SiC SoC SPICs SPWM ULSI UPS VLSI

Abbreviations

Root Mean Square Silicon Controlled Rectiﬁer Silicon Carbide System on Chip Smart Power Integrated Circuits Sinusoidal Pulse-Width Modulator Ultra-Large-Scale Integrated Uninterruptible Power Supply Very-Large-Scale Integration

Chapter 1

Introduction

Abstract Power electronics system integration is one of the effective approaches to reduce the cost of power electronics products, shorten the development cycle, and improve the reliability. Generally, power electronics system integration can be divided into three levels, namely, component-level integration, module-level integration, and system-level integration. One important branch of system-level integration is the series–parallel power conversion systems, in which multiple standardized converter modules are connected in series or parallel at the output and input sides to meet the requirements of different applications. In order to ensure the proper operation of series–parallel power conversion systems, proper sharing of the input/output voltage and proper sharing of the input/output current must be ensured. This chapter mainly introduces the control techniques for the series–parallel power conversion systems with DC–DC converter and DC–AC inverter as the basic module, respectively, which will lay a foundation for the narration of the following chapters. Keywords Power electronics system integration Series–parallel power conversion system · Input-parallel-output-parallel (IPOP) Input-parallel-output-series (IPOS) · Input-series-output-parallel (ISOP) Input-series-output-series (ISOS)

1.1 Development Trend of Power Electronics Power electronics started with the development of the mercury arc rectifier, which was invented by Peter Cooper Hewitt in 1902 and used for converting alternating current (AC) into direct current (DC) [1]. Other types of rectifiers, such as phanotrons and thyratrons, were gradually introduced in the 1930s. In 1956, the silicon controlled rectifier (SCR) was invented by Bell Laboratories and commercially introduced by General Electric Company [2], and after that, power electronics developed rapidly with the improved switching speed of semiconductor devices. A pioneer of power electronics, Dr. William E. Newell, pointed out that power electronics was interstitial to all three of the major disciplines of electrical engineering: electronics, power, © Springer Nature Singapore Pte Ltd. and Science Press, Beijing, China 2019 X. Ruan et al., Control of Series-Parallel Conversion Systems, CPSS Power Electronics Series, https://doi.org/10.1007/978-981-13-2760-5_1

1

2

1 Introduction

and control at the IEEE Power Electronics Specialists Conference (PESC) in 1973 [3]. In 1989, Dr. Bimal K. Bose expanded and refined the circumscription of power electronics, and demonstrated that power electronics was an interdisciplinary technology, including computer-aided design, power semiconductor devices, converter circuits, electrical machines, very large scale integration (VLSI) circuits, control theory, microcomputers, analog and digital electronics, and so on, as shown in Fig. 1.1 [4]. At present, more and more disciplines are involved in power electronics, such as modern control theory, microelectronics, material science, environmental science, and life science, and power electronics has become an independent interdisciplinary high-tech discipline. With the rapid development during the last nearly 60 years, power electronics technology has been widely used in almost all kinds of fields of industrial and agricultural production and daily life. Recently, more attention has been drawn to the renewable energy development and utilization globally due to the gap between the increasingly growth of energy demand and shortage of the traditional fossil energy, which also provides new opportunities for the development of power electronics. Power electronics technology will play an increasingly important role in renewable energy generation, flexible AC/DC transmission, electric vehicle and energy conservation, and environmental protection [5–8]. Power semiconductor devices, power conversion technology, and control technology are the three main fields involved in power electronics. The power semiconductor devices are the basis of the development of power electronics technology. Generally, the power semiconductor devices have gone through four stages of development: (1) Half-controlled devices: SCR is a typical representation of the half-controlled device, which has been widely used in low frequency, medium-to-high voltage AC power conversions. (2) Full-controlled devices: Gate turn-off thyristor (GTO), bipolar junction transistor (BJT), and metal-oxide-semiconductor field-effect transistor (MOSFET) are the main three kinds of full-controlled devices. The GTO and BJT are currentcontrolled devices, and the MOSFET is voltage-controlled device, which has higher switching speed than GTO and BJT.

Fig. 1.1 Power electronics—An interdisciplinary technology

Computer-aided design Analog and digital electronics Microcomputers

Power electronics and drives

Control theory

Power semiconductor devices Converter circuits Electrical machines

VLSI Circuits

1.1 Development Trend of Power Electronics

3

(3) Hybrid devices: Insulated gate bipolar transistor (IGBT) is a minority-carrier device, which integrates the advantages of the MOSFET and BJT, i.e., the high input impedance and high switching speed of a MOSFET and the low saturation voltage of a bipolar transistor. This hybrid combination is that the IGBT has the conduction characteristics of a bipolar transistor but is voltage-controlled like a MOSFET. (4) Power integrated devices: Smart power integrated circuits (SPICs) and high voltage integrated circuits (HVICs) are the main two kinds of power integrated devices, which integrate the power semiconductor devices, drive circuits, protection circuits, and even controllers. Recently, there has been significant progress in the development of wide bandgap semiconductor power devices, such as silicon carbide (SiC)-based and gallium nitride (GaN)-based power devices. SiC-based power devices have become strong competitors to Si-based power devices in medium power conversion applications, while GaN-based power devices can achieve ultra-fast switching. In summary, the power devices will be more large capacity, high frequency, modular, and intelligent in the future [9]. Power conversion technology can be classified into four types based on the forms of electric energy conversion: (1) AC-to-DC conversion, (2) DC-to-DC conversion, (3) DC-to-AC conversion, and (4) AC-to-AC conversion. Each electric energy conversion form can be achieved by various circuit topologies. The development of power conversion technology makes power electronics equipment have compact size, high conversion efficiency, and improved comprehensive economic benefit. Control technology is very important for power electronics technology. Various control technologies have been rapidly proposed and widely used, such as pulsewidth modulation (PWM) and pulse-frequency modulation (PFM). With the development of modern control theory, fuzzy control, genetic control, neural network control, and so on are introduced to the power electronics. With the growing applications of power electronics technology, it also faces various new challenges. At present, most power electronics products are specially designed according to different purposes and requirements of customers, i.e., customdesigned products. The design of power electronics products usually is a complex procedure with long development cycle time and high cost, since many aspects are involved, such as semiconductor devices, circuit topologies, control, magnetic material, package, mechanical structure, thermal design, and electromagnetic compatibility. Therefore, lots of repeated work exist when developing different power electronics products, resulting in lower labor efficiency and higher production cost. In order to reduce the cost and shorten the development cycle time, it is eager to improve the standardization of power electronics products for large-scale production to make the products can meet the rapidly changing market and customer needs, which is also one of major challenges for power electronics technology. The consensus among the world’s power electronics industry and academia is that the effective way to solve the problem is power electronics system integration, standardization and modularization [10].

4

1 Introduction

1.2 Power Electronics System Integration There are two main branches for the development of electronic technology after the appearance of semiconductor device. One is the microelectronic technology to process information, and the other is the power electronics technology to process power (energy). The development of microelectronic technology has experienced several stages, including discrete devices, large scale integrated (LSI) circuits, very large scale integrated (VLSI) circuits, ultra large scale integrated (ULSI) circuits, and system on chip (SoC), and now the integration is still the trend of microelectronic technology. Compared with the microelectronic technology, the development of power electronics integration is relatively slow. At present, most of power electronics products are composed of various discrete devices. Inspired by the success of the development of microelectronics technology, people will think that the cost of production will be reduced and the development cycle time will be shortened if the integration is applied in the power electronics, and this is the original intention of the proposal of power electronics system integration. The goal of power electronics system integration is to increase the integration levels in the components, including devices, circuits, controls, and sensors, which are integrated into standardized manufacturable modules, and then these standardized modules can be easily used to build various types of power electronics equipment. The idea of power electronics system integration was originally proposed by the US Navy, and the famous concept of power electronics building block (PEBB) was proposed with the vision of changing the paradigm for designing and manufacturing electrical high power conversion and control products [11]. After that, the Center for Power Electronics Systems (CPES) has been doing very well-known work on the PEBB, including the monolithic integration of high-voltage power devices and lowvoltage devices for controls and sensors, the interconnection of high-power devices and control devices on a common substrate, thermal management, and integration of passive high-power components, and CPES developed the integrated power electronics modules (IPEM), a standardized off-the-shelf module that has revolutionized power electronics [12]. Generally, the power electronics system integration can be divided into three levels. The first one is component-level integration, including active devices integration and passive components integration. The active power devices integration includes active power devices, drive circuits, protection circuits, and communication circuits, while the passive components integration includes inductors, capacitors, and transformers. The second one is module-level integration, in which a variety of integrated active devices and/or passive components are integrated together to form highly versatile standardized converter modules. The last one is system-level integration, in which a number of converter modules are put together to form different kinds of power conversion systems.

1.2 Power Electronics System Integration

5

1.2.1 Component-Level Integration The active devices integration refers to integrate active devices, drive circuits, control circuits, and protect circuits on a common substrate or into a module, leading to reduced volume and influence of the connection among devices. For instance, intelligent power module (IPM) is a high-function integrated module with IGBT chip and dedicated integrated circuits for self-protection (over-current, under-voltage, and over-heating). The passive components integration includes magnetic components integration (such as the integration of inductor and transformer, integration of inductors), integration of magnetic components and capacitors, and integration of magnetic components, capacitors, resistors, etc. Two or more discrete inductors and/or transformers can share a common magnetic core through proper coupling way and parameters design, leading to reduced magnetic components volume and total losses. The magnetic components and capacitive components can be integrated together, for instance, R. Reeves proposed an integration method for an inductor and a capacitor [13]. The resistors can also be integrated with magnetic components and capacitive components, in which the resistors are used for snubber circuit or control circuit [14].

1.2.2 Module-Level Integration Module-level integration refers to the integration of various active devices, passive components, control circuits, and protection circuits for a standardized module with strong versatility. CPES has proposed multi-chip packaging technology to achieve power electronics module-level integration. Kinds of power electronics devices in the form of bare dies are mounted on a ceramic or metal substrate, and the control circuits, drive circuits, protection circuits, snubber circuits, and sensors are also mounted on the same substrate in the form of IC bare dies or surface mount packages. All these devices and circuits are interconnected through thin-film power overlay technology and then connected with passive components through three-dimensional package technology to form a converter module to realize power conversion.

1.2.3 System-Level Integration System-level integration technology refers to integrate various standardized converter modules into systems to meet different power conversion demands. One important branch of system-level integration is series–parallel power conversion systems, in which multiple standardized converter modules are connected in series or parallel at the output and input sides to meet the demands of different applications [15]. For instance, the input sides of multiple converter modules can be connected in series for

6

1 Introduction

high input voltage applications, and the output sides of multiple converter modules can be connected in parallel for large output current applications. Series–parallel power conversion systems can be categorized into four possible architectures on the basis of the connection forms, namely, input-parallel-output-parallel (IPOP), inputparallel-output-series (IPOS), input-series-output-parallel (ISOP), and input-seriesoutput-series (ISOS), as shown schematically in Fig. 1.2. Each type of series–parallel power conversion system has its own specific application areas. With the rapid expansion of computer and telecommunication industries, distributed power systems consisting of paralleled modular converters delivering low output voltage and high output current are still the most popular practical configuration, where IPOP systems have been widely adopted. ISOP systems can be used in applications where the input voltage is relatively high and the output voltage is relatively low, such as in high-speed train power systems, industrial drives, and undersea observatories. For applications requiring high output voltages, such as X-ray equipment, photovoltaic systems, and electrostatic precipitator, IPOS systems become the desired choice. Moreover, ISOS systems are well suited for applications where both the input voltage and output voltage are high.

+ vin _

+ + vin1 Module 1 vo1 _ _

+ v_o

+ vin _

+ + vin1 Module 1 vo1 _ _

+ + v_in2 Module 2 vo2 _

+ + v_in2 Module 2 vo2 _

+ + v_inn Module n von _

+ + v_inn Module n von _

(a) IPOP

+

vin

_

+ + vin1 Module 1 vo1 _ _ + + v_in2 Module 2 vo2 _

+ + v_inn Module n von _ (c) ISOP

+

vo

_

(b) IPOS

+ v_o

+

vin

_

+ + vin1 Module 1 vo1 _ _ + + v_in2 Module 2 vo2 _

+ + v_inn Module n von _ (d) ISOS

Fig. 1.2 Four connection architectures of multiple converter modules

+

vo

_

1.2 Power Electronics System Integration

7

The advantages of the systems constructed from connecting multiple converter modules include (a) ease of thermal design as a result of each module handling only a part of the total power; (b) increased overall system reliability due to reduced thermal and electrical stresses on the power devices and components; (c) shortened design process and lowered cost for the system; (d) improved system reliability due to N + 1 redundancy; and (e) ease of expansion of power system capability [16]. According to the form of energy conversion, the modules in Fig. 1.2 can be DC–DC converter, DC–AC inverter, AC–DC rectifier, and AC–AC converter. This book focuses on the series–parallel power conversion systems with DC–DC converter and DC–AC inverter as the basic modules, respectively, and their recent development will be introduced as follows.

1.3 DC–DC Series–Parallel Power Conversion Systems 1.3.1 IPOP System Paralleling of multiple standardized converters has been extensively studied and widely applied in the power supplies for computers and communications. One basic objective of IPOP system is to share the load current among the constituent converters. To do this, a variety of approaches have been proposed, such as droop method, master–slave current sharing method, democratic current sharing method, etc. [17–20]. The droop method can be defined as one in which the output voltage droops as the load current is increased, as shown in Fig. 1.3. Its operation mechanism is to program the output impedance to achieve current sharing among converters. The droop method has the advantages of easy implementation, no wire connection among the control circuits of converters and high modularity and reliability. However, the current sharing and load regulation are relatively poor. In the master–slave current sharing method, one module is dedicated to be the master module and the remaining modules are the slave modules. A typical block

Io1 Ro1 V1 Converter 1

Io2

+ Vo RLd _

V1 V2 V Vo

Ro2 V2 Converter 2

(a) Equivalent circuit Fig. 1.3 Two-converter IPOP system

0

Io1

Io2

(b) Output characteristics

I

8

1 Introduction _ Vof + Voref

Gvo

Iref + _

Gi

Module 1 RS Flip-flop

VRAMP 1

ILf1_f

+ _

Comparator 1 − +

Gi

Module 2

Comparator 2 − +

RS Flip-flop

VRAMP 2

ILf2_f

Drive circuit 2

…

Currentsharing bus

Drive circuit 1

+ _ ILfn_f

Gi

Comparator n − +

Module n RS Flip-flop

Drive circuit n

VRAMP n

Fig. 1.4 A typical block diagram of master–slave current sharing method

diagram of master–slave current sharing method is shown in Fig. 1.4, where module 1 is the master module, which is voltage–current double closed-loop controlled, and modules 2 to n are the slave modules, which are only current closed-loop controlled. Gvo is the transfer function of the output voltage regulator, and Gi is the transfer function of the current regulator. The output of Gvo , I ref , is taken as the current sharing bus and becomes the common reference for all current closed loops to achieve current sharing. The limitations of the master–slave current sharing method are poor fault tolerance since the master failure will disable the entire system and degraded modularity of the system. Figure 1.5 shows a typical block diagram of the democratic current sharing method. As seen, the load current signal of each module is connected to the current sharing bus through an ideal diode, and the module with the highest output current is automatically selected to be the master module and its load current signal is taken as the current reference for all other modules. The current reference is then compared with the individual converter load current and the error signal is added to the voltage reference of each corresponding converter to regulate its load current to achieve current sharing. Obviously, in the democratic current sharing method, there is no dedicated master module, leading to good fault tolerance and modularity.

1.3 DC–DC Series–Parallel Power Conversion Systems _ io1 +

Gio

_ io2 +

Gio

Gvo

Voref + + _ vof

Gvo

Comparator 1 − +

Module 1 RS Flip-flop

VRAMP 1

Comparator 2 − +

Drive circuit 1 Module 2

RS Flip-flop

VRAMP 2

Drive circuit 2

…

Currentsharing bus

Voref + + _ vof

9

_ ion +

Gio

Voref + + _ vof

Gvo

Comparator n − +

Module n RS Flip-flop

VRAMP n

Drive circuit n

Fig. 1.5 A typical block diagram of democratic current sharing method

1.3.2 IPOS System IPOS system can be applied in the occasions requiring high output voltages or higher step-up ratios, such as X-ray equipment, photovoltaic systems, and electrostatic precipitator [21, 22]. The key issue of the IPOS system is to ensure output voltage sharing (OVS) among the constituent converters. Due to the output sides are connected in series, the constituent converters of IPOS system can regulate its own output voltages to be 1/n of the system output voltage (n is the number of converters) to share the output voltage. However, OVS is affected by the degree of matching among the sampling circuits and voltage references of constituent converters and it is difficult to obtain precise system output voltage. A simple common duty cycle control strategy was proposed in [23] for IPOS system; however, OVS cannot be ensured due to parameters mismatches such as turns ratio of individual power transformer. Ref. [24] proposed an OVS control strategy for IPOS system with a common system output voltage control loop and individual input current control loop for each module. The output of system output voltage regulator serves as the reference of input current control loop of each module to achieve input current sharing (ICS) and equal voltage regulation among the constituent modules, assuming that all the modules are identical. In [25], the module with maximum output voltage is regarded as the master module and its output voltage is taken as the reference, which is compared with the individual module output voltage and the error signal is added to the corresponding output of the output voltage regulator to regulate the module’s output voltage to achieve OVS. Hence, it can be seen that OVS of IPOS system can be achieved by both controlling ICS of modules and directly controlling OVS of modules. The relationship between ICS and OVS of IPOS system was revealed in [16]. It was pointed out that achieving ICS implies OVS and vice versa, and a novel OVS

10

1 Introduction Output voltage sharing loops Kvo

vo1

_ vo1f +

+

− + + VRAMP 1

Gvo_mod vo_EA1

Kvo

vo2

+

Comparator 1

_ vo2f

+

Gvo_mod

Comparator 2 − RS Flip-flop +

+ VRAMP 2

Kvo

_ +

vomod_ref

vo(n−1)f

vo

Kvo

Voref

vo_EA(n−1) … + + +

vof +

Comparator (n−1) − RS Flip-flop + + VRAMP (n−1) +

Gvo_mod

1/n

_ Gvo

Drive circuit 1

Drive circuit 2

…

…

vo_EA2

vo(n−1)

RS Flip-flop

vo_EA

_

Comparator n − RS Flip-flop +

+ VRAMP n

Drive circuit (n−1)

Drive circuit n

Output voltage loop

Fig. 1.6 OVS control strategy for IPOS system

control strategy was proposed, as shown in Fig. 1.6. It consists of a common output voltage control loop to regulate the system output voltage and n − 1 OVS control loops to achieve OVS for IPOS system. K vo is the sensing gain of the system output voltage and output voltage of each module. Gvo is the transfer function of the system output voltage regulator. Gvo_mod is the transfer function of OVS regulator. It can be seen that the reference voltage signal of each OVS loop is K vo vo /n, vo_EA is the control signal derived from the system output voltage regulator, and vo_EAj (j 1, 2,…, n − 1) is the control signal derived from each OVS regulator. For the first n − 1 modules, vo_EAj (j = 1, 2,…, n − 1) is added to vo_EA and then compared with the sawtooth signal to obtain the duty cycle signal of module j. For the module n, n−1 vo_EA j is subtracted from vo_EA and then compared with the sawtooth signal to j1

obtain the duty cycle signal of module n. With the control strategy, the duty cycles of modules with higher output voltages will decrease while the duty cycles of modules with lower output voltages will increase, and thus OVS is achieved. From Fig. 1.6, it can be found that the sum of the duty cycles caused by the OVS loops is zero, which means that the OVS control loops have no influence on the system output voltage regulation.

1.3 DC–DC Series–Parallel Power Conversion Systems

11

1.3.3 ISOP System ISOP system can be used in higher input voltage applications such as the electric railway system, where multiple converter modules with lower input voltage are adopted, which is beneficial for reducing the voltage stress of the power switches. In order to ensure the proper operation of ISOP system, the basic objective is to ensure input voltage sharing (IVS) and output current sharing (OCS) among the constituent modules. The common duty cycle control strategy can also be used for ISOP system [23]. With the common duty cycle for all modules, the input currents of the modules with higher input voltages will increase to reduce their input voltages, while the input currents of the modules with lower input voltages will decrease to boost their input voltages, and finally to achieve IVS. Similar to the IPOS system, IVS is also affected by the matching of modules’ parameters. A fault-tolerant democratic IVS control strategy for ISOP system was proposed in [26], as shown in Fig. 1.7. Each input voltage signal of module is connected to the voltage sharing bus through an idea diode, producing a common voltage reference signal, which is obtained from one of the converters with highest input voltage within the ISOP system. The voltage difference between the common voltage reference and each converter input voltage is amplified and added to the output of system output voltage regulator for minor control correction to ensure IVS. Gvcd is the transfer function of IVS loop regulator. If a converter fails with its input short-circuited, the common voltage reference is automatically increased to compensate for the loss of the failed converter, leading to high system reliability. An IVS control strategy was proposed for ISOP system in [27], as shown in Fig. 1.8. The control strategy consists of a common output voltage control loop for achieving desired system output voltage and n − 1 IVS control loops. V in /n is the reference signal for IVS control loops. vo_EA is the control signal derived from the system output voltage regulator, and vcd_EAj (j 1, 2,…, n − 1) is the control signal derived from each IVS regulator. For the first n − 1 modules, vcd_EAj (j 1, 2,…, n − 1) is subtracted from vo_EA and then compared with the sawtooth signal to obtain n−1 the duty cycle signal of module j. For the module n, vcd_EA j is added to vo_EA and j1

then compared with the sawtooth signal to obtain the duty cycle signal of module n. With the control strategy, the duty cycles of modules with high input voltages will increase and the duty cycles of modules with low input voltages will decrease, and finally to achieve IVS. It is proved that the IVS loops are decoupled from one another and are also decoupled from the system output voltage loop. Thus, the IVS regulator and system output voltage regulator can be designed independently. Based on the IVS control strategy shown in Fig. 1.8, an inner current control loop could be introduced for each module, for the purpose of improving the system dynamic performance. The feedback signal of inner current control loop can be the input currents or output currents of the modules. Charge control method could be also used for the inner current control loops, where the input currents of the modules are taken as the feedback signals for a two-module ISOP system [28]. In [29], the

12

1 Introduction

Input voltage sharing loops +

_ vin1

_ Gvcd

Voltage sharing bus +

_ vin2

+

_ Gvcd

Drive circuit 1

Comparator 2 − RS Flip-flop +

Drive circuit 2

VRAMP 1

VRAMP 2

…

…

+

Comparator 1 − RS Flip-flop +

+

_ vinn

_ Gvcd +

Voref

+ _

Comparator n − RS Flip-flop +

VRAMP n

Drive circuit n

Gvo vof

Output voltage loop Fig. 1.7 Democratic IVS control strategy for ISOP system

output filter inductor currents are taken as the feedback signals for the inner current control loops with average current control mode. As shown in Fig. 1.8, the duty cycle of each module regulates both the system output voltage and corresponding module input voltage. A simple IVS control strategy was proposed in [30], as shown in Fig. 1.9, where the regulation object of the duty cycle of each module is separated. For the first n − 1 modules, the duty cycles are used to regulate the input voltage of each module to be V in /n, and obviously the input voltage of module n is also V in /n. The system output voltage is regulated by module n. Sensing of the input voltages of the modules is necessary in the control strategies [26–30] to achieve IVS. A current cross-feedback control strategy was proposed in [31] avoiding the sensing of the input voltages. Figure 1.10 shows the control strategy for two-module ISOP system. Two modules use current-mode control with a common system output voltage regulator providing the current reference iref for both current loops. It should be noted that the feedback of current regulator for module 1 is the output current of module 2, while the feedback of current regulator for module 2 is the output current of module 1. Assuming the system input voltage is constant and the system works at steady state, if the input voltage V in1 increase while the input voltage V in2 decreases due to a disturbance, then filter inductor current iLf1 increases while iLf2 decreases. With the control strategy shown in Fig. 1.10, the duty

1.3 DC–DC Series–Parallel Power Conversion Systems

13

Input voltage sharing loops Vin/n

_ vin1 +

_

− + + VRAMP 1

Gvcd vcd_EA1

+

Comparator 1

_ vin2

_ Gvcd

Comparator 2 − RS Flip-flop +

+ VRAMP 2

vin(n−1)

+

_ Comparator (n−1) − RS Flip-flop + + VRAMP (n−1)

Gvcd vcd_EA(n−1) … + + +

+ Voref _

Gvo

Drive circuit 1

Drive circuit 2

…

…

vcd_EA2

_

RS Flip-flop

vo_EA

+

Comparator n − RS Flip-flop +

+ VRAMP n

Drive circuit (n−1)

Drive circuit n

vof

Output voltage loop Fig. 1.8 IVS control strategy for ISOP system

cycle of module 1 increases while the duty cycle of module 2 decreases, resulting in the decrease of V in1 and increase of V in2 , and finally IVS can be achieved. When the number of the modules increases, the control strategy will become complex due to mutual cross-feedback among the current signals and the system modularity is lowered.

1.3.4 ISOS System ISOS system is suitable for the applications where both the input and output voltages are relatively high, and it is critical to ensure both IVS and OVS for the stable operation of ISOS system. Ayyanar et al. proposed a voltage sharing control for ISOS system [32], which is similar to Fig. 1.8. References [33, 34] pointed out that a weak natural rebalancing ability exists in the ISOS system with interleaved switching for

14

1 Introduction

_+

vin1 Gvcd

Vin/n

Module 1

Comparator 1 − RS Flip-flop +

Drive circuit 1

VRAMP 1 _+

vin2 Gvcd

Vin/n

Module 2

Comparator 2 − RS Flip-flop +

Drive circuit 2

…

…

VRAMP 2 vin(n−1)

Comparator (n−1) − RS Flip-flop +

_+ Gvcd

Vin/n

VRAMP (n−1)

Module (n−1) Drive circuit (n−1) Module n

Comparator n +

Gvo

Voref _ vof

VRAMP n

− +

RS Flip-flop

Drive circuit n

Fig. 1.9 A simple IVS control strategy for ISOP system

_ iLf2 +

Output voltage regulator + Voref _

Gvo vof

iref +

Gi Current regulators _ iLf1 Gi

Comparator 1 − RS Flip-flop +

Drive circuit 1

Comparator 2 − RS Flip-flop +

Drive circuit 2

VRAMP 1

VRAMP 2

Fig. 1.10 Current cross-feedback control strategy for two-module ISOP system

different cells. Although the voltage-dependent losses, such as the switching loss, could enhance the rebalancing ability, the voltage sharing performance is considerably affected by the matching of module parameters and load current. The current cross-feedback control strategy can also be used for ISOS system to achieve voltage sharing [35, 36], and its operation principle is the same as that of

1.3 DC–DC Series–Parallel Power Conversion Systems

15

Fig. 1.10. Likewise, the control strategy will become complex when the number of the modules increases. In conclusion, these prior studies focus on one or two of the four connection architectures and do not reveal the inherent relationships between the four connection architectures. In this book, Chap. 2 will systematically analyze the inherent relationships between the sharing of input voltages/currents and the sharing of output voltages/currents from the view of power conservation, and discuss the stability of various possible control strategies. Based on the analysis, a general control strategy is proposed for all the four connection architectures. In Chap. 3, a small-signal model of ISOP system with phase-shifted PWM full-bridge converter as the basic module is derived to design the output voltage loop regulator and IVS loop regulator. Moreover, the design of the IVS loop regulator is analyzed when voltage–current double closed loop is employed for ISOP system. In Chap. 4, a wireless IVS control strategy for ISOP system based on positive output voltage gradient method is presented, in which no control interconnection among the constituent modules is required, leading to a truly modular design for ISOP system and high system reliability.

1.4 DC–AC Inverter Series–Parallel Power Conversion Systems 1.4.1 DC–AC Inverter DC–AC inverter is used for converting a DC voltage into an AC voltage, and there are various topologies. In this book, the popular single-phase full-bridge inverter is selected as the basic module to constitute the series–parallel combined systems. Figure 1.11a shows the single-phase full-bridge inverter, where V in is the input DC voltage, S 1 (D1 )–S 4 (D4 ) constitute the single-phase inverter bridge, L f is the output filter inductor, C f is the output filter capacitor, and Z Ld is the load impedance. For DC–AC inverter, there are mainly two control methods, namely, the voltage single closed-loop control and the voltage–current double closed-loop control. The block diagram of the voltage–current double closed-loop control is shown in Fig. 1.11b, where the error of the voltage reference voref and the sampling signal of output voltage vof is sent into the voltage regulator Gvo , and the output signal of Gvo serves as the current reference iref . Then, the error of the sampling signal of inductor current iLf and iref is sent to the current regulator Gi . The output signal of Gi is fed into the sinusoidal pulse-width modulator (SPWM) and it is compared with the triangular carrier to obtain the drive signals for the four power switches. By removing the inner current loop in Fig. 1.11b, the voltage single closed-loop control will be obtained, as shown in Fig. 1.11c. Compared with the voltage single closed-loop control, the voltage–current double closed-loop control has better dynamic performance and the current-limiting function.

16

1 Introduction

S1

D1 S3

D3 + vinv

Vin

Lf iLf

Cf

– S2

D2 S4

io

+

iCf ZLd

vo _

D4

(a) Single-phase full-bridge inverter +

voref

–

vo

Kv

iLf

Ki

Gvo

+ iref

–

Gi

SPWM

S1~S4

vof iLf_f

(b) Voltage-current double closed-loop control +

voref vo

–

Kv

Gvo

SPWM

S1~S4

vof

(c) Voltage single closed-loop control Fig. 1.11 Single-phase full-bridge inverter and its control block diagram

Regardless of any control method being adopted, multiple DC–AC inverter modules could be connected in series or parallel at the input and output sides to constitute the DC–AC inverter series–parallel combined system.

1.4.2 IPOP DC–AC Inverter System IPOP inverter system, which is commonly known as the parallel inverter system, has been extensively studied among the four series–parallel combined inverter systems. This system has been widely used in uninterruptible power supply (UPS), aeronautical static inverter, and so on. The crucial issue of the IPOP inverter system is to achieve output current sharing of each module. The research aiming at this control objective has experienced the evolution from centralized control, master–slave control to distributed control. The control methods for IPOP inverter system can be also categorized into interconnected control and wireless droop control, and the wireless droop control is essentially a kind of distributed control.

1.4 DC–AC Inverter Series–Parallel Power Conversion Systems

1.4.2.1

17

Centralized Control

Figure 1.12 shows the block diagram of the centralized control for IPOP inverter system [37, 38]. In the centralized control method, the phase-locked loop (PLL) in each inverter module is used to synchronize its output voltage in frequency and phase with the synchronization pulse signal, which is produced by the crystal oscillator in the parallel control unit. The total load current io is sensed and divided by the number of constituent inverter modules connected in parallel, and then serves as the current sharing reference, iref , for each inverter module. The sensed output current of each module is compared with iref , obtaining the corresponding current deviation. Thanks to being synchronized by one synchronization pulse signal, the output voltage of each inverter module can be almost the same in frequency and phase, and the current amplitude difference among the inverter modules is caused by the inconsistence of the voltage amplitude. The aforementioned current deviation is used to compensate any inequality in voltage amplitude of the inverter modules to achieve output current sharing of each module. The implementation of the centralized control is quite simple, and the current sharing is relatively good. However, the whole parallel system tends to collapse once the parallel control unit fails. Thus, the redundant and reliability is not high.

io

Fig. 1.12 Block diagram of centralized control for IPOP inverter system

18

1.4.2.2

1 Introduction

Master–Slave Control

Figure 1.13 shows the master–slave control [39–42], where one inverter is the master module and the others are the slave modules. In Fig. 1.13, K vo is the output voltage sense gain, K i1 –K in are the filter inductor current sense gain of each module, Gvo is the transfer function of the output voltage regulator, Gi1 –Gin are the transfer functions of the current regulator of each module, and Ginv1 –Ginvn are the transfer functions of the SPWM modulators. The output voltage regulator in the master module generates the unified current reference signal iref . Since the output filter inductor current of each module tracks iref , the output current sharing is achieved. The shortcoming of the master–slave control is that the whole system will be collapsed once the master module fails.

1.4.2.3

Distributed Control

1. Average Current Control Method Figure 1.14 shows the average current control, a kind of distributed control [43]. As seen, the voltage reference voref is sent to the output voltage regulators of all the inverter modules, and the average value of the outputs of the output voltage regulators of all the modules, iave , serves as the current reference for all the modules. With such arrangement, the instantaneous current sharing of each inverter module could be achieved without complex current sharing control circuit. This control method has the following merits: simple control circuit, prompt current sharing regulating speed,

Fig. 1.13 Block diagram of master–slave control for IPOP inverter system

1.4 DC–AC Inverter Series–Parallel Power Conversion Systems voref

19

iave + _

Gvo1

iref1

+ _

vof

Gi1 iLf1_f

vinv1 _ + Ginv1 Ki1

irefj +

+ +

+ _

iLf1 io1 +_ iCf1 sCf1

Kvo

Module 1 Output voltage synchronous bus

1 sLf1

Gvon vof

1 n

+ irefn

Module n

ioj

Current sharing bus

_

Gin iLfn_f

+

+ +

vinvn _ +

Ginvn

1 sLfn

Kin

iLfn

ZLd

vo

ion

+_ iCfn sC fn

Kvo

Fig. 1.14 Average current control method for IPOP inverter system

and ease of implementation of redundancy due to modular completely reciprocal. However, the current sharing bus among modules is susceptible to interference, which tends to decrease the current sharing performance. In addition, this control method has poor output characteristics. In order to improve the output characteristics of both single module and the whole system, the load current feed-forward control is introduced [44]. Here, the feed-forward point is located before the current reference averaging point. In the improved method, the current sharing performance and the current-limiting function are remained as the conventional method. 2. Wireless Droop Control In the aforementioned control methods, the current sharing bus is required, and it is susceptible to interference. Figure 1.15 shows the wireless droop control [45–49], which has no current sharing bus. The output characteristic droop circuit is introduced to make the frequency and amplitude of the output voltage of each module decrease with the increase of the active and reactive power, respectively, and thus the frequency and amplitude of the output voltage of each module stabilize at a new equilibrium point. By this means, the control method suppresses the circulating current and achieves the output current sharing. Since there is no connection among the constituent inverter modules, the wireless droop control is not susceptible to interference and has better redundancy. Nevertheless, due to the droop characteristic, the output characteristics are poor.

20

1 Introduction

Sine wave ω1 reference generator Vsinωt V1 +

Gvo1 voref1 – vo1f

iref1 +

P1 Q1 Gi1

– iLf1_f

Power calculation vinv1 Ginv1

_

iLf1 1 sLf1

+

+ _

io1

iCf1 sC f1

Ki1

vo

Kvo

Module 1

ioj

+

+ +

Sine wave reference generator Vsinωt +

Gvon vorefn – vonf Module n

ωn

Pn Qn

irefn +

Gin

– iLfn_f

vinvn +

_

Ginvn

Kin

vo

vo

Power calculation Vn

ZLd

1 sLfn

iLfn

+ ion _ iCfn sC fn

Kvo

Fig. 1.15 Wireless droop control strategy for IPOP inverter system

1.4.3 IPOS DC–AC Inverter System Similar with the output current sharing control for the IPOP inverter system, the output voltage sharing control is only needed for IPOS inverter system. Figure 1.16 shows a three-loop control strategy for IPOS inverter system [50]. The three loops here refer to the common system output voltage loop, the independent output voltage sharing loop of each module, and the inner current loop of each module. In Fig. 1.16, voref is the voltage reference of the system output voltage loop, Gvo is the output voltage regulator, Gi1 –Gin are the current regulators, L f1 –L fn are the filter inductors, C f1 –C fn are the filter capacitors, Gvo1 –Gvon are the regulators of output voltage sharing loop, and the voltage reference vomod_ref of output voltage sharing loop is equal to voref /n. The common output voltage loop regulates the system output voltage, and the output signal of its regulator Gvo serves as an initial reference signal for the inner current loop of each module. Since discrete parameters of each module are not well identical, the output voltage sharing loop is introduced to adjust the reference of the current loop to achieve output voltage sharing of each module. This control strategy

1.4 DC–AC Inverter Series–Parallel Power Conversion Systems

vomod_ref vo1

vo2

Kvo _ +

Kvo _ +

Gvo1

Gvo2

Output voltage sharing loops von

Kvo _ +

voref + _

+ +_

+ +_

21

Kvo

1/n

Inner current loops _ vinv1 Gi1 Ginv1 +

1 iLf1 + iCf1 1 vo1 sLf1 sCf1

iLf1_f

Ki1

Gi2 iLf2_f

Ginv2

vinv2 +

_

vo2 1 iLf2 + iCf2 1 sCf2 sLf2

io

Ki2 +

Gvon

Gvo

iref

vof Output voltage loop

+ +_

Gin iLfn_f

Ginvn

vinvn +

_

+ +

1 ZLd vo

1 iLfn + iCfn 1 von sLfn sCfn

Kin Kvo

Fig. 1.16 Control strategy for IPOS inverter system

can achieve low total harmonic distortion in the system output voltage, fast dynamic response, and good output voltage sharing performance whether the corresponding parameters among modules are equal or not.

1.4.4 ISOP DC–AC Inverter System The crucial issue of the ISOP inverter system is to ensure the input voltage sharing and the output current sharing among the constituent modules. Figure 1.17 shows a compound balanced control strategy [51], which consists of three control loops, namely, the common output voltage loop, the input voltage sharing loop, and the inner current loop of each module. Here, the current inner loop adopts inductor current feedback hysteretic control. In Fig. 1.17, Gvo is the output voltage regulator, Gvcd is the input voltage sharing regulator, and K vi and K vo are the sensor gains of the input voltage and output voltage, respectively. The common output voltage loop regulates the system output voltage, and the output signal of output voltage regulator Gvo , iref , serves as the initial current reference of all the inner current loops. The input voltage sharing loop ensures the even input voltage and the equivalence of the output active power of each module. Since the control object is the inverter whose current reference signal contains both amplitude and phase information, the multipliers are introduced here to synchronize the phase of the initial current reference and that of

22

1 Introduction

Vin n K + vi _ vin1 Kvi Vin n

Gvcd

isv1 _ iref1 +

+_

Input voltage sharing loops + _

Gvcd

Vin n K + vi _ vinn Kvi

Gvcd

vin2

ivd1

Kvi

Gvo vof

Output voltage loop

iLf1_f

_

Inner current loops isv2 _ vinv2 _ iref2 Gi2 Ginv2 + + +_ iLf2_f Ki2

ivdn

isvn

iref

_ + irefn +_

Gin iLfn_f

1 iLf1 + io1 _ sLf1 iCf1

Ki1

ivd2

Kvi

voref + _

Gi1

vinv1 Ginv1 +

vinvn

Ginvn + Kin

_

sCf1

+ vo 1 iLf2 + io2 io ZLd sLf2 _ ++ iCf2 sCf2

1 iLfn + ion sLfn _ iCfn

sCfn

Kvo

Fig. 1.17 Compound balanced control strategy for ISOP inverter system

the signal isvj (j 1, 2,…, n), which is used to alter the current reference by the input voltage sharing loop. Consequently, the output current amplitude is adjusted with the power factor angle of modular output current being kept the same simultaneously. Figure 1.18 shows an output-inductor-current cross-feedback control scheme [52], which consists of two control loops, namely, the common output voltage loop and the inner current loop of each module. The common output voltage loop regulates the system output voltage. In Fig. 1.18, Gvo is the output voltage regulator, and its output signal, iref , serves as the common reference signal of inner current loop of each module. Here, the current feedback for each individual inner current loop is the sum of other output inductor currents instead of its own, and so input voltage sharing (IVS) and output voltage sharing (OCS) are achieved for ISOP inverter system. Although the control scheme is simple, the need of cross-feedback for output inductor current is not conducive to the modularization of the system.

1.4.5 ISOS DC–AC Inverter System The crucial issue of ISOS inverter system is to ensure the input voltage sharing and the output voltage sharing of each module. A compound balanced control strategy is proposed [53], as shown in Fig. 1.19. It consists of three control loops, namely, the system output voltage loop, the input voltage sharing loop, and the inner output filter capacitor current loop of each module. The output signal of the output voltage regu-

1.4 DC–AC Inverter Series–Parallel Power Conversion Systems

+

voref _ Gvo vof

Module 1 iref + Gi1 _ n

∑i

Lfi_f

i=2

Module j + Gij _ n

∑

_ +

io1 1 iLf1 Ginv1 v sLf1 inv1 +_ iCf1 iLf1_f Ki1

Ginvj

+

_

vinvj iLfj_f

iLfi_f

i =1,i ≠ j

Module n +

_

Gin n −1

∑i

Lfi_f

i =1

23

sCf1

+ ioj 1 iLfj ZLd sLfj +_ + iCfj + sCfj Kij

_ ion + 1 iLfn Ginvn sLfn vinvn +_ iCfn iLfn_f Kin

vo

sCfn

Kvo Fig. 1.18 Output-inductor-current cross-feedback control for ISOP inverter system

lator Gvo , iref , serves as the initial current reference of all inner current loops. Since the current reference contains the amplitude and phase information, the multipliers are introduced here to synchronize the phase of the initial current reference and that of the signal isvj (j 1, 2,…, n), which is used to alter the current reference by input voltage sharing loop. Here, the output filter capacitor current control is adopted for the inner current loop, so the input voltage sharing loop only adjusts the amplitude of the output filter capacitor current with the phase of the output filter capacitor current of each module being kept the same. Therefore, only the amplitude of module output voltage is adjusted while the phase is synchronized. The input voltage sharing loop ensures the even input voltage and the equal of the output active power of each module. Thanks to the synchronization of the output voltage of each module, output voltage sharing (OVS) is achieved ultimately. An output-voltage cross-feedback control scheme is proposed [54], as shown in Fig. 1.20. It consists of two control loops, namely, the system output voltage loop and individual output voltage loop of each module. In Fig. 1.20, voref is the system output voltage reference, and the output signal of output voltage regulator Gvo , iref , is sent to all the modules. In the output voltage loop of each module, the reference signal is the amplitude of the system output voltage reference multiplied by (n − 1)/n, and the feedback signal is the sum of output voltage amplitude of other modules except itself. The output voltage regulators Gvo1 –Gvon generate regulating signals

24

1 Introduction

Vin n K + Gvcd1 vi _ vin1 Kvi Input voltage Vin sharing loops n K + Gvcd2 vi _ vin2 Kvi Vin n K + vi _ vinn Kvi

Gvcdn

voref

Gvo

+_

vof Output voltage loop

ivd1

isv1 _ iref1 +

ivd2

ivdn

+_

Gi1 iCf1_f

vinv1 Ginv1 +

_

1 iLf1 iCf1 sLf1 _

1 vo1 sCf1

Ki1

Inner current loops isv2 _ vinv2 _ iref2 Gi2 Ginv2 + + +_ iCf2_f Ki2

vo2 1 sLf2

iLf2 iCf2 _

1 sCf2

io 1 ZLd

isvn

vo iref

_ + irefn Gin +_ i Cfn_f

Ginvn

vinvn +

_

_

1 sLfn iLfn

iCfn

1 von sCfn

Kin

Kvo

Fig. 1.19 Compound balanced control strategy for ISOS inverter system

iref1 –irefn , which are related to magnitude difference. These signals are multiplied with iref to obtain the module adjusting signals isv1 –isvn with the same phase and different amplitudes, which are used to alter the amplitude of module output voltage, and thus input voltage sharing (IVS) and output voltage sharing (OVS) are achieved for ISOS inverter system. In this method, the feedback signal for every individual output voltage loop contains the information of other modules. So, all the modules are inevitably mutual coupling, which restricts the modularity of ISOS inverter system. In summary, for DC–AC inverter series–parallel combined system, existing literature has not analyzed the inherent relationship of the four types of series–parallel combination systems. So, Chap. 5 will analyze the relationship between the input and output powers of the above DC–AC inverter series–parallel combined system, and then put forward a general control strategy so as to form a set of systematic control theory for DC–AC inverter series–parallel combined system, and lay the theoretical foundation for the proposition of the specific control scheme of the four types of series–parallel combination systems. Based on the proposed general control strategy, Chaps. 6 and 7 will give specific control strategies for the ISOP and ISOS inverter systems, and put forward the distributed control based on the centralized control to realize the modularization of the systems. In addition, Chap. 8 will present an improved average current control strategy for IPOP inverter system to improve the output characteristics of both single module and the whole system. Moreover, in this improved method, the current sharing and the current-limiting function are remained unchanged compared with the conventional method.

1.5 Summary

voref

25

_ v of +

Gvo

1# Module

n −1 voref n

+_

Gvo1 Kvo

Kvo

iref isv1

iref1

Gi1

Ginv1

_ + vinv1

io1 1 iLf1 sLf1 +_ i Cf1

n

∑v

sCf1

oi

i=2

isvj

j# Module Gvoj

+_ Kvo

irefj

Gij

Ginvj

_ + vinvj

Ginvn

_ + vinvn

n

∑v

ioj+ 1 iLfj Z sLfj +_ + + Ld iCfj sCfj

vo

oi

i =1

i≠ j

isvn

n# Module +_

Gvon Kvo

irefn

n −1

Gin

ion 1 iLfn sLfn +_ iCfn

∑v

sCfn

oi

i =1

Fig. 1.20 Output-voltage cross-feedback control for ISOS inverter system

1.5 Summary This chapter briefly reviews the history and status of the development of power electronics. Power electronics system integration, standardization, and modularization are one of the effective ways to reduce the cost of producing power electronics products, shorten the development cycle, and improve the reliability. Standardized module series–parallel combination system is one of the significant constituent parts of power electronic system integration, and the research on its control technique to ensure the stability of the system operation and the power balance of each module is the primarily crucial issue. This chapter introduces the advantages and major categories of standardized module series–parallel combination system, and elaborates the control techniques for the series–parallel combination system with DC–DC converter and DC–AC inverter as the basic module, which will lay a foundation for the narration of the following chapters.

26

1 Introduction

References 1. Prince, D.C.: Mercury Arc Rectifier Phenomena. Trans. American Institute of Electrical Engineers, pp. 1064–1071 (1927) 2. Bose, B.K.: Power electronics—a technology review. Proc. IEEE 80(8), 1303–1334 (1992) 3. Newell, W.E.: Power electronics—emerging from Limbo. In: Proceedings of the IEEE Power Electronics Specialists Conference, pp. 6–12 (1973) 4. Bose, B.K.: Power electronics—an emerging technology. IEEE Trans. Ind. Electron. 36(2), 403–412 (1989) 5. Blaabjerg, F., Chen, Z., Kjaer, S.B.: Power electronics as efficient interface in dispersed power generation systems. IEEE Trans. Power Electron. 19(5), 1184–1194 (2004) 6. Emadi, A., Williamson, S.S., Khaligh, A.: Power electronics intensive solutions for advanced electric, hybrid electric, and fuel cell vehicular power systems. IEEE Trans. Power Electron. 21(3), 567–577 (2006) 7. Bose, B.K.: Power electronics and motor drives recent progress and perspective. IEEE Trans. Ind. Electron. 56(2), 581–588 (2009) 8. Blaabjerg, F., Ma, K., Yang, Y.: Power electronics—the key technology for renewable energy systems. In: Proceedings of the International Conference on Ecological Vehicles and Renewable Energies, pp. 1–11 (2014) 9. Lorenz, L.: Key power semiconductor devices and development trends. In: Proceedings of the International Conference on Electrical Machines and Systems, pp. 1137–1142 (2008) 10. Jacobus, D., Lee, F.C.: On a future for power electronics. IEEE J. Emerg. Sel. Top. Power Electron. 1(2), 59–72 (2013) 11. Kornegay, K.T.: Design issues in power electronic building block (PEBB) system integration. In: Proceedings of IEEE Computer Society Workshop on System Level Design, pp. 48–52 (1998) 12. Lee, F.C., Borojevic, D., Xing, K., Thandi, G., Zhu, H.: PEBBs and system integration at Virginia power electronics center, integrated power electronics module—a building block concept for system integration. In: Proceedings of the Center for Power Electronics Systems, pp. 1–4 (2000) 13. Reeves, R.: Inductor-capacitor hybrid. Proc. Inst. Electr. Eng. 122(11), 1323–1326 (1975) 14. Waffenschmidt, E., Ackermann, B., Ferreira, J.A.: Design method and material technologies for passives in printed circuit board embedded circuits. IEEE Trans. Power Electron. 20(3), 576–584 (2005) 15. Luo, S., Batarseh, I.: A review of distributed power systems part I: DC distributed power system. IEEE Aerosp. Electron. Syst. Mag. 20(8), 5–16 (2005) 16. Chen, W., Ruan, X., Yan, H., Tse, C.K.: DC/DC conversion systems consisting of multiple converter modules: stability, control and experimental verifications. IEEE Trans. Power Electron. 24(6), 1463–1474 (2009) 17. Kim, J., Choi, H., Cho, B.H.: A novel droop method for converter parallel operation. IEEE Trans. Power Electron. 17(1), 25–32 (2002) 18. Panov, Y., Jovanovi´c, M.M.: Stability and dynamic performance of current-sharing control for paralleled voltage regulator modules. IEEE Trans. Power Electron. 17(2), 172–179 (2002) 19. Lin, C., Chen, C.: Single-wire current-share paralleling of current-mode-controlled DC power supplies. IEEE Trans. Ind. Electron. 47(4), 780–786 (2000) 20. Huang, Y., Tse, C.K.: Circuit theoretic classification of parallel connected DC-DC converters. IEEE Trans. Circuits. Syst. 54(5), 1099–1108 (2007) 21. Du, Q., Qi, B., Wang, T., Zhang, T., Li, X.: A high-power input-parallel output-series buck and half-bridge converter and control methods. IEEE Trans. Power Electron. 27(6), 2703–2715 (2012) 22. Hu, X., Gong, C.: A high gain input-parallel output-series DC/DC converter with dual coupled inductors. IEEE Trans. Power Electron. 30(3), 1306–1317 (2015)

References

27

23. Giri, R., Choudhary, V., Ayyanar, R., Mohan, N.: Common-duty-ratio control of input-series connected modular DC-DC converters with active input voltage and load-current sharing. IEEE Trans. Ind. Appl. 42(4), 1101–1111 (2006) 24. Manias, S., Kostakis, G.: Modular DC–DC converter for high output voltage applications. IEE Electr. Power Appl. 140(2), 97–102 (1993) 25. Siri, K., Conner, K., Truong, C.: Uniform voltage distribution control for paralleled-input, series-output connected converters. In: Proceedings of the IEEE Aerospace Conference, pp. 1–11 (2005) 26. Siri, K., Willhoff, M., Conner, K.: Uniform voltage distribution control for series connected DC–DC converters. IEEE Trans. Power Electron. 22(4), 1269–1279 (2007) 27. Ruan, X., Chen, W., Cheng, L., Tse, C.K., Yan, H., Zhang, T.: Control strategy for input-seriesoutput-parallel converters. IEEE Trans. Ind. Electron. 56(4), 1174–1185 (2009) 28. Kim, J., You, J., Cho, B.H.: Modeling, control, and design of input-series-output-parallelconnected converter for high-speed-train power system. IEEE Trans. Ind. Electron. 48(3), 536–544 (2001) 29. Ayyanar, R., Giri, R., Mohan, N.: Active input-voltage and load-current sharing in inputseries and output-parallel connected modular DC-DC converters using dynamic input-voltage reference scheme. IEEE Trans. Power Electron. 19(6), 1462–1472 (2004) 30. Jang, Y., Jovanovic, M.M., Dillman, D., Li, S., Yang, C.: Input-voltage balancing of seriesconnected converters. In: Proceedings of the IEEE Applied Power Electronics Conference and Exposition, pp. 1153–1160 (2011) 31. Sha, D., Guo, Z., Liao, X.: Cross-feedback output-current-sharing control for input-seriesoutput-parallel modular DC–DC converters. IEEE Trans. Power Electron. 25(11), 2762–2771 (2010) 32. Giri, R., Ayyanar, R., Ledezma, E.: Input-series and output-series connected modular DC-DC converters with active input voltage and output voltage sharing. In: Proceedings of the IEEE Applied Power Electronics Conference and Exposition, pp. 1751–1756 (2004) 33. Merwe, J., Mouton, H.: An investigation of the natural balancing mechanisms of modular input-series-output-series DC-DC converters. In: Proceedings of the IEEE Energy Conversion Congress and Exposition, pp. 817–822 (2010) 34. Lu, Q., Yang, Z., Lin, S., Wang, S., Wang, C.: Research on voltage sharing for input-seriesoutput-series phase-shift full-bridge converters with common-duty-ratio. In: Proceedings of the Annual Conference of the IEEE Industrial Electronics Society, pp. 1548–1553 (2011) 35. Sha, D., Deng, K., Liao, X.: Duty cycle exchanging control for input-series-output-series connected two PS-FB DC-DC converters. IEEE Trans. Power Electron. 27(3), 1490–1501 (2012) 36. Sha, D., Guo, Z., Luo, T., Liao, X.: A general control strategy for input-series-output-series modular DC-DC converters. IEEE Trans. Power Electron. 29(7), 3766–3775 (2014) 37. Bekiarov, S.B., Emadi, A.: Uninterruptible power supplies: classification operation dynamics and control. In: Proceedings of the IEEE Applied Power Electronics Conference and Exposition, pp. 597–604 (2002) 38. Chen, X., Kang, Y., Chen, J.: Operation, control technique of parallel connected high power three-phase inverters. In: Proceedings of the IEEE International Power Electronics and Motion Control Conference, pp. 956–959 (2004) 39. Chen, J., Chu, C.: Combination voltage-controlled and current-controlled PWM inverter for UPS parallel operation. IEEE Trans. Power Electron. 10(5), 547–558 (1995) 40. Lee, W.C., Lee, T.K., Lee, S.H., Kim, K.H., Hyun, D.S., Suh, I.Y.: A master and slave control strategy for parallel operation of three-phase UPS systems with different ratings. In: Proceedings of the IEEE Applied Power Electronics Conference and Exposition, pp. 456–462 (2004) 41. Chiang, S.J., Lin, C.H., Yen, C.Y.: Current limitation control technique for parallel operation of UPS inverters. In: Proceedings of the IEEE Power Electronics Specialists Conference, pp. 1922–1926 (2004) 42. Perez-Ladron, G., Cardenas, V., Espinosa, G.: Analysis and implementation of a master-slave control based on a passivity approach for parallel inverters operation. In: Proceedings of the IEEE International Power Electronics Conference, pp. 1–5 (2006)

28

1 Introduction

43. Xing, Y., Huang, L., Yan, Y.: Redundant parallel control for current regulated inverters with instantaneous current sharing. In: Proceedings of the IEEE Power Electronics Specialists Conference, pp. 1438–1442 (2003) 44. Fang, T., Ruan, X., Xiao, L., Liu, A.: An improved distributed control strategy for parallel inverters. In: Proceedings of the IEEE Power Electronics Specialists Conference, pp. 3500–3505 (2008) 45. Tuladhar, A.: Advanced control techniques for parallel inverter operation without control interconnections. Ph.D. Dissertation, The University of British Columbia, Vancouver, Canada (2000) 46. Lin, X., Feng, F., Duan, S., Kang, Y., Chen, J.: The droop characteristic decoupling control of parallel connected UPS with no control interconnection. In: Proceedings of the IEEE International Electric Machines and Drives Conference, pp. 1777–1780 (2003) 47. De Brabandere, K., Bolsens, B., Van den Keybus, J., Woyte, A., Driesen, J., Belmans, R.: A voltage and frequency droop control method for parallel inverters. IEEE Trans. Power Electron. 22(4), 1107–1115 (2007) 48. Yao, W., Chen, M., Gao, M., Qian, Z.: A wireless load sharing controller to improve the performance of parallel-connected inverters. In: Proceedings of the IEEE Applied Power Electronics Conference and Exposition, pp. 1628–1631 (2008) 49. Guerrero, J., Vicuña, L., Matas, J., Castilla, M., Miret, J.: Output impedance design of parallelconnected UPS inverters with wireless load-sharing control. IEEE Trans. Ind. Electron. 52(4), 1126–1135 (2005) 50. Song, C., Zhao, R., Lin, W., Zeng, Z.: A novel control strategy for input-parallel-output-series inverter system. In: Proceedings of the IEEE International Conference on Electrical Machines and Systems, pp. 1–5 (2011) 51. Chen, W., Zhuang, K., Ruan, X.: Input-series and output-parallel connected inverter system for high input voltage applications. IEEE Trans. Power Electron. 24(9), 2127–2137 (2009) 52. Sha, D., Deng, K., Guo, Z., Liao, X.: Control strategy for input-series–output-parallel highfrequency ac link inverters. IEEE Trans. Ind. Electron. 59(11), 4101–4111 (2012) 53. Fang, T., Ruan, X., Tse, C.K.: Control strategy to achieve input and output voltage sharing for input-series-output-series connected inverters systems. IEEE Trans. Power Electron. 25(6), 1585–1596 (2010) 54. Sha, D., Xu, G., Liao, X.: Control strategy for input-series-output-series high-frequency ac-link inverters. IEEE Trans. Power Electron. 28(11), 5283–5292 (2012)

Chapter 2

A General Control Strategy for DC–DC Series–Parallel Power Conversion Systems

Abstract The inherent relationships between the sharing of input voltages/currents and the sharing of output voltages/currents are analyzed systematically from the view of power conservation in this chapter, and the stability of various possible control strategies is discussed. Based on the analysis, a general control strategy is proposed for the DC–DC series–parallel power conversion systems, which not only achieves equilibrium between the constituent modules but also decouples the output voltage control loop and the sharing control loop. Furthermore, a modularization architecture for the DC–DC series–parallel power conversion systems is proposed, achieving full modularization where all the modules are self-contained and standardized, and no extra control is needed to achieve sharing of voltage and/or current at the input and output sides. Experimental prototypes are built and tested to validate the proposed general control strategy and the modularization architecture. Keywords Sharing of input voltages/current · Sharing of output voltages/current Stability · General control strategy · Modularization architecture In Chap. 1, the four basic connection architectures (IPOP, IPOS, ISOP, and ISOS) of DC–DC series–parallel power conversion systems were introduced. In order to ensure proper operation of DC–DC series–parallel power conversion systems, the equilibrium between the constituent modules must be achieved, i.e., proper input voltage sharing (IVS) and output voltage sharing (OVS) among the constituent modules when they are connected in series at the input and output sides, respectively, and proper input current sharing (ICS) and output current sharing (OCS) among the constituent modules when they are connected in parallel at the input and output sides, respectively. Previous studies have addressed the issue of sharing of the input and/or output voltage and/or current. However, these prior studies focus on one or two of the four connection architectures and do not reveal the inherent relationships between the four connection architectures. In this chapter, the inherent relationships between the sharing of input voltages/currents and the sharing of output voltages/currents are analyzed systematically from the view of power conservation, and the stability of various possible control strategies are discussed. Based on the analysis, a general control strategy is proposed for the DC–DC series–parallel power conversion systems, which © Springer Nature Singapore Pte Ltd. and Science Press, Beijing, China 2019 X. Ruan et al., Control of Series-Parallel Conversion Systems, CPSS Power Electronics Series, https://doi.org/10.1007/978-981-13-2760-5_2

29

30

2 A General Control Strategy for DC–DC Series–Parallel Power …

not only achieves equilibrium between the constituent modules but also decouples the output voltage control loop and the sharing control loop. Furthermore, a modularization architecture for the DC–DC series–parallel power conversion systems is proposed, achieving full modularization where all the modules are self-contained and standardized, and no extra control is needed to achieve sharing of voltage and/or current at the input and output sides. Experimental prototypes were built and tested to validate the proposed general control strategy and the modularization architecture.

2.1 Possible Control Strategies for DC–DC Series–Parallel Power Conversion Systems For input-parallel-connected systems (IPOP and IPOS), the input voltages of modules are equal, therefore, we need to make the module output currents equal for IPOP system and the output module voltages equal for IPOS system, i.e., OCS and OVS should be ensured for IPOP and IPOS systems, respectively. For input-seriesconnected systems (ISOP and ISOS), we not only have to ensure OCS and OVS for ISOP and ISOS systems, respectively, but also ensure IVS for both ISOP and ISOS systems. In other words, both IVS and OCS are needed for ISOP system, and both IVS and OVS are needed for ISOS system. As discussed before, we can see that if the modules are connected in parallel at the input sides, only OCS or OVS is needed, and if the modules are connected in series at the input sides, both IVS and OCS or OVS are needed. Two questions are of fundamental importance: (1) for IPOP and IPOS systems, as all modules share the same input voltage, would ICS imply OCS or OVS so that we only need to direct control to ensure ICS? and (2) for ISOP and ISOS systems, what is the relationship between IVS and OCS or OVS? Figure 2.1 shows the four connection architectures of DC–DC series–parallel power conversion systems. For the j-th (j 1, 2, …, n, n is the number of modules) module, V in_j and I in_j denote the input voltage and the input current, respectively; V o_j and I o_j denote the output voltage and the output current, respectively; I cd_j and I cf_j denote the input and output filter capacitor currents, respectively; and I cin_j and I co_j denote the input and output currents of the power stage, not including the input and output filter capacitor currents, respectively. At steady state, since both the input voltage and the output voltage are constant, the average values of both the input filter capacitor current and the output filter capacitor current are zero, we have, Iin_ j Icin_ j ( j 1, 2, . . . , n)

(2.1)

Ico_ j Io_ j ( j 1, 2, . . . , n)

(2.2)

Then, by power conservation, we have,

2.1 Possible Control Strategies for DC–DC Series–Parallel Power … Iin

Iin_1

Icin_1

Ico_1

+ Icd_1 Vin_1 _ Cd1

Vin

Io_1 Icf_1

DC-DC Converter

Cf1

Io

Iin

+

+ Vo_1 _

Vo _

Iin_1

Icin_1

Ico_1

+ Icd_1 Vin_1 _ Cd1

Vin

31

DC-DC Converter

Icin_2

Ico_2

+ Icd_2 Vin_2 _ Cd2

Io_2 Icf_2

DC-DC Converter

Cf2

Iin_2

Icin_2

Ico_2

+ Icd_2 Vin_2 _ Cd2

+ Vo_2 _

DC-DC Converter

Icin_n

Ico_n

+ Icd_n Vin_n _ Cdn

DC-DC Converter

Io_n Icf_n

Cfn

Iin_n

Icin_n

Icin_1

+ Icd_1 Vin_1 _ Cd1

DC-DC Converter

Io_1 Icf_1

Cf1

+ Vo_1 _

Io

Iin

+

Iin_1

Icin_1

+ Icd_1 Vin_1 _ Cd1

Vin

Icin_2

+ Icd_2 Vin_2 _ Cd2

Ico_2 DC-DC Converter

Vo _

DC-DC Converter

Icin_n

+ Icd_n Vin_n _ Cdn

Icf_2 Cf2

Ico_n DC-DC Converter

+ Vo_2 _

Iin_2

Vin

Icin_2

+ Vo_n _

_

Cfn

Cf1

Io

+ Vo_1 _

+

+ Icd_2 Vin_2 _ Cd2

DC-DC Converter

Io_2 Icf_2

Cf2

+ Vo_2 _

Vo

Module 2

Io_n Icf_n

Io_1 Icf_1

Ico_2

+ Vo_n _

Iin_n

Icin_n

+ Icd_n Vin_n _ Cdn

Ico_n DC-DC Converter

Module n

(c) ISOP

Cfn

Module 1

Io_2

Module 2

Iin_n

Vo

Io_n Icf_n

Ico_1

Module 1

Iin_2

+ Vo_2 _

(b) IPOS Ico_1

DC-DC Converter

Cf2

Module n

(a) IPOP Iin_1

Io_2 Icf_2

Ico_n

+ Icd_n Vin_n _ Cdn

+ Vo_n _

Module n

Iin

+

Module 2

Module 2

Iin_n

Cf1

Io

+ Vo_1 _

Module 1

Module 1

Iin_2

Io_1 Icf_1

Io_n Icf_n

Cfn

+ Vo_n _

_

Module n

(d) ISOS

Fig. 2.1 Four connection architectures of DC–DC series–parallel power conversion systems

⎧ ⎪ Vo_1 · Io_1 Po1 Pin1 · η1 Vin_1 · Iin_1 · η1 ⎪ ⎪ ⎪ ⎪ ⎨ Vo_2 · Io_2 Po2 Pin2 · η2 Vin_2 · Iin_2 · η2 .. ⎪ ⎪ . ⎪ ⎪ ⎪ ⎩V · I P P ·η V o_n o_n on inn n in_n · Iin_n · ηn

(2.3)

where Po1 , Po2 , …, Pon are the output power of the modules, Pin1 , Pin2 , …, Pinn are the input power of the modules, and η1 , η2 , …, ηn are the conversion efficiencies of the modules.

32

2 A General Control Strategy for DC–DC Series–Parallel Power …

2.1.1 IPOP Systems For IPOP systems, the input voltages of the modules are equal, and the output voltages of the modules are equal, i.e., V in_1 V in_2 … V in_n and V o_1 V o_2 … V o_n . If the module input currents are equal, i.e., I in_1 I in_2 … I in_n , then the module input power are equal, i.e., Pin1 Pin2 … Pinn . According to (2.3), we can get, Io_1 /η1 Io_2 /η2 · · · Io_n /ηn

(2.4)

Equation (2.4) illustrates that the difference among the module output currents is dependent on the difference of the module efficiencies. In practice, as the topologies of the modules are identical, the mismatches in the power stages, such as the power devices, transformers, inductors, and capacitors, are not so large, so that the difference among the module efficiencies is very little and it will be ignored in the following. Thus, the module output currents will be nearly evenly shared for IPOP systems. In other words, achieving ICS implies OCS for IPOP systems.

2.1.2 IPOS Systems For IPOS systems, the input voltages of the modules are equal, and the output currents of the modules are equal, i.e., V in_1 V in_2 … V in_n and I o_1 I o_2 … I o_n . If the module input currents are equal, i.e., I in_1 I in_2 … I in_n , then the module input power are equal, i.e., Pin1 Pin2 … Pinn . According to (2.3), we can get, Vo_1 /η1 Vo_2 /η2 · · · Vo_n /ηn

(2.5)

Equation (2.5) illustrates that the difference among the module output voltages is dependent on the difference of the module efficiencies. Thus, the module output voltages will be nearly evenly shared for IPOS systems. In other words, achieving ICS implies OVS for IPOS systems.

2.1.3 ISOP Systems For ISOP systems, the input currents of the modules are equal, and the output voltages of the modules are equal, i.e., I in_1 I in_2 … I in_n and V o_1 V o_2 … V o_n . If the module output currents are equal, i.e., I o_1 I o_2 … I o_n , then the module output power are equal, i.e., Po1 Po2 … Pon . According to (2.3), we have, Vin_1 · η1 Vin_2 · η2 · · · Vin_n · ηn

(2.6)

2.1 Possible Control Strategies for DC–DC Series–Parallel Power …

33

Equation (2.6) illustrates that the difference in the input voltages among the modules is dependent on the difference of the module efficiencies if the module output currents are equal. If IVS control strategy is used to achieve V in_1 V in_2 … V in_n V in /n, then the module input power are equal, i.e., Pin1 Pin2 … Pinn . According to (2.3), we have, Io_1 /η1 Io_2 /η2 · · · Io_n /ηn

(2.7)

Equation (2.7) illustrates that the difference in the output currents among the modules is dependent on the difference of the module efficiencies if the module input voltages are equal. In summary, for ISOP systems, if the module input voltages are equal, the module output currents will be nearly equal, and on the other hand, if the module output currents are equal, the module input voltages will be nearly equal. So, we can conclude that once IVS is achieved, OCS is nearly achieved automatically and vice versa. That is to say, we need to ensure either IVS or OCS.

2.1.4 ISOS Systems For ISOS systems, the input currents of the modules are equal, and the output currents of the modules are equal, i.e., I in_1 I in_2 … I in_n and I o_1 I o_2 … I o_n . If the module output voltages are equal, i.e., V o_1 V o_2 … V o_n , then the module output power are equal, i.e., Po1 Po2 … Pon . According to (2.3), we have, Vin_1 · η1 Vin_2 · η2 · · · Vin_n · ηn

(2.8)

Equation (2.8) illustrates that the difference in the input voltages among the modules is dependent on the difference of the module efficiencies if the module output voltages are equal. If IVS control strategy is used to achieve V in_1 V in_2 … V in_n V in /n, then the module input power are equal, i.e., Pin1 Pin2 … Pinn . According to (2.3), we have, Vo_1 /η1 Vo_2 /η2 · · · Vo_n /ηn

(2.9)

Equation (2.9) illustrates that the difference in the output voltages among the modules is dependent on the difference of the module efficiencies if the module input voltages are equal. In summary, for ISOS systems, we can see that if the module input voltages are equal, the module output voltages will be nearly equal, and on the other hand, if the module output voltages are equal, the module input voltages will be nearly

34

2 A General Control Strategy for DC–DC Series–Parallel Power …

Table 2.1 Relationship between the input sides and output sides Architecture Input side Output side IPOP

Input voltages

Input currents

Output voltages

Output currents

Equal by connection

Equal sharing achievable by control

Equal by connection

Equal sharing achievable by control Equal by connection

IPOS

ISOP

Equal sharing achievable by control

Equal by connection

ISOS

Equal sharing achievable by control Equal by connection Equal sharing achievable by control

Equal sharing achievable by control Equal by connection

equal. Thus, we can conclude that once IVS is achieved, OVS is nearly achieved automatically and vice versa. That is to say, we need to ensure either IVS or OVS. Based on the above analysis, we have the following conclusions if the module efficiencies are the same: (1) (2) (3) (4)

for IPOP systems, achieving ICS implies OCS, for IPOS systems, achieving ICS implies OVS, for ISOP systems, achieving IVS implies OCS and vice versa, for ISOS systems, achieving IVS implies OVS and vice versa.

From the earlier discussion, we clearly see that for any of the four connection architectures, proper sharing of all voltages or currents can be nearly achieved by directing control efforts to ensure the proper sharing of either the input voltage or current, as shown in Table 2.1.

2.2 Stability of the Control Strategies In this section, we examine the stability of the system when the input voltages or currents are controlled to achieve IVS or ICS, and when the output voltages or currents are controlled to achieve OVS or OCS. In our analysis, we assume that the system’s input voltage, output voltage, and load current are kept unchanged when perturbations are imposed on the input voltage/current and output voltage/current of the modules.

2.2 Stability of the Control Strategies

35

2.2.1 Controlling at Input Side Regardless of the connection type at the input side, when ICS (for input-parallelconnected systems) or IVS (for input-series-connected systems) is achieved, each converter module behaves as a constant power source as seen at the output side, as shown in Fig. 2.2, and the module input powers are equal, i.e., Vo_ j · Ico_ j Pin j · η const ( j 1, 2, . . . , n)

(2.10)

The relationship between the currents at the output side is Icf_ j Ico_ j − Io_ j ( j 1, 2, . . . , n)

(2.11)

For the case where the output sides are connected in parallel, as shown in Fig. 2.2a, the output voltages of all modules are equal, and from (2.10), the output currents of all modules are identical. Thus, no perturbation occurs, i.e., the system is stable. We now examine the case where the output sides are connected in series, as shown in Fig. 2.2b. We assume that a perturbation occurs in the output filter capacitor voltages and the output voltages of some modules, e.g., V o_1 increases and V o_2 decreases, while the output voltages of other modules are unchanged. According to (2.10) and (2.11), I co_1 and I cf_1 vary as shown in Fig. 2.3. Suppose point A is the equilibrium operating point, when V o_1 increases under perturbation, the operating point moves to point B. At point B, I co_1 is smaller than I o and V o_1 is higher than V o /n. Hence, the filter capacitor current I cf_1 is negative, causing C f1 to be discharged Ico_1 Constant power source (Module 1)

Constant power source (Module 2)

Constant power source (Module n)

Io_1 Icf_1

Cf1

+ Vo_1 _

Ico_2

Io_2 Icf_2

Cf2

+ Vo_2 _

Ico_n

Io_n Icf_n

Cfn

Io

+ Vo_n _

(a) Output sides connected in parallel

Vo _

Io_1

Ico_1

+ Constant power source (Module 1)

Constant power source (Module 2)

Icf_1 Cf1

+ Vo_1 _

Ico_2

Io_2 Icf_2

Cf2

+ Vo_2 _

Ico_n Constant power source (Module n)

Io

+ Vo_n _

(b) Output sides connected in series

Fig. 2.2 Equivalent schematic diagram for the input side control strategies

Vo

Io_n Icf_n

Cfn

+

_

36

2 A General Control Strategy for DC–DC Series–Parallel Power …

Fig. 2.3 Currents and voltage of the output side of module 1

Ico_1 (Pin = const) A

Io

0

B Vo n

Vo_1 Icf_1

to supply the deficit current, and as a result, V o_1 decreases, and the operating point returns to point A. Thus, V o_1 returns to the equilibrium voltage. Similarly, I co_2 increases as V o_2 decreases, and as it becomes larger than I o , the excessive current charges C f2 to make V o_2 increase and return to the equilibrium voltage. Therefore, any control strategies intending to achieve ICS and IVS for input-parallel-connected systems (including IPOP and IPOS) and input-series-connected systems (including ISOP and ISOS), respectively, are stable.

2.2.2 Controlling at Output Side Regardless of the type of connection at the output side, when OCS (for outputparallel-connected systems) or OVS (for output-series-connected systems) is achieved, each of the converter modules behaves as a constant power sink as seen from the input side, as shown in Fig. 2.4, and the output powers of all modules are equal, i.e., Vin_ j · Icin_ j Po j /η const ( j 1, 2, . . . , n)

(2.12)

The relationship between the currents at the input side is Icd_ j Iin_ j − Icin_ j ( j 1, 2, . . . , n)

(2.13)

For the case where the input sides are connected in parallel, as shown in Fig. 2.4a, the input voltages of all modules are equal, and the input currents are thus identical. Hence, no perturbation occurs, i.e., the system is stable. To examine the case where the input sides are connected in series, as shown in Fig. 2.4b, we again consider perturbing the input filter capacitor voltages, the input voltages of some modules, e.g., increasing V in_1 and decreasing V in_2 , while keeping the input voltages of other modules unchanged. By virtue of constant power, I cin_1 will decrease and I cin_2 will increase, as illustrated in (2.13). According to (2.12) and

2.2 Stability of the Control Strategies

Iin Vin

Iin_1

Icin_1

+ Icd_1 Vin_1 Cd1 _ Iin_2

Iin

Iin_2 Constant power sink (Module 2)

Icin_n

+ Icd_n Vin_n Cdn _

Iin_1

Vin

(a) Input sides connected in parallel

Constant power sink (Module 1)

Icin_2

+ Icd_2 Vin_2 Cd2 _

Iin_n Constant power sink (Module n)

Icin_1

+ Icd_1 Vin_1 Cd1 _

Constant power sink (Module 1)

Icin_2

+ Icd_2 Vin_2 Cd2 _

Iin_n

37

Constant power sink (Module 2)

Icin_n

+ Icd_n Vin_n Cdn _

Constant power sink (Module n)

(b) Input sides connected in series

Fig. 2.4 Equivalent schematic diagram for the output side control strategies

(2.13), I cin_1 and I cd_1 can be shown to vary as given in Fig. 2.5. Suppose point A is the equilibrium operating point when V in_1 is perturbed to increase, the operating point moves to point B. At point B, I cin_1 is smaller than the system input current I in , and it can be derived from (2.13) that the input filter capacitor current I cd_1 is positive, causing C d1 to be charged up and its voltage increased. Thus, a positive feedback loop is formed and the operating point continues to move to the right and away from point A. So, the equilibrium point cannot be resumed. Similarly, if I cin_2 is larger than I in causing C d2 to be discharged by the unbalanced current, then V in_2 decreases and fails to return to the equilibrium point. If the module employs a buck-derived converter when V in_2 continues to decrease, it will not be high enough to provide the desired voltage, and the output current thus decreases, resulting in unbalanced output currents of the modules. Moreover, if the module employs a boost-derived or buck/boost-derived converter, V in_2 will continue to decrease until it is zero and provides no power to the load, again resulting in unbalanced output currents of the modules. From the above analysis, we can conclude that any control strategies intending to achieve OCS and OVS for ISOP and ISOS systems, respectively, are unstable. In addition to the above mentioned qualitative analysis, the tool of small-signal model can be used to prove that any control strategies intending to achieve OCS and OVS for ISOP and ISOS systems, respectively, are unstable. Referring to Fig. 2.4b, each converter module presents constant power sink characteristics as seen from the input side of power stage when OCS (for ISOP) or OVS (for ISOS) is achieved, and the input impedance of each converter module can be modeled as an equivalent negative resistance [1], i.e.,

38

2 A General Control Strategy for DC–DC Series–Parallel Power …

Fig. 2.5 Currents and voltage of the input side of module 1

Icin_1 ( Po = const) Iin

0

Rne j

2 Vin_ vˆin_ j Vin_ j j − − ˆi cin_ j Icin_ j Vin_ j · Icin_ j

A

B

Vin n

( j 1, 2, . . . , n)

Icd_1

Vin_1

(2.14)

At steady state, we have, Vin_ j Icin_ j η Po /n ( j 1, 2, . . . , n)

(2.15)

where Po is the output power of ISOP or ISOS system. Combining (2.14) and (2.15), yields, Rne j

2 Vin_ vˆin_ j j ·η·n − ( j 1, 2, . . . , n) Po iˆcin_ j

(2.16)

Figure 2.6 shows the small-signal model of input-series-connected systems employing OVS/OCS control strategies by replacing each module in Fig. 2.4b with a negative resistance Rnej (j 1, 2, …, n). The total input impedance of each module at the input terminal is, Z j Rne j //(1/sCd j )

Rne j 1 + s Rne j Cd j

( j 1, 2, . . . , n)

(2.17)

The individual input voltage perturbation caused by the system input voltage perturbation vˆin can be expressed as vˆin_ j

Zj · vˆin ( j 1, 2, . . . , n) n Zm

(2.18)

m1

According to (2.18), the difference of input voltage perturbations between module j and module k is vˆ jk vˆin_ j − vˆin_k

Z j − Zk vˆin ( j, k 1, 2, . . . , n; j k) n Zm m1

(2.19)

2.2 Stability of the Control Strategies

39

Fig. 2.6 Small-signal model of input-series-connected systems employing OVS/OCS control strategies

iˆcin_1

iˆin

Z1

+ ˆvin_1 _ Cd1

Rne1

iˆcin_2

Z2 vˆ in

+ vˆ in_2 _ Cd2

Rne2

iˆcin_n

Zn

+ vˆ in_n C _ dn

Rnen

Equation (2.19) can also be expressed as the response in the difference of input voltage perturbations of two modules due to perturbation in the system input voltage, i.e., vˆin_ j − vˆin_k Z j − Zk vˆ jk n ( j, k 1, 2, . . . , n; j k) vˆin vˆin Zm

(2.20)

m1

For simplicity, letting n 2, substitution of (2.17) into (2.20), yields, vˆ12 (Rne1 − Rne2 ) + s Rne1 Rne2 (Cd2 − Cd1 ) vˆin (Rne1 + Rne2 ) + s Rne1 Rne2 (Cd2 + Cd1 )

(2.21)

At steady state, V in_1 V in_2 V in / 2, substitution the equation into (2.16), we V 2η get Rne1 Rne2 − 2Pin o , then (2.21) can be rewritten as s(Cd1 − Cd2 )Vin2 η vˆ12 vˆin 4Po − s(Cd1 + Cd2 )Vin2 η

(2.22)

In practice, C d1 is impossible equal to C d2 precisely, it can be found from (2.22) that the transfer function of vˆ12 /vˆin has a right-half-plane pole, which implies unstable operation. Therefore, any control strategies intending to achieve OCS and OVS for ISOP and ISOS systems, respectively, are unstable, and the root cause is the negative resistance characteristics of the module’s input impedance with the control strategies to achieve OCS and OVS. From the foregoing discussion, we can draw the following conclusions: (1) for the input-parallel-connected systems, we can directly control efforts to achieve ICS that will automatically ensure OCS and OVS for IPOP and IPOS systems, respectively. Alternatively, we can directly control the output currents to achieve OCS for IPOP systems, or directly control the output voltages to achieve OVS for IPOS systems,

40

2 A General Control Strategy for DC–DC Series–Parallel Power …

(2) for the input-series-connected systems, we should control the input voltages to achieve IVS, and OCS or OVS will be achieved automatically. Any attempt to control the output currents or voltages to achieve OCS or OVS will not be successful because the control strategy is inherently unstable.

2.3 General Control Strategy for DC–DC Series–Parallel Power Conversion Systems In the following analysis, buck-derived converters with PWM control are taken as the basic modules (such as full-bridge converter, half-bridge converter, forward converter). Regardless of the type of the connection architectures, at least two control loops are needed, one is to regulate the system output voltage and the other is to control the sharing of input voltages/currents or output voltages/currents. Basically, the two control loops are exercised by varying the duty cycles of the modules. However, varying the duty cycles may affect both the output voltage of the system and the sharing of the input/output voltages/currents of the modules. Thus, the two control loops may be coupled, resulting in difficulties in the design of the loops. Here, we wish to propose a control strategy that decouples the two control loops. Referring to Fig. 2.1, and assuming the modules are buck-derived converters with transformer’s turns ratio being one, for the j-th module, the relationship between the input voltage vin_j and the output voltage vo_j is, vin_ j · dy_ j vo_ j

(2.23)

where d y_j is the duty cycle of the j-th module, and is composed of two components, i.e., dy_ j dvo_ j + dsh_ j

(2.24)

where d vo_j is produced by the system output voltage control loop and d sh_j is produced by the loop controlling the sharing of input voltages/currents or output voltages/currents. Perturbation and linearization of (2.23) leads to,

(2.25) Vin_ j + vˆin_ j · Dy + dˆvo_ j + dˆsh_ j Vo_ j + vˆo_ j where Dy , V in_j , and V o_j are the quiescent values, and dˆvo_ j , dˆsh_ j , vˆin_ j and vˆo_ j are the superimposed small AC perturbations. Neglecting the second-order nonlinear terms and noting that the DC terms on both sides of the equation are equal, we retain only the first-order AC terms on both sides of the equation, i.e.,

2.3 General Control Strategy for DC–DC Series–Parallel Power …

41

Vin_ j dˆvo_ j + Vin_ j dˆsh_ j + Dy vˆin_ j vˆo_ j

(2.26)

According to (2.26), we have, n

Vin_ j dˆvo_ j +

j1

n

Vin_ j dˆsh_ j + Dy ·

j1

n j1

vˆin_ j

n

vˆo_ j

(2.27)

j1

For input-parallel-connected systems, V in_j V in , and for input-series-connected system, V in_j V in /n. For output-parallel-connected systems, vˆo_ j vˆo , and for output-series-connected systems, vˆo_ j vˆo /n. Hence, (2.27) can be rewritten as n n Vin ˆ Vin ˆ m1 · dvo_ j + m 1 · dsh_ j + Dy · m 1 · vˆin m 2 · vˆo n j1 n j1

(2.28)

where m1 n and 1 for the input-parallel-connected and input-series-connected systems, respectively, and m2 n and 1 for the output-parallel-connected and outputseries-connected systems, respectively. If we have n

dˆsh_ j 0

(2.29)

j1

Equation (2.28) can be rewritten as m1 ·

n Vin ˆ dvo_ j + Dy · m 1 · vˆin m 2 · vˆo n j1

(2.30)

Equation (2.30) shows that if the sum of the duty cycle variations produced by the sharing control loops is zero, the system output voltage is only influenced by input voltage perturbation and the duty cycle perturbation produced by the system output voltage control loop. That is to say, the loop that controls the sharing of input voltages/currents or output voltages/currents has no effect on the system output voltage control loop. Thus, if condition (2.29) is satisfied, the two control loops will not interfere with each other. Despite this conclusion being drawn from a specific derivation process for buck-derived converters, it is also applicable to boost-derived or buck/boost-derived ones. According to (2.29), we have, dˆsh_n −

n−1

dˆsh_ j

(2.31)

j1

In practice, the duty cycle signal is produced by the intersection between the output signal of the regulator and saw tooth signal V RAMP , i.e.,

42

2 A General Control Strategy for DC–DC Series–Parallel Power …

vˆsh_EA j dˆsh_ j VRAMP

( j 1, 2, . . . , n)

(2.32)

where vˆsh_EA j is the output perturbation signal of the regulator for controlling the sharing of input voltages/currents or output voltages/currents. Combining (2.31) and (2.32), we have, vˆsh_EAn −

n−1

vˆsh_EA j

(2.33)

j1

According to (2.33), a general sharing control strategy for DC–DC series–parallel power conversion systems is proposed, as shown in Fig. 2.7. The system output voltage control loop ensures the desired system output voltage. vo_EA is the control signal derived from the output voltage regulator, which is used to produce common duty cycle signal d vo_j (j 1, 2, …, n) by intersecting with V RAMP . Input/output variables sharing control loops achieve the equilibrium among the constituent modules. The control signal V sh_g represents the average value of the total input/output voltages/currents, which needs to be shared among all the modules, and V sh_fj (j 1, 2, …, n) represents the sensed input or output variable of each module. For example, for input-series-connected systems, V sh_g is V in /n, and V sh_fj is the module input voltage. The representation of V sh_g and V sh_fj are shown in Table 2.2. Also, vsh_EAj (j 1, 2, …, n − 1) represents the control signal derived from the sharing control loop. For the first n − 1 modules, vsh_EAj is sent to the inverting input of PWM comparator to obtain the duty cycle signal d sh_j , for the module n, the sum of the first n − 1 vsh_EAj , which is denoted by vsh_EAn , is sent to the inverting input of PWM comparator to obtain the duty cycle signal d sh_n . It can be found that the proposed control strategy decouples the system output voltage regulating loop and input/output variables sharing loops. For the input-series-connected systems, the final duty cycle of module j is obtained by subtracting d sh_j from d vo_j (j 1, 2, …, n), while for input-parallel-connected systems, the final duty cycle of module j is the sum of d sh_j and d vo_j (j 1, 2, …, n).

Table 2.2 Representation of V sh_fj and V sh_g in four architectures Architecture

V sh_fj (j 1, 2, …, n − 1)

V sh_g

IPOP

I cin_j or I co_j

I in /n or I o /n

IPOS

I cin_j or V o_j

I in /n or V o /n

ISOP

V in_j

V in /n

ISOS

V in_j

V in /n

2.4 Modularization Architecture for DC–DC Series–Parallel Power …

Vsh_g

_ Vsh_f1 +

/+

Gsh

+

vsh_EA1 _ +

+

+

Input/output variables sharing control loops

Drive circuit 2

Gvo

vof Output voltage loop

vo_EA

+

Comparator (n−1) − RS Flip-flop +

VRAMP (n−1)

/+ vsh_EAn

_

Comparator 2 − RS Flip-flop +

VRAMP 2

/+

Gsh

+

Drive circuit 1

Vsh_f(n−1)

vsh_EA(n−1)

Voref

− +

RS Flip-flop

VRAMP 1

/+

vsh_EA2

+

Comparator 1

Vsh_f2 Gsh

_

43

Comparator n − + VRAMP n

RS Flip-flop

Drive circuit (n−1)

Drive circuit n

ISOP and ISOS /+ + IPOP and IPOS

Fig. 2.7 General sharing control strategy for DC–DC series–parallel power conversion systems

2.4 Modularization Architecture for DC–DC Series–Parallel Power Conversion Systems Among all the four series–parallel architectures, only IPOP has been successfully implemented in practice and commercially available in modular forms. In some previous designs of modular systems, e.g., those reported in [2–5], dedicated control methods are used for the constituent modules and the entire systems are not separable into standardized modules. In this section, the modularization architectures are proposed for IPOS, ISOP, and ISOS systems. The main features are: (a) the basic converter module has its own power stage and controller, and can operate in standalone mode; (b) no external controller is needed to achieve sharing of input/output voltages/currents; and (c) a relatively small number of wires are needed to connect up the modules to form the entire system. To achieve modularization, all power stages as well as controllers of all the modules should be identical. Hence, the sharing control signal vsh_EAn for the module n can be simplified as

44

2 A General Control Strategy for DC–DC Series–Parallel Power …

vsh_EAn −

n−1

vsh_EA j −

j1

n−1

Vsh_g − Vsh_f j G sh

j1

−G sh (n − 1)Vsh_g − (Vsh_f1 + · · · + Vsh_f(n−1) )

(2.34)

The average sharing control signal V sh_g is Vsh_g (Vsh_f1 + Vsh_f2 + · · · + Vsh_f(n−1) + Vsh_fn )/n

(2.35)

Substitution of (2.35) into (2.34), we have,

vsh_EAn −G sh (n − 1)Vsh_g − (n · Vsh_g − Vsh_fn ) G sh (Vsh_g − Vsh_fn ) (2.36) Based on (2.36), the control strategy depicted earlier in Fig. 2.7 can be redeveloped as shown in Fig. 2.8. Since the individual modules can be operated in stand-alone mode, each of them should have its own output voltage control loop. However, in practice, mismatches

Vsh_g

+

_ Vsh_f1

/+

Gsh

+

vsh_EA1

/+

Gsh

vsh_EA2 Input/output variables sharing control loops _

+

vsh_EA(n−1) _ Vsh_fn

+

Gvo vof Output voltage loop _

Drive circuit 2

Comparator (n−1)

VRAMP (n−1)

/+

Gsh

vsh_EA(n)

Voref

Comparator 2 − RS Flip-flop +

VRAMP 2

/+

Gsh

+

+

Vsh_f(n−1)

+

Drive circuit 1

VRAMP 1

_ Vsh_f2 +

Comparator 1 − RS Flip-flop +

+

− +

RS Flip-flop

Comparator n − RS Flip-flop +

VRAMP n

Drive circuit (n−1)

Drive circuit n

vo_EA ISOP and ISOS /+ + IPOP and IPOS

Fig. 2.8 Simplification of the general sharing control strategy for DC–DC series–parallel power conversion systems

2.4 Modularization Architecture for DC–DC Series–Parallel Power …

45

in the voltage references and output voltage sampling factors of the modules are inevitable. Thus, when the modules are connected to form a series–parallel power conversion system, we need to employ signal diodes to ensure only one output voltage control loop is active and regulates all the outputs of all modules. The final proposed modularization architecture is shown in Fig. 2.9. In Fig. 2.9, each dashed line block represents a self-contained module, which can operate in standalone mode. Taking module 1 as the example, when it is standalone, we have V sh_f1 V sh_g , i.e., vsh_EA1 is constant and the output voltage is regulated by its own output voltage control loop. Only two buses, namely, the input/output variables sharing bus and the common duty cycle bus, are needed when multiple modules are connected in one of the series–parallel architectures, and no external controller is required. It can be seen that the full modularization architecture for DC–DC series–parallel power conversion systems have the following advantages: (1) all modules are self-contained and standardized, (2) no extra control is needed to achieve sharing of voltage and/or current at the input and output sides, and (3) easy to expand system capacity.

Input/output variables sharing bus Vsh_g

R

Voref

R

ISOP and ISOS /+ + IPOP and IPOS

_ Vsh_f1 + + _

vsh_EA1

vof

+

Gvo vsh_EA2 /+

Gsh vof

Comparator 1 − RS Flip-flop +

VRAMP 1

Drive circuit 1 Module 1

_ Vsh_f2 +

+ Voref _

/+

Gsh

+

Gvo

Comparator 2 − RS Flip-flop +

VRAMP 2

Drive circuit 2 Module 2

Common duty cycle bus R

_ Vsh_f(n-1) +

+ Voref _

Gvo vof

_ Vsh_fn R + + Voref _

Gsh

vsh_EA(n-1) Comparator (n−1) /+ − RS Flip-flop + + VRAMP (n−1)

vof

Drive circuit (n−1) Module (n−1)

-vsh_EA(n) Gsh Gvo

/+ +

Comparator n − + VRAMP n

RS Flip-flop

Drive circuit n Module n

Fig. 2.9 Modularization architecture for DC–DC series–parallel power conversion systems

46

2 A General Control Strategy for DC–DC Series–Parallel Power …

From Fig. 2.9, it can be seen that each module regulates its own output voltage rather than system output voltage. For output-parallel-connected systems, the module output voltage is also the system output voltage, which will be precisely regulated, while for the output-series-connected systems, only one module output voltage is precisely regulated and the other module are indirectly regulated, leading to not very precise system output voltage. It should be noted that for IPOS system, the output voltage sharing bus and the common duty cycle bus are not mandatory because the system’s total output voltage will be well regulated if individual module manage to regulate its own output voltage; and for IPOP system, the common duty cycle bus is also unnecessary, and if it is removed, the current sharing method is similar to that proposed in [6], in which current-mode control is adopted as an inner current loop.

2.5 Experimental Verification In order to verify the effectiveness of the proposed general control strategy, we have constructed four DC–DC power conversion systems, i.e., IPOP, IPOS, ISOP, and ISOS systems. Each system consists of three basic converter modules, as shown in Fig. 2.10. The basic converter module is a phase-shifted full-bridge converter with the following specifications: • • • •

input voltage V in : 270 (±10%) Vdc; output voltage V o : 54 Vdc; rated output current I o : 20 A; switching frequency f s : 100 kHz. The power stage consists of the following devices and components:

(1) (2) (3) (4) (5)

switching device: IRFP460 (International Rectifier); clamping diodes: DSEI30-06 A (IXYS); rectifier diodes: APT60D30B (Advanced Power Technology); resonant inductance: 6 µH (Ferroxcube, E32/16/9, 8 turns); transformer: Ferroxcube, E42/21/20, primary winding is 15 turns, and secondary winding is 4 turns × 2 (center-tap full-wave rectifier); (6) output filter inductor: 30 µH (Ferroxcube, E42/21/15, 15 turns); (7) output filter capacitor: 4400 µF (2200 µF/100 V × 2—electrolytic). When the input sides and output sides of the three modules are connected in parallel, respectively, an IPOP system is constructed with the following specifications: system input voltage 270 Vdc ± 10%, system output voltage 54 Vdc, system rated output current 60 A (20 A × 3). Figure 2.11 shows the experimental waveforms of the IPOP system corresponding to a stepped load test and a stepped input voltage test. The individual output currents are sensed to provide information for the output current sharing bus. Figure 2.11a shows the output currents of the three modules and system’s output voltage (ac

2.5 Experimental Verification

47

Fig. 2.10 Prototype of DC–DC series–parallel power conversion systems consisting of three basic modules

coupled) when the system load changes between full load (60 A) and half load (30 A). We observe here that the output currents of the three modules are well shared. Figure 2.11b shows the waveforms when the input voltage steps up and down between 250 and 280 Vdc. Again we see that the output currents are well shared both at steady state and during transient. In order to verify that the OCS control is decoupled from the output voltage control, we use two sets of control parameters. Figures 2.11c, d show the dynamic responses with the same OCS control parameters but different output voltage control parameters. It can be seen that the output voltage responses differ. Figures 2.11e, f show the dynamic response with the same output voltage control parameters but different OCS control parameters. It can be seen that the output voltage response remains the same. In the prototype, the module duty cycle will be regulated fiercely during transient if the bandwidth of the OCS closed loop is high, which is easy to cause system oscillation. Hence, the bandwidth of the OCS closed loop is set to be relatively low in the prototype and the waveforms of output currents have little difference with different OCS control parameters.

48

2 A General Control Strategy for DC–DC Series–Parallel Power … C2 Io:[30A/div]

C1 Vin:[100V/div]

C2

M4 Io_1:[20A/div]

C1

M4

C4 Io_2:[20A/div]

M3

C4

C3 Io_3:[20A/div]

C3

C3 Io_2:[20A/div] C4 Io_3:[20A/div]

C4

C3 C1

M3 Io_1:[20A/div]

C1 Vo:[5V/div]

C2 Vo:[5V/div]

C2

Time:[20ms/div]

Time:[20ms/div]

(a) Response for stepped load

(b) Response for stepped input voltage C1Vin:[100V/div]

C2 Io:[30A/div]

C2

M4 Io_1:[20A/div]

M3

M4 C4 M1 C1

M3 Io_2:[20A/div]

C1

C4 Io_1:[20A/div] M1 Vo:[5V/div]

1# OCSR, 1# OVR

1# OCSR, 2# OVR

C1 Vo:[5V/div]

C3 Io_2:[20A/div]

C3 M2

M2 Vo:[5V/div]

C2

C2 Vo:[5V/div]

Time:[20ms/div]

Time:[20ms/div]

(c) Dynamic responses with same OCS parameters but different OVR parameters

(d) Dynamic responses with same OCS parameters but different OVR parameters C1 Vin:[100V/div]

C2 Io:[30A/div]

C2

M4 Io_1:[20A/div]

M3 Io_2:[20A/div]

C1 M3

M4 C4

1# OCSR, 1# OCSR, 1# OVR 2# OVR

C4 Io_1:[20A/div]

M1

M1 Vo:[5V/div]

C1

C1 Vo:[5V/div]

1# OCSR, 1# OVR

2# OCSR, 1# OVR

Time:[20ms/div]

(e) Dynamic responses with same OVR

C3 Io_2:[20A/div]

C3 M2

M2 Vo:[5V/div]

C2

C2 Vo:[5V/div]

1# OCSR, 1# OVR

2# OCSR, 1# OVR

Time:[20ms/div]

(f) Dynamic responses with same OVR

parameters but different OCSparameters

parameters but different OCS parameters

Fig. 2.11 Experimental waveforms for IPOP system

If the three modules are connected in parallel at the input sides and in series at the output sides, we have an IPOS system with the following specifications: system input voltage 270 Vdc ± 10%, system output voltage 162 Vdc (54 Vdc × 3), system rated output current 20 A.

2.5 Experimental Verification

49 C4 Vin:[100V/div]

C4 Io:[10A/div]

C4

C4

C2 Vo_2:[25V/div]

C1

C1 Vo_1:[5V/div]

C2

C2 Vo_2:[5V/div]

M3

M3 Vo_3:[5V/div]

C2

C3

C3 Vo:[10V/div]

M3

Time:[20ms/div]

C1 Vo_1:[25V/div]

C1

M3 Vo_3:[25V/div] C3 Vo:[100V/div]

C3 Time:[20ms/div]

(a) Response for stepped load

(b) Response for stepped input voltage

Fig. 2.12 Experimental waveforms for IPOS system

Figure 2.12 shows the experimental waveforms of the IPOS system corresponding to stepped load and stepped input voltage. The output voltage sharing bus and the common duty cycle bus are removed in the prototype system. Figure 2.12a shows the output voltages of the three modules (AC coupled) corresponding to a load stepping between full load (20 A) and half load (10 A), and Fig. 2.12b shows the waveforms of individual output voltages and the system’s total output voltage corresponding to an input voltage stepping between 250 and 280 Vdc. It can be seen that the output voltages are well shared both at steady state and during transient and the system’s output voltage is also well regulated. If the three modules are connected in series at the input sides and in parallel at the output sides, we have an ISOP system with the following specifications: system input voltage 810 Vdc ± 10% (270 Vdc × 3), system output voltage 54 V, system rated output current 60 A (20 A × 3). Figure 2.13 shows the experimental waveforms of the ISOP system corresponding to stepped load and stepped input voltage. Figure 2.13a shows the input voltages of the three modules and the system’s output voltage (AC coupled) corresponding to a load stepping between full load (60 A) and half load (30 A). Here, it is verified that the input voltages of the three modules are well shared. Figure 2.13b shows the waveforms corresponding to an input voltage stepping between 740 and 840 V. It can be seen that input voltages are well shared both at steady state and during transient, and the sharing of the output current is achieved automatically. In order to verify that the IVS control is decoupled from the output voltage control, we use two sets of control parameters. Figures 2.13c, d show the dynamic responses with the same IVS control parameters but different output voltage control parameters. It can be seen that the output voltage responses differ. Figures 2.13e, f show the dynamic response with the same output voltage control parameters but different IVS control parameters. It can be seen that the output voltage response remains the same. Similarly, the module duty cycle will be regulated violently during transient if the bandwidth of the IVS closed loop is high, which is easy to cause system oscillation. Hence, the bandwidth

50

2 A General Control Strategy for DC–DC Series–Parallel Power … C4 Vin:[500V/div] C4 Io:[30A/div]

C4 C1 C2

C4

C1Vin_1:[200V/div]

C1 Vin_1:[200V/div]

C2Vin_2:[200V/div]

C2 Vin_2:[200V/div]

C1

M3Vin_3:[200V/div]

M3Vin_3:[200V/div]

C2 M3

M3 C3

C3Vo:[5V/div]

C3 Vo:[5V/div]

C3

Time:[20ms/div]

Time:[20ms/div]

(a) Response for stepped load

(b) Response for stepped input voltage C4 Vin:[500V/div]

C4 Io:[30A/div]

C4

M2 Vin_2:[200V/div]

M1 Vin_1:[200V/div]

C4

C2 Vin_2:[200V/div]

C1 Vin_1:[200V/div]

M1 C1 M3

M3 Vo:[5V/div]

C3

C3 Vo:[5V/div]

M2

1# IVSR, 1# OVR

1# IVSR, 2# OVR

C2 M3

M3 Vo:[5V/div]

C3

C3 Vo:[5V/div]

1# IVSR, 1# OVR

1# IVSR, 2# OVR

Time:[20ms/div]

Time:[20ms/div]

(c) Dynamic responses with same IVS parameters but different OVR parameters

(d) Dynamic responses with same IVS parameters but different OVR parameters C4 Vin:[500V/div]

C4 Io:[30A/div] C4 M1

C4

M2 Vin_2:[200V/div]

M1 Vin_1:[200V/div]

C2 Vin_2:[200V/div]

C1Vin_1:[200V/div]

C1 M3

M3 Vo:[5V/div]

C3

C3 Vo:[5V/div]

M2

1# IVSR, 2# IVSR, 1# OVR 1# OVR

Time:[20ms/div]

(e) Dynamic responses with same OVR parameters but different IVSparameters

C2 M3 Vo:[5V/div]

M3

1# IVSR, 1# OVR

2# IVSR, 1# OVR

C3 Vo:[5V/div]

C3

Time:[20ms/div]

(f) Dynamic responses with same OVR parameters but different IVSparameters

Fig. 2.13 Experimental waveforms for the ISOP system

of the IVS closed loop is set to be relatively low in the prototype and the waveforms of input voltages have little difference with different IVS control parameters. When the input sides and output sides of the three modules are connected in series respectively, we have the ISOS system with the following specifications: system input

2.5 Experimental Verification

51

voltage 810 Vdc ± 10% (270 Vdc × 3), system output voltage 162 Vdc (54 Vdc × 3), system rated output current 20 A. Figure 2.14 shows the experimental waveforms of the ISOS system corresponding to stepped load and stepped input voltage. Figures 2.14a, b show the input voltage of the three modules and the system output voltage corresponding to a load stepping between full load (20 A) and half load (10 A) and an input voltage stepping between 740 and 840 Vdc, respectively. It can be seen that input voltages are shared both at steady state and during transient, and sharing of the output voltage is achieved automatically. In order to verify that the IVS control is decoupled from the output voltage control, we use two sets of control parameters. Figures 2.14c, d show the dynamic responses with the same IVS control parameters but different output voltage control parameters. It can be seen that the output voltage responses differ. Figures 2.14e, f show the dynamic response with the same output voltage control parameters but different IVS control parameters. It can be seen that the output voltage response remains the same.

2.6 Summary This chapter systematically analyzed the inherent relationships between the sharing of input voltages/currents and the sharing of output voltages/currents of the DC–DC series–parallel power conversion systems from the view of power conservation. It is pointed out that the sharing of input voltages/currents ensures the sharing of output voltages/currents and vice versa. For input-parallel-connected systems, the control strategies intending to achieve ICS or OCS (for IPOP) or OVS (for IPOS) can be used, however; the control strategies intending to achieve OCS (for ISOP) or OVS (for ISOS) for input-series-connected systems are unstable, and only the control strategies intending to achieve IVS can be adopted. A general control strategy was proposed for DC–DC series–parallel power conversion systems, which not only achieves equilibrium between the constituent modules but also decouples the output voltage control loop and the sharing control loop. Furthermore, modularization architecture for DC–DC series–parallel power conversion systems was proposed, achieving full modularization where all modules are selfcontained and standardized, and no extra control is needed to achieve sharing of voltage and/or current at the input and output sides. A prototype of DC–DC series–parallel power conversion systems consisting of three converter modules provided experimental results to verify the general control strategy and modularization architectures.

52

2 A General Control Strategy for DC–DC Series–Parallel Power …

C4 Vin:[500V/div]

C4 Io:[10A/div]

C4 C1 Vin_1:[200V/div]

C1 Vin_1:[200V/div]

C4

C2 Vin_2:[200V/div]

C2 Vin_2:[200V/div]

C1

C1 M3Vin_3:[200V/div]

M3 Vin_3:[200V/div]

C2

C2 M3

M3 C3 Vo:[10V/div]

C3

C3Vo:[100V/div]

C3 Time:[20ms/div]

Time:[20ms/div]

(a) Response for stepped load

(b) Response for stepped input voltage C4 Vin:[500V/div]

C4 Io:[10A/div] C4 M1

C4 M1Vin_2:[200V/div]

M1 Vin_1:[200V/div]

C1 Vin_2:[200V/div]

C1 Vin_1:[200V/div]

C1 M3

M3 Vo:[10V/div]

C3

C3 Vo:[10V/div]

M1 1# IVSR, 1# OVR

1# IVSR, 2# OVR

C1 M3 C3

M3 Vo:[10V/div]

1# IVSR, 1# OVR

1# IVSR, 2# OVR

C3 Vo:[10V/div] Time:[20ms/div]

Time:[20ms/div]

(c) Dynamic responses with same IVS

(d) Dynamic responses with same IVS

parameters but different OVR parameters

parameters but different OVR parameters C4 Vin:[500V/div]

C4 Io:[10A/div] C4 M1

C4 M1Vin_2:[200V/div]

M1 Vin_1:[200V/div] C1 Vin_1:[200V/div]

C1 M3

M3Vo:[10V/div]

C3

C3Vo:[10V/div]

M1 1# IVSR, 1# OVR

2# IVSR, 1# OVR

C1Vin_2:[200V/div]

C1 M3

M3 Vo:[10V/div]

C3

C3 Vo:[10V/div]

Time:[20ms/div]

(e) Dynamic responses with same OVR parameters but different IVS parameters Fig. 2.14 Experimental waveforms for ISOS system

1# IVSR, 1# OVR

2# IVSR, 1# OVR

Time:[20ms/div]

(f) Dynamic responses with same OVR parameters but different IVS parameters

References

53

References 1. Erickson, R., Maksimovic, D.: Fundmental of Power Electronics. Klurwer Academic, New York (2001) 2. Kim, J., Choi, H., Cho, B.H.: A novel droop method for converter parallel operation. IEEE Trans. Power Electron. 17(1), 25–32 (2002) 3. Ayyanar, R., Giri, R., Mohan, N.: Active input-voltage and load-current sharing in input-series and output-parallel connected modular DC-DC converters using dynamic input-voltage reference scheme. IEEE Trans. Power Electron. 19(6), 1462–1472 (2004) 4. Ruan, X., Chen, W., Cheng, L., Tse, C.K., Yan, H., Zhang, T.: Control strategy for input-seriesoutput-parallel converters. IEEE Trans. Ind. Electron. 56(4), 1174–1185 (2009) 5. Giri, R., Ayyanar, R., Ledezma, E.: Input-series and output-series connected modular DC-DC converters with active input voltage and output voltage sharing. In: Proceedings of the IEEE Applied Power Electronics Conference and Exposition, pp. 1751–1756 (2004) 6. Lin, C., Chen, C.: Single-wire current share paralleling of current mode-controlled dc power supplies. IEEE Trans. Ind. Electron. 47(4), 780–786 (2000)

Chapter 3

Mathematical Model and Closed-Loop Parameters Design for DC–DC ISOP System

Abstract This chapter derives the small-signal model of the phase-shifted fullbridge (PSFB)-based ISOP system, and based on the small-signal model, the general control strategy proposed in Chap. 2 is proved to be decoupled into n − 1 input voltage sharing (IVS) closed loops and one output voltage closed loop. The parameters design of the IVS control loop regulator and output voltage control loop regulator for a twomodule PSFB ISOP system is discussed and verified by the prototype. When the inner current loop is introduced into the general control strategy, the stability of the ISOP system with/without the IVS control loops is discussed and the root cause of system instability is also revealed. Furthermore, the design criteria of the IVS control loop regulator to keep ISOP system stable are derived. Keywords Small-signal model · Phase-shifted full-bridge converter Input voltage sharing · Parameter design · Inner current loop · Stability In Chap. 2, a general control strategy was proposed for DC–DC series–parallel power conversion systems, which not only achieves equilibrium between the constituent modules but also decouples the output voltage control loop and the sharing control loop. In this chapter, the small-signal model of ISOP system consisting of multiple phase-shifted full-bridge (PSFB) converters is derived to prove the decoupling of the system output voltage control loop and IVS control loop, and the design of the parameters of the two control loops is given. Furthermore, in order to improve the system dynamic response and realize output current short-circuit protection, the inner current loop is incorporated into the general control strategy, and the parameters design of the IVS control loop will be analyzed.

© Springer Nature Singapore Pte Ltd. and Science Press, Beijing, China 2019 X. Ruan et al., Control of Series-Parallel Conversion Systems, CPSS Power Electronics Series, https://doi.org/10.1007/978-981-13-2760-5_3

55

56

3 Mathematical Model and Closed-Loop Parameters Design …

3.1 Small-Signal Model of PSFB-Based ISOP System 3.1.1 Small-Signal Model of PSFB Converter The circuit configuration of the full-bridge converter is shown in Fig. 3.1, where Q1 ~ Q4 (including their body diodes D1 ~ D4 and their intrinsic capacitors C 1 ~ C 4 ) are four power switches, T r is the high-frequency isolation transformer with primaryand-secondary-winding ratio K, L r is the resonant inductor, DR1 and DR2 are two output rectifier diodes, L f is the output filter inductor, C f is the output filter capacitor, and RLd is the load. Figure 3.2 shows the key waveforms of the PSFB converter [1], where the switching process is ignored for simplicity. During [t 0 , t 1 ], Q1 and Q2 are conducting and the primary current ip is flowing through them in freewheeling state. In the secondary side, DR2 is conducting and the rectified voltage vrect is zero. At t 1 , Q2 is turned off and Q4 is turned on, then Q1 and Q4 conduct, and vAB is equal to V in . Both the primary and secondary winding currents decay and the secondary winding current is lower than the output filter inductor current, thus both DR1 and DR2 conduct, setting both the primary and secondary winding voltages to zero. The input voltage V in is applied on the resonant inductor L r , forcing ip to increase linearly. At t 2 , ip increases to the value of the output filter inductor current iLf reflected to the primary side, i.e., ip iLf /K. The rectifier diode DR2 turns off and DR1 continues conducting, vrect rises to the value of input voltage reflected to the secondary side, i.e., vrect V in /K, both ip and iLf increase linearly. At t 3 , Q1 is turned off and Q3 is turned on, ip is in freewheeling stage and vrect 0, starting the second half-switching period. As shown in Fig. 3.2, during [t 1 , t 2 ], the primary current ip increases from −I 2 to I 1 , the rectified voltage vrect is zero although vAB = V in . This means that the secondary side loses the input voltage reflected to the secondary during [t 1 , t 2 ]; in other words, the secondary effective duty cycle Deff is smaller than the primary duty cycle Dy [2]. Deff can be expressed as Deff Dy − Dloss

Q1 A

Q3

C1

Lr

ip

D 3 C3

Fig. 3.1 Full-bridge converter

Tr

Q2

Q4

DR1

Tr D2

C2

B

*

Vin

D1

(3.1)

D 4 C4

Lf + vrect iLf Cf _

* * DR2

RLd + Vo _

3.1 Small-Signal Model of PSFB-Based ISOP System

57

where Dloss is the secondary duty cycle loss. According to Fig. 3.2, we have [3] Dloss

t2 − t1 I1 + I2 Vin 1 Ts /2 · L r 2 fs

(3.2)

where T s is the switching period and f s 1/T s is the switching frequency. Referring to Fig. 3.2, the primary current ip at t 2 is 1 I Io − I1 K 2

(3.3)

where I o is the output current and I is the ripple current of the output filter inductor current. At t 4 , the primary current ip is equal to I 2 and it can be expressed as 1 − Dy 1 I (3.4) Io + − Vo I2 K 2 2 fs L f Substitution of (3.3) and (3.4) into (3.2) yields 1 − Dy 1 Dloss Vo 2Io − Vin 1 2 fs L f K L r 2 fs

vAB ip

(3.5)

I2

I1 Vin

t

0

Vin I2 DyTs/2 DeffTs/2 vrect

Vin/K t

0

I

iLf

Io t

0

t0

t1 t2

Fig. 3.2 Key waveforms of the PSFB converter

t3

t4 t5

t6

58

3 Mathematical Model and Closed-Loop Parameters Design …

Due to the fact that the PSFB converter is a buck-derived topology, its secondary effective duty cycle d eff is the summation of steady-state component Deff and transient component dˆeff , i.e., deff Deff + dˆeff

(3.6)

The transformer secondary rectified voltage vrect depends not only on the primary duty cycle Dy but also on the resonant inductor L r , the switching frequency f s , output filter inductor current iLf , and the input voltage V in . Therefore, the smallsignal transfer functions of the converter will depend on L r , f s , and the perturbations of the output filter inductor current iˆLf , input voltage, vˆin , and the primary duty cycle dˆy . To accurately model the dynamic behavior of the PSFB converter, it is necessary to find out the contributions of L r , f s , vˆin , iˆLf , and dˆy to dˆeff . These effects can be incorporated in the small-signal model of buck converter to obtain the model of the PSFB converter.

3.1.1.1

Effect of the Change of the Output Filter Inductor Current on Duty Cycle Modulation

Figure 3.3 shows the key waveforms of the PSFB converter when the output filter inductor current iLf is perturbed, where the solid lines represent the steady-state operation and the dashed line represents the primary current caused by an increase of iLf by the amount iˆLf . The perturbed primary current will reach the reflected filter inductor current at a later time than it would in the steady-state operation, and the additional delay time t can be expressed as t 2

iˆLf L r K Vin

(3.7)

Then, the change of d eff due to iˆLf , denoted as dˆi , is 4L r f s ˆ t − dˆi − i Lf Ts /2 K Vin

(3.8)

The negative sign in (3.8) represents that there will be a reduction in d eff if the output filter inductor current is increased. It should be noted that the primary duty cycle keeps unchanged in the perturbation.

3.1.1.2

Effect of the Change of the Input Voltage on Duty Cycle Modulation

Figure 3.4 shows the key waveforms of the PSFB converter when the input voltage V in is perturbed, where the solid lines represent the steady-state operation and the dashed

3.1 Small-Signal Model of PSFB-Based ISOP System

59

Fig. 3.3 Duty cycle modulation due to the change of the output filter inductor current Fig. 3.4 Duty cycle modulation due to the change of input voltage

line represents the primary current caused by an increase of V in by the amount vˆin . The perturbed primary current will reach the reflected filter inductor current sooner than it would in the steady-state operation, and the shortened time t can be expressed as

60

3 Mathematical Model and Closed-Loop Parameters Design …

1 2Io − K 1 2Io − K 1 2Io − ≈ K

t

1 − Dy Lr Lr Vo − 2 fs L f Vin Vin + vˆin 1 − Dy Lr vˆin Vo 2 fs L f Vin (Vin + vˆin ) 1 − Dy Lr Vo vˆin 2 fs L f Vin2

Then, the change of d eff due to vˆin , denoted as dˆv , is 1 − Dy 4L r f s t Vo vˆin dˆv Io − Ts /2 4 fs L f K Vin2

(3.9)

(3.10)

In practice, a relatively larger output filter inductance is usually used to reduce its current ripple; moreover, the primary duty cycle Dy is always made as large as possible to minimize the conduction loss caused by the circulating current in the primary side. Therefore, the second term in the bracket of (3.10) is much less than I o and (3.10) can be simplified as [2] 4L r f s Io vˆin dˆv K Vin2 The output current I o can be expressed as Io rewritten as 4L r f s Deff vˆin dˆv 2 K Vin RLd

(3.11) 1 Vin Deff and (3.11) can be K RLd

(3.12)

Equation (3.12) indicates that the effect of vˆin on d eff can be interpreted as an additional feed-forward of input voltage. The results of the previous analysis can be incorporated in the small-signal model of the buck converter, as shown in Fig. 3.5. The small-signal model of the PSFB converter can be derived by replacing dˆy in Fig. 3.5 by dˆeff , and dˆeff dˆy + dˆi + dˆv , as shown in Fig. 3.6. The contribution of dˆi and dˆv is represented by two controlled sources and the contribution of dˆv by two independent sources. This is to emphasize that dˆi and dˆv originate from the circuit itself (i.e., iˆLf and vˆin ) and are not controlled by the controller. It can be found that if L r 0, we have dˆi 0 and dˆv 0 according to (3.8) and (3.12), and if K = 1, the small-signal model of the PSFB converter is equivalent to that of the buck converter; in other words, the buck converter is a special case of the PSFB converter.

3.1 Small-Signal Model of PSFB-Based ISOP System

61

ˆ 1: Deff Vin d y

Fig. 3.5 Small-signal model of the buck converter

Lf

+ _

RLd +

vˆin

+ _

*

*

Vin dˆy RLd

Rcf Cf

vˆo _

Fig. 3.6 Small-signal model of the PSFB converter

3.1.2 Small-Signal Model of the PSFB-Based ISOP System The circuit architecture of the PSFB-based ISOP system is shown in Fig. 3.7, which consists of n PSFB converters connected in series at the input sides and in parallel at the output sides. At steady state, the input voltage and output current of each PSFB converter are V in /n and I o /n, respectively; thus, the load resistor of each converter is nRLd . Using the small-signal model of the PSFB converter, the small-signal model of the PSFB converter in ISOP system can be derived, as shown in Fig. 3.8. As shown in Fig. 3.8, we have the following equation:, Vin Vin /n 2 K · (n RLd ) n K RLd 4L r f s ˆ 4n L r f s ˆ i Lf j − i Lf j ( j 1, 2, . . . , n) dˆi j − K (Vin /n) K Vin 4L r Deff f s 4L r Deff f s vˆin j 2 2 vˆin j ( j 1, 2, . . . , n) K (Vin /n) · (n RLd ) K Vin RLd Ieq

dˆv j

(3.13.1) (3.13.2) (3.13.3)

where j represents module j and all the resonant inductors and turns ratios are assumed to be identical, respectively, i.e., L r1 L r2 ··· L rn L r , and K 1 K 2 ··· K n K. Using the small-signal model of the PSFB converter shown in Fig. 3.8, the smallsignal mode of the PSFB-based ISOP system can be obtained, as shown in Fig. 3.9.

62

3 Mathematical Model and Closed-Loop Parameters Design …

Lf1

iin1 Iin icd1 + vin1 _ C A1

Lr1

Tr1 B1

*

ip1

d1

Tr1 *

+ vrect1 iLf1 _

iinn

Vin

Cf

*

Lfn Lrn ipn

dn

Trn Bn

*

icdn + vinn _ C An

Trn *

+ vrectn iLfn _

*

Fig. 3.7 Circuit architecture of the PSFB-based ISOP system

Vin ˆ Vin ˆ d yj dij + dˆ vj nK nK 1: Deff /K

(

*

vˆinj

*

(

I eq dˆij + dˆ vj

)

)

iˆLfj

Lf

I eq dˆyj

nRLd + Rcf

vˆo

Cf

_

Fig. 3.8 Small-signal model of the PSFB converter in ISOP system

According to Fig. 3.9, we have ⎧ ⎪ Deff Vin ˆ ⎪ ˆv1 + dˆy1 s L f · iˆLf1 + vˆo ⎪ v ˆ d + + d in1 i1 ⎪ ⎪ nK ⎪ ⎪ K ⎪ ⎪ ⎪ Vin ˆ D eff ⎪ ⎨ vˆin2 + di2 + dˆv2 + dˆy2 s L f · iˆLf2 + vˆo K nK ⎪ ⎪ .. ⎪ ⎪ . ⎪ ⎪ ⎪ ⎪ ⎪ V D eff in ⎪ ⎪ vˆinn + dˆin + dˆvn + dˆyn s L f · iˆLfn + vˆo ⎩ K nK

(3.14.1) (3.14.2) .. . (3.14.n)

RLd + Vo _

3.1 Small-Signal Model of PSFB-Based ISOP System

Vin ˆ V d y1 in dˆ + dˆ i1 v1 nK 1: Deff /K nK

(

+ * vˆin1 _ Cd1

(

*

I eq dˆi1 + dˆ v1

63

)

)

I eq dˆy1

(

vˆin

(

*

I eq dˆi2 + dˆ v2

)

)

*

(

*

I eq dˆin + dˆ vn

vˆo _

iˆLf2 Lf2

)

)

……

(

Cdn

Rcf

I eq dˆy2

Vin ˆ Vin ˆ ˆ d yn din + d vn nK nK 1: Deff /K + vˆinn _

RLd +

Cf

Vin ˆ Vin ˆ ˆ d y2 di2 + d v2 nK nK 1: Deff /K + * vˆin2 _ Cd2

iˆLf1 Lf1

iˆLfn

Lfn

I eq dˆyn

Fig. 3.9 Small-signal model of the PSFB-based ISOP system

⎧ K ˆ ⎪ ⎪ i − sC v ˆ ⎪ in d in1 ⎪ ⎪ Deff ⎪ ⎪ ⎪ ⎪ K ˆ ⎪ ⎪ ⎨ i in − sCd vˆin2 Deff ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ K ˆ ⎪ ⎪ i in − sCd vˆinn ⎩ Deff

Ieq dˆy1 + dˆi1 + dˆv1 + iˆLf1

(3.15.1)

Ieq dˆy2 + dˆi2 + dˆv2 + iˆLf2

(3.15.2)

.. .

Ieq dˆyn + dˆin + dˆvn + iˆLfn n

vˆin j vˆin

.. . (3.15.n)

(3.16)

j1 n

j1

iˆLf j

vˆo

1 + Rcf //RLd sCf

(3.17)

It should be noted that in the above equations we made the following assuming: (1) L f1 L f2 … L fn L f , and (2) C d1 C d2 … C dn C d .

64

3 Mathematical Model and Closed-Loop Parameters Design …

The transfer functions and output impedance of the PSFB-based ISOP system can be derived with the aforementioned equations.

3.1.2.1

Control-to-Output Transfer Function

First, we consider the control-to-output transfer function of module j, and set vˆin 0, and dˆyk 0 (k = 1, 2, …, n, and k j). Then, adding up all the equations in (3.14) and combining (3.13.2), (3.13.3), (3.16), and (3.17), we obtain vˆo G vd j (s) vˆin 0 dˆy j dˆyk 0 (k1,2,...,n,

k j)

Vin (1 + sCf Rcf ) nK

Rp Rcf Lf Rcf +s + nCf Rcf + s 2 L f Cf 1 + + Rp C f 1 + +n RLd RLd RLd RLd G vd (s) ( j 1, 2, . . . , n)

(3.18)

where Rp 4L r f s /K 2 .

3.1.2.2

Control-to-Module-Input Transfer Function

To obtain the control-to-module-input transfer function, we first subtract (3.14.2) from (3.14.1) to obtain Vin Rp Deff vˆin1 − vˆin2 + 1+ dˆy1 − dˆy2 s L f + Rp iˆLf1 − iˆLf2 K n RLd nK (3.19) Moreover, subtracting (3.15.2) from (3.15.1) gives

Rp Vin ˆ dy1 − dˆy2 − 1 iˆLf1 − iˆLf2 2 n RLd n K RLd Rp Deff s K Cd vˆin1 − vˆin2 + + 2 2 Deff n K RLd

(3.20)

Putting (3.20) into (3.19) yields vˆin2 − vˆin1 A(s) dˆy2 − dˆy1 2 Deff n RLd Vin + s L f RLd Vin . where A(s) − 2 2 2 2 2 s Rp + s 2 L f n Deff RLd + s Deff L f Rp + n 2 K 2 Cd RLd

(3.21)

3.1 Small-Signal Model of PSFB-Based ISOP System

65

Likewise, we can obtain vˆin( j+1) − vˆin j A(s) dˆy( j+1) − dˆy j ( j 1, 2, . . . , n − 1)

(3.22)

From (3.22), we can derive vˆin j vˆin1 − A(s) dˆy1 − dˆy j ( j 1, 2, . . . , n)

(3.23)

Letting vˆin 0 and substituting (3.23) into (3.16), we obtain A(s) ˆ dy j n j1

(3.24)

A(s) ˆ dy j ( j 1, 2, . . . , n) n j1

(3.25)

vˆin1 A(s)dˆy1 −

n

Substitution of (3.24) into (3.23) yields vˆin j A(s)dˆy j −

n

Equation (3.25) can be rewritten in the following matrix form: ⎤⎡ ⎤ ⎡ ⎤ ⎡ A(s)(n−1) A(s) A(s) ˆy1 − . . . − d n n n vˆin1 ⎥⎢ ⎥ ⎢ ⎥ ⎢ ⎢ A(s) ⎥⎢ ˆ ⎥ A(s)(n−1) ⎢ ⎥ ⎥ ⎢ − A(s) ⎢ vˆin2 ⎥ . . . − d n n n ⎥⎢ y2 ⎥ ⎥ ⎢ ⎢ ⎢ ⎥⎢ ⎥ ⎢ . ⎥ .. .. .. .. ⎥⎢ .. ⎥ ⎢ ⎢ . ⎥ ⎥⎢ . ⎥ ⎢ ⎣ . ⎦ . . . . ⎦⎣ ⎦ ⎣ vˆinn vˆ 0 − A(s) − A(s) . . . A(s)(n−1) dˆyn in n

3.1.2.3

n

(3.26)

n

Input-Voltage-to-Output-Voltage Transfer Function

To find the input-voltage-to-output-voltage transfer function, we let dˆy j 0 (j = 1, 2, …, n). Then, adding up all the equations in (3.14) and combining (3.13.2), (3.13.3), (3.16), and (3.17), we can obtain vˆo G vg (s) vˆin dˆy j 0 ( j1,2,...,n) Rp Deff 1+ (1 + sCf Rcf ) K n RLd

Rp Rcf Lf Rcf +s + nCf Rcf + s 2 L f Cf 1 + + Rp C f 1 + +n RLd RLd RLd RLd (3.27)

66

3 Mathematical Model and Closed-Loop Parameters Design …

Fig. 3.10 Small-signal model representing the variation of output current

n

∑ iˆ

Lfj

j =1

PSFBbased ISOP System

3.1.2.4

Rcf Cf

RLd + vˆo _

iˆo

System-Input-to-Module-Input Transfer Function

We let dˆy j 0 (j = 1, 2, …, n) and use (3.23) to obtain vˆin1 vˆin2 · · · vˆinn

(3.28)

Substitution of (3.28) into (3.16) yields vˆin1 vˆin2 · · · vˆinn vˆin /n which can be rewritten as vˆin j G vin jg (s) vˆin dˆy j 0 3.1.2.5

( j1,2,...,n)

1 G ving (s) ( j 1, 2, . . . , n) n

(3.29)

(3.30)

System’s Output Impedance

The output impedance of the system can be readily found in Fig. 3.10. First, we observe that n

j 1

iˆLf j + iˆo

vˆo

1 + Rcf //RLd sCf

(3.31)

We let vˆin 0 and dˆy j 0 (j = 1, 2, …, n). Then, adding up all the equations in (3.14) and using (3.13.2), (3.13.3), (3.16), and (3.31) to obtain vˆo Z out (s) vˆin 0 iˆo dˆy j 0 ( j1,2,..., n) (3.32) s L f + Rp (1 + sCf Rcf )

Rp Rcf Lf Rcf +s + nCf Rcf + s 2 L f Cf 1 + + Rp C f 1 + +n RLd RLd RLd RLd

3.1 Small-Signal Model of PSFB-Based ISOP System

3.1.2.6

67

Output-Current-to-Module-Input Transfer Function

Putting vˆin 0 and dˆy j 0 (j = 1, 2, …, n) into (3.26), we can readily see that all the divided capacitor voltages are unrelated to the output current, i.e., vˆin j 0 (3.33) G vin j io (s) vˆin 0 iˆo dˆy j 0 ( j1,2,...,n)

3.2 Decoupling the Control Loops of the PSFB-Based ISOP System In Chap. 2, a general control strategy was proposed for DC–DC series–parallel power conversion systems. Thus, the control strategy for PSFB-based ISOP system can be obtained, as shown in Fig. 3.11, which consists of n − 1 IVS control loops and a common output voltage control loop. Gvcd and Gvo are the transfer functions of the input voltages sharing regulator (IVSR) and output voltage regulator (OVR), respectively, and vo_EA is the output signal of the OVR and vin_EAj (j = 1, 2, …, n−1) is the output signal of each IVSR. V RAMPj is the sawtooth signal and its amplitude is V pp . For the first n − 1 modules, vin_EAj is subtracted from vo_EA and then compares with V RAMPj to obtain the duty cycle of module j. For module n, the sum of the first n − 1 vin_EAj and vo_EA compares with V RAMPj to obtain its duty cycle. According to (3.18) and (3.26), we can obtain ⎤ ⎡ ⎡ ⎤ vˆin1_d dˆy1 ⎥ ⎢ ⎢ ⎥ ⎢ vˆin2_d ⎥ ⎢ ⎥ ⎥ ⎢ ˆ ⎢ d y2 ⎥ ⎥ ⎢ ⎥ .. ⎥ H⎢ ⎢ (3.34) ⎢ ⎥ ⎥ ⎢ . ⎢ .. ⎥ ⎥ ⎢ ⎢ ⎥ . ⎢ vˆin(n−1)_d ⎥ ⎣ ⎦ ⎦ ⎣ ˆyn d vˆo_d where vˆin j_d denotes the variations resulted from dˆy j (j = 1, 2, …, n), and H is ⎡ ⎤ A(s)(n−1) A(s) A(s) A(s) − . . . − − n n n n ⎢ ⎥ ⎢ A(s) A(s) A(s) ⎥ A(s)(n−1) − . . . − − ⎢ ⎥ n n n n ⎢ ⎥ ⎢ .. .. .. .. .. ⎥ H ⎢ ⎥. ⎢ . . . . . ⎥ ⎢ ⎥ ⎢ − A(s) − A(s) . . . A(s)(n−1) − A(s) ⎥ ⎣ n n n n ⎦ G vd (s) G vd (s) . . . G vd (s) G vd (s)

68

3 Mathematical Model and Closed-Loop Parameters Design …

+ Vin/n

_ v in1 Gvcd

_

vin_EA1

Comparator 1

+ VRAMP 1

+ Vin/n

_ v in2 Gvcd

vin_EA2 _ +

− +

Comparator 2 − +

RS Flip-flop

Drive circuit 1

RS Flip-flop

Drive circuit 2

VRAMP 2

+ Vin/n

_ v in(n−1) Gvcd IVSR vin_EA(n−1)

_ +

VRAMP (n−1)

+ + + Voref

_ vof

Gvo

Comparator (n−1) − RS Flip-flop +

vo_EA

+

+ +

Drive circuit (n−1)

Comparator n

VRAMP n

− +

RS Flip-flop

Drive circuit n

Fig. 3.11 Control strategy of the PSFB-based ISOP system

According to Fig. 3.11 and from (3.18), (3.27), (3.30), (3.32), and (3.34), we can develop the block diagram shown in Fig. 3.12, where 1/V pp is the transfer function of PWM modulator, and K vc and K vo are the sensor gains of input voltage and output voltage, respectively. Denoting the output signals of n − 1 IVSRs and OVR as xˆ1 , xˆ2 , …, xˆn , respectively, as shown in Fig. 3.12, we have ⎡ ⎤ ⎤⎡ xˆ ⎤ ⎡ dˆy1 ⎢ ⎥ 1 −1 0 0 · · · 0 1 ⎢ ⎥ ⎥ ⎢ dˆy2 ⎥ ⎢ 0 −1 0 · · · 0 1 ⎥⎢ x ˆ ⎢ ⎥ 2 ⎥⎢ ⎢ ⎥ ⎢ ⎥ ⎥ ⎢ ⎥ ⎢ 1 ⎢ .. ⎥ ⎢ .. ⎥ ⎢. . . . . . ⎥ (3.35) ⎢ . ⎥ V ⎢ .. .. .. .. .. .. ⎥⎢ . ⎥ ⎥ pp ⎢ ⎥⎢ ⎢ ⎥ ⎢ ⎥ ⎢ˆ ⎥ ⎣ 0 0 0 · · · −1 1 ⎦⎣ xˆn−1 ⎦ ⎢ dy(n−1) ⎥ ⎣ ⎦ 1 1 1 ··· 1 1 xˆn dˆyn Substituting (3.35) into (3.34) yields

3.2 Decoupling the Control Loops of the PSFB-Based ISOP System

vˆin(n -1)f

vˆin2f vˆin1f + + +

_ Gvcd(s) _ Gvcd(s) _ Gvcd(s)

xˆ1

_

xˆ2

+ _ + _

xˆn −1

+

Kvc/n

vˆin

+

vˆoref

_

Gvo(s)

xˆn

+ +

+

+ +

69

Kvc Kvc Kvc 1/Vpp 1/Vpp 1/Vpp

1/Vpp

dˆy1 dˆy2

+ + ˆvin1_d + +

vˆin1

+ + vˆin(n -1)_d

vˆin ( n −1)

vˆin2_d

dˆy( n −1)

H

Gving(s)

dˆyn

Gvg(s) ++

vˆo_d

vˆof

vˆin2

+

Zout(s)

Kvo

iˆo

vˆo

Fig. 3.12 Block diagram of the control strategy for PSFB-based ISOP system

⎡

vˆin1_d

⎤

⎡

−A(s) 0 ⎢ ⎥ ⎢ ⎢ vˆin2_d ⎥ ⎢ 0 −A(s) ⎢ ⎥ ⎢ ⎢ ⎥ 1 ⎢ . .. .. ⎢ ⎥ ⎢ .. ⎢ ⎥ . . V pp ⎢ ⎢ ⎥ ⎢ ⎢ vˆin(n−1)_d ⎥ 0 0 ⎣ ⎣ ⎦ 0 0 vˆo_d

⎤⎡ ⎤ xˆ1 0 ··· 0 0 ⎥⎢ ⎥ ··· 0 0 ⎥⎢ xˆ2 ⎥ ⎥⎢ ⎥ ⎥⎢ . ⎥ .. .. .. .. . ⎢ ⎥ ⎥ . . . . ⎥⎢ . ⎥ ⎥⎢ ⎥ 0 · · · −A(s) 0 ⎦⎣ xˆn−1 ⎦ 0 · · · 0 nG vd (s) xˆn

(3.36)

Incorporating (3.36) into Fig. 3.12, we can obtain the decoupled block diagram, as shown in Fig. 3.13. As seen, the original n-input-n-output control system is decoupled into n single-input-single-output closed loops, which consists of n − 1 IVS closed loops and one output voltage closed loop. Moreover, the n − 1 IVS closed loops are decoupled from one another and are also decoupled from the output voltage closed loop. Thus, we can design the parameters of IVSR and OVR independently.

3.3 Closed-Loop Design This section discusses the parameters design of Gvcd (s) and Gvo (s) based on a twomodule PSFB ISOP system. The specifications of the ISOP system are as follows. The input voltage V in 540 Vdc, the output voltage V o 60 Vdc, and the output current is 50 A (the equivalent load resistor RLd 1.2 ). For each PSFB module, the

70

3 Mathematical Model and Closed-Loop Parameters Design …

vˆin

Kvc/n

+ _

Gving(s) xˆ Gvcd(s) 1 1/Vpp

Kvc/n

+ _

Gving(s) xˆ Gvcd(s) 2 1/Vpp

Kvc/n

+ _

Gving(s) xˆ Gvcd(s) n −1 1/Vpp

vˆin ( n −1)f

vˆoref

+ _

Gvo(s)

vˆin1

A(s)

+

+

vˆin2

Zout(s)

1/Vpp

vˆof

A(s)

+

+

n−1 IVS control loops

vˆin ( n −1)

Kvc

iˆo

xˆn

+

Kvc

vˆin2f

vˆin

+

Kvc

vˆin1f

vˆin

A(s)

+ nGvd(s) +

+

vˆin

Gvg(s)

Kvo

vˆo

Output voltage control loop

Fig. 3.13 Decoupled block diagram of the control strategy for PSFB-based ISOP system

transformer turns ratio K = 12:4, the output filter inductor L f 26 µH, the output filter capacitor C f 3000 µF with equivalent series resistor Rcf 0.015 , the resonant inductor L r 6.5 µH, the switching frequency f s 100 kHz, the voltage sensor gains K vc 0.01 and K vo 0.03, and the peak-to-peak voltage of the sawtooth signal V pp 4 Vdc. The IVS and output voltage control loops are shown in Fig. 3.13. Here, n = 2.

3.3.1 Output Voltage Regulator It can be readily found in Fig. 3.13 that the loop gain of the output voltage control loop is Tvo_c (s) 2K vo G vd (s)G vo (s)/Vpp

(3.37)

By setting Gvo (s) 1, the uncompensated loop gain T vo_u (s) is obtained as

3.3 Closed-Loop Design

71

Fig. 3.14 Uncompensated and compensated output voltage loop gains

60 40

|Tvo_c|

20

(dB) 0

−20

|Tvo_u|

−40 −60 60 30 0

∠Tvo_u

(°) −30 −60 −90 −120 −150 10

∠Tvo_c 1⋅103 f (Hz)

100

1⋅104

1⋅105

|Gvo| (db) 20log|Gvo |

R2 vi vref

R1 R1

f f

0

C

0

∠Gvo (°)

_

fz =

−90

+

0.1fz

(a) Circuit diagram

fz

ωz 2π

10fz

(b) Bode plots

Fig. 3.15 PI compensator

Tvo_u (s) 2K vo G vd (s)/Vpp

(3.38)

With the given parameters of the ISOP system and substituting (3.18) into (3.38), the Bode plots of T vo _u (s) are depicted, as shown in Fig. 3.14 (dashed lines). As seen, the uncompensated loop gain is close to zero at low frequencies. In order to increase the low-frequency loop gain, a proportional–integral (PI) compensator is adopted as the output voltage regulator, as shown in Fig. 3.15a, which can be expressed as

G vo (s) G vo∞

R2 s + ωz1 s R1

s+

1 R2 C s

(3.39)

1 ωz1 . 2π 2π R2 C The crossover frequency f c1 is chosen to be 10 kHz, which is one-twentieth of the secondary ripple frequency. As shown in Fig. 3.14, the uncompensated loop gain where the corner frequency of the PI compensator is f z1

72

3 Mathematical Model and Closed-Loop Parameters Design …

at 10 kHz has a magnitude of −17.087 dB, and the compensator should then have a gain of 17.087 dB at 10 kHz, i.e., R2 R1 17.078 10 20

G vo∞

(3.40)

Setting R1 10 k and then R2 71 k. In practice, R2 68 k. To guarantee no phase angle lag after the compensation, it is usually set 10 f z1 ≤ f c1 . With Fig. 3.15b, we have C

1 5 ≥ 2.34 (nF) 2π f z1 R2 π f c1 R2

(3.41)

Here, C = 10 nF is chosen, and the transfer function of the PI compensator is G vo (s) 6.8

s + 1471 s

(3.42)

The compensated loop gain T vo_c (s) is shown as solid lines in Fig. 3.14. As shown, the compensated loop gain has a crossover frequency of 10 kHz with a phase margin of 85.9°.

3.3.2 IVS Regulator As shown in Fig. 3.13, the loop gain of the IVS control loop is Tcd_c (s) K vc G vcd (s)A(s)/Vpp

(3.43)

Since Gvinjg (s) 1/n (j = 1, 2, …, n) according to (3.30), it can be concluded from Fig. 3.13 that the IVS regulator gain should be zero, implying that the IVS control would be unnecessary when all the modules are identical. Obviously, when the modules are not identical, the IVS control would be needed. In fact, to facilitate an accurate design of the compensator, the small-signal model should be further modified, taking into account the differences of the modules, which is much more complex and will not be discussed here. By setting Gvcd (s) 1, the uncompensated loop gain T cd_u (s) is expressed as Tcd_u (s) K vc A(s)/Vpp

(3.44)

With the given parameters of the ISOP system, the Bode plots of T cd _u (s) are depicted, as shown in Fig. 3.16 (dashed lines). A PI compensator (see Fig. 3.15a) is employed and its transfer function is

3.3 Closed-Loop Design

73 60

Fig. 3.16 Uncompensated and compensated IVS loop gains

30

|Tcd_c|

0

(dB) −30

|Tcd_u|

−60 −90 −120 60 0

(°)

∠Tcd_u

−60 −120

∠Tcd_c

−180 10

100

G vcd (s) G vcd∞

1⋅103 f (Hz)

s + ωz2 s

1⋅104

1⋅105

(3.45)

Suppose that we aim to design a compensator to achieve a crossover frequency of 300 Hz, and the uncompensated loop gain at 300 Hz has a magnitude of approximately −13.313 dB. Thus, the compensator should have a gain of 13.313 dB at 300 Hz, i.e., G vcd∞ 10

13.313 20

(3.46)

Setting R1 10 k and then R2 46.13 k. Here, R2 47 k is chosen. To guarantee no phase angle lag after the compensation, 10f z2 ≤ f c2 is selected. With Fig. 3.15b, we have C

1 5 ≥ 0.113 (µF) 2π f z2 R2 π f c2 R2

(3.47)

Here, C 0.1 µF is chosen, and the transfer function of the PI compensator is G vcd (s) 4.7

s + 212.7 s

(3.48)

Substituting (3.48) into (3.43) and with the given system parameters, the compensated loop gain T cd_c (s) is depicted, as shown by the solid lines in Fig. 3.16. It can be found that the compensated loop gain has a crossover frequency of 300 Hz with a phase margin 80.5°.

74

3 Mathematical Model and Closed-Loop Parameters Design …

Table 3.1 Parameters of OVR and IVSR Parameters OVR

IVSR

1#

G vo (s) 6.8(s + 1471)/s

G vcd (s) 4.7(s + 212.7)/s

2#

G vo (s) 12(s + 833)/s

G vcd (s) 9.1(s + 161.6)/s

3.4 Experimental Verification of the General Control Strategy A prototype of ISOP system consisting of two PSFB modules was built in the lab to verify the correctness of the small-signal model and loop parameters design in Sects. 3.2 and 3.3, and the specifications are • • • •

input voltage V in 540 Vdc ± 10%, output voltage V o 60 Vdc, rated output current I o 50 A, and switching frequency f s 100 kHz.

Figure 3.17 shows the experimental waveforms of vAB1 , vAB2 , ip1 , ip2 , vrect1 , vrect2 , iLf1 , and iLf2 at full load condition, and it can be seen that both the input voltages and output currents of the two PSFB modules are shared. Figure 3.18 shows the dynamic responses of the input voltage and output voltage (AC coupled) of module 1 when the system input voltage is step changed between 486 and 596 Vdc. The parameters of OVR and IVSR are listed in Table 3.1. Figure 3.18a shows the dynamic responses with the same 1# IVSR parameter but different OVR parameters. It can be seen that the output voltage responses differ and no change on the module input voltage. Figure 3.18b shows the dynamic responses with the same 1# OVR parameter but different IVSR parameters. It can be seen that the output voltage responses keep unchanged. Figure 3.19 shows the dynamic responses of the input voltage (AC coupled) and output voltage (AC coupled) of module 1 when the load current is step changed between 25 and 50 A with different regulator parameters. It can be found that the load step has no influence on the module input voltage, verifying the correctness of (3.33). We can conclude from Figs. 3.18 and 3.19 that IVS control loop is decoupled from the output voltage control loop.

3.5 Three-Loop Control Strategy for ISOP System In Sects. 3.2 and 3.3, removing the IVS control loops will lead to a common duty cycle control strategy with a single-output voltage control loop for the ISOP system, and the ISOP system is stable. However, the mismatched parameters among the modules may cause input voltages unbalance, and thus the IVS control loops are needed. The

3.5 Three-Loop Control Strategy for ISOP System

75

vAB1:[400V/div] ip1:[20A/div]

vAB2:[400V/div]

Time:[2μs/div]

ip2:[20A/div]

(a) Primary voltages and currents of the two PSFB modules

vrect1:[100V/div]

iLf1:[15A/div] vrect2:[100V/div]

iLf2:[15A/div] Time:[2μs/div] (b) Secondary rectified voltages and filter inductor currents of the two PSFB modules Fig. 3.17 Experimental waveforms of the ISOP system

inner current loop can be introduced to form a voltage–current double closed loop to improve the system dynamic response and readily realize output current short-circuit protection. So, whether the ISOP system with only the voltage–current double closed loop is stable if the IVS control loops are removed too? If it is unstable, then the IVS control loops are required, but how to design the IVS regulators? This section will address these issues.

76

3 Mathematical Model and Closed-Loop Parameters Design …

vin:[200V/div] vin1:[100V/div]

1# IVSR 1# OVR

vin1:[100V/div] vo:[2V/div]

1# IVSR 2# OVR

vo:[2V/div] Time:[20ms/div] (a) Different OVR parameters

vin:[200V/div] vin1:[100V/div]

1# IVSR 1# OVR

vin1:[100V/div] vo:[2V/div]

2# IVSR 1# OVR

vo:[2V/div] Time:[20ms/div] (b) Different IVSR parameters

Fig. 3.18 Dynamic responses for stepped input voltage

3.5.1 Positive Resistance Figure 3.20 shows the control block diagram of the voltage–current double closed loop for PSFB ISOP system, where iLf1 , iLf2 , …, iLfn are the module output filter inductor currents, Gc is the gain of inner current loop regulator, and the current reference signal of the inner current loops is the output of the output voltage regulator. At steady state, all the inner current loops have the same current reference and all the modules have the same output filter inductor currents, as well as the output

3.5 Three-Loop Control Strategy for ISOP System

77

io:[25A/div] vin1:[10V/div]

1# IVSR 1# OVR

vin1:[10V/div]

vo:[2V/div]

1# IVSR 2# OVR

vo:[2V/div] Time:[20ms/div] (a) Different OVR parameters

io:[25A/div] vin1:[10V/div]

1# IVSR 1# OVR

vin1:[10V/div]

vo:[2V/div]

2# IVSR 1# OVR

vo:[2V/div] Time:[20ms/div] (b) Different IVSR parameters

Fig. 3.19 Dynamic responses for stepped load

currents, which is equivalent to an OCS control strategy. In Chap. 2, it is pointed out that the ISOP system is unstable with OCS control strategy and the root cause is that the input impedance of each converter module appears as negative resistance characteristic when OCS control strategy employed. Therefore, the ISOP system with only voltage–current double closed loop is unstable if the IVS control loops are removed. Thus, if the ISOP system is stable, the input impedance of each converter module must appear as positive resistance characteristic, and one direct approach is to connect

78

3 Mathematical Model and Closed-Loop Parameters Design … _ iLf1 +

Gc

Comparator 1 − RS Flip-flop +

Drive circuit 1

Comparator 2 − RS Flip-flop +

Drive circuit 2

Comparator n − RS Flip-flop +

Drive circuit n

VRAMP 1 _ iLf2 +

Gc

Inner current loops

+ Voref

_ vof

Gvo

_ iLfn + vo_EA

VRAMP 2

Gc

VRAMP n

Fig. 3.20 Block diagram of voltage–current double closed loop Fig. 3.21 Small-signal model of ISOP system with a resistor connected to the input side of each module

iˆcin_1

iˆin +

′ Z in1

vˆ in1 _ Cd1

iˆin_cs1 Rne1

Rp1

iˆcin _ 2 + vˆ in

′ Z in2

vˆ in2 _ Cd2

iˆin_cs2 Rne2

Rp2

iˆcin _ n +

′ Z inn vˆ inn _ Cdn

iˆin_csn Rnen

Rpn

a positive resistance Rpj (j = 1, 2, …, n) in parallel with the input side the each module, as shown in Fig. 3.21, where iˆin_cs j (j = 1, 2, …, n) represents the each module’s input current perturbation when the voltage–current double closed loop is adopted. The equivalent input impedance of each module is Z in j Rne j //Rp j

Rne j · Rp j > 0 ( j 1, 2, . . . , n) Rne j + Rp j

(3.49)

3.5 Three-Loop Control Strategy for ISOP System

+ Vin/n

+ Vin/n

_ v in1

vin_EA1 Gvcd IVSR

_ v in2

_

iref1+

_ vinn

_

iref2 + +

_ vof + Voref

vin_EAn Gvcd

Comparator 1 RS Flip-flop +

Drive circuit 1

Comparator 2 RS Flip-flop +

Drive circuit 2

Comparator n RS Flip-flop +

Drive circuit n

VRAMP 1

IVS loops

+ Vin/n

Gc

+

vin_EA2 Gvcd

_ iLf1

79

_

Gc Inner current loops

irefn + +

_ iLf2

_ iLfn Gc

VRAMP 2

VRAMP n Gvo

vo_EA

Output voltage loop

Fig. 3.22 Three-loop control strategy for ISOP system

Thus, we have Rp j < −Rne j ( j 1, 2, . . . , n)

(3.50)

The Rpj can be a real resistor and the resistance of Rpj < −Rnej (j = 1, 2, …, n), due to the −Rnej (j = 1, 2, …, n), is the equivalent resistance reflecting the input power of each module, if Rpj < −Rnej (j = 1, 2, …, n), the power dissipated on Rpj (j = 1, 2, …, n) will be larger than the input power of each module, which is absolutely unacceptable. When the input impedance of module appears as positive resistance characteristic, the module input current increases/decreases with the increase/decrease of input voltage, and consequently the module output current will also increase/decrease, as well as the module output filter inductor current. Therefore, we can compare the input voltage of each module with the reference of input voltage of each module V in /n, and when the input voltage of module is higher/lower than V in /n, the module output filter inductor current should increase/decrease by introducing a feedback signal to increase/decrease the current reference signal of the inner current loop. This function can be implemented by the IVS loops, as shown in Fig. 3.22. Taking module 1 for example, when its input voltage vin1 is higher than V in /n, leading to the increase of its current reference signal for inner current loop, thus the output filter inductor current and output power will increase, as well as the input current of module 1, resulting in the input impedance of module 1, appears as positive resistance characteristic.

80

3 Mathematical Model and Closed-Loop Parameters Design …

Fig. 3.23 Equivalent small-signal model for ISOP system with three-loop control strategy

iˆcin _1

iˆin

+ ′ Z in1 ˆvin1 _ Cd1

iˆin_cs1

iˆin_IVS1

Rne1

Rp1

iˆin_cs2

iˆin_IVS2

Rne2

Rp2

+

vˆ in

′ Z in2 vˆ in2 _ Cd2 iˆcin _ n +

′ Z inn vˆ inn _ Cdn

Ă

iˆcin _ 2

iˆin_csn

iˆin_IVSn

Rnen

Rpn

3.5.2 IVS Regulator Design In Sect. 3.2, it has been pointed out that the IVS control loops and output voltage control loop are decoupled, and thus we can design the IVS loops and voltage–current double closed loops independently. In Sect. 3.5.1, it is found that the function of the IVS loops is equivalent to a positive resistance and the equivalent small-signal model for ISOP system with three-loop control strategy is shown in Fig. 3.23, where iˆin_IVS j (j = 1, 2, …, n) is the module input current perturbation introduced by the IVS control loops. According to Fig. 3.22, the individual output filter inductor current perturbation iˆLf_IVS j (j = 1, 2, …, n) caused by the input voltage perturbationvˆin j (j = 1, 2, …, n) can be expressed as iˆLf_IVS j vˆin j · G vcd ( j 1, 2, . . . , n)

(3.51)

Then, the individual output power perturbation pˆ o j (j = 1, 2, …, n) is pˆ o j iˆLf_IVS j · Vo vˆin j · G vcd · Vo ( j 1, 2, . . . , n)

(3.52)

3.5 Three-Loop Control Strategy for ISOP System

81

Referring to Fig. 3.23, the individual input power is pin j vin j · i cin_ j ( j 1, 2, . . . , n)

(3.53)

The perturbation equation of (3.53) can be expressed as Pin j + pˆ in j Vin j + vˆin j Icin_ j + iˆcin_ j

Vin j Icin_ j + vˆin j Icin_ j + Vin j + vˆin j iˆin_IVS j + iˆin_cs j

( j 1, 2, . . . , n)

(3.54)

where Pinj , V cdj , and I cin_j (j = 1, 2, …, n) are the module input power, input voltage, and input current at steady state, respectively. Neglecting the second-order term of (3.54) and noting that the DC terms on both sides of the equation are equal, we obtain (3.55) pˆ in j Icin_ j vˆin j + Vin j iˆin_IVS j + iˆin_cs j ( j 1, 2, . . . , n) According to power conservation, we have pˆ o j pˆ in j · η j ( j 1, 2, . . . , n)

(3.56)

Combining (3.52), (3.55), and (3.56) yields Vo Icin_ j vˆin j + Vin j iˆin_IVS j + iˆin_cs j vˆin j · G vcd · ( j 1, 2, . . . , n) ηj

(3.57)

The negative resistance can be expressed as Vin j vˆin j − Rne j ( j 1, 2, . . . , n) ˆi in_cs j Icin

(3.58)

Substituting (3.58) into (3.57) yields Rp j

vˆin j iˆin_IVS j

Vin j G vcd · Vo η j

( j 1, 2, . . . , n)

(3.59)

Substituting (3.58) and (3.59) into (3.50) leads to G vcd >

Pin · η j ( j 1, 2, . . . , n) Vo · Vin

(3.60)

Neglecting the difference among the module efficiencies and assuming η1 η2 ··· ηn η, according to (3.60), we can obtain

82

3 Mathematical Model and Closed-Loop Parameters Design …

G vcd >

Po · η Vo · Vin

(3.61)

To ensure the stable operation of ISOP system over the whole input voltage and load range, Gvcd should satisfy the following formula: G vcd_min

Po_max · η Vo · Vin_min

(3.62)

where V in_min is the minimum input voltage and Po_max is the maximum output power.

3.6 Experimental Verification of the Three-Loop Control Strategy In order to verify the correctness of the IVS loop design, an ISOP system prototype consisting of two PSFB converter modules has been fabricated with the following specifications: • • • •

input voltage V in 540 Vdc ± 10%, output voltage V o 54 Vdc, rated output current I o 40 A, and switching frequency f s 100 kHz.

Figure 3.24 shows the experimental waveforms of the ISOP system with voltage–current double closed-loop control strategy (see Fig. 3.20). vin1 , vin2 , io1 , and io2 are the input voltages and output currents of the two modules, respectively, and vo is the system output voltage. Prior to t 1 , the three-loop control strategy (see Fig. 3.22) is employed and the ISOP system works stably, and IVS and OCS are achieved between the two modules. At t 1 , the IVS loops are removed and after that vin1 rises and vin2 declines, but the two modules still share the output current and system output voltage is still stable until t 2 . At t 2 , the duty cycle of module 2 reaches 1, and its inner current loop regulator saturates. Thus, the two modules cannot ensure the sharing of output current, and after a period of time the ISOP system enters the new steady state, where both the input voltages and output currents of the two modules are unbalanced. Due to the fact that the PSFB converter is a buck-derived topology, the input voltage of module 2 will not decline when it is not high enough to provide the desired output voltage. Figures 3.25 and 3.26 show the waveforms comparison of the ISOP system with three-loop control strategy (see Fig. 3.22) and general control strategy (see Fig. 3.11). Figure 3.25 shows the dynamic responses of input voltages of two modules and output voltage (AC coupled) for stepped input voltage between 500 and 600 Vdc. Figure 3.26 shows the dynamic responses of input voltages of two modules and output voltage (AC coupled) for stepped load current between 25 and 50 A. It can be seen that the input voltages are well shared, and the sharing of the output current is achieved

3.6 Experimental Verification of the Three-Loop Control Strategy

83

C2 vin1: [200V/div] C3 vin2: [200V/div] C2/ C3

M2 io1: [10A/div] M3 io2: [10A/div]

M2/ M3

C4 vo: [1V/div]

C4

t1 Time: [5ms/div]

t2

Fig. 3.24 Experimental waveforms of the ISOP system with voltage–current double closed-loop control strategy C1 vin:[500V/div] C1

C2

C2 vin1:[200V/div]

C3 vin2:[200V/div]

C3 C4

C1 vin:[500V/div] C1

C2 vin1:[200V/div]

C2

C3 vin2:[200V/div]

C3

C4 vo:[5V/div] Time:[20ms/div]

C4

C4 vo:[5V/div] Time:[20ms/div]

(a) General control strategy

(b) Three-loop control strategy

Fig. 3.25 Dynamic responses for stepped input voltage

automatically for the ISOP system regardless with the three-loop control strategy or the general control strategy. Furthermore, the ISOP system employing the three-loop control strategy has a better dynamic performance with smaller magnitude of output voltage overshoot and shorter regulation time in transient.

3.7 Summary In this chapter, the small-signal model of the PSFB-based ISOP system is derived. Based on the small-signal model, it is proved that the general control strategy proposed in Chap. 2 for n-module ISOP system can be decoupled into n single-input-

84

3 Mathematical Model and Closed-Loop Parameters Design … C1 io:[20A/div]

C1 io:[20A/div] C1

C2 vin1:[200V/div]

C1

C2 vin1:[200V/div]

C2

C2

C3 vin2:[200V/div]

C3 vin2:[200V/div]

C3

C3 C4

C4 vo:[2V/div] Time:[20ms/div]

C4 vo:[2V/div]

C4

Time:[20ms/div]

(a) General control strategy

(b) Three-loop control strategy

Fig. 3.26 Dynamic responses for stepped load current

single-output closed loops, which consists of n − 1 IVS closed loops and one output voltage closed loop. Moreover, the n − 1 IVS closed loops are decoupled from one another and are also decoupled from the output voltage closed loop. The parameters design of the IVS control loop regulator and output voltage control loop regulator based on a two-module PSFB ISOP system is discussed and verified by the prototype. Inner current loop can be introduced into the general control strategy to improve the system dynamic response and readily realize output current short-circuit protection. If the IVS loops are removed and only voltage–current double closed loop is used for the ISOP system, the system is unstable and the root cause is that the input impedance of each module appears as negative resistance characteristic. Hence, the IVS loops are indispensable for ensuring the system stability to change the negative resistance characteristic to the positive resistance characteristic, and this is also the design criteria IVS control loop regulator.

References 1. Ruan, X., Yan, Y.: Soft-switching techniques for PWM full bridge converters. In: Proceedings of the IEEE Power Electronics Specialists Conference, pp. 634–639 (2000) 2. Vlatkovic, V., Sabate, J.A., Ridley, R.B., Lee, F.C., Cho, B.H.: Small-signal analysis of the phase-shifted PWM converter. IEEE Trans. Power Electron. 17(1), 128–135 (1992) 3. Sabate, J.A., Vlatkovic, V., Ridley, R.B., Lee, F.C., Cho, B.H.: Design considerations for highvoltage high-power full-bridge zero-voltage-switched PWM converter. In: Proceedings of the IEEE Applied Power Electronics Conference and Exposition, pp. 275–284 (1990)

Chapter 4

Wireless IVS Control Strategies for Input-Series-Connected Systems Based on Positive Output Voltage Gradient Method

Abstract This chapter proposes the wireless input voltage sharing (IVS) control strategies for the input-series-connected systems based on positive output voltage gradient method, which increases the output voltage of each module with the increase of module input voltage. The constituent modules of input-series-connected systems with the proposed control strategy merely need to sense their own input/output voltages and there is no control interconnection among modules, leading to a truly modular design and high system reliability. The operation principle of the control strategy and the system stability are analyzed in this chapter. Experimental prototypes of the input-series-connected systems consisting of three two-transistor forward modules have been fabricated and tested to validate the proposed wireless IVS control strategies. Keywords Input voltage sharing · Input-series-connected systems Positive output voltage gradient method · Truly modular design Multi-module series–parallel power conversion systems can be categorized into IPOP, IPOS, ISOP, and ISOS systems, where ISOP and ISOS systems are the inputseries-connected systems. It is pointed out in Chap. 2 that only control strategies intending to achieve IVS for the input-series-connected systems are stable, and the OCS (for ISOP) and OVS (for ISOS) are achieved automatically. A modularization architecture for the input-series-connected systems was proposed by distributing the IVS control loops and output voltage control loop into the individual module and then interconnecting each module through a common duty cycle bus and an IVS bus to achieve IVS and the desired output voltage. This implies that the common duty cycle bus and IVS bus are essential for the system stable operation. When the buses are disrupted or failed, the operation of systems will be affected or even go to paralysis which reduces the system reliability to some extent. Therefore, eliminating the interconnection among the modules is the desired choice. This chapter proposes the wireless IVS control strategies for the input-seriesconnected systems based on positive output voltage gradient method, which increases the output voltage of each module with the increase of module input voltage. The

© Springer Nature Singapore Pte Ltd. and Science Press, Beijing, China 2019 X. Ruan et al., Control of Series-Parallel Conversion Systems, CPSS Power Electronics Series, https://doi.org/10.1007/978-981-13-2760-5_4

85

86

4 Wireless IVS Control Strategies for Input-Series-Connected …

constituent modules of input-series-connected systems with the proposed control strategy merely need to sense their own input/output voltages and there is no control interconnection among modules leading to a truly modular design and high system reliability [1, 2]. This chapter will analyze the operation principle of the control strategy and system stability. Experimental prototypes of the input-series-connected systems consisting of three two-transistor forward modules have been fabricated and tested to validate the proposed wireless IVS control strategies.

4.1 Genesis of the Wireless IVS Control Strategies Based on Positive Output Voltage Gradient Method The DC–DC ISOP system shown in Fig. 4.1 is taken as an example to derive the wireless IVS control strategy. In Chap. 2, the modularization architecture for DC–DC series–parallel power conversion systems is proposed (see Fig. 2.9). For the convenience of illustration, the control block diagram for ISOP system is given here, as shown in Fig. 4.2, where V inj (j 1, 2, …, n) is the input voltage of module j, vofj (j 1, 2, …, n) is the sensed output voltage of each module (also is the system output voltage), V oref is the output voltage reference, V RAMPj (j 1, 2, …, n) is the sawtooth signal of the modulator j, Gvo is the transfer function of output voltage control loop regulator and Gvcd is the transfer function of IVS control loop regulator. The control strategy of each module consists of an output voltage closed loop and an IVS closed loop. The signal of common duty cycle bus, vo_EA , is formed by interconnecting the output signal of the individual output voltage control loop regulator through a diode. The signal of IVS bus V in_ref is formed by interconnecting the input voltage of each module through a resistor R, and it is easy to know that V in_ref V in /n, where V in is the system input voltage. V in_ref serves as the reference voltage of each module input voltage. The error signal between V in_ref and V inj (j = 1, 2, …, n) is sent to the corresponding IVS control loop regulator and its output is subtracted from vo_EA and then compared with the sawtooth signal generating the duty cycle signal of each module. Obviously, the reference voltage V in_ref is variable when the system input voltage V in varies. If we set the V in_ref as a constant value, denoted as V inmod_ref , the IVS bus can be eliminated. However, if the system input voltage is not equal to n·V inmod_ref , the IVS control loop regulators will be saturated and fail to fulfill their function. To achieve IVS, the error signal between the reference voltage V inmod_ref and V inj (j 1, 2, …, n) can be amplified by a proportional coefficient k up and added to the reference output voltage V oref as shown in Fig. 4.3. The common duty cycle bus and corresponding diodes can also be removed and finally, the wireless IVS control strategy for ISOP system is obtained as shown in Fig. 4.4. It can be seen that only the input and output voltages are sensed for each module and there is no control interconnection among the constituent modules, which implies that a wireless controller is obtained and the system reliability can be significantly enhanced. Moreover, all the modules

4.1 Genesis of the Wireless IVS Control Strategies Based …

Iin

87

Iin1 + Vin1 C _ d1

Io1 DC-DC Module 1

Cf1

Iin2 Vin

+ Vin2 C _ d2

+ RLd Vo _

Io2 DC-DC Module 2

Cf2

Iinn + Vinn C _ dn

Io

Ion DC-DC Module n

Cfn

Fig. 4.1 Circuit diagram of ISOP system

are identical and can operate in stand-alone mode, which means that the loop design is very simple compared with the other control strategies for the ISOP systems. A truly modular design can be realized for the ISOP system with the proposed control strategy. At steady state, the input voltage of each module varies in the range of [V inmin /n, V inmax /n], where V inmin and V inmax are the minimum and maximum system input voltages, respectively. If we set V inmod_ref to be a constant value, such as V inmin /n, the error signal (V inj − V inmod_ref ) (j 1, 2, …, n) will rise if the module input voltage increases from V inmin /n. Since the error signal is added to V oref , the system output voltage will increase. This means that the output voltage increases with the increase of input voltage and the output voltage has a positive gradient regulation characteristic. Here, k up is defined as gradient gain. The operation principle will be analyzed as follows. The ISOP system operates at steady state and the constituent modules share the system input voltage. Assuming a perturbation occurs on the input voltages, which causes V in1 to decrease and V in2 to increase and the input voltages of other modules keep unchanged. Thus, we have V in1 < V in /n < V in2 . Then, the real reference voltage of 1# module output voltage control loop, [V oref + k up1 (V in1 − V inmod_ref )], will lower than that of 2# module output voltage control loop, [V oref + k up2 (V in2 − V inmod_ref )]. Due to the fact that the output sides of the modules are connected in parallel and all the modules have the same output voltage, for the controller of module 1, the output voltage is considered to be higher than its corresponding reference voltage, and the duty cycle will be regulated to reduce the input power of module 1, I in1 will decrease, then the input voltage of module 1 increases. In the meanwhile, for the controller of module 2, the

88

4 Wireless IVS Control Strategies for Input-Series-Connected …

R + Voref

Vin1 _

_ Gvcd

+ _

+

Gvo

Comparator 1 − +

RS Flip-flop

Drive circuit 1

VRAMP 1

vof1

Vin_ref

Module 1 R

+ Voref

Vin2 _

_

Gvcd

+ _

+

Comparator 2 − +

RS Flip-flop

VRAMP 2

Gvo vof2

R + vo_EA Voref

Module 2

Vinn _

_ Gvcd

+ _

Drive circuit 2

+

Gvo

Comparator n − RS Flip-flop +

VRAMP n

vofn

Drive circuit n Module n

Fig. 4.2 Modularization control block diagram of ISOP system

+

Vin1 _ kup1

Vinmod_ref

+

_ vof1 _

V_in2

+

Gvo

Voref

+ Vinmod_ref

Vinn _

+ Voref

Gvo

Comparator 2 − RS Flip-flop +

Comparator n − +

Drive circuit 1 Module 1

VRAMP 2

_ v _ ofn kupn

Comparator 1 − RS Flip-flop +

VRAMP 1

_ vof2 _ kup2

Vinmod_ref

+ Voref

Gvo

Drive circuit 2 Module 2

RS Flip-flop

VRAMP n

Fig. 4.3 Simplified modularization control block diagram of ISOP system

Drive circuit n Module n

4.1 Genesis of the Wireless IVS Control Strategies Based …

Vin1 _ + Vinmod_ref

kup1

_ vof1 + Gvo + Voref

Vin2 _ + Vinmod_ref

kup2

VRAMP 1

_ vof2 + Gvo + Voref

Vinn _ + Vinmod_ref

kupn

+ Voref

Drive circuit 1 Module 1

Comparator 2 − RS Flip-flop +

VRAMP 2

_ vofn +

Comparator 1 − RS Flip-flop +

89

Drive circuit 2 Module 2

Comparator n Gvo VRAMP n

− +

RS Flip-flop

Drive circuit n Module n

Fig. 4.4 Wireless IVS control strategy of ISOP system

output voltage is considered to be lower than its corresponding reference voltage, and its duty cycle will be regulated to increase the input power of module 2, I in2 will increase, then the input voltage of module 2 decreases. Finally, two modules return to the steady state. From the above analysis, for the ISOP system employing the wireless IVS control strategy, the module input current increases when the module input voltage increases, and the module input current decreases when the module input voltage decreases. Thus, the input impedance of the module appears as positive resistance characteristic and the control strategy is stable. According to Fig. 4.4, the output voltage of module j at steady state is Vo j

1 Voref + kup j · Vin j − Vin mod_ref ( j 1, 2, . . . , n) K vo

(4.1)

where K vo is the output voltage sensing gain. Defining V omin V oref /K vo and setting V inmod_ref V inmin /n, (4.1) can be rewritten as kup j Vinmin Vin j − Vo j Vomin + (4.2) ( j 1, 2, . . . , n) K vo n According to (4.2), the curve of the output voltage versus module input voltage can be depicted as shown in Fig. 4.5. Obviously, each module has a positive output voltage gradient regulation characteristic.

90

4 Wireless IVS Control Strategies for Input-Series-Connected …

Fig. 4.5 Positive output voltage gradient regulation characteristic

Voj Vomin

Vinmin /n

Vinj

4.2 Basic Features of the Input-Series-Connected Systems with Wireless IVS Control Strategies 4.2.1 ISOP System For simplicity, a two-module ISOP system is taken as an example to discuss the basic features of the wireless IVS control strategies. According to (4.2), the output voltages of two modules can be expressed as kup1 Vinmin (4.3) · Vin1 − Vo1 Vo1min + K vo 2 kup2 Vinmin (4.4) Vo2 Vo2min + · Vin2 − K vo 2 It should be noted that here, V o1min and V o2min are used for two modules, respectively, because the reference voltages of the two modules may be mismatched. For a two-module ISOP system, we have Vin Vin1 + Vin2

(4.5)

Vo1 Vo2 Vo

(4.6)

According to (4.3)–(4.6), we have Vinmin kup2 Vin + K vo (Vo2min − Vo1min ) + kup1 − kup2 2 Vin1 kup1 + kup2 Vinmin kup1 Vin + K vo (Vo1min − Vo2min ) + kup2 − kup1 2 Vin2 kup1 + kup2 kup1 Vo2min + kup2 Vo1min kup1 kup2 Vinmin Vo − kup1 + kup2 K vo kup1 + kup2 kup1 kup2 Vin + K vo kup1 + kup2

(4.7)

(4.8)

(4.9)

4.2 Basic Features of the Input-Series-Connected Systems …

91

Combining (4.7) and (4.8), the input voltage sharing error is expressed as kup2 − kup1 (Vin − Vinmin ) + 2K vo (Vo2min − Vo1min ) Vin1 − Vin2 Vin12 Vin Vin kup1 + kup2 Vin (4.10) Equation (4.10) implies that the input sharing error is determined by the gradient gains, minimum input voltage, and minimum output voltages of the two modules. According to (4.9), the gradient gain of ISOP system is kup1 kup2 1 dVo 1 1 1 dVin K vo K vo kup1 + kup2 + kup1 kup2

(4.11)

Likewise, the gradient gain of ISOP system consisting of n modules is 1 1 dVo 1 1 dVin K vo 1 + + ··· + kup1 kup2 kupn

(4.12)

It can be seen that the gradient gain of ISOP system is only determined by the module gradient gains, and the larger the n, the smaller the gradient gain of ISOP system, which means that the system has better output voltage regulation performance. In particular, when kup1 kup2 · · · kupn kup , the gradient gain of ISOP system is equal to k up /(nK vo ). Figure 4.6 shows the dependence of the IVS accuracy on the mismatch of output voltage set-point (V o1min and V o2min ) for a two-module ISOP system, and Fig. 4.7 shows the dependence of the IVS accuracy on the output voltage gradient gain. As seen in Fig. 4.6, the IVS accuracy is improved as the output voltage set-point mismatching decreases, and when the output voltage gradient gains are perfectly matched, as shown in Fig. 4.6c, the two modules share the input voltage evenly. From Fig. 4.7, it can be seen that a larger output voltage gradient gain results in better IVS performance, but the system output voltage regulation performance is deteriorated. According to (4.10), (4.11), Figs. 4.6 and 4.7, we can obtain the following conclusions: the larger the module gradient gain, the better the IVS characteristic, but the system output voltage regulation performance is deteriorated. On the contrary, the smaller the module gradient gain, the worse the IVS characteristic, but the system output voltage regulation performance is improved. Therefore, a trade-off must be made between the IVS and the output voltage regulation performance when designing the control circuit. Hence, the proposed control strategy is more suitable for the applications with a narrow input voltage range.

92

4 Wireless IVS Control Strategies for Input-Series-Connected …

Vo

Vo 1# 2#

1# 2#

Vo= Vo1= Vo2

Vin1

Vin/2

Vin2

Vo= Vo1= Vo2

Vin1 Vin/2 Vin2

Vin

(a) Large mismatch

Vin

(b) Small mismatch

Vo 1#, 2# Vo= Vo1= Vo2 Vin

Vin1=Vin2=Vin/2

(c) No mismatch Fig. 4.6 Effect of output voltage set-point mismatching on IVS accuracy

1# 2#

Vo

Vo 1# 2#

Vo= Vo1= Vo2

Vin1 Vin/2

Vin2

(a) Large slope

Vo= Vo1= Vo2

Vin1 Vin/2 Vin2

Vin

Vin

(b) Small slope

Fig. 4.7 Effect of output voltage gradient gain on IVS accuracy

4.2.2 ISOS System Likewise, the wireless IVS control strategy for ISOS system based on positive output voltage gradient method is shown in Fig. 4.8. It is worth noting that the feedback signal for each module is the system output voltage V o .

4.2 Basic Features of the Input-Series-Connected Systems …

Vin1 _ + Vinmod_ref

kup1

_ vof + +

Gvo

Voref Vin2 _ + Vinmod_ref

kup2

_ vof + +

VRAMP 1

Gvo

Voref

Vinn _ + Vinmod_ref

kupn

+ Voref

Comparator 2 − RS Flip-flop +

VRAMP 2

_ vof +

Comparator 1 − RS Flip-flop +

Gvo

Comparator n − RS Flip-flop +

VRAMP n

93

Drive Circuit 1 Module 1

Drive circuit 2 Module 2

Drive Circuit n Module n

Fig. 4.8 Wireless IVS control strategy for ISOS system

A two-module ISOS system is also taken as an example to analyze its features. Similar to ISOP system, we set V omin V oref /K vo and V inmod_ref V inmin /n according to Fig. 4.8, we have kup1 kup2 Vinmin Vinmin Vo2min + (4.13) · Vin1 − · Vin2 − Vo Vo1min + K vo 2 K vo 2 It should be noted that here, V o1min and V o2min are used for two modules, respectively, because the reference voltages of the two modules may be mismatched. For ISOS system, V in V in1 + V in2 . By combining this expression and (4.13), we have K vo (Vo2min − Vo1min ) + kup2 Vin − Vinmin 2 + kup1 Vinmin 2 (4.14) Vin1 kup1 + kup2 K vo (Vo1min − Vo2min ) + kup1 Vin − Vinmin 2 + kup2 Vinmin 2 Vin2 (4.15) kup1 + kup2 kup1 Vo2min + kup2 Vo1min kup1 kup2 (Vin − Vinmin ) Vo + (4.16) kup1 + kup2 K vo kup1 + kup2 According to (4.14) and (4.15), the input voltage sharing error is obtained and is expressed as

94

4 Wireless IVS Control Strategies for Input-Series-Connected …

2K vo (Vo2min − Vo1min ) + kup2 − kup1 (Vin − Vinmin ) Vin12 Vin1 − Vin2 Vin Vin kup1 + kup2 Vin (4.17) Equation (4.17) implies that the input sharing error is determined by the gradient gains (k up1 and k up2 ) and minimum output voltages of the two modules (V o1min and V o2min ). According to (4.16), the gradient gain of two-module ISOS system is kup1 kup2 1 dVo 1 1 dVin K vo 1 K vo kup1 + kup2 + kup1 kup2

(4.18)

It can be seen that the gradient gain of two-module ISOS system is the same as that of two-module ISOP system. Similar to ISOP system, we also can obtain the following conclusions for ISOS system: the larger the module gradient gain, the better the IVS characteristic, but the system output voltage regulation performance is deteriorated. On the contrary, the smaller the module gradient gain, the worse the IVS characteristic, but the system output voltage regulation performance is improved.

4.3 System Stability The purpose of this section is to study the stability of the ISOP system with the wireless IVS control strategy. With no loss of generality and for ease of analysis, the ISOP system under study consists of two two-transistor forward converter modules as shown in Fig. 4.9, where C d1 and C d2 are two input divided capacitors, power switches S 1 and S 2 , diodes D1 and D2 , transformer T r1 , rectifier diodes DR1 and DR2 , and output filter inductor L f1 form two-transistor forward converter 1, power switches S 3 and S 4 , diodes D3 and D4 , transformer T r2 , rectifier diodes DR3 and DR4 , and output filter inductor L f2 form two-transistor forward converter 2, C f is the output filter capacitor. The corresponding small-signal model the ISOP system is shown in Fig. 4.10 [3]. According to Fig. 4.10, we have K1 ˆ Io 1 ˆ dy1 i in − sCd1 vˆin1 − iˆo1 (4.19) Dy1 K1 ˆi o2 K 2 iˆin − sCd2 vˆin2 − Io 2 dˆy2 (4.20) Dy2 K2

RLd vˆo iˆo1 + iˆo2 (4.21) sCf RLd + 1

4.3 System Stability

S1

D2

Cd2

+ Vin2 _

S3

D4

+ Cf

DR2

Vo _

S2 DR3

Iin2

RLd

io1

Tr1

D3

K2:1 *

Vin

K1:1

Lf1

Lf2 io2

*

+ Vin1 _

D R1 D1

*

Cd1

Iin1

*

Iin

95

Tr2

DR4

S4

Fig. 4.9 Circuit diagram of two-module ISOP system

iˆin

1:

Dy1

+ Cd1 vˆin1 I o1 ˆ _ d y1 K1

Vin1 ˆ d y1 K1

1: + Cd2 ˆvin2 I o2 ˆ _ d y2 K2

iˆo

RLd +

iˆo1 Cf

vˆo _

Dy2 K2

Lf2

+ _

vˆin

Lf1

+ _

K1

Vin2 ˆ d y2 K2

iˆo2

Fig. 4.10 Small-signal model of the ISOP system shown in Fig. 4.9

where dˆy1 and dˆy2 are the perturbations of the duty cycles of two modules, and vˆin1 and vˆin2 are the perturbations of individual input voltages, iˆo1 and iˆo2 are the perturbations of individual output currents, K 1 and K 2 are turns ratios of the individual transformers, Dy1 , Dy2 , I o1, and I o2 are steady-state duty cycles and output currents, respectively. Assuming k up1 k up2 k up , K 1 K 2 K, and Dy1 Dy2 Dy , we have I o1 I o2 V o /(2RLd ) and V in1 V in2 V in /2. From Fig. 4.4, the expressions of dˆy1 and dˆy2 are obtained as

96

4 Wireless IVS Control Strategies for Input-Series-Connected …

dˆy1 G vo kup vˆin1 − K vo vˆo Vpp dˆy2 G vo kup vˆin2 − K vo vˆo Vpp

(4.22) (4.23)

where V pp is the peak-to-peak value of sawtooth signal V RAMP . Assuming the bandwidths of the output voltage closed loops are high enough, the perturbation of output voltage vˆo caused by the perturbation of input voltage vˆin can be expressed as vˆo

kup vˆin 2K vo

(4.24)

Substitution of (4.19), (4.20), (4.22), (4.23), and (4.24) into (4.21), yields kup Dy (sCf RLd + 1) vˆin 2iˆin − sCd1 vˆin1 − sCd2 vˆin2 2K vo K RLd

(4.25)

According to Fig. 4.9, the input power and output power of the ISOP system can be expressed as Pin + pˆ in (Vin + vˆin )(Iin + iˆin ) Vin Iin + Vin iˆin + Iin vˆin + vˆin iˆin Po + pˆ o

(Vo + vˆo ) 2Vo vˆo + + RLd RLd RLd RLd 2

Vo2

vˆo2

(4.26) (4.27)

where pˆ in and pˆ o are input and output power perturbations, respectively. Assuming the conversion efficiency is 100%, the input power is equal to the output power. So, according to (4.26) and (4.27), we have V 2 2Vo vˆo vˆ 2 + o Vin Iin + Vin iˆin + Iin vˆin + vˆin iˆin o + RLd RLd RLd

(4.28)

Neglecting the second-order terms of (4.28) and noting that the DC terms on both V2 1 sides of (4.28) are equal, and knowing that Iin o · , substitution of (4.24) into RLd Vin (4.28), gives kup Vin Vo − Vo2 K vo vˆin iˆin RLd Vin2

(4.29)

Substituting (4.29) into (4.25) and considering vˆin vˆin1 + vˆin2 yields vˆin1 vˆin vˆin2 vˆin

kup Vin Vo − K vo Vo2 kup Dy (sCf RLd + 1) −2 + sCd2 2K vo K RLd K vo R Ld Vin2 s(Cd2 − Cd1 ) kup Vin Vo − K vo Vo2 kup Dy (sCf RLd + 1) −2 + sCd1 2K vo K RLd K vo R Ld Vin2 s(Cd1 − Cd2 )

(4.30)

(4.31)

4.3 System Stability

97

From (4.30) and (4.31), the transfer function of the input voltage difference of the two modules to the total input voltage can be derived as vˆin12 vˆin1 − vˆin2 vˆin vˆin

s

kup Dy kup Vin Vo − K vo Vo2 kup Dy Cf −4 + Cd1 + Cd2 + K vo K K vo K RLd K vo RLd V 2 s(Cd2 − Cd1 )

in

(4.32) In practice, C d1 is impossible to equal to C d2 precisely. Hence, there is a root at the origin in the characteristic equation of the transfer function vˆin12 /vˆin , which means that the system is critically stable, i.e., if the input voltage difference exists between the two modules when there is a system input voltage perturbation, the difference will neither to converge nor to diverge. However, the actual system is surely damped by the parasitic resistance of the switches or inductors, so the perturbation will eventually converge and the ISOP system with the wireless IVS control strategy will be stable in practice.

4.4 Experimental Verification In order to verify the theoretical analysis in the previous sections, we have constructed the input-series-connected systems (ISOP and ISOS) prototype including three twotransistor forward converter modules in lab, and the specifications of each module are as follows: • • • • • •

input voltage V in 100–150 Vdc, output voltage V o 50 Vdc, rated output current I o 5 A, switching frequency f s 100 kHz, gradient gain k up 5.68 × 10−3 , sensing gain K vo 0.1.

4.4.1 Experimental Results of ISOP System The specifications of ISOP system are: system input voltage 300–450 Vdc, system output voltage 50 Vdc, system rated output current 15 A. Figure 4.11 shows the voltages across the primary windings of the individual transformers under normal input voltage (400 Vdc) and full load. It can be seen that all the voltages have almost the same positive amplitudes, which means that the module input voltages are equal and the module output current sharing automatically. It also should be noted that the switching signals of the three modules are not interleaved or synchronized to realize the goal of no control interconnection, and hence, the

98

4 Wireless IVS Control Strategies for Input-Series-Connected …

C1: Vtr1 [100V/div] C2: Vtr2 [100V/div]

C3: Vtr3 [100V/div] C4: Vo [50V/div] Time: [2μs/div] Ave

55

165

50

150

45

135

40

120 Vo Vin1 Vin2 Vin3

35 30 290

310

330

350

370 Vin (V)

390

410

430

Vin1 / Vin2 / Vin3 (V)

Vo (V)

Fig. 4.11 Voltages across the primary windings of individual transformers

105 90 450

Fig. 4.12 Measured system output voltage and three module input voltages

switching frequencies of the three modules are slightly different due to tolerance of RC parameters, which determine the switching frequency. Figure 4.12 shows the measured curves of system output voltage and input voltage of each module when ISOP system input voltage varies from 300 to 450 Vdc. It can be seen that the system output voltage increases with the increase of system input voltage, and the output voltage regulation is 5.6% in the whole system input voltage range from 300 to 450 Vdc. The three module input voltage curves are very close to each other. Due to small mismatches in the controller parameters among the three prototypes, such as reference voltages, voltage sensors, and operational amplifiers, the measured maximum input voltage difference among the three modules is 4 Vdc.

4.4 Experimental Verification

99

C1: Vin1 [100V/div]

C2: Vin2 [100V/div]

C3: Vin3 [100V/div] C4: Vo [50V/div] Time: [10ms/div] Ave

Ave

Fig. 4.13 Responses in individual input voltages and output voltage for stepped system input voltage

C1: Vin1 [100V/div]

C2: Vin2 [100V/div] C3 Vin3 [100V/div] C4: Io [10A/div] Time: [4ms/div] Ave

Ave

Fig. 4.14 Responses in individual input voltages to stepped load

Figure 4.13 shows the experimental waveforms of the input voltages of the three modules and the system output voltage corresponding to the system input voltage stepping between 300 and 450 Vdc. Figure 4.14 shows the experimental waveforms of the input voltages of the three modules corresponding to load stepping between full load (15 A) and half load (7.5 A). It can be seen that input voltages are well shared both at steady state and during transient and sharing of the output current is achieved automatically.

100

4 Wireless IVS Control Strategies for Input-Series-Connected …

C1: io1 [5A/div] C2: io2 [5A/div]

C3: io3 [5A/div]

C4 Vo [50V/div] Time: [40ms/div]

(a) Output currents of each module

C2: iLf2 [5A/div]

C1: iLf1 [5A/div]

C3: iLf3 [5A/div]

C4: Vo [50V/div] Time: [40ms/div]

(b) Output filter inductor currents of each module Fig. 4.15 Responses in individual output currents and output filter inductor currents to stepped load

Figure 4.15 shows the experimental waveforms of the output currents and output filter inductor currents of the three modules corresponding to the load stepping between full load (15 A) and half load (7.5 A). It can be seen that the load sharing is achieved both at steady state and during transient. In order to verify the redundancy of the ISOP system with the wireless IVS control strategy (see Fig. 4.4), a fault is imitated in the three-module ISOP system by shorting the input capacitor of one module with a switch in series with a 0.5 resistor, which

4.4 Experimental Verification

101

C1: Vin1 [100V/div]

C2: Vin2 [100V/div]

C3: Vin3 [100V/div] C4: Vo [50V/div] Time: [10ms/div] Ave

(a) Module 1 is isolated

C1: Vin1 [100V/div] C2: Vin2 [100V/div]

C3: Vin3 [100V/div]

C4: Vo [50V/div] Time: [20ms/div] Ave

(b) Module 1 is inserted Fig. 4.16 Response of individual input voltages and output voltage when module 1 is isolated and inserted, respectively

limits the discharging current of the capacitor. Figure 4.16 shows the system response to a fault of module 1. Figure 4.16a shows the input voltages of the three modules and the system output voltage corresponding to a short fault of module 1 under system input voltage V in 310 Vdc and full load condition. It can be seen that the system input voltage is evenly shared by the remaining two modules when the short switch

102

4 Wireless IVS Control Strategies for Input-Series-Connected …

C1: vtr1 [100V/div]

Time : [4μs/div]

C2: vtr2 [100V/div]

C4: Vo [100V/div]

C3: vtr3 [100V/div]

Fig. 4.17 Voltages across the primary windings of individual transformers

is closed. The system output voltage increases slightly because the input voltage changes from about 100–150 Vdc for module 2 and module 3 and the output voltage increases according to the positive output voltage regulation. Figure 4.16b shows the system response when module 1 is inserted into the system by opening the short switch under system input voltage V in 310 Vdc and full load condition. After opening the short switch, the input capacitor of module 1 is charged by the system input current and the input voltage increases, and in the meanwhile, the input voltages of modules 2 and 3 decrease. After a certain amount of regulation time, the three modules share the system input voltage and the output voltage is restored to be stable. The system output voltage decreases slightly because the input voltage changes from about 155–100 Vdc for modules 2 and 3 and the output voltage decreases according to the positive output voltage regulation. It can be seen that without any control interconnection among the modules, redundancy in combination with hot-swap capability of modules can be easily achieved.

4.4.2 Experimental Results of ISOS System The specifications of ISOS system are: system input voltage 300–450 Vdc, system output voltage 150 Vdc, system rated output current 5 A. Figure 4.17 shows the experimental waveforms of the voltages across the primary windings of individual transformers under normal input voltage (400 Vdc) and full load. As seen, all the voltages have almost the same positive amplitudes, which means that the module input voltages are equal. Figures 4.18 and 4.19 show the experimental waveforms of the ISOS system corresponding to stepped input voltage and stepped load, respectively. Figure 4.18

4.4 Experimental Verification

103

C1: Vin1 [100V/div] C2: Vin2 [100V/div]

C3: Vin3 [100V/div]

C4: Vo [100V/div] Time : [20 ms/div] Ave

Ave

(a) Individual input voltages and system output voltage

C1: Vin1 [100V/div] C2: Vo1 [50V/div]

C3: Vo2 [50V/div]

C4: Vo3 [50V/div] Time : [20 ms/div] Ave

(b) Individual output voltages and input voltage of module 1 Fig. 4.18 Responses to stepped input voltage

shows the input voltages and output voltages of the three modules and the system’s output voltage corresponding to an input voltage stepping between 300 and 450 Vdc under full load condition. Figure 4.19 shows the input voltages and output voltages of the three modules corresponding to a load stepping between half load (2.5 A) and full load (5 A) when system input voltage is 400 Vdc. It can be seen that input voltages are well shared both at steady state and during transient and the sharing of the output voltage is achieved automatically.

104

4 Wireless IVS Control Strategies for Input-Series-Connected …

C1: Vin1 [100V/div] C2: Vin2 [100V/div]

C3: Vin3 [100V/div]

C4: Io [5A/div] Time : [10 ms/div] Ave

(a) Individual input voltages and output current

C1: Vo1 [50V/div] C2: Vo2 [50V/div]

C3: Vo3 [50V/div]

C4: Io [5A/div] Time : [10 ms/div] Ave

(b) Individual output voltages and output current Fig. 4.19 Responses to stepped load

4.5 Summary

105

4.5 Summary In order to eliminate the control signal interconnection among the modules and improve the system reliability, the wireless IVS control strategies based on positive output voltage gradient method are proposed for the input-series-connected systems in this chapter. Truly modular input-series-connected systems can be achieved with the proposed control strategies, where all the modules are totally identical for both power stages and control stages. Each module is self-contained and no extra supervisory controller is needed to achieve the input voltage sharing among the modules. Moreover, there is no any control interconnection among the modules, which considerably increases the redundancy with the improved system reliability and maintainability. Experimental results of three-module ISOP and ISOS systems verify the effectiveness of the wireless IVS control strategies.

References 1. Chen, W., Wang, G., Ruan, X., Jiang, W., Gu, W.: Wireless input-voltage-sharing control strategy for input-series output-parallel (ISOP) system based on positive output-voltage gradient method. IEEE Trans. Ind. Electron. 61(11), 6022–6030 (2014) 2. Chen, W., Wang, G.: Decentralized voltage sharing control strategy for fully modular inputseries output-series system with improved voltage regulation. IEEE Trans. Ind. Electron. 62(5), 2777–2787 (2015) 3. Giri, R., Choudhary, V., Ayyanar, R., Mohan, N.: Common-duty-ratio control of input-series connected modular DC-DC converters with active input voltage and load-current sharing. IEEE Trans. Ind. Appl. 42(4), 1101–1111 (2006)

Chapter 5

A General Control Strategy for DC–AC Series–Parallel Power Conversion Systems

Abstract For the DC–AC series–parallel power conversion systems, the output power balance includes not only the equaling of the amplitude and frequency of the modular output voltages and currents but also the equaling of the phase of them. This means that the equilibrium of both the real and reactive powers should be actualized. In the light of this perspective, this chapter first presents a systematical analysis of the relationship between the input voltage/current sharing (IVS/ICS) and output voltage/current sharing (OVS/OCS) of four kinds of DC–AC series–parallel power conversion system. Then, the control strategies to achieve IVS/ICS and OVS/OCS and the system stability are discussed. Finally, a general control strategy for the DC–AC series–parallel power conversion system is presented. Keywords DC–AC series–parallel power conversion systems Input voltage sharing · Input current sharing · Output voltage sharing Output current sharing · Stability · General control strategy Similar to the DC–DC series–parallel power conversion systems, the DC–AC series–parallel power conversion systems, which are composed of multiple DC–AC inverter modules, can also be divided into four kinds, namely, input-parallel-outputparallel (IPOP), input-parallel-output-series (IPOS), input-series-output-parallel (ISOP), and input-series-output-series (ISOS). To guarantee normal operation of these systems, power balance among the constituent DC–AC inverter modules should be ensured. That is to say, input voltage sharing (IVS) and output voltage sharing (OVS) should be achieved when the inverter modules are connected in series at the input and output sides, respectively, and input current sharing (ICS) and output current sharing (OCS) should be realized when the inverter modules are connected in parallel at the input or output sides, respectively. Different from the DC–DC converters, the output of the DC–AC inverters is in the form of alternating current. Therefore, in the DC–AC series–parallel power conversion systems, the output power balance includes not only the equaling of the amplitude and frequency of the modular output voltages and currents but also the equaling of the phase of them. This means that the equilibrium of both the real and reactive powers should be actualized. In this chapter, the relationship between the input © Springer Nature Singapore Pte Ltd. and Science Press, Beijing, China 2019 X. Ruan et al., Control of Series-Parallel Conversion Systems, CPSS Power Electronics Series, https://doi.org/10.1007/978-981-13-2760-5_5

107

108

5 A General Control Strategy for DC–AC Series–Parallel …

voltage/current sharing (IVS/ICS) and output voltage/current sharing (OVS/OCS) of the four kinds of DC–AC series–parallel power conversion systems will be analyzed systematically. Also, the control strategies to achieve IVS/ICS and OVS/OCS and the stability are discussed. Then, a general control strategy for the four kinds of DC–AC series–parallel power conversion system is proposed [1, 2].

5.1 Relationship Between Input Voltage/Current Sharing and Output Voltage/Current Sharing Figure 5.1 shows the four connection architectures of DC–AC series–parallel power conversion system with n DC–AC inverter modules. As shown, for the jth module (j 1, 2, …, n), V in_j and I in_j denote the input voltage and the input current, respectively; vo_j and io_j denote the output voltage and the output current, respectively; I cd_j and icf_j denote the input and output filter capacitor currents, respectively; and I cin_j and ico_j denote the input and output currents of the power stage. At steady state, since both the input voltage and the output voltage are constant, the average values of both the input filter capacitor current and the output filter capacitor current are zero. Therefore, we have Iin_ j Icin_ j ( j 1, 2, . . . , n)

(5.1)

i o_ j i co_ j ( j 1, 2, . . . , n)

(5.2)

Then, the input power of each DC–AC inverter module can be derived as ⎧ ⎪ P Vin_1 · Iin_1 ⎪ ⎪ in1 ⎪ ⎪ ⎨ Pin2 Vin_2 · Iin_2 ⎪ ⎪ ⎪ ⎪ ⎪ ⎩P

inn

.. . Vin_n · Iin_n

(5.3)

And the output real power and reactive power could be expressed as ⎧ ⎪ Po1 Vo_1 · Io_1 · cos θ1 ⎪ ⎪ ⎪ ⎪ ⎨ Po2 Vo_2 · Io_2 · cos θ2 ⎪ ⎪ ⎪ ⎪ ⎪ ⎩

.. . Pon Vo_n · Io_n · cos θn

(5.4)

5.1 Relationship Between Input Voltage/Current Sharing … Iin Vin

Iin_1 Icin_1 ico_1 io_1 + Icd_1 + icf_1 DC-AC vo_1 Cd1 Cf1 Vin_1 inverter _ _ _

io ZLd

+ vo _ Vin

Iin

Module 1

109

Iin_1 Icin_1 ico_1 io_1 io + Icd_1 + icf_1 DC-AC v C C d1 f1 o_1 Vin_1 inverter _ __ Module 1

Iin_2 Icin_2 ico_2 io_2 + Icd_2 + icf_2 DC-AC vo_2 C C d2 f2 Vin_2 inverter _ __

Iin_2 Icin_2 ico_2 io_2 + Icd_2 + icf_2 + DC-AC vo_2 ZLd vo Cd2 Cf2 Vin_2 inverter _ __ _

Module 2

Module 2

Iin_n Icin_n ico_n io_n + Icd_n + icf_n DC-AC Vin_n vo_n Cdn Cfn inverter _ _

Iin_n Icin_n ico_n io_n + Icd_n + icf_n DC-AC Vin_n vo_n Cdn Cfn inverter _ _

Module n

Module n

(a) IPOP Iin

(b) IPOS

Iin_1 Icin_1 ico_1 io_1 + Icd_1 + icf_1 DC-AC Cd1 Cf1 vo_1 Vin_1 inverter _ __

io ZLd

Iin

+ vo _

Iin_1 Icin_1 ico_1 io_1 io + Icd_1 + icf_1 DC-AC Cd1 Cf1 vo_1 Vin_1 inverter _ __

Module 1

Vin

Module 1

Iin_2 Icin_2 ico_2 io_2 + Icd_2 + icf_2 DC-AC Cd2 Cf2 vo_2 Vin_2 inverter _ _ _ Module 2

Iin_n Icin_n ico_n io_n + Icd_n + icf_n DC-AC Vin_n vo_n Cdn Cfn inverter _ _

Vin

Iin_2 Icin_2 ico_2 io_2 + + Icd_2 + icf_2 DC-AC vo_2 ZLd vo Cd2 Cf2 Vin_2 inverter _ _ _ _ Module 2

Iin_n Icin_n ico_n io_n + Icd_n + icf_n DC-AC Vin_n Cdn Cfn vo_n inverter _ _

Module n

Module n

(c) ISOP

(d) ISOS

Fig. 5.1 Four connection architectures of DC–AC series–parallel power conversion system

⎧ ⎪ ⎪ Q o1 Vo_1 · Io_1 · sin θ1 ⎪ ⎪ ⎪ ⎨ Q o2 Vo_2 · Io_2 · sin θ2 ⎪ ⎪ ⎪ ⎪ ⎪ ⎩Q

on

.. . Vo_n · Io_n · sin θn

(5.5)

where V o_1 , V o_2 , …, V o_n are the root-mean-square (RMS) values of the output voltage of each module; I o_1 , I o_2 , …, I o_n are the RMS values of the output current of each module; and θ 1 , θ 2 , …, θ n are the output power factor angles of each module. Suppose the efficiencies of the constituent module are equal and the values are η. So, by power conservation, we have Po j η × Pin j ( j 1, 2, . . . , n). Then, combining (5.3) and (5.4), we can get

110

5 A General Control Strategy for DC–AC Series–Parallel …

⎧ ⎪ V · I · cos θ1 Vin_1 · Iin_1 · η ⎪ ⎪ o_1 o_1 ⎪ ⎪ ⎨ Vo_2 · Io_2 · cos θ2 Vin_2 · Iin_2 · η ⎪ ⎪ ⎪ ⎪ ⎪ ⎩

Vo_n · Io_n

.. . · cos θn Vin_n · Iin_n · η

(5.6)

5.1.1 IPOP Systems As shown in Fig. 5.1a, for DC–AC IPOP systems, the input voltage and the output voltage are equal, respectively, i.e., Vin_1 Vin_2 · · · Vin_n , and vo_1 vo_2 · · · vo_n . If input current sharing (ICS) among the DC–AC inverter modules is achieved, we have Iin_1 Iin_2 · · · Iin_n . Substitution of these three equations into (5.6) leads to Io_1 · cos θ1 Io_2 · cos θ2 · · · Io_n · cos θn

(5.7)

According to (5.7), if we wish to realize output current sharing (OCS) among the constituent DC–AC inverter modules, i.e., i o_1 i o_2 · · · i o_n

(5.8)

One of the following conditions should be satisfied, i.e., Io_1 Io_2 · · · Io_n

(5.9)

θ1 θ2 · · · θn

(5.10)

That is to say, according to (5.7), if (5.9) is valid, then (5.10) is valid; or if (5.10) is established, then (5.9) is established. According to the above analysis, it can be concluded that achieving ICS among the constituent modules can only ensure that the output real powers of the modules in IPOP systems are equal, and the reactive powers among the modules could not be guaranteed. In order to ensure the reactive powers among the modules are equal, the condition (5.9) or (5.10) should be valid, i.e., the amplitudes or phases of the output currents of the inverter modules should be equal. Thus, OCS could be ensured. This is quite different from that of DC–DC IPOP systems since the output of DC–DC converter is in the form of direct current and only the real power is involved. On the other hand, for the DC–AC IPOP systems, if OCS is ensured, (5.9) and (5.10) are valid. Substituting these two equations together with Vin_1 Vin_2 · · · Vin_n and vo_1 vo_2 · · · vo_n into (5.6) leads to Iin_1 Iin_2 · · · Iin_n , which means that ICS is achieved.

5.1 Relationship Between Input Voltage/Current Sharing …

111

5.1.2 IPOS Systems As shown in Fig. 5.1b, for IPOS systems, the input voltages and the output currents of the constituent modules are equal, respectively, i.e., Vin_1 Vin_2 · · · Vin_n , and i o_1 i o_2 · · · i o_n . If ICS of the constituent modules is achieved, we have Iin_1 Iin_2 · · · Iin_n . Substitution of these three equations into (5.6) yields Vo_1 · cos θ1 Vo_2 · cos θ2 · · · Vo_n · cos θn

(5.11)

From (5.11), if OVS among the modules are expected, i.e., vo_1 vo_2 · · · vo_n

(5.12)

One of the following equations should be satisfied, i.e., Vo_1 Vo_2 · · · Vo_n

(5.13)

θ1 θ2 · · · θn

(5.14)

That is to say, according to (5.11), if (5.13) is valid, then (5.14) is valid; or if (5.14) is established, then (5.13) is established. According to the above analysis, it can be concluded that achieving ICS among the modules can only ensure the output real powers of the modules in IPOS systems are equal, but the equaling of the reactive powers of the modules could not be guaranteed. For the purpose of ensuring the reactive powers of the modules are equal, the condition (5.13) or (5.14) should be valid, i.e., the amplitudes or phases of the output voltage of the modules should be equal. Thus, OVS could be guaranteed. This is quite different from DC–DC IPOS systems since the outputs of DC–DC converters are in the form of direct current and only real power is considered. On the other hand, for DC–AC IPOS systems, if OVS among the constituent modules is ensured, (5.13) and (5.14) are valid. Substituting these two equations together with Vin_1 Vin_2 · · · Vin_n and i o_1 i o_2 · · · i o_n into (5.6), we can obtain Iin_1 Iin_2 · · · Iin_n , which means that ICS is achieved.

5.1.3 ISOP Systems As shown in Fig. 5.1c, for ISOP systems, the input currents and the output voltages of the modules are equal, respectively, i.e., Iin_1 Iin_2 · · · Iin_n , and vo_1 vo_2 · · · vo_n . If IVS among the constituent modules is achieved, we have Vin_1 Vin_2 · · · Vin_n . Substitution of these three equations into (5.6) yields Io_1 · cos θ1 Io_2 · cos θ2 · · · Io_n · cos θn

(5.15)

112

5 A General Control Strategy for DC–AC Series–Parallel …

From (5.15), if we intend to achieve OCS among the modules, i.e., i o_1 i o_2 · · · i o_n

(5.16)

One of the following equations should be satisfied, i.e., Io_1 Io_2 · · · Io_n

(5.17)

θ1 θ2 · · · θn

(5.18)

That is to say, according to (5.15), if (5.17) is valid, then (5.18) is valid and vice versa. According to the above analysis, it can be concluded that achieving IVS among the modules can only ensure the module output real powers are equal, but the module reactive powers could not be guaranteed. If the condition (5.17) or (5.18) is valid, i.e., the amplitudes or phases of the output currents of the modules are equal, the module reactive powers could be equal, thus achieving OCS [1]. This is quite different from that of DC–DC ISOP system in which only the real power is involved. On the other hand, for DC–AC ISOP systems, if OCS among the modules is ensured, both (5.17) and (5.18) are valid. Substituting these two equations together with Iin_1 Iin_2 · · · Iin_n and vo_1 vo_2 · · · vo_n into (5.6), we can get Vin_1 Vin_2 · · · Vin_n , which means that IVS is achieved.

5.1.4 ISOS Systems As shown in Fig. 5.1d, for ISOS system, the input currents and the output currents of the modules are equal, respectively, i.e., Iin_1 Iin_2 · · · Iin_n , and i o_1 i o_2 · · · i o_n . If IVS among the modules is achieved, we have Vin_1 Vin_2 · · · Vin_n . Substitution of these three equations into (5.6) leads to Vo_1 · cos θ1 Vo_2 · cos θ2 · · · Vo_n · cos θn

(5.19)

From (5.19), if OVS among the modules is expected, i.e., vo_1 vo_2 · · · vo_n

(5.20)

One of the following equations should be satisfied, i.e., Vo_1 Vo_2 · · · Vo_n

(5.21)

θ1 θ2 · · · θn

(5.22)

That is to say, according to (5.19), if (5.21) is valid, then (5.22) is valid; or if (5.22) is established, then (5.21) is established.

5.1 Relationship Between Input Voltage/Current Sharing …

113

According to the above analysis, it can be concluded that achieving IVS among the modules can only ensure the module output real powers are equal, but the module reactive powers could not be guaranteed. If the condition (5.21) or (5.22) is valid, i.e., the amplitudes or phases of the output voltages of the modules are equal, the module reactive powers could be equal, achieving OVS [2]. This is quite different from that of DC–DC ISOS systems which have only real power. On the other hand, for DC–AC ISOS systems, if OVS among the modules is ensured, both (5.21) and (5.22) are valid. Substituting these two equations together with Iin_1 Iin_2 · · · Iin_n and i o_1 i o_2 · · · i o_n into (5.6), we can get Vin_1 Vin_2 · · · Vin_n , which means that IVS is achieved. According to the above analysis, we can find that, for DC–AC series–parallel power conversion systems, achieving IVS/ICS among the constituent modules can only ensure the module output real powers are equal, but the equaling of the module reactive powers could not be guaranteed. In order to achieve the output reactive power equilibrium, in addition to achieving IVS/ICS, the amplitudes or phases of the module output currents (in IPOP and ISOP systems) or the amplitudes or phases of the module output voltages (in IPOS and ISOS systems) should be ensured. Hereinafter, the control strategy that controls IVS/ICS and makes the amplitudes or phases of module output currents or module output voltages be equal is defined as the compound balanced control strategy [1, 2]. On the other hand, for DC–AC series–parallel power conversion systems, if OVS/OCS among the modules could be guaranteed, IVS/ICS will be ensured automatically.

5.2 Stability of the Control Strategies for DC–AC Series–Parallel Power Conversion Systems It is pointed out in Chap. 2 that, for DC–DC series–parallel power conversion systems, if IVS/ICS among the modules could be guaranteed, OVS/OCS will be ensured automatically, and vice versa. Moreover, it is further pointed out that the IVS/ICS control strategy could ensure the stable operation of all of four kinds of DC–DC series–parallel power conversion systems including IPOP, IPOS, ISOP, and ISOS systems. While the OVS/OCS control strategy could ensure the stable operation of the input-parallel-connected systems (including IPOP and IPOS systems), the operation of the input-series-connected systems (including ISOP and ISOS systems) is unstable. So, do the DC–AC series–parallel power conversion systems have the similar characteristics as mentioned above? In this section, we examine the stability of the systems with the possible control strategies discussed in Sect. 5.1. Before the analysis, we suppose that the system’s input voltage, output voltage, and load current are kept unchanged when perturbations are imposed on the input voltage/current and output voltage/current of the modules, and the modules have the same efficiency.

114

5 A General Control Strategy for DC–AC Series–Parallel …

5.2.1 Compound Balanced Control Strategy Regardless of the connection type at the input side of the DC–AC series–parallel power conversion system, if the compound balanced control strategy is adopted, each DC–AC inverter module behaves as a constant power source as seen at the output side, as shown in Fig. 5.2, and the modular input powers are equal. For the case where the output sides are connected in parallel, as shown in Fig. 5.2a, the output voltages of all the modules are equal. With the compound balanced control strategy, both the module output real powers and reactive powers are, respectively, equal. Thus, the module output currents are identical. So, no perturbation occurs, i.e., the system is stable. For the case where the output sides are connected in series, as shown in Fig. 5.2b, the output currents of all the modules are equal. Assume that a perturbation occurs in the output voltages of some two modules, e.g., the amplitude of vo_1 increases and the amplitude of vo_2 decreases (here we suppose the scheme of controlling the phase of output voltages to be equal is adopted), while the output voltages of other modules are unchanged. Since the system output current io is unchanged, the rise of the amplitude of vo_1 makes module 1 output more real power than other inverter modules. Hence, the output filter capacitor C f1 will be discharged, making the voltage amplitude of module 1 descend and return to the equilibrium point. In the meanwhile, the output filter capacitor C f2 will be charged, and the voltage amplitude of module 2 will increase and returns to the equilibrium point. In conclusion, with the compound balanced control strategy, all the four series— parallel power conversion systems could work stably.

ico_1 Constant Power Source (Inverter 1)

Constant Power Source (Inverter 2)

Constant Power Source (Inverter n)

io_1 icf_1

Cf1

+ vo_1 _

ico_2

io_2 icf_2

Cf2

+ vo_2 _

ico_n

io_n icf_n

Cfn

io ZLd

+ vo_n _

(a) Output sides connected in parallel

+ vo _

ico_1 Constant Power Source (Inverter 1)

Constant Power Source (Inverter 2)

Constant Power Source (Inverter n)

io_1 icf_1

Cf1

+ vo_1 _

ico_2

io_2 icf_2

io

Cf2

+ vo_2 ZLd _

ico_n

io_n icf_n

Cfn

+ vo_n _

(b) Output sides connected in series

Fig. 5.2 Equivalent schematic diagrams for the compound balanced control strategy

+

vo

_

5.2 Stability of the Control Strategies for DC–AC Series–Parallel … Iin

Vin

Iin_1

Icin_1

+ Icd_1 Vin_1 Cd1 _ Iin_2

Iin_2 Constant Power Sink (Inverter 2)

Icin_n

+ Icd_n Vin_n Cdn _

Iin_1

Vin

(a) Input-parallel-connected systems

Constant Power Sink (Inverter 1)

Icin_2

+ Icd_2 Vin_2 Cd2 _

Iin_n Constant Power Sink (Inverter n)

Icin_1

+ Icd_1 Vin_1 Cd1 _

Icin_2

+ Icd_2 Vin_2 Cd2 _

Iin_n

Iin Constant Power Sink (Inverter 1)

115

Constant Power Sink (Inverter 2)

Icin_n

+ Icd_n Vin_n Cdn _

Constant Power Sink (Inverter n)

(b) Input-series-connected systems

Fig. 5.3 Equivalent schematic diagrams for the output side control strategy

5.2.2 Output Voltage/Current Sharing (OVS/OCS) Control Strategy Regardless of the connection type at the output side for the DC–AC series–parallel power conversion systems, if the OVS/OCS control strategy is employed, each inverter module behaves as a constant power sink as seen at the input side, as shown in Fig. 5.3, and the modular output powers are equal (which means the active and reactive powers are, respectively, equal). It is noted that we can only see the input real power at the input side, so we could draw the same conclusion with the DC–DC series–parallel power conversion systems, i.e., the OVS/OCS control strategy could ensure stable operation of the input-parallel-connected systems (including IPOP and IPOS systems), but the operation of the input-series-connected systems (including ISOP and ISOS systems) are unstable. The detailed analysis has been given in Sect. 2.2.2 in Chap. 2 and it is not repeated here.

5.3 General Voltage/Current Sharing Control Strategy for DC–AC Series–Parallel Power Conversion Systems According to the analysis of the above two sections, we can draw the following conclusions: 1. For the input-parallel-connected DC–AC systems (including IPOP and IPOS systems), both compound balanced control strategy and OVS/OCS control strat-

116

5 A General Control Strategy for DC–AC Series–Parallel …

Fig. 5.4 Main circuit of buck-derived inverter

+

Vin

vinv –

Lf iLf iCf Cf

io + vo

ZLd

–

egy could ensure stable operation of the systems. With the compound balanced control strategy, the amplitudes or phases of the modular output currents (for IPOP system) or modular output voltages (for IPOS system) should be guaranteed together with controlling for ICS among the modules. Compared with the OVS/OCS control strategy, the compound balanced control strategy is quite complicated. Thus, the OCS control strategy is more convenient for DC–AC IPOP systems, and the OVS control strategy is preferred for IPOS DC–AC systems. 2. For the input-series-connected DC–AC systems (including ISOP and ISOS systems), OVS/OCS control strategy could not ensure stable operation of the systems. So, we can only adopt the compound control strategy, i.e., the equaling of the amplitude or phase of modular output current (for ISOP system) or modular output voltage (for ISOS system) should be guaranteed together with controlling for IVS among the modules. Regardless of the type of the connection architectures, at least two control loops are needed. One is the output voltage loop for regulating the system output voltage, and the other one is the input voltages/currents or output voltages/currents sharing loop, which will be referred to as voltage/current sharing loop hereinafter. With the buckderived inverter as the basic module, as shown in Fig. 5.4, a general voltage/current sharing control strategy for four kinds of DC–AC series–parallel power conversion systems is proposed, as shown in Fig. 5.5. Likewise, the general voltage/current sharing control strategy for the DC–AC series–parallel power conversion systems with boost-derived or buck–boost-derived converter as the basic module can be derived, and it is not repeated here. The general control strategy includes three closed loops, namely, the system output voltage loop, which is common for all the inverter modules, the voltage/current sharing loop, and the current inner loop of each module. Different from the twoloop control for the DC–DC series–parallel power conversion systems in Sect. 2.3 in Chap. 2, the current inner loop of each module is introduced here. This is because the equaling of the amplitude or phase of modular output current (for ISOP system) or modular output voltage (for ISOS system) should be guaranteed except for IVS among the constituent modules for the purpose of achieving the power balance for the input-series-connected DC–AC power conversion systems (including ISOP and ISOS systems). In order to make the phase of output current (for ISOP system) or output voltage (for ISOS system) of each module identical, the output of the voltage/current sharing loop of each module is multiplied by the output of the common system output voltage loop. In the three-loop control strategy, the system output voltage loop aims

5.3 General Voltage/Current Sharing Control Strategy … Voltage/current sharing loop Vsh_g _ V sh_f 1 + Gsh vsh_EA1 _ Vsh_f 2 +

if_1 −/+

ish_EA1

iref1

if_2 −/+

Gsh vsh_EA2

Current regulator 1

SPWM modulator

Drive signal 1

Current regulator 2

SPWM modulator

Drive signal 2

Current regulator (n−1)

SPWM modulator

Drive signal (n−1)

Current regulator n

SPWM modulator

Drive signal n

_

+

+

…

iref2

…

ish_EA2

_

+

+

if_(n−1)

_ Vsh_f (n−1) + Gsh vsh_EA(n−1) _ Vsh_f n +

−/+

if_n −/+

Gsh

+

_ vof

Gvo

_

+

+ iref(n−1)

ish_EA(n−1)

vsh_EAn

voref

117

+

+

ish_EAn

irefn

_

iref

System output voltage loop

if_ j

−/+

− ISOP and ISOS + IPOP and IPOS

io_ j

ISOP and IPOP

iCf_ j

IPOS and ISOS

iLf_ j

Four kinds of system

(j = 1, 2, …, n)

Fig. 5.5 General voltage/current sharing control strategy for DC–AC series–parallel power conversion systems

for regulating the system output voltage, and its output signal, iref , serves as the initial common reference for all the inner current loops. The voltage/current sharing loop of each module adjusts the magnitude of iref , leading to the final current reference for each module, irefj (j 1, 2, …, n). For the voltage/current sharing loop shown in Fig. 5.5, V sh_g represents the average value of the total system input/output voltage/current, and V sh_fj (j 1, 2, …, n) represents the sensed input or output variable of each module. The representations of V sh_g and V sh_fj for the four kinds of DC–AC series–parallel power conversion systems are given in Table 5.1. As shown in Table 5.1, for the input-series-connected DC–AC power conversion systems, V sh_g is V in /n, and V sh_fj is the sensed input voltage of each module; While for the input-parallel-connected DC–AC power conversion systems, if the compound balanced control is adopted, V sh_fj is I cin_j , and if the output side control is employed, V sh_fj is vo_j for IPOS system and io_j for IPOP system. Furthermore, for the IPOP system, if io_j is taken as the feedback signal of each current inner loop, all the module output currents will track the common cur-

118

5 A General Control Strategy for DC–AC Series–Parallel …

Table 5.1 Representation of V sh_fj and V sh_g in four connection architectures Architecture

V sh_fj (j = 1, 2, …, n)

V sh_g

IPOP

I cin_j

I in /n

IPOS

I cin_j or vo_j

I in /n or vo /n

ISOP

V in_j

V in /n

ISOS

V in_j

V in /n

rent reference iref , and OCS will be achieved without any additional voltage/current sharing loop. Thus, the voltage/current sharing loop could be removed. For the inner current loop of each module in Fig. 5.5, the current feedback signal if_j can be the output current io_j , the output filter inductor current iLf_j , and the output filter capacitor current iCf_j (j 1, 2, …, n) of each module. With different current feedback signals, the sharing of the output voltage/current for the four DC–AC series–parallel power conversion systems will be different. For example, for the ISOP system, if the output filter inductor current is chosen as the feedback signal, the sharing of the output filter inductor current is achieved, and there is output reactive power difference among the modules when the output filter capacitors are not well matched; and if the output current is chosen as the feedback signal, OCS could be achieved even if the output filter capacitors are not well matched [1]. This issue will be discussed in detail in Chap. 6. For the ISOS system, if the output filter inductor current is selected as the feedback signal, there is output reactive power difference among the modules when the output filter capacitors are not well matched; and if the output filter capacitor current is selected as the feedback signal, the sharing of the output voltage will be ensured no matter the output filter capacitors are matched or not [2], which will be elaborated specifically in Chap. 7.

5.4 Summary This chapter systematically analyzes the inherent relationships between the input voltages/current sharing and output voltages/current sharing of the DC–AC series— parallel power conversion systems from the view of power conservation, and discusses the system stability with the compound balanced control strategy and output side control strategy. Then, a general voltage/current sharing control strategy is proposed for the DC–AC series–parallel power conversion systems. (1) For the DC–AC series–parallel power conversion systems, controlling for the input voltage/current sharing (IVS/ICS) among the constituent modules can only ensure the equaling of the output real power of each module, but the equaling of the reactive power could not be guaranteed, i.e., output voltage/current sharing (OVS/OCS) among the constituent modules could not be achieved. This is quite

5.4 Summary

119

different from the DC–DC series–parallel power conversion systems since it only has the real power at the output sides. (2) For the DC–AC series–parallel power conversion systems, controlling at the output sides to achieve output voltage/current sharing (OVS/OCS) among the modules can ensure the module output real and reactive powers are equal, respectively. As a consequence, the input power balance among the modules is guaranteed, and the input voltage/current sharing (IVS/ICS) is achieved. This is the same as DC–DC series–parallel power conversion systems. (3) For the DC–AC series–parallel power conversion systems, the control strategy that controls the input voltage/current sharing (IVS/ICS) and makes the amplitudes or phases of module output currents or module output voltages be equal is defined as the compound balanced control, which can ensure stable operation, and the output side control strategy can only guarantee stable operation of the input-parallel-connected systems (including IPOP and IPOS systems) but cannot guarantee stable operation of the input-series-connected systems (including ISOP and ISOS systems). (4) For the input-parallel-connected DC–AC power conversion systems (including IPOP and IPOS systems), both the compound balanced control strategy and output voltage/current sharing control strategy are effective. Compared to the compound balanced control strategy, the output voltage/current sharing control is quite simple and it is usually employed. For the input-series-connected DC–AC power conversion systems (including ISOP and ISOS systems), only the compound balanced control strategy could be adopted, that is, the equaling of the amplitude or phase of the modular output current (for ISOP system) or modular output voltage (for ISOS system) should be guaranteed together with the control for achieving input voltage sharing (IVS) among the modules.

References 1. Chen, W.: Research on series-parallel conversion systems consisting of multiple converter modules (in Chinese). Ph.D. Dissertation, Nanjing University of Aeronautics and Astronautics, Nanjing, China (2009) 2. Fang, T.: Control strategies for input-parallel-output-parallel and input-series-output-series connected combined inverter systems (in Chinese). Ph.D. Dissertation, Nanjing University of Aeronautics and Astronautics, Nanjing, China (2008)

Chapter 6

Compound Balanced Control Strategy for DC–AC Input-Series Output-Parallel Systems

Abstract In this paper, the implementation of the compound balanced control strategy is presented for the DC–AC input-series output-parallel (ISOP) systems to achieve input voltage sharing (IVS) and output current sharing (OCS). The inductor current and the output current adopted as the feedback variables for the inner current loop of each module, respectively, is analyzed and compared. Taking the outputcurrent-feedback scheme as an example, the decoupling relationship between the control loops is analyzed and the design of the control loops is given. The simulation and experimental results are provided to verify the effectiveness of the proposed control method. Furthermore, this chapter presents a new solution dedicated to distributed configuration on the basis of centralized control, which can improve the reliability of the systems. Keywords Input-series output-parallel system Compound balanced control strategy · Input voltage sharing Output current sharing · Closed-loops design DC–AC input-series-connected systems (including DC–AC ISOP systems and DC–AC ISOS inverter systems) are very suitable for high input DC voltage applications such as the urban railway systems, ship-electric-power-distribution systems, high-speed railway electrical systems, etc. The DC–AC input-series-connected systems have the following advantages: (1) easy selection of power switches due to the reduced voltage stress, (2) easy design due to the reduced power level of individual modules, and (3) high modularity and reliability [1, 2]. The DC–AC ISOP systems is quite suitable for high input voltage and high output current power conversion applications, and the primary control objective of the systems is to ensure input voltage sharing (IVS) and output current sharing (OCS) for all the constituent modules. Unlike DC–DC ISOP system, OCS for DC–AC ISOP system means both the magnitudes and phases of the modular output currents are equal due to AC output of DC–AC ISOP systems. In Chap. 5, it is pointed out that DC–AC ISOP system needs to adopt the compound balanced control strategy, i.e., to achieve IVS while ensuring that the magnitudes or phases of modular currents are equal. In this chapter, the implementation of the © Springer Nature Singapore Pte Ltd. and Science Press, Beijing, China 2019 X. Ruan et al., Control of Series-Parallel Conversion Systems, CPSS Power Electronics Series, https://doi.org/10.1007/978-981-13-2760-5_6

121

122

6 Compound Balanced Control Strategy for DC–AC …

compound balanced control strategy is presented first. In this method, either the inductor current or the output current can be adopted as the feedback variable for the inner current loop. The comparative analysis shows that the method adopting the output current as the feedback variable can realize IVS and OCS unconditionally. Then, taking the output-current-feedback method as an example, the decoupling relationship between the control loops is analyzed and the loop design procedure is given. Finally, the simulation and experimental results are provided to verify the effectiveness of the proposed control method. Furthermore, this chapter put forward a new solution dedicated to distributed configuration on the basis of centralized control, which can improve the reliability of the systems.

6.1 DC–AC Inverter Topology for the ISOP Systems In the DC–AC ISOP systems, the DC–AC inverter modules are connected in series at the input side and in parallel at the output side. This requires the DC–AC inverter modules that should have the function of galvanic isolation, which could be realized by a low-frequency transformer or a high-frequency transformer [3–5] as shown in Fig. 6.1. As shown in Fig. 6.1a, a line-frequency transformer is introduced at the output side of the DC–AC inverter. In addition to galvanic isolation, this transformer can step up/down the output voltage. However, the line-frequency transformer is bulky and heavy since the frequency is quite low, which is 50 Hz or 60 Hz. As shown in Fig. 6.1b, the isolated DC–AC inverter is a two-stage configuration. The first stage is an isolated DC–DC converter, which converts a DC input voltage

*

*

Vin

DC-AC Inverter

+ vo

Line-Frequency Transformer

_

(a) Galvanic isolation realized by alow-frequency transformer

*

* Vin

+

DC-DC Converter with High-Frequency Transformer

DC-AC Inverter

vo _

(b) Galvanic isolation realized by ahigh-frequency transformer Fig. 6.1 Isolated DC–AC inverter

6.1 DC–AC Inverter Topology for the ISOP Systems

123

into a regulated DC output voltage. Since the DC–DC converter is operated at high frequency, the transformer is quite small and light. The second stage is a DC–AC inverter, which converts the DC voltage obtained from the first stage into an AC output voltage. With the two-stage configuration, the DC–DC converter and DC–AC inverter could be optimally designed separately, and it is also contributed to the integration and modularization of the system. So, the high-frequency isolated inverter is chosen here as the basic inverter module. The isolated DC–DC converters can be buck-derived, boost-derived, and buckboost-derived ones. The buck-derived isolated DC–DC converters include forward, push-pull, half-bridge, full-bridge, etc. For the full-bridge converter, the power switches only sustains the input voltage, the utilization rate of the magnetic core is high, and it is suitable for medium-to-high power conversions. When employing the phase-shift control, the full-bridge converter could realize soft-switching of the power devices, leading to reduced switching loss and thus high conversion efficiency [6]. So, the phase-shifted full-bridge converter is adopted here for the DC–DC stage. Full-bridge inverter is suitable for high power applications and it can output high quality voltage since three voltage levels are available. Therefore, the full-bridge inverter is used here as the DC–AC stage. Figure 6.2 shows the main circuit of the two-stage inverter module.

Module 1 Full-Bridge DC-DC Converter Ldc1 iin1 Q1

icd1 + vin1 _

+

Q4

icdn Cdn

S1 A1 +

iin2 + vin2 _

Module 2

iinn + vinn _

Fig. 6.2 Main circuit of DC–AC ISOP system

Module n

Lf1

S3 iLf1

io1

_ B1 S4

io

iCf 1

vinv1

S2

DR2 DR4

…

Cd2

Tr

Vdc1 Cdc1 _

Q2

Vin icd2

Q3

* *

Cd1

iLdc1 DR1 DR3

…

Iin

Full-Bridge Inverter

Cf1

ZLd

+ vo _

124

6 Compound Balanced Control Strategy for DC–AC …

6.2 Implementation of the Compound Balanced Control Strategy Figure 6.3 shows the block diagram of the compound balanced control strategy for the DC–AC ISOP systems [7], which has three control loops, namely the common system output voltage loop, the input voltage sharing (IVS) loop, and the inner current loop for each inverter module. In Fig. 6.3, K vi is the sensor gain of module input voltage, K io is the sensor gain of module output current, K vo is the sensor gain of the system output voltage, Ginv is the transfer function from the modulating signal to the voltage across the phase-leg mid-points of the inverter, Gvo is the system output voltage regulator, Gvcd is the IVS regulator, and Gi is the inner current loop regulator. The system output voltage loop aims for regulating the system output voltage. The output of Gvo provides the common initial current reference iref for all the inner current loops. The IVS loop is used to achieve IVS among the constituent inverter modules. Generated by Gvcd , the dc deviation signal ivdj (j 1, 2, …, n) along with the initial current reference iref are sent to the multiplier to obtain the IVS regulating signal isvj (j 1, 2, …, n), which is then superimposed on the initial current reference iref to obtain irefj , which serves as the actual current reference of respective modules. The output current of each module follows irefj , and thus the module output current amplitude is regulated and the equivalence of the power factor of the module output current is also ensured. It can be seen that the control scheme can realize IVS and synchronization of the output current phase of each inverter module simultaneously, and consequently, IVS and OCS of the system can be achieved. At steady state, the compound balanced control strategy can realize IVS and OCS of each inverter module for the DC–AC ISOP system. Suppose a perturbation occurs to cause the variation of the input-dividing capacitor voltage. For example, vin1 ascends with its value higher than that of the average input voltage V in /n, which will result in the decrease of the output of IVS regulator ivd1 and the increase of the amplitude of corresponding current reference iref1 . Then, the amplitude of output current io1 traces that of iref1 to rise, which will result in the rise of the output active power of module 1. Therefore, the input power of module 1 will also increase and so is the input current iin1 , i.e., iin1 > iin . Since the input-dividing capacitor current icd1 equals to iin − iin1 , it is obvious that icd1 < 0, which means that the capacitor C d1 is discharged and vin1 will decrease until it returns to V in /n. On the other hand, if a perturbation leads to the descent of input voltage of some module with its value less than V in /n, the input voltage will also be adjusted and ultimately rise back to V in /n. It can be seen that the compound balanced control strategy could also achieve IVS and OCS at dynamic state. It is noteworthy that the feedback signal of modular inner current loop in the control strategy proposed in Fig. 6.3 is the output current ioj , and it also could be the output filter inductor current iLfj . The DC–AC ISOP systems uses different current feedback signal will get different OCS effect. If the output filter inductor current feedback is adopted [8], that is, the output filter inductor current of each module is employed as the inner current loop feedback on the basis of IVS, the output filter

6.2 Implementation of the Compound Balanced Control Strategy Inner Current Loop

IVS Loop Vin n

Kvi vin1

+ _

Gvcd

ivd1

isv1

iref1 _

Kvi

Gi

Ginv

+ vinv1

Inner Current Loop

+ _

Gvcd

ivd2

isv2

iref2 _

Kvi

+ + _

Gi

Ginv

iCf1 1 sLf1

iLf1

_

io1

+

+ vinv2

_

iCf2 _ io2 1 iLf2 sLf2 +

…

…

+

Inner Current Loop

IVS Loop Kvi vinn

sCf1

sCf2

Kio …

vin2

+ + _

_

Kio

IVS Loop Kvi

125

+ _

Gvcd

ivdn

isvn

Kvi

+ voref _

irefn _

+ + _

Gi

Ginv

+ vinvn

_

iCfn 1 sLfn

iLfn

_

ion

+ io ZLd +

vo

sCfn

+

Kio Gvo

iref

Kvo Output Voltage Loop

Fig. 6.3 Block diagram of compound balanced control strategy for DC–AC ISOP system

capacitor of each inverter module needs to be identical to ensure OCS (which means the equivalence of both the active and reactive current). However, the output filter capacitors of different modules are not matched in the actual prototype. In order to simplify the analysis, a two-module DC–AC ISOP system is taken as an example to analyze the effect of unmatched output filter capacitors on the sharing of output current. The equivalent circuit of two-module DC–AC ISOP system is shown in Fig. 6.4, where it is supposed that C f1 > C f2 . Figure 6.5a shows the phasor diagram of voltages and currents of two-module DC–AC ISOP system with output filter inductor current feedback, and here, the output filter inductor currents of two inverter modules are equal, i.e., I˙Lf1 I˙Lf2

(6.1)

I˙o1 I˙Lf1 − I˙Cf1 I˙o2 I˙Lf2 − I˙Cf2

(6.2a)

From Fig. 6.4, we have

Combining (6.1) with (6.2a, 6.2b), we have

(6.2b)

126

6 Compound Balanced Control Strategy for DC–AC …

Fig. 6.4 Equivalent circuit of two-module DC–AC ISOP system

· ICf2

· · ILf1–ILf2

· ICf1

· ICf1 · ICf2

· · Io1–Io2 · Vo

0 · Io1

(a) With output filter inductor current feedback

0

· Vo

(b) With output current feedback

Fig. 6.5 Phasor diagrams for different current feedback

I˙o1 − I˙o2 I˙Cf2 − I˙Cf1 jωo (Cf2 − Cf1 )V˙o

(6.3)

where ωo is the fundamental angular frequency of the inverter system. Since C f1 > C f2 , it can be derived from (6.3) that the output current difference of the two inverter modules, I˙o1 − I˙o2 , lags the output voltage V˙o by 90°, and the inductive reactive power is presented at the output side of the system. On the contrary, the capacitive reactive power will be presented if C f1 < C f2 . So, it can be seen that the unmatched output filter capacitors induce the reactive circulating current between the inverter modules and OCS cannot be truly realized when the output filter inductor current feedback control is adopted. For DC–AC ISOP systems, if the output current feedback is adopted directly [7], that is, the output current of each inverter module is used instead of the output filter inductor current as the feedback of the inner current loop of the inverter module on the basis of IVS, whether the output filter capacitors are matched or not will have no effects on OCS (which means both the active and reactive currents of the modules are equal, respectively). Figure 6.5b shows the phasor diagram of the voltages and currents of two-module DC–AC ISOP system with the output current feedback control, and here we have I˙o1 I˙o2 at steady state. Thus, it can be derived from (6.2) that I˙Lf1 − I˙Lf2 I˙Cf1 − I˙Cf2 jωo (Cf1 − Cf2 )V˙o . Since C f1 > C f2 , the difference of the output filter inductor currents of two inverter modules will lead the output voltage V˙o by 90°, and the capacitive reactive power is presented at the output side of the system. On the contrary, the inductive reactive power will be presented if C f1 < C f2 . Through the above analysis, we can find that for the DC–AC ISOP systems, when the output filter inductor current feedback control is adopted, OCS can be achieved

6.2 Implementation of the Compound Balanced Control Strategy

+

Vdcg Vdcj

_ Kvdc

Gvdc

Comparator − Phase-Shifted Control Circuit +

127

Drive Signals

VRAMP

Fig. 6.6 Closed-loop control block diagram of full-bridge converter as the first stage

only if the output filter capacitors are matched. If the output current feedback control is employed, OCS could be truly achieved whether the output filter capacitors are matched or not.

6.3 Closed-Loops Design for DC–AC ISOP Systems Taking the output current feedback control given in Sect. 6.2 as an example, the relationship of the control loops and the design of the control loops will be presented. As shown in Fig. 6.3, the compound balanced control strategy includes three closed loops, namely the system output voltage loop, IVS loop, and inner current loop. In addition, for the phase-shifted full-bridge converter as the first stage of each module, Fig. 6.6 shows the control block diagram, where the output voltage of the full-bridge converter, V dcj , is sensed with the sensor gain of K vdc , and then is compared with voltage reference V dcg , and the error is sent into the output voltage regulator Gvdc . The output of Gvdc is compared with the sawtooth carrier V RAMP , and the output of the comparator is sent to the phase-shifted control circuit, generating the drive signals.

6.3.1 Decoupling Relationship Between IVS Loop and Output Voltage Loop As seen from Fig. 6.3, the output of the IVS loop of each inverter module, isvj (j 1, 2, …, n), is a part of the reference of the inner current loop of each inverter module, so the IVS loop and the output voltage loop are coupled each other, and it is very difficult to design the system directly. So, the analysis of the control block diagram will be given below. From Fig. 6.3, the reference of the inner current loop of each inverter module is the difference between the output of the output voltage regulator iref and the output of each IVS loop isvj (j 1, 2, …, n), i.e., Iref j (s) Iref (s) − Isv j (s) ( j 1, 2, . . . , n)

(6.4)

128

6 Compound Balanced Control Strategy for DC–AC …

Assuming the parameters of the inverter modules are consistent and simplifying the inner current loop of each inverter module, the transfer function of each inner current loop is equal, expressed as G c j (s)

Io j (s) G c (s) ( j 1, 2, . . . , n) Iref j (s)

(6.5)

Then, Fig. 6.3 can be equivalent to Fig. 6.7a, and the output current of each inverter module is Io j (s) G c (s) · Iref j (s) G c (s) · Iref (s) − Isv j (s) G c (s) · Iref (s) − G c (s) · Isv j (s)

( j 1, 2, . . . , n)

(6.6)

According to (6.6), the total output current of the DC–AC ISOP system can be obtained as Io (s)

n

Io j (s) n · G c (s) · Iref (s) − G c (s) ·

j1

n

Isv j (s)

(6.7)

j1

According to (6.7), Fig. 6.7a can be equivalent to Fig. 6.7b, from which, the influence of the IVS loops on the output voltage loop of the DC–AC ISOP system can be derived as n

Isv j (s)

j 1

n j1

K vi

⎛

Vin − vin j G vcd (s)Iref (s) K vi G vcd (s)Iref (s) ⎝Vin − n

n

⎞ vin j ⎠

j1

0 (6.8) Equation (6.8) implies that the total effect of the IVS loops on the system output voltage loop is zero (as shown in Fig. 6.7c). That is to say, the output voltage loop is not affected by the IVS loops. On the other hand, the output signal of the output voltage regulator iref provides the initial reference for each inner current loop. When iref changes, the actual reference of each module, irefj , will also change simultaneously so that the output power and input power of each module vary accordingly. For example, when IVS and OCS are both achieved initially, the abruptly increasing load would result in a decrease of the output voltage. In order to ensure the stability of the output voltage, iref will increase and the output current of each module will also rise. In the meantime, OCS and the balanced output power of all modules are maintained. By power conservation, the input current of each module and hence the total input current of the system will all

6.3 Closed-Loops Design for DC–AC ISOP Systems

129

Fig. 6.7 Equivalent block diagram of the output current feedback control for DC–AC ISOP system

130

6 Compound Balanced Control Strategy for DC–AC …

Fig. 6.7 (continued)

increase. Since the input current is still being shared and the average current of each input-dividing capacitor is zero, IVS is maintained. It can be seen that the operation of the output voltage loop does not change the state of the input voltage of each module, and so does not affect the operation of the IVS loop. The above analysis shows that the IVS loop and output voltage loop are decoupled and can be designed separately when all the inverter modules are identical.

6.3.2 Design of Control Loop of the First-Stage Full-Bridge Converter Figure 6.8 shows the topology of the basic inverter module for the DC–AC ISOP system, which is composed of a full-bridge converter as the first stage and a fullbridge inverter as the second stage. If the output voltage frequency of the inverter module is 400 Hz, the input current of the second-stage, iac , contains not only highfrequency (switching frequency) ripple current, but also the low-frequency (800 Hz) ripple current. Therefore, the first-stage converter needs a larger output filter capacitor C dc to filter the low-frequency ripple current for the purpose of reducing the ripple

6.3 Closed-Loops Design for DC–AC ISOP Systems

131 800Hz

Ldc

Iin

Q1

Q3

Tr

DR1 DR3 +

S1 A +

* *

Q2

Q4

DR2 DR4

Lf

S3 iLf vinv

Vdc Cdc _

Vin

400Hz

iac

iLdc

_ B S2

io

iCf Cf

ZLd

+ vo _

S4

Fig. 6.8 The main circuit of single inverter module

of intermediate bus voltage, V dc . However, C dc cannot be infinite, so there is still low-frequency ripple component of 800 Hz in V dc . In order to reduce the influence of 800 Hz ripple on the voltage loop, the cutoff frequency of its voltage loop should be designed to be lower than 800 Hz and it is generally about 1/10 of 800 Hz [9, 10]. In the following, simulation is carried out for the two-stage inverter as shown in Fig. 6.8. The key specifications are as follows: the input voltage V in 270 V, the average value of the intermediate bus voltage V dc 180 V, the output voltage vo 115 V/400 Hz, and the output capacity S o 1 kVA. For the first-stage full-bridge converter, the key parameters are as follows: the transformer turns ratio K 15:13, the output filter inductor L dc 230 µH, the output filter capacitor C dc 800 µF (with the equivalent series resistance RCf 0.03 ), the output voltage sensor gain K vo 0.01, the peak-to-peak value of sawtooth carrier V pp 2.1 V, and the switching frequency f s 100 kHz. For the second-stage full-bridge inverter, the key parameters are as follows: the output filter inductor L f 400 µH and the output filter capacitor C f 16 µF. According to Fig. 6.6, the control block diagram of the output voltage loop of the first-stage full-bridge converter is shown in Fig. 6.9, where the control-to-output transfer function Gdc (s) is [11] Vin (1 + sCdc RCf ) K G dc (s) Rp RCf L dc RCf +s + Cdc RCf + s 2 L dc Cdc 1 + + Rp Cdc 1 + +1 RLd RLd RLd RLd (6.9) where Rp 4L r f s /K 2 , and RLd is the equivalent load of the first-stage full-bridge converter, which is obtained from the second-stage inverter and whose value is determined by the output active power of the inverter. Considering that the inverter operates with full resistive load, RLd V 2dc /Po 1802 / 1000 32.4 . From Fig. 6.9, the output voltage loop gain of the first-stage full-bridge converter can be obtained as

132

6 Compound Balanced Control Strategy for DC–AC …

vˆdcg

+ _

Gvdc(s)

1/Vpp

dˆy

Gdc(s)

vˆdc

Kvdc

Fig. 6.9 Control block diagram of the output voltage loop of the first-stage full-bridge converter Fig. 6.10 Bode diagram of the output voltage loop gain of the full-bridge converter at different cutoff frequencies

Table 6.1 Output voltage regulator Gvdc (s) at different cutoff frequencies Cutoff frequency/Hz Gvdc (s) 1000

9 + 1.4 × 104 /s

300

2 + 2500/s

80

0.5 + 500/s

Tvo (s) K vdc G vdc (s)G dc (s)/Vpp

(6.10)

Figure 6.10 gives the bode diagram of the output voltage loop gain of the fullbridge converter with different cutoff frequencies, i.e., 1 kHz, 300, and 80 Hz, and the corresponding transfer functions of output voltage regulator Gvdc (s) are shown in Table 6.1 Figure 6.11 shows the simulation waveforms of the primary current of transformer ip , the filter inductor current iLdc , the intermediate bus voltage V dc , and the inverter output voltage at different cutoff frequencies. Figure 6.11a shows simulation waveforms at the cutoff frequency of 1 kHz. As seen, the fluctuation of ip and iLdc of the first-stage full-bridge converter is relatively large, and the maximum value is almost twice of the average value. This is because the voltage loop tends to follow and compensate for the ripple voltage at 800 Hz, which will increase the current stress of the power devices and magnetic components. Figures 6.11b, c show the simulation waveforms at the cutoff frequencies of 300 and 80 Hz, respectively. It can be seen

6.3 Closed-Loops Design for DC–AC ISOP Systems

133

that the pulsation of 800 Hz in ip and iLdc is reduced with the decrease of the cutoff frequency. Therefore, in order to reduce the current stress of the power devices and the magnetic components of the full-bridge converter, the cutoff frequency of the voltage loop is designed relatively low, and it is set at about 1/10 of the output frequency.

6.3.3 Design of IVS Loop As illustrated in Sect. 6.3.1, the IVS loop and the system output voltage loop are decoupled each other, so the IVS loop can be designed separately. According to Fig. 6.3, the load of each inverter module can be considered to be composed of two parts, which are determined, respectively, by iref , the output of the system output voltage loop, and isvj (j 1, 2, …, n) introduced by each IVS loop. So, from the input side of the inverter module, each inverter can be regarded as resistances Rvo and RIVSj paralleled as shown in Fig. 6.12. The value of Rvo and RIVSj are determined by iref and isvj (j 1, 2, …, n), respectively.

6.3.3.1

Derivation of Rvo

If the IVS loops are ignored (see Fig. 6.3), i.e., iˆsv1 iˆsv2 . . . iˆsvn 0, iref is the common reference of the inner current loop of all inverter modules, and the method will be equivalent to OCS control. So, from the input side, under the role of iref , each inverter module can be equivalent to a constant power load. Within the cutoff frequency range of the current loop, the small-signal model of a constant power load presents the negative resistance characteristic. From (6.8), it can be seen that the sum of isvj (j 1, 2, …, n) of each inverter module equals to zero, which means that the sum of each module power corresponding to isvj also equals to zero and thus the sum of each inverter module power corresponding to iref serves as the power of the DC–AC ISOP systems. Therefore, the active power of each inverter module corresponding to iref is equal, and it is 1/n of the system power Po . So the resistance Rvo can be expressed as Rvo

6.3.3.2

vˆin j iˆin_vo j

−Rn −

(Vin /n)2 ( j 1, 2, . . . , n) Po /n

(6.11)

Derivation of RIVSj

From Fig. 6.7a, the individual output current perturbation Iˆo_IVS j ( j 1, 2, . . . , n) of inverter introduced by the input voltage perturbation vˆin j ( j 1, 2, . . . , n) can be expressed as

6 Compound Balanced Control Strategy for DC–AC …

(V)

(V)

(A)

(A)

134

10 0 −10 12 6 0 185 180 175 200 0 −200

ip

iLdc Vdc

vo 10

11

12

13

t (ms)

14

15

16

17

16

17

16

17

(V)

(V )

(A)

(A)

(a) With voltage loopcut-off frequencyof 1 kHz

10 0 −10 10 5 0 185 180 175 200 0 −200

ip

iLdc Vdc

vo 10

11

12

13

t (ms)

14

15

(V)

(V)

(A)

(A)

(b) With voltage loop cut-off frequency of 300 Hz

10 0 −10 10 5 0 185 180 175 200 0 −200

ip

iLdc Vdc vo 10

11

12

13

t (ms)

14

15

(c) With voltage loop cut-off frequency of 80 Hz

Fig. 6.11 Simulation waveforms at different cutoff frequencies

6.3 Closed-Loops Design for DC–AC ISOP Systems

135

iˆin1

Fig. 6.12 Equivalent small-signal model of DC–AC ISOP system

+ ˆvin1 _

Z1′

Cd1

iˆin_vo1 Rvo

iˆin_IVS1 RIVS1

iˆin_vo2 Rvo

iˆin_IVS2 RIVS2

iˆin_von Rvo

iˆin_IVSn RIVSn

iˆin2 + vˆ in2 _

Cd2 ˆiinn

…

vˆ in

Z 2′

+ ˆvinn _

Z n′

Cdn

Iˆo_IVS j vˆin j · K vi · G vcd · Iref (s) · G c (s) ( j 1, 2, . . . , n)

(6.12)

where Iˆo_IVS j has the same phase with the corresponding load current Iˆo j ( j 1, 2, . . . , n). The individual active output power perturbation pˆ o_IVS j ( j 1, 2, . . . , n) of module j corresponding to I o_IVS j can be expressed as

pˆ o_IVS j I o_IVS j · Vo · cos θ ( j 1, 2, . . . , n)

(6.13)

The individual input power perturbation pˆ in_IVS j ( j 1, 2, . . . , n) of module j corresponding to vˆin j can be expressed as pˆ in_IVS j Iin vˆin j +

Vin ˆ i in_IVS j ( j 1, 2, . . . , n) n

(6.14)

Assuming the conversion efficiency for each individual module is 100%, we have pˆ in_IVS j pˆ o_IVS j ( j 1, 2, . . . , n)

(6.15)

Substituting (6.12) into (6.13), and then substituting (6.13) and (6.14) into (6.15), yields RIVS j

vˆin j iˆin_IVS j

Vin /n ( j 1, 2, . . . , n) K vi G vcd Iref (s)G c (s) · Vo cos θ − Iin (6.16)

According to Fig. 6.7c, we have I ref (s) Gc (s) I o /n. So, I ref (s) Gc (s) V o cosθ equals to Po /n. Then, (6.16) can be rewritten as

136

6 Compound Balanced Control Strategy for DC–AC …

RIVS j

6.3.3.3

Vin /n Vin ( j 1, 2, . . . , n) (6.17) K vi G vcd Po /n − Iin K vi G vcd Po − n Po /Vin

Design of IVS Regulator Gvcd

As pointed out in Chap. 3, the OCS control could not ensure the stability of the DC–DC ISOP system. The essential reason is that the input impedance of each module is negative resistance. Similarly, for the DC–AC ISOP system, the OCS control will lead to an unstable system since the modules present the negative resistance at the input sides. Equation (6.17) implies that the role of IVS loops is equivalent to introduce a positive resistance to the input terminal of each module as shown in Fig. 6.12. So, the equivalent input impedance of each inverter module is Z j Rvo //RIVS j

−Rn · RIVS j Rn · RIVS j −Rn + RIVS j Rn − RIVS j

( j 1, 2, . . . , n)

(6.18)

In order to ensure the stability of DC–AC ISOP system, the equivalent input impedance of each inverter module should present positive resistance characteristics, i.e., Z j > 0. According to (6.18), we can obtain the condition for stability as Rn > RIVS j ( j 1, 2, . . . , n)

(6.19)

Substitution of (6.11) and (6.17) into (6.19), yields G vcd >

2n G vcd min K vi · Vin

(6.20)

It can be seen that the gain of IVS loop compensator must satisfy (6.20) so that each inverter module presents positive resistance characteristics and the DC–AC ISOP system with IVS control strategy is stable. For simplicity, a two-module DC–AC ISOP system is simulated to verify the correctness of the IVS loop design. The specifications of the DC–AC ISOP system are as follows: the system input voltage is 540 V (±10%), the output voltage is 115 VAC/400 Hz, the output capacity is 2 kVA; the module input voltage is 270 V (±10%), the module output voltage is 115 VAC/400 Hz, the output capacity of each module is 1 kVA; the input voltage sensor gain is K vi 0.01. The input-dividing capacitors of the two modules are intentionally set at different values, i.e., C d1 1000 µF and C d2 800 µF. Figure 6.13 shows the response in the individual input voltages and output currents of two modules when the system input voltage is step changed from 500 to 600 V. According to (6.20), it can be obtained that Gvcdmin 0.666 ~ 0.815. As shown in Figs. 6.13a, b, for Gvcd > Gvcdmin , the two input voltages converge finally, and the higher values of Gvcd result in a faster correction in the input voltages after a system input voltage disturbance. As shown in Fig. 6.13c, for Gvcd < Gvcdmin , the two input

6.3 Closed-Loops Design for DC–AC ISOP Systems

320

vin2 vin1

(V) 280 240 20

io1

10 (A)

137

io2

0 10 20

10

0

20 t (ms)

30

40

30

40

(a) G vcd = 8

320

vin2 vin1

(V) 280 240 20

io1

io2

10 (A)

0 10 20

0

10

20 t (ms) (b) G vcd = 3

Fig. 6.13 Response to a step change in the system input voltage with different Gvcd

voltages diverge, resulting in a runaway mode. The system cannot work normally and it is unstable.

6.3.4 Design of Output Voltage Loop As shown in Fig. 6.7c, the system output voltage loop is decoupled from the IVS loops. In order to design the compensator for system output voltage loop, it is required

138

6 Compound Balanced Control Strategy for DC–AC …

320

vin2 11.66V

15.43V

vin1

(V) 280 240 20

io1

io2

10 (A)

0

10 20

0

5

15

25

35 t (ms)

45

55

65 70

(c) G vcd = 0.25

Fig. 6.13 (continued)

to derive the transfer function of inner current loop Gc (s) of a single inverter first. Based on the control strategy shown in Fig. 6.3, when the IVS loops are not considered, the control block diagram of a single inverter module can be obtained as shown in Fig. 6.14a. Relocating the feedback node of io backward to the output voltage, Fig. 6.14b is obtained. Simplifying the part of the diagram in the dashed block of Fig. 6.14b, c can be obtained. From Fig. 6.14c, the transfer function of inner current loop Gc (s) can be derived as G c (s)

G i G inv Io (s) 2 Iref (s) s n Z Ld L f Cf + s L f + n Z Ld + K io G i G inv

(6.21)

From (6.21) and Fig. 6.7c, the loop gain of the system output voltage loop can be derived as Tvo (s)

s2n Z

n K vo G vo G i G inv Z Ld Ld L f C f + s L f + n Z Ld + K io G i G inv

(6.22)

6.4 Simulation Results In order to verify the effectiveness of the IVS and OCS control strategy based on the output current feedback and the output filter inductor current feedback for the DC–AC ISOP system, the simulation of two-module DC–AC ISOP system is carried out. The parameters used in the simulation are the same as those given in Sect. 6.3.3. To obtain the effects of output filter capacitor mismatch on two kinds of control

6.4 Simulation Results

139

iCf

(a) voref +

_

Gvo

iref + _

Gi

Ginv

vinv

_

1 iLf sLf +

+

_

sCf nZLd

io

vo

Kio Kvo

(b) voref + _

Gvo

iref + _

Gi Kio

Ginv io

vinv

_

iCf sCf _ 1 iLf nZLd sLf + io

vo

1 nZLd

Kvo

(c) voref +

_

Gvo

iref + _

Gi Kio

Ginv io

vinv

nZ Ld s 2 nZ Ld Lf Cf sLf

nZ Ld

vo

1 nZLd

Kvo Fig. 6.14 Equivalent control block diagram of output voltage loop of individual module

methods, a 4.7 µF capacitor is intentionally connected in parallel with the output filter capacitor of module 1. Figure 6.15 shows the simulation waveforms of the DC–AC ISOP system with output current feedback control. The waveforms of the individual input voltage and the difference between input voltages are shown in Fig. 6.15a. As seen, the input voltages of the two inverter modules are equal. The waveforms of the system output voltage, individual output filter inductor current and the difference between these two currents are shown in Fig. 6.15b, from which it can be seen that the output filter inductor current of the two inverter modules are different, and the difference leads the output voltage by 90° with the capacitive reactive power being presented. Figure 6.15c shows the waveforms of individual output current and the difference between these two currents. Due to the adoption of the output current feedback, the output filter capacitor mismatch does not affect the OCS effect, and the output current of the two inverter modules is equal. Figure 6.16 shows the simulation waveforms of the DC–AC ISOP system with output filter inductor current feedback control. The waveforms of individual input voltage and the difference between input voltages are shown in Fig. 6.16a. As seen, the input voltages of the two inverter modules are equal. The waveforms of the output voltage of the system, individual output current and the difference between these two currents are shown in Fig. 6.16b, from which it can be seen that the output currents

140

6 Compound Balanced Control Strategy for DC–AC … 280 270 (V) 260 250 280 (V) 270 260 250 5 (V)

vin1

vin2

vin1 –vin2

0 −5

0

2

4

t (ms)

6

8

10

(a) Simulationwaveformsof input voltages

200 (V)

vo

0 −200 20

(A)

0 −20 20

(A)

0

iLf1

iLf2

−20 5 (A) 0 −5

iLf1 – iLf2 0

2

4

t (ms)

6

8

10

(b) Simulationwaveformsof output filter inductor currents

15 (A)

0 −15 15

(A)

io1

io2

0 −15 5

(A)

0 −5 0

io1 –io2

2

4

t (ms)

6

8

(c) Simulation waveforms of output currents

Fig. 6.15 Simulation waveforms with output current feedback control

10

6.4 Simulation Results

141

of the two inverter modules are different, and the difference lags the output voltage by 90° with the inductive reactive power being presented. Figure 6.16c shows the waveforms of individual output filter inductor current and the difference between these two currents. Due to the adoption of the output filter inductor current feedback, the output filter inductor current of the two inverter modules are equal. Figures 6.15 and 6.16 verify the validity of the previous theoretical analysis.

6.5 Experimental Results In order to verify the compound balanced control strategy with output current feedback controlled, a 2-kVA two-module DC–AC ISOP system prototype is fabricated in the laboratory. The specifications of the DC–AC ISOP system and the single inverter module are the same as the parameters in the simulation given in Sect. 6.4.

6.5.1 Steady-State Experiment Figure 6.17 gives the steady-state experimental waveforms of the system at full resistive load (2 kW). The waveforms of individual input voltage, individual output current, and the difference between these two currents are shown in Fig. 6.17a. As seen, the input voltage and output current of the two inverter modules are the same, respectively, which means that IVS and OCS are achieved. The waveforms of the system output voltage, total output current, and individual output current are shown in Fig. 6.17b, which indicate that the system outputs a stable sinusoidal voltage. Figure 6.18 gives the steady-state experimental waveforms of the system at full inductive load (2 kW, PF 0.75). The waveforms of the individual input voltage, individual output current, and the difference between these two currents are shown in Fig. 6.18a, from which we can see that IVS and OCS are well achieved. Figure 6.18b shows the waveforms of the system output voltage, total output current, and individual output current. Due to the inductive load, the output voltage leads the total output current.

6.5.2 Dynamic Experiment Figures 6.19 and 6.20 give the experimental waveforms of the system at nominal input voltage when the load current steps up and down between half and full resistive load, and between half and full inductive load, respectively. Figures 6.19a and 6.20a show the waveforms of the individual input voltage, system output voltage, and total output current. As seen, the input voltages are shared equally during the transient condition. It should be noted that the input voltage source of the system is not an ideal one,

142

6 Compound Balanced Control Strategy for DC–AC … 280 270 (V) 260 250 280 (V) 270 260 250 5 (V)

vin1

vin2

vin1 –vin2

0 −5

0

2

4

t (ms)

6

8

10

8

10

(a) Simulationwaveformsof input voltages

200 (V)

vo

0 −200 15

(A)

0 −15 15

(A)

0

io1

io2

−15 5 (A) 0 −5

io1 – io2 0

2

4

t (ms)

6

(b) Simulationwaveformsof output currents

20 (A)

iLf1

0 −20 20

(A)

iLf2

0 −20 5

(A)

iLf1 –iLf2

0 −5

0

2

4

t (ms)

6

8

10

(c) Simulationwaveformsof output filter inductor currents

Fig. 6.16 Simulation waveforms with output filter inductor current feedback control

6.5 Experimental Results

143

C1 vin1 : [100V/div] C2 vin2 : [100V/div] C1

C3 io1 : [25A/div]

C3 C2

C4 io2 : [25A/div]

C4 F2

F2 io1 io2 : [25A/div] Time: [2ms/div] (a) Experimental waveformsof input voltages and output currents

C1 vo : [200V/div] C1

F2

C3

F2 io : [25A/div]

C3 io1 : [25A/div]

C4

C4 io2 : [25A/div] Time: [2ms/div] (b) Experimental waveformsof output voltages and output currents

Fig. 6.17 Experimental waveforms at full resistive load

the input voltage has slight drop or rise when the load current steps up and down. Figures 6.19b and 6.20b show the waveforms of the individual input voltage and output current. As seen, the output currents are shared equally both at steady state and transient. Figures 6.21 and 6.22 give the experimental waveforms of the input voltage stepping between 486 V (90%V in ) and 594 V (110%V in ) at full resistive and inductive load, respectively. Figures 6.21a and 6.22a show the waveforms of individual input voltage, system output voltage, and total output current. As seen, the input voltages of the two modules are shared equally during the transient condition and the system

144

6 Compound Balanced Control Strategy for DC–AC …

C1 vin1 : [100V/div] C2 vin2 : [100V/div] C1

C3 io1 : [25A/div]

C3 C2

C4 io2 : [25A/div]

C4 F2

F2 io1 – io2 : [25A/div] Time: [2ms/div] (a) Experimental waveformsof input voltages and output currents

C1 vo : [200V/div] C1

F2

F2 io : [25A/div] C3

C3 io1 : [25A/div]

C4

Time: [2ms/div]

C4 io2 : [25A/div]

(b) Experimental waveformsof output voltages and output currents

Fig. 6.18 Experimental waveforms at full inductive load

output voltage is stable. Figures 6.21b and 6.22b show the waveforms of the individual input voltage and output current. As seen, the output currents of the two modules are shared equally during the transient condition.

6.5 Experimental Results

145

C1 vin1 : [100V/div] C2 vin2 : [100V/div] C1

C3 vo : [200V/div]

C3 C2 C4

C4 io : [25A/div] Time: [5ms/div] (a) Experimental waveforms of input voltages, output voltage and output currents

C1 vin1 : [100V/div] C2 vin2 : [100V/div] C4 io1 : [10A/div]

C1 C4 C2

C3

Time: [5ms/div]

C3 io2 : [10A/div]

(b) Experimental waveforms of input voltages and individual output currents

Fig. 6.19 Experimental waveforms under step change of resistive load

6.5.3 Comparison of the Output Current Feedback and Output Filter Inductor Current Feedback Further comparative experiments on the compound balanced control strategies with the output current feedback and the output filter inductor current feedback are conducted. Figure 6.23a gives the experimental waveforms of the system with output current feedback control. It can be seen that the mismatch of output filter capac-

146

6 Compound Balanced Control Strategy for DC–AC …

C1 vin1 : [100V/div] C2 vin2 : [100V/div] C1

C3 vo : [200V/div]

C3 C2 C4

C4 io : [25A/div] Time: [5ms/div] (a) Experimental waveforms of input voltages, output voltage and output currents

C1 vin1 : [100V/div] C2 vin2 : [100V/div] C1

C4 io1 : [10A/div]

C4 C2

C3

C3 io2 : [10A/div] Time: [5ms/div] (b) Experimental waveforms of input voltages and individual output currents

Fig. 6.20 Experimental waveforms under step change of inductive load

itor has no influence on the sharing of output currents between the two modules. However, the unmatched output filter capacitors introduce the difference between the output filter inductor currents. Figure 6.23b gives the experimental waveforms of the system with output filter inductor current feedback control. As seen, the output filter inductor current of the two inverter modules are almost the same, but there is reactive current between the two inverter modules, which is consistent with the previous analysis.

6.5 Experimental Results

147

C4 vin1 : [100V/div] C3 vin2 : [100V/div] C2 vo : [200V/div] C2/C4 C3 C1

C1 io : [25A/div] Time: [10ms/div] (a) Experimental waveforms of input voltages, output voltage and output currents

C4 vin1 : [100V/div] C2 io1 : [10A/div]

C3 vin2 : [100V/div]

C2/C4 C3

C1

Time: [10ms/div]

C1 io2 : [10A/div]

(b) Experimental waveforms of input voltages and individual output currents

Fig. 6.21 Experimental waveforms with step change of system input voltage at full resistive load

The steady state and dynamic experiments prove that the IVS and OCS control strategy based on output current feedback can achieve IVS and OCS under various circumstances. Furthermore, with the method adopted, the disturbance of input voltage and load can be restrained and the reliability of the system can also be guaranteed.

148

6 Compound Balanced Control Strategy for DC–AC …

C1 vin1 : [100V/div] C2 vin2 : [100V/div] C3 vo : [200V/div] C1/C3 C2 C4

C4 io : [25A/div] Time: [10ms/div] (a) Experimental waveforms of input voltages, output voltage and output currents

C1 vin1 : [100V/div] C3 io1 : [10A/div]

C2 vin2 : [100V/div]

C1/C3 C2

C4

Time: [10ms/div]

C4 io2 : [10A/div]

(b) Experimental waveforms of input voltages and individual output currents Fig. 6.22 Experimental waveforms with step change of system input voltage at full inductive load

6.6 Distributed IVS/OCS Control Strategy In the control scheme shown in Fig. 6.3, all the inverter modules share a common voltage loop, and this control scheme is actually a kind of centralized control strategy. Therefore, a distributed control is presented here to make sure that all the inverter modules are fully modularized. The failure of any module will not affect the nor-

6.6 Distributed IVS/OCS Control Strategy

149 C3 vo : [200V/div] C1 io1 : [25A/div] C2 io2 : [25A/div] M4 io1 – io2: [5A/div] M2 iLf1: [25A/div] M1 iLf2: [25A/div] M3 iLf1 – iLf2: [10A/div]

Time: [2ms/div] (a) Experimental waveforms with output current feedback control

C2 vo : [200V/div] M1 io1 : [25A/div] M2 io2 : [25A/div] M3 io1 – io2: [10A/div] C1 iLf1: [25A/div] C3– iLf2: [25A/div] F1 iLf1 – iLf2: [5A/div] Time: [2ms/div] (b) Experimental waveforms with output filter inductor current feedback control

Fig. 6.23 Experimental waveforms of two control strategies with different current feedback when the output filter capacitors are unmatched

mal operation of the DC–AC ISOP systems, so as to improve the reliability and redundancy of the DC–AC ISOP systems. Figure 6.24 presents the distributed control strategy for the DC–AC ISOP system [12], which also includes the output voltage loop, IVS loop, and inner current loop. Different from the centralized control strategy, the output voltage loops are scattered in each inverter module. All the modules communicate with each other by three buses, namely the output voltage reference synchronous bus (vref synchronous bus), the IVS bus, and the average current bus (iave bus). It can be seen that all the inverter modules in the system are completely reciprocal, and so distribution and modularity are fully achieved. In this distributed control strategy, the output voltage of each inverter module is regulated by its respective output voltage loop. All the output voltage references are synchronized by the digital signal processor (DSP) to generate the output voltage reference synchronous bus. Since the components of the feedback networks and regulators maybe not well matched, the output signals of every output

150

6 Compound Balanced Control Strategy for DC–AC … iave ivd1

+ Kvi _ Gvcd vin1 Kvi IVS Loop ig1 voref + Gvo _

ivd2

ig2

…

…

+

Gvo

io1

isv2 Inner Current Loop iCf2 _ iref2 iLf2 _ sCf2 vinv2 _ + + 1 Gi Ginv _ sLf2 + io2 + Ki

+ Kvi _ Gvcd vinn Kvi IVS Loop voref + _

1 sLf1 +

iCf1 sC _ f1

Kvo

Output Voltage Loop

vref IVS Synchronous Bus Bus

iLf1

Kvo

+ Kvi _ Gvcd vin2 Kvi IVS Loop Gvo

_

Ki

Output Voltage Loop

voref + _

isv1 Inner Current Loop _ iref1 vinv1 + + Gi Ginv _ +

ign

Output Voltage Loop

+ +

1 n

ivdn

iave Bus

+

+ io

…

Vin n

…

voref

+

ZLd

vo

isvn Inner Current Loop iCfn _ irefn iLfn _ sCfn vinvn _ + + 1 Gi Ginv _ sLfn + ion + Ki Kvo

Fig. 6.24 Block diagram of distributed control strategy for DC–AC ISOP system

voltage regulators (namely igj ) are not exactly the same. So, they are averaged, forming the iave bus. The signal iave is sent to all the inverter modules and served as the common initial current reference. The IVS bus signal comes from all the modules through an averaging circuit and its value is V in /n, which provides the input voltage reference for all the inverter modules. The output signals of modular IVS loops ivdj (j 1, 2, …, n) multiply iave and then superimpose on iave to produce the actual current reference signal of each module, irefj . The purpose of the multiplier circuit is to ensure that output filter inductor current of each inverter module tracks the reference irefj to attain the same phase. In the meanwhile, the IVS loop modifies the magnitude of the output filter inductor current (and adjusts the output active power of each module) to ensure IVS and OCS. Basically, the distributed control strategy still adopts the conception of compound balanced control to realize the power balance of the system. Similar with the centralized control mentioned before, the OCS effect is different for the DC–AC ISOP system in the distributed control system when taking the output

6.6 Distributed IVS/OCS Control Strategy

151

filter inductor current and the output current as feedback signal. And, the decoupling relationships between loops and parameters design is similar to that in the centralized control strategy, which are not repeated here.

6.7 Summary This chapter proposes the compound control scheme for the DC–AC ISOP systems to obtain IVS and OCS. In the proposed control scheme, the inner current loop of each inverter module can adopt the output filter inductor current feedback and the output current feedback as well. Among them, the former scheme can realize OCS when the output filter capacitors are matched, and the latter scheme can achieve OCS unconditionally. Taking the latter scheme as an example, the IVS loop and the system output voltage loop are analyzed and it is pointed out that they are decoupled and can be designed separately. A two-module DC–AC ISOP system prototype is fabricated and tested in the laboratory, and the experimental results verify the effectiveness of the proposed control strategy and the parameters design of the control loops. Furthermore, a distributed control scheme is proposed on the basis of the centralized control, and it disperses all the control loops into each inverter module and achieves modularity completely.

References 1. Chen, W., Zhuang, K., Ruan, X.: Input-series and output-parallel connected inverter system for high input voltage applications. IEEE Trans. Power Electron. 24(9), 2127–2137 (2009) 2. Fang, T., Ruan, X., Tse, C.K.: Control strategy to achieve input and output voltage sharing for input-series-output-series connected inverters systems. IEEE Trans. Power Electron. 25(6), 1585–1596 (2010) 3. Silva, C., Oliveira, D., Bascope R.: A dc-ac converter with high frequency isolation. In: Proceedings of IEEE International Symposium on Industrial Electronics, pp. 953–958 (2007) 4. Xue, Y., Chang, L., Kjaer, S.B., Bordonau, J., Shimizu, T.: Topologies of single-phase inverters for small distributed power generators: an overview. IEEE Trans. Power Electron. 19(5), 1305–1314 (2004) 5. Kjaer, S.B., Pedersen, J.K., Blaabjerg, F.: A review of single-phase grid-connected inverters for photovoltaic modules. IEEE Trans. Ind. Appl. 41(5), 1292–1306 (2005) 6. Ruan, X.: Soft-Switching PWM Full-Bridge Converters. Wiley, Singapore (2014) 7. Chen, W.: Research on series-parallel conversion systems consisting of multiple converter modules (in Chinese). PhD Dissertation. Nanjing University of Aeronautics and Astronautics, Nanjing, China (2009) 8. Zhuang, K., Ruan, X.: Control strategy to achieve input voltage sharing and output current sharing for input-series-output-parallel inverter. In: Proceedings of IEEE Power Electronics Specialists Conference, pp. 3655–3658 (2008) 9. Andersen, G.K., Klumpner, C., Kjaer, S.B., Blaabjerg, F.: A new green power inverter for fuel cells. In: Proceedings of IEEE Power Electronics Specialists Conference, pp. 727–733 (2002) 10. Liu, C., Lai, J.: Low frequency current ripple reduction technique with active control in a fuel cell power system with inverter load. IEEE Trans. Power Electron. 22(4), 1429–1436 (2007)

152

6 Compound Balanced Control Strategy for DC–AC …

11. Vlatkovic, V., Sabate, J.A., Ridley, R.B., Lee, F.C., Cho, B.H.: Small-signal analysis of the phase-shifted PWM converter. IEEE Trans. Power Electron. 17(1), 128–135 (1992) 12. Fang, T., Shen, L., He, W., Ruan, X.: Distributed control and redundant technique to achieve superior reliability for fully modular input-series-output-parallel inverter system. IEEE Trans. Power Electron. 32(1), 723–735 (2017)

Chapter 7

Compound Balanced Control Strategy for DC–AC Input-Series-Output-Series Systems

Abstract DC–AC input-series-output-series (ISOS) systems are very suitable for high input voltage and high output voltage power conversion applications. The implementation of the compound balanced control strategy is presented for the DC–AC ISOS system to achieve input voltage sharing (IVS) and output voltage sharing (OVS). The output filter inductor current or the output filter capacitor current adopted as the feedback signal for the inner current loops of each module, respectively, is analyzed and compared. Taking the output-filter-capacitor-current method as an example, the decoupling relationship between the control loops is analyzed and the loop design procedure is given. The simulation and experimental results are provided to verify the effectiveness of the proposed control method. Furthermore, this chapter presents a new solution dedicated to distributed configuration on the basis of centralized control, which can improve the reliability of the system. Keywords Input-series-output-parallel system Compound balanced control strategy · Input voltage sharing Output voltage sharing · Closed-loop designs As one of the input-series-connected DC–AC power conversion systems, the DC–AC input-series-output-series (ISOS) system is composed of multiple DC–AC inverter modules with low input voltage and low output voltage, which are connected in series both at the input side and output side, and thus it is suitable for high input voltage and high output voltage applications. The key issue of the DC–AC ISOS systems is to achieve input voltage sharing (IVS) and output voltage sharing (OVS) of the constituent modules simultaneously. Different from the DC–DC ISOS systems, the output voltage of the DC–AC ISOS system is an AC one. So, in order to achieve OVS, it is required to control both the magnitude and phase of output voltages of all the modules to be equal. As a consequence, both the active and reactive powers of the modules will be balanced.

© Springer Nature Singapore Pte Ltd. and Science Press, Beijing, China 2019 X. Ruan et al., Control of Series-Parallel Conversion Systems, CPSS Power Electronics Series, https://doi.org/10.1007/978-981-13-2760-5_7

153

154

7 Compound Balanced Control Strategy …

In Chap. 5, it is pointed out that the DC–AC ISOS systems needs to adopt the compound balanced control strategy, i.e., to realize IVS and ensure the magnitude or phase of the modular output voltages be equal. In this chapter, an implementation of the compound balanced control strategy is proposed. In the proposed control strategy, the output filter inductor current or the output filter capacitor current could be the feedback signal for the inner current loops of each module. The analysis shows that the output filter capacitor current feedback can realize IVS and OVS among the constituent modules no matter the output filter capacitors of modules are well matched or not. As an example, the compound balanced control strategy with output filter capacitor current feedback is analyzed, and it is shown that the IVS control loops and the system output voltage loop are decoupled each other. Then, the parameters design for all the control loops is presented. Simulation and experimental results are given to verify the validity of the proposed compound balanced control strategy. Finally, based on the compound balanced control strategy, a distributed control scheme is proposed to achieve full modularity of the DC–AC ISOS systems.

7.1 Implementation of the Compound Balanced Control Strategy Figure 7.1 gives the configuration of the DC–AC ISOS systems consisting of n inverter modules. The basic inverter module in the DC–AC ISOS systems adopts the two-stage configuration consisting of the first-stage DC–DC converter and the second-stage DC–AC inverter. Here, the phase-shifted full-bridge converter is employed as the first-stage, and the full-bridge inverter is adopted as the second stage. As described in Sect. 5.1.4, the compound balanced control strategy means ensuring the equaling the amplitude or phase of modular output voltages while controlling the input voltage sharing. According to this idea, Fig. 7.2 shows an implementation of the compound balanced control strategy for the DC–AC ISOS system [1], which includes three control loops, namely, the common system output voltage loop, the IVS loop, and the inner current loop of each inverter module (Here the module output filter capacitor current is taken as the feedback signal for each inner current loop). The common system output voltage loop guarantees the stability of the system output voltage, and its output signal iref acts as the common initial current reference for all the inner current loops. Here iref contains both the amplitude and the phase information. In the IVS loop, the system input voltage divided by n and the input voltage of each module are sensed with sensor gain of K vi and compared. The voltage error is sent into the IVS regulator Gvcd , whose output signal ivdj (j 1, 2,…, n) serves as the current reference amplitude correction signal of each module. By multiplying ivdj and iref , the current correction signal isvj (j 1, 2, …, n) is obtained with the same phase as that of iref . Then, isvj is superimposed on the initial reference iref to adjust the amplitude of the current reference of each module. Since the current inner loop adopts the capacitor current feedback of each module, the IVS loop only alters

7.1 Implementation of the Compound Balanced Control Strategy Module 1 Full-Bridge DC-DC Converter L dc1 DR1 DR3

iin1 Q1

icd1

icd2

S1 A1 +

Q4

_ B1 S2

DR2 DR4 + vin2 _

iLf1 vinv1

Vdc1 Cdc1 _

Q2

Lf1

S3

Module 2

S4

io1

iCf1 Cf1

io + vo1 _ +

ZLd v_o + vo2 _

…

Cd2

+

+ v _in1

Vin

iLdc1

Tr

* *

Cd1

Q3

Full-Bridge Inverter

…

Iin

155

icdn Cdn

iinn + vinn _

Module n

+ von _

Fig. 7.1 Configuration of the DC–AC ISOS system

the amplitude of the output filter capacitor current reference of each module, and the phase of the current reference is always synchronized. As a result, all the output filter capacitor currents have the same phase by tracking the synchronized reference. Since the filter capacitor current of each module leads the corresponding module output voltage by 90°, the module output voltages have the same phase, i.e., θ1 θ2 · · · θn

(7.1)

where θ 1 , θ 2, …, θ n are the output power factor angle of each inverter module. With the IVS loop, we have Vin1 Vin2 · · · Vinn

(7.2)

At steady state, the average current of the input dividing capacitor of each module is zero, so we have Iin1 Iin2 · · · Iinn

(7.3)

It is supposed that the conversion efficiency of each inverter module is 100%. So, by power conservation, we have

156

7 Compound Balanced Control Strategy … IVS Loop

vin1

Gvcd

ivd1

isv1

_

Kvi

+ + _

Gi

Ginv

_

+

1 iLf1 sLf +_

iCf1 1 vo1 sCf

Ki Inner Current Loop

IVS Loop + _

Kvi vin2 Kvi

Gvcd

ivd2

isv2

iref2 _

+

+ _

Gi

Ginv

vinv2

vo2

_ 1 sLf

+

iLf2 +_

iCf2 1 sCf

io

…

…

Ki

+ Kvi _

+

Inner Current Loop

IVS Loop

vinn

vinv1

…

Vin n

+ Kvi _

Inner Current Loop iref1

Gvcd

ivdn

isvn

Kvi

irefn _

+ + _

Gi

Ginv

vinvn +

_

_ + 1 sLf iLfn

1 ZLd + + vo

iCfn 1 von sCf

Ki

+ voref _

Gvo

iref

Kvo System Output Voltage Loop

Fig. 7.2 Compound balanced control strategy for DC–AC ISOS system

Pin j Vin j · Iin j Vo j · Io · cos θ j Po j ( j 1, 2, . . . , n)

(7.4)

where Pinj and Poj are the input power and output active power of each module, respectively. Substitution of (7.2) and (7.3) into (7.4) leads to Vo1 · cos θ1 Vo2 · cos θ2 · · · Von · cos θn

(7.5)

Substituting (7.1) into (7.5) yields Vo1 Vo2 · · · Von

(7.6)

According to (7.1) and (7.6), we have vo1 vo2 · · · von

(7.7)

As discussed above, the implementation of the compound balanced control strategy as shown in Fig. 7.2 can achieve IVS and OVS simultaneously, and it is referred to as the combined control of IVS and output phase synchronization. For the DC–AC ISOS systems, the feedback signal of the inner current loop of each module in Fig. 7.2 can also be the output filter inductor current. In this

7.1 Implementation of the Compound Balanced Control Strategy · · · ILf2 Vo2 ICf1 · ICf2 0 · Vo1

· f1 IL

157

· ICf1 · Vo

· Io

· · Vo1 Vo2

(a) With filter inductor current feedback

· ICf2

· Vo · Io

0 · · Vo1=Vo2

(b) With filter capacitor current feedback

Fig. 7.3 Steady-state phasor diagram of the output variables of DC–AC ISOS system with different current feedback when C f1 > C f2

case, the IVS loop can still ensure IVS as well as the output active power balance among the inverter modules [1]. However, when the output filter inductor current is selected as the feedback signal for the inner current loop, all of module filter inductor currents could have the same phase. For achieving OVS, all the module output filter capacitor should be well matched. Here, a two-module DC–AC ISOS system is illustrated as an example. For this system, we have iLfj iCfj + io (j = 1, 2), where, iLfj is the output filter inductor current, iCfj is the output filter capacitor current, and io is output current of the system. So, if the output filter capacitors are equal, the equaling of the output filter inductor currents (iLf1 iLf2 ) will ensure that the output filter capacitor currents are equal, i.e., iCf1 iCf2 , leading to vo1 vo2 . As a result, OVS can be achieved. If the output filter capacitors of two modules are not matched, for example, C f1 > C f2 , the IVS loop regulates so that each module’s filter inductor current shares the same phase while differing in amplitude, and the final steadystate phasor diagram is shown in Fig. 7.3a. It can be seen that the output voltage difference between the two modules, vo1 − vo2 , lags load current io by 90°, which means there is a difference of the reactive power between the two modules, and it presents the capacitive reactive power. On the contrary, if the output filter capacitor current is selected as the feedback signal, when there is a difference between the output filter capacitors of two modules (also assuming C f1 > C f2 ), the IVS loop could ensure active power balance (i.e., (7.5) is established), and the role of the output synchronization could guarantee the establishment of (7.1). Thus, OVS is achieved (i.e. (7.7) is established). So, if the output filter capacitors is not matched, the IVS loop will adjust the capacitor current to attain the difference amplitude but the same phase at steady state as shown in Fig. 7.3b. From the above analysis, with the output filtering inductor current feedback, OVS can be achieved only if the output filter capacitors of the module are matched; and with the output filter capacitor current feedback, OVS can be achieved no matter the output filter capacitors are matched or not.

158

7 Compound Balanced Control Strategy …

irefj +

-h

_ if

Vdc

vinv

h

_

iLf

1 sLf

+

+_ io

-Vdc

iCf

1 sCf

voj

Ki Fig. 7.4 Control block diagram of inverter with three-state hysteresis modulation

7.2 Decoupling the Control Loops As seen in Fig. 7.2, in the combined control of IVS and output phase synchronization, there are n + 1 variables including the system output voltage and n modular input voltages, which need to be controlled. These variables seem to interfere with each other and it is very difficult to analyze and design regulators of the system directly. It is, thus, necessary to explore the relationship of these variables to reduce the complexity of the system. In Fig. 7.2, Gvcd and Gvo are the transfer functions of the IVS regulator and the output voltage regulator, respectively. In this chapter, the three-state hysteresis modulation is employed in the inner current loop, which will be described in detail in Chap. 8. Figure 7.4 shows the control block diagram of each inverter module. With well-designed current tracking, the inner current loop, as shown in the dashed box in Fig. 7.4, can be reduced to a current follower with a gain equal to 1/K i [2, 3]. The equivalent control block diagram of Fig. 7.2 is reorganized in Fig. 7.5a, from which we have vo

n j1

⎛

vo j

n i ref − i sv j · j1

⎞ n 1 1 n ⎝ni ref − i sv j ⎠ · ⎝i ref − i sv j ⎠ · s K i Cf n j1 s K i Cf j1 n

⎞

1 s K i Cf ⎛

(7.8)

According to (7.8), Fig. 7.5a can be further simplified as shown in Fig. 7.5b, and we have i sv1 + i sv2 + · · · + i svn

n

(Vin n − vin j ) · K vi · G vcd · i ref j1

⎛

⎝Vin −

n

⎞ vin j ⎠ · K vi · G vcd · i ref

j1

0

(7.9)

7.2 Decoupling the Control Loops

+ Kvi _

vin1

Gvcd

isv1

_ +

Kvi IVS Loop

+ Kvi _

Gvcd

ivd2

isv2

_ +

Kvi

iref1

1 Ki

1 sCf1

vo1

iref2

1 Ki

1 sCf2

vo2 +

1 sCfn

von

…

IVS Loop

…

vin2

ivd1

+ Kvi _

vinn

Gvcd

Kvi

ivdn

isvn

_ +

IVS Loop

+ _

voref

…

Vin n

159

Gvo

irefn

1 Ki

+ vo +

Inner Current Loop

iref

Kvo System Output Voltage Loop

(a) Equivalent control block diagram of Figure 7.2 + Kvi _

vin1

Gvcd

isv1

Kvi IVS Loop

+ Kvi _ Kvi

Gvcd

ivd2

isv2

+

+

IVS Loop

+

1 n

…

vin2

ivd1

…

Vin n

+ Kvi _

vinn

Kvi

voref

+ _

Gvcd

ivdn

isvn

IVS Loop Gvo

iref

_ +

n sKiCf

vo

Kvo System Output Voltage Loop (b) Further equivalent contro l block diagram Fig. 7.5 Decoupling analysis of DC–AC ISOS system with compound balanced control strategy

160

+ Kvi _

vin1

Gvcd

isv1

Kvi IVS Loop

+ Kvi _ Kvi

Gvcd

ivd2

isv2

+

+

IVS Loop

=0 +

…

vin2

ivd1

…

Vin n

7 Compound Balanced Control Strategy …

+ Kvi _

vinn

Kvi

Gvcd

ivdn

isvn

IVS Loop

+ voref _

Gvo

iref

n sKiCf

vo

Kvo System Output Voltage Loop (c) Decoupling block diagram Fig. 7.5 (continued)

According to (7.9), Fig. 7.5b can be further simplified as shown in Fig. 7.5c. Equation (7.9) means that the IVS loops have no effect on the system output voltage loop. The output of the system output voltage regulator iref serves as the initial current reference for the inner current loop of each module. When iref changes, the current reference of each module, irefj , will change simultaneously so that the output power and input power of each module vary accordingly. For example, when IVS and OVS are both achieved initially, the abruptly increasing load would result in a decrease of the output voltage. In order to keep the output voltage unchanged, iref will increase, and the output filter capacitor current and then the output voltage of each module will rise. In the meantime, OVS and the balanced output power of all modules are maintained. According to the principle of power conservation, the input current of each module and hence the total input current of the system will all increase. Since the input current is still being shared and the average current of each input dividing capacitor is zero, IVS is maintained. It can be seen that the operation of the IVS loops is not affected by the system output voltage loop. From the previous analysis,

7.2 Decoupling the Control Loops

161

Fig. 7.6 Block diagram of the output voltage loop

+ voref _

Gvo vof

n sKiCf

vo

Kvo

it is clear that the IVS loop and the system output voltage loop work independently and they are decoupled from each other. Thus, the two control loops can be designed separately.

7.3 Design of the Control Loops for the DC–AC ISOS Systems With no loss of generality, a two-module DC–AC ISOS system is taken as an example in the following to illustrate the design of the IVS loop and output voltage loop. The main parameters of the system are as follows: the number of modules n is 2, the input voltage is 540 V (±10%), the output voltage is 230 V AC/400 Hz, the output capacity is 2 kVA, the input dividing capacitor C d is 1000 µF, the system output voltage sensor gain K vo is 0.031, the input voltage sensor gain K vi is 0.03. The main parameters of each module are as follows: the input voltage is 270 V (±10%), the output voltage is 115 V AC/400 Hz, the output capacity is 1 kVA, the output filter inductor L f is 600 µH, the output filter capacitor C f is 30 µF, the inner current loop sensor gain K i is 0.2. In addition, the parameters of the first-stage fullbridge converter of each module are given in Sect. 6.3.2 and they are not repeated here.

7.3.1 Design of the Output Voltage Loop From Fig. 7.5c, the control block diagram of the system output voltage loop is extracted as shown in Fig. 7.6. From Fig. 7.6, the loop gain of the system output voltage loop can be derived as Tvo_c (s)

n K vo G vo (s) s K i Cf

(7.10)

Substituting the above related parameters into (7.10) and letting Gvo (s) 1, the uncompensated loop gain can be obtained which is expressed as Tvo_u (s)

n K vo 1.03 × 104 s K i Cf s

(7.11)

162

7 Compound Balanced Control Strategy …

Fig. 7.7 PI regulator

For the sake of improving the tracking precision of the output voltage with the reference, a PI regulator is adopted as the output voltage regulator (see Fig. 7.7a), with the transfer function expressed as R2 · G vo (s) R1

s+

1 s + ωz R2 C G vo∞ s s

(7.12)

1 . 2π R2 C The three-state hysteresis current controlled inverter has a broad distribution of harmonics in the output voltage and its switching frequency is variable. Denoting its sampling frequency by f k , the three-state hysteresis current controlled inverter can be equivalent to the PWM inverter whose switching frequency is a quarter of f k [4, 5]. In this chapter, the sampling frequency of the inverter f k is 100 kHz. Then, the switching frequency of the equivalent PWM inverter f s is 25 kHz. One-fifth of f s , i.e., f c 5 kHz [6], is set as the cutoff frequency of the compensated output voltage loop. Figure 7.8 gives the Bode diagram of the uncompensated output voltage loop gain, from which we can find the magnitude of the uncompensated loop gain is −9.66 dB at 5 kHz. Thus, the PI compensator should have a gain of 9.66 dB at 5 kHz, i.e., where the turning frequency of PI regulator is f z

G vo∞

R2 9.66 10 20 3.04 R1

(7.13)

Here, R1 is chosen as 10 k, then we have R2 30.4 k. Finally, R2 30 k is chosen. It can be found in Fig. 7.7b that the PI regulator will not bring large phase lag when 10f z ≤ f c . So, we have C

1 5 ≥ 10.61 (nF) 2π f z R2 π f c R2

(7.14)

We choose C = 10 nF. Then, the transfer function of the PI regulator can be derived as

7.3 Design of the Control Loops for the DC–AC ISOS Systems

163

Fig. 7.8 Bode diagram of the uncompensated and compensated loop gains of the system output voltage closed loop

G vo (s) 3 ·

s + 3.33 × 103 s

(7.15)

The Bode diagram of the compensated output voltage loop gain is sketched in Fig. 7.8. It can be found that the compensated loop gain has a cutoff frequency of 5 kHz with a phase margin of 83.9°.

7.3.2 Design of the Control Loop of the First-Stage Converter and IVS Loop For the first-stage phase-shifted full-bridge converter, the intermediate DC bus voltage closed-loop control is adopted, and its design is similar to that in Sect. 6.3.2. Meanwhile, the design of the IVS loop can also refer to Sect. 6.3.3. Similarly, we can conclude that, if the IVS loop compensator gain satisfies (6.20), each module of the DC–AC ISOS systems could show positive resistance characteristic at the input side and then the system can be stable.

7.4 Simulation and Experimental Verification In order to verify the effectiveness of the compound balanced control strategy, a two-module DC–AC ISOS system is taken as an example in the simulation and experiments. The main parameters of this system have been given in Sect. 7.2.

164

7 Compound Balanced Control Strategy …

(V)

280

vin1

vin2

270

(V)

260 200

(V)

(A)

-20

vo1-vo2

0 -200 400

20 0

vo1 vo2

vo io

0 -400 0.110

0.112

0.114

t (ms)

0.116

0.118

0.120

0.118

0.120

(a) Full resistive load (V)

280

vin1

vin2

270

(V)

260 200

(V)

(A)

-20

vo1-vo2

0 -200 400

20 0

vo1 vo2

vo io

0 -400 0.110

0.112

0.114

t (ms)

0.116

(b) Full inductive load Fig. 7.9 Simulation waveforms at steady state

7.4.1 Simulation Results Figures 7.9a, b gives the steady-state simulation waveforms of the system with the full resistive and full inductive load (cos θ = 0.75), respectively. As shown, the system achieves IVS and OVS under various load conditions. Figure 7.10 shows the system dynamic simulation waveforms. Figure 7.10a gives the system simulation waveforms at full resistive load (2 kW) with the input voltage stepping between 486 V (90% of the rated input voltage) and 594 V (110% of the rated input voltage). Figure 7.10b gives the system simulation waveforms at rated input voltage (540 V) conditions with the load current stepping between 1/3 full load and full load. It can be seen from the waveforms that IVS and OVS are well achieved during the transients.

7.4 Simulation and Experimental Verification

165

(V)

640 560

vin

480

(V)

320 280

vin1

vin2

240

(V)

200

vo1 vo2

0 -200

58

60

66

64

62

t (ms) (a) Step-input-voltage condition

20

(A)

vo1-vo2

68

70

72

io

0

(V)

-20 280 270

vin1

(A)

260 200

vin2 vo1 vo2

vo1-vo2

0 -200

82

86

84

88

90

92

t (ms) (b) Under step-load Fig. 7.10 Dynamic simulation waveforms

Figure 7.11 shows the steady-state simulation waveforms of the system at full resistive load with an unmatched output filter capacitor (C f1 30 µF, C f2 25 µF). It can be found in Fig. 7.11a that when the output filter capacitor current feedback is adopted, the output filter capacitor currents of two modules have the same phase but the different amplitude, and the output voltage of each module are equal. This implies that the unmatched output filter capacitors do not influence OVS when employing the output filter capacitor current feedback. From Fig. 7.11b, it can be found that when the output filter inductor current feedback is employed, the output filter inductor currents of two modules have the same phase but different amplitude, and the output

166

7 Compound Balanced Control Strategy …

Fig. 7.11 Steady-state simulation waveforms of the two kinds of current feedback with unmatched output filter capacitor (at full resistive load)

voltages of two modules have phase difference and vo1 lags behind vo2 . The difference between the two voltages vo1 − vo2 lags load current io by 90°, which indicates that there is the reactive power difference at the outputs of two modules. The simulation results in Fig. 7.11 are in agreement with the theoretical analysis in Sect. 7.1.

7.4 Simulation and Experimental Verification

167

7.4.2 Experimental Results In order to further validate the effectiveness of the proposed compound balanced control strategy, a two-module DC–AC ISOS system prototype is built in the lab and the experiment has been carried out. The parameters are the same as that of the simulation. Figure 7.12 shows the experimental waveforms at steady state with full resistive and full inductive load (cos θ = 0.75). As seen, the system achieves IVS and OVS under various load conditions. Figure 7.13a, b shows the dynamic experimental waveforms when the input voltage of the system is step changed between 486 and 594 V and when the load is step changed between 1/3 full load and full load. As seen, IVS and OCS are well achieved during the transient. Obviously, the compound balanced control strategy can effectively realize IVS and OVS both at steady state and during the transient. Figure 7.14a, b shows the steady-state experimental waveforms of the two-module DC–AC ISOS system (at full resistive load) with output filter capacitor current feedback and output filter inductor current feedback, respectively. Here, the output filter capacitors are intentionally set at different values, i.e., C f1 30 µF, C f2 25 µF. As shown in Fig. 7.14a, despite the difference between the output filter capacitors, the two modules can still effectively achieve OVS when using the output filter capacitor current feedback. While from Fig. 7.14b we can find that when adopting the output filter inductor current feedback, the difference between the output filter capacitors leads to a phase difference between the output voltages of the two modules, and vo1 lags vo2. The difference between the two voltages vo1 − vo2 lags the load current io by 90°, indicating the existence of output reactive power difference between the two modules. Obviously, even if the output filter capacitors are not matched, the filter capacitor current feedback control can still achieve OVS among the constituent modules.

7.5 Distributed Control Strategy for DC–AC ISOS Systems The compound balanced control strategy proposed in Sect. 7.1 includes three control loops, namely, the system output voltage loop, the IVS loop and inner current loop of each module. The output voltage loop is common used for all modules, and this scheme is a centralized control without realizing full modularity of the system. Therefore, in this section, a distributed control method [7] is given and its control block diagram is shown in Fig. 7.15, where each module has three independent control loops: IVS loop, output voltage loop, and inner current loop. And the communication among the modules is achieved through the buses, in which the input voltage sharing bus signal is obtained by the average circuit and its value is V in /n.

168

7 Compound Balanced Control Strategy …

M1 vin1 : [100V/div] M2 vin2 : [100V/div]

C1 vo1 : [200V/div]

C1/M1 C2/M2

F1

F1 vo1-vo2 : [200V/div] C3: vo[400V/div]

C2 vo2 : [200V/div] M4: io[20A/div]

C3/M4

Time: [2ms/div]

(a) At full resistive load M1 vin1 : [100V/div] M2 vin2 : [100V/div]

C1 vo1 : [200V/div]

C1/M1 C2/M2

F1

F1 vo1-vo2 : [200V/div] C3: vo[400V/div]

C2 vo2 : [200V/div] M4: io[20A/div]

C3/M4

Time: [2ms/div]

(b) At full inductive load Fig. 7.12 Experimental waveforms under steady state

Additionally, the output voltage references of all the modules can be synchronized with DSP to form the output voltage reference bus denoted as voref . Moreover, the output voltage feedback signals of all the modules, vofj , are averaged to generate the output voltage average bus, denoted as vof_ave . The error between voref and vof_ave is sent to each output voltage regulator. It should be noted that from Fig. 7.15, the average feedback signal vof_ave of each module is obtained as vof_ave (K vo · vo1 + K vo · vo2 + · · · + K vo · von ) K vo · vo /n

(7.16)

7.5 Distributed Control Strategy for DC–AC ISOS Systems

169

M1 vin : [100V/div] M2 vin1 : [100V/div] M1

M3 vin2 : [100V/div] C2 vo1 : [200V/div]

C2/M2 M3 C3

F1

F1 vo1-vo2 : [200V/div]

C3 vo2 : [200V/div]

Time: [10ms/div]

(a) Under step-input-voltage condition M1 io : [20A/div]

M1

M2 vin1 : [100V/div] M3 vin2 : [100V/div]

C2 vo1 : [200V/div]

C2 M2 C3/M3

F1

F1 vo1-vo2 : [200V/div]

C3 vo2 : [200V/div]

Time: [5ms/div]

(b) Under step-load Fig. 7.13 Dynamic experimental waveforms.

Equation (7.16) shows that it is the average of all the module’s output voltages that tracks the synchronous voltage reference voref in each independent output voltage loop. So, the control effect here is equivalent to that of the previous centralized control strategy in which the output voltage of the system is adjusted to realize its sinusoidal output. However, different from the centralized control strategy, the proposed distributed control strategy has a set of independent output voltage loop for each inverter module. In addition, the common current reference bus signal iave of each module is also obtained by means of the average circuit, and the inner current loop of each module is independent of each other. It can be seen that in the distributed

170

7 Compound Balanced Control Strategy … M1 vin1 : [100V/div] M2 vin2 : [100V/div]

C1 vo1 : [200V/div]

C1/M1 C2/M2

F1

F1 vo1-vo2 : [200V/div] C3: vo[400V/div]

C2 vo2 : [200V/div] M4: io[20A/div]

C3/M4

Time: [2ms/div]

(a) With the output filter capacitor current feedback M1 vin1 : [100V/div] M2 vin2 : [100V/div]

C1 vo1 : [200V/div]

C1/M1 C2/M2

F1

F1 vo1-vo2 : [200V/div] C3: vo[400V/div]

C2 vo2 : [200V/div] M4: io[20A/div]

C3/M4

Time: [2ms/div]

(b) With the output filter inductor current feedback Fig. 7.14 Steady-state experimental waveforms of the two kinds of current feedback with unmatched output filter capacitor (at full resistive load)

control method, the control circuits are scattered in individual modules so that each module can work independently and the true modularity is achieved for the system. The control mechanism of the input and output voltage sharing of the proposed strategy can be further illustrated as follows. Actually, the IVS loop alters the real power at the output terminal to achieve IVS. Since the output signals of the IVS regulator ivdj are DC components while the variables at the output terminal are AC quantities. So the multipliers are introduced here to multiply the DC deviation ivdj with the AC variable iave . In this way, the phase of the signal that is used to tune the current reference by the IVS loop is kept the same with that of the original common current reference iave so as to ensure the synchronization of the current reference, irefj , with the IVS loop only slightly adjusting the magnitude of it. From the compound balanced control conception, we know that both IVS and OVS could be achieved under IVS control combined with the synchronization of the power factor angle of each module simultaneously. So here the output filter capacitor feedback

ig1

Gvcd

Gvcd

+

ign voref + Gvo _ vof_ave Output Voltage Loop

IVS Loop + Kvi _ vinn Kvi

…

ig2 voref + Gvo _ vof_ave Output Voltage Loop

IVS Loop + Kvi _ vin2 Kvi

…

+

+

ivdn

1 n

ivd2

ivd1 _ iLf1 1 sLf +_

iCf2

iCf1

+ +

1 sCf vo2

1 vo1 sCf

+

Kvo

Ki

Inner Current Loop _ irefn iLfn _ iCfn vinvn _ + + 1 1 von G G i inv _ sLf + sCf +

isvn

Common Current Reference Bus

Kvo

Ki

Inner Current Loop _ iref2 iLf2 vinv2 _ + + 1 G G i inv _ sLf +_ +

isv2

Kvo

Ki

isv1 Inner Current Loop _ iref1 vinv1 + + G G i inv _ +

iave

Fig. 7.15 Block diagram of distributed control strategy for DC–AC ISOS system

Output IVS Output Voltage 1 + + Bus Voltage Average n + Reference Bus Bus

Gvo

Gvcd

Output Voltage Loop

voref + _ vof_ave

IVS Loop + Kvi _ vin1 Kvi

…

vof_ave

…

…

voref Vin n

io

vo

1 ZLd

7.5 Distributed Control Strategy for DC–AC ISOS Systems 171

172

7 Compound Balanced Control Strategy …

is adopted for the inner current loops. Since the output voltage lags the output filter capacitor current in phase by 90° and all the output filter capacitor currents track the synchronous current references, phase synchronization of the output voltages of all the modules is fulfilled. Similar with the aforementioned centralized control, the output filter inductor current feedback can also be used in the distributed control of DC–AC ISOS systems, which is not repeated here.

7.6 Summary Aiming for achieving IVS and OVS for the DC–AC ISOS systems, this chapter presents an implementation of the compound balanced control strategy, i.e., the combined control strategy of IVS and output phase synchronization. For the inner current loop, when adopting the output filter inductor current feedback, OVS can be realized only if the output filter capacitors are well matched; while when the output filter capacitor current feedback is adopted, OVS can be achieved no matter the output filter capacitors are matched or not. The compound balanced control strategy with the output filter capacitor current feedback is analyzed, and it is pointed out that the IVS loop and the system output voltage loop are decoupled each other, and these control loops can be designed independantly. A prototype of two-module DC–AC ISOS systems has been fabricated in the lab. The simulation and experimental results show that the proposed combined control of IVS and output phase synchronization is effective. Moreover, based on the centralized control, a distributed control scheme is proposed, in which all the control loops are dispersed into each module, and hence the modularity of the system is achieved.

References 1. Fang, T., Ruan, X., Tse, C.K.: Control strategy to achieve input and output voltage sharing for input-series-output-series connected inverters systems. IEEE Trans. Power Electron. 25(6), 1585–1596 (2010) 2. Xing, Y., Huang, L., Yan, Y.: Redundant parallel control for current regulated inverters with instantaneous current sharing. In: Proceedings of the IEEE Power Electronics Specialists Conference, pp. 1438–1442 (2003) 3. Chen, L., Xiao, L., Gong, C., Yan, Y.: Circulating current’s characteristics analysis and the control strategy of parallel system based on double close-loop controlled VSI. In: Proceedings of the IEEE Power Electronics Specialists Conference, pp. 4791–4797 (2004) 4. Venkataramanan, G., Divan, D.M., Jahns, T.M.: Discrete pulse modulation strategies for highfrequency inverter systems. IEEE Trans. Power Electron. 8(3), 279–287 (1993)

References

173

5. Venkataramanan, G., Divan, D.M.: Pulse width modulation with resonant dc link converters. IEEE Trans. Ind. Electron. 9(1), 113–120 (1993) 6. Li, A., Zhang, C.: Modern Inverter Technology and its Application. Science Press, Beijing, China (2000) (in Chinese) 7. Zhu, H., Fang, T., Ruan, X.: Distributed voltage sharing control strategy for input-series-outputseries inverters system. In: Proceedings of the Annual Conference of the IEEE Industrial Electronics Society, pp. 647–652 (2013)

Chapter 8

An Improved Average Current Control Strategy for DC–AC Input-Parallel-Output-Parallel Systems

Abstract For the DC–AC input-parallel-output-parallel (IPOP) systems, the output current sharing control strategy is popular and effective to achieve output current sharing (OCS) among the constituent modules, and the average current control is usually adopted. The load current feed-forward scheme is incorporated into the average current control and the load current feed-forward node is located before the current-limiting unit, for not only improving the DC–AC IPOP system output voltage characteristics but also limiting the load current. The prototype of two-module DC–AC IPOP systems is built in the lab, and the steady-state and dynamic experiments are provided to verify the superiority of the proposed improved strategy. Keywords Input-parallel-output-parallel systems · Average current control Load current feed-forward · Output characteristics Three-state hysteresis modulation The DC–AC input-parallel-output-parallel (IPOP) systems, commonly known as the parallel inverter system, features ease of system capacity expansion and achievement of N + 1 redundancy of system, improving the system reliability. It has been widely used in uninterruptible power supply (UPS), aeronautical static inverters, etc. It has been pointed out in Chap. 5 that the output current sharing control strategy or the compound input current sharing control strategy could be applied to DC–AC IPOP system. Comparatively, the output current sharing control is quite simpler than the compound input current sharing control strategy, and it is usually employed. In Sect. 1.4.1, various implementations of the output current sharing control strategy have been presented [1–13]. Centralized control and master–slave control are simple, but they have poor redundancy owing to relying on the central control unit or the master module to operate. Wireless droop control is not susceptible to interference and is easy to implement the redundancy due to no connection among the constituent modules, but the output characteristics are poor. As one of the distributed control schemes, the average current control has the merits of simple control circuit, prompt current sharing regulating speed, and ease of implementation of hot-swap and redundant control. However, the DC–AC IPOP system with the average current control also has the defect of poor output characteristics. For tackling this issue, the chapter proposes © Springer Nature Singapore Pte Ltd. and Science Press, Beijing, China 2019 X. Ruan et al., Control of Series-Parallel Conversion Systems, CPSS Power Electronics Series, https://doi.org/10.1007/978-981-13-2760-5_8

175

176

8 An Improved Average Current Control …

a control strategy of the inductor current feedback combined with the load current feed forward, which can greatly improve the output characteristics with preservation of the original current-limiting function.

8.1 Current Three-State Hysteresis Modulation In this chapter, the basic module in the DC–AC IPOP systems still adopts the two-stage configuration consisting of full-bridge DC–DC converter and full-bridge inverter, as shown in Fig. 8.1. Same as that in Chaps. 6 and 7, the first-stage fullbridge DC–DC converter controls its output voltage V dc . The second-stage full-bridge inverter adopts double closed-loop control, composed of outer voltage control and inner current loop. The inner current loop adopts the three-state hysteresis modulation and the control variable is the filter inductor current iLf . The operating waveforms of the three-state hysteresis modulation are given in Fig. 8.2, where iref is the current reference, if is the sensed filter inductor current, 2h is the current hysteresis, and Clk is the sampling clock. If if < iref − h at the sampling instant, power switches S 1 and S 4 are turned on, and the output voltage of inverter bridge, vinv , is equal to +V dc , which is defined as +1 state. Because V dc > vo , iLf is forced to increase. If if > iref + h at the sampling instant, S 2 and S 3 are turned on, and thus vinv −V dc , which is defined as −1 state. Because −V dc < vo , iLf declines. If if is within the hysteresis width at the sampling instant, i.e., iref − h < if < iref + h, S 2 and S 4 are turn on, or S 1 and S 3 are turned on, and thus vinv 0, which is defined as zero state. At zero state, iLf is freewheeling, and whether it increases or decreases depends on the polarity of the output voltage vo , namely, if vo is positive, iLf declines; and if vo is negative, iLf rises. The rates of the rise and decline of iLf in zero state are smaller than that of +1 and −1 state, respectively. Since the inverter could be

Iin

Full-Bridge DC-DC Converter DR1 DR3

Full-Bridge Inverter

Ldc iLdc

Q1

Q3

Tr +

S1 A +

* *

Q2

_ B S2

Q4 DR2 DR4

Fig. 8.1 Main circuit topology of basic inverter module

io

iLf vinv

Vdc Cdc _

Vin

Lf

S3

S4

Cf

+ ZLd vo _

8.1 Current Three-State Hysteresis Modulation

177

operated at +1, 0, and −1 states, the above current control is called the three-state hysteresis modulation. In summary, at the sampling instant, the operating state S is determined as ⎧ ⎪ i f < i ref − h ⎨ +1, (8.1) S 0, i ref − h < i f < i ref + h ⎪ ⎩ −1, i f > i ref + h As shown in Fig. 8.2, it is noteworthy that the operating state does not change immediately once the inductor current exceeds the hysteresis band, and it does only when the sampling clock comes. In addition, when the output voltage is in positive (negative) half cycle, the inverter switches between +1 (or −1) and 0 states. This implies that the inverter is operated with unipolar modulation. Compared with the bipolar modulation where the state of the inverter is modulated between +1 state and −1 state [14], the three-state hysteresis modulation has the following advantages: (1) The ripple of filter inductor current is relatively small. (2) Due to the presence of freewheeling mode, the feedback energy to the intermediate bus is relatively small and thus has little influence on the DC bus voltage. Figure 8.3a shows the double closed-loop control block diagram of the secondstage full-bridge inverter, where the outer voltage loop Gvo adopts PI regulator, and the inner current loop adopts the three-state hysteresis modulation. voref is the output voltage reference, vo is the output voltage, K vo is the sensor gain of the output voltage, and C f is the output filter capacitor, and Z Ld is the load impedance. Generally,

Fig. 8.2 Operating waveforms of three-state hysteresis modulation

178

8 An Improved Average Current Control …

Fig. 8.3 Double closed-loop control block diagram

the switching frequency is much higher than the output frequency, the inner loop (referring to the dashed frame in Fig. 8.3a) can be simplified as the current follower whose magnification is 1/K iL [15]. Thus, Fig. 8.3a can be reduced to Fig. 8.3b when the load current feed forward is not included.

8.2 Improvement of the Output Characteristics of Inverter As mentioned above, a PI regulator is chosen as the outer voltage regulator for the inverter module. Although the output voltage static error could be very small with properly designed PI regulator, it cannot be fully eliminated. In order to improve the output characteristics, the load current feed forward [16] is introduced as shown with the dashed lines in Fig. 8.3b, where K io is load current sensor gain. The load current feed-forward could be located at point ➀ or ➁, and the function of the current-limiting unit is different, which will be analyzed later. Obviously, under normal operating condition when the current-limiting unit is not activated, the output characteristics of the inverter are the same. With the current-limiting unit omitted, by replacing the feed-forward signal io with vo , Fig. 8.3b can be equivalent to Fig. 8.4a. By simplifying the transfer function from iLf to vo in the dashed box, Fig. 8.4a can be further reduced to Fig. 8.4b. As mentioned above, the voltage regulator adopted a PI regulator is expressed as 1 . Hence, according to Fig. 8.4b, the transfer function from the G vo (s) K PI + sτPI output voltage reference to the output voltage can be derived as

8.2 Improvement of the Output Characteristics of Inverter

io

Kio voref + _

Gvo

+ + iref

1/KiL

179

1 ZLd

iLf +

io _

ZLd

vo

sCf Kvo

(a) Equivalent control block diagram with current-limiting unit omitted io

Kio voref + _

Gvo

+ + iref

1/KiL

iLf

1 ZLd ZLd 1+sCfZLd

vo

Kvo

(b) Further equivalent control block diagram Fig. 8.4 Transfer function derivation of the inverter module

Vo (s) Voref (s)

s 2 K iL Cf + s

K PI s +

1 τPI

K vo K iL − K io + K vo K PI + Z Ld τPI

(8.2)

where the load Z Ld can be expressed in the following form: Z Ld |Z Ld |(cos ϕ + j sin ϕ)

(8.3)

where |Z Ld | is the magnitude of the load, and ϕ is the power factor angle of the load. If ϕ = 0, the load is purely resistive; if ϕ > 0, the load is inductive; and If ϕ < 0, the load is capacitive. Substituting (8.3) into (8.2), we have 2 ω2 + 1 K PI 2 τPI Vo ( jω) 2 2 Voref ( jω) K vo ω(K iL − K io ) sin ϕ (K iL − K io ) cos ϕ + K vo K PI ω2 − Cf K iL ω2 + + |Z Ld | |Z Ld | τPI

(8.4) According to (8.4), when the load current feed forward is not introduced, i.e., K io 0, with the parameters of the inverters given in Sect. 8.4, the curves of the output voltage versus the load current with different ϕ can be depicted as shown in Fig. 8.5. From Fig. 8.5, it can be found that, (1) When the load is resistive or inductive (ϕ ≥ 0°), the output voltage declines with the increase of the load current, and the larger the ϕ is the more the output voltage declines; and (2) When the load is capacitive (ϕ < 0°), the output voltage rises with the increase of the load current, and the larger

180

8 An Improved Average Current Control …

Fig. 8.5 Output voltage characteristics with and without the load current feed forward

the ϕ is the more the output voltage rises. In conclusion, the output characteristics are relatively poor when the load current feed-forward is not incorporated. When the load current feed forward is introduced and K io is set to be equal to K iL , it can be found from (8.4) V o is independent of the load as shown in Fig. 8.5, which is the desired output characteristic. In conclusion, when introducing the load current feed-forward and letting K io K iL , the output voltage of inverter is independent of the load and the optimal output characteristics is achieved. At this time, the feedback signal is the capacitor current in the current loop [16]. When the load current feed-forward node is located at point ➀, the current-limiting unit (as shown in Fig. 8.3b) actually limits the capacitor current, and the load current cannot be limited since the capacitor current does not contain the information of the load current. When the load current feed-forward node is located at point ➁, the current-limiting unit limits the inductor current and thus the load current. So, it is desirable to locate the load current feed-forward node at point ➁.

8.3 Configuration and Control Scheme for DC–AC IPOP Systems 8.3.1 Configuration of the DC–AC IPOP Systems As illustrated above, the DC–AC IPOP systems generally adopts the OCS control strategy, which could be implemented with centralized, master–slave, or distributed control. The distributed control includes average current control and wireless droop control. In this chapter, the average current distributed control is adopted [7], and the

8.3 Configuration and Control Scheme for DC–AC IPOP Systems

181

system configuration is shown in Fig. 8.6a. As seen, there are two signals among the control circuits of each module. One is the synchronizing signal syn, which ensures the sinusoidal voltage references of modules synchronous, and the synchronizing circuit is shown in Fig. 8.6b. The other is the averaging current reference iave , which is obtained by averaging all the outputs of modular voltage regulators as shown in Fig. 8.6c. iave is regarded as the common current reference of all the current regulators achieving the current sharing among the modules. K 1 is the output relay and K 2 is the current reference access-control relay. Once a module fails, its K 1 and K 2 will be turned off, and the driving signals for the switches will be disabled. The other modules will supply the load.

8.3.1.1

Synchronous Circuit

To ensure stable operation of the DC–AC IPOP systems, the amplitude, frequency and phase of the output voltage of each inverter module are all required to be consistent. The voltage reference of each module is generated by the high-precision D/A chip (MAX507) of the respective module, the frequencies of the output voltages of modules can be the same. Figure 8.6b gives the phase synchronizing circuit for the voltage reference [17], in which each voltage reference tracks the phase information of the synchronizing bus, syn. In each inverter module, a rising edge is emitted via signal port PB1 at the positive zero-crossing point of the voltage reference, and it induces a rising edge on the synchronizing bus. For the synchronizing bus, the first rising edge signal is regarded as the valid signal, and it is sent to the capture port CAP1 as a common synchronizing signal to fulfill the transmission and judgment of synchronizing signal. CAP1 captures the trailing edge and generates the synchronous interrupt request, and the DSP enters capture interrupt service routine. Here, the phase of the corresponding voltage reference starts from the positive zero-crossing of sinusoidal wave, achieving the zero-crossing phase-lock. From Fig. 8.6b, it can be found that as long as port PB1 of DSP of any inverter module is high, synchronous bus syn is high, which is tracked by the other inverter modules to achieve synchronization. With the synchronizing circuit, the phase difference of the output voltage in each cycle can be corrected, and the safety and reliable operation of the DC–AC IPOP systems will be ensured.

8.3.1.2

Instantaneous Average Circuit

In addition to ensuring synchronization of the output voltage phase, the output current sharing control is required for achieving power sharing among the inverter modules for DC–AC IPOP systems. So, it is better to produce the averaging instantaneous current reference for the modules, and the averaging circuit is shown in Fig. 8.6c, where irefj is the inductor current reference signal of jth inverter module (j = 1, 2, …, n). The average value of the current reference signals can be obtained as

182

8 An Improved Average Current Control … Vin

vo

syn iave

K1 Inverter 1

iref1 K2 DSP 1

syn

K1 Inverter 2

iref2 K2 DSP 2

syn

K1 Inverter n

irefn K2 DSP n

syn

(a) Block diagram of system structure syn +15V

+15V

INV1

+15V

INV2

INVn

PB11

PB12

PB1n

CAP11

CAP12

CAP1n

(b) Synchronizing circuit _ iref1 iave

R

+

_ iave

iref2

R

+

_ iave

irefn

R

(c) Current reference instantaneous averaging circuit Fig. 8.6 Block diagram of IPOP inverter

+

iave

8.3 Configuration and Control Scheme for DC–AC IPOP Systems

i ave (i ref1 + i ref2 + · · · + i refn ) n

183

(8.5)

And iave serves as the common current reference for each module. From (8.5) and Fig. 8.6c, we can also see that this averaging circuit, which is equally distributed in each module, is independent on the number of parallel inverter modules. Thus, the distributed control method can achieve redundant control of the DC–AC IPOP systems.

8.3.2 Control Scheme As indicated in [7], the outputs of voltage regulators of all the inverter modules are averaged, and the average signal serves as the current reference for the inductor current of all inverter modules and it is independent on a particular module, allowing hot-swap of any module. The block diagram of the average current distributed control for DC–AC IPOP systems is shown in Fig. 8.7, where each inner current loop of threestate hysteresis modulation has been simplified as a current follower, and the load current feed forward is incorporated to improve the output characteristics. According to the analysis of Sect. 8.2, when the load current feed forward is located before the current-limiting unit K ioj K iLj (j = 1, 2, …, n), the output voltage will be independent of the load, while the current-limiting function is preserved.

8.4 Experimental Results To verify the validity of the theoretical analysis, a 2-kVA two-module DC–AC IPOP system prototype is fabricated in the laboratory. The specifications of the inverter module are as follows: Rated output power: Po_mod 1 kVA, input voltage V in 270 VDC, output voltage V o 115 VAC/400 Hz, output filter inductor L f 600 µH, sensor gain K ioj K iLj 0.2. Figure 8.8 gives the measured output voltage versus the output current of the system with and without the load current feed forward at the resistive load. As seen, the output characteristics of the system have been greatly improved with the introduction of the load current feed-forward scheme. Figure 8.9 gives the experimental waveforms of a single inverter when the load is step changed between the rated resistive load current and over current. Here, the output current of the inverter is limited at 1.5 times of the rated load. As seen, the inductor current is limited and thus the load current is limited, and the inverter has good dynamic performance. Figure 8.10 gives the experimental waveforms of the DC–AC IPOP system when the load steps up and down between the rated resistive load current and overcurrent. It also can be seen that the load currents of the two inverter modules are successfully

184

8 An Improved Average Current Control …

voref

iave Kio

voref +_

Gvo1

+

+ iref1

1/KiL1

vof

sCf1

iCf1 _

iLf1 +

io1

Kvo

Module # 1

Average Current Bus irefj +

+ +

voref +_

Gvon

+

ioj

1 n

+

vo

+ +

ZLd

Kio

+ irefn

1/KiLn

vof

iLfn + _ iCfn

ion sCfn

Kvo

Module #n

Fig. 8.7 Block diagram of the control in DC–AC IPOP systems 116.0

Output voltage (V)

115.5 115.0 114.5 114.0 113.5 113.0 112.5

Without the load current feed-forward With the load current feed-forward 0

5

10 Output current (A)

15

20

Fig. 8.8 Measured output characteristics curves of the system

limited, and the DC–AC IPOP system exhibits excellent dynamic performance. It is evident that the current-limiting function is effective, which is well in agreement with the theoretical analysis.

8.4 Experimental Results

185

vo [100V/div]

iL [10A/div]

io [10A/div]

Time: [5ms/div] Fig. 8.9 Experimental waveforms of a single inverter module with stepped load current

vo [100V/div]

io1 [10A/div]

io2 [10A/div]

Time: [5ms/div] Fig. 8.10 Experimental waveforms of the DC–AC IPOP system with stepped load current

Figure 8.11 gives steady-state experimental waveforms of the DC–AC IPOP system with the improved method at full resistive load. As seen, the difference between the output currents of the two inverter modules is fully eliminated, which means OCS is achieved. Figure 8.12 shows the dynamic experimental waveforms of the DC–AC IPOP system when the load current is step changed between the full resistive load and no-load. It can be seen that the output currents of the two modules are equal during the transient showing that OCS is also achieved. Figure 8.13a shows the experimental waveforms of the system output voltage vo , output current of module 1, io1 , output current of module 2, io2 , when module 2 is plugged into the DC–AC IPOP system. As seen, before module 2 is plugged into

186

8 An Improved Average Current Control …

–

Fig. 8.11 Steady-state waveforms of the DC–AC IPOP system at full resistive load

Fig. 8.12 Dynamic waveforms of the DC–AC IPOP system

the system, module 1 provides all the load current; and when module 2 is plugged into the system, io1 declines, and io2 increases from zero, and finally, they provide the load current equally, and there is no obvious distortion in vo . Figure 8.13b gives the experimental waveforms when module 2 is plugged out from the DC–AC IPOP system. When the control circuit produces the plug-out command, power switches in module 2 are turned off, and module 2 is plugged out from the DC–AC IPOP system after about eight output cycles due to the delay of the output relay operation. So, during the period, module 1 supplies power to the load and the output filter capacitor of module 2, and io2 initially reduces to the output filter capacitor current and then reduces to zero after around 20 ms. It can be seen that there is no obvious distortion in vo in the above process. Figure 8.13 illustrates that the DC–AC IPOP system has excellent dynamic performance, and hot-swap is also achieved.

8.5 Summary

187

v o [100V/div]

i o1 [5A/div]

i o2 [5A/div]

Time: [2ms/div]

(a) Module 2 is plugged into the IPOP operation

v o [100V/div]

io1[5A/div] io2[5A/div] Time: [2ms/div]

(b) Module 2 is cut off from thedc-ac IPOP system Fig. 8.13 Hot-swap waveforms of the DC–AC IPOP system

8.5 Summary This chapter first introduces the main circuit topology and control method of the inverter module, and the operation principle of the current three-state hysteresis modulation is also presented. The load current feed-forward scheme is incorporated for making the output voltage of the inverter independent of load current. It

188

8 An Improved Average Current Control …

is worth noting that the load current feed-forward node should be located before the current-limiting unit, for not only improving the inverter output characteristics but also limiting the load current. This idea is also extended to the DC–AC IPOP systems. Second, this chapter puts forward the system architecture of average current distributed control strategy, which includes two critical signal buses, namely, synchronous voltage reference bus and average current reference bus. The voltage reference of each inverter module tracks the phase information of synchronous bus for the purpose of synchronization. In the meanwhile, the current references of all modules are averaged to obtain the common current reference signal for each module so as to realize output current sharing among the modules. With this average current distributed control, the DC–AC IPOP system achieves the hot-swap function and redundant ability. Finally, the prototype of two-module DC–AC IPOP systems is built in the lab, and the steady-state and dynamic experiments are provided. The experimental results are consistent with the theoretical analysis verifying the superiority of the proposed improved strategy and also realizing hot-swap of DC–AC IPOP system.

References 1. Bekiarov, S.B., Emadi, A.: Uninterruptible power supplies: classification operation dynamics and control. In: Proceedings of IEEE Applied Power Electronics Conference and Exposition, pp. 597–604 (2002) 2. Chen, X., Kang, Y., Chen, J.: Operation, control technique of parallel connected high power three-phase inverters. In: Proceedings of IEEE International Power Electronics and Motion Control Conference, pp. 956–959 (2004) 3. Chen, J., Chu, C.: Combination voltage-controlled and current-controlled PWM inverter for UPS parallel operation. IEEE Trans. Power Electron. 10(5), 547–558 (1995) 4. Lee, W.C., Lee, T.K., Lee, S.H., Kim, K.H., Hyun, D.S., Suh, I.Y.: A master and slave control strategy for parallel operation of three-phase UPS systems with different ratings. In: Proceedings of IEEE Applied Power Electronics Conference and Exposition, pp. 456–462 (2004) 5. Chiang, S.J., Lin, C.H., Yen, C.Y.: Current limitation control technique for parallel operation of UPS inverters. In: Proceedings of IEEE Power Electronics Specialists Conference, pp. 1922–1926 (2004) 6. Perez-Ladron, G., Cardenas, V., Espinosa, G.: Analysis and implementation of a master-slave control based on a passivity approach for parallel inverters operation. In: Proceedings of IEEE International Power Electronics Conference, pp. 1–5 (2006) 7. Xing, Y., Huang, L., Yan, Y.: Redundant parallel control for current regulated inverters with instantaneous current sharing. In: Proceedings of IEEE Power Electronics Specialists Conference, pp. 1438–1442 (2003) 8. Fang, T., Ruan, X., Xiao, L., Liu, A.: An improved distributed control strategy for parallel inverters. In: Proceedings of IEEE Power Electronics Specialists Conference, pp. 3500–3505 (2008) 9. Tuladhar, A.: Advanced control techniques for parallel inverter operation without control interconnections. Ph.D. dissertation, The University of British Columbia, Vancouver, Canada (2000) 10. Lin, X., Feng, F., Duan, S., Kang, Y., Chen, J.: The droop characteristic decoupling control of parallel connected UPS with no control interconnection. In: Proceedings of IEEE International Electric Machines and Drives Conference, pp. 1777–1780 (2003)

References

189

11. De Brabandere, K., Bolsens, B., Van den Keybus, J., Woyte, A., Driesen, J., Belmans, R.: A voltage and frequency droop control method for parallel inverters. IEEE Trans. Power Electron. 22(4), 1107–1115 (2007) 12. Yao, W., Chen, M., Gao, M., Qian, Z.: A wireless load sharing controller to improve the performance of parallel-connected inverters. In: Proceedings of IEEE Applied Power Electronics Conference and Exposition, pp. 1628–1631 (2008) 13. Guerrero, J., Vicuña, L., Matas, J., Castilla, M., Miret, J.: Output impedance design of parallelconnected UPS inverters with wireless load-sharing control. IEEE Trans. Ind. Electron. 52(4), 1126–1135 (2005) 14. Kheraluwala, M.H., Divan, D.M.: Delta modulation strategies for resonant link inverters. IEEE Trans. Power Electron. 5(2), 220–228 (1990) 15. Chen, L., Xiao, L., Gong, C., Yan, Y.: Circulating current’s characteristics analysis and the control strategy of parallel system based on double close-loop controlled VSI. In: Proceedings of IEEE Power Electronics Specialists Conference, pp. 4791–4797 (2004) 16. Fang, T., Ruan, X., Xiao, L., Liu, A.: An improved distributed control strategy for parallel inverters. In: Proceedings of IEEE Power Electronics Specialists Conference, pp. 3500–3505 (2008) 17. Fang, T.: Control strategies for input-parallel-output-parallel and input-series-output- series connected combined inverter systems (in Chinese). Ph.D. dissertation, Nanjing University of Aeronautics and Astronautics, Nanjing, China (2008)

Chapter 9

Input Voltage Sharing Control Strategy for ISOP Systems Under Extreme Load Conditions

Abstract Input-series and output-parallel (ISOP) systems are very suitable for high input voltage and high power applications. Input voltage sharing (IVS) and output current sharing (OCS) of the constituent modules in the ISOP systems should be ensured. This chapter reveals the failure mechanism of IVS of the ISOP systems under extreme load conditions. Based on the existed IVS control strategies, this chapter proposes two methods, including changing the weighted averages of IVS loops and adding an IVS auxiliary circuit, to achieve IVS of the modules under the extreme load conditions. The experimental results are presented to verify the effectiveness of the proposed methods. Keywords Input-series output-parallel system · Extreme load conditions Input voltage sharing · Weighted average · Auxiliary circuit For the ISOP systems, including the DC–DC ISOP system and DC–AC ISOP system, the control strategies proposed from Chaps. 2 to 7 realize IVS all by adjusting the module output power by means of regulating the duty cycle of the constituent modules. When the ISOP systems operate under extreme load conditions such as light load or short-circuit, the module output powers are very small, and the duty cycles are also very small. With the small duty cycle, the adjustment range of the duty cycle is limited, and the abovementioned control strategies cannot ensure IVS anymore, which may cause the module with higher input voltage damaged. This chapter will reveal the failure mechanism of IVS of the ISOP systems under extreme load conditions. Furthermore, this chapter will propose two methods, including changing the weighted averages of IVS loop and adding an auxiliary IVS circuit to achieve IVS for ISOP systems under extreme load conditions [1].

© Springer Nature Singapore Pte Ltd. and Science Press, Beijing, China 2019 X. Ruan et al., Control of Series-Parallel Conversion Systems, CPSS Power Electronics Series, https://doi.org/10.1007/978-981-13-2760-5_9

191

192

9 Input Voltage Sharing Control Strategy for ISOP Systems …

Iin

Iin1

+ Icd1 Vin1 _ Cd1

Io1

Io +

DC-DC (DC-AC) Module

Cf1

RLd

Vo _

Module 1

Vin

Iin2 + Icd2 Vin2 _ Cd2

Io2 DC-DC (DC-AC) Module

Cf2 Module 2

Fig. 9.1 Two-module ISOP system

9.1 Reason of Input Voltages Imbalance of ISOP Systems Under Extreme Load Conditions In Chaps. 2 and 5, the conversion efficiencies of the modules are assumed to be equal. However, this is impossible in practical circuits. This section will analyze the effect of mismatched conversion efficiencies on IVS under extreme load conditions. For simplification of analysis, a two-module ISOP system is taken as an example as shown in Fig. 9.1. Assuming the ISOP system operates at steady state and denoting the loss of module as Plossj (j = 1, 2), we have Pin1 Po1 + Ploss1 (9.1) Pin2 Po2 + Ploss2 where Pin1 and Po1 are the input power and output power of module 1, respectively and Pin2 and Po2 are the input power and output power of module 2, respectively. At steady state, the average currents of the input dividing capacitors are zero, i.e., I cd1 = I cd2 0, then the input currents of two modules are equal, i.e., Iin1 Iin2 Iin

(9.2)

With the help of IVS loops, the input voltages of two modules are equal, i.e., Vin1 Vin2 Vin /2

(9.3)

Combining (9.2) and (9.3), we have Pin1 Vin1 Iin1 Vin2 Iin2 Pin2

(9.4)

9.1 Reason of Input Voltages Imbalance of ISOP Systems …

193

This means that if IVS is achieved, the module input powers will be balanced. Combining (9.1) and (9.4), we have Po2 − Po1 Ploss1 − Ploss2

(9.5)

Assuming Ploss1 > Ploss2 , we have Po1 < Po2

(9.6)

From the above analysis, if the module losses are not identical, the IVS loops will adjust the module output power. The output power of the module with larger losses will be reduced and the output power of the module with lower losses will be increased. By doing so, the module input powers will be balanced and thus IVS between the modules is ensured. Obviously, the difference of the module output powers equals the difference of the module losses. Therefore, in order to realize IVS, the total output power of the ISOP system must be higher than the difference of the module losses. Only with this condition, it is possible for the IVS loops to regulate the module output power to make the module input powers be balanced. In the meanwhile, the IVS loops must have enough regulation ability to ensure that the difference of the module output powers can compensate the difference of the module losses. The output power of ISOP systems is close to zero under extreme load conditions, i.e., light load, no–load, or short-circuit, but the difference of the module losses may still exist and larger than the system output power, which means that the IVS loops cannot compensate for the difference of the module losses by regulating the module output power, leading to input voltages imbalance of the ISOP systems. According to (9.5), we have Io2 − Io1

Po2 − Po1 Ploss1 − Ploss2 Vo Vo

(9.7)

Equation (9.7) indicates that control strategies to ensure IVS for ISOP systems will cause output currents imbalance among the modules when the module conversion efficiencies are different. The output voltage of ISOP systems is close to zero under the short-circuit condition with current-limiting. It will cause large output current different among the modules even if the difference of the module losses is smaller when IVS is ensured. The module with larger output current may fail due to the overcurrent of the power devices. Therefore, we need to not only limit the system total output current but also limit the individual module output current under the condition of short-circuit current-limiting.

194

9 Input Voltage Sharing Control Strategy for ISOP Systems …

Fig. 9.2 Equivalent loss model under extreme load conditions

Cd1

Rloss1

Cd2

Rloss2

Cdn

Rlossn

Vin

9.2 Input Voltage Sharing Control Strategy for ISOP Systems Under Extreme Load Conditions In Sect. 9.1, it is pointed out that under extreme load conditions, the difference among the module losses is the main reason of the input voltages imbalance of ISOP systems with the control strategies mentioned in the previous chapters. Under extreme load conditions, the output power of each module is very small and approximately equal to its loss, therefore, the equivalent loss model of ISOP system under extreme load conditions can be illustrated in Fig. 9.2, where Rlossj (j = 1, 2, …, n) represents the module loss and it should be noted that Rlossj varies depending on the input voltage. At steady state, all Rlossj (j = 1, 2, …, n) have the same flowing current. Hence, the larger the Rlossj , the higher the module input voltage is, and the smaller Rlossj , the lower the module input voltage is. Thus, it is required that Rloss1 Rloss2 ··· Rlossn for achieving IVS. There are two methods to achieve IVS for ISOP systems under extreme load conditions, that is, (1) regulating Rlossj (j = 1, 2, …, n) to be equal and (2) connecting a parallel adjustable resistance to Rlossj (j = 1, 2, …, n) to make the all the equivalent resistances to be equal as shown in Fig. 9.3.

9.2.1 Changing the Weighted Values of the IVS Loops At light load, there is no need for all modules in the ISOP systems to work, and only a few of modules can be activated for supplying the needed load power. For the disabled modules, their equivalent resistances Rloss are infinite. If we can make the modules work intermittently and adjust their working time, then the equivalent resistances of modules can be adjusted to be equal to realize IVS under extreme load conditions.

9.2 Input Voltage Sharing Control Strategy for ISOP …

195

Fig. 9.3 Method of achieving equal equivalent resistance

Cd1

Rloss1

Cd2

Rloss2

Cdn

Rlossn

R1

R2

Vin

Rn

The DC–DC ISOP system is taken as an example to discuss the IVS control strategy under extreme load conditions, which is also suitable for the DC–AC ISOP system. For the convenience of illustration, the IVS control strategy for DC–DC ISOP system (see Fig. 3.19) is redrawn here as shown in Fig. 9.4. In order to explain the proposed control strategy with changeable weighted values of the IVS loops, the weighted value K vcd is added to the output of IVS regulators. When K vcd = 1, Fig. 9.4 is the same as Fig. 3.19. When the input voltage of module j (j = 1, 2,…, n) is lower than V in /n, the output of the IVS regulator vin_EAj increases and the current reference of the inner current loop declines, and when the current reference of inner current loop declines to be lower than zero, vEAj will be also lower than zero, which is lower than the minimum setting value of PWM controller, then the drive signals will be shut down and this module is disabled [2]. Then, the input current I inj 0 and the current of the dividing capacitor I cdj = I in , which charges the dividing capacitor and the module input voltage rises. When the module input voltage rises to be higher V in /n, vin_EAj declines, the current reference of inner current loop as well as the corresponding vEAj increases to be higher than the minimum setting value of PWM controller, then the drive signals will be turned on and the module works again. It can be seen that the modules will work intermittently if the weighted value K vcd is large enough to achieve IVS. To avoid severely regulating the module output current in transient, the weighted value K vcd cannot be too large. However, this cannot ensure the intermittent work of the modules under extreme load conditions. In other words, we hope that K vcd 1 at heavy load and K vcd > 1 at light load. So, two weighted values are set in the control strategy, i.e., 1 and K vcd_light , corresponding to the heavy load and light load conditions, respectively as shown in Fig. 9.5. The two weighted values are selected by a single-pole double-throw, which is controlled by a comparator. When the system output current io is larger than the threshold current I th , the output of comparator is low level and the weighted value of 1 is selected; and if io is smaller than I th , the output of comparator is high level and the weighted value K vcd_light is selected. Generally, the module loss is very small at light load, and the difference among the

196

9 Input Voltage Sharing Control Strategy for ISOP Systems …

+ Vin/n

+ Vin/n

+ Vin/n

– vin1

ivcdn

Gvcd vin_EAn

Kvcd

Gc

vEA1 Comparator 1 _ RS Flip-flop

Drive circuit 1

RS Flip-flop

Drive circuit 2

vEAn Comparator n _ RS Flip-flop + VRAMP n

Drive circuit n

VRAMP 1 – iLf2 +

–

Kvcd

vinn

–

+ ivcd2

Gvcd vin_EA2 IVS loops

– iLf1 +

–

Kvcd

– vin2

–

+

ivcd1 Gvcd vin_EA1

+

Gc Inner current loops

–

– iLfn +

+

Gc

+

vEA2 Comparator 2 _ VRAMP 2

+

vof

Gvo Voref vo_EA Output voltage loop

Fig. 9.4 IVS control strategy for ISOP system Con

+ Vin/n

+ Vin/n

– vin1 Gvcd vin_EA1 – vin2

Gvcd vin_EA2 IVS loops

+ Vin/n

+

–

–

vinn Gvcd vin_EAn

–

io

+

Ith

Comp

1

–

Kvcd_light

+

1

–

Kvcd_light

+

1

–

Kvcd_light

+

– iLf1 +

Gc

Vcl_mod

vEA1 Comparator1 _ VRAMP 1

+

RS Flip-flop

Drive circuit 1

RS Flip-flop

Drive circuit 2

RS Flip-flop

Drive circuit n

vEA2 Comparator 2 _

– iLf2 +

Gc Inner current + VRAMP 2 loops

Vcl_mod

– iLfn + Vcl_mod

Gc

vEAn Comparator n _ VRAMP n

+

vof

Gvo Voref vo_EA Output voltage loop

Vcl_total

Fig. 9.5 Control strategy with changeable weighted values of IVS loops

module losses can be considered to be less than 5%I o , where I o is the system rated output current. So, we set I th 5%I o . The output current of ISOP system has to be limited to protect the system. So, a Zener diode is introduced to the output of the output voltage regulator, and the Zener voltage V cl_total corresponds to the limited system output current as shown in Fig. 9.5. In the meanwhile, Zener diodes are introduced to the current references of the inner current loops to limit the individual module output current, and the Zener

9.2 Input Voltage Sharing Control Strategy for ISOP …

197

voltage V cl_mod corresponds to the limited module output current. V cl_mod must be higher than V cl_total for ensuring that the IVS loops could adjust the inner current loops during the system current-limiting.

9.2.2 Input Voltage Sharing Auxiliary Circuit Referring to the control strategy with changeable weighted values of IVS loops as shown in Fig. 9.5, when the output of ISOP system is shorted, the adjustment of the IVS loops to the current reference of inner current loops is limited by V cl_mod . This means that the IVS loops fail to adjust the current reference of the inner current loop, and cannot achieve IVS under short-circuit condition, and it is only effective at light load. Therefore, we have to introduce IVS auxiliary circuits to the input sides of the modules when short-circuit occurs to regulate the equivalent resistances of all the modules to be equal for IVS. Because the module losses are different and vary with the input voltage, the IVS auxiliary circuits should have the following features: (1) the equivalent resistance of the IVS auxiliary circuit is adjustable; (2) ease of modularization; (3) the IVS auxiliary circuits are only enabled when IVS is not achieved under extreme load conditions to avoid extra loss under normal load conditions. The IVS auxiliary circuits with adjustable resistance and the corresponding control circuits are shown in Fig. 9.6. Each IVS auxiliary circuit is composed of a resistor Rreg and a switch Qaj (j = 1, 2, …, n). Qaj is PWM controlled and the equivalent resistance of the IVS auxiliary circuit is Rreg /DQaj , where DQaj is the duty cycle of Qaj . When DQaj 1, the power dissipated by Rreg must be larger than the maximum difference among the module losses. When the ISOP system works at normal load, the IVS loops can realize IVS among modules, and Qaj is shut off. When the system total output power is lower than the difference among the module losses, the input voltage of the module with larger Rloss will increase from the average value. Until the module input voltage is higher than the allowed maximum input voltage V mod_max , the IVS auxiliary circuit starts to work and regulate DQaj to adjust the power dissipated by Rreg , limiting the module input voltage at V mod_max . The IVS auxiliary circuit can ensure the ISOP system working normally under extreme load conditions, and its structure is simple, but the loss under extreme load conditions is increased. To ensure the stable operation of ISOP system in the full load range, we can utilize the IVS auxiliary circuit or combine the IVS auxiliary circuit with the control strategy with changeable weighted values of the IVS loops. When the output current is lower than I th , we can use the control strategy with changeable weighted values of the IVS loops to realize IVS and the auxiliary circuit is disabled. When the ISOP system output is shorted with current-limiting, the output current is higher than I th , the IVS auxiliary circuit will be activated to limit the module input voltage at V mod_max to ensure the system work stably.

198

9 Input Voltage Sharing Control Strategy for ISOP Systems …

RLd Rreg Cd1

Cf

Qa1

Rreg Vin

Module 1

Cd2

Qa2

Rreg Cdn

Qan

+ Vo –

Module 2

Module n

(a) IVS auxiliary circuit VRAMPComparator _

_

Vmod_max Vinj

+

EA

+

Qaj Driver

(b) Control circuit of IVS auxiliary circuit Fig. 9.6 IVS auxiliary circuit and control circuit

9.3 Simulation and Experimental Verification A DC–DC ISOP system consisting of two phase-shifted full-bridge (PSFB) converter modules is taken as an example as shown in Fig. 9.7, to verify the effectiveness of the proposed methods to realize IVS under extreme load conditions. The specifications of the ISOP system are as follows: input voltage V in 540 Vdc, output voltage V o 48 Vdc, and rated system output current I o 104 A. The main parameters of the PS-FB converters are as follows: switching frequency f s 100 kHz, filter inductance L f1 L f2 10 µH, resonant inductance L r1 L r2 6 µH, and transformer turns ratio K = 14:3. To simulate the module loss, resistors with two different resistances are connected in series with the power switches, i.e., Rl1 0.5 and Rl2 2 , and it is readily to know that Rloss1 < Rloss2 when the input voltages of two modules are equal under extreme load conditions. The threshold current I th is set at 10%I o , i.e., I th 10.4 A. Rreg is chosen as 1 k, and its dissipated power is 73 W when Qaj (j = 1, 2) is always ON and the module input voltage is 270 Vdc. V mod_max is set at 110%V in /2, i.e., V mod_max 297 Vdc. Figure 9.8 shows the simulation waveforms of the ISOP system employing the IVS control strategy (see Fig. 9.4 and K vcd 1) under no-load condition. vin1 and vin2 are the input voltages of modules 1 and 2, respectively, vAB1 and vAB2 are the

9.3 Simulation and Experimental Verification

199

Module 1

Lf1 T1

Q11

Rreg

+ Vin1 –

Rl1 A1 Rl1

Q12 T1

Lr1 ip1

Rl1 B1 Rl1

*

* DR11

iLf1 Cf

RLd + Vo –

* DR12

Qa1 Q13

+ vrect1 –

Q14 Lf2

Vin T2 Rreg

Q21 + Vin2 –

Rl2 A2 Rl2

Q22 T2

Lr2 ip2

Rl2 B2 Rl2

*

Qa2 Q23

+ * DR21 v rect2 –

iLf2

* DR22

Q24 Module 2

Fig. 9.7 Circuit diagram of ISOP system consisting of two PSFB modules

(V)

400

vin2

300

vin1

200

(V)

0 −200 −400 400 200

(V)

100 400 200

0 −200 −400 9.500m

vAB1

vAB2

9.504m

9.508m

9.512m

9.516m

9.520m

9.524m

9.528m

t (s) Fig. 9.8 Simulation waveforms with the IVS control strategy (see Fig. 9.4 and K vcd 1)

voltages across the middle points of the phase-legs of the two modules. It can be seen that vin2 is much higher than vin1 due to Rloss1 < Rloss2 and the voltage difference reaches 220 Vdc, and IVS is not realized under no-load condition.

200

9 Input Voltage Sharing Control Strategy for ISOP Systems … 275

(V)

vin1 270

vin2 265

(V)

400 200

vAB1

0 −200 −400

(V)

400 200

vAB2

0 −200 −400 38.77m

38.78m

38.79m

38.80m

38.81m

38.82m

38.83m

38.84m

38.85m

t (s) Fig. 9.9 Simulation waveforms with the IVS control strategy (see Fig. 9.5 and K vcd_light 2) under no-load condition

Figure 9.9 shows the simulation waveforms of the ISOP system when employing the control strategy with changeable weighted values of the IVS loops (see Fig. 9.5 and K vcd_light 2) under no-load condition. It can be seen that module 1 works intermittently due to Rloss1 < Rloss2 and the average input voltages of two modules are approximately the same. Figure 9.10 shows the simulation waveforms of the ISOP system when employing the control strategy (see Fig. 9.4) and IVS auxiliary circuit under no-load condition. Figure 9.10a shows the waveforms under no-load condition at steady state, where Qa2 is the drive signal of power switch in the IVS auxiliary circuit of module 2. The IVS auxiliary circuit of module 2 works due to Rloss1 < Rloss2 and controls the input voltage of module 2 at 110%V in /2, i.e., 297 Vdc. Figure 9.10b shows the waveforms corresponding to load stepping between half load and no load. At half load, the system output power is larger than the difference between the module losses, the IVS loops can realize IVS and the auxiliary circuit is disabled. At no load, the input voltage of module 2 rises and when it reaches 297 Vdc, the IVS auxiliary circuit of module 2 works and controls the input voltage of module 2 at 297 Vdc, and consequently, the input voltage of module 1 is controlled at 243 Vdc, and the system can operate steadily. Figure 9.11a shows the simulation waveforms of the ISOP system with the control strategy (see Fig. 9.5) and without the IVS auxiliary circuit under short-circuit current-limiting condition. The system output current is limited at 110 A. It can be seen that vin2 is higher than vin1 due to Rloss1 < Rloss2 , the current reference of inner current loop of module 2 increases and when its IVS loop regulator is saturated,

9.3 Simulation and Experimental Verification

201

(V)

(V)

(V)

(V)

400

vin2

300

vin1

200 100 400 200 0 −200 −400 400 200 0 −200 −400 20

vAB1

vAB2

Qa2

10 0 21.662m

21.666m

21.670m

21.674m

21.678m

21.682m

21.686m

21.690m

t (s)

(V)

(A)

(a) Waveforms at steady state 60 40 20 0 −20 60

io

40

vo

20

(V)

0 350

vin2

300 250

vin1

(V)

200 20

Qa2

10 0

2.0m

4.0m

6.0m

8.0m

10.0m

12.0m

14.0m

t (s)

(b) Waveforms corresponding to stepped load Fig. 9.10 Simulation waveforms with the control strategy (see Fig. 9.4) and IVS auxiliary circuit under no-load condition

the output current of module 2 is 65 A, which is the sum of average limited output current 55 and 10 A corresponding to the regulation of IVS loop. The output current of module 1 is 45 A and the difference of the input voltages of two modules reaches 100 Vdc.

(V)

(V)

(V)

(A)

202

9 Input Voltage Sharing Control Strategy for ISOP Systems … 70 60 50 40 30 350

iLf2 iLf1

300

vin2

250

vin1

200 400 200 0 -200 -400 400 200 0 -200 -400 19.76m

vAB1

vAB2

19.78m

19.8m

19.82m

19.84m

19.86m

19.88m

17.34m

17.36m

t (s)

(V)

(V)

(V)

(V)

(A)

(a) Without IVS auxiliary circuit 70 60 50 40 30 350

iLf2 iLf1

vin2

300 250

vin1

200 400 200 0 -200 -400 400 200 0 -200 -400 20

vAB1

vAB2

Qa2

10 0

17.24m

17.26m

17.28m

17.30m t (s)

17.32m

(b) With IVS auxiliary circuit Fig. 9.11 Simulation waveforms of ISOP system with the control strategy (see Fig. 9.5) under short-circuit current-limiting condition

9.3 Simulation and Experimental Verification

203

Figure 9.11b shows the simulation waveforms of the ISOP system with both the control strategy (see Fig. 9.5) and the IVS auxiliary circuit under short-circuit currentlimiting condition. It can be seen that the input voltage of module 2 is controlled at 297 Vdc and its output current is 60 A. An ISOP system prototype is also built in our lab with the same specifications as that in the simulation. For each module with rated input voltage 270 Vdc and under no-load condition, the measured module losses are 48.7 and 33.4 W for modules 1 and 2, respectively, and the power difference is 15.3 W. Figure 9.12a shows the experimental waveforms of the ISOP system with the IVS control strategy (see Fig. 9.4 and K vcd 1) when the system output current is 1 A, i.e., the system output power is 48 W. It can be seen that two modules share the system input voltage well because the system output power is larger than the difference of two module losses. Figure 9.12b shows the experimental waveforms of the ISOP system with the IVS control strategy (see Fig. 9.4 and K vcd 1) under no-load condition. It can be seen that the input voltage of module 2 is much higher than that of module 1 because the no-load loss of module 2 is smaller than that of module 1 with the same input voltage, and the difference of the module input voltages reaches 180 Vdc. Therefore, the ISOP system must adopt special IVS control strategy under extreme load conditions. Figure 9.13 shows the experimental waveforms of ISOP system when employing the control strategy with changeable weighted values of the IVS loops under no-load condition. Figure 9.13a shows the waveforms at steady state. The waveforms of vAB are distorted because the frequency of vAB is 100 kHz but the time scale of the figure is 500 ms/div. The waveforms of vin1 and vin2 indicate that the two modules work intermittently because the two dividing capacitors charge and discharge alternately, leading to the input voltages rise and fall alternately. The average values of the module input voltages are approximately equal. Figure 9.13b shows the waveforms corresponding to load stepping between quarter load and no load. The experimental results shown in Fig. 9.13 verify that the control strategy with changeable weighted values of the IVS loops can ensure the IVS in the full load range. Figure 9.14 shows the experimental waveforms of the ISOP system with the control strategy (see Fig. 9.4 and K vcd 1) and IVS auxiliary circuit under no-load condition. Here, V mod_max is set at 110%V in /2, i.e., V mod_max 297 Vdc. Figure 9.14a, b show the waveforms of vin1 , vin2 , vAB1 , vAB2 , and the drive signal of the auxiliary circuit of module 2 at steady state. As seen, the input voltage of module 2 is controlled at 297 Vdc with the help of IVS auxiliary circuit, and the input voltage of module 1 is controlled at 243 Vdc. Figure 9.14c shows the waveforms corresponding to load stepping between quarter load (25 A) and no load. It can be seen that the IVS is achieved by the control strategy, and the IVS auxiliary circuit is disabled when the output power is larger than the difference between the module losses. Under noload condition, the IVS auxiliary circuit of module 2 starts to work when the input voltage of module 2 rises to 297 Vdc and regulates the input voltage module 2 at 297 Vdc, and consequently, the input voltage of module 1 is controlled at 243 Vdc. The experimental results shown in Fig. 9.14 verify that the IVS auxiliary circuit works well under no-load condition.

204

9 Input Voltage Sharing Control Strategy for ISOP Systems …

C1: vin2 [100V/div] C2: vin1 [100V/div]

C1/C2

C3: vAB1 [400V/div]

C3

C4: vAB2 [400V/div]

C4

Time: [2μs/div]

(a) Output current is 1 A C1: vin2 [100V/div] C2: vin1 [100V/div]

C1/C2 C3

C3: vAB1 [400V/div]

C4: vAB2 [400V/div]

C4

Time: [2μs/div]

(b) No-load Fig. 9.12 Experimental waveforms with the IVS control strategy (see Fig. 9.4 and K vcd 1)

Figure 9.15 shows the experimental waveforms of the ISOP system with the control strategy (see Fig. 9.4 and K vcd 1) and IVS auxiliary circuit under shortcircuit current-limiting condition. For the sake of safety, the system output current is limited at 50 A and output current of each module is limited at 27.5 A (110%I o /2). From Fig. 9.15a, it can be seen that the output current of module 2 is limited at 27.5 A because the no-load loss of module 2 is smaller than that of module 1 under the same input voltage, and output current of module 1 is 22.5 A. Figure 9.15b shows that the input voltage of module 2 is controlled at 297 Vdc with the help of IVS

9.3 Simulation and Experimental Verification

205

C1: vin2 [100V/div] C2: vin1 [100V/div]

C1/C2 C3

C3: vAB1 [400V/div] C4: vAB2 [400V/div]

C4

Time: [500ms/div] (a) Waveforms of steady state

C3

C3: io [20A/div]

C1: vin2 [100V/div] C2: vin1 [100V/div]

C1/C2

Time: [500ms/div] (b) Waveforms corresponding to stepped load Fig. 9.13 Experimental waveforms of ISOP system when employing the control strategy with changeable weighted values of the IVS loops under no-load condition

auxiliary circuit, and the input voltage of module 1 is 243 Vdc. Figure 9.15c shows the waveforms corresponding to load stepping between 25 A and limited current (50 A). It can be seen that the IVS is achieved by the control strategy and the IVS auxiliary circuit is disabled under normal load condition. The experimental results shown in Fig. 9.15 verify that the IVS auxiliary circuit works well under short-circuit current-limiting condition.

206

9 Input Voltage Sharing Control Strategy for ISOP Systems … C1: vin2 [100V/div] C2: vin1 [100V/div]

C1/C2 C3

C4

C3: vAB1 [400V/div]

C4: vAB2 [400V/div] Time: [2μs/div]

(a) Waveforms of vin1, vin2, vAB1, and vAB2 of steady state C1: vin2 [100V/div] C2: vin1 [100V/div]

C1/C2

C3 C3: Qa2 [10V/div] Time: [2μs/div]

(b) Waveforms of vin1, vin2 and drive signal of auxiliary circuit of module 2

C4

C4: io [20A/div]

C1: vin2 [100V/div]

C2: vin1 [100V/div] C3 C1/C2

C3: Qa2 [20V/div]

Time: [500ms/div]

(c) Waveforms corresponding to load stepping Fig. 9.14 Experimental waveforms of ISOP system with the control strategy (see Fig. 9.4) and IVS auxiliary circuit under no-load condition

9.3 Simulation and Experimental Verification

M1: iLf1 [10A/div]

207

C2: iLf2 [10A/div]

M1/C2

C3

C4

C3: vAB1 [400V/div]

C4: vAB2 [400V/div] Time: [2μs/div] (a) Waveforms of vAB1, vAB2, iLf1, and iLf2 of steady state

C2: vin2 [100V/div] C1: vin1 [100V/div]

C1/C2

C3 C3: Qa2 [10V/div] Time: [2μs/div]

(b) Waveforms of vin1, vin2 and drive signal of auxiliary circuit of module 2

C4: vo [40V/div]

C4

C2: vin2 [100V/div] C1: vin1 [100V/div] C3 C3: Qa2 [20V/div]

C1/C2 Time: [500 ms/div]

(c) Waveforms corresponding to load stepping

Fig. 9.15 Experimental waveforms of ISOP system with the control strategy (see Fig. 9.4) and IVS auxiliary circuit under short-circuit current-limiting condition

208

9 Input Voltage Sharing Control Strategy for ISOP Systems …

9.4 Summary This chapter reveals the failure mechanism of IVS of the ISOP systems employing the control strategies proposed from Chaps. 2 to 7 under extreme load conditions, and points out that when the output power of the ISOP systems is smaller than the difference among the module losses, these control strategies cannot regulate enough module output power to compensate for the loss difference, leading to the IVS failure. Moreover, output currents imbalance will occur under short-circuit current-limiting condition. To solve these problems, two methods of achieving IVS under extreme load conditions was proposed, namely the control strategy with changeable weighted values of the IVS loops and the IVS auxiliary circuit. It is recommended that the control strategy with changeable weighted values of the IVS loops should be used under light-load or no-load condition, and the IVS auxiliary circuit should be used under short-circuit current-limiting condition. The effectiveness of the two methods is verified by simulation and experimental results.

References 1. Yan, H., Ruan, X., Chen, W.: The input voltage sharing control strategy for input-series and output-parallel converter under extreme conditions. In: Proceedings of IEEE Energy Conversion Congress and Exposition, pp. 662–667 (2009) 2. UCC3895 BiCMOS advanced phase-shift PWM controller datasheet, http://www.ti.com/lit/ds/ symlink/ucc3895.pdf

Index

A Active current, 125, 126 Active power, 4, 21, 23, 124, 126, 131, 133, 150, 156, 157 Aeronautical static inverter, 16, 175 Alternating current (AC), 1, 107 Auxiliary circuit, 197, 200, 203, 208 Average current control, 12, 18, 24, 175, 180 B Bipolar junction transistor, 2 Bipolar modulation, 177 BJT, see Bipolar junction transistor Body diode, 56 Boost-derived converter, 116 Buck/boost-derived converter, 37 Buck-derived converter, 37, 40, 41 Bus voltage, 130–132, 164, 177 C Central control unit, 175 Centralized control, 16, 17, 24, 122, 148–151, 169, 171, 172, 175 Changeable weighted values, 195–197, 200, 203, 208 Circulating current, 19, 60, 126 Common current reference, 118, 171, 181, 183, 188 Common output voltage control loop, 10, 11, 67 Compensator, 71–73, 136, 137, 163, 164 Compound balanced control strategy, 21, 22, 113–115, 118, 119, 121, 124, 125, 127, 141, 154, 156, 164–166, 168, 172 Constant power source, 35, 114

Constituent inverter modules, 17, 19, 124 Controlled sources, 60 Converter circuit, 2 Corner frequency, 71 Cross-feedback control, 12, 22, 23 Cross-feedback for output inductor current, 22 Crossover frequency, 71–73 Crystal oscillator, 17 Current closed-loop controlled, 8 cross-feedback control strategy, 12, 14 deviation, 17 follower, 158, 178, 183 hysteresis, 158, 162, 176, 177, 187 imbalance, 193, 208 reference, 8, 12, 15, 18–21, 23, 76, 79, 117, 124, 150, 154, 161, 171, 172, 176, 181, 183, 195, 196, 200 ripple, 57, 60, 130, 177 sharing, 7, 8, 17–19, 51, 116, 118, 119, 175, 181, 188 sharing reference, 41 sharing regulating speed, 18 Current-limiting, 19, 24, 176, 178, 180, 183, 185, 188, 193, 197, 200, 202–204, 207, 208 Current sharing bus, 8 D DC, see Direct current DC–AC inverter, 7, 15, 16, 24, 25, 107, 108, 110, 114, 122, 153, 154 Decoupling relationship, 122, 151 Deﬁcit current, 36 Democratic current sharing, 7–9

© Springer Nature Singapore Pte Ltd. and Science Press, Beijing, China 2019 X. Ruan et al., Control of Series-Parallel Conversion Systems, CPSS Power Electronics Series, https://doi.org/10.1007/978-981-13-2760-5

209

210 Device, 3–5, 123, 132, 193 Digital signal processor, 149 Direct current, 110, 111 Discharging current, 101 Distributed control, 16, 18, 24, 148–151, 154, 169, 171, 172, 175, 180, 183, 188 Drive circuit, 3–5 signal, 15, 127, 195, 200, 203 DSP, see Digital signal processor Duty cycle, 9–13, 40–42, 45–47, 49, 56, 58, 59, 67, 74, 82, 85–87, 95, 191, 197 Dynamic response, 21, 47, 49, 51, 55, 74, 75, 82, 84 E Electric vehicle, 2 Electromagnetic compatibility, 3 Equilibrium point, 19, 37, 114 voltage, 36 Equivalent negative resistance, 37 Equivalent resistance, 79, 194, 197 Excessive current, 36 Extreme load condition, 191–195, 197, 198, 203, 208 F Fault tolerance, 8 Feedback signal, 11, 23, 79, 92, 117, 118, 124, 150, 154, 156, 157, 170 Filter capacitor, 15, 20, 22, 23, 30, 35, 36, 46, 56, 70, 94, 108, 114, 118, 125, 126, 130, 131, 138, 139, 146, 151, 154, 155, 157, 164, 166, 167, 170, 172, 177, 186 Filter inductor, 12, 15, 18, 20, 46, 56–59, 70, 76, 79, 80, 94, 100, 118, 124–126, 131, 132, 138, 139, 141, 142, 145, 150, 151, 154, 156, 157, 161, 164, 166, 167, 170, 172, 176, 177, 183 Freewheeling, 56, 176, 177 Full-bridge, 15, 40, 46, 55, 56, 123, 127, 130–132, 154, 161, 164, 176, 177, 198 Full-controlled device, 2 Full load, 47, 49, 51, 74, 97, 99–103, 164, 166, 197, 203 Full resistive load, 131, 141, 143, 147, 164, 166, 185, 186 Fundamental angular frequency, 126 Fuzzy control, 3

Index G Gallium nitride, 3 Galvanic isolation, 122 GaN, see Gallium nitride Gate turn-off thyristor, 2 Genetic control, 3 Gradient gain, 87, 91, 92, 94, 97 GTO, see Gate turn-off thyristor H Half-bridge, 40, 123 Half-controlled device, 2 Half load, 47, 49, 51, 99, 100, 103 Heavy load, 195 High voltage integrated circuit, 3 HIVC, see High voltage integrated circuit Hot-swap, 102, 175, 183, 187, 188 Hysteresis width, 176 I ICS, see Input current sharing IGBT, see Insulated gate bipolar transistor Inductive load, 141, 143, 144, 164, 165, 168 Inductor-current cross-feedback control, 22, 23 Inductor current feedback hysteretic control, 21 Inner current control loop, 11, 12 Inner output ﬁlter capacitor current loop, 22 Input current perturbation, 78, 80 current sharing, 9, 29, 107, 110, 119, 175 impedance, 3, 37–39, 77–79, 84, 89, 136 power, 31–33, 35, 79, 81, 87, 96, 108, 114, 119, 124, 128, 135, 156, 161, 192, 193 side, 5, 30, 34–37, 46, 48, 50, 61, 78, 114, 115, 122, 133, 153, 164, 197 terminal, 38, 136, 171 voltage, 6, 11, 12, 15, 21–23, 29, 30, 32–34, 36, 38–42, 46–51, 56, 58, 60, 61, 67–69, 74, 79, 81, 82, 85–87, 89, 91, 95, 97–99, 101, 102, 105, 108, 110, 111, 113, 117–119, 121, 123, 124, 128, 131, 136–139, 141, 143, 150, 153, 154, 158, 161, 164, 166, 169, 183, 192–195, 197, 198, 200, 203, 204 voltage perturbations, 38, 80, 97, 133 voltage sharing, 9, 11, 21–23, 29, 107, 121, 153 voltage sharing error, 91, 93 voltage sharing regulator, 21 Input-dividing capacitor voltage, 124

Index Input-parallel-connected systems, 30, 35, 36, 39, 41, 42, 51, 113, 115, 119 Input-parallel-output-parallel, 6, 107, 175 Input-parallel-output-series, 6, 9, 11, 30, 32, 39, 46, 48, 85, 107, 111, 113, 117, 119 Input-series-connected systems, 35, 36, 38–42, 51, 85, 86, 90, 97, 105, 113, 115, 119, 121 Input-series-output-parallel, 6, 107 Input-series-output-series, 6, 13, 22, 24, 30, 33, 34, 36–39, 42, 43, 46, 50, 52, 85, 92, 94, 102, 107, 112, 113, 115, 116, 118, 119, 153, 154, 156, 159, 161, 164–166, 172 Insulated gate bipolar transistor (IGBT), 3 Intelligent power module (IPM), 5 Interconnected control, 16 Interleaved switching, 13 Intrinsic capacitors, 56 IPM, see Intelligent power module IPOP, see Input-parallel-output-parallel IPOS, see Input-parallel-output-series ISOP, see Input-series-output-parallel ISOS, see Input-series-output-series IVS, see Input voltage sharing IVSR, see Input voltages sharing regulator L Light load, 191, 193–195, 197 Linearization, 40 Line-frequency transformer, 122 Load current, 7, 8, 19, 34, 74, 113, 135, 141, 157, 164, 167, 176, 178–180, 183, 185–187 current feed-forward control, 178, 180 regulation, 7 sharing, 100 M Magnetic core, 5, 123 Master–slave current sharing, 7, 8 Mercury arc rectiﬁer, 1 Microelectronic, 2, 4 Minority-carrier device, 3 Modularization architecture, 30, 43, 45, 51, 85, 86 Module efﬁciencies, 32–34, 81 Mutual cross-feedback, 13 N Natural rebalancing ability, 13 Negative resistance, 38, 39, 77, 81, 84, 133, 136 Negative resistance characteristics, 84

211 Neural network, 3 Neural network control, 3 O OCS, see Output current sharing Output characteristics, 19, 24, 175, 178, 180, 183, 184, 188 current feedback control, 126, 127, 129, 138, 141, 145 current sharing, 11, 16, 17, 19–21, 24, 29, 46, 97, 107, 110, 121, 175, 181 voltage, 6, 7, 9–13, 15, 17–24, 29, 30, 32–35, 40–42, 44–47, 49, 51, 55, 65, 69–71, 74, 82, 84–87, 89–92, 94, 97–99, 102, 103, 105, 108–111, 113, 114, 116–119, 122–124, 126–128, 131–133, 136–139, 141, 143, 149, 151, 153–155, 158, 161, 162, 164, 167, 169, 171, 172, 176–181, 183, 187, 196, 198 voltage control loop regulator, 86 voltage/current sharing control, 119 voltage regulator, 8–11, 18, 20–23, 42, 67, 127, 128, 132, 150, 158, 162, 170 voltage sharing, 20, 22–24, 29, 153 Output-voltage cross-feedback control, 23, 25 OVR, see Output voltage regulator OVS, see Output voltage sharing P Parallel control unit, 17 Parasitic resistance, 97 Peak-to-peak voltage, 70 Perturbation, 34–36, 38–42, 58, 78, 80, 81, 87, 95–97, 113, 114, 124, 133, 135 PFC, see Power factor correction PFM, see Pulse-frequency modulation Phase lag, 163 Phase-leg mid-points, 124 Phase-locked loop, 17 Phase-shift control, 123 Phase-shifted full-bridge converter, 123 Phase synchronization, 156, 158, 172 Phasotron Photovoltaic system, 6, 9 PI, see Proportional-integral PLL, see Phase-locked loop Plugged into, 186 Plugged out, 186 Positive feedback loop, 37 Positive gradient regulation characteristic, 87 Positive resistance, 77, 79, 80, 84, 89, 136, 164 Positive zero-crossing point, 181

212 Power balance, 25, 107, 116, 150, 157 Power electronics building block, 4 Power electronics system integration, 3, 4, 25 modularization, 3, 25 standardization, 3, 25 Power factor angles, 22, 109, 155, 172, 179 correction, 154 Power integrated device, 3 Power switch, 11, 15, 56, 94, 121, 123, 176, 186, 198, 200 Primary-and-secondary-windings ratio, 56 Primary current, 56–59, 132 Primary duty cycle, 56, 58, 60 Proportional coefﬁcient, 86 Proportional–integral, 71 Protection circuit, 3–5 PSFB, see Phase-shifted full-bridge converter Pulse-frequency modulation, 3 Pulse-width modulation, 3 Push-pull, 123 PWM comparator, 42 PWM, see Pulse width modulation R Rated output current, 46, 48, 49, 51, 97, 196 Reactive current, 125, 126, 146 Reactive power, 19, 107, 108, 110–115, 118, 126, 139, 153, 157, 165, 167 Real power, 108, 110–115, 118, 171 Rebalancing ability, 14 Rectiﬁer diodes, 46, 56, 94 Redundancy, 7, 19, 100, 102, 105, 149, 175 Regulator, 9–12, 15, 18, 20, 41, 71, 72, 76, 84, 86, 124, 150, 158, 171, 178, 181, 183, 195, 196, 200 Resonant inductance, 46, 198 Right-half-plane pole, 39 Ripple current, 130 Rising edge, 181 RMS, see Root-mean-square Root-mean-square, 109 S Sampled output voltage Sampling clock, 176, 177 factor, 45 frequency, 162 signal, 15 Saturation voltage, 3 Sawtooth carrier, 127, 131

Index signal, 10, 11, 67, 70, 86, 96 SCR, see Silicon controlled rectiﬁer Secondary rectiﬁed voltage, 58 Secondary side, 56 Second stage, 123, 130 Semiconductor devices, 1–4 Series–parallel combined systems, 15 Series–parallel power conversion system, 5–7, 29, 30, 42, 45, 51, 55, 67, 85, 86, 107–109, 113–119 Short-circuit, 193, 197, 202–205, 208 Short fault, 101 Short switch, 101, 102 SiC, see Silicon carbide Silicon carbide, 3 Silicon controlled rectiﬁer, 1 Single-phase full-bridge inverter, 15, 16 Sinusoidal pulse-width modulator, 18 Smart power integrated circuit, 3 Small-signal model, 15, 37–39, 55, 58, 60, 61, 72, 74, 78, 80, 83, 94, 133 Soft-switching, 123 SPIC, see Smart power integrated circuit SPWM, see Sinusoidal pulse-width modulator Static error, 178 Stepped input voltage, 46, 49, 51, 82, 102 Stepped load, 46, 49, 51, 82, 99, 102, 185 Switching frequency, 57, 58, 70, 98, 130, 131, 162, 178, 198 Switching loss, 14, 123 Synchronization of the output current phase, 124 Synchronization pulse signal, 17 Synchronizing bus, 181 T Three control loops, 21, 22, 124, 154, 168 Three-dimensional package technology, 5 Three-loop control strategy, 20, 80, 82, 116 Three-state hysteresis modulation, 158, 176, 177, 183 Three voltage level, 123 Threshold current, 195, 198 Thyratron, 1 Total harmonic distortion, 21 Total load current, 17 Turning frequency, 162 Turns ratio, 9, 40, 61, 70, 95, 131 Two-stage inverter, 123, 131 Two-transistor forward converter, 94, 97 U Unbalanced current, 37 Uncompensated loop gain, 70–72, 162, 163

Index Uninterruptible power supply, 16 Unipolar modulation, 177 UPS, see Uninterruptible power supply V Voltage amplitude, 17, 23, 114 disturbance, 12, 136, 147 diverge, 136 Voltage-controlled device, 2 Voltage–current double closed-loop, 8, 15, 75, 76, 80, 84

213 W Wire connection, 7 Wireless droop control, 16, 19, 20, 175, 181 Wireless IVS control strategy, 15, 86, 89, 92–94, 97, 100 Z Zero-crossing phase-lock, 181 Zero state, 176

Xinbo Ruan · Wu Chen Tianzhi Fang · Kai Zhuang Tao Zhang · Hong Yan

Control of Series-Parallel Conversion Systems

CPSS Power Electronics Series Series editors Wei Chen, Fuzhou University, Fuzhou, Fujian, China Yongzheng Chen, Liaoning University of Technology, Jinzhou, Liaoning, China Xiangning He, Zhejiang University, Hangzhou, Zhejiang, China Yongdong Li, Tsinghua University, Beijing, China Jingjun Liu, Xi’an Jiaotong University, Xi’an, Shaanxi, China An Luo, Hunan University, Changsha, Hunan, China Xikui Ma, Xi’an Jiaotong University, Xi’an, Shaanxi, China Xinbo Ruan, Nanjing University of Aeronautics and Astronautics, Nanjing, Jiangsu, China Kuang Shen, Zhejiang University, Hangzhou, Zhejiang, China Dianguo Xu, Harbin Institute of Technology, Harbin, Heilongjiang, China Jianping Xu, Xinan Jiaotong University, Chengdu, Sichuan, China Mark Dehong Xu, Zhejiang University, Hangzhou, Zhejiang, China Xiaoming Zha, Wuhan University, Wuhan, Hubei, China Bo Zhang, South China University of Technology, Guangzhou, Guangdong, China Lei Zhang, China Power Supply Society, Tianjin, China Xin Zhang, Hefei University of Technology, Hefei, Anhui, China Zhengming Zhao, Tsinghua University, Beijing, China Qionglin Zheng, Beijing Jiaotong University, Beijing, China Luowei Zhou, Chongqing University, Chongqing, China

This series comprises advanced textbooks, research monographs, professional books, and reference works covering different aspects of power electronics, such as Variable Frequency Power Supply, DC Power Supply, Magnetic Technology, New Energy Power Conversion, Electromagnetic Compatibility as well as Wireless Power Transfer Technology and Equipment. The series features leading Chinese scholars and researchers and publishes authored books as well as edited compilations. It aims to provide critical reviews of important subjects in the ﬁeld, publish new discoveries and signiﬁcant progress that has been made in development of applications and the advancement of principles, theories and designs, and report cutting-edge research and relevant technologies. The CPSS Power Electronics series has an editorial board with members from the China Power Supply Society and a consulting editor from Springer. Readership: Research scientists in universities, research institutions and the industry, graduate students, and senior undergraduates.

More information about this series at http://www.springer.com/series/15422

Xinbo Ruan Wu Chen Tianzhi Fang Kai Zhuang Tao Zhang Hong Yan •

•

•

Control of Series-Parallel Conversion Systems

123

Xinbo Ruan College of Automation Engineering Nanjing University of Aeronautics and Astronautics Nanjing, Jiangsu, China Wu Chen School of Electrical Engineering Southeast University Nanjing, Jiangsu, China Tianzhi Fang College of Automation Engineering Nanjing University of Aeronautics and Astronautics Nanjing, Jiangsu, China

Kai Zhuang SDAAC Automotive Air-Conditioning System Co., Ltd. Shanghai, China Tao Zhang Flextronics R&D (ShenZhen) Co., Ltd. Shenzhen, Guangdong, China Hong Yan School of Mechanical and Electronic Engineering Bengbu University Bengbu, Anhui, China

ISSN 2520-8853 ISSN 2520-8861 (electronic) CPSS Power Electronics Series ISBN 978-981-13-2759-9 ISBN 978-981-13-2760-5 (eBook) https://doi.org/10.1007/978-981-13-2760-5 Jointly published with Science Press, Beijing, China The print edition is not for sale in China Mainland. Customers from China Mainland please order the print book from: Science Press. Library of Congress Control Number: 2018957246 © Springer Nature Singapore Pte Ltd. and Science Press, Beijing, China 2019 This work is subject to copyright. All rights are reserved by the Publishers, whether the whole or part of the material is concerned, speciﬁcally the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microﬁlms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a speciﬁc statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publishers, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publishers nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publishers remains neutral with regard to jurisdictional claims in published maps and institutional afﬁliations. This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd. The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721, Singapore

Preface

In the past three decades, power electronics technology has been rapidly developing and has found wide applications in almost all kinds of ﬁelds of industrial and agricultural production and daily life. Recently, the development and utilization of renewable energy for addressing the traditional fossil energy shortage and environmental pollution has been received more and more attention, which provides new opportunities for power electronics technology. Power electronics technology has been playing an increasingly important role in renewable energy generation, flexible AC/DC transmission, smart grid, and electric vehicle. At present, most of power electronics products are composed of various discrete devices. Inspired by the success of the development of microelectronics technology, the concept of power electronics system integration is emerging for the purpose of reducing the manufacture cost and improving the reliability. Generally, the power electronics system integration can be divided into three levels. The ﬁrst one is component-level integration including active devices integration and passive components integration. The second one is module-level integration, in which a variety of integrated active devices and/or passive components are integrated together to form highly versatile standardized converter modules. The last one is system-level integration, in which a number of converter modules are put together to form different kinds of power conversion systems. In the system-level integration, series–parallel power conversion system, in which multiple standardized converter modules are connected in series or parallel at the input and output sides, is one of the typical forms. Series–parallel power conversion systems have the following advantages, including reduced design difﬁculties, ease of capacity expansion, redundancy, and improved reliability. Series– parallel power conversion systems can be categorized into four possible architectures on the basis of the connection forms, namely, input-parallel-output-parallel (IPOP), input-parallel-output-series (IPOS), input-series-output-parallel (ISOP), and input-series-output-series (ISOS), and each type of series–parallel power conversion system has its own speciﬁc application areas. The IPOP system is suitable for the applications with high output current, the IPOS is suitable for the applications with low input voltage and high output voltage, the ISOP system is suitable for the v

vi

Preface

applications with high input voltage and high output current, and the ISOS system is attractive for the applications with high input voltage and high output voltage. This book focuses on the control techniques for the series–parallel power conversion systems with DC–DC converters and DC–AC inverters as the basic modules, respectively, to achieve equilibrium among the constituent modules. The detailed theoretical analysis with design examples and experimental validations are presented. This book includes nine chapters. In Chap. 1, the classiﬁcation and characteristics of the series–parallel power conversion systems are presented, and the control strategies for the series–parallel power conversion systems with DC–DC converters and DC–AC inverters as the basic modules, respectively, are briefly reviewed. Chapters 2–4 focus on the control strategies for the DC–DC series–parallel power conversion systems. In Chap. 2, the inherent relationship between the sharing of input voltages/ currents and the sharing of output voltages/currents among the constituent modules in the four types of DC–DC series–parallel power conversion systems is analyzed systematically from the view of power conservation, and the stability of various possible control strategies is discussed. Based on the analysis, a general control strategy is proposed for the DC–DC series–parallel power conversion systems, which not only achieve equilibrium among the constituent modules, but also decouples the output voltage control loop and the sharing control loop. In Chap. 3, the small-signal model of ISOP system consisting of multiple phase-shifted full-bridge converters is derived to prove the decoupling of the system output voltage control loop and input voltage sharing (IVS) control loop, and the parameters design of the two control loops is given. Furthermore, in order to improve the system dynamic response and realize output current short-circuit protection, the inner current loop is incorporated into the general control strategy, and the parameters design of the IVS control loop is analyzed. In Chap. 4, the wireless IVS control strategies for the input-series-connected power conversion systems based on positive output voltage gradient method are presented, which increases the output voltage of each module with the increase of module input voltage. With this control strategy, there is no control interconnection among the constituent modules leading to a truly modular design and high system reliability. Chapters 5–8 are dedicated to the control strategies for the DC–AC series– parallel power conversion systems. In Chap. 5, the inherent relationship between the input voltage/current sharing (IVS/ICS) and output voltage/current sharing (OVS/OCS) of the four kinds of DC– AC series–parallel power conversion systems is analyzed systematically, and the control strategies to achieve IVS/ICS and OVS/OCS and the system stability are discussed. A general control strategy for the four kinds of DC–AC series–parallel power conversion system is proposed.

Preface

vii

Chapters 6 and 7 present the implementation of the compound balanced control strategies for the DC–AC ISOP system and DC–AC ISOS system, respectively, which control the module input voltages to be equal for achieving IVS and control the magnitudes or phases of modular output currents (for ISOP system) or output voltages (for ISOS system) to be identical for achieving OCS/OVS. In Chap. 8, the load current feed-forward scheme is incorporated for the inverter module in the DC–AC IPOP system, and the load current feed-forward node is intentionally located before the current-limiting unit, for not only making the output voltage of the inverter independent of load current but also limiting the load current. In Chap. 9, the control strategy with changeable weighted values of the IVS loops and the IVS auxiliary circuit are presented for the input-series-connected power conversion system for achieving IVS under extreme load conditions, including light-load, no-load, and short-circuit current-limiting. This book is essential and valuable reference for graduate students and academics majoring in power electronics and engineers engaged in developing DC–DC converters, DC–AC inverters, series–parallel power conversion systems and power electronics transformers. Senior undergraduate students majoring in electrical engineering and automation engineering would also ﬁnd this book useful. Nanjing, China Nanjing, China Nanjing, China Shanghai, China Zhuhai, China Bengbu, China

Xinbo Ruan Wu Chen Tianzhi Fang Kai Zhuang Tao Zhang Hong Yan

Acknowledgements

This research monograph systematically summarizes the research work on the control techniques for series–parallel power conversion systems since 2003. The work in this book was partially supported by the Delta Power Electronics Science and Education Development Foundation under Award DREO2006006. We would like to express our sincere thanks to the Executive Committee of Delta Power Electronics Science and Education Development Foundation. It has been a great pleasure to work with the colleagues of Springer, Science Press, China, and China Power Supply Society (CPSS). The support and help from Mr. Wayne Hu (the Project Editor) are greatly appreciated. June 2018

ix

Contents

. . . . . . . . . . . . . . . . . . .

1 1 4 5 5 5 7 7 9 11 13 15 15 16 20 21 22 25 26

........

29

. . . . .

30 32 32 32 33

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1 Development Trend of Power Electronics . . . . . . . . . . . . . . . 1.2 Power Electronics System Integration . . . . . . . . . . . . . . . . . . 1.2.1 Component-Level Integration . . . . . . . . . . . . . . . . . . . 1.2.2 Module-Level Integration . . . . . . . . . . . . . . . . . . . . . . 1.2.3 System-Level Integration . . . . . . . . . . . . . . . . . . . . . . 1.3 DC–DC Series–Parallel Power Conversion Systems . . . . . . . . 1.3.1 IPOP System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3.2 IPOS System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3.3 ISOP System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3.4 ISOS System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4 DC–AC Inverter Series–Parallel Power Conversion Systems . . . . 1.4.1 DC–AC Inverter . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4.2 IPOP DC–AC Inverter System . . . . . . . . . . . . . . . . . . 1.4.3 IPOS DC–AC Inverter System . . . . . . . . . . . . . . . . . . 1.4.4 ISOP DC–AC Inverter System . . . . . . . . . . . . . . . . . . 1.4.5 ISOS DC–AC Inverter System . . . . . . . . . . . . . . . . . . 1.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 A General Control Strategy for DC–DC Series–Parallel Power Conversion Systems . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Possible Control Strategies for DC–DC Series–Parallel Power Conversion Systems . . . . . . . . . . . . . . . . . . . . . 2.1.1 IPOP Systems . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.2 IPOS Systems . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.3 ISOP Systems . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.4 ISOS Systems . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . . . . . . . . . . . . . . . .

. . . . .

. . . . . . . . . . . . . . . . . . .

. . . . .

. . . . .

xi

xii

Contents

2.2 Stability of the Control Strategies . . . . . . . . . . . . . . . . . . . . . . 2.2.1 Controlling at Input Side . . . . . . . . . . . . . . . . . . . . . . . 2.2.2 Controlling at Output Side . . . . . . . . . . . . . . . . . . . . . . 2.3 General Control Strategy for DC–DC Series–Parallel Power Conversion Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4 Modularization Architecture for DC–DC Series–Parallel Power Conversion Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5 Experimental Veriﬁcation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Mathematical Model and Closed-Loop Parameters Design for DC–DC ISOP System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 Small-Signal Model of PSFB-Based ISOP System . . . . . . . . . 3.1.1 Small-Signal Model of PSFB Converter . . . . . . . . . . . 3.1.2 Small-Signal Model of the PSFB-Based ISOP System . 3.2 Decoupling the Control Loops of the PSFB-Based ISOP System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3 Closed-Loop Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1 Output Voltage Regulator . . . . . . . . . . . . . . . . . . . . . . 3.3.2 IVS Regulator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4 Experimental Veriﬁcation of the General Control Strategy . . . 3.5 Three-Loop Control Strategy for ISOP System . . . . . . . . . . . . 3.5.1 Positive Resistance . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5.2 IVS Regulator Design . . . . . . . . . . . . . . . . . . . . . . . . 3.6 Experimental Veriﬁcation of the Three-Loop Control Strategy 3.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

.. .. ..

34 35 36

..

40

. . . .

. . . .

43 46 51 53

. . . .

. . . .

. . . .

55 56 56 61

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

67 69 70 72 74 74 76 80 82 83 84

..

85

..

86

4 Wireless IVS Control Strategies for Input-Series-Connected Systems Based on Positive Output Voltage Gradient Method . . . . 4.1 Genesis of the Wireless IVS Control Strategies Based on Positive Output Voltage Gradient Method . . . . . . . . . . . . . . . . . . . . . . . 4.2 Basic Features of the Input-Series-Connected Systems with Wireless IVS Control Strategies . . . . . . . . . . . . . . . . . . . . . . . 4.2.1 ISOP System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.2 ISOS System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3 System Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4 Experimental Veriﬁcation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4.1 Experimental Results of ISOP System . . . . . . . . . . . . . . 4.4.2 Experimental Results of ISOS System . . . . . . . . . . . . . . 4.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . .

. 90 . 90 . 92 . 94 . 97 . 97 . 102 . 105 . 105

Contents

xiii

5 A General Control Strategy for DC–AC Series–Parallel Power Conversion Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1 Relationship Between Input Voltage/Current Sharing and Output Voltage/Current Sharing . . . . . . . . . . . . . . . . . . . 5.1.1 IPOP Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1.2 IPOS Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1.3 ISOP Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1.4 ISOS Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2 Stability of the Control Strategies for DC–AC Series–Parallel Power Conversion Systems . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2.1 Compound Balanced Control Strategy . . . . . . . . . . . . . 5.2.2 Output Voltage/Current Sharing (OVS/OCS) Control Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3 General Voltage/Current Sharing Control Strategy for DC–AC Series–Parallel Power Conversion Systems . . . . . . . . . . . . . . . 5.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . 107 . . . . .

. . . . .

. . . . .

108 110 111 111 112

. . . 113 . . . 114 . . . 115 . . . 115 . . . 118 . . . 119

6 Compound Balanced Control Strategy for DC–AC Input-Series Output-Parallel Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.1 DC–AC Inverter Topology for the ISOP Systems . . . . . . . . . . . 6.2 Implementation of the Compound Balanced Control Strategy . . 6.3 Closed-Loops Design for DC–AC ISOP Systems . . . . . . . . . . . 6.3.1 Decoupling Relationship Between IVS Loop and Output Voltage Loop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3.2 Design of Control Loop of the First-Stage Full-Bridge Converter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3.3 Design of IVS Loop . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3.4 Design of Output Voltage Loop . . . . . . . . . . . . . . . . . . 6.4 Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.5 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.5.1 Steady-State Experiment . . . . . . . . . . . . . . . . . . . . . . . 6.5.2 Dynamic Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . 6.5.3 Comparison of the Output Current Feedback and Output Filter Inductor Current Feedback . . . . . . . . . . . . . . . . . 6.6 Distributed IVS/OCS Control Strategy . . . . . . . . . . . . . . . . . . . 6.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . .

. . . . . . .

130 133 137 138 141 141 141

. . . .

. . . .

145 148 151 151

7 Compound Balanced Control Strategy for DC–AC Input-Series-Output-Series Systems . . . . . . . . . . . . . . . . . . . . . 7.1 Implementation of the Compound Balanced Control Strategy 7.2 Decoupling the Control Loops . . . . . . . . . . . . . . . . . . . . . . . 7.3 Design of the Control Loops for the DC–AC ISOS Systems .

. . . .

. . . .

153 154 158 161

. . . .

. . . .

. . . .

. . . .

121 122 124 127

. . 127

xiv

Contents

7.3.1 Design of the Output Voltage Loop . . . . . . . . . . . . . . . 7.3.2 Design of the Control Loop of the First-Stage Converter and IVS Loop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.4 Simulation and Experimental Veriﬁcation . . . . . . . . . . . . . . . . 7.4.1 Simulation Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.4.2 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . 7.5 Distributed Control Strategy for DC–AC ISOS Systems . . . . . . 7.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 An Improved Average Current Control Strategy for DC–AC Input-Parallel-Output-Parallel Systems . . . . . . . . . . . . . . . . . . . . 8.1 Current Three-State Hysteresis Modulation . . . . . . . . . . . . . . . 8.2 Improvement of the Output Characteristics of Inverter . . . . . . 8.3 Conﬁguration and Control Scheme for DC–AC IPOP Systems . . 8.3.1 Conﬁguration of the DC–AC IPOP Systems . . . . . . . . 8.3.2 Control Scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.4 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . .

9 Input Voltage Sharing Control Strategy for ISOP Systems Under Extreme Load Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.1 Reason of Input Voltages Imbalance of ISOP Systems Under Extreme Load Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2 Input Voltage Sharing Control Strategy for ISOP Systems Under Extreme Load Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2.1 Changing the Weighted Values of the IVS Loops . . . . . 9.2.2 Input Voltage Sharing Auxiliary Circuit . . . . . . . . . . . . 9.3 Simulation and Experimental Veriﬁcation . . . . . . . . . . . . . . . . 9.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . 161 . . . . . . .

. . . . . . .

163 163 164 167 167 172 172

. . . . . . . . .

. . . . . . . . .

175 176 178 180 180 183 183 187 188

. . 191 . . 192 . . . . . .

. . . . . .

194 194 197 198 208 208

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209

About the Authors

Xinbo Ruan was born in Hubei Province, China, in 1970. He received B.S. and Ph.D. in electrical engineering from Nanjing University of Aeronautics and Astronautics (NUAA), Nanjing, China, in 1991 and 1996, respectively. In 1996, he joined the Faculty of Electrical Engineering Teaching and Research Division, NUAA, where he became a Professor in the College of Automation Engineering in 2002 and has been engaged in teaching and research in the ﬁeld of power electronics. From August to October 2007, he was a Research Fellow in the Department of Electronic and Information Engineering, Hong Kong Polytechnic University, Hong Kong, China. From March 2008 to August 2011, he was also with the School of Electrical and Electronic Engineering, Huazhong University of Science and Technology, China. He is a Guest Professor with Beijing Jiaotong University, Beijing, China, Hefei University of Technology, Hefei, China, and Wuhan University, Wuhan, China. He is the author or co-author of nine books and more than 300 technical papers published in journals and conferences. His main research interests include soft-switching DC–DC converters, soft-switching inverters, power factor correction converters, modeling the converters, power electronics system integration, and renewable energy generation system. Dr. Ruan was a recipient of the Delta Scholarship by the Delta Environment and Education Fund in 2003 and was a recipient of the Special Appointed Professor of the Chang Jiang Scholars Program by the Ministry of Education, China, in 2007. From 2005 to 2013, again from 2017, he served as Vice President of the China Power Supply Society. From 2014 to 2016, he served as Vice Chair of the Technical Committee on Renewable Energy Systems within the IEEE Industrial Electronics Society. Currently, he is an Associate Editor for the IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, IEEE TRANSACTIONS ON POWER ELECTRONICS, IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS-II, and the IEEE JOURNAL OF EMERGING AND SELECTED TOPICS ON POWER ELECTRONICS. He was elevated to IEEE Fellow in 2015. Wu Chen was born in Jiangsu Province, China, in 1981. He received B.S., M.S., and Ph.D. in electrical engineering from Nanjing University of Aeronautics and Astronautics (NUAA), Nanjing, China, in 2003, 2006, and 2009, respectively. xv

xvi

About the Authors

From 2009 to 2010, he was a Senior Research Assistant in the Department of Electronic Engineering, City University of Hong Kong, Kowloon, Hong Kong, China. From 2010 to 2011, he was a Postdoctoral Researcher in Future Electric Energy Delivery and Management Systems Center, North Carolina State University, Raleigh. Since September 2011, he has been an Associate Research Fellow with the School of Electrical Engineering, Southeast University, Nanjing, China, where he has been a Professor since 2016. His main research interests include soft-switching converters, power delivery, and power electronic system integration. Tianzhi Fang was born in Jiangsu Province, China, in 1977. He received B.S. in electrical engineering and automation from Nanjing University of Aeronautics and Astronautics (NUAA), Nanjing, China, in 2000, M.S. in control theory and engineering from Jiangsu University, Zhenjiang, China, in 2003, and Ph.D. in electrical engineering from NUAA, Nanjing, China, in 2009. In 2009, he joined the Faculty of the College of Automation Engineering, NUAA, where he is currently an Associate Professor. His main research interests include series–parallel inverters, grid-connected inverter, and power electronics system integration. Kai Zhuang was born in Shandong Province, China, in 1975. He received B.S in electrical engineering from Jinan University, Jinan, China, in 1998, M.S. in electrical engineering from Anhui University of Technology, Ma’anshan, China, in 2001, and Ph.D. in electrical engineering from Nanjing University of Aeronautics and Astronautics (NUAA), Nanjing, China, in 2008. He had worked in GE (China) research center, LG Electronics (Shanghai) Research Center, Shanghai Delphi Automotive Air-Conditioning System Co., Ltd. In Nov. 2017, he joined Valeo Powertrain (Shanghai) Co., Ltd, Shanghai, China, as hardware department manager. His current research interests include iBSG, DC– DC converter, onboard charger, and wireless power transfer system for electric vehicles. Tao Zhang was born in Hubei Province, China, in 1977. He received B.S. in industrial automation from Daqing Petroleum Institute, Daqing, China, in 2000, and M.S. in electrical engineering from Nanjing University of Aeronautics and Astronautics (NUAA), Nanjing, China, in 2005. In 2005, he joined Astec Power Co., Ltd. and in 2008 joined TDI Power Co., Ltd. for design and development of power products. At the beginning of 2011, he joined the Flex R&D (Shenzhen) Co., Ltd. for the research and development of new power technology. He joined Innoscience (Zhuhai) Technology Co., Ltd. in 2018, as AE director responsible for the application development of wide band gap semiconductor products GaN. He holds one US patent and two pended US patents.

About the Authors

xvii

Hong Yan was born in Hunan Province, China, in 1979. She received B.S. in automation from Hefei University of Technology, Hefei, China, in 2001, and M.S. in electrical engineering from Nanjing University of Aeronautics and Astronautics (NUAA), Nanjing, China, in 2009. Since April 2009, she joined the College of Electronic and Electrical Engineering at Bengbu University, where she is currently a Lecturer. Her main research interests include power electronics system integration and DC–DC converters.

Abbreviations

AC BJT CPES DC DSP GaN GTO HVICs ICS IGBT IPEM IPM IPOP IPOS ISOP ISOS IVS IVSR LSI MOSFET OCS OVR OVS PEBB PFC PFM PI PLL PSFB PWM

Alternating Current Bipolar Junction Transistor Center for Power Electronics Systems Direct Current Digital Signal Processor Gallium Nitride Gate Turn-off Thyristor High-Voltage Integrated Circuits Input Current Sharing Insulated Gate Bipolar Transistor Integrated Power Electronics Modules Intelligent Power Module Input-Parallel-Output-Parallel Input-Parallel-Output-Series Input-Series-Output-Parallel Input-Series-Output-Series Input Voltage Sharing Input Voltages Sharing Regulator Large-Scale Integrated Metal-Oxide-Semiconductor Field-Effect Transistor Output Current Sharing Output Voltage Regulator Output Voltage Sharing Power Electronics Building Block Power Factor Correction Pulse Frequency Modulation Proportional-Integral Phase-Locked Loop Phase-Shifted Full-Bridge Pulse-Width Modulation

xix

xx

RMS SCR SiC SoC SPICs SPWM ULSI UPS VLSI

Abbreviations

Root Mean Square Silicon Controlled Rectiﬁer Silicon Carbide System on Chip Smart Power Integrated Circuits Sinusoidal Pulse-Width Modulator Ultra-Large-Scale Integrated Uninterruptible Power Supply Very-Large-Scale Integration

Chapter 1

Introduction

Abstract Power electronics system integration is one of the effective approaches to reduce the cost of power electronics products, shorten the development cycle, and improve the reliability. Generally, power electronics system integration can be divided into three levels, namely, component-level integration, module-level integration, and system-level integration. One important branch of system-level integration is the series–parallel power conversion systems, in which multiple standardized converter modules are connected in series or parallel at the output and input sides to meet the requirements of different applications. In order to ensure the proper operation of series–parallel power conversion systems, proper sharing of the input/output voltage and proper sharing of the input/output current must be ensured. This chapter mainly introduces the control techniques for the series–parallel power conversion systems with DC–DC converter and DC–AC inverter as the basic module, respectively, which will lay a foundation for the narration of the following chapters. Keywords Power electronics system integration Series–parallel power conversion system · Input-parallel-output-parallel (IPOP) Input-parallel-output-series (IPOS) · Input-series-output-parallel (ISOP) Input-series-output-series (ISOS)

1.1 Development Trend of Power Electronics Power electronics started with the development of the mercury arc rectifier, which was invented by Peter Cooper Hewitt in 1902 and used for converting alternating current (AC) into direct current (DC) [1]. Other types of rectifiers, such as phanotrons and thyratrons, were gradually introduced in the 1930s. In 1956, the silicon controlled rectifier (SCR) was invented by Bell Laboratories and commercially introduced by General Electric Company [2], and after that, power electronics developed rapidly with the improved switching speed of semiconductor devices. A pioneer of power electronics, Dr. William E. Newell, pointed out that power electronics was interstitial to all three of the major disciplines of electrical engineering: electronics, power, © Springer Nature Singapore Pte Ltd. and Science Press, Beijing, China 2019 X. Ruan et al., Control of Series-Parallel Conversion Systems, CPSS Power Electronics Series, https://doi.org/10.1007/978-981-13-2760-5_1

1

2

1 Introduction

and control at the IEEE Power Electronics Specialists Conference (PESC) in 1973 [3]. In 1989, Dr. Bimal K. Bose expanded and refined the circumscription of power electronics, and demonstrated that power electronics was an interdisciplinary technology, including computer-aided design, power semiconductor devices, converter circuits, electrical machines, very large scale integration (VLSI) circuits, control theory, microcomputers, analog and digital electronics, and so on, as shown in Fig. 1.1 [4]. At present, more and more disciplines are involved in power electronics, such as modern control theory, microelectronics, material science, environmental science, and life science, and power electronics has become an independent interdisciplinary high-tech discipline. With the rapid development during the last nearly 60 years, power electronics technology has been widely used in almost all kinds of fields of industrial and agricultural production and daily life. Recently, more attention has been drawn to the renewable energy development and utilization globally due to the gap between the increasingly growth of energy demand and shortage of the traditional fossil energy, which also provides new opportunities for the development of power electronics. Power electronics technology will play an increasingly important role in renewable energy generation, flexible AC/DC transmission, electric vehicle and energy conservation, and environmental protection [5–8]. Power semiconductor devices, power conversion technology, and control technology are the three main fields involved in power electronics. The power semiconductor devices are the basis of the development of power electronics technology. Generally, the power semiconductor devices have gone through four stages of development: (1) Half-controlled devices: SCR is a typical representation of the half-controlled device, which has been widely used in low frequency, medium-to-high voltage AC power conversions. (2) Full-controlled devices: Gate turn-off thyristor (GTO), bipolar junction transistor (BJT), and metal-oxide-semiconductor field-effect transistor (MOSFET) are the main three kinds of full-controlled devices. The GTO and BJT are currentcontrolled devices, and the MOSFET is voltage-controlled device, which has higher switching speed than GTO and BJT.

Fig. 1.1 Power electronics—An interdisciplinary technology

Computer-aided design Analog and digital electronics Microcomputers

Power electronics and drives

Control theory

Power semiconductor devices Converter circuits Electrical machines

VLSI Circuits

1.1 Development Trend of Power Electronics

3

(3) Hybrid devices: Insulated gate bipolar transistor (IGBT) is a minority-carrier device, which integrates the advantages of the MOSFET and BJT, i.e., the high input impedance and high switching speed of a MOSFET and the low saturation voltage of a bipolar transistor. This hybrid combination is that the IGBT has the conduction characteristics of a bipolar transistor but is voltage-controlled like a MOSFET. (4) Power integrated devices: Smart power integrated circuits (SPICs) and high voltage integrated circuits (HVICs) are the main two kinds of power integrated devices, which integrate the power semiconductor devices, drive circuits, protection circuits, and even controllers. Recently, there has been significant progress in the development of wide bandgap semiconductor power devices, such as silicon carbide (SiC)-based and gallium nitride (GaN)-based power devices. SiC-based power devices have become strong competitors to Si-based power devices in medium power conversion applications, while GaN-based power devices can achieve ultra-fast switching. In summary, the power devices will be more large capacity, high frequency, modular, and intelligent in the future [9]. Power conversion technology can be classified into four types based on the forms of electric energy conversion: (1) AC-to-DC conversion, (2) DC-to-DC conversion, (3) DC-to-AC conversion, and (4) AC-to-AC conversion. Each electric energy conversion form can be achieved by various circuit topologies. The development of power conversion technology makes power electronics equipment have compact size, high conversion efficiency, and improved comprehensive economic benefit. Control technology is very important for power electronics technology. Various control technologies have been rapidly proposed and widely used, such as pulsewidth modulation (PWM) and pulse-frequency modulation (PFM). With the development of modern control theory, fuzzy control, genetic control, neural network control, and so on are introduced to the power electronics. With the growing applications of power electronics technology, it also faces various new challenges. At present, most power electronics products are specially designed according to different purposes and requirements of customers, i.e., customdesigned products. The design of power electronics products usually is a complex procedure with long development cycle time and high cost, since many aspects are involved, such as semiconductor devices, circuit topologies, control, magnetic material, package, mechanical structure, thermal design, and electromagnetic compatibility. Therefore, lots of repeated work exist when developing different power electronics products, resulting in lower labor efficiency and higher production cost. In order to reduce the cost and shorten the development cycle time, it is eager to improve the standardization of power electronics products for large-scale production to make the products can meet the rapidly changing market and customer needs, which is also one of major challenges for power electronics technology. The consensus among the world’s power electronics industry and academia is that the effective way to solve the problem is power electronics system integration, standardization and modularization [10].

4

1 Introduction

1.2 Power Electronics System Integration There are two main branches for the development of electronic technology after the appearance of semiconductor device. One is the microelectronic technology to process information, and the other is the power electronics technology to process power (energy). The development of microelectronic technology has experienced several stages, including discrete devices, large scale integrated (LSI) circuits, very large scale integrated (VLSI) circuits, ultra large scale integrated (ULSI) circuits, and system on chip (SoC), and now the integration is still the trend of microelectronic technology. Compared with the microelectronic technology, the development of power electronics integration is relatively slow. At present, most of power electronics products are composed of various discrete devices. Inspired by the success of the development of microelectronics technology, people will think that the cost of production will be reduced and the development cycle time will be shortened if the integration is applied in the power electronics, and this is the original intention of the proposal of power electronics system integration. The goal of power electronics system integration is to increase the integration levels in the components, including devices, circuits, controls, and sensors, which are integrated into standardized manufacturable modules, and then these standardized modules can be easily used to build various types of power electronics equipment. The idea of power electronics system integration was originally proposed by the US Navy, and the famous concept of power electronics building block (PEBB) was proposed with the vision of changing the paradigm for designing and manufacturing electrical high power conversion and control products [11]. After that, the Center for Power Electronics Systems (CPES) has been doing very well-known work on the PEBB, including the monolithic integration of high-voltage power devices and lowvoltage devices for controls and sensors, the interconnection of high-power devices and control devices on a common substrate, thermal management, and integration of passive high-power components, and CPES developed the integrated power electronics modules (IPEM), a standardized off-the-shelf module that has revolutionized power electronics [12]. Generally, the power electronics system integration can be divided into three levels. The first one is component-level integration, including active devices integration and passive components integration. The active power devices integration includes active power devices, drive circuits, protection circuits, and communication circuits, while the passive components integration includes inductors, capacitors, and transformers. The second one is module-level integration, in which a variety of integrated active devices and/or passive components are integrated together to form highly versatile standardized converter modules. The last one is system-level integration, in which a number of converter modules are put together to form different kinds of power conversion systems.

1.2 Power Electronics System Integration

5

1.2.1 Component-Level Integration The active devices integration refers to integrate active devices, drive circuits, control circuits, and protect circuits on a common substrate or into a module, leading to reduced volume and influence of the connection among devices. For instance, intelligent power module (IPM) is a high-function integrated module with IGBT chip and dedicated integrated circuits for self-protection (over-current, under-voltage, and over-heating). The passive components integration includes magnetic components integration (such as the integration of inductor and transformer, integration of inductors), integration of magnetic components and capacitors, and integration of magnetic components, capacitors, resistors, etc. Two or more discrete inductors and/or transformers can share a common magnetic core through proper coupling way and parameters design, leading to reduced magnetic components volume and total losses. The magnetic components and capacitive components can be integrated together, for instance, R. Reeves proposed an integration method for an inductor and a capacitor [13]. The resistors can also be integrated with magnetic components and capacitive components, in which the resistors are used for snubber circuit or control circuit [14].

1.2.2 Module-Level Integration Module-level integration refers to the integration of various active devices, passive components, control circuits, and protection circuits for a standardized module with strong versatility. CPES has proposed multi-chip packaging technology to achieve power electronics module-level integration. Kinds of power electronics devices in the form of bare dies are mounted on a ceramic or metal substrate, and the control circuits, drive circuits, protection circuits, snubber circuits, and sensors are also mounted on the same substrate in the form of IC bare dies or surface mount packages. All these devices and circuits are interconnected through thin-film power overlay technology and then connected with passive components through three-dimensional package technology to form a converter module to realize power conversion.

1.2.3 System-Level Integration System-level integration technology refers to integrate various standardized converter modules into systems to meet different power conversion demands. One important branch of system-level integration is series–parallel power conversion systems, in which multiple standardized converter modules are connected in series or parallel at the output and input sides to meet the demands of different applications [15]. For instance, the input sides of multiple converter modules can be connected in series for

6

1 Introduction

high input voltage applications, and the output sides of multiple converter modules can be connected in parallel for large output current applications. Series–parallel power conversion systems can be categorized into four possible architectures on the basis of the connection forms, namely, input-parallel-output-parallel (IPOP), inputparallel-output-series (IPOS), input-series-output-parallel (ISOP), and input-seriesoutput-series (ISOS), as shown schematically in Fig. 1.2. Each type of series–parallel power conversion system has its own specific application areas. With the rapid expansion of computer and telecommunication industries, distributed power systems consisting of paralleled modular converters delivering low output voltage and high output current are still the most popular practical configuration, where IPOP systems have been widely adopted. ISOP systems can be used in applications where the input voltage is relatively high and the output voltage is relatively low, such as in high-speed train power systems, industrial drives, and undersea observatories. For applications requiring high output voltages, such as X-ray equipment, photovoltaic systems, and electrostatic precipitator, IPOS systems become the desired choice. Moreover, ISOS systems are well suited for applications where both the input voltage and output voltage are high.

+ vin _

+ + vin1 Module 1 vo1 _ _

+ v_o

+ vin _

+ + vin1 Module 1 vo1 _ _

+ + v_in2 Module 2 vo2 _

+ + v_in2 Module 2 vo2 _

+ + v_inn Module n von _

+ + v_inn Module n von _

(a) IPOP

+

vin

_

+ + vin1 Module 1 vo1 _ _ + + v_in2 Module 2 vo2 _

+ + v_inn Module n von _ (c) ISOP

+

vo

_

(b) IPOS

+ v_o

+

vin

_

+ + vin1 Module 1 vo1 _ _ + + v_in2 Module 2 vo2 _

+ + v_inn Module n von _ (d) ISOS

Fig. 1.2 Four connection architectures of multiple converter modules

+

vo

_

1.2 Power Electronics System Integration

7

The advantages of the systems constructed from connecting multiple converter modules include (a) ease of thermal design as a result of each module handling only a part of the total power; (b) increased overall system reliability due to reduced thermal and electrical stresses on the power devices and components; (c) shortened design process and lowered cost for the system; (d) improved system reliability due to N + 1 redundancy; and (e) ease of expansion of power system capability [16]. According to the form of energy conversion, the modules in Fig. 1.2 can be DC–DC converter, DC–AC inverter, AC–DC rectifier, and AC–AC converter. This book focuses on the series–parallel power conversion systems with DC–DC converter and DC–AC inverter as the basic modules, respectively, and their recent development will be introduced as follows.

1.3 DC–DC Series–Parallel Power Conversion Systems 1.3.1 IPOP System Paralleling of multiple standardized converters has been extensively studied and widely applied in the power supplies for computers and communications. One basic objective of IPOP system is to share the load current among the constituent converters. To do this, a variety of approaches have been proposed, such as droop method, master–slave current sharing method, democratic current sharing method, etc. [17–20]. The droop method can be defined as one in which the output voltage droops as the load current is increased, as shown in Fig. 1.3. Its operation mechanism is to program the output impedance to achieve current sharing among converters. The droop method has the advantages of easy implementation, no wire connection among the control circuits of converters and high modularity and reliability. However, the current sharing and load regulation are relatively poor. In the master–slave current sharing method, one module is dedicated to be the master module and the remaining modules are the slave modules. A typical block

Io1 Ro1 V1 Converter 1

Io2

+ Vo RLd _

V1 V2 V Vo

Ro2 V2 Converter 2

(a) Equivalent circuit Fig. 1.3 Two-converter IPOP system

0

Io1

Io2

(b) Output characteristics

I

8

1 Introduction _ Vof + Voref

Gvo

Iref + _

Gi

Module 1 RS Flip-flop

VRAMP 1

ILf1_f

+ _

Comparator 1 − +

Gi

Module 2

Comparator 2 − +

RS Flip-flop

VRAMP 2

ILf2_f

Drive circuit 2

…

Currentsharing bus

Drive circuit 1

+ _ ILfn_f

Gi

Comparator n − +

Module n RS Flip-flop

Drive circuit n

VRAMP n

Fig. 1.4 A typical block diagram of master–slave current sharing method

diagram of master–slave current sharing method is shown in Fig. 1.4, where module 1 is the master module, which is voltage–current double closed-loop controlled, and modules 2 to n are the slave modules, which are only current closed-loop controlled. Gvo is the transfer function of the output voltage regulator, and Gi is the transfer function of the current regulator. The output of Gvo , I ref , is taken as the current sharing bus and becomes the common reference for all current closed loops to achieve current sharing. The limitations of the master–slave current sharing method are poor fault tolerance since the master failure will disable the entire system and degraded modularity of the system. Figure 1.5 shows a typical block diagram of the democratic current sharing method. As seen, the load current signal of each module is connected to the current sharing bus through an ideal diode, and the module with the highest output current is automatically selected to be the master module and its load current signal is taken as the current reference for all other modules. The current reference is then compared with the individual converter load current and the error signal is added to the voltage reference of each corresponding converter to regulate its load current to achieve current sharing. Obviously, in the democratic current sharing method, there is no dedicated master module, leading to good fault tolerance and modularity.

1.3 DC–DC Series–Parallel Power Conversion Systems _ io1 +

Gio

_ io2 +

Gio

Gvo

Voref + + _ vof

Gvo

Comparator 1 − +

Module 1 RS Flip-flop

VRAMP 1

Comparator 2 − +

Drive circuit 1 Module 2

RS Flip-flop

VRAMP 2

Drive circuit 2

…

Currentsharing bus

Voref + + _ vof

9

_ ion +

Gio

Voref + + _ vof

Gvo

Comparator n − +

Module n RS Flip-flop

VRAMP n

Drive circuit n

Fig. 1.5 A typical block diagram of democratic current sharing method

1.3.2 IPOS System IPOS system can be applied in the occasions requiring high output voltages or higher step-up ratios, such as X-ray equipment, photovoltaic systems, and electrostatic precipitator [21, 22]. The key issue of the IPOS system is to ensure output voltage sharing (OVS) among the constituent converters. Due to the output sides are connected in series, the constituent converters of IPOS system can regulate its own output voltages to be 1/n of the system output voltage (n is the number of converters) to share the output voltage. However, OVS is affected by the degree of matching among the sampling circuits and voltage references of constituent converters and it is difficult to obtain precise system output voltage. A simple common duty cycle control strategy was proposed in [23] for IPOS system; however, OVS cannot be ensured due to parameters mismatches such as turns ratio of individual power transformer. Ref. [24] proposed an OVS control strategy for IPOS system with a common system output voltage control loop and individual input current control loop for each module. The output of system output voltage regulator serves as the reference of input current control loop of each module to achieve input current sharing (ICS) and equal voltage regulation among the constituent modules, assuming that all the modules are identical. In [25], the module with maximum output voltage is regarded as the master module and its output voltage is taken as the reference, which is compared with the individual module output voltage and the error signal is added to the corresponding output of the output voltage regulator to regulate the module’s output voltage to achieve OVS. Hence, it can be seen that OVS of IPOS system can be achieved by both controlling ICS of modules and directly controlling OVS of modules. The relationship between ICS and OVS of IPOS system was revealed in [16]. It was pointed out that achieving ICS implies OVS and vice versa, and a novel OVS

10

1 Introduction Output voltage sharing loops Kvo

vo1

_ vo1f +

+

− + + VRAMP 1

Gvo_mod vo_EA1

Kvo

vo2

+

Comparator 1

_ vo2f

+

Gvo_mod

Comparator 2 − RS Flip-flop +

+ VRAMP 2

Kvo

_ +

vomod_ref

vo(n−1)f

vo

Kvo

Voref

vo_EA(n−1) … + + +

vof +

Comparator (n−1) − RS Flip-flop + + VRAMP (n−1) +

Gvo_mod

1/n

_ Gvo

Drive circuit 1

Drive circuit 2

…

…

vo_EA2

vo(n−1)

RS Flip-flop

vo_EA

_

Comparator n − RS Flip-flop +

+ VRAMP n

Drive circuit (n−1)

Drive circuit n

Output voltage loop

Fig. 1.6 OVS control strategy for IPOS system

control strategy was proposed, as shown in Fig. 1.6. It consists of a common output voltage control loop to regulate the system output voltage and n − 1 OVS control loops to achieve OVS for IPOS system. K vo is the sensing gain of the system output voltage and output voltage of each module. Gvo is the transfer function of the system output voltage regulator. Gvo_mod is the transfer function of OVS regulator. It can be seen that the reference voltage signal of each OVS loop is K vo vo /n, vo_EA is the control signal derived from the system output voltage regulator, and vo_EAj (j 1, 2,…, n − 1) is the control signal derived from each OVS regulator. For the first n − 1 modules, vo_EAj (j = 1, 2,…, n − 1) is added to vo_EA and then compared with the sawtooth signal to obtain the duty cycle signal of module j. For the module n, n−1 vo_EA j is subtracted from vo_EA and then compared with the sawtooth signal to j1

obtain the duty cycle signal of module n. With the control strategy, the duty cycles of modules with higher output voltages will decrease while the duty cycles of modules with lower output voltages will increase, and thus OVS is achieved. From Fig. 1.6, it can be found that the sum of the duty cycles caused by the OVS loops is zero, which means that the OVS control loops have no influence on the system output voltage regulation.

1.3 DC–DC Series–Parallel Power Conversion Systems

11

1.3.3 ISOP System ISOP system can be used in higher input voltage applications such as the electric railway system, where multiple converter modules with lower input voltage are adopted, which is beneficial for reducing the voltage stress of the power switches. In order to ensure the proper operation of ISOP system, the basic objective is to ensure input voltage sharing (IVS) and output current sharing (OCS) among the constituent modules. The common duty cycle control strategy can also be used for ISOP system [23]. With the common duty cycle for all modules, the input currents of the modules with higher input voltages will increase to reduce their input voltages, while the input currents of the modules with lower input voltages will decrease to boost their input voltages, and finally to achieve IVS. Similar to the IPOS system, IVS is also affected by the matching of modules’ parameters. A fault-tolerant democratic IVS control strategy for ISOP system was proposed in [26], as shown in Fig. 1.7. Each input voltage signal of module is connected to the voltage sharing bus through an idea diode, producing a common voltage reference signal, which is obtained from one of the converters with highest input voltage within the ISOP system. The voltage difference between the common voltage reference and each converter input voltage is amplified and added to the output of system output voltage regulator for minor control correction to ensure IVS. Gvcd is the transfer function of IVS loop regulator. If a converter fails with its input short-circuited, the common voltage reference is automatically increased to compensate for the loss of the failed converter, leading to high system reliability. An IVS control strategy was proposed for ISOP system in [27], as shown in Fig. 1.8. The control strategy consists of a common output voltage control loop for achieving desired system output voltage and n − 1 IVS control loops. V in /n is the reference signal for IVS control loops. vo_EA is the control signal derived from the system output voltage regulator, and vcd_EAj (j 1, 2,…, n − 1) is the control signal derived from each IVS regulator. For the first n − 1 modules, vcd_EAj (j 1, 2,…, n − 1) is subtracted from vo_EA and then compared with the sawtooth signal to obtain n−1 the duty cycle signal of module j. For the module n, vcd_EA j is added to vo_EA and j1

then compared with the sawtooth signal to obtain the duty cycle signal of module n. With the control strategy, the duty cycles of modules with high input voltages will increase and the duty cycles of modules with low input voltages will decrease, and finally to achieve IVS. It is proved that the IVS loops are decoupled from one another and are also decoupled from the system output voltage loop. Thus, the IVS regulator and system output voltage regulator can be designed independently. Based on the IVS control strategy shown in Fig. 1.8, an inner current control loop could be introduced for each module, for the purpose of improving the system dynamic performance. The feedback signal of inner current control loop can be the input currents or output currents of the modules. Charge control method could be also used for the inner current control loops, where the input currents of the modules are taken as the feedback signals for a two-module ISOP system [28]. In [29], the

12

1 Introduction

Input voltage sharing loops +

_ vin1

_ Gvcd

Voltage sharing bus +

_ vin2

+

_ Gvcd

Drive circuit 1

Comparator 2 − RS Flip-flop +

Drive circuit 2

VRAMP 1

VRAMP 2

…

…

+

Comparator 1 − RS Flip-flop +

+

_ vinn

_ Gvcd +

Voref

+ _

Comparator n − RS Flip-flop +

VRAMP n

Drive circuit n

Gvo vof

Output voltage loop Fig. 1.7 Democratic IVS control strategy for ISOP system

output filter inductor currents are taken as the feedback signals for the inner current control loops with average current control mode. As shown in Fig. 1.8, the duty cycle of each module regulates both the system output voltage and corresponding module input voltage. A simple IVS control strategy was proposed in [30], as shown in Fig. 1.9, where the regulation object of the duty cycle of each module is separated. For the first n − 1 modules, the duty cycles are used to regulate the input voltage of each module to be V in /n, and obviously the input voltage of module n is also V in /n. The system output voltage is regulated by module n. Sensing of the input voltages of the modules is necessary in the control strategies [26–30] to achieve IVS. A current cross-feedback control strategy was proposed in [31] avoiding the sensing of the input voltages. Figure 1.10 shows the control strategy for two-module ISOP system. Two modules use current-mode control with a common system output voltage regulator providing the current reference iref for both current loops. It should be noted that the feedback of current regulator for module 1 is the output current of module 2, while the feedback of current regulator for module 2 is the output current of module 1. Assuming the system input voltage is constant and the system works at steady state, if the input voltage V in1 increase while the input voltage V in2 decreases due to a disturbance, then filter inductor current iLf1 increases while iLf2 decreases. With the control strategy shown in Fig. 1.10, the duty

1.3 DC–DC Series–Parallel Power Conversion Systems

13

Input voltage sharing loops Vin/n

_ vin1 +

_

− + + VRAMP 1

Gvcd vcd_EA1

+

Comparator 1

_ vin2

_ Gvcd

Comparator 2 − RS Flip-flop +

+ VRAMP 2

vin(n−1)

+

_ Comparator (n−1) − RS Flip-flop + + VRAMP (n−1)

Gvcd vcd_EA(n−1) … + + +

+ Voref _

Gvo

Drive circuit 1

Drive circuit 2

…

…

vcd_EA2

_

RS Flip-flop

vo_EA

+

Comparator n − RS Flip-flop +

+ VRAMP n

Drive circuit (n−1)

Drive circuit n

vof

Output voltage loop Fig. 1.8 IVS control strategy for ISOP system

cycle of module 1 increases while the duty cycle of module 2 decreases, resulting in the decrease of V in1 and increase of V in2 , and finally IVS can be achieved. When the number of the modules increases, the control strategy will become complex due to mutual cross-feedback among the current signals and the system modularity is lowered.

1.3.4 ISOS System ISOS system is suitable for the applications where both the input and output voltages are relatively high, and it is critical to ensure both IVS and OVS for the stable operation of ISOS system. Ayyanar et al. proposed a voltage sharing control for ISOS system [32], which is similar to Fig. 1.8. References [33, 34] pointed out that a weak natural rebalancing ability exists in the ISOS system with interleaved switching for

14

1 Introduction

_+

vin1 Gvcd

Vin/n

Module 1

Comparator 1 − RS Flip-flop +

Drive circuit 1

VRAMP 1 _+

vin2 Gvcd

Vin/n

Module 2

Comparator 2 − RS Flip-flop +

Drive circuit 2

…

…

VRAMP 2 vin(n−1)

Comparator (n−1) − RS Flip-flop +

_+ Gvcd

Vin/n

VRAMP (n−1)

Module (n−1) Drive circuit (n−1) Module n

Comparator n +

Gvo

Voref _ vof

VRAMP n

− +

RS Flip-flop

Drive circuit n

Fig. 1.9 A simple IVS control strategy for ISOP system

_ iLf2 +

Output voltage regulator + Voref _

Gvo vof

iref +

Gi Current regulators _ iLf1 Gi

Comparator 1 − RS Flip-flop +

Drive circuit 1

Comparator 2 − RS Flip-flop +

Drive circuit 2

VRAMP 1

VRAMP 2

Fig. 1.10 Current cross-feedback control strategy for two-module ISOP system

different cells. Although the voltage-dependent losses, such as the switching loss, could enhance the rebalancing ability, the voltage sharing performance is considerably affected by the matching of module parameters and load current. The current cross-feedback control strategy can also be used for ISOS system to achieve voltage sharing [35, 36], and its operation principle is the same as that of

1.3 DC–DC Series–Parallel Power Conversion Systems

15

Fig. 1.10. Likewise, the control strategy will become complex when the number of the modules increases. In conclusion, these prior studies focus on one or two of the four connection architectures and do not reveal the inherent relationships between the four connection architectures. In this book, Chap. 2 will systematically analyze the inherent relationships between the sharing of input voltages/currents and the sharing of output voltages/currents from the view of power conservation, and discuss the stability of various possible control strategies. Based on the analysis, a general control strategy is proposed for all the four connection architectures. In Chap. 3, a small-signal model of ISOP system with phase-shifted PWM full-bridge converter as the basic module is derived to design the output voltage loop regulator and IVS loop regulator. Moreover, the design of the IVS loop regulator is analyzed when voltage–current double closed loop is employed for ISOP system. In Chap. 4, a wireless IVS control strategy for ISOP system based on positive output voltage gradient method is presented, in which no control interconnection among the constituent modules is required, leading to a truly modular design for ISOP system and high system reliability.

1.4 DC–AC Inverter Series–Parallel Power Conversion Systems 1.4.1 DC–AC Inverter DC–AC inverter is used for converting a DC voltage into an AC voltage, and there are various topologies. In this book, the popular single-phase full-bridge inverter is selected as the basic module to constitute the series–parallel combined systems. Figure 1.11a shows the single-phase full-bridge inverter, where V in is the input DC voltage, S 1 (D1 )–S 4 (D4 ) constitute the single-phase inverter bridge, L f is the output filter inductor, C f is the output filter capacitor, and Z Ld is the load impedance. For DC–AC inverter, there are mainly two control methods, namely, the voltage single closed-loop control and the voltage–current double closed-loop control. The block diagram of the voltage–current double closed-loop control is shown in Fig. 1.11b, where the error of the voltage reference voref and the sampling signal of output voltage vof is sent into the voltage regulator Gvo , and the output signal of Gvo serves as the current reference iref . Then, the error of the sampling signal of inductor current iLf and iref is sent to the current regulator Gi . The output signal of Gi is fed into the sinusoidal pulse-width modulator (SPWM) and it is compared with the triangular carrier to obtain the drive signals for the four power switches. By removing the inner current loop in Fig. 1.11b, the voltage single closed-loop control will be obtained, as shown in Fig. 1.11c. Compared with the voltage single closed-loop control, the voltage–current double closed-loop control has better dynamic performance and the current-limiting function.

16

1 Introduction

S1

D1 S3

D3 + vinv

Vin

Lf iLf

Cf

– S2

D2 S4

io

+

iCf ZLd

vo _

D4

(a) Single-phase full-bridge inverter +

voref

–

vo

Kv

iLf

Ki

Gvo

+ iref

–

Gi

SPWM

S1~S4

vof iLf_f

(b) Voltage-current double closed-loop control +

voref vo

–

Kv

Gvo

SPWM

S1~S4

vof

(c) Voltage single closed-loop control Fig. 1.11 Single-phase full-bridge inverter and its control block diagram

Regardless of any control method being adopted, multiple DC–AC inverter modules could be connected in series or parallel at the input and output sides to constitute the DC–AC inverter series–parallel combined system.

1.4.2 IPOP DC–AC Inverter System IPOP inverter system, which is commonly known as the parallel inverter system, has been extensively studied among the four series–parallel combined inverter systems. This system has been widely used in uninterruptible power supply (UPS), aeronautical static inverter, and so on. The crucial issue of the IPOP inverter system is to achieve output current sharing of each module. The research aiming at this control objective has experienced the evolution from centralized control, master–slave control to distributed control. The control methods for IPOP inverter system can be also categorized into interconnected control and wireless droop control, and the wireless droop control is essentially a kind of distributed control.

1.4 DC–AC Inverter Series–Parallel Power Conversion Systems

1.4.2.1

17

Centralized Control

Figure 1.12 shows the block diagram of the centralized control for IPOP inverter system [37, 38]. In the centralized control method, the phase-locked loop (PLL) in each inverter module is used to synchronize its output voltage in frequency and phase with the synchronization pulse signal, which is produced by the crystal oscillator in the parallel control unit. The total load current io is sensed and divided by the number of constituent inverter modules connected in parallel, and then serves as the current sharing reference, iref , for each inverter module. The sensed output current of each module is compared with iref , obtaining the corresponding current deviation. Thanks to being synchronized by one synchronization pulse signal, the output voltage of each inverter module can be almost the same in frequency and phase, and the current amplitude difference among the inverter modules is caused by the inconsistence of the voltage amplitude. The aforementioned current deviation is used to compensate any inequality in voltage amplitude of the inverter modules to achieve output current sharing of each module. The implementation of the centralized control is quite simple, and the current sharing is relatively good. However, the whole parallel system tends to collapse once the parallel control unit fails. Thus, the redundant and reliability is not high.

io

Fig. 1.12 Block diagram of centralized control for IPOP inverter system

18

1.4.2.2

1 Introduction

Master–Slave Control

Figure 1.13 shows the master–slave control [39–42], where one inverter is the master module and the others are the slave modules. In Fig. 1.13, K vo is the output voltage sense gain, K i1 –K in are the filter inductor current sense gain of each module, Gvo is the transfer function of the output voltage regulator, Gi1 –Gin are the transfer functions of the current regulator of each module, and Ginv1 –Ginvn are the transfer functions of the SPWM modulators. The output voltage regulator in the master module generates the unified current reference signal iref . Since the output filter inductor current of each module tracks iref , the output current sharing is achieved. The shortcoming of the master–slave control is that the whole system will be collapsed once the master module fails.

1.4.2.3

Distributed Control

1. Average Current Control Method Figure 1.14 shows the average current control, a kind of distributed control [43]. As seen, the voltage reference voref is sent to the output voltage regulators of all the inverter modules, and the average value of the outputs of the output voltage regulators of all the modules, iave , serves as the current reference for all the modules. With such arrangement, the instantaneous current sharing of each inverter module could be achieved without complex current sharing control circuit. This control method has the following merits: simple control circuit, prompt current sharing regulating speed,

Fig. 1.13 Block diagram of master–slave control for IPOP inverter system

1.4 DC–AC Inverter Series–Parallel Power Conversion Systems voref

19

iave + _

Gvo1

iref1

+ _

vof

Gi1 iLf1_f

vinv1 _ + Ginv1 Ki1

irefj +

+ +

+ _

iLf1 io1 +_ iCf1 sCf1

Kvo

Module 1 Output voltage synchronous bus

1 sLf1

Gvon vof

1 n

+ irefn

Module n

ioj

Current sharing bus

_

Gin iLfn_f

+

+ +

vinvn _ +

Ginvn

1 sLfn

Kin

iLfn

ZLd

vo

ion

+_ iCfn sC fn

Kvo

Fig. 1.14 Average current control method for IPOP inverter system

and ease of implementation of redundancy due to modular completely reciprocal. However, the current sharing bus among modules is susceptible to interference, which tends to decrease the current sharing performance. In addition, this control method has poor output characteristics. In order to improve the output characteristics of both single module and the whole system, the load current feed-forward control is introduced [44]. Here, the feed-forward point is located before the current reference averaging point. In the improved method, the current sharing performance and the current-limiting function are remained as the conventional method. 2. Wireless Droop Control In the aforementioned control methods, the current sharing bus is required, and it is susceptible to interference. Figure 1.15 shows the wireless droop control [45–49], which has no current sharing bus. The output characteristic droop circuit is introduced to make the frequency and amplitude of the output voltage of each module decrease with the increase of the active and reactive power, respectively, and thus the frequency and amplitude of the output voltage of each module stabilize at a new equilibrium point. By this means, the control method suppresses the circulating current and achieves the output current sharing. Since there is no connection among the constituent inverter modules, the wireless droop control is not susceptible to interference and has better redundancy. Nevertheless, due to the droop characteristic, the output characteristics are poor.

20

1 Introduction

Sine wave ω1 reference generator Vsinωt V1 +

Gvo1 voref1 – vo1f

iref1 +

P1 Q1 Gi1

– iLf1_f

Power calculation vinv1 Ginv1

_

iLf1 1 sLf1

+

+ _

io1

iCf1 sC f1

Ki1

vo

Kvo

Module 1

ioj

+

+ +

Sine wave reference generator Vsinωt +

Gvon vorefn – vonf Module n

ωn

Pn Qn

irefn +

Gin

– iLfn_f

vinvn +

_

Ginvn

Kin

vo

vo

Power calculation Vn

ZLd

1 sLfn

iLfn

+ ion _ iCfn sC fn

Kvo

Fig. 1.15 Wireless droop control strategy for IPOP inverter system

1.4.3 IPOS DC–AC Inverter System Similar with the output current sharing control for the IPOP inverter system, the output voltage sharing control is only needed for IPOS inverter system. Figure 1.16 shows a three-loop control strategy for IPOS inverter system [50]. The three loops here refer to the common system output voltage loop, the independent output voltage sharing loop of each module, and the inner current loop of each module. In Fig. 1.16, voref is the voltage reference of the system output voltage loop, Gvo is the output voltage regulator, Gi1 –Gin are the current regulators, L f1 –L fn are the filter inductors, C f1 –C fn are the filter capacitors, Gvo1 –Gvon are the regulators of output voltage sharing loop, and the voltage reference vomod_ref of output voltage sharing loop is equal to voref /n. The common output voltage loop regulates the system output voltage, and the output signal of its regulator Gvo serves as an initial reference signal for the inner current loop of each module. Since discrete parameters of each module are not well identical, the output voltage sharing loop is introduced to adjust the reference of the current loop to achieve output voltage sharing of each module. This control strategy

1.4 DC–AC Inverter Series–Parallel Power Conversion Systems

vomod_ref vo1

vo2

Kvo _ +

Kvo _ +

Gvo1

Gvo2

Output voltage sharing loops von

Kvo _ +

voref + _

+ +_

+ +_

21

Kvo

1/n

Inner current loops _ vinv1 Gi1 Ginv1 +

1 iLf1 + iCf1 1 vo1 sLf1 sCf1

iLf1_f

Ki1

Gi2 iLf2_f

Ginv2

vinv2 +

_

vo2 1 iLf2 + iCf2 1 sCf2 sLf2

io

Ki2 +

Gvon

Gvo

iref

vof Output voltage loop

+ +_

Gin iLfn_f

Ginvn

vinvn +

_

+ +

1 ZLd vo

1 iLfn + iCfn 1 von sLfn sCfn

Kin Kvo

Fig. 1.16 Control strategy for IPOS inverter system

can achieve low total harmonic distortion in the system output voltage, fast dynamic response, and good output voltage sharing performance whether the corresponding parameters among modules are equal or not.

1.4.4 ISOP DC–AC Inverter System The crucial issue of the ISOP inverter system is to ensure the input voltage sharing and the output current sharing among the constituent modules. Figure 1.17 shows a compound balanced control strategy [51], which consists of three control loops, namely, the common output voltage loop, the input voltage sharing loop, and the inner current loop of each module. Here, the current inner loop adopts inductor current feedback hysteretic control. In Fig. 1.17, Gvo is the output voltage regulator, Gvcd is the input voltage sharing regulator, and K vi and K vo are the sensor gains of the input voltage and output voltage, respectively. The common output voltage loop regulates the system output voltage, and the output signal of output voltage regulator Gvo , iref , serves as the initial current reference of all the inner current loops. The input voltage sharing loop ensures the even input voltage and the equivalence of the output active power of each module. Since the control object is the inverter whose current reference signal contains both amplitude and phase information, the multipliers are introduced here to synchronize the phase of the initial current reference and that of

22

1 Introduction

Vin n K + vi _ vin1 Kvi Vin n

Gvcd

isv1 _ iref1 +

+_

Input voltage sharing loops + _

Gvcd

Vin n K + vi _ vinn Kvi

Gvcd

vin2

ivd1

Kvi

Gvo vof

Output voltage loop

iLf1_f

_

Inner current loops isv2 _ vinv2 _ iref2 Gi2 Ginv2 + + +_ iLf2_f Ki2

ivdn

isvn

iref

_ + irefn +_

Gin iLfn_f

1 iLf1 + io1 _ sLf1 iCf1

Ki1

ivd2

Kvi

voref + _

Gi1

vinv1 Ginv1 +

vinvn

Ginvn + Kin

_

sCf1

+ vo 1 iLf2 + io2 io ZLd sLf2 _ ++ iCf2 sCf2

1 iLfn + ion sLfn _ iCfn

sCfn

Kvo

Fig. 1.17 Compound balanced control strategy for ISOP inverter system

the signal isvj (j 1, 2,…, n), which is used to alter the current reference by the input voltage sharing loop. Consequently, the output current amplitude is adjusted with the power factor angle of modular output current being kept the same simultaneously. Figure 1.18 shows an output-inductor-current cross-feedback control scheme [52], which consists of two control loops, namely, the common output voltage loop and the inner current loop of each module. The common output voltage loop regulates the system output voltage. In Fig. 1.18, Gvo is the output voltage regulator, and its output signal, iref , serves as the common reference signal of inner current loop of each module. Here, the current feedback for each individual inner current loop is the sum of other output inductor currents instead of its own, and so input voltage sharing (IVS) and output voltage sharing (OCS) are achieved for ISOP inverter system. Although the control scheme is simple, the need of cross-feedback for output inductor current is not conducive to the modularization of the system.

1.4.5 ISOS DC–AC Inverter System The crucial issue of ISOS inverter system is to ensure the input voltage sharing and the output voltage sharing of each module. A compound balanced control strategy is proposed [53], as shown in Fig. 1.19. It consists of three control loops, namely, the system output voltage loop, the input voltage sharing loop, and the inner output filter capacitor current loop of each module. The output signal of the output voltage regu-

1.4 DC–AC Inverter Series–Parallel Power Conversion Systems

+

voref _ Gvo vof

Module 1 iref + Gi1 _ n

∑i

Lfi_f

i=2

Module j + Gij _ n

∑

_ +

io1 1 iLf1 Ginv1 v sLf1 inv1 +_ iCf1 iLf1_f Ki1

Ginvj

+

_

vinvj iLfj_f

iLfi_f

i =1,i ≠ j

Module n +

_

Gin n −1

∑i

Lfi_f

i =1

23

sCf1

+ ioj 1 iLfj ZLd sLfj +_ + iCfj + sCfj Kij

_ ion + 1 iLfn Ginvn sLfn vinvn +_ iCfn iLfn_f Kin

vo

sCfn

Kvo Fig. 1.18 Output-inductor-current cross-feedback control for ISOP inverter system

lator Gvo , iref , serves as the initial current reference of all inner current loops. Since the current reference contains the amplitude and phase information, the multipliers are introduced here to synchronize the phase of the initial current reference and that of the signal isvj (j 1, 2,…, n), which is used to alter the current reference by input voltage sharing loop. Here, the output filter capacitor current control is adopted for the inner current loop, so the input voltage sharing loop only adjusts the amplitude of the output filter capacitor current with the phase of the output filter capacitor current of each module being kept the same. Therefore, only the amplitude of module output voltage is adjusted while the phase is synchronized. The input voltage sharing loop ensures the even input voltage and the equal of the output active power of each module. Thanks to the synchronization of the output voltage of each module, output voltage sharing (OVS) is achieved ultimately. An output-voltage cross-feedback control scheme is proposed [54], as shown in Fig. 1.20. It consists of two control loops, namely, the system output voltage loop and individual output voltage loop of each module. In Fig. 1.20, voref is the system output voltage reference, and the output signal of output voltage regulator Gvo , iref , is sent to all the modules. In the output voltage loop of each module, the reference signal is the amplitude of the system output voltage reference multiplied by (n − 1)/n, and the feedback signal is the sum of output voltage amplitude of other modules except itself. The output voltage regulators Gvo1 –Gvon generate regulating signals

24

1 Introduction

Vin n K + Gvcd1 vi _ vin1 Kvi Input voltage Vin sharing loops n K + Gvcd2 vi _ vin2 Kvi Vin n K + vi _ vinn Kvi

Gvcdn

voref

Gvo

+_

vof Output voltage loop

ivd1

isv1 _ iref1 +

ivd2

ivdn

+_

Gi1 iCf1_f

vinv1 Ginv1 +

_

1 iLf1 iCf1 sLf1 _

1 vo1 sCf1

Ki1

Inner current loops isv2 _ vinv2 _ iref2 Gi2 Ginv2 + + +_ iCf2_f Ki2

vo2 1 sLf2

iLf2 iCf2 _

1 sCf2

io 1 ZLd

isvn

vo iref

_ + irefn Gin +_ i Cfn_f

Ginvn

vinvn +

_

_

1 sLfn iLfn

iCfn

1 von sCfn

Kin

Kvo

Fig. 1.19 Compound balanced control strategy for ISOS inverter system

iref1 –irefn , which are related to magnitude difference. These signals are multiplied with iref to obtain the module adjusting signals isv1 –isvn with the same phase and different amplitudes, which are used to alter the amplitude of module output voltage, and thus input voltage sharing (IVS) and output voltage sharing (OVS) are achieved for ISOS inverter system. In this method, the feedback signal for every individual output voltage loop contains the information of other modules. So, all the modules are inevitably mutual coupling, which restricts the modularity of ISOS inverter system. In summary, for DC–AC inverter series–parallel combined system, existing literature has not analyzed the inherent relationship of the four types of series–parallel combination systems. So, Chap. 5 will analyze the relationship between the input and output powers of the above DC–AC inverter series–parallel combined system, and then put forward a general control strategy so as to form a set of systematic control theory for DC–AC inverter series–parallel combined system, and lay the theoretical foundation for the proposition of the specific control scheme of the four types of series–parallel combination systems. Based on the proposed general control strategy, Chaps. 6 and 7 will give specific control strategies for the ISOP and ISOS inverter systems, and put forward the distributed control based on the centralized control to realize the modularization of the systems. In addition, Chap. 8 will present an improved average current control strategy for IPOP inverter system to improve the output characteristics of both single module and the whole system. Moreover, in this improved method, the current sharing and the current-limiting function are remained unchanged compared with the conventional method.

1.5 Summary

voref

25

_ v of +

Gvo

1# Module

n −1 voref n

+_

Gvo1 Kvo

Kvo

iref isv1

iref1

Gi1

Ginv1

_ + vinv1

io1 1 iLf1 sLf1 +_ i Cf1

n

∑v

sCf1

oi

i=2

isvj

j# Module Gvoj

+_ Kvo

irefj

Gij

Ginvj

_ + vinvj

Ginvn

_ + vinvn

n

∑v

ioj+ 1 iLfj Z sLfj +_ + + Ld iCfj sCfj

vo

oi

i =1

i≠ j

isvn

n# Module +_

Gvon Kvo

irefn

n −1

Gin

ion 1 iLfn sLfn +_ iCfn

∑v

sCfn

oi

i =1

Fig. 1.20 Output-voltage cross-feedback control for ISOS inverter system

1.5 Summary This chapter briefly reviews the history and status of the development of power electronics. Power electronics system integration, standardization, and modularization are one of the effective ways to reduce the cost of producing power electronics products, shorten the development cycle, and improve the reliability. Standardized module series–parallel combination system is one of the significant constituent parts of power electronic system integration, and the research on its control technique to ensure the stability of the system operation and the power balance of each module is the primarily crucial issue. This chapter introduces the advantages and major categories of standardized module series–parallel combination system, and elaborates the control techniques for the series–parallel combination system with DC–DC converter and DC–AC inverter as the basic module, which will lay a foundation for the narration of the following chapters.

26

1 Introduction

References 1. Prince, D.C.: Mercury Arc Rectifier Phenomena. Trans. American Institute of Electrical Engineers, pp. 1064–1071 (1927) 2. Bose, B.K.: Power electronics—a technology review. Proc. IEEE 80(8), 1303–1334 (1992) 3. Newell, W.E.: Power electronics—emerging from Limbo. In: Proceedings of the IEEE Power Electronics Specialists Conference, pp. 6–12 (1973) 4. Bose, B.K.: Power electronics—an emerging technology. IEEE Trans. Ind. Electron. 36(2), 403–412 (1989) 5. Blaabjerg, F., Chen, Z., Kjaer, S.B.: Power electronics as efficient interface in dispersed power generation systems. IEEE Trans. Power Electron. 19(5), 1184–1194 (2004) 6. Emadi, A., Williamson, S.S., Khaligh, A.: Power electronics intensive solutions for advanced electric, hybrid electric, and fuel cell vehicular power systems. IEEE Trans. Power Electron. 21(3), 567–577 (2006) 7. Bose, B.K.: Power electronics and motor drives recent progress and perspective. IEEE Trans. Ind. Electron. 56(2), 581–588 (2009) 8. Blaabjerg, F., Ma, K., Yang, Y.: Power electronics—the key technology for renewable energy systems. In: Proceedings of the International Conference on Ecological Vehicles and Renewable Energies, pp. 1–11 (2014) 9. Lorenz, L.: Key power semiconductor devices and development trends. In: Proceedings of the International Conference on Electrical Machines and Systems, pp. 1137–1142 (2008) 10. Jacobus, D., Lee, F.C.: On a future for power electronics. IEEE J. Emerg. Sel. Top. Power Electron. 1(2), 59–72 (2013) 11. Kornegay, K.T.: Design issues in power electronic building block (PEBB) system integration. In: Proceedings of IEEE Computer Society Workshop on System Level Design, pp. 48–52 (1998) 12. Lee, F.C., Borojevic, D., Xing, K., Thandi, G., Zhu, H.: PEBBs and system integration at Virginia power electronics center, integrated power electronics module—a building block concept for system integration. In: Proceedings of the Center for Power Electronics Systems, pp. 1–4 (2000) 13. Reeves, R.: Inductor-capacitor hybrid. Proc. Inst. Electr. Eng. 122(11), 1323–1326 (1975) 14. Waffenschmidt, E., Ackermann, B., Ferreira, J.A.: Design method and material technologies for passives in printed circuit board embedded circuits. IEEE Trans. Power Electron. 20(3), 576–584 (2005) 15. Luo, S., Batarseh, I.: A review of distributed power systems part I: DC distributed power system. IEEE Aerosp. Electron. Syst. Mag. 20(8), 5–16 (2005) 16. Chen, W., Ruan, X., Yan, H., Tse, C.K.: DC/DC conversion systems consisting of multiple converter modules: stability, control and experimental verifications. IEEE Trans. Power Electron. 24(6), 1463–1474 (2009) 17. Kim, J., Choi, H., Cho, B.H.: A novel droop method for converter parallel operation. IEEE Trans. Power Electron. 17(1), 25–32 (2002) 18. Panov, Y., Jovanovi´c, M.M.: Stability and dynamic performance of current-sharing control for paralleled voltage regulator modules. IEEE Trans. Power Electron. 17(2), 172–179 (2002) 19. Lin, C., Chen, C.: Single-wire current-share paralleling of current-mode-controlled DC power supplies. IEEE Trans. Ind. Electron. 47(4), 780–786 (2000) 20. Huang, Y., Tse, C.K.: Circuit theoretic classification of parallel connected DC-DC converters. IEEE Trans. Circuits. Syst. 54(5), 1099–1108 (2007) 21. Du, Q., Qi, B., Wang, T., Zhang, T., Li, X.: A high-power input-parallel output-series buck and half-bridge converter and control methods. IEEE Trans. Power Electron. 27(6), 2703–2715 (2012) 22. Hu, X., Gong, C.: A high gain input-parallel output-series DC/DC converter with dual coupled inductors. IEEE Trans. Power Electron. 30(3), 1306–1317 (2015)

References

27

23. Giri, R., Choudhary, V., Ayyanar, R., Mohan, N.: Common-duty-ratio control of input-series connected modular DC-DC converters with active input voltage and load-current sharing. IEEE Trans. Ind. Appl. 42(4), 1101–1111 (2006) 24. Manias, S., Kostakis, G.: Modular DC–DC converter for high output voltage applications. IEE Electr. Power Appl. 140(2), 97–102 (1993) 25. Siri, K., Conner, K., Truong, C.: Uniform voltage distribution control for paralleled-input, series-output connected converters. In: Proceedings of the IEEE Aerospace Conference, pp. 1–11 (2005) 26. Siri, K., Willhoff, M., Conner, K.: Uniform voltage distribution control for series connected DC–DC converters. IEEE Trans. Power Electron. 22(4), 1269–1279 (2007) 27. Ruan, X., Chen, W., Cheng, L., Tse, C.K., Yan, H., Zhang, T.: Control strategy for input-seriesoutput-parallel converters. IEEE Trans. Ind. Electron. 56(4), 1174–1185 (2009) 28. Kim, J., You, J., Cho, B.H.: Modeling, control, and design of input-series-output-parallelconnected converter for high-speed-train power system. IEEE Trans. Ind. Electron. 48(3), 536–544 (2001) 29. Ayyanar, R., Giri, R., Mohan, N.: Active input-voltage and load-current sharing in inputseries and output-parallel connected modular DC-DC converters using dynamic input-voltage reference scheme. IEEE Trans. Power Electron. 19(6), 1462–1472 (2004) 30. Jang, Y., Jovanovic, M.M., Dillman, D., Li, S., Yang, C.: Input-voltage balancing of seriesconnected converters. In: Proceedings of the IEEE Applied Power Electronics Conference and Exposition, pp. 1153–1160 (2011) 31. Sha, D., Guo, Z., Liao, X.: Cross-feedback output-current-sharing control for input-seriesoutput-parallel modular DC–DC converters. IEEE Trans. Power Electron. 25(11), 2762–2771 (2010) 32. Giri, R., Ayyanar, R., Ledezma, E.: Input-series and output-series connected modular DC-DC converters with active input voltage and output voltage sharing. In: Proceedings of the IEEE Applied Power Electronics Conference and Exposition, pp. 1751–1756 (2004) 33. Merwe, J., Mouton, H.: An investigation of the natural balancing mechanisms of modular input-series-output-series DC-DC converters. In: Proceedings of the IEEE Energy Conversion Congress and Exposition, pp. 817–822 (2010) 34. Lu, Q., Yang, Z., Lin, S., Wang, S., Wang, C.: Research on voltage sharing for input-seriesoutput-series phase-shift full-bridge converters with common-duty-ratio. In: Proceedings of the Annual Conference of the IEEE Industrial Electronics Society, pp. 1548–1553 (2011) 35. Sha, D., Deng, K., Liao, X.: Duty cycle exchanging control for input-series-output-series connected two PS-FB DC-DC converters. IEEE Trans. Power Electron. 27(3), 1490–1501 (2012) 36. Sha, D., Guo, Z., Luo, T., Liao, X.: A general control strategy for input-series-output-series modular DC-DC converters. IEEE Trans. Power Electron. 29(7), 3766–3775 (2014) 37. Bekiarov, S.B., Emadi, A.: Uninterruptible power supplies: classification operation dynamics and control. In: Proceedings of the IEEE Applied Power Electronics Conference and Exposition, pp. 597–604 (2002) 38. Chen, X., Kang, Y., Chen, J.: Operation, control technique of parallel connected high power three-phase inverters. In: Proceedings of the IEEE International Power Electronics and Motion Control Conference, pp. 956–959 (2004) 39. Chen, J., Chu, C.: Combination voltage-controlled and current-controlled PWM inverter for UPS parallel operation. IEEE Trans. Power Electron. 10(5), 547–558 (1995) 40. Lee, W.C., Lee, T.K., Lee, S.H., Kim, K.H., Hyun, D.S., Suh, I.Y.: A master and slave control strategy for parallel operation of three-phase UPS systems with different ratings. In: Proceedings of the IEEE Applied Power Electronics Conference and Exposition, pp. 456–462 (2004) 41. Chiang, S.J., Lin, C.H., Yen, C.Y.: Current limitation control technique for parallel operation of UPS inverters. In: Proceedings of the IEEE Power Electronics Specialists Conference, pp. 1922–1926 (2004) 42. Perez-Ladron, G., Cardenas, V., Espinosa, G.: Analysis and implementation of a master-slave control based on a passivity approach for parallel inverters operation. In: Proceedings of the IEEE International Power Electronics Conference, pp. 1–5 (2006)

28

1 Introduction

43. Xing, Y., Huang, L., Yan, Y.: Redundant parallel control for current regulated inverters with instantaneous current sharing. In: Proceedings of the IEEE Power Electronics Specialists Conference, pp. 1438–1442 (2003) 44. Fang, T., Ruan, X., Xiao, L., Liu, A.: An improved distributed control strategy for parallel inverters. In: Proceedings of the IEEE Power Electronics Specialists Conference, pp. 3500–3505 (2008) 45. Tuladhar, A.: Advanced control techniques for parallel inverter operation without control interconnections. Ph.D. Dissertation, The University of British Columbia, Vancouver, Canada (2000) 46. Lin, X., Feng, F., Duan, S., Kang, Y., Chen, J.: The droop characteristic decoupling control of parallel connected UPS with no control interconnection. In: Proceedings of the IEEE International Electric Machines and Drives Conference, pp. 1777–1780 (2003) 47. De Brabandere, K., Bolsens, B., Van den Keybus, J., Woyte, A., Driesen, J., Belmans, R.: A voltage and frequency droop control method for parallel inverters. IEEE Trans. Power Electron. 22(4), 1107–1115 (2007) 48. Yao, W., Chen, M., Gao, M., Qian, Z.: A wireless load sharing controller to improve the performance of parallel-connected inverters. In: Proceedings of the IEEE Applied Power Electronics Conference and Exposition, pp. 1628–1631 (2008) 49. Guerrero, J., Vicuña, L., Matas, J., Castilla, M., Miret, J.: Output impedance design of parallelconnected UPS inverters with wireless load-sharing control. IEEE Trans. Ind. Electron. 52(4), 1126–1135 (2005) 50. Song, C., Zhao, R., Lin, W., Zeng, Z.: A novel control strategy for input-parallel-output-series inverter system. In: Proceedings of the IEEE International Conference on Electrical Machines and Systems, pp. 1–5 (2011) 51. Chen, W., Zhuang, K., Ruan, X.: Input-series and output-parallel connected inverter system for high input voltage applications. IEEE Trans. Power Electron. 24(9), 2127–2137 (2009) 52. Sha, D., Deng, K., Guo, Z., Liao, X.: Control strategy for input-series–output-parallel highfrequency ac link inverters. IEEE Trans. Ind. Electron. 59(11), 4101–4111 (2012) 53. Fang, T., Ruan, X., Tse, C.K.: Control strategy to achieve input and output voltage sharing for input-series-output-series connected inverters systems. IEEE Trans. Power Electron. 25(6), 1585–1596 (2010) 54. Sha, D., Xu, G., Liao, X.: Control strategy for input-series-output-series high-frequency ac-link inverters. IEEE Trans. Power Electron. 28(11), 5283–5292 (2012)

Chapter 2

A General Control Strategy for DC–DC Series–Parallel Power Conversion Systems

Abstract The inherent relationships between the sharing of input voltages/currents and the sharing of output voltages/currents are analyzed systematically from the view of power conservation in this chapter, and the stability of various possible control strategies is discussed. Based on the analysis, a general control strategy is proposed for the DC–DC series–parallel power conversion systems, which not only achieves equilibrium between the constituent modules but also decouples the output voltage control loop and the sharing control loop. Furthermore, a modularization architecture for the DC–DC series–parallel power conversion systems is proposed, achieving full modularization where all the modules are self-contained and standardized, and no extra control is needed to achieve sharing of voltage and/or current at the input and output sides. Experimental prototypes are built and tested to validate the proposed general control strategy and the modularization architecture. Keywords Sharing of input voltages/current · Sharing of output voltages/current Stability · General control strategy · Modularization architecture In Chap. 1, the four basic connection architectures (IPOP, IPOS, ISOP, and ISOS) of DC–DC series–parallel power conversion systems were introduced. In order to ensure proper operation of DC–DC series–parallel power conversion systems, the equilibrium between the constituent modules must be achieved, i.e., proper input voltage sharing (IVS) and output voltage sharing (OVS) among the constituent modules when they are connected in series at the input and output sides, respectively, and proper input current sharing (ICS) and output current sharing (OCS) among the constituent modules when they are connected in parallel at the input and output sides, respectively. Previous studies have addressed the issue of sharing of the input and/or output voltage and/or current. However, these prior studies focus on one or two of the four connection architectures and do not reveal the inherent relationships between the four connection architectures. In this chapter, the inherent relationships between the sharing of input voltages/currents and the sharing of output voltages/currents are analyzed systematically from the view of power conservation, and the stability of various possible control strategies are discussed. Based on the analysis, a general control strategy is proposed for the DC–DC series–parallel power conversion systems, which © Springer Nature Singapore Pte Ltd. and Science Press, Beijing, China 2019 X. Ruan et al., Control of Series-Parallel Conversion Systems, CPSS Power Electronics Series, https://doi.org/10.1007/978-981-13-2760-5_2

29

30

2 A General Control Strategy for DC–DC Series–Parallel Power …

not only achieves equilibrium between the constituent modules but also decouples the output voltage control loop and the sharing control loop. Furthermore, a modularization architecture for the DC–DC series–parallel power conversion systems is proposed, achieving full modularization where all the modules are self-contained and standardized, and no extra control is needed to achieve sharing of voltage and/or current at the input and output sides. Experimental prototypes were built and tested to validate the proposed general control strategy and the modularization architecture.

2.1 Possible Control Strategies for DC–DC Series–Parallel Power Conversion Systems For input-parallel-connected systems (IPOP and IPOS), the input voltages of modules are equal, therefore, we need to make the module output currents equal for IPOP system and the output module voltages equal for IPOS system, i.e., OCS and OVS should be ensured for IPOP and IPOS systems, respectively. For input-seriesconnected systems (ISOP and ISOS), we not only have to ensure OCS and OVS for ISOP and ISOS systems, respectively, but also ensure IVS for both ISOP and ISOS systems. In other words, both IVS and OCS are needed for ISOP system, and both IVS and OVS are needed for ISOS system. As discussed before, we can see that if the modules are connected in parallel at the input sides, only OCS or OVS is needed, and if the modules are connected in series at the input sides, both IVS and OCS or OVS are needed. Two questions are of fundamental importance: (1) for IPOP and IPOS systems, as all modules share the same input voltage, would ICS imply OCS or OVS so that we only need to direct control to ensure ICS? and (2) for ISOP and ISOS systems, what is the relationship between IVS and OCS or OVS? Figure 2.1 shows the four connection architectures of DC–DC series–parallel power conversion systems. For the j-th (j 1, 2, …, n, n is the number of modules) module, V in_j and I in_j denote the input voltage and the input current, respectively; V o_j and I o_j denote the output voltage and the output current, respectively; I cd_j and I cf_j denote the input and output filter capacitor currents, respectively; and I cin_j and I co_j denote the input and output currents of the power stage, not including the input and output filter capacitor currents, respectively. At steady state, since both the input voltage and the output voltage are constant, the average values of both the input filter capacitor current and the output filter capacitor current are zero, we have, Iin_ j Icin_ j ( j 1, 2, . . . , n)

(2.1)

Ico_ j Io_ j ( j 1, 2, . . . , n)

(2.2)

Then, by power conservation, we have,

2.1 Possible Control Strategies for DC–DC Series–Parallel Power … Iin

Iin_1

Icin_1

Ico_1

+ Icd_1 Vin_1 _ Cd1

Vin

Io_1 Icf_1

DC-DC Converter

Cf1

Io

Iin

+

+ Vo_1 _

Vo _

Iin_1

Icin_1

Ico_1

+ Icd_1 Vin_1 _ Cd1

Vin

31

DC-DC Converter

Icin_2

Ico_2

+ Icd_2 Vin_2 _ Cd2

Io_2 Icf_2

DC-DC Converter

Cf2

Iin_2

Icin_2

Ico_2

+ Icd_2 Vin_2 _ Cd2

+ Vo_2 _

DC-DC Converter

Icin_n

Ico_n

+ Icd_n Vin_n _ Cdn

DC-DC Converter

Io_n Icf_n

Cfn

Iin_n

Icin_n

Icin_1

+ Icd_1 Vin_1 _ Cd1

DC-DC Converter

Io_1 Icf_1

Cf1

+ Vo_1 _

Io

Iin

+

Iin_1

Icin_1

+ Icd_1 Vin_1 _ Cd1

Vin

Icin_2

+ Icd_2 Vin_2 _ Cd2

Ico_2 DC-DC Converter

Vo _

DC-DC Converter

Icin_n

+ Icd_n Vin_n _ Cdn

Icf_2 Cf2

Ico_n DC-DC Converter

+ Vo_2 _

Iin_2

Vin

Icin_2

+ Vo_n _

_

Cfn

Cf1

Io

+ Vo_1 _

+

+ Icd_2 Vin_2 _ Cd2

DC-DC Converter

Io_2 Icf_2

Cf2

+ Vo_2 _

Vo

Module 2

Io_n Icf_n

Io_1 Icf_1

Ico_2

+ Vo_n _

Iin_n

Icin_n

+ Icd_n Vin_n _ Cdn

Ico_n DC-DC Converter

Module n

(c) ISOP

Cfn

Module 1

Io_2

Module 2

Iin_n

Vo

Io_n Icf_n

Ico_1

Module 1

Iin_2

+ Vo_2 _

(b) IPOS Ico_1

DC-DC Converter

Cf2

Module n

(a) IPOP Iin_1

Io_2 Icf_2

Ico_n

+ Icd_n Vin_n _ Cdn

+ Vo_n _

Module n

Iin

+

Module 2

Module 2

Iin_n

Cf1

Io

+ Vo_1 _

Module 1

Module 1

Iin_2

Io_1 Icf_1

Io_n Icf_n

Cfn

+ Vo_n _

_

Module n

(d) ISOS

Fig. 2.1 Four connection architectures of DC–DC series–parallel power conversion systems

⎧ ⎪ Vo_1 · Io_1 Po1 Pin1 · η1 Vin_1 · Iin_1 · η1 ⎪ ⎪ ⎪ ⎪ ⎨ Vo_2 · Io_2 Po2 Pin2 · η2 Vin_2 · Iin_2 · η2 .. ⎪ ⎪ . ⎪ ⎪ ⎪ ⎩V · I P P ·η V o_n o_n on inn n in_n · Iin_n · ηn

(2.3)

where Po1 , Po2 , …, Pon are the output power of the modules, Pin1 , Pin2 , …, Pinn are the input power of the modules, and η1 , η2 , …, ηn are the conversion efficiencies of the modules.

32

2 A General Control Strategy for DC–DC Series–Parallel Power …

2.1.1 IPOP Systems For IPOP systems, the input voltages of the modules are equal, and the output voltages of the modules are equal, i.e., V in_1 V in_2 … V in_n and V o_1 V o_2 … V o_n . If the module input currents are equal, i.e., I in_1 I in_2 … I in_n , then the module input power are equal, i.e., Pin1 Pin2 … Pinn . According to (2.3), we can get, Io_1 /η1 Io_2 /η2 · · · Io_n /ηn

(2.4)

Equation (2.4) illustrates that the difference among the module output currents is dependent on the difference of the module efficiencies. In practice, as the topologies of the modules are identical, the mismatches in the power stages, such as the power devices, transformers, inductors, and capacitors, are not so large, so that the difference among the module efficiencies is very little and it will be ignored in the following. Thus, the module output currents will be nearly evenly shared for IPOP systems. In other words, achieving ICS implies OCS for IPOP systems.

2.1.2 IPOS Systems For IPOS systems, the input voltages of the modules are equal, and the output currents of the modules are equal, i.e., V in_1 V in_2 … V in_n and I o_1 I o_2 … I o_n . If the module input currents are equal, i.e., I in_1 I in_2 … I in_n , then the module input power are equal, i.e., Pin1 Pin2 … Pinn . According to (2.3), we can get, Vo_1 /η1 Vo_2 /η2 · · · Vo_n /ηn

(2.5)

Equation (2.5) illustrates that the difference among the module output voltages is dependent on the difference of the module efficiencies. Thus, the module output voltages will be nearly evenly shared for IPOS systems. In other words, achieving ICS implies OVS for IPOS systems.

2.1.3 ISOP Systems For ISOP systems, the input currents of the modules are equal, and the output voltages of the modules are equal, i.e., I in_1 I in_2 … I in_n and V o_1 V o_2 … V o_n . If the module output currents are equal, i.e., I o_1 I o_2 … I o_n , then the module output power are equal, i.e., Po1 Po2 … Pon . According to (2.3), we have, Vin_1 · η1 Vin_2 · η2 · · · Vin_n · ηn

(2.6)

2.1 Possible Control Strategies for DC–DC Series–Parallel Power …

33

Equation (2.6) illustrates that the difference in the input voltages among the modules is dependent on the difference of the module efficiencies if the module output currents are equal. If IVS control strategy is used to achieve V in_1 V in_2 … V in_n V in /n, then the module input power are equal, i.e., Pin1 Pin2 … Pinn . According to (2.3), we have, Io_1 /η1 Io_2 /η2 · · · Io_n /ηn

(2.7)

Equation (2.7) illustrates that the difference in the output currents among the modules is dependent on the difference of the module efficiencies if the module input voltages are equal. In summary, for ISOP systems, if the module input voltages are equal, the module output currents will be nearly equal, and on the other hand, if the module output currents are equal, the module input voltages will be nearly equal. So, we can conclude that once IVS is achieved, OCS is nearly achieved automatically and vice versa. That is to say, we need to ensure either IVS or OCS.

2.1.4 ISOS Systems For ISOS systems, the input currents of the modules are equal, and the output currents of the modules are equal, i.e., I in_1 I in_2 … I in_n and I o_1 I o_2 … I o_n . If the module output voltages are equal, i.e., V o_1 V o_2 … V o_n , then the module output power are equal, i.e., Po1 Po2 … Pon . According to (2.3), we have, Vin_1 · η1 Vin_2 · η2 · · · Vin_n · ηn

(2.8)

Equation (2.8) illustrates that the difference in the input voltages among the modules is dependent on the difference of the module efficiencies if the module output voltages are equal. If IVS control strategy is used to achieve V in_1 V in_2 … V in_n V in /n, then the module input power are equal, i.e., Pin1 Pin2 … Pinn . According to (2.3), we have, Vo_1 /η1 Vo_2 /η2 · · · Vo_n /ηn

(2.9)

Equation (2.9) illustrates that the difference in the output voltages among the modules is dependent on the difference of the module efficiencies if the module input voltages are equal. In summary, for ISOS systems, we can see that if the module input voltages are equal, the module output voltages will be nearly equal, and on the other hand, if the module output voltages are equal, the module input voltages will be nearly

34

2 A General Control Strategy for DC–DC Series–Parallel Power …

Table 2.1 Relationship between the input sides and output sides Architecture Input side Output side IPOP

Input voltages

Input currents

Output voltages

Output currents

Equal by connection

Equal sharing achievable by control

Equal by connection

Equal sharing achievable by control Equal by connection

IPOS

ISOP

Equal sharing achievable by control

Equal by connection

ISOS

Equal sharing achievable by control Equal by connection Equal sharing achievable by control

Equal sharing achievable by control Equal by connection

equal. Thus, we can conclude that once IVS is achieved, OVS is nearly achieved automatically and vice versa. That is to say, we need to ensure either IVS or OVS. Based on the above analysis, we have the following conclusions if the module efficiencies are the same: (1) (2) (3) (4)

for IPOP systems, achieving ICS implies OCS, for IPOS systems, achieving ICS implies OVS, for ISOP systems, achieving IVS implies OCS and vice versa, for ISOS systems, achieving IVS implies OVS and vice versa.

From the earlier discussion, we clearly see that for any of the four connection architectures, proper sharing of all voltages or currents can be nearly achieved by directing control efforts to ensure the proper sharing of either the input voltage or current, as shown in Table 2.1.

2.2 Stability of the Control Strategies In this section, we examine the stability of the system when the input voltages or currents are controlled to achieve IVS or ICS, and when the output voltages or currents are controlled to achieve OVS or OCS. In our analysis, we assume that the system’s input voltage, output voltage, and load current are kept unchanged when perturbations are imposed on the input voltage/current and output voltage/current of the modules.

2.2 Stability of the Control Strategies

35

2.2.1 Controlling at Input Side Regardless of the connection type at the input side, when ICS (for input-parallelconnected systems) or IVS (for input-series-connected systems) is achieved, each converter module behaves as a constant power source as seen at the output side, as shown in Fig. 2.2, and the module input powers are equal, i.e., Vo_ j · Ico_ j Pin j · η const ( j 1, 2, . . . , n)

(2.10)

The relationship between the currents at the output side is Icf_ j Ico_ j − Io_ j ( j 1, 2, . . . , n)

(2.11)

For the case where the output sides are connected in parallel, as shown in Fig. 2.2a, the output voltages of all modules are equal, and from (2.10), the output currents of all modules are identical. Thus, no perturbation occurs, i.e., the system is stable. We now examine the case where the output sides are connected in series, as shown in Fig. 2.2b. We assume that a perturbation occurs in the output filter capacitor voltages and the output voltages of some modules, e.g., V o_1 increases and V o_2 decreases, while the output voltages of other modules are unchanged. According to (2.10) and (2.11), I co_1 and I cf_1 vary as shown in Fig. 2.3. Suppose point A is the equilibrium operating point, when V o_1 increases under perturbation, the operating point moves to point B. At point B, I co_1 is smaller than I o and V o_1 is higher than V o /n. Hence, the filter capacitor current I cf_1 is negative, causing C f1 to be discharged Ico_1 Constant power source (Module 1)

Constant power source (Module 2)

Constant power source (Module n)

Io_1 Icf_1

Cf1

+ Vo_1 _

Ico_2

Io_2 Icf_2

Cf2

+ Vo_2 _

Ico_n

Io_n Icf_n

Cfn

Io

+ Vo_n _

(a) Output sides connected in parallel

Vo _

Io_1

Ico_1

+ Constant power source (Module 1)

Constant power source (Module 2)

Icf_1 Cf1

+ Vo_1 _

Ico_2

Io_2 Icf_2

Cf2

+ Vo_2 _

Ico_n Constant power source (Module n)

Io

+ Vo_n _

(b) Output sides connected in series

Fig. 2.2 Equivalent schematic diagram for the input side control strategies

Vo

Io_n Icf_n

Cfn

+

_

36

2 A General Control Strategy for DC–DC Series–Parallel Power …

Fig. 2.3 Currents and voltage of the output side of module 1

Ico_1 (Pin = const) A

Io

0

B Vo n

Vo_1 Icf_1

to supply the deficit current, and as a result, V o_1 decreases, and the operating point returns to point A. Thus, V o_1 returns to the equilibrium voltage. Similarly, I co_2 increases as V o_2 decreases, and as it becomes larger than I o , the excessive current charges C f2 to make V o_2 increase and return to the equilibrium voltage. Therefore, any control strategies intending to achieve ICS and IVS for input-parallel-connected systems (including IPOP and IPOS) and input-series-connected systems (including ISOP and ISOS), respectively, are stable.

2.2.2 Controlling at Output Side Regardless of the type of connection at the output side, when OCS (for outputparallel-connected systems) or OVS (for output-series-connected systems) is achieved, each of the converter modules behaves as a constant power sink as seen from the input side, as shown in Fig. 2.4, and the output powers of all modules are equal, i.e., Vin_ j · Icin_ j Po j /η const ( j 1, 2, . . . , n)

(2.12)

The relationship between the currents at the input side is Icd_ j Iin_ j − Icin_ j ( j 1, 2, . . . , n)

(2.13)

For the case where the input sides are connected in parallel, as shown in Fig. 2.4a, the input voltages of all modules are equal, and the input currents are thus identical. Hence, no perturbation occurs, i.e., the system is stable. To examine the case where the input sides are connected in series, as shown in Fig. 2.4b, we again consider perturbing the input filter capacitor voltages, the input voltages of some modules, e.g., increasing V in_1 and decreasing V in_2 , while keeping the input voltages of other modules unchanged. By virtue of constant power, I cin_1 will decrease and I cin_2 will increase, as illustrated in (2.13). According to (2.12) and

2.2 Stability of the Control Strategies

Iin Vin

Iin_1

Icin_1

+ Icd_1 Vin_1 Cd1 _ Iin_2

Iin

Iin_2 Constant power sink (Module 2)

Icin_n

+ Icd_n Vin_n Cdn _

Iin_1

Vin

(a) Input sides connected in parallel

Constant power sink (Module 1)

Icin_2

+ Icd_2 Vin_2 Cd2 _

Iin_n Constant power sink (Module n)

Icin_1

+ Icd_1 Vin_1 Cd1 _

Constant power sink (Module 1)

Icin_2

+ Icd_2 Vin_2 Cd2 _

Iin_n

37

Constant power sink (Module 2)

Icin_n

+ Icd_n Vin_n Cdn _

Constant power sink (Module n)

(b) Input sides connected in series

Fig. 2.4 Equivalent schematic diagram for the output side control strategies

(2.13), I cin_1 and I cd_1 can be shown to vary as given in Fig. 2.5. Suppose point A is the equilibrium operating point when V in_1 is perturbed to increase, the operating point moves to point B. At point B, I cin_1 is smaller than the system input current I in , and it can be derived from (2.13) that the input filter capacitor current I cd_1 is positive, causing C d1 to be charged up and its voltage increased. Thus, a positive feedback loop is formed and the operating point continues to move to the right and away from point A. So, the equilibrium point cannot be resumed. Similarly, if I cin_2 is larger than I in causing C d2 to be discharged by the unbalanced current, then V in_2 decreases and fails to return to the equilibrium point. If the module employs a buck-derived converter when V in_2 continues to decrease, it will not be high enough to provide the desired voltage, and the output current thus decreases, resulting in unbalanced output currents of the modules. Moreover, if the module employs a boost-derived or buck/boost-derived converter, V in_2 will continue to decrease until it is zero and provides no power to the load, again resulting in unbalanced output currents of the modules. From the above analysis, we can conclude that any control strategies intending to achieve OCS and OVS for ISOP and ISOS systems, respectively, are unstable. In addition to the above mentioned qualitative analysis, the tool of small-signal model can be used to prove that any control strategies intending to achieve OCS and OVS for ISOP and ISOS systems, respectively, are unstable. Referring to Fig. 2.4b, each converter module presents constant power sink characteristics as seen from the input side of power stage when OCS (for ISOP) or OVS (for ISOS) is achieved, and the input impedance of each converter module can be modeled as an equivalent negative resistance [1], i.e.,

38

2 A General Control Strategy for DC–DC Series–Parallel Power …

Fig. 2.5 Currents and voltage of the input side of module 1

Icin_1 ( Po = const) Iin

0

Rne j

2 Vin_ vˆin_ j Vin_ j j − − ˆi cin_ j Icin_ j Vin_ j · Icin_ j

A

B

Vin n

( j 1, 2, . . . , n)

Icd_1

Vin_1

(2.14)

At steady state, we have, Vin_ j Icin_ j η Po /n ( j 1, 2, . . . , n)

(2.15)

where Po is the output power of ISOP or ISOS system. Combining (2.14) and (2.15), yields, Rne j

2 Vin_ vˆin_ j j ·η·n − ( j 1, 2, . . . , n) Po iˆcin_ j

(2.16)

Figure 2.6 shows the small-signal model of input-series-connected systems employing OVS/OCS control strategies by replacing each module in Fig. 2.4b with a negative resistance Rnej (j 1, 2, …, n). The total input impedance of each module at the input terminal is, Z j Rne j //(1/sCd j )

Rne j 1 + s Rne j Cd j

( j 1, 2, . . . , n)

(2.17)

The individual input voltage perturbation caused by the system input voltage perturbation vˆin can be expressed as vˆin_ j

Zj · vˆin ( j 1, 2, . . . , n) n Zm

(2.18)

m1

According to (2.18), the difference of input voltage perturbations between module j and module k is vˆ jk vˆin_ j − vˆin_k

Z j − Zk vˆin ( j, k 1, 2, . . . , n; j k) n Zm m1

(2.19)

2.2 Stability of the Control Strategies

39

Fig. 2.6 Small-signal model of input-series-connected systems employing OVS/OCS control strategies

iˆcin_1

iˆin

Z1

+ ˆvin_1 _ Cd1

Rne1

iˆcin_2

Z2 vˆ in

+ vˆ in_2 _ Cd2

Rne2

iˆcin_n

Zn

+ vˆ in_n C _ dn

Rnen

Equation (2.19) can also be expressed as the response in the difference of input voltage perturbations of two modules due to perturbation in the system input voltage, i.e., vˆin_ j − vˆin_k Z j − Zk vˆ jk n ( j, k 1, 2, . . . , n; j k) vˆin vˆin Zm

(2.20)

m1

For simplicity, letting n 2, substitution of (2.17) into (2.20), yields, vˆ12 (Rne1 − Rne2 ) + s Rne1 Rne2 (Cd2 − Cd1 ) vˆin (Rne1 + Rne2 ) + s Rne1 Rne2 (Cd2 + Cd1 )

(2.21)

At steady state, V in_1 V in_2 V in / 2, substitution the equation into (2.16), we V 2η get Rne1 Rne2 − 2Pin o , then (2.21) can be rewritten as s(Cd1 − Cd2 )Vin2 η vˆ12 vˆin 4Po − s(Cd1 + Cd2 )Vin2 η

(2.22)

In practice, C d1 is impossible equal to C d2 precisely, it can be found from (2.22) that the transfer function of vˆ12 /vˆin has a right-half-plane pole, which implies unstable operation. Therefore, any control strategies intending to achieve OCS and OVS for ISOP and ISOS systems, respectively, are unstable, and the root cause is the negative resistance characteristics of the module’s input impedance with the control strategies to achieve OCS and OVS. From the foregoing discussion, we can draw the following conclusions: (1) for the input-parallel-connected systems, we can directly control efforts to achieve ICS that will automatically ensure OCS and OVS for IPOP and IPOS systems, respectively. Alternatively, we can directly control the output currents to achieve OCS for IPOP systems, or directly control the output voltages to achieve OVS for IPOS systems,

40

2 A General Control Strategy for DC–DC Series–Parallel Power …

(2) for the input-series-connected systems, we should control the input voltages to achieve IVS, and OCS or OVS will be achieved automatically. Any attempt to control the output currents or voltages to achieve OCS or OVS will not be successful because the control strategy is inherently unstable.

2.3 General Control Strategy for DC–DC Series–Parallel Power Conversion Systems In the following analysis, buck-derived converters with PWM control are taken as the basic modules (such as full-bridge converter, half-bridge converter, forward converter). Regardless of the type of the connection architectures, at least two control loops are needed, one is to regulate the system output voltage and the other is to control the sharing of input voltages/currents or output voltages/currents. Basically, the two control loops are exercised by varying the duty cycles of the modules. However, varying the duty cycles may affect both the output voltage of the system and the sharing of the input/output voltages/currents of the modules. Thus, the two control loops may be coupled, resulting in difficulties in the design of the loops. Here, we wish to propose a control strategy that decouples the two control loops. Referring to Fig. 2.1, and assuming the modules are buck-derived converters with transformer’s turns ratio being one, for the j-th module, the relationship between the input voltage vin_j and the output voltage vo_j is, vin_ j · dy_ j vo_ j

(2.23)

where d y_j is the duty cycle of the j-th module, and is composed of two components, i.e., dy_ j dvo_ j + dsh_ j

(2.24)

where d vo_j is produced by the system output voltage control loop and d sh_j is produced by the loop controlling the sharing of input voltages/currents or output voltages/currents. Perturbation and linearization of (2.23) leads to,

(2.25) Vin_ j + vˆin_ j · Dy + dˆvo_ j + dˆsh_ j Vo_ j + vˆo_ j where Dy , V in_j , and V o_j are the quiescent values, and dˆvo_ j , dˆsh_ j , vˆin_ j and vˆo_ j are the superimposed small AC perturbations. Neglecting the second-order nonlinear terms and noting that the DC terms on both sides of the equation are equal, we retain only the first-order AC terms on both sides of the equation, i.e.,

2.3 General Control Strategy for DC–DC Series–Parallel Power …

41

Vin_ j dˆvo_ j + Vin_ j dˆsh_ j + Dy vˆin_ j vˆo_ j

(2.26)

According to (2.26), we have, n

Vin_ j dˆvo_ j +

j1

n

Vin_ j dˆsh_ j + Dy ·

j1

n j1

vˆin_ j

n

vˆo_ j

(2.27)

j1

For input-parallel-connected systems, V in_j V in , and for input-series-connected system, V in_j V in /n. For output-parallel-connected systems, vˆo_ j vˆo , and for output-series-connected systems, vˆo_ j vˆo /n. Hence, (2.27) can be rewritten as n n Vin ˆ Vin ˆ m1 · dvo_ j + m 1 · dsh_ j + Dy · m 1 · vˆin m 2 · vˆo n j1 n j1

(2.28)

where m1 n and 1 for the input-parallel-connected and input-series-connected systems, respectively, and m2 n and 1 for the output-parallel-connected and outputseries-connected systems, respectively. If we have n

dˆsh_ j 0

(2.29)

j1

Equation (2.28) can be rewritten as m1 ·

n Vin ˆ dvo_ j + Dy · m 1 · vˆin m 2 · vˆo n j1

(2.30)

Equation (2.30) shows that if the sum of the duty cycle variations produced by the sharing control loops is zero, the system output voltage is only influenced by input voltage perturbation and the duty cycle perturbation produced by the system output voltage control loop. That is to say, the loop that controls the sharing of input voltages/currents or output voltages/currents has no effect on the system output voltage control loop. Thus, if condition (2.29) is satisfied, the two control loops will not interfere with each other. Despite this conclusion being drawn from a specific derivation process for buck-derived converters, it is also applicable to boost-derived or buck/boost-derived ones. According to (2.29), we have, dˆsh_n −

n−1

dˆsh_ j

(2.31)

j1

In practice, the duty cycle signal is produced by the intersection between the output signal of the regulator and saw tooth signal V RAMP , i.e.,

42

2 A General Control Strategy for DC–DC Series–Parallel Power …

vˆsh_EA j dˆsh_ j VRAMP

( j 1, 2, . . . , n)

(2.32)

where vˆsh_EA j is the output perturbation signal of the regulator for controlling the sharing of input voltages/currents or output voltages/currents. Combining (2.31) and (2.32), we have, vˆsh_EAn −

n−1

vˆsh_EA j

(2.33)

j1

According to (2.33), a general sharing control strategy for DC–DC series–parallel power conversion systems is proposed, as shown in Fig. 2.7. The system output voltage control loop ensures the desired system output voltage. vo_EA is the control signal derived from the output voltage regulator, which is used to produce common duty cycle signal d vo_j (j 1, 2, …, n) by intersecting with V RAMP . Input/output variables sharing control loops achieve the equilibrium among the constituent modules. The control signal V sh_g represents the average value of the total input/output voltages/currents, which needs to be shared among all the modules, and V sh_fj (j 1, 2, …, n) represents the sensed input or output variable of each module. For example, for input-series-connected systems, V sh_g is V in /n, and V sh_fj is the module input voltage. The representation of V sh_g and V sh_fj are shown in Table 2.2. Also, vsh_EAj (j 1, 2, …, n − 1) represents the control signal derived from the sharing control loop. For the first n − 1 modules, vsh_EAj is sent to the inverting input of PWM comparator to obtain the duty cycle signal d sh_j , for the module n, the sum of the first n − 1 vsh_EAj , which is denoted by vsh_EAn , is sent to the inverting input of PWM comparator to obtain the duty cycle signal d sh_n . It can be found that the proposed control strategy decouples the system output voltage regulating loop and input/output variables sharing loops. For the input-series-connected systems, the final duty cycle of module j is obtained by subtracting d sh_j from d vo_j (j 1, 2, …, n), while for input-parallel-connected systems, the final duty cycle of module j is the sum of d sh_j and d vo_j (j 1, 2, …, n).

Table 2.2 Representation of V sh_fj and V sh_g in four architectures Architecture

V sh_fj (j 1, 2, …, n − 1)

V sh_g

IPOP

I cin_j or I co_j

I in /n or I o /n

IPOS

I cin_j or V o_j

I in /n or V o /n

ISOP

V in_j

V in /n

ISOS

V in_j

V in /n

2.4 Modularization Architecture for DC–DC Series–Parallel Power …

Vsh_g

_ Vsh_f1 +

/+

Gsh

+

vsh_EA1 _ +

+

+

Input/output variables sharing control loops

Drive circuit 2

Gvo

vof Output voltage loop

vo_EA

+

Comparator (n−1) − RS Flip-flop +

VRAMP (n−1)

/+ vsh_EAn

_

Comparator 2 − RS Flip-flop +

VRAMP 2

/+

Gsh

+

Drive circuit 1

Vsh_f(n−1)

vsh_EA(n−1)

Voref

− +

RS Flip-flop

VRAMP 1

/+

vsh_EA2

+

Comparator 1

Vsh_f2 Gsh

_

43

Comparator n − + VRAMP n

RS Flip-flop

Drive circuit (n−1)

Drive circuit n

ISOP and ISOS /+ + IPOP and IPOS

Fig. 2.7 General sharing control strategy for DC–DC series–parallel power conversion systems

2.4 Modularization Architecture for DC–DC Series–Parallel Power Conversion Systems Among all the four series–parallel architectures, only IPOP has been successfully implemented in practice and commercially available in modular forms. In some previous designs of modular systems, e.g., those reported in [2–5], dedicated control methods are used for the constituent modules and the entire systems are not separable into standardized modules. In this section, the modularization architectures are proposed for IPOS, ISOP, and ISOS systems. The main features are: (a) the basic converter module has its own power stage and controller, and can operate in standalone mode; (b) no external controller is needed to achieve sharing of input/output voltages/currents; and (c) a relatively small number of wires are needed to connect up the modules to form the entire system. To achieve modularization, all power stages as well as controllers of all the modules should be identical. Hence, the sharing control signal vsh_EAn for the module n can be simplified as

44

2 A General Control Strategy for DC–DC Series–Parallel Power …

vsh_EAn −

n−1

vsh_EA j −

j1

n−1

Vsh_g − Vsh_f j G sh

j1

−G sh (n − 1)Vsh_g − (Vsh_f1 + · · · + Vsh_f(n−1) )

(2.34)

The average sharing control signal V sh_g is Vsh_g (Vsh_f1 + Vsh_f2 + · · · + Vsh_f(n−1) + Vsh_fn )/n

(2.35)

Substitution of (2.35) into (2.34), we have,

vsh_EAn −G sh (n − 1)Vsh_g − (n · Vsh_g − Vsh_fn ) G sh (Vsh_g − Vsh_fn ) (2.36) Based on (2.36), the control strategy depicted earlier in Fig. 2.7 can be redeveloped as shown in Fig. 2.8. Since the individual modules can be operated in stand-alone mode, each of them should have its own output voltage control loop. However, in practice, mismatches

Vsh_g

+

_ Vsh_f1

/+

Gsh

+

vsh_EA1

/+

Gsh

vsh_EA2 Input/output variables sharing control loops _

+

vsh_EA(n−1) _ Vsh_fn

+

Gvo vof Output voltage loop _

Drive circuit 2

Comparator (n−1)

VRAMP (n−1)

/+

Gsh

vsh_EA(n)

Voref

Comparator 2 − RS Flip-flop +

VRAMP 2

/+

Gsh

+

+

Vsh_f(n−1)

+

Drive circuit 1

VRAMP 1

_ Vsh_f2 +

Comparator 1 − RS Flip-flop +

+

− +

RS Flip-flop

Comparator n − RS Flip-flop +

VRAMP n

Drive circuit (n−1)

Drive circuit n

vo_EA ISOP and ISOS /+ + IPOP and IPOS

Fig. 2.8 Simplification of the general sharing control strategy for DC–DC series–parallel power conversion systems

2.4 Modularization Architecture for DC–DC Series–Parallel Power …

45

in the voltage references and output voltage sampling factors of the modules are inevitable. Thus, when the modules are connected to form a series–parallel power conversion system, we need to employ signal diodes to ensure only one output voltage control loop is active and regulates all the outputs of all modules. The final proposed modularization architecture is shown in Fig. 2.9. In Fig. 2.9, each dashed line block represents a self-contained module, which can operate in standalone mode. Taking module 1 as the example, when it is standalone, we have V sh_f1 V sh_g , i.e., vsh_EA1 is constant and the output voltage is regulated by its own output voltage control loop. Only two buses, namely, the input/output variables sharing bus and the common duty cycle bus, are needed when multiple modules are connected in one of the series–parallel architectures, and no external controller is required. It can be seen that the full modularization architecture for DC–DC series–parallel power conversion systems have the following advantages: (1) all modules are self-contained and standardized, (2) no extra control is needed to achieve sharing of voltage and/or current at the input and output sides, and (3) easy to expand system capacity.

Input/output variables sharing bus Vsh_g

R

Voref

R

ISOP and ISOS /+ + IPOP and IPOS

_ Vsh_f1 + + _

vsh_EA1

vof

+

Gvo vsh_EA2 /+

Gsh vof

Comparator 1 − RS Flip-flop +

VRAMP 1

Drive circuit 1 Module 1

_ Vsh_f2 +

+ Voref _

/+

Gsh

+

Gvo

Comparator 2 − RS Flip-flop +

VRAMP 2

Drive circuit 2 Module 2

Common duty cycle bus R

_ Vsh_f(n-1) +

+ Voref _

Gvo vof

_ Vsh_fn R + + Voref _

Gsh

vsh_EA(n-1) Comparator (n−1) /+ − RS Flip-flop + + VRAMP (n−1)

vof

Drive circuit (n−1) Module (n−1)

-vsh_EA(n) Gsh Gvo

/+ +

Comparator n − + VRAMP n

RS Flip-flop

Drive circuit n Module n

Fig. 2.9 Modularization architecture for DC–DC series–parallel power conversion systems

46

2 A General Control Strategy for DC–DC Series–Parallel Power …

From Fig. 2.9, it can be seen that each module regulates its own output voltage rather than system output voltage. For output-parallel-connected systems, the module output voltage is also the system output voltage, which will be precisely regulated, while for the output-series-connected systems, only one module output voltage is precisely regulated and the other module are indirectly regulated, leading to not very precise system output voltage. It should be noted that for IPOS system, the output voltage sharing bus and the common duty cycle bus are not mandatory because the system’s total output voltage will be well regulated if individual module manage to regulate its own output voltage; and for IPOP system, the common duty cycle bus is also unnecessary, and if it is removed, the current sharing method is similar to that proposed in [6], in which current-mode control is adopted as an inner current loop.

2.5 Experimental Verification In order to verify the effectiveness of the proposed general control strategy, we have constructed four DC–DC power conversion systems, i.e., IPOP, IPOS, ISOP, and ISOS systems. Each system consists of three basic converter modules, as shown in Fig. 2.10. The basic converter module is a phase-shifted full-bridge converter with the following specifications: • • • •

input voltage V in : 270 (±10%) Vdc; output voltage V o : 54 Vdc; rated output current I o : 20 A; switching frequency f s : 100 kHz. The power stage consists of the following devices and components:

(1) (2) (3) (4) (5)

switching device: IRFP460 (International Rectifier); clamping diodes: DSEI30-06 A (IXYS); rectifier diodes: APT60D30B (Advanced Power Technology); resonant inductance: 6 µH (Ferroxcube, E32/16/9, 8 turns); transformer: Ferroxcube, E42/21/20, primary winding is 15 turns, and secondary winding is 4 turns × 2 (center-tap full-wave rectifier); (6) output filter inductor: 30 µH (Ferroxcube, E42/21/15, 15 turns); (7) output filter capacitor: 4400 µF (2200 µF/100 V × 2—electrolytic). When the input sides and output sides of the three modules are connected in parallel, respectively, an IPOP system is constructed with the following specifications: system input voltage 270 Vdc ± 10%, system output voltage 54 Vdc, system rated output current 60 A (20 A × 3). Figure 2.11 shows the experimental waveforms of the IPOP system corresponding to a stepped load test and a stepped input voltage test. The individual output currents are sensed to provide information for the output current sharing bus. Figure 2.11a shows the output currents of the three modules and system’s output voltage (ac

2.5 Experimental Verification

47

Fig. 2.10 Prototype of DC–DC series–parallel power conversion systems consisting of three basic modules

coupled) when the system load changes between full load (60 A) and half load (30 A). We observe here that the output currents of the three modules are well shared. Figure 2.11b shows the waveforms when the input voltage steps up and down between 250 and 280 Vdc. Again we see that the output currents are well shared both at steady state and during transient. In order to verify that the OCS control is decoupled from the output voltage control, we use two sets of control parameters. Figures 2.11c, d show the dynamic responses with the same OCS control parameters but different output voltage control parameters. It can be seen that the output voltage responses differ. Figures 2.11e, f show the dynamic response with the same output voltage control parameters but different OCS control parameters. It can be seen that the output voltage response remains the same. In the prototype, the module duty cycle will be regulated fiercely during transient if the bandwidth of the OCS closed loop is high, which is easy to cause system oscillation. Hence, the bandwidth of the OCS closed loop is set to be relatively low in the prototype and the waveforms of output currents have little difference with different OCS control parameters.

48

2 A General Control Strategy for DC–DC Series–Parallel Power … C2 Io:[30A/div]

C1 Vin:[100V/div]

C2

M4 Io_1:[20A/div]

C1

M4

C4 Io_2:[20A/div]

M3

C4

C3 Io_3:[20A/div]

C3

C3 Io_2:[20A/div] C4 Io_3:[20A/div]

C4

C3 C1

M3 Io_1:[20A/div]

C1 Vo:[5V/div]

C2 Vo:[5V/div]

C2

Time:[20ms/div]

Time:[20ms/div]

(a) Response for stepped load

(b) Response for stepped input voltage C1Vin:[100V/div]

C2 Io:[30A/div]

C2

M4 Io_1:[20A/div]

M3

M4 C4 M1 C1

M3 Io_2:[20A/div]

C1

C4 Io_1:[20A/div] M1 Vo:[5V/div]

1# OCSR, 1# OVR

1# OCSR, 2# OVR

C1 Vo:[5V/div]

C3 Io_2:[20A/div]

C3 M2

M2 Vo:[5V/div]

C2

C2 Vo:[5V/div]

Time:[20ms/div]

Time:[20ms/div]

(c) Dynamic responses with same OCS parameters but different OVR parameters

(d) Dynamic responses with same OCS parameters but different OVR parameters C1 Vin:[100V/div]

C2 Io:[30A/div]

C2

M4 Io_1:[20A/div]

M3 Io_2:[20A/div]

C1 M3

M4 C4

1# OCSR, 1# OCSR, 1# OVR 2# OVR

C4 Io_1:[20A/div]

M1

M1 Vo:[5V/div]

C1

C1 Vo:[5V/div]

1# OCSR, 1# OVR

2# OCSR, 1# OVR

Time:[20ms/div]

(e) Dynamic responses with same OVR

C3 Io_2:[20A/div]

C3 M2

M2 Vo:[5V/div]

C2

C2 Vo:[5V/div]

1# OCSR, 1# OVR

2# OCSR, 1# OVR

Time:[20ms/div]

(f) Dynamic responses with same OVR

parameters but different OCSparameters

parameters but different OCS parameters

Fig. 2.11 Experimental waveforms for IPOP system

If the three modules are connected in parallel at the input sides and in series at the output sides, we have an IPOS system with the following specifications: system input voltage 270 Vdc ± 10%, system output voltage 162 Vdc (54 Vdc × 3), system rated output current 20 A.

2.5 Experimental Verification

49 C4 Vin:[100V/div]

C4 Io:[10A/div]

C4

C4

C2 Vo_2:[25V/div]

C1

C1 Vo_1:[5V/div]

C2

C2 Vo_2:[5V/div]

M3

M3 Vo_3:[5V/div]

C2

C3

C3 Vo:[10V/div]

M3

Time:[20ms/div]

C1 Vo_1:[25V/div]

C1

M3 Vo_3:[25V/div] C3 Vo:[100V/div]

C3 Time:[20ms/div]

(a) Response for stepped load

(b) Response for stepped input voltage

Fig. 2.12 Experimental waveforms for IPOS system

Figure 2.12 shows the experimental waveforms of the IPOS system corresponding to stepped load and stepped input voltage. The output voltage sharing bus and the common duty cycle bus are removed in the prototype system. Figure 2.12a shows the output voltages of the three modules (AC coupled) corresponding to a load stepping between full load (20 A) and half load (10 A), and Fig. 2.12b shows the waveforms of individual output voltages and the system’s total output voltage corresponding to an input voltage stepping between 250 and 280 Vdc. It can be seen that the output voltages are well shared both at steady state and during transient and the system’s output voltage is also well regulated. If the three modules are connected in series at the input sides and in parallel at the output sides, we have an ISOP system with the following specifications: system input voltage 810 Vdc ± 10% (270 Vdc × 3), system output voltage 54 V, system rated output current 60 A (20 A × 3). Figure 2.13 shows the experimental waveforms of the ISOP system corresponding to stepped load and stepped input voltage. Figure 2.13a shows the input voltages of the three modules and the system’s output voltage (AC coupled) corresponding to a load stepping between full load (60 A) and half load (30 A). Here, it is verified that the input voltages of the three modules are well shared. Figure 2.13b shows the waveforms corresponding to an input voltage stepping between 740 and 840 V. It can be seen that input voltages are well shared both at steady state and during transient, and the sharing of the output current is achieved automatically. In order to verify that the IVS control is decoupled from the output voltage control, we use two sets of control parameters. Figures 2.13c, d show the dynamic responses with the same IVS control parameters but different output voltage control parameters. It can be seen that the output voltage responses differ. Figures 2.13e, f show the dynamic response with the same output voltage control parameters but different IVS control parameters. It can be seen that the output voltage response remains the same. Similarly, the module duty cycle will be regulated violently during transient if the bandwidth of the IVS closed loop is high, which is easy to cause system oscillation. Hence, the bandwidth

50

2 A General Control Strategy for DC–DC Series–Parallel Power … C4 Vin:[500V/div] C4 Io:[30A/div]

C4 C1 C2

C4

C1Vin_1:[200V/div]

C1 Vin_1:[200V/div]

C2Vin_2:[200V/div]

C2 Vin_2:[200V/div]

C1

M3Vin_3:[200V/div]

M3Vin_3:[200V/div]

C2 M3

M3 C3

C3Vo:[5V/div]

C3 Vo:[5V/div]

C3

Time:[20ms/div]

Time:[20ms/div]

(a) Response for stepped load

(b) Response for stepped input voltage C4 Vin:[500V/div]

C4 Io:[30A/div]

C4

M2 Vin_2:[200V/div]

M1 Vin_1:[200V/div]

C4

C2 Vin_2:[200V/div]

C1 Vin_1:[200V/div]

M1 C1 M3

M3 Vo:[5V/div]

C3

C3 Vo:[5V/div]

M2

1# IVSR, 1# OVR

1# IVSR, 2# OVR

C2 M3

M3 Vo:[5V/div]

C3

C3 Vo:[5V/div]

1# IVSR, 1# OVR

1# IVSR, 2# OVR

Time:[20ms/div]

Time:[20ms/div]

(c) Dynamic responses with same IVS parameters but different OVR parameters

(d) Dynamic responses with same IVS parameters but different OVR parameters C4 Vin:[500V/div]

C4 Io:[30A/div] C4 M1

C4

M2 Vin_2:[200V/div]

M1 Vin_1:[200V/div]

C2 Vin_2:[200V/div]

C1Vin_1:[200V/div]

C1 M3

M3 Vo:[5V/div]

C3

C3 Vo:[5V/div]

M2

1# IVSR, 2# IVSR, 1# OVR 1# OVR

Time:[20ms/div]

(e) Dynamic responses with same OVR parameters but different IVSparameters

C2 M3 Vo:[5V/div]

M3

1# IVSR, 1# OVR

2# IVSR, 1# OVR

C3 Vo:[5V/div]

C3

Time:[20ms/div]

(f) Dynamic responses with same OVR parameters but different IVSparameters

Fig. 2.13 Experimental waveforms for the ISOP system

of the IVS closed loop is set to be relatively low in the prototype and the waveforms of input voltages have little difference with different IVS control parameters. When the input sides and output sides of the three modules are connected in series respectively, we have the ISOS system with the following specifications: system input

2.5 Experimental Verification

51

voltage 810 Vdc ± 10% (270 Vdc × 3), system output voltage 162 Vdc (54 Vdc × 3), system rated output current 20 A. Figure 2.14 shows the experimental waveforms of the ISOS system corresponding to stepped load and stepped input voltage. Figures 2.14a, b show the input voltage of the three modules and the system output voltage corresponding to a load stepping between full load (20 A) and half load (10 A) and an input voltage stepping between 740 and 840 Vdc, respectively. It can be seen that input voltages are shared both at steady state and during transient, and sharing of the output voltage is achieved automatically. In order to verify that the IVS control is decoupled from the output voltage control, we use two sets of control parameters. Figures 2.14c, d show the dynamic responses with the same IVS control parameters but different output voltage control parameters. It can be seen that the output voltage responses differ. Figures 2.14e, f show the dynamic response with the same output voltage control parameters but different IVS control parameters. It can be seen that the output voltage response remains the same.

2.6 Summary This chapter systematically analyzed the inherent relationships between the sharing of input voltages/currents and the sharing of output voltages/currents of the DC–DC series–parallel power conversion systems from the view of power conservation. It is pointed out that the sharing of input voltages/currents ensures the sharing of output voltages/currents and vice versa. For input-parallel-connected systems, the control strategies intending to achieve ICS or OCS (for IPOP) or OVS (for IPOS) can be used, however; the control strategies intending to achieve OCS (for ISOP) or OVS (for ISOS) for input-series-connected systems are unstable, and only the control strategies intending to achieve IVS can be adopted. A general control strategy was proposed for DC–DC series–parallel power conversion systems, which not only achieves equilibrium between the constituent modules but also decouples the output voltage control loop and the sharing control loop. Furthermore, modularization architecture for DC–DC series–parallel power conversion systems was proposed, achieving full modularization where all modules are selfcontained and standardized, and no extra control is needed to achieve sharing of voltage and/or current at the input and output sides. A prototype of DC–DC series–parallel power conversion systems consisting of three converter modules provided experimental results to verify the general control strategy and modularization architectures.

52

2 A General Control Strategy for DC–DC Series–Parallel Power …

C4 Vin:[500V/div]

C4 Io:[10A/div]

C4 C1 Vin_1:[200V/div]

C1 Vin_1:[200V/div]

C4

C2 Vin_2:[200V/div]

C2 Vin_2:[200V/div]

C1

C1 M3Vin_3:[200V/div]

M3 Vin_3:[200V/div]

C2

C2 M3

M3 C3 Vo:[10V/div]

C3

C3Vo:[100V/div]

C3 Time:[20ms/div]

Time:[20ms/div]

(a) Response for stepped load

(b) Response for stepped input voltage C4 Vin:[500V/div]

C4 Io:[10A/div] C4 M1

C4 M1Vin_2:[200V/div]

M1 Vin_1:[200V/div]

C1 Vin_2:[200V/div]

C1 Vin_1:[200V/div]

C1 M3

M3 Vo:[10V/div]

C3

C3 Vo:[10V/div]

M1 1# IVSR, 1# OVR

1# IVSR, 2# OVR

C1 M3 C3

M3 Vo:[10V/div]

1# IVSR, 1# OVR

1# IVSR, 2# OVR

C3 Vo:[10V/div] Time:[20ms/div]

Time:[20ms/div]

(c) Dynamic responses with same IVS

(d) Dynamic responses with same IVS

parameters but different OVR parameters

parameters but different OVR parameters C4 Vin:[500V/div]

C4 Io:[10A/div] C4 M1

C4 M1Vin_2:[200V/div]

M1 Vin_1:[200V/div] C1 Vin_1:[200V/div]

C1 M3

M3Vo:[10V/div]

C3

C3Vo:[10V/div]

M1 1# IVSR, 1# OVR

2# IVSR, 1# OVR

C1Vin_2:[200V/div]

C1 M3

M3 Vo:[10V/div]

C3

C3 Vo:[10V/div]

Time:[20ms/div]

(e) Dynamic responses with same OVR parameters but different IVS parameters Fig. 2.14 Experimental waveforms for ISOS system

1# IVSR, 1# OVR

2# IVSR, 1# OVR

Time:[20ms/div]

(f) Dynamic responses with same OVR parameters but different IVS parameters

References

53

References 1. Erickson, R., Maksimovic, D.: Fundmental of Power Electronics. Klurwer Academic, New York (2001) 2. Kim, J., Choi, H., Cho, B.H.: A novel droop method for converter parallel operation. IEEE Trans. Power Electron. 17(1), 25–32 (2002) 3. Ayyanar, R., Giri, R., Mohan, N.: Active input-voltage and load-current sharing in input-series and output-parallel connected modular DC-DC converters using dynamic input-voltage reference scheme. IEEE Trans. Power Electron. 19(6), 1462–1472 (2004) 4. Ruan, X., Chen, W., Cheng, L., Tse, C.K., Yan, H., Zhang, T.: Control strategy for input-seriesoutput-parallel converters. IEEE Trans. Ind. Electron. 56(4), 1174–1185 (2009) 5. Giri, R., Ayyanar, R., Ledezma, E.: Input-series and output-series connected modular DC-DC converters with active input voltage and output voltage sharing. In: Proceedings of the IEEE Applied Power Electronics Conference and Exposition, pp. 1751–1756 (2004) 6. Lin, C., Chen, C.: Single-wire current share paralleling of current mode-controlled dc power supplies. IEEE Trans. Ind. Electron. 47(4), 780–786 (2000)

Chapter 3

Mathematical Model and Closed-Loop Parameters Design for DC–DC ISOP System

Abstract This chapter derives the small-signal model of the phase-shifted fullbridge (PSFB)-based ISOP system, and based on the small-signal model, the general control strategy proposed in Chap. 2 is proved to be decoupled into n − 1 input voltage sharing (IVS) closed loops and one output voltage closed loop. The parameters design of the IVS control loop regulator and output voltage control loop regulator for a twomodule PSFB ISOP system is discussed and verified by the prototype. When the inner current loop is introduced into the general control strategy, the stability of the ISOP system with/without the IVS control loops is discussed and the root cause of system instability is also revealed. Furthermore, the design criteria of the IVS control loop regulator to keep ISOP system stable are derived. Keywords Small-signal model · Phase-shifted full-bridge converter Input voltage sharing · Parameter design · Inner current loop · Stability In Chap. 2, a general control strategy was proposed for DC–DC series–parallel power conversion systems, which not only achieves equilibrium between the constituent modules but also decouples the output voltage control loop and the sharing control loop. In this chapter, the small-signal model of ISOP system consisting of multiple phase-shifted full-bridge (PSFB) converters is derived to prove the decoupling of the system output voltage control loop and IVS control loop, and the design of the parameters of the two control loops is given. Furthermore, in order to improve the system dynamic response and realize output current short-circuit protection, the inner current loop is incorporated into the general control strategy, and the parameters design of the IVS control loop will be analyzed.

© Springer Nature Singapore Pte Ltd. and Science Press, Beijing, China 2019 X. Ruan et al., Control of Series-Parallel Conversion Systems, CPSS Power Electronics Series, https://doi.org/10.1007/978-981-13-2760-5_3

55

56

3 Mathematical Model and Closed-Loop Parameters Design …

3.1 Small-Signal Model of PSFB-Based ISOP System 3.1.1 Small-Signal Model of PSFB Converter The circuit configuration of the full-bridge converter is shown in Fig. 3.1, where Q1 ~ Q4 (including their body diodes D1 ~ D4 and their intrinsic capacitors C 1 ~ C 4 ) are four power switches, T r is the high-frequency isolation transformer with primaryand-secondary-winding ratio K, L r is the resonant inductor, DR1 and DR2 are two output rectifier diodes, L f is the output filter inductor, C f is the output filter capacitor, and RLd is the load. Figure 3.2 shows the key waveforms of the PSFB converter [1], where the switching process is ignored for simplicity. During [t 0 , t 1 ], Q1 and Q2 are conducting and the primary current ip is flowing through them in freewheeling state. In the secondary side, DR2 is conducting and the rectified voltage vrect is zero. At t 1 , Q2 is turned off and Q4 is turned on, then Q1 and Q4 conduct, and vAB is equal to V in . Both the primary and secondary winding currents decay and the secondary winding current is lower than the output filter inductor current, thus both DR1 and DR2 conduct, setting both the primary and secondary winding voltages to zero. The input voltage V in is applied on the resonant inductor L r , forcing ip to increase linearly. At t 2 , ip increases to the value of the output filter inductor current iLf reflected to the primary side, i.e., ip iLf /K. The rectifier diode DR2 turns off and DR1 continues conducting, vrect rises to the value of input voltage reflected to the secondary side, i.e., vrect V in /K, both ip and iLf increase linearly. At t 3 , Q1 is turned off and Q3 is turned on, ip is in freewheeling stage and vrect 0, starting the second half-switching period. As shown in Fig. 3.2, during [t 1 , t 2 ], the primary current ip increases from −I 2 to I 1 , the rectified voltage vrect is zero although vAB = V in . This means that the secondary side loses the input voltage reflected to the secondary during [t 1 , t 2 ]; in other words, the secondary effective duty cycle Deff is smaller than the primary duty cycle Dy [2]. Deff can be expressed as Deff Dy − Dloss

Q1 A

Q3

C1

Lr

ip

D 3 C3

Fig. 3.1 Full-bridge converter

Tr

Q2

Q4

DR1

Tr D2

C2

B

*

Vin

D1

(3.1)

D 4 C4

Lf + vrect iLf Cf _

* * DR2

RLd + Vo _

3.1 Small-Signal Model of PSFB-Based ISOP System

57

where Dloss is the secondary duty cycle loss. According to Fig. 3.2, we have [3] Dloss

t2 − t1 I1 + I2 Vin 1 Ts /2 · L r 2 fs

(3.2)

where T s is the switching period and f s 1/T s is the switching frequency. Referring to Fig. 3.2, the primary current ip at t 2 is 1 I Io − I1 K 2

(3.3)

where I o is the output current and I is the ripple current of the output filter inductor current. At t 4 , the primary current ip is equal to I 2 and it can be expressed as 1 − Dy 1 I (3.4) Io + − Vo I2 K 2 2 fs L f Substitution of (3.3) and (3.4) into (3.2) yields 1 − Dy 1 Dloss Vo 2Io − Vin 1 2 fs L f K L r 2 fs

vAB ip

(3.5)

I2

I1 Vin

t

0

Vin I2 DyTs/2 DeffTs/2 vrect

Vin/K t

0

I

iLf

Io t

0

t0

t1 t2

Fig. 3.2 Key waveforms of the PSFB converter

t3

t4 t5

t6

58

3 Mathematical Model and Closed-Loop Parameters Design …

Due to the fact that the PSFB converter is a buck-derived topology, its secondary effective duty cycle d eff is the summation of steady-state component Deff and transient component dˆeff , i.e., deff Deff + dˆeff

(3.6)

The transformer secondary rectified voltage vrect depends not only on the primary duty cycle Dy but also on the resonant inductor L r , the switching frequency f s , output filter inductor current iLf , and the input voltage V in . Therefore, the smallsignal transfer functions of the converter will depend on L r , f s , and the perturbations of the output filter inductor current iˆLf , input voltage, vˆin , and the primary duty cycle dˆy . To accurately model the dynamic behavior of the PSFB converter, it is necessary to find out the contributions of L r , f s , vˆin , iˆLf , and dˆy to dˆeff . These effects can be incorporated in the small-signal model of buck converter to obtain the model of the PSFB converter.

3.1.1.1

Effect of the Change of the Output Filter Inductor Current on Duty Cycle Modulation

Figure 3.3 shows the key waveforms of the PSFB converter when the output filter inductor current iLf is perturbed, where the solid lines represent the steady-state operation and the dashed line represents the primary current caused by an increase of iLf by the amount iˆLf . The perturbed primary current will reach the reflected filter inductor current at a later time than it would in the steady-state operation, and the additional delay time t can be expressed as t 2

iˆLf L r K Vin

(3.7)

Then, the change of d eff due to iˆLf , denoted as dˆi , is 4L r f s ˆ t − dˆi − i Lf Ts /2 K Vin

(3.8)

The negative sign in (3.8) represents that there will be a reduction in d eff if the output filter inductor current is increased. It should be noted that the primary duty cycle keeps unchanged in the perturbation.

3.1.1.2

Effect of the Change of the Input Voltage on Duty Cycle Modulation

Figure 3.4 shows the key waveforms of the PSFB converter when the input voltage V in is perturbed, where the solid lines represent the steady-state operation and the dashed

3.1 Small-Signal Model of PSFB-Based ISOP System

59

Fig. 3.3 Duty cycle modulation due to the change of the output filter inductor current Fig. 3.4 Duty cycle modulation due to the change of input voltage

line represents the primary current caused by an increase of V in by the amount vˆin . The perturbed primary current will reach the reflected filter inductor current sooner than it would in the steady-state operation, and the shortened time t can be expressed as

60

3 Mathematical Model and Closed-Loop Parameters Design …

1 2Io − K 1 2Io − K 1 2Io − ≈ K

t

1 − Dy Lr Lr Vo − 2 fs L f Vin Vin + vˆin 1 − Dy Lr vˆin Vo 2 fs L f Vin (Vin + vˆin ) 1 − Dy Lr Vo vˆin 2 fs L f Vin2

Then, the change of d eff due to vˆin , denoted as dˆv , is 1 − Dy 4L r f s t Vo vˆin dˆv Io − Ts /2 4 fs L f K Vin2

(3.9)

(3.10)

In practice, a relatively larger output filter inductance is usually used to reduce its current ripple; moreover, the primary duty cycle Dy is always made as large as possible to minimize the conduction loss caused by the circulating current in the primary side. Therefore, the second term in the bracket of (3.10) is much less than I o and (3.10) can be simplified as [2] 4L r f s Io vˆin dˆv K Vin2 The output current I o can be expressed as Io rewritten as 4L r f s Deff vˆin dˆv 2 K Vin RLd

(3.11) 1 Vin Deff and (3.11) can be K RLd

(3.12)

Equation (3.12) indicates that the effect of vˆin on d eff can be interpreted as an additional feed-forward of input voltage. The results of the previous analysis can be incorporated in the small-signal model of the buck converter, as shown in Fig. 3.5. The small-signal model of the PSFB converter can be derived by replacing dˆy in Fig. 3.5 by dˆeff , and dˆeff dˆy + dˆi + dˆv , as shown in Fig. 3.6. The contribution of dˆi and dˆv is represented by two controlled sources and the contribution of dˆv by two independent sources. This is to emphasize that dˆi and dˆv originate from the circuit itself (i.e., iˆLf and vˆin ) and are not controlled by the controller. It can be found that if L r 0, we have dˆi 0 and dˆv 0 according to (3.8) and (3.12), and if K = 1, the small-signal model of the PSFB converter is equivalent to that of the buck converter; in other words, the buck converter is a special case of the PSFB converter.

3.1 Small-Signal Model of PSFB-Based ISOP System

61

ˆ 1: Deff Vin d y

Fig. 3.5 Small-signal model of the buck converter

Lf

+ _

RLd +

vˆin

+ _

*

*

Vin dˆy RLd

Rcf Cf

vˆo _

Fig. 3.6 Small-signal model of the PSFB converter

3.1.2 Small-Signal Model of the PSFB-Based ISOP System The circuit architecture of the PSFB-based ISOP system is shown in Fig. 3.7, which consists of n PSFB converters connected in series at the input sides and in parallel at the output sides. At steady state, the input voltage and output current of each PSFB converter are V in /n and I o /n, respectively; thus, the load resistor of each converter is nRLd . Using the small-signal model of the PSFB converter, the small-signal model of the PSFB converter in ISOP system can be derived, as shown in Fig. 3.8. As shown in Fig. 3.8, we have the following equation:, Vin Vin /n 2 K · (n RLd ) n K RLd 4L r f s ˆ 4n L r f s ˆ i Lf j − i Lf j ( j 1, 2, . . . , n) dˆi j − K (Vin /n) K Vin 4L r Deff f s 4L r Deff f s vˆin j 2 2 vˆin j ( j 1, 2, . . . , n) K (Vin /n) · (n RLd ) K Vin RLd Ieq

dˆv j

(3.13.1) (3.13.2) (3.13.3)

where j represents module j and all the resonant inductors and turns ratios are assumed to be identical, respectively, i.e., L r1 L r2 ··· L rn L r , and K 1 K 2 ··· K n K. Using the small-signal model of the PSFB converter shown in Fig. 3.8, the smallsignal mode of the PSFB-based ISOP system can be obtained, as shown in Fig. 3.9.

62

3 Mathematical Model and Closed-Loop Parameters Design …

Lf1

iin1 Iin icd1 + vin1 _ C A1

Lr1

Tr1 B1

*

ip1

d1

Tr1 *

+ vrect1 iLf1 _

iinn

Vin

Cf

*

Lfn Lrn ipn

dn

Trn Bn

*

icdn + vinn _ C An

Trn *

+ vrectn iLfn _

*

Fig. 3.7 Circuit architecture of the PSFB-based ISOP system

Vin ˆ Vin ˆ d yj dij + dˆ vj nK nK 1: Deff /K

(

*

vˆinj

*

(

I eq dˆij + dˆ vj

)

)

iˆLfj

Lf

I eq dˆyj

nRLd + Rcf

vˆo

Cf

_

Fig. 3.8 Small-signal model of the PSFB converter in ISOP system

According to Fig. 3.9, we have ⎧ ⎪ Deff Vin ˆ ⎪ ˆv1 + dˆy1 s L f · iˆLf1 + vˆo ⎪ v ˆ d + + d in1 i1 ⎪ ⎪ nK ⎪ ⎪ K ⎪ ⎪ ⎪ Vin ˆ D eff ⎪ ⎨ vˆin2 + di2 + dˆv2 + dˆy2 s L f · iˆLf2 + vˆo K nK ⎪ ⎪ .. ⎪ ⎪ . ⎪ ⎪ ⎪ ⎪ ⎪ V D eff in ⎪ ⎪ vˆinn + dˆin + dˆvn + dˆyn s L f · iˆLfn + vˆo ⎩ K nK

(3.14.1) (3.14.2) .. . (3.14.n)

RLd + Vo _

3.1 Small-Signal Model of PSFB-Based ISOP System

Vin ˆ V d y1 in dˆ + dˆ i1 v1 nK 1: Deff /K nK

(

+ * vˆin1 _ Cd1

(

*

I eq dˆi1 + dˆ v1

63

)

)

I eq dˆy1

(

vˆin

(

*

I eq dˆi2 + dˆ v2

)

)

*

(

*

I eq dˆin + dˆ vn

vˆo _

iˆLf2 Lf2

)

)

……

(

Cdn

Rcf

I eq dˆy2

Vin ˆ Vin ˆ ˆ d yn din + d vn nK nK 1: Deff /K + vˆinn _

RLd +

Cf

Vin ˆ Vin ˆ ˆ d y2 di2 + d v2 nK nK 1: Deff /K + * vˆin2 _ Cd2

iˆLf1 Lf1

iˆLfn

Lfn

I eq dˆyn

Fig. 3.9 Small-signal model of the PSFB-based ISOP system

⎧ K ˆ ⎪ ⎪ i − sC v ˆ ⎪ in d in1 ⎪ ⎪ Deff ⎪ ⎪ ⎪ ⎪ K ˆ ⎪ ⎪ ⎨ i in − sCd vˆin2 Deff ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ ⎪ K ˆ ⎪ ⎪ i in − sCd vˆinn ⎩ Deff

Ieq dˆy1 + dˆi1 + dˆv1 + iˆLf1

(3.15.1)

Ieq dˆy2 + dˆi2 + dˆv2 + iˆLf2

(3.15.2)

.. .

Ieq dˆyn + dˆin + dˆvn + iˆLfn n

vˆin j vˆin

.. . (3.15.n)

(3.16)

j1 n

j1

iˆLf j

vˆo

1 + Rcf //RLd sCf

(3.17)

It should be noted that in the above equations we made the following assuming: (1) L f1 L f2 … L fn L f , and (2) C d1 C d2 … C dn C d .

64

3 Mathematical Model and Closed-Loop Parameters Design …

The transfer functions and output impedance of the PSFB-based ISOP system can be derived with the aforementioned equations.

3.1.2.1

Control-to-Output Transfer Function

First, we consider the control-to-output transfer function of module j, and set vˆin 0, and dˆyk 0 (k = 1, 2, …, n, and k j). Then, adding up all the equations in (3.14) and combining (3.13.2), (3.13.3), (3.16), and (3.17), we obtain vˆo G vd j (s) vˆin 0 dˆy j dˆyk 0 (k1,2,...,n,

k j)

Vin (1 + sCf Rcf ) nK

Rp Rcf Lf Rcf +s + nCf Rcf + s 2 L f Cf 1 + + Rp C f 1 + +n RLd RLd RLd RLd G vd (s) ( j 1, 2, . . . , n)

(3.18)

where Rp 4L r f s /K 2 .

3.1.2.2

Control-to-Module-Input Transfer Function

To obtain the control-to-module-input transfer function, we first subtract (3.14.2) from (3.14.1) to obtain Vin Rp Deff vˆin1 − vˆin2 + 1+ dˆy1 − dˆy2 s L f + Rp iˆLf1 − iˆLf2 K n RLd nK (3.19) Moreover, subtracting (3.15.2) from (3.15.1) gives

Rp Vin ˆ dy1 − dˆy2 − 1 iˆLf1 − iˆLf2 2 n RLd n K RLd Rp Deff s K Cd vˆin1 − vˆin2 + + 2 2 Deff n K RLd

(3.20)

Putting (3.20) into (3.19) yields vˆin2 − vˆin1 A(s) dˆy2 − dˆy1 2 Deff n RLd Vin + s L f RLd Vin . where A(s) − 2 2 2 2 2 s Rp + s 2 L f n Deff RLd + s Deff L f Rp + n 2 K 2 Cd RLd

(3.21)

3.1 Small-Signal Model of PSFB-Based ISOP System

65

Likewise, we can obtain vˆin( j+1) − vˆin j A(s) dˆy( j+1) − dˆy j ( j 1, 2, . . . , n − 1)

(3.22)

From (3.22), we can derive vˆin j vˆin1 − A(s) dˆy1 − dˆy j ( j 1, 2, . . . , n)

(3.23)

Letting vˆin 0 and substituting (3.23) into (3.16), we obtain A(s) ˆ dy j n j1

(3.24)

A(s) ˆ dy j ( j 1, 2, . . . , n) n j1

(3.25)

vˆin1 A(s)dˆy1 −

n

Substitution of (3.24) into (3.23) yields vˆin j A(s)dˆy j −

n

Equation (3.25) can be rewritten in the following matrix form: ⎤⎡ ⎤ ⎡ ⎤ ⎡ A(s)(n−1) A(s) A(s) ˆy1 − . . . − d n n n vˆin1 ⎥⎢ ⎥ ⎢ ⎥ ⎢ ⎢ A(s) ⎥⎢ ˆ ⎥ A(s)(n−1) ⎢ ⎥ ⎥ ⎢ − A(s) ⎢ vˆin2 ⎥ . . . − d n n n ⎥⎢ y2 ⎥ ⎥ ⎢ ⎢ ⎢ ⎥⎢ ⎥ ⎢ . ⎥ .. .. .. .. ⎥⎢ .. ⎥ ⎢ ⎢ . ⎥ ⎥⎢ . ⎥ ⎢ ⎣ . ⎦ . . . . ⎦⎣ ⎦ ⎣ vˆinn vˆ 0 − A(s) − A(s) . . . A(s)(n−1) dˆyn in n

3.1.2.3

n

(3.26)

n

Input-Voltage-to-Output-Voltage Transfer Function

To find the input-voltage-to-output-voltage transfer function, we let dˆy j 0 (j = 1, 2, …, n). Then, adding up all the equations in (3.14) and combining (3.13.2), (3.13.3), (3.16), and (3.17), we can obtain vˆo G vg (s) vˆin dˆy j 0 ( j1,2,...,n) Rp Deff 1+ (1 + sCf Rcf ) K n RLd

Rp Rcf Lf Rcf +s + nCf Rcf + s 2 L f Cf 1 + + Rp C f 1 + +n RLd RLd RLd RLd (3.27)

66

3 Mathematical Model and Closed-Loop Parameters Design …

Fig. 3.10 Small-signal model representing the variation of output current

n

∑ iˆ

Lfj

j =1

PSFBbased ISOP System

3.1.2.4

Rcf Cf

RLd + vˆo _

iˆo

System-Input-to-Module-Input Transfer Function

We let dˆy j 0 (j = 1, 2, …, n) and use (3.23) to obtain vˆin1 vˆin2 · · · vˆinn

(3.28)

Substitution of (3.28) into (3.16) yields vˆin1 vˆin2 · · · vˆinn vˆin /n which can be rewritten as vˆin j G vin jg (s) vˆin dˆy j 0 3.1.2.5

( j1,2,...,n)

1 G ving (s) ( j 1, 2, . . . , n) n

(3.29)

(3.30)

System’s Output Impedance

The output impedance of the system can be readily found in Fig. 3.10. First, we observe that n

j 1

iˆLf j + iˆo

vˆo

1 + Rcf //RLd sCf

(3.31)

We let vˆin 0 and dˆy j 0 (j = 1, 2, …, n). Then, adding up all the equations in (3.14) and using (3.13.2), (3.13.3), (3.16), and (3.31) to obtain vˆo Z out (s) vˆin 0 iˆo dˆy j 0 ( j1,2,..., n) (3.32) s L f + Rp (1 + sCf Rcf )

Rp Rcf Lf Rcf +s + nCf Rcf + s 2 L f Cf 1 + + Rp C f 1 + +n RLd RLd RLd RLd

3.1 Small-Signal Model of PSFB-Based ISOP System

3.1.2.6

67

Output-Current-to-Module-Input Transfer Function

Putting vˆin 0 and dˆy j 0 (j = 1, 2, …, n) into (3.26), we can readily see that all the divided capacitor voltages are unrelated to the output current, i.e., vˆin j 0 (3.33) G vin j io (s) vˆin 0 iˆo dˆy j 0 ( j1,2,...,n)

3.2 Decoupling the Control Loops of the PSFB-Based ISOP System In Chap. 2, a general control strategy was proposed for DC–DC series–parallel power conversion systems. Thus, the control strategy for PSFB-based ISOP system can be obtained, as shown in Fig. 3.11, which consists of n − 1 IVS control loops and a common output voltage control loop. Gvcd and Gvo are the transfer functions of the input voltages sharing regulator (IVSR) and output voltage regulator (OVR), respectively, and vo_EA is the output signal of the OVR and vin_EAj (j = 1, 2, …, n−1) is the output signal of each IVSR. V RAMPj is the sawtooth signal and its amplitude is V pp . For the first n − 1 modules, vin_EAj is subtracted from vo_EA and then compares with V RAMPj to obtain the duty cycle of module j. For module n, the sum of the first n − 1 vin_EAj and vo_EA compares with V RAMPj to obtain its duty cycle. According to (3.18) and (3.26), we can obtain ⎤ ⎡ ⎡ ⎤ vˆin1_d dˆy1 ⎥ ⎢ ⎢ ⎥ ⎢ vˆin2_d ⎥ ⎢ ⎥ ⎥ ⎢ ˆ ⎢ d y2 ⎥ ⎥ ⎢ ⎥ .. ⎥ H⎢ ⎢ (3.34) ⎢ ⎥ ⎥ ⎢ . ⎢ .. ⎥ ⎥ ⎢ ⎢ ⎥ . ⎢ vˆin(n−1)_d ⎥ ⎣ ⎦ ⎦ ⎣ ˆyn d vˆo_d where vˆin j_d denotes the variations resulted from dˆy j (j = 1, 2, …, n), and H is ⎡ ⎤ A(s)(n−1) A(s) A(s) A(s) − . . . − − n n n n ⎢ ⎥ ⎢ A(s) A(s) A(s) ⎥ A(s)(n−1) − . . . − − ⎢ ⎥ n n n n ⎢ ⎥ ⎢ .. .. .. .. .. ⎥ H ⎢ ⎥. ⎢ . . . . . ⎥ ⎢ ⎥ ⎢ − A(s) − A(s) . . . A(s)(n−1) − A(s) ⎥ ⎣ n n n n ⎦ G vd (s) G vd (s) . . . G vd (s) G vd (s)

68

3 Mathematical Model and Closed-Loop Parameters Design …

+ Vin/n

_ v in1 Gvcd

_

vin_EA1

Comparator 1

+ VRAMP 1

+ Vin/n

_ v in2 Gvcd

vin_EA2 _ +

− +

Comparator 2 − +

RS Flip-flop

Drive circuit 1

RS Flip-flop

Drive circuit 2

VRAMP 2

+ Vin/n

_ v in(n−1) Gvcd IVSR vin_EA(n−1)

_ +

VRAMP (n−1)

+ + + Voref

_ vof

Gvo

Comparator (n−1) − RS Flip-flop +

vo_EA

+

+ +

Drive circuit (n−1)

Comparator n

VRAMP n

− +

RS Flip-flop

Drive circuit n

Fig. 3.11 Control strategy of the PSFB-based ISOP system

According to Fig. 3.11 and from (3.18), (3.27), (3.30), (3.32), and (3.34), we can develop the block diagram shown in Fig. 3.12, where 1/V pp is the transfer function of PWM modulator, and K vc and K vo are the sensor gains of input voltage and output voltage, respectively. Denoting the output signals of n − 1 IVSRs and OVR as xˆ1 , xˆ2 , …, xˆn , respectively, as shown in Fig. 3.12, we have ⎡ ⎤ ⎤⎡ xˆ ⎤ ⎡ dˆy1 ⎢ ⎥ 1 −1 0 0 · · · 0 1 ⎢ ⎥ ⎥ ⎢ dˆy2 ⎥ ⎢ 0 −1 0 · · · 0 1 ⎥⎢ x ˆ ⎢ ⎥ 2 ⎥⎢ ⎢ ⎥ ⎢ ⎥ ⎥ ⎢ ⎥ ⎢ 1 ⎢ .. ⎥ ⎢ .. ⎥ ⎢. . . . . . ⎥ (3.35) ⎢ . ⎥ V ⎢ .. .. .. .. .. .. ⎥⎢ . ⎥ ⎥ pp ⎢ ⎥⎢ ⎢ ⎥ ⎢ ⎥ ⎢ˆ ⎥ ⎣ 0 0 0 · · · −1 1 ⎦⎣ xˆn−1 ⎦ ⎢ dy(n−1) ⎥ ⎣ ⎦ 1 1 1 ··· 1 1 xˆn dˆyn Substituting (3.35) into (3.34) yields

3.2 Decoupling the Control Loops of the PSFB-Based ISOP System

vˆin(n -1)f

vˆin2f vˆin1f + + +

_ Gvcd(s) _ Gvcd(s) _ Gvcd(s)

xˆ1

_

xˆ2

+ _ + _

xˆn −1

+

Kvc/n

vˆin

+

vˆoref

_

Gvo(s)

xˆn

+ +

+

+ +

69

Kvc Kvc Kvc 1/Vpp 1/Vpp 1/Vpp

1/Vpp

dˆy1 dˆy2

+ + ˆvin1_d + +

vˆin1

+ + vˆin(n -1)_d

vˆin ( n −1)

vˆin2_d

dˆy( n −1)

H

Gving(s)

dˆyn

Gvg(s) ++

vˆo_d

vˆof

vˆin2

+

Zout(s)

Kvo

iˆo

vˆo

Fig. 3.12 Block diagram of the control strategy for PSFB-based ISOP system

⎡

vˆin1_d

⎤

⎡

−A(s) 0 ⎢ ⎥ ⎢ ⎢ vˆin2_d ⎥ ⎢ 0 −A(s) ⎢ ⎥ ⎢ ⎢ ⎥ 1 ⎢ . .. .. ⎢ ⎥ ⎢ .. ⎢ ⎥ . . V pp ⎢ ⎢ ⎥ ⎢ ⎢ vˆin(n−1)_d ⎥ 0 0 ⎣ ⎣ ⎦ 0 0 vˆo_d

⎤⎡ ⎤ xˆ1 0 ··· 0 0 ⎥⎢ ⎥ ··· 0 0 ⎥⎢ xˆ2 ⎥ ⎥⎢ ⎥ ⎥⎢ . ⎥ .. .. .. .. . ⎢ ⎥ ⎥ . . . . ⎥⎢ . ⎥ ⎥⎢ ⎥ 0 · · · −A(s) 0 ⎦⎣ xˆn−1 ⎦ 0 · · · 0 nG vd (s) xˆn

(3.36)

Incorporating (3.36) into Fig. 3.12, we can obtain the decoupled block diagram, as shown in Fig. 3.13. As seen, the original n-input-n-output control system is decoupled into n single-input-single-output closed loops, which consists of n − 1 IVS closed loops and one output voltage closed loop. Moreover, the n − 1 IVS closed loops are decoupled from one another and are also decoupled from the output voltage closed loop. Thus, we can design the parameters of IVSR and OVR independently.

3.3 Closed-Loop Design This section discusses the parameters design of Gvcd (s) and Gvo (s) based on a twomodule PSFB ISOP system. The specifications of the ISOP system are as follows. The input voltage V in 540 Vdc, the output voltage V o 60 Vdc, and the output current is 50 A (the equivalent load resistor RLd 1.2 ). For each PSFB module, the

70

3 Mathematical Model and Closed-Loop Parameters Design …

vˆin

Kvc/n

+ _

Gving(s) xˆ Gvcd(s) 1 1/Vpp

Kvc/n

+ _

Gving(s) xˆ Gvcd(s) 2 1/Vpp

Kvc/n

+ _

Gving(s) xˆ Gvcd(s) n −1 1/Vpp

vˆin ( n −1)f

vˆoref

+ _

Gvo(s)

vˆin1

A(s)

+

+

vˆin2

Zout(s)

1/Vpp

vˆof

A(s)

+

+

n−1 IVS control loops

vˆin ( n −1)

Kvc

iˆo

xˆn

+

Kvc

vˆin2f

vˆin

+

Kvc

vˆin1f

vˆin

A(s)

+ nGvd(s) +

+

vˆin

Gvg(s)

Kvo

vˆo

Output voltage control loop

Fig. 3.13 Decoupled block diagram of the control strategy for PSFB-based ISOP system

transformer turns ratio K = 12:4, the output filter inductor L f 26 µH, the output filter capacitor C f 3000 µF with equivalent series resistor Rcf 0.015 , the resonant inductor L r 6.5 µH, the switching frequency f s 100 kHz, the voltage sensor gains K vc 0.01 and K vo 0.03, and the peak-to-peak voltage of the sawtooth signal V pp 4 Vdc. The IVS and output voltage control loops are shown in Fig. 3.13. Here, n = 2.

3.3.1 Output Voltage Regulator It can be readily found in Fig. 3.13 that the loop gain of the output voltage control loop is Tvo_c (s) 2K vo G vd (s)G vo (s)/Vpp

(3.37)

By setting Gvo (s) 1, the uncompensated loop gain T vo_u (s) is obtained as

3.3 Closed-Loop Design

71

Fig. 3.14 Uncompensated and compensated output voltage loop gains

60 40

|Tvo_c|

20

(dB) 0

−20

|Tvo_u|

−40 −60 60 30 0

∠Tvo_u

(°) −30 −60 −90 −120 −150 10

∠Tvo_c 1⋅103 f (Hz)

100

1⋅104

1⋅105

|Gvo| (db) 20log|Gvo |

R2 vi vref

R1 R1

f f

0

C

0

∠Gvo (°)

_

fz =

−90

+

0.1fz

(a) Circuit diagram

fz

ωz 2π

10fz

(b) Bode plots

Fig. 3.15 PI compensator

Tvo_u (s) 2K vo G vd (s)/Vpp

(3.38)

With the given parameters of the ISOP system and substituting (3.18) into (3.38), the Bode plots of T vo _u (s) are depicted, as shown in Fig. 3.14 (dashed lines). As seen, the uncompensated loop gain is close to zero at low frequencies. In order to increase the low-frequency loop gain, a proportional–integral (PI) compensator is adopted as the output voltage regulator, as shown in Fig. 3.15a, which can be expressed as

G vo (s) G vo∞

R2 s + ωz1 s R1

s+

1 R2 C s

(3.39)

1 ωz1 . 2π 2π R2 C The crossover frequency f c1 is chosen to be 10 kHz, which is one-twentieth of the secondary ripple frequency. As shown in Fig. 3.14, the uncompensated loop gain where the corner frequency of the PI compensator is f z1

72

3 Mathematical Model and Closed-Loop Parameters Design …

at 10 kHz has a magnitude of −17.087 dB, and the compensator should then have a gain of 17.087 dB at 10 kHz, i.e., R2 R1 17.078 10 20

G vo∞

(3.40)

Setting R1 10 k and then R2 71 k. In practice, R2 68 k. To guarantee no phase angle lag after the compensation, it is usually set 10 f z1 ≤ f c1 . With Fig. 3.15b, we have C

1 5 ≥ 2.34 (nF) 2π f z1 R2 π f c1 R2

(3.41)

Here, C = 10 nF is chosen, and the transfer function of the PI compensator is G vo (s) 6.8

s + 1471 s

(3.42)

The compensated loop gain T vo_c (s) is shown as solid lines in Fig. 3.14. As shown, the compensated loop gain has a crossover frequency of 10 kHz with a phase margin of 85.9°.

3.3.2 IVS Regulator As shown in Fig. 3.13, the loop gain of the IVS control loop is Tcd_c (s) K vc G vcd (s)A(s)/Vpp

(3.43)

Since Gvinjg (s) 1/n (j = 1, 2, …, n) according to (3.30), it can be concluded from Fig. 3.13 that the IVS regulator gain should be zero, implying that the IVS control would be unnecessary when all the modules are identical. Obviously, when the modules are not identical, the IVS control would be needed. In fact, to facilitate an accurate design of the compensator, the small-signal model should be further modified, taking into account the differences of the modules, which is much more complex and will not be discussed here. By setting Gvcd (s) 1, the uncompensated loop gain T cd_u (s) is expressed as Tcd_u (s) K vc A(s)/Vpp

(3.44)

With the given parameters of the ISOP system, the Bode plots of T cd _u (s) are depicted, as shown in Fig. 3.16 (dashed lines). A PI compensator (see Fig. 3.15a) is employed and its transfer function is

3.3 Closed-Loop Design

73 60

Fig. 3.16 Uncompensated and compensated IVS loop gains

30

|Tcd_c|

0

(dB) −30

|Tcd_u|

−60 −90 −120 60 0

(°)

∠Tcd_u

−60 −120

∠Tcd_c

−180 10

100

G vcd (s) G vcd∞

1⋅103 f (Hz)

s + ωz2 s

1⋅104

1⋅105

(3.45)

Suppose that we aim to design a compensator to achieve a crossover frequency of 300 Hz, and the uncompensated loop gain at 300 Hz has a magnitude of approximately −13.313 dB. Thus, the compensator should have a gain of 13.313 dB at 300 Hz, i.e., G vcd∞ 10

13.313 20

(3.46)

Setting R1 10 k and then R2 46.13 k. Here, R2 47 k is chosen. To guarantee no phase angle lag after the compensation, 10f z2 ≤ f c2 is selected. With Fig. 3.15b, we have C

1 5 ≥ 0.113 (µF) 2π f z2 R2 π f c2 R2

(3.47)

Here, C 0.1 µF is chosen, and the transfer function of the PI compensator is G vcd (s) 4.7

s + 212.7 s

(3.48)

Substituting (3.48) into (3.43) and with the given system parameters, the compensated loop gain T cd_c (s) is depicted, as shown by the solid lines in Fig. 3.16. It can be found that the compensated loop gain has a crossover frequency of 300 Hz with a phase margin 80.5°.

74

3 Mathematical Model and Closed-Loop Parameters Design …

Table 3.1 Parameters of OVR and IVSR Parameters OVR

IVSR

1#

G vo (s) 6.8(s + 1471)/s

G vcd (s) 4.7(s + 212.7)/s

2#

G vo (s) 12(s + 833)/s

G vcd (s) 9.1(s + 161.6)/s

3.4 Experimental Verification of the General Control Strategy A prototype of ISOP system consisting of two PSFB modules was built in the lab to verify the correctness of the small-signal model and loop parameters design in Sects. 3.2 and 3.3, and the specifications are • • • •

input voltage V in 540 Vdc ± 10%, output voltage V o 60 Vdc, rated output current I o 50 A, and switching frequency f s 100 kHz.

Figure 3.17 shows the experimental waveforms of vAB1 , vAB2 , ip1 , ip2 , vrect1 , vrect2 , iLf1 , and iLf2 at full load condition, and it can be seen that both the input voltages and output currents of the two PSFB modules are shared. Figure 3.18 shows the dynamic responses of the input voltage and output voltage (AC coupled) of module 1 when the system input voltage is step changed between 486 and 596 Vdc. The parameters of OVR and IVSR are listed in Table 3.1. Figure 3.18a shows the dynamic responses with the same 1# IVSR parameter but different OVR parameters. It can be seen that the output voltage responses differ and no change on the module input voltage. Figure 3.18b shows the dynamic responses with the same 1# OVR parameter but different IVSR parameters. It can be seen that the output voltage responses keep unchanged. Figure 3.19 shows the dynamic responses of the input voltage (AC coupled) and output voltage (AC coupled) of module 1 when the load current is step changed between 25 and 50 A with different regulator parameters. It can be found that the load step has no influence on the module input voltage, verifying the correctness of (3.33). We can conclude from Figs. 3.18 and 3.19 that IVS control loop is decoupled from the output voltage control loop.

3.5 Three-Loop Control Strategy for ISOP System In Sects. 3.2 and 3.3, removing the IVS control loops will lead to a common duty cycle control strategy with a single-output voltage control loop for the ISOP system, and the ISOP system is stable. However, the mismatched parameters among the modules may cause input voltages unbalance, and thus the IVS control loops are needed. The

3.5 Three-Loop Control Strategy for ISOP System

75

vAB1:[400V/div] ip1:[20A/div]

vAB2:[400V/div]

Time:[2μs/div]

ip2:[20A/div]

(a) Primary voltages and currents of the two PSFB modules

vrect1:[100V/div]

iLf1:[15A/div] vrect2:[100V/div]

iLf2:[15A/div] Time:[2μs/div] (b) Secondary rectified voltages and filter inductor currents of the two PSFB modules Fig. 3.17 Experimental waveforms of the ISOP system

inner current loop can be introduced to form a voltage–current double closed loop to improve the system dynamic response and readily realize output current short-circuit protection. So, whether the ISOP system with only the voltage–current double closed loop is stable if the IVS control loops are removed too? If it is unstable, then the IVS control loops are required, but how to design the IVS regulators? This section will address these issues.

76

3 Mathematical Model and Closed-Loop Parameters Design …

vin:[200V/div] vin1:[100V/div]

1# IVSR 1# OVR

vin1:[100V/div] vo:[2V/div]

1# IVSR 2# OVR

vo:[2V/div] Time:[20ms/div] (a) Different OVR parameters

vin:[200V/div] vin1:[100V/div]

1# IVSR 1# OVR

vin1:[100V/div] vo:[2V/div]

2# IVSR 1# OVR

vo:[2V/div] Time:[20ms/div] (b) Different IVSR parameters

Fig. 3.18 Dynamic responses for stepped input voltage

3.5.1 Positive Resistance Figure 3.20 shows the control block diagram of the voltage–current double closed loop for PSFB ISOP system, where iLf1 , iLf2 , …, iLfn are the module output filter inductor currents, Gc is the gain of inner current loop regulator, and the current reference signal of the inner current loops is the output of the output voltage regulator. At steady state, all the inner current loops have the same current reference and all the modules have the same output filter inductor currents, as well as the output

3.5 Three-Loop Control Strategy for ISOP System

77

io:[25A/div] vin1:[10V/div]

1# IVSR 1# OVR

vin1:[10V/div]

vo:[2V/div]

1# IVSR 2# OVR

vo:[2V/div] Time:[20ms/div] (a) Different OVR parameters

io:[25A/div] vin1:[10V/div]

1# IVSR 1# OVR

vin1:[10V/div]

vo:[2V/div]

2# IVSR 1# OVR

vo:[2V/div] Time:[20ms/div] (b) Different IVSR parameters

Fig. 3.19 Dynamic responses for stepped load

currents, which is equivalent to an OCS control strategy. In Chap. 2, it is pointed out that the ISOP system is unstable with OCS control strategy and the root cause is that the input impedance of each converter module appears as negative resistance characteristic when OCS control strategy employed. Therefore, the ISOP system with only voltage–current double closed loop is unstable if the IVS control loops are removed. Thus, if the ISOP system is stable, the input impedance of each converter module must appear as positive resistance characteristic, and one direct approach is to connect

78

3 Mathematical Model and Closed-Loop Parameters Design … _ iLf1 +

Gc

Comparator 1 − RS Flip-flop +

Drive circuit 1

Comparator 2 − RS Flip-flop +

Drive circuit 2

Comparator n − RS Flip-flop +

Drive circuit n

VRAMP 1 _ iLf2 +

Gc

Inner current loops

+ Voref

_ vof

Gvo

_ iLfn + vo_EA

VRAMP 2

Gc

VRAMP n

Fig. 3.20 Block diagram of voltage–current double closed loop Fig. 3.21 Small-signal model of ISOP system with a resistor connected to the input side of each module

iˆcin_1

iˆin +

′ Z in1

vˆ in1 _ Cd1

iˆin_cs1 Rne1

Rp1

iˆcin _ 2 + vˆ in

′ Z in2

vˆ in2 _ Cd2

iˆin_cs2 Rne2

Rp2

iˆcin _ n +

′ Z inn vˆ inn _ Cdn

iˆin_csn Rnen

Rpn

a positive resistance Rpj (j = 1, 2, …, n) in parallel with the input side the each module, as shown in Fig. 3.21, where iˆin_cs j (j = 1, 2, …, n) represents the each module’s input current perturbation when the voltage–current double closed loop is adopted. The equivalent input impedance of each module is Z in j Rne j //Rp j

Rne j · Rp j > 0 ( j 1, 2, . . . , n) Rne j + Rp j

(3.49)

3.5 Three-Loop Control Strategy for ISOP System

+ Vin/n

+ Vin/n

_ v in1

vin_EA1 Gvcd IVSR

_ v in2

_

iref1+

_ vinn

_

iref2 + +

_ vof + Voref

vin_EAn Gvcd

Comparator 1 RS Flip-flop +

Drive circuit 1

Comparator 2 RS Flip-flop +

Drive circuit 2

Comparator n RS Flip-flop +

Drive circuit n

VRAMP 1

IVS loops

+ Vin/n

Gc

+

vin_EA2 Gvcd

_ iLf1

79

_

Gc Inner current loops

irefn + +

_ iLf2

_ iLfn Gc

VRAMP 2

VRAMP n Gvo

vo_EA

Output voltage loop

Fig. 3.22 Three-loop control strategy for ISOP system

Thus, we have Rp j < −Rne j ( j 1, 2, . . . , n)

(3.50)

The Rpj can be a real resistor and the resistance of Rpj < −Rnej (j = 1, 2, …, n), due to the −Rnej (j = 1, 2, …, n), is the equivalent resistance reflecting the input power of each module, if Rpj < −Rnej (j = 1, 2, …, n), the power dissipated on Rpj (j = 1, 2, …, n) will be larger than the input power of each module, which is absolutely unacceptable. When the input impedance of module appears as positive resistance characteristic, the module input current increases/decreases with the increase/decrease of input voltage, and consequently the module output current will also increase/decrease, as well as the module output filter inductor current. Therefore, we can compare the input voltage of each module with the reference of input voltage of each module V in /n, and when the input voltage of module is higher/lower than V in /n, the module output filter inductor current should increase/decrease by introducing a feedback signal to increase/decrease the current reference signal of the inner current loop. This function can be implemented by the IVS loops, as shown in Fig. 3.22. Taking module 1 for example, when its input voltage vin1 is higher than V in /n, leading to the increase of its current reference signal for inner current loop, thus the output filter inductor current and output power will increase, as well as the input current of module 1, resulting in the input impedance of module 1, appears as positive resistance characteristic.

80

3 Mathematical Model and Closed-Loop Parameters Design …

Fig. 3.23 Equivalent small-signal model for ISOP system with three-loop control strategy

iˆcin _1

iˆin

+ ′ Z in1 ˆvin1 _ Cd1

iˆin_cs1

iˆin_IVS1

Rne1

Rp1

iˆin_cs2

iˆin_IVS2

Rne2

Rp2

+

vˆ in

′ Z in2 vˆ in2 _ Cd2 iˆcin _ n +

′ Z inn vˆ inn _ Cdn

Ă

iˆcin _ 2

iˆin_csn

iˆin_IVSn

Rnen

Rpn

3.5.2 IVS Regulator Design In Sect. 3.2, it has been pointed out that the IVS control loops and output voltage control loop are decoupled, and thus we can design the IVS loops and voltage–current double closed loops independently. In Sect. 3.5.1, it is found that the function of the IVS loops is equivalent to a positive resistance and the equivalent small-signal model for ISOP system with three-loop control strategy is shown in Fig. 3.23, where iˆin_IVS j (j = 1, 2, …, n) is the module input current perturbation introduced by the IVS control loops. According to Fig. 3.22, the individual output filter inductor current perturbation iˆLf_IVS j (j = 1, 2, …, n) caused by the input voltage perturbationvˆin j (j = 1, 2, …, n) can be expressed as iˆLf_IVS j vˆin j · G vcd ( j 1, 2, . . . , n)

(3.51)

Then, the individual output power perturbation pˆ o j (j = 1, 2, …, n) is pˆ o j iˆLf_IVS j · Vo vˆin j · G vcd · Vo ( j 1, 2, . . . , n)

(3.52)

3.5 Three-Loop Control Strategy for ISOP System

81

Referring to Fig. 3.23, the individual input power is pin j vin j · i cin_ j ( j 1, 2, . . . , n)

(3.53)

The perturbation equation of (3.53) can be expressed as Pin j + pˆ in j Vin j + vˆin j Icin_ j + iˆcin_ j

Vin j Icin_ j + vˆin j Icin_ j + Vin j + vˆin j iˆin_IVS j + iˆin_cs j

( j 1, 2, . . . , n)

(3.54)

where Pinj , V cdj , and I cin_j (j = 1, 2, …, n) are the module input power, input voltage, and input current at steady state, respectively. Neglecting the second-order term of (3.54) and noting that the DC terms on both sides of the equation are equal, we obtain (3.55) pˆ in j Icin_ j vˆin j + Vin j iˆin_IVS j + iˆin_cs j ( j 1, 2, . . . , n) According to power conservation, we have pˆ o j pˆ in j · η j ( j 1, 2, . . . , n)

(3.56)

Combining (3.52), (3.55), and (3.56) yields Vo Icin_ j vˆin j + Vin j iˆin_IVS j + iˆin_cs j vˆin j · G vcd · ( j 1, 2, . . . , n) ηj

(3.57)

The negative resistance can be expressed as Vin j vˆin j − Rne j ( j 1, 2, . . . , n) ˆi in_cs j Icin

(3.58)

Substituting (3.58) into (3.57) yields Rp j

vˆin j iˆin_IVS j

Vin j G vcd · Vo η j

( j 1, 2, . . . , n)

(3.59)

Substituting (3.58) and (3.59) into (3.50) leads to G vcd >

Pin · η j ( j 1, 2, . . . , n) Vo · Vin

(3.60)

Neglecting the difference among the module efficiencies and assuming η1 η2 ··· ηn η, according to (3.60), we can obtain

82

3 Mathematical Model and Closed-Loop Parameters Design …

G vcd >

Po · η Vo · Vin

(3.61)

To ensure the stable operation of ISOP system over the whole input voltage and load range, Gvcd should satisfy the following formula: G vcd_min

Po_max · η Vo · Vin_min

(3.62)

where V in_min is the minimum input voltage and Po_max is the maximum output power.

3.6 Experimental Verification of the Three-Loop Control Strategy In order to verify the correctness of the IVS loop design, an ISOP system prototype consisting of two PSFB converter modules has been fabricated with the following specifications: • • • •

input voltage V in 540 Vdc ± 10%, output voltage V o 54 Vdc, rated output current I o 40 A, and switching frequency f s 100 kHz.

Figure 3.24 shows the experimental waveforms of the ISOP system with voltage–current double closed-loop control strategy (see Fig. 3.20). vin1 , vin2 , io1 , and io2 are the input voltages and output currents of the two modules, respectively, and vo is the system output voltage. Prior to t 1 , the three-loop control strategy (see Fig. 3.22) is employed and the ISOP system works stably, and IVS and OCS are achieved between the two modules. At t 1 , the IVS loops are removed and after that vin1 rises and vin2 declines, but the two modules still share the output current and system output voltage is still stable until t 2 . At t 2 , the duty cycle of module 2 reaches 1, and its inner current loop regulator saturates. Thus, the two modules cannot ensure the sharing of output current, and after a period of time the ISOP system enters the new steady state, where both the input voltages and output currents of the two modules are unbalanced. Due to the fact that the PSFB converter is a buck-derived topology, the input voltage of module 2 will not decline when it is not high enough to provide the desired output voltage. Figures 3.25 and 3.26 show the waveforms comparison of the ISOP system with three-loop control strategy (see Fig. 3.22) and general control strategy (see Fig. 3.11). Figure 3.25 shows the dynamic responses of input voltages of two modules and output voltage (AC coupled) for stepped input voltage between 500 and 600 Vdc. Figure 3.26 shows the dynamic responses of input voltages of two modules and output voltage (AC coupled) for stepped load current between 25 and 50 A. It can be seen that the input voltages are well shared, and the sharing of the output current is achieved

3.6 Experimental Verification of the Three-Loop Control Strategy

83

C2 vin1: [200V/div] C3 vin2: [200V/div] C2/ C3

M2 io1: [10A/div] M3 io2: [10A/div]

M2/ M3

C4 vo: [1V/div]

C4

t1 Time: [5ms/div]

t2

Fig. 3.24 Experimental waveforms of the ISOP system with voltage–current double closed-loop control strategy C1 vin:[500V/div] C1

C2

C2 vin1:[200V/div]

C3 vin2:[200V/div]

C3 C4

C1 vin:[500V/div] C1

C2 vin1:[200V/div]

C2

C3 vin2:[200V/div]

C3

C4 vo:[5V/div] Time:[20ms/div]

C4

C4 vo:[5V/div] Time:[20ms/div]

(a) General control strategy

(b) Three-loop control strategy

Fig. 3.25 Dynamic responses for stepped input voltage

automatically for the ISOP system regardless with the three-loop control strategy or the general control strategy. Furthermore, the ISOP system employing the three-loop control strategy has a better dynamic performance with smaller magnitude of output voltage overshoot and shorter regulation time in transient.

3.7 Summary In this chapter, the small-signal model of the PSFB-based ISOP system is derived. Based on the small-signal model, it is proved that the general control strategy proposed in Chap. 2 for n-module ISOP system can be decoupled into n single-input-

84

3 Mathematical Model and Closed-Loop Parameters Design … C1 io:[20A/div]

C1 io:[20A/div] C1

C2 vin1:[200V/div]

C1

C2 vin1:[200V/div]

C2

C2

C3 vin2:[200V/div]

C3 vin2:[200V/div]

C3

C3 C4

C4 vo:[2V/div] Time:[20ms/div]

C4 vo:[2V/div]

C4

Time:[20ms/div]

(a) General control strategy

(b) Three-loop control strategy

Fig. 3.26 Dynamic responses for stepped load current

single-output closed loops, which consists of n − 1 IVS closed loops and one output voltage closed loop. Moreover, the n − 1 IVS closed loops are decoupled from one another and are also decoupled from the output voltage closed loop. The parameters design of the IVS control loop regulator and output voltage control loop regulator based on a two-module PSFB ISOP system is discussed and verified by the prototype. Inner current loop can be introduced into the general control strategy to improve the system dynamic response and readily realize output current short-circuit protection. If the IVS loops are removed and only voltage–current double closed loop is used for the ISOP system, the system is unstable and the root cause is that the input impedance of each module appears as negative resistance characteristic. Hence, the IVS loops are indispensable for ensuring the system stability to change the negative resistance characteristic to the positive resistance characteristic, and this is also the design criteria IVS control loop regulator.

References 1. Ruan, X., Yan, Y.: Soft-switching techniques for PWM full bridge converters. In: Proceedings of the IEEE Power Electronics Specialists Conference, pp. 634–639 (2000) 2. Vlatkovic, V., Sabate, J.A., Ridley, R.B., Lee, F.C., Cho, B.H.: Small-signal analysis of the phase-shifted PWM converter. IEEE Trans. Power Electron. 17(1), 128–135 (1992) 3. Sabate, J.A., Vlatkovic, V., Ridley, R.B., Lee, F.C., Cho, B.H.: Design considerations for highvoltage high-power full-bridge zero-voltage-switched PWM converter. In: Proceedings of the IEEE Applied Power Electronics Conference and Exposition, pp. 275–284 (1990)

Chapter 4

Wireless IVS Control Strategies for Input-Series-Connected Systems Based on Positive Output Voltage Gradient Method

Abstract This chapter proposes the wireless input voltage sharing (IVS) control strategies for the input-series-connected systems based on positive output voltage gradient method, which increases the output voltage of each module with the increase of module input voltage. The constituent modules of input-series-connected systems with the proposed control strategy merely need to sense their own input/output voltages and there is no control interconnection among modules, leading to a truly modular design and high system reliability. The operation principle of the control strategy and the system stability are analyzed in this chapter. Experimental prototypes of the input-series-connected systems consisting of three two-transistor forward modules have been fabricated and tested to validate the proposed wireless IVS control strategies. Keywords Input voltage sharing · Input-series-connected systems Positive output voltage gradient method · Truly modular design Multi-module series–parallel power conversion systems can be categorized into IPOP, IPOS, ISOP, and ISOS systems, where ISOP and ISOS systems are the inputseries-connected systems. It is pointed out in Chap. 2 that only control strategies intending to achieve IVS for the input-series-connected systems are stable, and the OCS (for ISOP) and OVS (for ISOS) are achieved automatically. A modularization architecture for the input-series-connected systems was proposed by distributing the IVS control loops and output voltage control loop into the individual module and then interconnecting each module through a common duty cycle bus and an IVS bus to achieve IVS and the desired output voltage. This implies that the common duty cycle bus and IVS bus are essential for the system stable operation. When the buses are disrupted or failed, the operation of systems will be affected or even go to paralysis which reduces the system reliability to some extent. Therefore, eliminating the interconnection among the modules is the desired choice. This chapter proposes the wireless IVS control strategies for the input-seriesconnected systems based on positive output voltage gradient method, which increases the output voltage of each module with the increase of module input voltage. The

© Springer Nature Singapore Pte Ltd. and Science Press, Beijing, China 2019 X. Ruan et al., Control of Series-Parallel Conversion Systems, CPSS Power Electronics Series, https://doi.org/10.1007/978-981-13-2760-5_4

85

86

4 Wireless IVS Control Strategies for Input-Series-Connected …

constituent modules of input-series-connected systems with the proposed control strategy merely need to sense their own input/output voltages and there is no control interconnection among modules leading to a truly modular design and high system reliability [1, 2]. This chapter will analyze the operation principle of the control strategy and system stability. Experimental prototypes of the input-series-connected systems consisting of three two-transistor forward modules have been fabricated and tested to validate the proposed wireless IVS control strategies.

4.1 Genesis of the Wireless IVS Control Strategies Based on Positive Output Voltage Gradient Method The DC–DC ISOP system shown in Fig. 4.1 is taken as an example to derive the wireless IVS control strategy. In Chap. 2, the modularization architecture for DC–DC series–parallel power conversion systems is proposed (see Fig. 2.9). For the convenience of illustration, the control block diagram for ISOP system is given here, as shown in Fig. 4.2, where V inj (j 1, 2, …, n) is the input voltage of module j, vofj (j 1, 2, …, n) is the sensed output voltage of each module (also is the system output voltage), V oref is the output voltage reference, V RAMPj (j 1, 2, …, n) is the sawtooth signal of the modulator j, Gvo is the transfer function of output voltage control loop regulator and Gvcd is the transfer function of IVS control loop regulator. The control strategy of each module consists of an output voltage closed loop and an IVS closed loop. The signal of common duty cycle bus, vo_EA , is formed by interconnecting the output signal of the individual output voltage control loop regulator through a diode. The signal of IVS bus V in_ref is formed by interconnecting the input voltage of each module through a resistor R, and it is easy to know that V in_ref V in /n, where V in is the system input voltage. V in_ref serves as the reference voltage of each module input voltage. The error signal between V in_ref and V inj (j = 1, 2, …, n) is sent to the corresponding IVS control loop regulator and its output is subtracted from vo_EA and then compared with the sawtooth signal generating the duty cycle signal of each module. Obviously, the reference voltage V in_ref is variable when the system input voltage V in varies. If we set the V in_ref as a constant value, denoted as V inmod_ref , the IVS bus can be eliminated. However, if the system input voltage is not equal to n·V inmod_ref , the IVS control loop regulators will be saturated and fail to fulfill their function. To achieve IVS, the error signal between the reference voltage V inmod_ref and V inj (j 1, 2, …, n) can be amplified by a proportional coefficient k up and added to the reference output voltage V oref as shown in Fig. 4.3. The common duty cycle bus and corresponding diodes can also be removed and finally, the wireless IVS control strategy for ISOP system is obtained as shown in Fig. 4.4. It can be seen that only the input and output voltages are sensed for each module and there is no control interconnection among the constituent modules, which implies that a wireless controller is obtained and the system reliability can be significantly enhanced. Moreover, all the modules

4.1 Genesis of the Wireless IVS Control Strategies Based …

Iin

87

Iin1 + Vin1 C _ d1

Io1 DC-DC Module 1

Cf1

Iin2 Vin

+ Vin2 C _ d2

+ RLd Vo _

Io2 DC-DC Module 2

Cf2

Iinn + Vinn C _ dn

Io

Ion DC-DC Module n

Cfn

Fig. 4.1 Circuit diagram of ISOP system

are identical and can operate in stand-alone mode, which means that the loop design is very simple compared with the other control strategies for the ISOP systems. A truly modular design can be realized for the ISOP system with the proposed control strategy. At steady state, the input voltage of each module varies in the range of [V inmin /n, V inmax /n], where V inmin and V inmax are the minimum and maximum system input voltages, respectively. If we set V inmod_ref to be a constant value, such as V inmin /n, the error signal (V inj − V inmod_ref ) (j 1, 2, …, n) will rise if the module input voltage increases from V inmin /n. Since the error signal is added to V oref , the system output voltage will increase. This means that the output voltage increases with the increase of input voltage and the output voltage has a positive gradient regulation characteristic. Here, k up is defined as gradient gain. The operation principle will be analyzed as follows. The ISOP system operates at steady state and the constituent modules share the system input voltage. Assuming a perturbation occurs on the input voltages, which causes V in1 to decrease and V in2 to increase and the input voltages of other modules keep unchanged. Thus, we have V in1 < V in /n < V in2 . Then, the real reference voltage of 1# module output voltage control loop, [V oref + k up1 (V in1 − V inmod_ref )], will lower than that of 2# module output voltage control loop, [V oref + k up2 (V in2 − V inmod_ref )]. Due to the fact that the output sides of the modules are connected in parallel and all the modules have the same output voltage, for the controller of module 1, the output voltage is considered to be higher than its corresponding reference voltage, and the duty cycle will be regulated to reduce the input power of module 1, I in1 will decrease, then the input voltage of module 1 increases. In the meanwhile, for the controller of module 2, the

88

4 Wireless IVS Control Strategies for Input-Series-Connected …

R + Voref

Vin1 _

_ Gvcd

+ _

+

Gvo

Comparator 1 − +

RS Flip-flop

Drive circuit 1

VRAMP 1

vof1

Vin_ref

Module 1 R

+ Voref

Vin2 _

_

Gvcd

+ _

+

Comparator 2 − +

RS Flip-flop

VRAMP 2

Gvo vof2

R + vo_EA Voref

Module 2

Vinn _

_ Gvcd

+ _

Drive circuit 2

+

Gvo

Comparator n − RS Flip-flop +

VRAMP n

vofn

Drive circuit n Module n

Fig. 4.2 Modularization control block diagram of ISOP system

+

Vin1 _ kup1

Vinmod_ref

+

_ vof1 _

V_in2

+

Gvo

Voref

+ Vinmod_ref

Vinn _

+ Voref

Gvo

Comparator 2 − RS Flip-flop +

Comparator n − +

Drive circuit 1 Module 1

VRAMP 2

_ v _ ofn kupn

Comparator 1 − RS Flip-flop +

VRAMP 1

_ vof2 _ kup2

Vinmod_ref

+ Voref

Gvo

Drive circuit 2 Module 2

RS Flip-flop

VRAMP n

Fig. 4.3 Simplified modularization control block diagram of ISOP system

Drive circuit n Module n

4.1 Genesis of the Wireless IVS Control Strategies Based …

Vin1 _ + Vinmod_ref

kup1

_ vof1 + Gvo + Voref

Vin2 _ + Vinmod_ref

kup2

VRAMP 1

_ vof2 + Gvo + Voref

Vinn _ + Vinmod_ref

kupn

+ Voref

Drive circuit 1 Module 1

Comparator 2 − RS Flip-flop +

VRAMP 2

_ vofn +

Comparator 1 − RS Flip-flop +

89

Drive circuit 2 Module 2

Comparator n Gvo VRAMP n

− +

RS Flip-flop

Drive circuit n Module n

Fig. 4.4 Wireless IVS control strategy of ISOP system

output voltage is considered to be lower than its corresponding reference voltage, and its duty cycle will be regulated to increase the input power of module 2, I in2 will increase, then the input voltage of module 2 decreases. Finally, two modules return to the steady state. From the above analysis, for the ISOP system employing the wireless IVS control strategy, the module input current increases when the module input voltage increases, and the module input current decreases when the module input voltage decreases. Thus, the input impedance of the module appears as positive resistance characteristic and the control strategy is stable. According to Fig. 4.4, the output voltage of module j at steady state is Vo j

1 Voref + kup j · Vin j − Vin mod_ref ( j 1, 2, . . . , n) K vo

(4.1)

where K vo is the output voltage sensing gain. Defining V omin V oref /K vo and setting V inmod_ref V inmin /n, (4.1) can be rewritten as kup j Vinmin Vin j − Vo j Vomin + (4.2) ( j 1, 2, . . . , n) K vo n According to (4.2), the curve of the output voltage versus module input voltage can be depicted as shown in Fig. 4.5. Obviously, each module has a positive output voltage gradient regulation characteristic.

90

4 Wireless IVS Control Strategies for Input-Series-Connected …

Fig. 4.5 Positive output voltage gradient regulation characteristic

Voj Vomin

Vinmin /n

Vinj

4.2 Basic Features of the Input-Series-Connected Systems with Wireless IVS Control Strategies 4.2.1 ISOP System For simplicity, a two-module ISOP system is taken as an example to discuss the basic features of the wireless IVS control strategies. According to (4.2), the output voltages of two modules can be expressed as kup1 Vinmin (4.3) · Vin1 − Vo1 Vo1min + K vo 2 kup2 Vinmin (4.4) Vo2 Vo2min + · Vin2 − K vo 2 It should be noted that here, V o1min and V o2min are used for two modules, respectively, because the reference voltages of the two modules may be mismatched. For a two-module ISOP system, we have Vin Vin1 + Vin2

(4.5)

Vo1 Vo2 Vo

(4.6)

According to (4.3)–(4.6), we have Vinmin kup2 Vin + K vo (Vo2min − Vo1min ) + kup1 − kup2 2 Vin1 kup1 + kup2 Vinmin kup1 Vin + K vo (Vo1min − Vo2min ) + kup2 − kup1 2 Vin2 kup1 + kup2 kup1 Vo2min + kup2 Vo1min kup1 kup2 Vinmin Vo − kup1 + kup2 K vo kup1 + kup2 kup1 kup2 Vin + K vo kup1 + kup2

(4.7)

(4.8)

(4.9)

4.2 Basic Features of the Input-Series-Connected Systems …

91

Combining (4.7) and (4.8), the input voltage sharing error is expressed as kup2 − kup1 (Vin − Vinmin ) + 2K vo (Vo2min − Vo1min ) Vin1 − Vin2 Vin12 Vin Vin kup1 + kup2 Vin (4.10) Equation (4.10) implies that the input sharing error is determined by the gradient gains, minimum input voltage, and minimum output voltages of the two modules. According to (4.9), the gradient gain of ISOP system is kup1 kup2 1 dVo 1 1 1 dVin K vo K vo kup1 + kup2 + kup1 kup2

(4.11)

Likewise, the gradient gain of ISOP system consisting of n modules is 1 1 dVo 1 1 dVin K vo 1 + + ··· + kup1 kup2 kupn

(4.12)

It can be seen that the gradient gain of ISOP system is only determined by the module gradient gains, and the larger the n, the smaller the gradient gain of ISOP system, which means that the system has better output voltage regulation performance. In particular, when kup1 kup2 · · · kupn kup , the gradient gain of ISOP system is equal to k up /(nK vo ). Figure 4.6 shows the dependence of the IVS accuracy on the mismatch of output voltage set-point (V o1min and V o2min ) for a two-module ISOP system, and Fig. 4.7 shows the dependence of the IVS accuracy on the output voltage gradient gain. As seen in Fig. 4.6, the IVS accuracy is improved as the output voltage set-point mismatching decreases, and when the output voltage gradient gains are perfectly matched, as shown in Fig. 4.6c, the two modules share the input voltage evenly. From Fig. 4.7, it can be seen that a larger output voltage gradient gain results in better IVS performance, but the system output voltage regulation performance is deteriorated. According to (4.10), (4.11), Figs. 4.6 and 4.7, we can obtain the following conclusions: the larger the module gradient gain, the better the IVS characteristic, but the system output voltage regulation performance is deteriorated. On the contrary, the smaller the module gradient gain, the worse the IVS characteristic, but the system output voltage regulation performance is improved. Therefore, a trade-off must be made between the IVS and the output voltage regulation performance when designing the control circuit. Hence, the proposed control strategy is more suitable for the applications with a narrow input voltage range.

92

4 Wireless IVS Control Strategies for Input-Series-Connected …

Vo

Vo 1# 2#

1# 2#

Vo= Vo1= Vo2

Vin1

Vin/2

Vin2

Vo= Vo1= Vo2

Vin1 Vin/2 Vin2

Vin

(a) Large mismatch

Vin

(b) Small mismatch

Vo 1#, 2# Vo= Vo1= Vo2 Vin

Vin1=Vin2=Vin/2

(c) No mismatch Fig. 4.6 Effect of output voltage set-point mismatching on IVS accuracy

1# 2#

Vo

Vo 1# 2#

Vo= Vo1= Vo2

Vin1 Vin/2

Vin2

(a) Large slope

Vo= Vo1= Vo2

Vin1 Vin/2 Vin2

Vin

Vin

(b) Small slope

Fig. 4.7 Effect of output voltage gradient gain on IVS accuracy

4.2.2 ISOS System Likewise, the wireless IVS control strategy for ISOS system based on positive output voltage gradient method is shown in Fig. 4.8. It is worth noting that the feedback signal for each module is the system output voltage V o .

4.2 Basic Features of the Input-Series-Connected Systems …

Vin1 _ + Vinmod_ref

kup1

_ vof + +

Gvo

Voref Vin2 _ + Vinmod_ref

kup2

_ vof + +

VRAMP 1

Gvo

Voref

Vinn _ + Vinmod_ref

kupn

+ Voref

Comparator 2 − RS Flip-flop +

VRAMP 2

_ vof +

Comparator 1 − RS Flip-flop +

Gvo

Comparator n − RS Flip-flop +

VRAMP n

93

Drive Circuit 1 Module 1

Drive circuit 2 Module 2

Drive Circuit n Module n

Fig. 4.8 Wireless IVS control strategy for ISOS system

A two-module ISOS system is also taken as an example to analyze its features. Similar to ISOP system, we set V omin V oref /K vo and V inmod_ref V inmin /n according to Fig. 4.8, we have kup1 kup2 Vinmin Vinmin Vo2min + (4.13) · Vin1 − · Vin2 − Vo Vo1min + K vo 2 K vo 2 It should be noted that here, V o1min and V o2min are used for two modules, respectively, because the reference voltages of the two modules may be mismatched. For ISOS system, V in V in1 + V in2 . By combining this expression and (4.13), we have K vo (Vo2min − Vo1min ) + kup2 Vin − Vinmin 2 + kup1 Vinmin 2 (4.14) Vin1 kup1 + kup2 K vo (Vo1min − Vo2min ) + kup1 Vin − Vinmin 2 + kup2 Vinmin 2 Vin2 (4.15) kup1 + kup2 kup1 Vo2min + kup2 Vo1min kup1 kup2 (Vin − Vinmin ) Vo + (4.16) kup1 + kup2 K vo kup1 + kup2 According to (4.14) and (4.15), the input voltage sharing error is obtained and is expressed as

94

4 Wireless IVS Control Strategies for Input-Series-Connected …

2K vo (Vo2min − Vo1min ) + kup2 − kup1 (Vin − Vinmin ) Vin12 Vin1 − Vin2 Vin Vin kup1 + kup2 Vin (4.17) Equation (4.17) implies that the input sharing error is determined by the gradient gains (k up1 and k up2 ) and minimum output voltages of the two modules (V o1min and V o2min ). According to (4.16), the gradient gain of two-module ISOS system is kup1 kup2 1 dVo 1 1 dVin K vo 1 K vo kup1 + kup2 + kup1 kup2

(4.18)

It can be seen that the gradient gain of two-module ISOS system is the same as that of two-module ISOP system. Similar to ISOP system, we also can obtain the following conclusions for ISOS system: the larger the module gradient gain, the better the IVS characteristic, but the system output voltage regulation performance is deteriorated. On the contrary, the smaller the module gradient gain, the worse the IVS characteristic, but the system output voltage regulation performance is improved.

4.3 System Stability The purpose of this section is to study the stability of the ISOP system with the wireless IVS control strategy. With no loss of generality and for ease of analysis, the ISOP system under study consists of two two-transistor forward converter modules as shown in Fig. 4.9, where C d1 and C d2 are two input divided capacitors, power switches S 1 and S 2 , diodes D1 and D2 , transformer T r1 , rectifier diodes DR1 and DR2 , and output filter inductor L f1 form two-transistor forward converter 1, power switches S 3 and S 4 , diodes D3 and D4 , transformer T r2 , rectifier diodes DR3 and DR4 , and output filter inductor L f2 form two-transistor forward converter 2, C f is the output filter capacitor. The corresponding small-signal model the ISOP system is shown in Fig. 4.10 [3]. According to Fig. 4.10, we have K1 ˆ Io 1 ˆ dy1 i in − sCd1 vˆin1 − iˆo1 (4.19) Dy1 K1 ˆi o2 K 2 iˆin − sCd2 vˆin2 − Io 2 dˆy2 (4.20) Dy2 K2

RLd vˆo iˆo1 + iˆo2 (4.21) sCf RLd + 1

4.3 System Stability

S1

D2

Cd2

+ Vin2 _

S3

D4

+ Cf

DR2

Vo _

S2 DR3

Iin2

RLd

io1

Tr1

D3

K2:1 *

Vin

K1:1

Lf1

Lf2 io2

*

+ Vin1 _

D R1 D1

*

Cd1

Iin1

*

Iin

95

Tr2

DR4

S4

Fig. 4.9 Circuit diagram of two-module ISOP system

iˆin

1:

Dy1

+ Cd1 vˆin1 I o1 ˆ _ d y1 K1

Vin1 ˆ d y1 K1

1: + Cd2 ˆvin2 I o2 ˆ _ d y2 K2

iˆo

RLd +

iˆo1 Cf

vˆo _

Dy2 K2

Lf2

+ _

vˆin

Lf1

+ _

K1

Vin2 ˆ d y2 K2

iˆo2

Fig. 4.10 Small-signal model of the ISOP system shown in Fig. 4.9

where dˆy1 and dˆy2 are the perturbations of the duty cycles of two modules, and vˆin1 and vˆin2 are the perturbations of individual input voltages, iˆo1 and iˆo2 are the perturbations of individual output currents, K 1 and K 2 are turns ratios of the individual transformers, Dy1 , Dy2 , I o1, and I o2 are steady-state duty cycles and output currents, respectively. Assuming k up1 k up2 k up , K 1 K 2 K, and Dy1 Dy2 Dy , we have I o1 I o2 V o /(2RLd ) and V in1 V in2 V in /2. From Fig. 4.4, the expressions of dˆy1 and dˆy2 are obtained as

96

4 Wireless IVS Control Strategies for Input-Series-Connected …

dˆy1 G vo kup vˆin1 − K vo vˆo Vpp dˆy2 G vo kup vˆin2 − K vo vˆo Vpp

(4.22) (4.23)

where V pp is the peak-to-peak value of sawtooth signal V RAMP . Assuming the bandwidths of the output voltage closed loops are high enough, the perturbation of output voltage vˆo caused by the perturbation of input voltage vˆin can be expressed as vˆo

kup vˆin 2K vo

(4.24)

Substitution of (4.19), (4.20), (4.22), (4.23), and (4.24) into (4.21), yields kup Dy (sCf RLd + 1) vˆin 2iˆin − sCd1 vˆin1 − sCd2 vˆin2 2K vo K RLd

(4.25)

According to Fig. 4.9, the input power and output power of the ISOP system can be expressed as Pin + pˆ in (Vin + vˆin )(Iin + iˆin ) Vin Iin + Vin iˆin + Iin vˆin + vˆin iˆin Po + pˆ o

(Vo + vˆo ) 2Vo vˆo + + RLd RLd RLd RLd 2

Vo2

vˆo2

(4.26) (4.27)

where pˆ in and pˆ o are input and output power perturbations, respectively. Assuming the conversion efficiency is 100%, the input power is equal to the output power. So, according to (4.26) and (4.27), we have V 2 2Vo vˆo vˆ 2 + o Vin Iin + Vin iˆin + Iin vˆin + vˆin iˆin o + RLd RLd RLd

(4.28)

Neglecting the second-order terms of (4.28) and noting that the DC terms on both V2 1 sides of (4.28) are equal, and knowing that Iin o · , substitution of (4.24) into RLd Vin (4.28), gives kup Vin Vo − Vo2 K vo vˆin iˆin RLd Vin2

(4.29)

Substituting (4.29) into (4.25) and considering vˆin vˆin1 + vˆin2 yields vˆin1 vˆin vˆin2 vˆin

kup Vin Vo − K vo Vo2 kup Dy (sCf RLd + 1) −2 + sCd2 2K vo K RLd K vo R Ld Vin2 s(Cd2 − Cd1 ) kup Vin Vo − K vo Vo2 kup Dy (sCf RLd + 1) −2 + sCd1 2K vo K RLd K vo R Ld Vin2 s(Cd1 − Cd2 )

(4.30)

(4.31)

4.3 System Stability

97

From (4.30) and (4.31), the transfer function of the input voltage difference of the two modules to the total input voltage can be derived as vˆin12 vˆin1 − vˆin2 vˆin vˆin

s

kup Dy kup Vin Vo − K vo Vo2 kup Dy Cf −4 + Cd1 + Cd2 + K vo K K vo K RLd K vo RLd V 2 s(Cd2 − Cd1 )

in

(4.32) In practice, C d1 is impossible to equal to C d2 precisely. Hence, there is a root at the origin in the characteristic equation of the transfer function vˆin12 /vˆin , which means that the system is critically stable, i.e., if the input voltage difference exists between the two modules when there is a system input voltage perturbation, the difference will neither to converge nor to diverge. However, the actual system is surely damped by the parasitic resistance of the switches or inductors, so the perturbation will eventually converge and the ISOP system with the wireless IVS control strategy will be stable in practice.

4.4 Experimental Verification In order to verify the theoretical analysis in the previous sections, we have constructed the input-series-connected systems (ISOP and ISOS) prototype including three twotransistor forward converter modules in lab, and the specifications of each module are as follows: • • • • • •

input voltage V in 100–150 Vdc, output voltage V o 50 Vdc, rated output current I o 5 A, switching frequency f s 100 kHz, gradient gain k up 5.68 × 10−3 , sensing gain K vo 0.1.

4.4.1 Experimental Results of ISOP System The specifications of ISOP system are: system input voltage 300–450 Vdc, system output voltage 50 Vdc, system rated output current 15 A. Figure 4.11 shows the voltages across the primary windings of the individual transformers under normal input voltage (400 Vdc) and full load. It can be seen that all the voltages have almost the same positive amplitudes, which means that the module input voltages are equal and the module output current sharing automatically. It also should be noted that the switching signals of the three modules are not interleaved or synchronized to realize the goal of no control interconnection, and hence, the

98

4 Wireless IVS Control Strategies for Input-Series-Connected …

C1: Vtr1 [100V/div] C2: Vtr2 [100V/div]

C3: Vtr3 [100V/div] C4: Vo [50V/div] Time: [2μs/div] Ave

55

165

50

150

45

135

40

120 Vo Vin1 Vin2 Vin3

35 30 290

310

330

350

370 Vin (V)

390

410

430

Vin1 / Vin2 / Vin3 (V)

Vo (V)

Fig. 4.11 Voltages across the primary windings of individual transformers

105 90 450

Fig. 4.12 Measured system output voltage and three module input voltages

switching frequencies of the three modules are slightly different due to tolerance of RC parameters, which determine the switching frequency. Figure 4.12 shows the measured curves of system output voltage and input voltage of each module when ISOP system input voltage varies from 300 to 450 Vdc. It can be seen that the system output voltage increases with the increase of system input voltage, and the output voltage regulation is 5.6% in the whole system input voltage range from 300 to 450 Vdc. The three module input voltage curves are very close to each other. Due to small mismatches in the controller parameters among the three prototypes, such as reference voltages, voltage sensors, and operational amplifiers, the measured maximum input voltage difference among the three modules is 4 Vdc.

4.4 Experimental Verification

99

C1: Vin1 [100V/div]

C2: Vin2 [100V/div]

C3: Vin3 [100V/div] C4: Vo [50V/div] Time: [10ms/div] Ave

Ave

Fig. 4.13 Responses in individual input voltages and output voltage for stepped system input voltage

C1: Vin1 [100V/div]

C2: Vin2 [100V/div] C3 Vin3 [100V/div] C4: Io [10A/div] Time: [4ms/div] Ave

Ave

Fig. 4.14 Responses in individual input voltages to stepped load

Figure 4.13 shows the experimental waveforms of the input voltages of the three modules and the system output voltage corresponding to the system input voltage stepping between 300 and 450 Vdc. Figure 4.14 shows the experimental waveforms of the input voltages of the three modules corresponding to load stepping between full load (15 A) and half load (7.5 A). It can be seen that input voltages are well shared both at steady state and during transient and sharing of the output current is achieved automatically.

100

4 Wireless IVS Control Strategies for Input-Series-Connected …

C1: io1 [5A/div] C2: io2 [5A/div]

C3: io3 [5A/div]

C4 Vo [50V/div] Time: [40ms/div]

(a) Output currents of each module

C2: iLf2 [5A/div]

C1: iLf1 [5A/div]

C3: iLf3 [5A/div]

C4: Vo [50V/div] Time: [40ms/div]

(b) Output filter inductor currents of each module Fig. 4.15 Responses in individual output currents and output filter inductor currents to stepped load

Figure 4.15 shows the experimental waveforms of the output currents and output filter inductor currents of the three modules corresponding to the load stepping between full load (15 A) and half load (7.5 A). It can be seen that the load sharing is achieved both at steady state and during transient. In order to verify the redundancy of the ISOP system with the wireless IVS control strategy (see Fig. 4.4), a fault is imitated in the three-module ISOP system by shorting the input capacitor of one module with a switch in series with a 0.5 resistor, which

4.4 Experimental Verification

101

C1: Vin1 [100V/div]

C2: Vin2 [100V/div]

C3: Vin3 [100V/div] C4: Vo [50V/div] Time: [10ms/div] Ave

(a) Module 1 is isolated

C1: Vin1 [100V/div] C2: Vin2 [100V/div]

C3: Vin3 [100V/div]

C4: Vo [50V/div] Time: [20ms/div] Ave

(b) Module 1 is inserted Fig. 4.16 Response of individual input voltages and output voltage when module 1 is isolated and inserted, respectively

limits the discharging current of the capacitor. Figure 4.16 shows the system response to a fault of module 1. Figure 4.16a shows the input voltages of the three modules and the system output voltage corresponding to a short fault of module 1 under system input voltage V in 310 Vdc and full load condition. It can be seen that the system input voltage is evenly shared by the remaining two modules when the short switch

102

4 Wireless IVS Control Strategies for Input-Series-Connected …

C1: vtr1 [100V/div]

Time : [4μs/div]

C2: vtr2 [100V/div]

C4: Vo [100V/div]

C3: vtr3 [100V/div]

Fig. 4.17 Voltages across the primary windings of individual transformers

is closed. The system output voltage increases slightly because the input voltage changes from about 100–150 Vdc for module 2 and module 3 and the output voltage increases according to the positive output voltage regulation. Figure 4.16b shows the system response when module 1 is inserted into the system by opening the short switch under system input voltage V in 310 Vdc and full load condition. After opening the short switch, the input capacitor of module 1 is charged by the system input current and the input voltage increases, and in the meanwhile, the input voltages of modules 2 and 3 decrease. After a certain amount of regulation time, the three modules share the system input voltage and the output voltage is restored to be stable. The system output voltage decreases slightly because the input voltage changes from about 155–100 Vdc for modules 2 and 3 and the output voltage decreases according to the positive output voltage regulation. It can be seen that without any control interconnection among the modules, redundancy in combination with hot-swap capability of modules can be easily achieved.

4.4.2 Experimental Results of ISOS System The specifications of ISOS system are: system input voltage 300–450 Vdc, system output voltage 150 Vdc, system rated output current 5 A. Figure 4.17 shows the experimental waveforms of the voltages across the primary windings of individual transformers under normal input voltage (400 Vdc) and full load. As seen, all the voltages have almost the same positive amplitudes, which means that the module input voltages are equal. Figures 4.18 and 4.19 show the experimental waveforms of the ISOS system corresponding to stepped input voltage and stepped load, respectively. Figure 4.18

4.4 Experimental Verification

103

C1: Vin1 [100V/div] C2: Vin2 [100V/div]

C3: Vin3 [100V/div]

C4: Vo [100V/div] Time : [20 ms/div] Ave

Ave

(a) Individual input voltages and system output voltage

C1: Vin1 [100V/div] C2: Vo1 [50V/div]

C3: Vo2 [50V/div]

C4: Vo3 [50V/div] Time : [20 ms/div] Ave

(b) Individual output voltages and input voltage of module 1 Fig. 4.18 Responses to stepped input voltage

shows the input voltages and output voltages of the three modules and the system’s output voltage corresponding to an input voltage stepping between 300 and 450 Vdc under full load condition. Figure 4.19 shows the input voltages and output voltages of the three modules corresponding to a load stepping between half load (2.5 A) and full load (5 A) when system input voltage is 400 Vdc. It can be seen that input voltages are well shared both at steady state and during transient and the sharing of the output voltage is achieved automatically.

104

4 Wireless IVS Control Strategies for Input-Series-Connected …

C1: Vin1 [100V/div] C2: Vin2 [100V/div]

C3: Vin3 [100V/div]

C4: Io [5A/div] Time : [10 ms/div] Ave

(a) Individual input voltages and output current

C1: Vo1 [50V/div] C2: Vo2 [50V/div]

C3: Vo3 [50V/div]

C4: Io [5A/div] Time : [10 ms/div] Ave

(b) Individual output voltages and output current Fig. 4.19 Responses to stepped load

4.5 Summary

105

4.5 Summary In order to eliminate the control signal interconnection among the modules and improve the system reliability, the wireless IVS control strategies based on positive output voltage gradient method are proposed for the input-series-connected systems in this chapter. Truly modular input-series-connected systems can be achieved with the proposed control strategies, where all the modules are totally identical for both power stages and control stages. Each module is self-contained and no extra supervisory controller is needed to achieve the input voltage sharing among the modules. Moreover, there is no any control interconnection among the modules, which considerably increases the redundancy with the improved system reliability and maintainability. Experimental results of three-module ISOP and ISOS systems verify the effectiveness of the wireless IVS control strategies.

References 1. Chen, W., Wang, G., Ruan, X., Jiang, W., Gu, W.: Wireless input-voltage-sharing control strategy for input-series output-parallel (ISOP) system based on positive output-voltage gradient method. IEEE Trans. Ind. Electron. 61(11), 6022–6030 (2014) 2. Chen, W., Wang, G.: Decentralized voltage sharing control strategy for fully modular inputseries output-series system with improved voltage regulation. IEEE Trans. Ind. Electron. 62(5), 2777–2787 (2015) 3. Giri, R., Choudhary, V., Ayyanar, R., Mohan, N.: Common-duty-ratio control of input-series connected modular DC-DC converters with active input voltage and load-current sharing. IEEE Trans. Ind. Appl. 42(4), 1101–1111 (2006)

Chapter 5

A General Control Strategy for DC–AC Series–Parallel Power Conversion Systems

Abstract For the DC–AC series–parallel power conversion systems, the output power balance includes not only the equaling of the amplitude and frequency of the modular output voltages and currents but also the equaling of the phase of them. This means that the equilibrium of both the real and reactive powers should be actualized. In the light of this perspective, this chapter first presents a systematical analysis of the relationship between the input voltage/current sharing (IVS/ICS) and output voltage/current sharing (OVS/OCS) of four kinds of DC–AC series–parallel power conversion system. Then, the control strategies to achieve IVS/ICS and OVS/OCS and the system stability are discussed. Finally, a general control strategy for the DC–AC series–parallel power conversion system is presented. Keywords DC–AC series–parallel power conversion systems Input voltage sharing · Input current sharing · Output voltage sharing Output current sharing · Stability · General control strategy Similar to the DC–DC series–parallel power conversion systems, the DC–AC series–parallel power conversion systems, which are composed of multiple DC–AC inverter modules, can also be divided into four kinds, namely, input-parallel-outputparallel (IPOP), input-parallel-output-series (IPOS), input-series-output-parallel (ISOP), and input-series-output-series (ISOS). To guarantee normal operation of these systems, power balance among the constituent DC–AC inverter modules should be ensured. That is to say, input voltage sharing (IVS) and output voltage sharing (OVS) should be achieved when the inverter modules are connected in series at the input and output sides, respectively, and input current sharing (ICS) and output current sharing (OCS) should be realized when the inverter modules are connected in parallel at the input or output sides, respectively. Different from the DC–DC converters, the output of the DC–AC inverters is in the form of alternating current. Therefore, in the DC–AC series–parallel power conversion systems, the output power balance includes not only the equaling of the amplitude and frequency of the modular output voltages and currents but also the equaling of the phase of them. This means that the equilibrium of both the real and reactive powers should be actualized. In this chapter, the relationship between the input © Springer Nature Singapore Pte Ltd. and Science Press, Beijing, China 2019 X. Ruan et al., Control of Series-Parallel Conversion Systems, CPSS Power Electronics Series, https://doi.org/10.1007/978-981-13-2760-5_5

107

108

5 A General Control Strategy for DC–AC Series–Parallel …

voltage/current sharing (IVS/ICS) and output voltage/current sharing (OVS/OCS) of the four kinds of DC–AC series–parallel power conversion systems will be analyzed systematically. Also, the control strategies to achieve IVS/ICS and OVS/OCS and the stability are discussed. Then, a general control strategy for the four kinds of DC–AC series–parallel power conversion system is proposed [1, 2].

5.1 Relationship Between Input Voltage/Current Sharing and Output Voltage/Current Sharing Figure 5.1 shows the four connection architectures of DC–AC series–parallel power conversion system with n DC–AC inverter modules. As shown, for the jth module (j 1, 2, …, n), V in_j and I in_j denote the input voltage and the input current, respectively; vo_j and io_j denote the output voltage and the output current, respectively; I cd_j and icf_j denote the input and output filter capacitor currents, respectively; and I cin_j and ico_j denote the input and output currents of the power stage. At steady state, since both the input voltage and the output voltage are constant, the average values of both the input filter capacitor current and the output filter capacitor current are zero. Therefore, we have Iin_ j Icin_ j ( j 1, 2, . . . , n)

(5.1)

i o_ j i co_ j ( j 1, 2, . . . , n)

(5.2)

Then, the input power of each DC–AC inverter module can be derived as ⎧ ⎪ P Vin_1 · Iin_1 ⎪ ⎪ in1 ⎪ ⎪ ⎨ Pin2 Vin_2 · Iin_2 ⎪ ⎪ ⎪ ⎪ ⎪ ⎩P

inn

.. . Vin_n · Iin_n

(5.3)

And the output real power and reactive power could be expressed as ⎧ ⎪ Po1 Vo_1 · Io_1 · cos θ1 ⎪ ⎪ ⎪ ⎪ ⎨ Po2 Vo_2 · Io_2 · cos θ2 ⎪ ⎪ ⎪ ⎪ ⎪ ⎩

.. . Pon Vo_n · Io_n · cos θn

(5.4)

5.1 Relationship Between Input Voltage/Current Sharing … Iin Vin

Iin_1 Icin_1 ico_1 io_1 + Icd_1 + icf_1 DC-AC vo_1 Cd1 Cf1 Vin_1 inverter _ _ _

io ZLd

+ vo _ Vin

Iin

Module 1

109

Iin_1 Icin_1 ico_1 io_1 io + Icd_1 + icf_1 DC-AC v C C d1 f1 o_1 Vin_1 inverter _ __ Module 1

Iin_2 Icin_2 ico_2 io_2 + Icd_2 + icf_2 DC-AC vo_2 C C d2 f2 Vin_2 inverter _ __

Iin_2 Icin_2 ico_2 io_2 + Icd_2 + icf_2 + DC-AC vo_2 ZLd vo Cd2 Cf2 Vin_2 inverter _ __ _

Module 2

Module 2

Iin_n Icin_n ico_n io_n + Icd_n + icf_n DC-AC Vin_n vo_n Cdn Cfn inverter _ _

Iin_n Icin_n ico_n io_n + Icd_n + icf_n DC-AC Vin_n vo_n Cdn Cfn inverter _ _

Module n

Module n

(a) IPOP Iin

(b) IPOS

Iin_1 Icin_1 ico_1 io_1 + Icd_1 + icf_1 DC-AC Cd1 Cf1 vo_1 Vin_1 inverter _ __

io ZLd

Iin

+ vo _

Iin_1 Icin_1 ico_1 io_1 io + Icd_1 + icf_1 DC-AC Cd1 Cf1 vo_1 Vin_1 inverter _ __

Module 1

Vin

Module 1

Iin_2 Icin_2 ico_2 io_2 + Icd_2 + icf_2 DC-AC Cd2 Cf2 vo_2 Vin_2 inverter _ _ _ Module 2

Iin_n Icin_n ico_n io_n + Icd_n + icf_n DC-AC Vin_n vo_n Cdn Cfn inverter _ _

Vin

Iin_2 Icin_2 ico_2 io_2 + + Icd_2 + icf_2 DC-AC vo_2 ZLd vo Cd2 Cf2 Vin_2 inverter _ _ _ _ Module 2

Iin_n Icin_n ico_n io_n + Icd_n + icf_n DC-AC Vin_n Cdn Cfn vo_n inverter _ _

Module n

Module n

(c) ISOP

(d) ISOS

Fig. 5.1 Four connection architectures of DC–AC series–parallel power conversion system

⎧ ⎪ ⎪ Q o1 Vo_1 · Io_1 · sin θ1 ⎪ ⎪ ⎪ ⎨ Q o2 Vo_2 · Io_2 · sin θ2 ⎪ ⎪ ⎪ ⎪ ⎪ ⎩Q

on

.. . Vo_n · Io_n · sin θn

(5.5)

where V o_1 , V o_2 , …, V o_n are the root-mean-square (RMS) values of the output voltage of each module; I o_1 , I o_2 , …, I o_n are the RMS values of the output current of each module; and θ 1 , θ 2 , …, θ n are the output power factor angles of each module. Suppose the efficiencies of the constituent module are equal and the values are η. So, by power conservation, we have Po j η × Pin j ( j 1, 2, . . . , n). Then, combining (5.3) and (5.4), we can get

110

5 A General Control Strategy for DC–AC Series–Parallel …

⎧ ⎪ V · I · cos θ1 Vin_1 · Iin_1 · η ⎪ ⎪ o_1 o_1 ⎪ ⎪ ⎨ Vo_2 · Io_2 · cos θ2 Vin_2 · Iin_2 · η ⎪ ⎪ ⎪ ⎪ ⎪ ⎩

Vo_n · Io_n

.. . · cos θn Vin_n · Iin_n · η

(5.6)

5.1.1 IPOP Systems As shown in Fig. 5.1a, for DC–AC IPOP systems, the input voltage and the output voltage are equal, respectively, i.e., Vin_1 Vin_2 · · · Vin_n , and vo_1 vo_2 · · · vo_n . If input current sharing (ICS) among the DC–AC inverter modules is achieved, we have Iin_1 Iin_2 · · · Iin_n . Substitution of these three equations into (5.6) leads to Io_1 · cos θ1 Io_2 · cos θ2 · · · Io_n · cos θn

(5.7)

According to (5.7), if we wish to realize output current sharing (OCS) among the constituent DC–AC inverter modules, i.e., i o_1 i o_2 · · · i o_n

(5.8)

One of the following conditions should be satisfied, i.e., Io_1 Io_2 · · · Io_n

(5.9)

θ1 θ2 · · · θn

(5.10)

That is to say, according to (5.7), if (5.9) is valid, then (5.10) is valid; or if (5.10) is established, then (5.9) is established. According to the above analysis, it can be concluded that achieving ICS among the constituent modules can only ensure that the output real powers of the modules in IPOP systems are equal, and the reactive powers among the modules could not be guaranteed. In order to ensure the reactive powers among the modules are equal, the condition (5.9) or (5.10) should be valid, i.e., the amplitudes or phases of the output currents of the inverter modules should be equal. Thus, OCS could be ensured. This is quite different from that of DC–DC IPOP systems since the output of DC–DC converter is in the form of direct current and only the real power is involved. On the other hand, for the DC–AC IPOP systems, if OCS is ensured, (5.9) and (5.10) are valid. Substituting these two equations together with Vin_1 Vin_2 · · · Vin_n and vo_1 vo_2 · · · vo_n into (5.6) leads to Iin_1 Iin_2 · · · Iin_n , which means that ICS is achieved.

5.1 Relationship Between Input Voltage/Current Sharing …

111

5.1.2 IPOS Systems As shown in Fig. 5.1b, for IPOS systems, the input voltages and the output currents of the constituent modules are equal, respectively, i.e., Vin_1 Vin_2 · · · Vin_n , and i o_1 i o_2 · · · i o_n . If ICS of the constituent modules is achieved, we have Iin_1 Iin_2 · · · Iin_n . Substitution of these three equations into (5.6) yields Vo_1 · cos θ1 Vo_2 · cos θ2 · · · Vo_n · cos θn

(5.11)

From (5.11), if OVS among the modules are expected, i.e., vo_1 vo_2 · · · vo_n

(5.12)

One of the following equations should be satisfied, i.e., Vo_1 Vo_2 · · · Vo_n

(5.13)

θ1 θ2 · · · θn

(5.14)

That is to say, according to (5.11), if (5.13) is valid, then (5.14) is valid; or if (5.14) is established, then (5.13) is established. According to the above analysis, it can be concluded that achieving ICS among the modules can only ensure the output real powers of the modules in IPOS systems are equal, but the equaling of the reactive powers of the modules could not be guaranteed. For the purpose of ensuring the reactive powers of the modules are equal, the condition (5.13) or (5.14) should be valid, i.e., the amplitudes or phases of the output voltage of the modules should be equal. Thus, OVS could be guaranteed. This is quite different from DC–DC IPOS systems since the outputs of DC–DC converters are in the form of direct current and only real power is considered. On the other hand, for DC–AC IPOS systems, if OVS among the constituent modules is ensured, (5.13) and (5.14) are valid. Substituting these two equations together with Vin_1 Vin_2 · · · Vin_n and i o_1 i o_2 · · · i o_n into (5.6), we can obtain Iin_1 Iin_2 · · · Iin_n , which means that ICS is achieved.

5.1.3 ISOP Systems As shown in Fig. 5.1c, for ISOP systems, the input currents and the output voltages of the modules are equal, respectively, i.e., Iin_1 Iin_2 · · · Iin_n , and vo_1 vo_2 · · · vo_n . If IVS among the constituent modules is achieved, we have Vin_1 Vin_2 · · · Vin_n . Substitution of these three equations into (5.6) yields Io_1 · cos θ1 Io_2 · cos θ2 · · · Io_n · cos θn

(5.15)

112

5 A General Control Strategy for DC–AC Series–Parallel …

From (5.15), if we intend to achieve OCS among the modules, i.e., i o_1 i o_2 · · · i o_n

(5.16)

One of the following equations should be satisfied, i.e., Io_1 Io_2 · · · Io_n

(5.17)

θ1 θ2 · · · θn

(5.18)

That is to say, according to (5.15), if (5.17) is valid, then (5.18) is valid and vice versa. According to the above analysis, it can be concluded that achieving IVS among the modules can only ensure the module output real powers are equal, but the module reactive powers could not be guaranteed. If the condition (5.17) or (5.18) is valid, i.e., the amplitudes or phases of the output currents of the modules are equal, the module reactive powers could be equal, thus achieving OCS [1]. This is quite different from that of DC–DC ISOP system in which only the real power is involved. On the other hand, for DC–AC ISOP systems, if OCS among the modules is ensured, both (5.17) and (5.18) are valid. Substituting these two equations together with Iin_1 Iin_2 · · · Iin_n and vo_1 vo_2 · · · vo_n into (5.6), we can get Vin_1 Vin_2 · · · Vin_n , which means that IVS is achieved.

5.1.4 ISOS Systems As shown in Fig. 5.1d, for ISOS system, the input currents and the output currents of the modules are equal, respectively, i.e., Iin_1 Iin_2 · · · Iin_n , and i o_1 i o_2 · · · i o_n . If IVS among the modules is achieved, we have Vin_1 Vin_2 · · · Vin_n . Substitution of these three equations into (5.6) leads to Vo_1 · cos θ1 Vo_2 · cos θ2 · · · Vo_n · cos θn

(5.19)

From (5.19), if OVS among the modules is expected, i.e., vo_1 vo_2 · · · vo_n

(5.20)

One of the following equations should be satisfied, i.e., Vo_1 Vo_2 · · · Vo_n

(5.21)

θ1 θ2 · · · θn

(5.22)

That is to say, according to (5.19), if (5.21) is valid, then (5.22) is valid; or if (5.22) is established, then (5.21) is established.

5.1 Relationship Between Input Voltage/Current Sharing …

113

According to the above analysis, it can be concluded that achieving IVS among the modules can only ensure the module output real powers are equal, but the module reactive powers could not be guaranteed. If the condition (5.21) or (5.22) is valid, i.e., the amplitudes or phases of the output voltages of the modules are equal, the module reactive powers could be equal, achieving OVS [2]. This is quite different from that of DC–DC ISOS systems which have only real power. On the other hand, for DC–AC ISOS systems, if OVS among the modules is ensured, both (5.21) and (5.22) are valid. Substituting these two equations together with Iin_1 Iin_2 · · · Iin_n and i o_1 i o_2 · · · i o_n into (5.6), we can get Vin_1 Vin_2 · · · Vin_n , which means that IVS is achieved. According to the above analysis, we can find that, for DC–AC series–parallel power conversion systems, achieving IVS/ICS among the constituent modules can only ensure the module output real powers are equal, but the equaling of the module reactive powers could not be guaranteed. In order to achieve the output reactive power equilibrium, in addition to achieving IVS/ICS, the amplitudes or phases of the module output currents (in IPOP and ISOP systems) or the amplitudes or phases of the module output voltages (in IPOS and ISOS systems) should be ensured. Hereinafter, the control strategy that controls IVS/ICS and makes the amplitudes or phases of module output currents or module output voltages be equal is defined as the compound balanced control strategy [1, 2]. On the other hand, for DC–AC series–parallel power conversion systems, if OVS/OCS among the modules could be guaranteed, IVS/ICS will be ensured automatically.

5.2 Stability of the Control Strategies for DC–AC Series–Parallel Power Conversion Systems It is pointed out in Chap. 2 that, for DC–DC series–parallel power conversion systems, if IVS/ICS among the modules could be guaranteed, OVS/OCS will be ensured automatically, and vice versa. Moreover, it is further pointed out that the IVS/ICS control strategy could ensure the stable operation of all of four kinds of DC–DC series–parallel power conversion systems including IPOP, IPOS, ISOP, and ISOS systems. While the OVS/OCS control strategy could ensure the stable operation of the input-parallel-connected systems (including IPOP and IPOS systems), the operation of the input-series-connected systems (including ISOP and ISOS systems) is unstable. So, do the DC–AC series–parallel power conversion systems have the similar characteristics as mentioned above? In this section, we examine the stability of the systems with the possible control strategies discussed in Sect. 5.1. Before the analysis, we suppose that the system’s input voltage, output voltage, and load current are kept unchanged when perturbations are imposed on the input voltage/current and output voltage/current of the modules, and the modules have the same efficiency.

114

5 A General Control Strategy for DC–AC Series–Parallel …

5.2.1 Compound Balanced Control Strategy Regardless of the connection type at the input side of the DC–AC series–parallel power conversion system, if the compound balanced control strategy is adopted, each DC–AC inverter module behaves as a constant power source as seen at the output side, as shown in Fig. 5.2, and the modular input powers are equal. For the case where the output sides are connected in parallel, as shown in Fig. 5.2a, the output voltages of all the modules are equal. With the compound balanced control strategy, both the module output real powers and reactive powers are, respectively, equal. Thus, the module output currents are identical. So, no perturbation occurs, i.e., the system is stable. For the case where the output sides are connected in series, as shown in Fig. 5.2b, the output currents of all the modules are equal. Assume that a perturbation occurs in the output voltages of some two modules, e.g., the amplitude of vo_1 increases and the amplitude of vo_2 decreases (here we suppose the scheme of controlling the phase of output voltages to be equal is adopted), while the output voltages of other modules are unchanged. Since the system output current io is unchanged, the rise of the amplitude of vo_1 makes module 1 output more real power than other inverter modules. Hence, the output filter capacitor C f1 will be discharged, making the voltage amplitude of module 1 descend and return to the equilibrium point. In the meanwhile, the output filter capacitor C f2 will be charged, and the voltage amplitude of module 2 will increase and returns to the equilibrium point. In conclusion, with the compound balanced control strategy, all the four series— parallel power conversion systems could work stably.

ico_1 Constant Power Source (Inverter 1)

Constant Power Source (Inverter 2)

Constant Power Source (Inverter n)

io_1 icf_1

Cf1

+ vo_1 _

ico_2

io_2 icf_2

Cf2

+ vo_2 _

ico_n

io_n icf_n

Cfn

io ZLd

+ vo_n _

(a) Output sides connected in parallel

+ vo _

ico_1 Constant Power Source (Inverter 1)

Constant Power Source (Inverter 2)

Constant Power Source (Inverter n)

io_1 icf_1

Cf1

+ vo_1 _

ico_2

io_2 icf_2

io

Cf2

+ vo_2 ZLd _

ico_n

io_n icf_n

Cfn

+ vo_n _

(b) Output sides connected in series

Fig. 5.2 Equivalent schematic diagrams for the compound balanced control strategy

+

vo

_

5.2 Stability of the Control Strategies for DC–AC Series–Parallel … Iin

Vin

Iin_1

Icin_1

+ Icd_1 Vin_1 Cd1 _ Iin_2

Iin_2 Constant Power Sink (Inverter 2)

Icin_n

+ Icd_n Vin_n Cdn _

Iin_1

Vin

(a) Input-parallel-connected systems

Constant Power Sink (Inverter 1)

Icin_2

+ Icd_2 Vin_2 Cd2 _

Iin_n Constant Power Sink (Inverter n)

Icin_1

+ Icd_1 Vin_1 Cd1 _

Icin_2

+ Icd_2 Vin_2 Cd2 _

Iin_n

Iin Constant Power Sink (Inverter 1)

115

Constant Power Sink (Inverter 2)

Icin_n

+ Icd_n Vin_n Cdn _

Constant Power Sink (Inverter n)

(b) Input-series-connected systems

Fig. 5.3 Equivalent schematic diagrams for the output side control strategy

5.2.2 Output Voltage/Current Sharing (OVS/OCS) Control Strategy Regardless of the connection type at the output side for the DC–AC series–parallel power conversion systems, if the OVS/OCS control strategy is employed, each inverter module behaves as a constant power sink as seen at the input side, as shown in Fig. 5.3, and the modular output powers are equal (which means the active and reactive powers are, respectively, equal). It is noted that we can only see the input real power at the input side, so we could draw the same conclusion with the DC–DC series–parallel power conversion systems, i.e., the OVS/OCS control strategy could ensure stable operation of the input-parallel-connected systems (including IPOP and IPOS systems), but the operation of the input-series-connected systems (including ISOP and ISOS systems) are unstable. The detailed analysis has been given in Sect. 2.2.2 in Chap. 2 and it is not repeated here.

5.3 General Voltage/Current Sharing Control Strategy for DC–AC Series–Parallel Power Conversion Systems According to the analysis of the above two sections, we can draw the following conclusions: 1. For the input-parallel-connected DC–AC systems (including IPOP and IPOS systems), both compound balanced control strategy and OVS/OCS control strat-

116

5 A General Control Strategy for DC–AC Series–Parallel …

Fig. 5.4 Main circuit of buck-derived inverter

+

Vin

vinv –

Lf iLf iCf Cf

io + vo

ZLd

–

egy could ensure stable operation of the systems. With the compound balanced control strategy, the amplitudes or phases of the modular output currents (for IPOP system) or modular output voltages (for IPOS system) should be guaranteed together with controlling for ICS among the modules. Compared with the OVS/OCS control strategy, the compound balanced control strategy is quite complicated. Thus, the OCS control strategy is more convenient for DC–AC IPOP systems, and the OVS control strategy is preferred for IPOS DC–AC systems. 2. For the input-series-connected DC–AC systems (including ISOP and ISOS systems), OVS/OCS control strategy could not ensure stable operation of the systems. So, we can only adopt the compound control strategy, i.e., the equaling of the amplitude or phase of modular output current (for ISOP system) or modular output voltage (for ISOS system) should be guaranteed together with controlling for IVS among the modules. Regardless of the type of the connection architectures, at least two control loops are needed. One is the output voltage loop for regulating the system output voltage, and the other one is the input voltages/currents or output voltages/currents sharing loop, which will be referred to as voltage/current sharing loop hereinafter. With the buckderived inverter as the basic module, as shown in Fig. 5.4, a general voltage/current sharing control strategy for four kinds of DC–AC series–parallel power conversion systems is proposed, as shown in Fig. 5.5. Likewise, the general voltage/current sharing control strategy for the DC–AC series–parallel power conversion systems with boost-derived or buck–boost-derived converter as the basic module can be derived, and it is not repeated here. The general control strategy includes three closed loops, namely, the system output voltage loop, which is common for all the inverter modules, the voltage/current sharing loop, and the current inner loop of each module. Different from the twoloop control for the DC–DC series–parallel power conversion systems in Sect. 2.3 in Chap. 2, the current inner loop of each module is introduced here. This is because the equaling of the amplitude or phase of modular output current (for ISOP system) or modular output voltage (for ISOS system) should be guaranteed except for IVS among the constituent modules for the purpose of achieving the power balance for the input-series-connected DC–AC power conversion systems (including ISOP and ISOS systems). In order to make the phase of output current (for ISOP system) or output voltage (for ISOS system) of each module identical, the output of the voltage/current sharing loop of each module is multiplied by the output of the common system output voltage loop. In the three-loop control strategy, the system output voltage loop aims

5.3 General Voltage/Current Sharing Control Strategy … Voltage/current sharing loop Vsh_g _ V sh_f 1 + Gsh vsh_EA1 _ Vsh_f 2 +

if_1 −/+

ish_EA1

iref1

if_2 −/+

Gsh vsh_EA2

Current regulator 1

SPWM modulator

Drive signal 1

Current regulator 2

SPWM modulator

Drive signal 2

Current regulator (n−1)

SPWM modulator

Drive signal (n−1)

Current regulator n

SPWM modulator

Drive signal n

_

+

+

…

iref2

…

ish_EA2

_

+

+

if_(n−1)

_ Vsh_f (n−1) + Gsh vsh_EA(n−1) _ Vsh_f n +

−/+

if_n −/+

Gsh

+

_ vof

Gvo

_

+

+ iref(n−1)

ish_EA(n−1)

vsh_EAn

voref

117

+

+

ish_EAn

irefn

_

iref

System output voltage loop

if_ j

−/+

− ISOP and ISOS + IPOP and IPOS

io_ j

ISOP and IPOP

iCf_ j

IPOS and ISOS

iLf_ j

Four kinds of system

(j = 1, 2, …, n)

Fig. 5.5 General voltage/current sharing control strategy for DC–AC series–parallel power conversion systems

for regulating the system output voltage, and its output signal, iref , serves as the initial common reference for all the inner current loops. The voltage/current sharing loop of each module adjusts the magnitude of iref , leading to the final current reference for each module, irefj (j 1, 2, …, n). For the voltage/current sharing loop shown in Fig. 5.5, V sh_g represents the average value of the total system input/output voltage/current, and V sh_fj (j 1, 2, …, n) represents the sensed input or output variable of each module. The representations of V sh_g and V sh_fj for the four kinds of DC–AC series–parallel power conversion systems are given in Table 5.1. As shown in Table 5.1, for the input-series-connected DC–AC power conversion systems, V sh_g is V in /n, and V sh_fj is the sensed input voltage of each module; While for the input-parallel-connected DC–AC power conversion systems, if the compound balanced control is adopted, V sh_fj is I cin_j , and if the output side control is employed, V sh_fj is vo_j for IPOS system and io_j for IPOP system. Furthermore, for the IPOP system, if io_j is taken as the feedback signal of each current inner loop, all the module output currents will track the common cur-

118

5 A General Control Strategy for DC–AC Series–Parallel …

Table 5.1 Representation of V sh_fj and V sh_g in four connection architectures Architecture

V sh_fj (j = 1, 2, …, n)

V sh_g

IPOP

I cin_j

I in /n

IPOS

I cin_j or vo_j

I in /n or vo /n

ISOP

V in_j

V in /n

ISOS

V in_j

V in /n

rent reference iref , and OCS will be achieved without any additional voltage/current sharing loop. Thus, the voltage/current sharing loop could be removed. For the inner current loop of each module in Fig. 5.5, the current feedback signal if_j can be the output current io_j , the output filter inductor current iLf_j , and the output filter capacitor current iCf_j (j 1, 2, …, n) of each module. With different current feedback signals, the sharing of the output voltage/current for the four DC–AC series–parallel power conversion systems will be different. For example, for the ISOP system, if the output filter inductor current is chosen as the feedback signal, the sharing of the output filter inductor current is achieved, and there is output reactive power difference among the modules when the output filter capacitors are not well matched; and if the output current is chosen as the feedback signal, OCS could be achieved even if the output filter capacitors are not well matched [1]. This issue will be discussed in detail in Chap. 6. For the ISOS system, if the output filter inductor current is selected as the feedback signal, there is output reactive power difference among the modules when the output filter capacitors are not well matched; and if the output filter capacitor current is selected as the feedback signal, the sharing of the output voltage will be ensured no matter the output filter capacitors are matched or not [2], which will be elaborated specifically in Chap. 7.

5.4 Summary This chapter systematically analyzes the inherent relationships between the input voltages/current sharing and output voltages/current sharing of the DC–AC series— parallel power conversion systems from the view of power conservation, and discusses the system stability with the compound balanced control strategy and output side control strategy. Then, a general voltage/current sharing control strategy is proposed for the DC–AC series–parallel power conversion systems. (1) For the DC–AC series–parallel power conversion systems, controlling for the input voltage/current sharing (IVS/ICS) among the constituent modules can only ensure the equaling of the output real power of each module, but the equaling of the reactive power could not be guaranteed, i.e., output voltage/current sharing (OVS/OCS) among the constituent modules could not be achieved. This is quite

5.4 Summary

119

different from the DC–DC series–parallel power conversion systems since it only has the real power at the output sides. (2) For the DC–AC series–parallel power conversion systems, controlling at the output sides to achieve output voltage/current sharing (OVS/OCS) among the modules can ensure the module output real and reactive powers are equal, respectively. As a consequence, the input power balance among the modules is guaranteed, and the input voltage/current sharing (IVS/ICS) is achieved. This is the same as DC–DC series–parallel power conversion systems. (3) For the DC–AC series–parallel power conversion systems, the control strategy that controls the input voltage/current sharing (IVS/ICS) and makes the amplitudes or phases of module output currents or module output voltages be equal is defined as the compound balanced control, which can ensure stable operation, and the output side control strategy can only guarantee stable operation of the input-parallel-connected systems (including IPOP and IPOS systems) but cannot guarantee stable operation of the input-series-connected systems (including ISOP and ISOS systems). (4) For the input-parallel-connected DC–AC power conversion systems (including IPOP and IPOS systems), both the compound balanced control strategy and output voltage/current sharing control strategy are effective. Compared to the compound balanced control strategy, the output voltage/current sharing control is quite simple and it is usually employed. For the input-series-connected DC–AC power conversion systems (including ISOP and ISOS systems), only the compound balanced control strategy could be adopted, that is, the equaling of the amplitude or phase of the modular output current (for ISOP system) or modular output voltage (for ISOS system) should be guaranteed together with the control for achieving input voltage sharing (IVS) among the modules.

References 1. Chen, W.: Research on series-parallel conversion systems consisting of multiple converter modules (in Chinese). Ph.D. Dissertation, Nanjing University of Aeronautics and Astronautics, Nanjing, China (2009) 2. Fang, T.: Control strategies for input-parallel-output-parallel and input-series-output-series connected combined inverter systems (in Chinese). Ph.D. Dissertation, Nanjing University of Aeronautics and Astronautics, Nanjing, China (2008)

Chapter 6

Compound Balanced Control Strategy for DC–AC Input-Series Output-Parallel Systems

Abstract In this paper, the implementation of the compound balanced control strategy is presented for the DC–AC input-series output-parallel (ISOP) systems to achieve input voltage sharing (IVS) and output current sharing (OCS). The inductor current and the output current adopted as the feedback variables for the inner current loop of each module, respectively, is analyzed and compared. Taking the outputcurrent-feedback scheme as an example, the decoupling relationship between the control loops is analyzed and the design of the control loops is given. The simulation and experimental results are provided to verify the effectiveness of the proposed control method. Furthermore, this chapter presents a new solution dedicated to distributed configuration on the basis of centralized control, which can improve the reliability of the systems. Keywords Input-series output-parallel system Compound balanced control strategy · Input voltage sharing Output current sharing · Closed-loops design DC–AC input-series-connected systems (including DC–AC ISOP systems and DC–AC ISOS inverter systems) are very suitable for high input DC voltage applications such as the urban railway systems, ship-electric-power-distribution systems, high-speed railway electrical systems, etc. The DC–AC input-series-connected systems have the following advantages: (1) easy selection of power switches due to the reduced voltage stress, (2) easy design due to the reduced power level of individual modules, and (3) high modularity and reliability [1, 2]. The DC–AC ISOP systems is quite suitable for high input voltage and high output current power conversion applications, and the primary control objective of the systems is to ensure input voltage sharing (IVS) and output current sharing (OCS) for all the constituent modules. Unlike DC–DC ISOP system, OCS for DC–AC ISOP system means both the magnitudes and phases of the modular output currents are equal due to AC output of DC–AC ISOP systems. In Chap. 5, it is pointed out that DC–AC ISOP system needs to adopt the compound balanced control strategy, i.e., to achieve IVS while ensuring that the magnitudes or phases of modular currents are equal. In this chapter, the implementation of the © Springer Nature Singapore Pte Ltd. and Science Press, Beijing, China 2019 X. Ruan et al., Control of Series-Parallel Conversion Systems, CPSS Power Electronics Series, https://doi.org/10.1007/978-981-13-2760-5_6

121

122

6 Compound Balanced Control Strategy for DC–AC …

compound balanced control strategy is presented first. In this method, either the inductor current or the output current can be adopted as the feedback variable for the inner current loop. The comparative analysis shows that the method adopting the output current as the feedback variable can realize IVS and OCS unconditionally. Then, taking the output-current-feedback method as an example, the decoupling relationship between the control loops is analyzed and the loop design procedure is given. Finally, the simulation and experimental results are provided to verify the effectiveness of the proposed control method. Furthermore, this chapter put forward a new solution dedicated to distributed configuration on the basis of centralized control, which can improve the reliability of the systems.

6.1 DC–AC Inverter Topology for the ISOP Systems In the DC–AC ISOP systems, the DC–AC inverter modules are connected in series at the input side and in parallel at the output side. This requires the DC–AC inverter modules that should have the function of galvanic isolation, which could be realized by a low-frequency transformer or a high-frequency transformer [3–5] as shown in Fig. 6.1. As shown in Fig. 6.1a, a line-frequency transformer is introduced at the output side of the DC–AC inverter. In addition to galvanic isolation, this transformer can step up/down the output voltage. However, the line-frequency transformer is bulky and heavy since the frequency is quite low, which is 50 Hz or 60 Hz. As shown in Fig. 6.1b, the isolated DC–AC inverter is a two-stage configuration. The first stage is an isolated DC–DC converter, which converts a DC input voltage

*

*

Vin

DC-AC Inverter

+ vo

Line-Frequency Transformer

_

(a) Galvanic isolation realized by alow-frequency transformer

*

* Vin

+

DC-DC Converter with High-Frequency Transformer

DC-AC Inverter

vo _

(b) Galvanic isolation realized by ahigh-frequency transformer Fig. 6.1 Isolated DC–AC inverter

6.1 DC–AC Inverter Topology for the ISOP Systems

123

into a regulated DC output voltage. Since the DC–DC converter is operated at high frequency, the transformer is quite small and light. The second stage is a DC–AC inverter, which converts the DC voltage obtained from the first stage into an AC output voltage. With the two-stage configuration, the DC–DC converter and DC–AC inverter could be optimally designed separately, and it is also contributed to the integration and modularization of the system. So, the high-frequency isolated inverter is chosen here as the basic inverter module. The isolated DC–DC converters can be buck-derived, boost-derived, and buckboost-derived ones. The buck-derived isolated DC–DC converters include forward, push-pull, half-bridge, full-bridge, etc. For the full-bridge converter, the power switches only sustains the input voltage, the utilization rate of the magnetic core is high, and it is suitable for medium-to-high power conversions. When employing the phase-shift control, the full-bridge converter could realize soft-switching of the power devices, leading to reduced switching loss and thus high conversion efficiency [6]. So, the phase-shifted full-bridge converter is adopted here for the DC–DC stage. Full-bridge inverter is suitable for high power applications and it can output high quality voltage since three voltage levels are available. Therefore, the full-bridge inverter is used here as the DC–AC stage. Figure 6.2 shows the main circuit of the two-stage inverter module.

Module 1 Full-Bridge DC-DC Converter Ldc1 iin1 Q1

icd1 + vin1 _

+

Q4

icdn Cdn

S1 A1 +

iin2 + vin2 _

Module 2

iinn + vinn _

Fig. 6.2 Main circuit of DC–AC ISOP system

Module n

Lf1

S3 iLf1

io1

_ B1 S4

io

iCf 1

vinv1

S2

DR2 DR4

…

Cd2

Tr

Vdc1 Cdc1 _

Q2

Vin icd2

Q3

* *

Cd1

iLdc1 DR1 DR3

…

Iin

Full-Bridge Inverter

Cf1

ZLd

+ vo _

124

6 Compound Balanced Control Strategy for DC–AC …

6.2 Implementation of the Compound Balanced Control Strategy Figure 6.3 shows the block diagram of the compound balanced control strategy for the DC–AC ISOP systems [7], which has three control loops, namely the common system output voltage loop, the input voltage sharing (IVS) loop, and the inner current loop for each inverter module. In Fig. 6.3, K vi is the sensor gain of module input voltage, K io is the sensor gain of module output current, K vo is the sensor gain of the system output voltage, Ginv is the transfer function from the modulating signal to the voltage across the phase-leg mid-points of the inverter, Gvo is the system output voltage regulator, Gvcd is the IVS regulator, and Gi is the inner current loop regulator. The system output voltage loop aims for regulating the system output voltage. The output of Gvo provides the common initial current reference iref for all the inner current loops. The IVS loop is used to achieve IVS among the constituent inverter modules. Generated by Gvcd , the dc deviation signal ivdj (j 1, 2, …, n) along with the initial current reference iref are sent to the multiplier to obtain the IVS regulating signal isvj (j 1, 2, …, n), which is then superimposed on the initial current reference iref to obtain irefj , which serves as the actual current reference of respective modules. The output current of each module follows irefj , and thus the module output current amplitude is regulated and the equivalence of the power factor of the module output current is also ensured. It can be seen that the control scheme can realize IVS and synchronization of the output current phase of each inverter module simultaneously, and consequently, IVS and OCS of the system can be achieved. At steady state, the compound balanced control strategy can realize IVS and OCS of each inverter module for the DC–AC ISOP system. Suppose a perturbation occurs to cause the variation of the input-dividing capacitor voltage. For example, vin1 ascends with its value higher than that of the average input voltage V in /n, which will result in the decrease of the output of IVS regulator ivd1 and the increase of the amplitude of corresponding current reference iref1 . Then, the amplitude of output current io1 traces that of iref1 to rise, which will result in the rise of the output active power of module 1. Therefore, the input power of module 1 will also increase and so is the input current iin1 , i.e., iin1 > iin . Since the input-dividing capacitor current icd1 equals to iin − iin1 , it is obvious that icd1 < 0, which means that the capacitor C d1 is discharged and vin1 will decrease until it returns to V in /n. On the other hand, if a perturbation leads to the descent of input voltage of some module with its value less than V in /n, the input voltage will also be adjusted and ultimately rise back to V in /n. It can be seen that the compound balanced control strategy could also achieve IVS and OCS at dynamic state. It is noteworthy that the feedback signal of modular inner current loop in the control strategy proposed in Fig. 6.3 is the output current ioj , and it also could be the output filter inductor current iLfj . The DC–AC ISOP systems uses different current feedback signal will get different OCS effect. If the output filter inductor current feedback is adopted [8], that is, the output filter inductor current of each module is employed as the inner current loop feedback on the basis of IVS, the output filter

6.2 Implementation of the Compound Balanced Control Strategy Inner Current Loop

IVS Loop Vin n

Kvi vin1

+ _

Gvcd

ivd1

isv1

iref1 _

Kvi

Gi

Ginv

+ vinv1

Inner Current Loop

+ _

Gvcd

ivd2

isv2

iref2 _

Kvi

+ + _

Gi

Ginv

iCf1 1 sLf1

iLf1

_

io1

+

+ vinv2

_

iCf2 _ io2 1 iLf2 sLf2 +

…

…

+

Inner Current Loop

IVS Loop Kvi vinn

sCf1

sCf2

Kio …

vin2

+ + _

_

Kio

IVS Loop Kvi

125

+ _

Gvcd

ivdn

isvn

Kvi

+ voref _

irefn _

+ + _

Gi

Ginv

+ vinvn

_

iCfn 1 sLfn

iLfn

_

ion

+ io ZLd +

vo

sCfn

+

Kio Gvo

iref

Kvo Output Voltage Loop

Fig. 6.3 Block diagram of compound balanced control strategy for DC–AC ISOP system

capacitor of each inverter module needs to be identical to ensure OCS (which means the equivalence of both the active and reactive current). However, the output filter capacitors of different modules are not matched in the actual prototype. In order to simplify the analysis, a two-module DC–AC ISOP system is taken as an example to analyze the effect of unmatched output filter capacitors on the sharing of output current. The equivalent circuit of two-module DC–AC ISOP system is shown in Fig. 6.4, where it is supposed that C f1 > C f2 . Figure 6.5a shows the phasor diagram of voltages and currents of two-module DC–AC ISOP system with output filter inductor current feedback, and here, the output filter inductor currents of two inverter modules are equal, i.e., I˙Lf1 I˙Lf2

(6.1)

I˙o1 I˙Lf1 − I˙Cf1 I˙o2 I˙Lf2 − I˙Cf2

(6.2a)

From Fig. 6.4, we have

Combining (6.1) with (6.2a, 6.2b), we have

(6.2b)

126

6 Compound Balanced Control Strategy for DC–AC …

Fig. 6.4 Equivalent circuit of two-module DC–AC ISOP system

· ICf2

· · ILf1–ILf2

· ICf1

· ICf1 · ICf2

· · Io1–Io2 · Vo

0 · Io1

(a) With output filter inductor current feedback

0

· Vo

(b) With output current feedback

Fig. 6.5 Phasor diagrams for different current feedback

I˙o1 − I˙o2 I˙Cf2 − I˙Cf1 jωo (Cf2 − Cf1 )V˙o

(6.3)

where ωo is the fundamental angular frequency of the inverter system. Since C f1 > C f2 , it can be derived from (6.3) that the output current difference of the two inverter modules, I˙o1 − I˙o2 , lags the output voltage V˙o by 90°, and the inductive reactive power is presented at the output side of the system. On the contrary, the capacitive reactive power will be presented if C f1 < C f2 . So, it can be seen that the unmatched output filter capacitors induce the reactive circulating current between the inverter modules and OCS cannot be truly realized when the output filter inductor current feedback control is adopted. For DC–AC ISOP systems, if the output current feedback is adopted directly [7], that is, the output current of each inverter module is used instead of the output filter inductor current as the feedback of the inner current loop of the inverter module on the basis of IVS, whether the output filter capacitors are matched or not will have no effects on OCS (which means both the active and reactive currents of the modules are equal, respectively). Figure 6.5b shows the phasor diagram of the voltages and currents of two-module DC–AC ISOP system with the output current feedback control, and here we have I˙o1 I˙o2 at steady state. Thus, it can be derived from (6.2) that I˙Lf1 − I˙Lf2 I˙Cf1 − I˙Cf2 jωo (Cf1 − Cf2 )V˙o . Since C f1 > C f2 , the difference of the output filter inductor currents of two inverter modules will lead the output voltage V˙o by 90°, and the capacitive reactive power is presented at the output side of the system. On the contrary, the inductive reactive power will be presented if C f1 < C f2 . Through the above analysis, we can find that for the DC–AC ISOP systems, when the output filter inductor current feedback control is adopted, OCS can be achieved

6.2 Implementation of the Compound Balanced Control Strategy

+

Vdcg Vdcj

_ Kvdc

Gvdc

Comparator − Phase-Shifted Control Circuit +

127

Drive Signals

VRAMP

Fig. 6.6 Closed-loop control block diagram of full-bridge converter as the first stage

only if the output filter capacitors are matched. If the output current feedback control is employed, OCS could be truly achieved whether the output filter capacitors are matched or not.

6.3 Closed-Loops Design for DC–AC ISOP Systems Taking the output current feedback control given in Sect. 6.2 as an example, the relationship of the control loops and the design of the control loops will be presented. As shown in Fig. 6.3, the compound balanced control strategy includes three closed loops, namely the system output voltage loop, IVS loop, and inner current loop. In addition, for the phase-shifted full-bridge converter as the first stage of each module, Fig. 6.6 shows the control block diagram, where the output voltage of the full-bridge converter, V dcj , is sensed with the sensor gain of K vdc , and then is compared with voltage reference V dcg , and the error is sent into the output voltage regulator Gvdc . The output of Gvdc is compared with the sawtooth carrier V RAMP , and the output of the comparator is sent to the phase-shifted control circuit, generating the drive signals.

6.3.1 Decoupling Relationship Between IVS Loop and Output Voltage Loop As seen from Fig. 6.3, the output of the IVS loop of each inverter module, isvj (j 1, 2, …, n), is a part of the reference of the inner current loop of each inverter module, so the IVS loop and the output voltage loop are coupled each other, and it is very difficult to design the system directly. So, the analysis of the control block diagram will be given below. From Fig. 6.3, the reference of the inner current loop of each inverter module is the difference between the output of the output voltage regulator iref and the output of each IVS loop isvj (j 1, 2, …, n), i.e., Iref j (s) Iref (s) − Isv j (s) ( j 1, 2, . . . , n)

(6.4)

128

6 Compound Balanced Control Strategy for DC–AC …

Assuming the parameters of the inverter modules are consistent and simplifying the inner current loop of each inverter module, the transfer function of each inner current loop is equal, expressed as G c j (s)

Io j (s) G c (s) ( j 1, 2, . . . , n) Iref j (s)

(6.5)

Then, Fig. 6.3 can be equivalent to Fig. 6.7a, and the output current of each inverter module is Io j (s) G c (s) · Iref j (s) G c (s) · Iref (s) − Isv j (s) G c (s) · Iref (s) − G c (s) · Isv j (s)

( j 1, 2, . . . , n)

(6.6)

According to (6.6), the total output current of the DC–AC ISOP system can be obtained as Io (s)

n

Io j (s) n · G c (s) · Iref (s) − G c (s) ·

j1

n

Isv j (s)

(6.7)

j1

According to (6.7), Fig. 6.7a can be equivalent to Fig. 6.7b, from which, the influence of the IVS loops on the output voltage loop of the DC–AC ISOP system can be derived as n

Isv j (s)

j 1

n j1

K vi

⎛

Vin − vin j G vcd (s)Iref (s) K vi G vcd (s)Iref (s) ⎝Vin − n

n

⎞ vin j ⎠

j1

0 (6.8) Equation (6.8) implies that the total effect of the IVS loops on the system output voltage loop is zero (as shown in Fig. 6.7c). That is to say, the output voltage loop is not affected by the IVS loops. On the other hand, the output signal of the output voltage regulator iref provides the initial reference for each inner current loop. When iref changes, the actual reference of each module, irefj , will also change simultaneously so that the output power and input power of each module vary accordingly. For example, when IVS and OCS are both achieved initially, the abruptly increasing load would result in a decrease of the output voltage. In order to ensure the stability of the output voltage, iref will increase and the output current of each module will also rise. In the meantime, OCS and the balanced output power of all modules are maintained. By power conservation, the input current of each module and hence the total input current of the system will all

6.3 Closed-Loops Design for DC–AC ISOP Systems

129

Fig. 6.7 Equivalent block diagram of the output current feedback control for DC–AC ISOP system

130

6 Compound Balanced Control Strategy for DC–AC …

Fig. 6.7 (continued)

increase. Since the input current is still being shared and the average current of each input-dividing capacitor is zero, IVS is maintained. It can be seen that the operation of the output voltage loop does not change the state of the input voltage of each module, and so does not affect the operation of the IVS loop. The above analysis shows that the IVS loop and output voltage loop are decoupled and can be designed separately when all the inverter modules are identical.

6.3.2 Design of Control Loop of the First-Stage Full-Bridge Converter Figure 6.8 shows the topology of the basic inverter module for the DC–AC ISOP system, which is composed of a full-bridge converter as the first stage and a fullbridge inverter as the second stage. If the output voltage frequency of the inverter module is 400 Hz, the input current of the second-stage, iac , contains not only highfrequency (switching frequency) ripple current, but also the low-frequency (800 Hz) ripple current. Therefore, the first-stage converter needs a larger output filter capacitor C dc to filter the low-frequency ripple current for the purpose of reducing the ripple

6.3 Closed-Loops Design for DC–AC ISOP Systems

131 800Hz

Ldc

Iin

Q1

Q3

Tr

DR1 DR3 +

S1 A +

* *

Q2

Q4

DR2 DR4

Lf

S3 iLf vinv

Vdc Cdc _

Vin

400Hz

iac

iLdc

_ B S2

io

iCf Cf

ZLd

+ vo _

S4

Fig. 6.8 The main circuit of single inverter module

of intermediate bus voltage, V dc . However, C dc cannot be infinite, so there is still low-frequency ripple component of 800 Hz in V dc . In order to reduce the influence of 800 Hz ripple on the voltage loop, the cutoff frequency of its voltage loop should be designed to be lower than 800 Hz and it is generally about 1/10 of 800 Hz [9, 10]. In the following, simulation is carried out for the two-stage inverter as shown in Fig. 6.8. The key specifications are as follows: the input voltage V in 270 V, the average value of the intermediate bus voltage V dc 180 V, the output voltage vo 115 V/400 Hz, and the output capacity S o 1 kVA. For the first-stage full-bridge converter, the key parameters are as follows: the transformer turns ratio K 15:13, the output filter inductor L dc 230 µH, the output filter capacitor C dc 800 µF (with the equivalent series resistance RCf 0.03 ), the output voltage sensor gain K vo 0.01, the peak-to-peak value of sawtooth carrier V pp 2.1 V, and the switching frequency f s 100 kHz. For the second-stage full-bridge inverter, the key parameters are as follows: the output filter inductor L f 400 µH and the output filter capacitor C f 16 µF. According to Fig. 6.6, the control block diagram of the output voltage loop of the first-stage full-bridge converter is shown in Fig. 6.9, where the control-to-output transfer function Gdc (s) is [11] Vin (1 + sCdc RCf ) K G dc (s) Rp RCf L dc RCf +s + Cdc RCf + s 2 L dc Cdc 1 + + Rp Cdc 1 + +1 RLd RLd RLd RLd (6.9) where Rp 4L r f s /K 2 , and RLd is the equivalent load of the first-stage full-bridge converter, which is obtained from the second-stage inverter and whose value is determined by the output active power of the inverter. Considering that the inverter operates with full resistive load, RLd V 2dc /Po 1802 / 1000 32.4 . From Fig. 6.9, the output voltage loop gain of the first-stage full-bridge converter can be obtained as

132

6 Compound Balanced Control Strategy for DC–AC …

vˆdcg

+ _

Gvdc(s)

1/Vpp

dˆy

Gdc(s)

vˆdc

Kvdc

Fig. 6.9 Control block diagram of the output voltage loop of the first-stage full-bridge converter Fig. 6.10 Bode diagram of the output voltage loop gain of the full-bridge converter at different cutoff frequencies

Table 6.1 Output voltage regulator Gvdc (s) at different cutoff frequencies Cutoff frequency/Hz Gvdc (s) 1000

9 + 1.4 × 104 /s

300

2 + 2500/s

80

0.5 + 500/s

Tvo (s) K vdc G vdc (s)G dc (s)/Vpp

(6.10)

Figure 6.10 gives the bode diagram of the output voltage loop gain of the fullbridge converter with different cutoff frequencies, i.e., 1 kHz, 300, and 80 Hz, and the corresponding transfer functions of output voltage regulator Gvdc (s) are shown in Table 6.1 Figure 6.11 shows the simulation waveforms of the primary current of transformer ip , the filter inductor current iLdc , the intermediate bus voltage V dc , and the inverter output voltage at different cutoff frequencies. Figure 6.11a shows simulation waveforms at the cutoff frequency of 1 kHz. As seen, the fluctuation of ip and iLdc of the first-stage full-bridge converter is relatively large, and the maximum value is almost twice of the average value. This is because the voltage loop tends to follow and compensate for the ripple voltage at 800 Hz, which will increase the current stress of the power devices and magnetic components. Figures 6.11b, c show the simulation waveforms at the cutoff frequencies of 300 and 80 Hz, respectively. It can be seen

6.3 Closed-Loops Design for DC–AC ISOP Systems

133

that the pulsation of 800 Hz in ip and iLdc is reduced with the decrease of the cutoff frequency. Therefore, in order to reduce the current stress of the power devices and the magnetic components of the full-bridge converter, the cutoff frequency of the voltage loop is designed relatively low, and it is set at about 1/10 of the output frequency.

6.3.3 Design of IVS Loop As illustrated in Sect. 6.3.1, the IVS loop and the system output voltage loop are decoupled each other, so the IVS loop can be designed separately. According to Fig. 6.3, the load of each inverter module can be considered to be composed of two parts, which are determined, respectively, by iref , the output of the system output voltage loop, and isvj (j 1, 2, …, n) introduced by each IVS loop. So, from the input side of the inverter module, each inverter can be regarded as resistances Rvo and RIVSj paralleled as shown in Fig. 6.12. The value of Rvo and RIVSj are determined by iref and isvj (j 1, 2, …, n), respectively.

6.3.3.1

Derivation of Rvo

If the IVS loops are ignored (see Fig. 6.3), i.e., iˆsv1 iˆsv2 . . . iˆsvn 0, iref is the common reference of the inner current loop of all inverter modules, and the method will be equivalent to OCS control. So, from the input side, under the role of iref , each inverter module can be equivalent to a constant power load. Within the cutoff frequency range of the current loop, the small-signal model of a constant power load presents the negative resistance characteristic. From (6.8), it can be seen that the sum of isvj (j 1, 2, …, n) of each inverter module equals to zero, which means that the sum of each module power corresponding to isvj also equals to zero and thus the sum of each inverter module power corresponding to iref serves as the power of the DC–AC ISOP systems. Therefore, the active power of each inverter module corresponding to iref is equal, and it is 1/n of the system power Po . So the resistance Rvo can be expressed as Rvo

6.3.3.2

vˆin j iˆin_vo j

−Rn −

(Vin /n)2 ( j 1, 2, . . . , n) Po /n

(6.11)

Derivation of RIVSj

From Fig. 6.7a, the individual output current perturbation Iˆo_IVS j ( j 1, 2, . . . , n) of inverter introduced by the input voltage perturbation vˆin j ( j 1, 2, . . . , n) can be expressed as

6 Compound Balanced Control Strategy for DC–AC …

(V)

(V)

(A)

(A)

134

10 0 −10 12 6 0 185 180 175 200 0 −200

ip

iLdc Vdc

vo 10

11

12

13

t (ms)

14

15

16

17

16

17

16

17

(V)

(V )

(A)

(A)

(a) With voltage loopcut-off frequencyof 1 kHz

10 0 −10 10 5 0 185 180 175 200 0 −200

ip

iLdc Vdc

vo 10

11

12

13

t (ms)

14

15

(V)

(V)

(A)

(A)

(b) With voltage loop cut-off frequency of 300 Hz

10 0 −10 10 5 0 185 180 175 200 0 −200

ip

iLdc Vdc vo 10

11

12

13

t (ms)

14

15

(c) With voltage loop cut-off frequency of 80 Hz

Fig. 6.11 Simulation waveforms at different cutoff frequencies

6.3 Closed-Loops Design for DC–AC ISOP Systems

135

iˆin1

Fig. 6.12 Equivalent small-signal model of DC–AC ISOP system

+ ˆvin1 _

Z1′

Cd1

iˆin_vo1 Rvo

iˆin_IVS1 RIVS1

iˆin_vo2 Rvo

iˆin_IVS2 RIVS2

iˆin_von Rvo

iˆin_IVSn RIVSn

iˆin2 + vˆ in2 _

Cd2 ˆiinn

…

vˆ in

Z 2′

+ ˆvinn _

Z n′

Cdn

Iˆo_IVS j vˆin j · K vi · G vcd · Iref (s) · G c (s) ( j 1, 2, . . . , n)

(6.12)

where Iˆo_IVS j has the same phase with the corresponding load current Iˆo j ( j 1, 2, . . . , n). The individual active output power perturbation pˆ o_IVS j ( j 1, 2, . . . , n) of module j corresponding to I o_IVS j can be expressed as

pˆ o_IVS j I o_IVS j · Vo · cos θ ( j 1, 2, . . . , n)

(6.13)

The individual input power perturbation pˆ in_IVS j ( j 1, 2, . . . , n) of module j corresponding to vˆin j can be expressed as pˆ in_IVS j Iin vˆin j +

Vin ˆ i in_IVS j ( j 1, 2, . . . , n) n

(6.14)

Assuming the conversion efficiency for each individual module is 100%, we have pˆ in_IVS j pˆ o_IVS j ( j 1, 2, . . . , n)

(6.15)

Substituting (6.12) into (6.13), and then substituting (6.13) and (6.14) into (6.15), yields RIVS j

vˆin j iˆin_IVS j

Vin /n ( j 1, 2, . . . , n) K vi G vcd Iref (s)G c (s) · Vo cos θ − Iin (6.16)

According to Fig. 6.7c, we have I ref (s) Gc (s) I o /n. So, I ref (s) Gc (s) V o cosθ equals to Po /n. Then, (6.16) can be rewritten as

136

6 Compound Balanced Control Strategy for DC–AC …

RIVS j

6.3.3.3

Vin /n Vin ( j 1, 2, . . . , n) (6.17) K vi G vcd Po /n − Iin K vi G vcd Po − n Po /Vin

Design of IVS Regulator Gvcd

As pointed out in Chap. 3, the OCS control could not ensure the stability of the DC–DC ISOP system. The essential reason is that the input impedance of each module is negative resistance. Similarly, for the DC–AC ISOP system, the OCS control will lead to an unstable system since the modules present the negative resistance at the input sides. Equation (6.17) implies that the role of IVS loops is equivalent to introduce a positive resistance to the input terminal of each module as shown in Fig. 6.12. So, the equivalent input impedance of each inverter module is Z j Rvo //RIVS j

−Rn · RIVS j Rn · RIVS j −Rn + RIVS j Rn − RIVS j

( j 1, 2, . . . , n)

(6.18)

In order to ensure the stability of DC–AC ISOP system, the equivalent input impedance of each inverter module should present positive resistance characteristics, i.e., Z j > 0. According to (6.18), we can obtain the condition for stability as Rn > RIVS j ( j 1, 2, . . . , n)

(6.19)

Substitution of (6.11) and (6.17) into (6.19), yields G vcd >

2n G vcd min K vi · Vin

(6.20)

It can be seen that the gain of IVS loop compensator must satisfy (6.20) so that each inverter module presents positive resistance characteristics and the DC–AC ISOP system with IVS control strategy is stable. For simplicity, a two-module DC–AC ISOP system is simulated to verify the correctness of the IVS loop design. The specifications of the DC–AC ISOP system are as follows: the system input voltage is 540 V (±10%), the output voltage is 115 VAC/400 Hz, the output capacity is 2 kVA; the module input voltage is 270 V (±10%), the module output voltage is 115 VAC/400 Hz, the output capacity of each module is 1 kVA; the input voltage sensor gain is K vi 0.01. The input-dividing capacitors of the two modules are intentionally set at different values, i.e., C d1 1000 µF and C d2 800 µF. Figure 6.13 shows the response in the individual input voltages and output currents of two modules when the system input voltage is step changed from 500 to 600 V. According to (6.20), it can be obtained that Gvcdmin 0.666 ~ 0.815. As shown in Figs. 6.13a, b, for Gvcd > Gvcdmin , the two input voltages converge finally, and the higher values of Gvcd result in a faster correction in the input voltages after a system input voltage disturbance. As shown in Fig. 6.13c, for Gvcd < Gvcdmin , the two input

6.3 Closed-Loops Design for DC–AC ISOP Systems

320

vin2 vin1

(V) 280 240 20

io1

10 (A)

137

io2

0 10 20

10

0

20 t (ms)

30

40

30

40

(a) G vcd = 8

320

vin2 vin1

(V) 280 240 20

io1

io2

10 (A)

0 10 20

0

10

20 t (ms) (b) G vcd = 3

Fig. 6.13 Response to a step change in the system input voltage with different Gvcd

voltages diverge, resulting in a runaway mode. The system cannot work normally and it is unstable.

6.3.4 Design of Output Voltage Loop As shown in Fig. 6.7c, the system output voltage loop is decoupled from the IVS loops. In order to design the compensator for system output voltage loop, it is required

138

6 Compound Balanced Control Strategy for DC–AC …

320

vin2 11.66V

15.43V

vin1

(V) 280 240 20

io1

io2

10 (A)

0

10 20

0

5

15

25

35 t (ms)

45

55

65 70

(c) G vcd = 0.25

Fig. 6.13 (continued)

to derive the transfer function of inner current loop Gc (s) of a single inverter first. Based on the control strategy shown in Fig. 6.3, when the IVS loops are not considered, the control block diagram of a single inverter module can be obtained as shown in Fig. 6.14a. Relocating the feedback node of io backward to the output voltage, Fig. 6.14b is obtained. Simplifying the part of the diagram in the dashed block of Fig. 6.14b, c can be obtained. From Fig. 6.14c, the transfer function of inner current loop Gc (s) can be derived as G c (s)

G i G inv Io (s) 2 Iref (s) s n Z Ld L f Cf + s L f + n Z Ld + K io G i G inv

(6.21)

From (6.21) and Fig. 6.7c, the loop gain of the system output voltage loop can be derived as Tvo (s)

s2n Z

n K vo G vo G i G inv Z Ld Ld L f C f + s L f + n Z Ld + K io G i G inv

(6.22)

6.4 Simulation Results In order to verify the effectiveness of the IVS and OCS control strategy based on the output current feedback and the output filter inductor current feedback for the DC–AC ISOP system, the simulation of two-module DC–AC ISOP system is carried out. The parameters used in the simulation are the same as those given in Sect. 6.3.3. To obtain the effects of output filter capacitor mismatch on two kinds of control

6.4 Simulation Results

139

iCf

(a) voref +

_

Gvo

iref + _

Gi

Ginv

vinv

_

1 iLf sLf +

+

_

sCf nZLd

io

vo

Kio Kvo

(b) voref + _

Gvo

iref + _

Gi Kio

Ginv io

vinv

_

iCf sCf _ 1 iLf nZLd sLf + io

vo

1 nZLd

Kvo

(c) voref +

_

Gvo

iref + _

Gi Kio

Ginv io

vinv

nZ Ld s 2 nZ Ld Lf Cf sLf

nZ Ld

vo

1 nZLd

Kvo Fig. 6.14 Equivalent control block diagram of output voltage loop of individual module

methods, a 4.7 µF capacitor is intentionally connected in parallel with the output filter capacitor of module 1. Figure 6.15 shows the simulation waveforms of the DC–AC ISOP system with output current feedback control. The waveforms of the individual input voltage and the difference between input voltages are shown in Fig. 6.15a. As seen, the input voltages of the two inverter modules are equal. The waveforms of the system output voltage, individual output filter inductor current and the difference between these two currents are shown in Fig. 6.15b, from which it can be seen that the output filter inductor current of the two inverter modules are different, and the difference leads the output voltage by 90° with the capacitive reactive power being presented. Figure 6.15c shows the waveforms of individual output current and the difference between these two currents. Due to the adoption of the output current feedback, the output filter capacitor mismatch does not affect the OCS effect, and the output current of the two inverter modules is equal. Figure 6.16 shows the simulation waveforms of the DC–AC ISOP system with output filter inductor current feedback control. The waveforms of individual input voltage and the difference between input voltages are shown in Fig. 6.16a. As seen, the input voltages of the two inverter modules are equal. The waveforms of the output voltage of the system, individual output current and the difference between these two currents are shown in Fig. 6.16b, from which it can be seen that the output currents

140

6 Compound Balanced Control Strategy for DC–AC … 280 270 (V) 260 250 280 (V) 270 260 250 5 (V)

vin1

vin2

vin1 –vin2

0 −5

0

2

4

t (ms)

6

8

10

(a) Simulationwaveformsof input voltages

200 (V)

vo

0 −200 20

(A)

0 −20 20

(A)

0

iLf1

iLf2

−20 5 (A) 0 −5

iLf1 – iLf2 0

2

4

t (ms)

6

8

10

(b) Simulationwaveformsof output filter inductor currents

15 (A)

0 −15 15

(A)

io1

io2

0 −15 5

(A)

0 −5 0

io1 –io2

2

4

t (ms)

6

8

(c) Simulation waveforms of output currents

Fig. 6.15 Simulation waveforms with output current feedback control

10

6.4 Simulation Results

141

of the two inverter modules are different, and the difference lags the output voltage by 90° with the inductive reactive power being presented. Figure 6.16c shows the waveforms of individual output filter inductor current and the difference between these two currents. Due to the adoption of the output filter inductor current feedback, the output filter inductor current of the two inverter modules are equal. Figures 6.15 and 6.16 verify the validity of the previous theoretical analysis.

6.5 Experimental Results In order to verify the compound balanced control strategy with output current feedback controlled, a 2-kVA two-module DC–AC ISOP system prototype is fabricated in the laboratory. The specifications of the DC–AC ISOP system and the single inverter module are the same as the parameters in the simulation given in Sect. 6.4.

6.5.1 Steady-State Experiment Figure 6.17 gives the steady-state experimental waveforms of the system at full resistive load (2 kW). The waveforms of individual input voltage, individual output current, and the difference between these two currents are shown in Fig. 6.17a. As seen, the input voltage and output current of the two inverter modules are the same, respectively, which means that IVS and OCS are achieved. The waveforms of the system output voltage, total output current, and individual output current are shown in Fig. 6.17b, which indicate that the system outputs a stable sinusoidal voltage. Figure 6.18 gives the steady-state experimental waveforms of the system at full inductive load (2 kW, PF 0.75). The waveforms of the individual input voltage, individual output current, and the difference between these two currents are shown in Fig. 6.18a, from which we can see that IVS and OCS are well achieved. Figure 6.18b shows the waveforms of the system output voltage, total output current, and individual output current. Due to the inductive load, the output voltage leads the total output current.

6.5.2 Dynamic Experiment Figures 6.19 and 6.20 give the experimental waveforms of the system at nominal input voltage when the load current steps up and down between half and full resistive load, and between half and full inductive load, respectively. Figures 6.19a and 6.20a show the waveforms of the individual input voltage, system output voltage, and total output current. As seen, the input voltages are shared equally during the transient condition. It should be noted that the input voltage source of the system is not an ideal one,

142

6 Compound Balanced Control Strategy for DC–AC … 280 270 (V) 260 250 280 (V) 270 260 250 5 (V)

vin1

vin2

vin1 –vin2

0 −5

0

2

4

t (ms)

6

8

10

8

10

(a) Simulationwaveformsof input voltages

200 (V)

vo

0 −200 15

(A)

0 −15 15

(A)

0

io1

io2

−15 5 (A) 0 −5

io1 – io2 0

2

4

t (ms)

6

(b) Simulationwaveformsof output currents

20 (A)

iLf1

0 −20 20

(A)

iLf2

0 −20 5

(A)

iLf1 –iLf2

0 −5

0

2

4

t (ms)

6

8

10

(c) Simulationwaveformsof output filter inductor currents

Fig. 6.16 Simulation waveforms with output filter inductor current feedback control

6.5 Experimental Results

143

C1 vin1 : [100V/div] C2 vin2 : [100V/div] C1

C3 io1 : [25A/div]

C3 C2

C4 io2 : [25A/div]

C4 F2

F2 io1 io2 : [25A/div] Time: [2ms/div] (a) Experimental waveformsof input voltages and output currents

C1 vo : [200V/div] C1

F2

C3

F2 io : [25A/div]

C3 io1 : [25A/div]

C4

C4 io2 : [25A/div] Time: [2ms/div] (b) Experimental waveformsof output voltages and output currents

Fig. 6.17 Experimental waveforms at full resistive load

the input voltage has slight drop or rise when the load current steps up and down. Figures 6.19b and 6.20b show the waveforms of the individual input voltage and output current. As seen, the output currents are shared equally both at steady state and transient. Figures 6.21 and 6.22 give the experimental waveforms of the input voltage stepping between 486 V (90%V in ) and 594 V (110%V in ) at full resistive and inductive load, respectively. Figures 6.21a and 6.22a show the waveforms of individual input voltage, system output voltage, and total output current. As seen, the input voltages of the two modules are shared equally during the transient condition and the system

144

6 Compound Balanced Control Strategy for DC–AC …

C1 vin1 : [100V/div] C2 vin2 : [100V/div] C1

C3 io1 : [25A/div]

C3 C2

C4 io2 : [25A/div]

C4 F2

F2 io1 – io2 : [25A/div] Time: [2ms/div] (a) Experimental waveformsof input voltages and output currents

C1 vo : [200V/div] C1

F2

F2 io : [25A/div] C3

C3 io1 : [25A/div]

C4

Time: [2ms/div]

C4 io2 : [25A/div]

(b) Experimental waveformsof output voltages and output currents

Fig. 6.18 Experimental waveforms at full inductive load

output voltage is stable. Figures 6.21b and 6.22b show the waveforms of the individual input voltage and output current. As seen, the output currents of the two modules are shared equally during the transient condition.

6.5 Experimental Results

145

C1 vin1 : [100V/div] C2 vin2 : [100V/div] C1

C3 vo : [200V/div]

C3 C2 C4

C4 io : [25A/div] Time: [5ms/div] (a) Experimental waveforms of input voltages, output voltage and output currents

C1 vin1 : [100V/div] C2 vin2 : [100V/div] C4 io1 : [10A/div]

C1 C4 C2

C3

Time: [5ms/div]

C3 io2 : [10A/div]

(b) Experimental waveforms of input voltages and individual output currents

Fig. 6.19 Experimental waveforms under step change of resistive load

6.5.3 Comparison of the Output Current Feedback and Output Filter Inductor Current Feedback Further comparative experiments on the compound balanced control strategies with the output current feedback and the output filter inductor current feedback are conducted. Figure 6.23a gives the experimental waveforms of the system with output current feedback control. It can be seen that the mismatch of output filter capac-

146

6 Compound Balanced Control Strategy for DC–AC …

C1 vin1 : [100V/div] C2 vin2 : [100V/div] C1

C3 vo : [200V/div]

C3 C2 C4

C4 io : [25A/div] Time: [5ms/div] (a) Experimental waveforms of input voltages, output voltage and output currents

C1 vin1 : [100V/div] C2 vin2 : [100V/div] C1

C4 io1 : [10A/div]

C4 C2

C3

C3 io2 : [10A/div] Time: [5ms/div] (b) Experimental waveforms of input voltages and individual output currents

Fig. 6.20 Experimental waveforms under step change of inductive load

itor has no influence on the sharing of output currents between the two modules. However, the unmatched output filter capacitors introduce the difference between the output filter inductor currents. Figure 6.23b gives the experimental waveforms of the system with output filter inductor current feedback control. As seen, the output filter inductor current of the two inverter modules are almost the same, but there is reactive current between the two inverter modules, which is consistent with the previous analysis.

6.5 Experimental Results

147

C4 vin1 : [100V/div] C3 vin2 : [100V/div] C2 vo : [200V/div] C2/C4 C3 C1

C1 io : [25A/div] Time: [10ms/div] (a) Experimental waveforms of input voltages, output voltage and output currents

C4 vin1 : [100V/div] C2 io1 : [10A/div]

C3 vin2 : [100V/div]

C2/C4 C3

C1

Time: [10ms/div]

C1 io2 : [10A/div]

(b) Experimental waveforms of input voltages and individual output currents

Fig. 6.21 Experimental waveforms with step change of system input voltage at full resistive load

The steady state and dynamic experiments prove that the IVS and OCS control strategy based on output current feedback can achieve IVS and OCS under various circumstances. Furthermore, with the method adopted, the disturbance of input voltage and load can be restrained and the reliability of the system can also be guaranteed.

148

6 Compound Balanced Control Strategy for DC–AC …

C1 vin1 : [100V/div] C2 vin2 : [100V/div] C3 vo : [200V/div] C1/C3 C2 C4

C4 io : [25A/div] Time: [10ms/div] (a) Experimental waveforms of input voltages, output voltage and output currents

C1 vin1 : [100V/div] C3 io1 : [10A/div]

C2 vin2 : [100V/div]

C1/C3 C2

C4

Time: [10ms/div]

C4 io2 : [10A/div]

(b) Experimental waveforms of input voltages and individual output currents Fig. 6.22 Experimental waveforms with step change of system input voltage at full inductive load

6.6 Distributed IVS/OCS Control Strategy In the control scheme shown in Fig. 6.3, all the inverter modules share a common voltage loop, and this control scheme is actually a kind of centralized control strategy. Therefore, a distributed control is presented here to make sure that all the inverter modules are fully modularized. The failure of any module will not affect the nor-

6.6 Distributed IVS/OCS Control Strategy

149 C3 vo : [200V/div] C1 io1 : [25A/div] C2 io2 : [25A/div] M4 io1 – io2: [5A/div] M2 iLf1: [25A/div] M1 iLf2: [25A/div] M3 iLf1 – iLf2: [10A/div]

Time: [2ms/div] (a) Experimental waveforms with output current feedback control

C2 vo : [200V/div] M1 io1 : [25A/div] M2 io2 : [25A/div] M3 io1 – io2: [10A/div] C1 iLf1: [25A/div] C3– iLf2: [25A/div] F1 iLf1 – iLf2: [5A/div] Time: [2ms/div] (b) Experimental waveforms with output filter inductor current feedback control

Fig. 6.23 Experimental waveforms of two control strategies with different current feedback when the output filter capacitors are unmatched

mal operation of the DC–AC ISOP systems, so as to improve the reliability and redundancy of the DC–AC ISOP systems. Figure 6.24 presents the distributed control strategy for the DC–AC ISOP system [12], which also includes the output voltage loop, IVS loop, and inner current loop. Different from the centralized control strategy, the output voltage loops are scattered in each inverter module. All the modules communicate with each other by three buses, namely the output voltage reference synchronous bus (vref synchronous bus), the IVS bus, and the average current bus (iave bus). It can be seen that all the inverter modules in the system are completely reciprocal, and so distribution and modularity are fully achieved. In this distributed control strategy, the output voltage of each inverter module is regulated by its respective output voltage loop. All the output voltage references are synchronized by the digital signal processor (DSP) to generate the output voltage reference synchronous bus. Since the components of the feedback networks and regulators maybe not well matched, the output signals of every output

150

6 Compound Balanced Control Strategy for DC–AC … iave ivd1

+ Kvi _ Gvcd vin1 Kvi IVS Loop ig1 voref + Gvo _

ivd2

ig2

…

…

+

Gvo

io1

isv2 Inner Current Loop iCf2 _ iref2 iLf2 _ sCf2 vinv2 _ + + 1 Gi Ginv _ sLf2 + io2 + Ki

+ Kvi _ Gvcd vinn Kvi IVS Loop voref + _

1 sLf1 +

iCf1 sC _ f1

Kvo

Output Voltage Loop

vref IVS Synchronous Bus Bus

iLf1

Kvo

+ Kvi _ Gvcd vin2 Kvi IVS Loop Gvo

_

Ki

Output Voltage Loop

voref + _

isv1 Inner Current Loop _ iref1 vinv1 + + Gi Ginv _ +

ign

Output Voltage Loop

+ +

1 n

ivdn

iave Bus

+

+ io

…

Vin n

…

voref

+

ZLd

vo

isvn Inner Current Loop iCfn _ irefn iLfn _ sCfn vinvn _ + + 1 Gi Ginv _ sLfn + ion + Ki Kvo

Fig. 6.24 Block diagram of distributed control strategy for DC–AC ISOP system

voltage regulators (namely igj ) are not exactly the same. So, they are averaged, forming the iave bus. The signal iave is sent to all the inverter modules and served as the common initial current reference. The IVS bus signal comes from all the modules through an averaging circuit and its value is V in /n, which provides the input voltage reference for all the inverter modules. The output signals of modular IVS loops ivdj (j 1, 2, …, n) multiply iave and then superimpose on iave to produce the actual current reference signal of each module, irefj . The purpose of the multiplier circuit is to ensure that output filter inductor current of each inverter module tracks the reference irefj to attain the same phase. In the meanwhile, the IVS loop modifies the magnitude of the output filter inductor current (and adjusts the output active power of each module) to ensure IVS and OCS. Basically, the distributed control strategy still adopts the conception of compound balanced control to realize the power balance of the system. Similar with the centralized control mentioned before, the OCS effect is different for the DC–AC ISOP system in the distributed control system when taking the output

6.6 Distributed IVS/OCS Control Strategy

151

filter inductor current and the output current as feedback signal. And, the decoupling relationships between loops and parameters design is similar to that in the centralized control strategy, which are not repeated here.

6.7 Summary This chapter proposes the compound control scheme for the DC–AC ISOP systems to obtain IVS and OCS. In the proposed control scheme, the inner current loop of each inverter module can adopt the output filter inductor current feedback and the output current feedback as well. Among them, the former scheme can realize OCS when the output filter capacitors are matched, and the latter scheme can achieve OCS unconditionally. Taking the latter scheme as an example, the IVS loop and the system output voltage loop are analyzed and it is pointed out that they are decoupled and can be designed separately. A two-module DC–AC ISOP system prototype is fabricated and tested in the laboratory, and the experimental results verify the effectiveness of the proposed control strategy and the parameters design of the control loops. Furthermore, a distributed control scheme is proposed on the basis of the centralized control, and it disperses all the control loops into each inverter module and achieves modularity completely.

References 1. Chen, W., Zhuang, K., Ruan, X.: Input-series and output-parallel connected inverter system for high input voltage applications. IEEE Trans. Power Electron. 24(9), 2127–2137 (2009) 2. Fang, T., Ruan, X., Tse, C.K.: Control strategy to achieve input and output voltage sharing for input-series-output-series connected inverters systems. IEEE Trans. Power Electron. 25(6), 1585–1596 (2010) 3. Silva, C., Oliveira, D., Bascope R.: A dc-ac converter with high frequency isolation. In: Proceedings of IEEE International Symposium on Industrial Electronics, pp. 953–958 (2007) 4. Xue, Y., Chang, L., Kjaer, S.B., Bordonau, J., Shimizu, T.: Topologies of single-phase inverters for small distributed power generators: an overview. IEEE Trans. Power Electron. 19(5), 1305–1314 (2004) 5. Kjaer, S.B., Pedersen, J.K., Blaabjerg, F.: A review of single-phase grid-connected inverters for photovoltaic modules. IEEE Trans. Ind. Appl. 41(5), 1292–1306 (2005) 6. Ruan, X.: Soft-Switching PWM Full-Bridge Converters. Wiley, Singapore (2014) 7. Chen, W.: Research on series-parallel conversion systems consisting of multiple converter modules (in Chinese). PhD Dissertation. Nanjing University of Aeronautics and Astronautics, Nanjing, China (2009) 8. Zhuang, K., Ruan, X.: Control strategy to achieve input voltage sharing and output current sharing for input-series-output-parallel inverter. In: Proceedings of IEEE Power Electronics Specialists Conference, pp. 3655–3658 (2008) 9. Andersen, G.K., Klumpner, C., Kjaer, S.B., Blaabjerg, F.: A new green power inverter for fuel cells. In: Proceedings of IEEE Power Electronics Specialists Conference, pp. 727–733 (2002) 10. Liu, C., Lai, J.: Low frequency current ripple reduction technique with active control in a fuel cell power system with inverter load. IEEE Trans. Power Electron. 22(4), 1429–1436 (2007)

152

6 Compound Balanced Control Strategy for DC–AC …

11. Vlatkovic, V., Sabate, J.A., Ridley, R.B., Lee, F.C., Cho, B.H.: Small-signal analysis of the phase-shifted PWM converter. IEEE Trans. Power Electron. 17(1), 128–135 (1992) 12. Fang, T., Shen, L., He, W., Ruan, X.: Distributed control and redundant technique to achieve superior reliability for fully modular input-series-output-parallel inverter system. IEEE Trans. Power Electron. 32(1), 723–735 (2017)

Chapter 7

Compound Balanced Control Strategy for DC–AC Input-Series-Output-Series Systems

Abstract DC–AC input-series-output-series (ISOS) systems are very suitable for high input voltage and high output voltage power conversion applications. The implementation of the compound balanced control strategy is presented for the DC–AC ISOS system to achieve input voltage sharing (IVS) and output voltage sharing (OVS). The output filter inductor current or the output filter capacitor current adopted as the feedback signal for the inner current loops of each module, respectively, is analyzed and compared. Taking the output-filter-capacitor-current method as an example, the decoupling relationship between the control loops is analyzed and the loop design procedure is given. The simulation and experimental results are provided to verify the effectiveness of the proposed control method. Furthermore, this chapter presents a new solution dedicated to distributed configuration on the basis of centralized control, which can improve the reliability of the system. Keywords Input-series-output-parallel system Compound balanced control strategy · Input voltage sharing Output voltage sharing · Closed-loop designs As one of the input-series-connected DC–AC power conversion systems, the DC–AC input-series-output-series (ISOS) system is composed of multiple DC–AC inverter modules with low input voltage and low output voltage, which are connected in series both at the input side and output side, and thus it is suitable for high input voltage and high output voltage applications. The key issue of the DC–AC ISOS systems is to achieve input voltage sharing (IVS) and output voltage sharing (OVS) of the constituent modules simultaneously. Different from the DC–DC ISOS systems, the output voltage of the DC–AC ISOS system is an AC one. So, in order to achieve OVS, it is required to control both the magnitude and phase of output voltages of all the modules to be equal. As a consequence, both the active and reactive powers of the modules will be balanced.

© Springer Nature Singapore Pte Ltd. and Science Press, Beijing, China 2019 X. Ruan et al., Control of Series-Parallel Conversion Systems, CPSS Power Electronics Series, https://doi.org/10.1007/978-981-13-2760-5_7

153

154

7 Compound Balanced Control Strategy …

In Chap. 5, it is pointed out that the DC–AC ISOS systems needs to adopt the compound balanced control strategy, i.e., to realize IVS and ensure the magnitude or phase of the modular output voltages be equal. In this chapter, an implementation of the compound balanced control strategy is proposed. In the proposed control strategy, the output filter inductor current or the output filter capacitor current could be the feedback signal for the inner current loops of each module. The analysis shows that the output filter capacitor current feedback can realize IVS and OVS among the constituent modules no matter the output filter capacitors of modules are well matched or not. As an example, the compound balanced control strategy with output filter capacitor current feedback is analyzed, and it is shown that the IVS control loops and the system output voltage loop are decoupled each other. Then, the parameters design for all the control loops is presented. Simulation and experimental results are given to verify the validity of the proposed compound balanced control strategy. Finally, based on the compound balanced control strategy, a distributed control scheme is proposed to achieve full modularity of the DC–AC ISOS systems.

7.1 Implementation of the Compound Balanced Control Strategy Figure 7.1 gives the configuration of the DC–AC ISOS systems consisting of n inverter modules. The basic inverter module in the DC–AC ISOS systems adopts the two-stage configuration consisting of the first-stage DC–DC converter and the second-stage DC–AC inverter. Here, the phase-shifted full-bridge converter is employed as the first-stage, and the full-bridge inverter is adopted as the second stage. As described in Sect. 5.1.4, the compound balanced control strategy means ensuring the equaling the amplitude or phase of modular output voltages while controlling the input voltage sharing. According to this idea, Fig. 7.2 shows an implementation of the compound balanced control strategy for the DC–AC ISOS system [1], which includes three control loops, namely, the common system output voltage loop, the IVS loop, and the inner current loop of each inverter module (Here the module output filter capacitor current is taken as the feedback signal for each inner current loop). The common system output voltage loop guarantees the stability of the system output voltage, and its output signal iref acts as the common initial current reference for all the inner current loops. Here iref contains both the amplitude and the phase information. In the IVS loop, the system input voltage divided by n and the input voltage of each module are sensed with sensor gain of K vi and compared. The voltage error is sent into the IVS regulator Gvcd , whose output signal ivdj (j 1, 2,…, n) serves as the current reference amplitude correction signal of each module. By multiplying ivdj and iref , the current correction signal isvj (j 1, 2, …, n) is obtained with the same phase as that of iref . Then, isvj is superimposed on the initial reference iref to adjust the amplitude of the current reference of each module. Since the current inner loop adopts the capacitor current feedback of each module, the IVS loop only alters

7.1 Implementation of the Compound Balanced Control Strategy Module 1 Full-Bridge DC-DC Converter L dc1 DR1 DR3

iin1 Q1

icd1

icd2

S1 A1 +

Q4

_ B1 S2

DR2 DR4 + vin2 _

iLf1 vinv1

Vdc1 Cdc1 _

Q2

Lf1

S3

Module 2

S4

io1

iCf1 Cf1

io + vo1 _ +

ZLd v_o + vo2 _

…

Cd2

+

+ v _in1

Vin

iLdc1

Tr

* *

Cd1

Q3

Full-Bridge Inverter

…

Iin

155

icdn Cdn

iinn + vinn _

Module n

+ von _

Fig. 7.1 Configuration of the DC–AC ISOS system

the amplitude of the output filter capacitor current reference of each module, and the phase of the current reference is always synchronized. As a result, all the output filter capacitor currents have the same phase by tracking the synchronized reference. Since the filter capacitor current of each module leads the corresponding module output voltage by 90°, the module output voltages have the same phase, i.e., θ1 θ2 · · · θn

(7.1)

where θ 1 , θ 2, …, θ n are the output power factor angle of each inverter module. With the IVS loop, we have Vin1 Vin2 · · · Vinn

(7.2)

At steady state, the average current of the input dividing capacitor of each module is zero, so we have Iin1 Iin2 · · · Iinn

(7.3)

It is supposed that the conversion efficiency of each inverter module is 100%. So, by power conservation, we have

156

7 Compound Balanced Control Strategy … IVS Loop

vin1

Gvcd

ivd1

isv1

_

Kvi

+ + _

Gi

Ginv

_

+

1 iLf1 sLf +_

iCf1 1 vo1 sCf

Ki Inner Current Loop

IVS Loop + _

Kvi vin2 Kvi

Gvcd

ivd2

isv2

iref2 _

+

+ _

Gi

Ginv

vinv2

vo2

_ 1 sLf

+

iLf2 +_

iCf2 1 sCf

io

…

…

Ki

+ Kvi _

+

Inner Current Loop

IVS Loop

vinn

vinv1

…

Vin n

+ Kvi _

Inner Current Loop iref1

Gvcd

ivdn

isvn

Kvi

irefn _

+ + _

Gi

Ginv

vinvn +

_

_ + 1 sLf iLfn

1 ZLd + + vo

iCfn 1 von sCf

Ki

+ voref _

Gvo

iref

Kvo System Output Voltage Loop

Fig. 7.2 Compound balanced control strategy for DC–AC ISOS system

Pin j Vin j · Iin j Vo j · Io · cos θ j Po j ( j 1, 2, . . . , n)

(7.4)

where Pinj and Poj are the input power and output active power of each module, respectively. Substitution of (7.2) and (7.3) into (7.4) leads to Vo1 · cos θ1 Vo2 · cos θ2 · · · Von · cos θn

(7.5)

Substituting (7.1) into (7.5) yields Vo1 Vo2 · · · Von

(7.6)

According to (7.1) and (7.6), we have vo1 vo2 · · · von

(7.7)

As discussed above, the implementation of the compound balanced control strategy as shown in Fig. 7.2 can achieve IVS and OVS simultaneously, and it is referred to as the combined control of IVS and output phase synchronization. For the DC–AC ISOS systems, the feedback signal of the inner current loop of each module in Fig. 7.2 can also be the output filter inductor current. In this

7.1 Implementation of the Compound Balanced Control Strategy · · · ILf2 Vo2 ICf1 · ICf2 0 · Vo1

· f1 IL

157

· ICf1 · Vo

· Io

· · Vo1 Vo2

(a) With filter inductor current feedback

· ICf2

· Vo · Io

0 · · Vo1=Vo2

(b) With filter capacitor current feedback

Fig. 7.3 Steady-state phasor diagram of the output variables of DC–AC ISOS system with different current feedback when C f1 > C f2

case, the IVS loop can still ensure IVS as well as the output active power balance among the inverter modules [1]. However, when the output filter inductor current is selected as the feedback signal for the inner current loop, all of module filter inductor currents could have the same phase. For achieving OVS, all the module output filter capacitor should be well matched. Here, a two-module DC–AC ISOS system is illustrated as an example. For this system, we have iLfj iCfj + io (j = 1, 2), where, iLfj is the output filter inductor current, iCfj is the output filter capacitor current, and io is output current of the system. So, if the output filter capacitors are equal, the equaling of the output filter inductor currents (iLf1 iLf2 ) will ensure that the output filter capacitor currents are equal, i.e., iCf1 iCf2 , leading to vo1 vo2 . As a result, OVS can be achieved. If the output filter capacitors of two modules are not matched, for example, C f1 > C f2 , the IVS loop regulates so that each module’s filter inductor current shares the same phase while differing in amplitude, and the final steadystate phasor diagram is shown in Fig. 7.3a. It can be seen that the output voltage difference between the two modules, vo1 − vo2 , lags load current io by 90°, which means there is a difference of the reactive power between the two modules, and it presents the capacitive reactive power. On the contrary, if the output filter capacitor current is selected as the feedback signal, when there is a difference between the output filter capacitors of two modules (also assuming C f1 > C f2 ), the IVS loop could ensure active power balance (i.e., (7.5) is established), and the role of the output synchronization could guarantee the establishment of (7.1). Thus, OVS is achieved (i.e. (7.7) is established). So, if the output filter capacitors is not matched, the IVS loop will adjust the capacitor current to attain the difference amplitude but the same phase at steady state as shown in Fig. 7.3b. From the above analysis, with the output filtering inductor current feedback, OVS can be achieved only if the output filter capacitors of the module are matched; and with the output filter capacitor current feedback, OVS can be achieved no matter the output filter capacitors are matched or not.

158

7 Compound Balanced Control Strategy …

irefj +

-h

_ if

Vdc

vinv

h

_

iLf

1 sLf

+

+_ io

-Vdc

iCf

1 sCf

voj

Ki Fig. 7.4 Control block diagram of inverter with three-state hysteresis modulation

7.2 Decoupling the Control Loops As seen in Fig. 7.2, in the combined control of IVS and output phase synchronization, there are n + 1 variables including the system output voltage and n modular input voltages, which need to be controlled. These variables seem to interfere with each other and it is very difficult to analyze and design regulators of the system directly. It is, thus, necessary to explore the relationship of these variables to reduce the complexity of the system. In Fig. 7.2, Gvcd and Gvo are the transfer functions of the IVS regulator and the output voltage regulator, respectively. In this chapter, the three-state hysteresis modulation is employed in the inner current loop, which will be described in detail in Chap. 8. Figure 7.4 shows the control block diagram of each inverter module. With well-designed current tracking, the inner current loop, as shown in the dashed box in Fig. 7.4, can be reduced to a current follower with a gain equal to 1/K i [2, 3]. The equivalent control block diagram of Fig. 7.2 is reorganized in Fig. 7.5a, from which we have vo

n j1

⎛

vo j

n i ref − i sv j · j1

⎞ n 1 1 n ⎝ni ref − i sv j ⎠ · ⎝i ref − i sv j ⎠ · s K i Cf n j1 s K i Cf j1 n

⎞

1 s K i Cf ⎛

(7.8)

According to (7.8), Fig. 7.5a can be further simplified as shown in Fig. 7.5b, and we have i sv1 + i sv2 + · · · + i svn

n

(Vin n − vin j ) · K vi · G vcd · i ref j1

⎛

⎝Vin −

n

⎞ vin j ⎠ · K vi · G vcd · i ref

j1

0

(7.9)

7.2 Decoupling the Control Loops

+ Kvi _

vin1

Gvcd

isv1

_ +

Kvi IVS Loop

+ Kvi _

Gvcd

ivd2

isv2

_ +

Kvi

iref1

1 Ki

1 sCf1

vo1

iref2

1 Ki

1 sCf2

vo2 +

1 sCfn

von

…

IVS Loop

…

vin2

ivd1

+ Kvi _

vinn

Gvcd

Kvi

ivdn

isvn

_ +

IVS Loop

+ _

voref

…

Vin n

159

Gvo

irefn

1 Ki

+ vo +

Inner Current Loop

iref

Kvo System Output Voltage Loop

(a) Equivalent control block diagram of Figure 7.2 + Kvi _

vin1

Gvcd

isv1

Kvi IVS Loop

+ Kvi _ Kvi

Gvcd

ivd2

isv2

+

+

IVS Loop

+

1 n

…

vin2

ivd1

…

Vin n

+ Kvi _

vinn

Kvi

voref

+ _

Gvcd

ivdn

isvn

IVS Loop Gvo

iref

_ +

n sKiCf

vo

Kvo System Output Voltage Loop (b) Further equivalent contro l block diagram Fig. 7.5 Decoupling analysis of DC–AC ISOS system with compound balanced control strategy

160

+ Kvi _

vin1

Gvcd

isv1

Kvi IVS Loop

+ Kvi _ Kvi

Gvcd

ivd2

isv2

+

+

IVS Loop

=0 +

…

vin2

ivd1

…

Vin n

7 Compound Balanced Control Strategy …

+ Kvi _

vinn

Kvi

Gvcd

ivdn

isvn

IVS Loop

+ voref _

Gvo

iref

n sKiCf

vo

Kvo System Output Voltage Loop (c) Decoupling block diagram Fig. 7.5 (continued)

According to (7.9), Fig. 7.5b can be further simplified as shown in Fig. 7.5c. Equation (7.9) means that the IVS loops have no effect on the system output voltage loop. The output of the system output voltage regulator iref serves as the initial current reference for the inner current loop of each module. When iref changes, the current reference of each module, irefj , will change simultaneously so that the output power and input power of each module vary accordingly. For example, when IVS and OVS are both achieved initially, the abruptly increasing load would result in a decrease of the output voltage. In order to keep the output voltage unchanged, iref will increase, and the output filter capacitor current and then the output voltage of each module will rise. In the meantime, OVS and the balanced output power of all modules are maintained. According to the principle of power conservation, the input current of each module and hence the total input current of the system will all increase. Since the input current is still being shared and the average current of each input dividing capacitor is zero, IVS is maintained. It can be seen that the operation of the IVS loops is not affected by the system output voltage loop. From the previous analysis,

7.2 Decoupling the Control Loops

161

Fig. 7.6 Block diagram of the output voltage loop

+ voref _

Gvo vof

n sKiCf

vo

Kvo

it is clear that the IVS loop and the system output voltage loop work independently and they are decoupled from each other. Thus, the two control loops can be designed separately.

7.3 Design of the Control Loops for the DC–AC ISOS Systems With no loss of generality, a two-module DC–AC ISOS system is taken as an example in the following to illustrate the design of the IVS loop and output voltage loop. The main parameters of the system are as follows: the number of modules n is 2, the input voltage is 540 V (±10%), the output voltage is 230 V AC/400 Hz, the output capacity is 2 kVA, the input dividing capacitor C d is 1000 µF, the system output voltage sensor gain K vo is 0.031, the input voltage sensor gain K vi is 0.03. The main parameters of each module are as follows: the input voltage is 270 V (±10%), the output voltage is 115 V AC/400 Hz, the output capacity is 1 kVA, the output filter inductor L f is 600 µH, the output filter capacitor C f is 30 µF, the inner current loop sensor gain K i is 0.2. In addition, the parameters of the first-stage fullbridge converter of each module are given in Sect. 6.3.2 and they are not repeated here.

7.3.1 Design of the Output Voltage Loop From Fig. 7.5c, the control block diagram of the system output voltage loop is extracted as shown in Fig. 7.6. From Fig. 7.6, the loop gain of the system output voltage loop can be derived as Tvo_c (s)

n K vo G vo (s) s K i Cf

(7.10)

Substituting the above related parameters into (7.10) and letting Gvo (s) 1, the uncompensated loop gain can be obtained which is expressed as Tvo_u (s)

n K vo 1.03 × 104 s K i Cf s

(7.11)

162

7 Compound Balanced Control Strategy …

Fig. 7.7 PI regulator

For the sake of improving the tracking precision of the output voltage with the reference, a PI regulator is adopted as the output voltage regulator (see Fig. 7.7a), with the transfer function expressed as R2 · G vo (s) R1

s+

1 s + ωz R2 C G vo∞ s s

(7.12)

1 . 2π R2 C The three-state hysteresis current controlled inverter has a broad distribution of harmonics in the output voltage and its switching frequency is variable. Denoting its sampling frequency by f k , the three-state hysteresis current controlled inverter can be equivalent to the PWM inverter whose switching frequency is a quarter of f k [4, 5]. In this chapter, the sampling frequency of the inverter f k is 100 kHz. Then, the switching frequency of the equivalent PWM inverter f s is 25 kHz. One-fifth of f s , i.e., f c 5 kHz [6], is set as the cutoff frequency of the compensated output voltage loop. Figure 7.8 gives the Bode diagram of the uncompensated output voltage loop gain, from which we can find the magnitude of the uncompensated loop gain is −9.66 dB at 5 kHz. Thus, the PI compensator should have a gain of 9.66 dB at 5 kHz, i.e., where the turning frequency of PI regulator is f z

G vo∞

R2 9.66 10 20 3.04 R1

(7.13)

Here, R1 is chosen as 10 k, then we have R2 30.4 k. Finally, R2 30 k is chosen. It can be found in Fig. 7.7b that the PI regulator will not bring large phase lag when 10f z ≤ f c . So, we have C

1 5 ≥ 10.61 (nF) 2π f z R2 π f c R2

(7.14)

We choose C = 10 nF. Then, the transfer function of the PI regulator can be derived as

7.3 Design of the Control Loops for the DC–AC ISOS Systems

163

Fig. 7.8 Bode diagram of the uncompensated and compensated loop gains of the system output voltage closed loop

G vo (s) 3 ·

s + 3.33 × 103 s

(7.15)

The Bode diagram of the compensated output voltage loop gain is sketched in Fig. 7.8. It can be found that the compensated loop gain has a cutoff frequency of 5 kHz with a phase margin of 83.9°.

7.3.2 Design of the Control Loop of the First-Stage Converter and IVS Loop For the first-stage phase-shifted full-bridge converter, the intermediate DC bus voltage closed-loop control is adopted, and its design is similar to that in Sect. 6.3.2. Meanwhile, the design of the IVS loop can also refer to Sect. 6.3.3. Similarly, we can conclude that, if the IVS loop compensator gain satisfies (6.20), each module of the DC–AC ISOS systems could show positive resistance characteristic at the input side and then the system can be stable.

7.4 Simulation and Experimental Verification In order to verify the effectiveness of the compound balanced control strategy, a two-module DC–AC ISOS system is taken as an example in the simulation and experiments. The main parameters of this system have been given in Sect. 7.2.

164

7 Compound Balanced Control Strategy …

(V)

280

vin1

vin2

270

(V)

260 200

(V)

(A)

-20

vo1-vo2

0 -200 400

20 0

vo1 vo2

vo io

0 -400 0.110

0.112

0.114

t (ms)

0.116

0.118

0.120

0.118

0.120

(a) Full resistive load (V)

280

vin1

vin2

270

(V)

260 200

(V)

(A)

-20

vo1-vo2

0 -200 400

20 0

vo1 vo2

vo io

0 -400 0.110

0.112

0.114

t (ms)

0.116

(b) Full inductive load Fig. 7.9 Simulation waveforms at steady state

7.4.1 Simulation Results Figures 7.9a, b gives the steady-state simulation waveforms of the system with the full resistive and full inductive load (cos θ = 0.75), respectively. As shown, the system achieves IVS and OVS under various load conditions. Figure 7.10 shows the system dynamic simulation waveforms. Figure 7.10a gives the system simulation waveforms at full resistive load (2 kW) with the input voltage stepping between 486 V (90% of the rated input voltage) and 594 V (110% of the rated input voltage). Figure 7.10b gives the system simulation waveforms at rated input voltage (540 V) conditions with the load current stepping between 1/3 full load and full load. It can be seen from the waveforms that IVS and OVS are well achieved during the transients.

7.4 Simulation and Experimental Verification

165

(V)

640 560

vin

480

(V)

320 280

vin1

vin2

240

(V)

200

vo1 vo2

0 -200

58

60

66

64

62

t (ms) (a) Step-input-voltage condition

20

(A)

vo1-vo2

68

70

72

io

0

(V)

-20 280 270

vin1

(A)

260 200

vin2 vo1 vo2

vo1-vo2

0 -200

82

86

84

88

90

92

t (ms) (b) Under step-load Fig. 7.10 Dynamic simulation waveforms

Figure 7.11 shows the steady-state simulation waveforms of the system at full resistive load with an unmatched output filter capacitor (C f1 30 µF, C f2 25 µF). It can be found in Fig. 7.11a that when the output filter capacitor current feedback is adopted, the output filter capacitor currents of two modules have the same phase but the different amplitude, and the output voltage of each module are equal. This implies that the unmatched output filter capacitors do not influence OVS when employing the output filter capacitor current feedback. From Fig. 7.11b, it can be found that when the output filter inductor current feedback is employed, the output filter inductor currents of two modules have the same phase but different amplitude, and the output

166

7 Compound Balanced Control Strategy …

Fig. 7.11 Steady-state simulation waveforms of the two kinds of current feedback with unmatched output filter capacitor (at full resistive load)

voltages of two modules have phase difference and vo1 lags behind vo2 . The difference between the two voltages vo1 − vo2 lags load current io by 90°, which indicates that there is the reactive power difference at the outputs of two modules. The simulation results in Fig. 7.11 are in agreement with the theoretical analysis in Sect. 7.1.

7.4 Simulation and Experimental Verification

167

7.4.2 Experimental Results In order to further validate the effectiveness of the proposed compound balanced control strategy, a two-module DC–AC ISOS system prototype is built in the lab and the experiment has been carried out. The parameters are the same as that of the simulation. Figure 7.12 shows the experimental waveforms at steady state with full resistive and full inductive load (cos θ = 0.75). As seen, the system achieves IVS and OVS under various load conditions. Figure 7.13a, b shows the dynamic experimental waveforms when the input voltage of the system is step changed between 486 and 594 V and when the load is step changed between 1/3 full load and full load. As seen, IVS and OCS are well achieved during the transient. Obviously, the compound balanced control strategy can effectively realize IVS and OVS both at steady state and during the transient. Figure 7.14a, b shows the steady-state experimental waveforms of the two-module DC–AC ISOS system (at full resistive load) with output filter capacitor current feedback and output filter inductor current feedback, respectively. Here, the output filter capacitors are intentionally set at different values, i.e., C f1 30 µF, C f2 25 µF. As shown in Fig. 7.14a, despite the difference between the output filter capacitors, the two modules can still effectively achieve OVS when using the output filter capacitor current feedback. While from Fig. 7.14b we can find that when adopting the output filter inductor current feedback, the difference between the output filter capacitors leads to a phase difference between the output voltages of the two modules, and vo1 lags vo2. The difference between the two voltages vo1 − vo2 lags the load current io by 90°, indicating the existence of output reactive power difference between the two modules. Obviously, even if the output filter capacitors are not matched, the filter capacitor current feedback control can still achieve OVS among the constituent modules.

7.5 Distributed Control Strategy for DC–AC ISOS Systems The compound balanced control strategy proposed in Sect. 7.1 includes three control loops, namely, the system output voltage loop, the IVS loop and inner current loop of each module. The output voltage loop is common used for all modules, and this scheme is a centralized control without realizing full modularity of the system. Therefore, in this section, a distributed control method [7] is given and its control block diagram is shown in Fig. 7.15, where each module has three independent control loops: IVS loop, output voltage loop, and inner current loop. And the communication among the modules is achieved through the buses, in which the input voltage sharing bus signal is obtained by the average circuit and its value is V in /n.

168

7 Compound Balanced Control Strategy …

M1 vin1 : [100V/div] M2 vin2 : [100V/div]

C1 vo1 : [200V/div]

C1/M1 C2/M2

F1

F1 vo1-vo2 : [200V/div] C3: vo[400V/div]

C2 vo2 : [200V/div] M4: io[20A/div]

C3/M4

Time: [2ms/div]

(a) At full resistive load M1 vin1 : [100V/div] M2 vin2 : [100V/div]

C1 vo1 : [200V/div]

C1/M1 C2/M2

F1

F1 vo1-vo2 : [200V/div] C3: vo[400V/div]

C2 vo2 : [200V/div] M4: io[20A/div]

C3/M4

Time: [2ms/div]

(b) At full inductive load Fig. 7.12 Experimental waveforms under steady state

Additionally, the output voltage references of all the modules can be synchronized with DSP to form the output voltage reference bus denoted as voref . Moreover, the output voltage feedback signals of all the modules, vofj , are averaged to generate the output voltage average bus, denoted as vof_ave . The error between voref and vof_ave is sent to each output voltage regulator. It should be noted that from Fig. 7.15, the average feedback signal vof_ave of each module is obtained as vof_ave (K vo · vo1 + K vo · vo2 + · · · + K vo · von ) K vo · vo /n

(7.16)

7.5 Distributed Control Strategy for DC–AC ISOS Systems

169

M1 vin : [100V/div] M2 vin1 : [100V/div] M1

M3 vin2 : [100V/div] C2 vo1 : [200V/div]

C2/M2 M3 C3

F1

F1 vo1-vo2 : [200V/div]

C3 vo2 : [200V/div]

Time: [10ms/div]

(a) Under step-input-voltage condition M1 io : [20A/div]

M1

M2 vin1 : [100V/div] M3 vin2 : [100V/div]

C2 vo1 : [200V/div]

C2 M2 C3/M3

F1

F1 vo1-vo2 : [200V/div]

C3 vo2 : [200V/div]

Time: [5ms/div]

(b) Under step-load Fig. 7.13 Dynamic experimental waveforms.

Equation (7.16) shows that it is the average of all the module’s output voltages that tracks the synchronous voltage reference voref in each independent output voltage loop. So, the control effect here is equivalent to that of the previous centralized control strategy in which the output voltage of the system is adjusted to realize its sinusoidal output. However, different from the centralized control strategy, the proposed distributed control strategy has a set of independent output voltage loop for each inverter module. In addition, the common current reference bus signal iave of each module is also obtained by means of the average circuit, and the inner current loop of each module is independent of each other. It can be seen that in the distributed

170

7 Compound Balanced Control Strategy … M1 vin1 : [100V/div] M2 vin2 : [100V/div]

C1 vo1 : [200V/div]

C1/M1 C2/M2

F1

F1 vo1-vo2 : [200V/div] C3: vo[400V/div]

C2 vo2 : [200V/div] M4: io[20A/div]

C3/M4

Time: [2ms/div]

(a) With the output filter capacitor current feedback M1 vin1 : [100V/div] M2 vin2 : [100V/div]

C1 vo1 : [200V/div]

C1/M1 C2/M2

F1

F1 vo1-vo2 : [200V/div] C3: vo[400V/div]

C2 vo2 : [200V/div] M4: io[20A/div]

C3/M4

Time: [2ms/div]

(b) With the output filter inductor current feedback Fig. 7.14 Steady-state experimental waveforms of the two kinds of current feedback with unmatched output filter capacitor (at full resistive load)

control method, the control circuits are scattered in individual modules so that each module can work independently and the true modularity is achieved for the system. The control mechanism of the input and output voltage sharing of the proposed strategy can be further illustrated as follows. Actually, the IVS loop alters the real power at the output terminal to achieve IVS. Since the output signals of the IVS regulator ivdj are DC components while the variables at the output terminal are AC quantities. So the multipliers are introduced here to multiply the DC deviation ivdj with the AC variable iave . In this way, the phase of the signal that is used to tune the current reference by the IVS loop is kept the same with that of the original common current reference iave so as to ensure the synchronization of the current reference, irefj , with the IVS loop only slightly adjusting the magnitude of it. From the compound balanced control conception, we know that both IVS and OVS could be achieved under IVS control combined with the synchronization of the power factor angle of each module simultaneously. So here the output filter capacitor feedback

ig1

Gvcd

Gvcd

+

ign voref + Gvo _ vof_ave Output Voltage Loop

IVS Loop + Kvi _ vinn Kvi

…

ig2 voref + Gvo _ vof_ave Output Voltage Loop

IVS Loop + Kvi _ vin2 Kvi

…

+

+

ivdn

1 n

ivd2

ivd1 _ iLf1 1 sLf +_

iCf2

iCf1

+ +

1 sCf vo2

1 vo1 sCf

+

Kvo

Ki

Inner Current Loop _ irefn iLfn _ iCfn vinvn _ + + 1 1 von G G i inv _ sLf + sCf +

isvn

Common Current Reference Bus

Kvo

Ki

Inner Current Loop _ iref2 iLf2 vinv2 _ + + 1 G G i inv _ sLf +_ +

isv2

Kvo

Ki

isv1 Inner Current Loop _ iref1 vinv1 + + G G i inv _ +

iave

Fig. 7.15 Block diagram of distributed control strategy for DC–AC ISOS system

Output IVS Output Voltage 1 + + Bus Voltage Average n + Reference Bus Bus

Gvo

Gvcd

Output Voltage Loop

voref + _ vof_ave

IVS Loop + Kvi _ vin1 Kvi

…

vof_ave

…

…

voref Vin n

io

vo

1 ZLd

7.5 Distributed Control Strategy for DC–AC ISOS Systems 171

172

7 Compound Balanced Control Strategy …

is adopted for the inner current loops. Since the output voltage lags the output filter capacitor current in phase by 90° and all the output filter capacitor currents track the synchronous current references, phase synchronization of the output voltages of all the modules is fulfilled. Similar with the aforementioned centralized control, the output filter inductor current feedback can also be used in the distributed control of DC–AC ISOS systems, which is not repeated here.

7.6 Summary Aiming for achieving IVS and OVS for the DC–AC ISOS systems, this chapter presents an implementation of the compound balanced control strategy, i.e., the combined control strategy of IVS and output phase synchronization. For the inner current loop, when adopting the output filter inductor current feedback, OVS can be realized only if the output filter capacitors are well matched; while when the output filter capacitor current feedback is adopted, OVS can be achieved no matter the output filter capacitors are matched or not. The compound balanced control strategy with the output filter capacitor current feedback is analyzed, and it is pointed out that the IVS loop and the system output voltage loop are decoupled each other, and these control loops can be designed independantly. A prototype of two-module DC–AC ISOS systems has been fabricated in the lab. The simulation and experimental results show that the proposed combined control of IVS and output phase synchronization is effective. Moreover, based on the centralized control, a distributed control scheme is proposed, in which all the control loops are dispersed into each module, and hence the modularity of the system is achieved.

References 1. Fang, T., Ruan, X., Tse, C.K.: Control strategy to achieve input and output voltage sharing for input-series-output-series connected inverters systems. IEEE Trans. Power Electron. 25(6), 1585–1596 (2010) 2. Xing, Y., Huang, L., Yan, Y.: Redundant parallel control for current regulated inverters with instantaneous current sharing. In: Proceedings of the IEEE Power Electronics Specialists Conference, pp. 1438–1442 (2003) 3. Chen, L., Xiao, L., Gong, C., Yan, Y.: Circulating current’s characteristics analysis and the control strategy of parallel system based on double close-loop controlled VSI. In: Proceedings of the IEEE Power Electronics Specialists Conference, pp. 4791–4797 (2004) 4. Venkataramanan, G., Divan, D.M., Jahns, T.M.: Discrete pulse modulation strategies for highfrequency inverter systems. IEEE Trans. Power Electron. 8(3), 279–287 (1993)

References

173

5. Venkataramanan, G., Divan, D.M.: Pulse width modulation with resonant dc link converters. IEEE Trans. Ind. Electron. 9(1), 113–120 (1993) 6. Li, A., Zhang, C.: Modern Inverter Technology and its Application. Science Press, Beijing, China (2000) (in Chinese) 7. Zhu, H., Fang, T., Ruan, X.: Distributed voltage sharing control strategy for input-series-outputseries inverters system. In: Proceedings of the Annual Conference of the IEEE Industrial Electronics Society, pp. 647–652 (2013)

Chapter 8

An Improved Average Current Control Strategy for DC–AC Input-Parallel-Output-Parallel Systems

Abstract For the DC–AC input-parallel-output-parallel (IPOP) systems, the output current sharing control strategy is popular and effective to achieve output current sharing (OCS) among the constituent modules, and the average current control is usually adopted. The load current feed-forward scheme is incorporated into the average current control and the load current feed-forward node is located before the current-limiting unit, for not only improving the DC–AC IPOP system output voltage characteristics but also limiting the load current. The prototype of two-module DC–AC IPOP systems is built in the lab, and the steady-state and dynamic experiments are provided to verify the superiority of the proposed improved strategy. Keywords Input-parallel-output-parallel systems · Average current control Load current feed-forward · Output characteristics Three-state hysteresis modulation The DC–AC input-parallel-output-parallel (IPOP) systems, commonly known as the parallel inverter system, features ease of system capacity expansion and achievement of N + 1 redundancy of system, improving the system reliability. It has been widely used in uninterruptible power supply (UPS), aeronautical static inverters, etc. It has been pointed out in Chap. 5 that the output current sharing control strategy or the compound input current sharing control strategy could be applied to DC–AC IPOP system. Comparatively, the output current sharing control is quite simpler than the compound input current sharing control strategy, and it is usually employed. In Sect. 1.4.1, various implementations of the output current sharing control strategy have been presented [1–13]. Centralized control and master–slave control are simple, but they have poor redundancy owing to relying on the central control unit or the master module to operate. Wireless droop control is not susceptible to interference and is easy to implement the redundancy due to no connection among the constituent modules, but the output characteristics are poor. As one of the distributed control schemes, the average current control has the merits of simple control circuit, prompt current sharing regulating speed, and ease of implementation of hot-swap and redundant control. However, the DC–AC IPOP system with the average current control also has the defect of poor output characteristics. For tackling this issue, the chapter proposes © Springer Nature Singapore Pte Ltd. and Science Press, Beijing, China 2019 X. Ruan et al., Control of Series-Parallel Conversion Systems, CPSS Power Electronics Series, https://doi.org/10.1007/978-981-13-2760-5_8

175

176

8 An Improved Average Current Control …

a control strategy of the inductor current feedback combined with the load current feed forward, which can greatly improve the output characteristics with preservation of the original current-limiting function.

8.1 Current Three-State Hysteresis Modulation In this chapter, the basic module in the DC–AC IPOP systems still adopts the two-stage configuration consisting of full-bridge DC–DC converter and full-bridge inverter, as shown in Fig. 8.1. Same as that in Chaps. 6 and 7, the first-stage fullbridge DC–DC converter controls its output voltage V dc . The second-stage full-bridge inverter adopts double closed-loop control, composed of outer voltage control and inner current loop. The inner current loop adopts the three-state hysteresis modulation and the control variable is the filter inductor current iLf . The operating waveforms of the three-state hysteresis modulation are given in Fig. 8.2, where iref is the current reference, if is the sensed filter inductor current, 2h is the current hysteresis, and Clk is the sampling clock. If if < iref − h at the sampling instant, power switches S 1 and S 4 are turned on, and the output voltage of inverter bridge, vinv , is equal to +V dc , which is defined as +1 state. Because V dc > vo , iLf is forced to increase. If if > iref + h at the sampling instant, S 2 and S 3 are turned on, and thus vinv −V dc , which is defined as −1 state. Because −V dc < vo , iLf declines. If if is within the hysteresis width at the sampling instant, i.e., iref − h < if < iref + h, S 2 and S 4 are turn on, or S 1 and S 3 are turned on, and thus vinv 0, which is defined as zero state. At zero state, iLf is freewheeling, and whether it increases or decreases depends on the polarity of the output voltage vo , namely, if vo is positive, iLf declines; and if vo is negative, iLf rises. The rates of the rise and decline of iLf in zero state are smaller than that of +1 and −1 state, respectively. Since the inverter could be

Iin

Full-Bridge DC-DC Converter DR1 DR3

Full-Bridge Inverter

Ldc iLdc

Q1

Q3

Tr +

S1 A +

* *

Q2

_ B S2

Q4 DR2 DR4

Fig. 8.1 Main circuit topology of basic inverter module

io

iLf vinv

Vdc Cdc _

Vin

Lf

S3

S4

Cf

+ ZLd vo _

8.1 Current Three-State Hysteresis Modulation

177

operated at +1, 0, and −1 states, the above current control is called the three-state hysteresis modulation. In summary, at the sampling instant, the operating state S is determined as ⎧ ⎪ i f < i ref − h ⎨ +1, (8.1) S 0, i ref − h < i f < i ref + h ⎪ ⎩ −1, i f > i ref + h As shown in Fig. 8.2, it is noteworthy that the operating state does not change immediately once the inductor current exceeds the hysteresis band, and it does only when the sampling clock comes. In addition, when the output voltage is in positive (negative) half cycle, the inverter switches between +1 (or −1) and 0 states. This implies that the inverter is operated with unipolar modulation. Compared with the bipolar modulation where the state of the inverter is modulated between +1 state and −1 state [14], the three-state hysteresis modulation has the following advantages: (1) The ripple of filter inductor current is relatively small. (2) Due to the presence of freewheeling mode, the feedback energy to the intermediate bus is relatively small and thus has little influence on the DC bus voltage. Figure 8.3a shows the double closed-loop control block diagram of the secondstage full-bridge inverter, where the outer voltage loop Gvo adopts PI regulator, and the inner current loop adopts the three-state hysteresis modulation. voref is the output voltage reference, vo is the output voltage, K vo is the sensor gain of the output voltage, and C f is the output filter capacitor, and Z Ld is the load impedance. Generally,

Fig. 8.2 Operating waveforms of three-state hysteresis modulation

178

8 An Improved Average Current Control …

Fig. 8.3 Double closed-loop control block diagram

the switching frequency is much higher than the output frequency, the inner loop (referring to the dashed frame in Fig. 8.3a) can be simplified as the current follower whose magnification is 1/K iL [15]. Thus, Fig. 8.3a can be reduced to Fig. 8.3b when the load current feed forward is not included.

8.2 Improvement of the Output Characteristics of Inverter As mentioned above, a PI regulator is chosen as the outer voltage regulator for the inverter module. Although the output voltage static error could be very small with properly designed PI regulator, it cannot be fully eliminated. In order to improve the output characteristics, the load current feed forward [16] is introduced as shown with the dashed lines in Fig. 8.3b, where K io is load current sensor gain. The load current feed-forward could be located at point ➀ or ➁, and the function of the current-limiting unit is different, which will be analyzed later. Obviously, under normal operating condition when the current-limiting unit is not activated, the output characteristics of the inverter are the same. With the current-limiting unit omitted, by replacing the feed-forward signal io with vo , Fig. 8.3b can be equivalent to Fig. 8.4a. By simplifying the transfer function from iLf to vo in the dashed box, Fig. 8.4a can be further reduced to Fig. 8.4b. As mentioned above, the voltage regulator adopted a PI regulator is expressed as 1 . Hence, according to Fig. 8.4b, the transfer function from the G vo (s) K PI + sτPI output voltage reference to the output voltage can be derived as

8.2 Improvement of the Output Characteristics of Inverter

io

Kio voref + _

Gvo

+ + iref

1/KiL

179

1 ZLd

iLf +

io _

ZLd

vo

sCf Kvo

(a) Equivalent control block diagram with current-limiting unit omitted io

Kio voref + _

Gvo

+ + iref

1/KiL

iLf

1 ZLd ZLd 1+sCfZLd

vo

Kvo

(b) Further equivalent control block diagram Fig. 8.4 Transfer function derivation of the inverter module

Vo (s) Voref (s)

s 2 K iL Cf + s

K PI s +

1 τPI

K vo K iL − K io + K vo K PI + Z Ld τPI

(8.2)

where the load Z Ld can be expressed in the following form: Z Ld |Z Ld |(cos ϕ + j sin ϕ)

(8.3)

where |Z Ld | is the magnitude of the load, and ϕ is the power factor angle of the load. If ϕ = 0, the load is purely resistive; if ϕ > 0, the load is inductive; and If ϕ < 0, the load is capacitive. Substituting (8.3) into (8.2), we have 2 ω2 + 1 K PI 2 τPI Vo ( jω) 2 2 Voref ( jω) K vo ω(K iL − K io ) sin ϕ (K iL − K io ) cos ϕ + K vo K PI ω2 − Cf K iL ω2 + + |Z Ld | |Z Ld | τPI

(8.4) According to (8.4), when the load current feed forward is not introduced, i.e., K io 0, with the parameters of the inverters given in Sect. 8.4, the curves of the output voltage versus the load current with different ϕ can be depicted as shown in Fig. 8.5. From Fig. 8.5, it can be found that, (1) When the load is resistive or inductive (ϕ ≥ 0°), the output voltage declines with the increase of the load current, and the larger the ϕ is the more the output voltage declines; and (2) When the load is capacitive (ϕ < 0°), the output voltage rises with the increase of the load current, and the larger

180

8 An Improved Average Current Control …

Fig. 8.5 Output voltage characteristics with and without the load current feed forward

the ϕ is the more the output voltage rises. In conclusion, the output characteristics are relatively poor when the load current feed-forward is not incorporated. When the load current feed forward is introduced and K io is set to be equal to K iL , it can be found from (8.4) V o is independent of the load as shown in Fig. 8.5, which is the desired output characteristic. In conclusion, when introducing the load current feed-forward and letting K io K iL , the output voltage of inverter is independent of the load and the optimal output characteristics is achieved. At this time, the feedback signal is the capacitor current in the current loop [16]. When the load current feed-forward node is located at point ➀, the current-limiting unit (as shown in Fig. 8.3b) actually limits the capacitor current, and the load current cannot be limited since the capacitor current does not contain the information of the load current. When the load current feed-forward node is located at point ➁, the current-limiting unit limits the inductor current and thus the load current. So, it is desirable to locate the load current feed-forward node at point ➁.

8.3 Configuration and Control Scheme for DC–AC IPOP Systems 8.3.1 Configuration of the DC–AC IPOP Systems As illustrated above, the DC–AC IPOP systems generally adopts the OCS control strategy, which could be implemented with centralized, master–slave, or distributed control. The distributed control includes average current control and wireless droop control. In this chapter, the average current distributed control is adopted [7], and the

8.3 Configuration and Control Scheme for DC–AC IPOP Systems

181

system configuration is shown in Fig. 8.6a. As seen, there are two signals among the control circuits of each module. One is the synchronizing signal syn, which ensures the sinusoidal voltage references of modules synchronous, and the synchronizing circuit is shown in Fig. 8.6b. The other is the averaging current reference iave , which is obtained by averaging all the outputs of modular voltage regulators as shown in Fig. 8.6c. iave is regarded as the common current reference of all the current regulators achieving the current sharing among the modules. K 1 is the output relay and K 2 is the current reference access-control relay. Once a module fails, its K 1 and K 2 will be turned off, and the driving signals for the switches will be disabled. The other modules will supply the load.

8.3.1.1

Synchronous Circuit

To ensure stable operation of the DC–AC IPOP systems, the amplitude, frequency and phase of the output voltage of each inverter module are all required to be consistent. The voltage reference of each module is generated by the high-precision D/A chip (MAX507) of the respective module, the frequencies of the output voltages of modules can be the same. Figure 8.6b gives the phase synchronizing circuit for the voltage reference [17], in which each voltage reference tracks the phase information of the synchronizing bus, syn. In each inverter module, a rising edge is emitted via signal port PB1 at the positive zero-crossing point of the voltage reference, and it induces a rising edge on the synchronizing bus. For the synchronizing bus, the first rising edge signal is regarded as the valid signal, and it is sent to the capture port CAP1 as a common synchronizing signal to fulfill the transmission and judgment of synchronizing signal. CAP1 captures the trailing edge and generates the synchronous interrupt request, and the DSP enters capture interrupt service routine. Here, the phase of the corresponding voltage reference starts from the positive zero-crossing of sinusoidal wave, achieving the zero-crossing phase-lock. From Fig. 8.6b, it can be found that as long as port PB1 of DSP of any inverter module is high, synchronous bus syn is high, which is tracked by the other inverter modules to achieve synchronization. With the synchronizing circuit, the phase difference of the output voltage in each cycle can be corrected, and the safety and reliable operation of the DC–AC IPOP systems will be ensured.

8.3.1.2

Instantaneous Average Circuit

In addition to ensuring synchronization of the output voltage phase, the output current sharing control is required for achieving power sharing among the inverter modules for DC–AC IPOP systems. So, it is better to produce the averaging instantaneous current reference for the modules, and the averaging circuit is shown in Fig. 8.6c, where irefj is the inductor current reference signal of jth inverter module (j = 1, 2, …, n). The average value of the current reference signals can be obtained as

182

8 An Improved Average Current Control … Vin

vo

syn iave

K1 Inverter 1

iref1 K2 DSP 1

syn

K1 Inverter 2

iref2 K2 DSP 2

syn

K1 Inverter n

irefn K2 DSP n

syn

(a) Block diagram of system structure syn +15V

+15V

INV1

+15V

INV2

INVn

PB11

PB12

PB1n

CAP11

CAP12

CAP1n

(b) Synchronizing circuit _ iref1 iave

R

+

_ iave

iref2

R

+

_ iave

irefn

R

(c) Current reference instantaneous averaging circuit Fig. 8.6 Block diagram of IPOP inverter

+

iave

8.3 Configuration and Control Scheme for DC–AC IPOP Systems

i ave (i ref1 + i ref2 + · · · + i refn ) n

183

(8.5)

And iave serves as the common current reference for each module. From (8.5) and Fig. 8.6c, we can also see that this averaging circuit, which is equally distributed in each module, is independent on the number of parallel inverter modules. Thus, the distributed control method can achieve redundant control of the DC–AC IPOP systems.

8.3.2 Control Scheme As indicated in [7], the outputs of voltage regulators of all the inverter modules are averaged, and the average signal serves as the current reference for the inductor current of all inverter modules and it is independent on a particular module, allowing hot-swap of any module. The block diagram of the average current distributed control for DC–AC IPOP systems is shown in Fig. 8.7, where each inner current loop of threestate hysteresis modulation has been simplified as a current follower, and the load current feed forward is incorporated to improve the output characteristics. According to the analysis of Sect. 8.2, when the load current feed forward is located before the current-limiting unit K ioj K iLj (j = 1, 2, …, n), the output voltage will be independent of the load, while the current-limiting function is preserved.

8.4 Experimental Results To verify the validity of the theoretical analysis, a 2-kVA two-module DC–AC IPOP system prototype is fabricated in the laboratory. The specifications of the inverter module are as follows: Rated output power: Po_mod 1 kVA, input voltage V in 270 VDC, output voltage V o 115 VAC/400 Hz, output filter inductor L f 600 µH, sensor gain K ioj K iLj 0.2. Figure 8.8 gives the measured output voltage versus the output current of the system with and without the load current feed forward at the resistive load. As seen, the output characteristics of the system have been greatly improved with the introduction of the load current feed-forward scheme. Figure 8.9 gives the experimental waveforms of a single inverter when the load is step changed between the rated resistive load current and over current. Here, the output current of the inverter is limited at 1.5 times of the rated load. As seen, the inductor current is limited and thus the load current is limited, and the inverter has good dynamic performance. Figure 8.10 gives the experimental waveforms of the DC–AC IPOP system when the load steps up and down between the rated resistive load current and overcurrent. It also can be seen that the load currents of the two inverter modules are successfully

184

8 An Improved Average Current Control …

voref

iave Kio

voref +_

Gvo1

+

+ iref1

1/KiL1

vof

sCf1

iCf1 _

iLf1 +

io1

Kvo

Module # 1

Average Current Bus irefj +

+ +

voref +_

Gvon

+

ioj

1 n

+

vo

+ +

ZLd

Kio

+ irefn

1/KiLn

vof

iLfn + _ iCfn

ion sCfn

Kvo

Module #n

Fig. 8.7 Block diagram of the control in DC–AC IPOP systems 116.0

Output voltage (V)

115.5 115.0 114.5 114.0 113.5 113.0 112.5

Without the load current feed-forward With the load current feed-forward 0

5

10 Output current (A)

15

20

Fig. 8.8 Measured output characteristics curves of the system

limited, and the DC–AC IPOP system exhibits excellent dynamic performance. It is evident that the current-limiting function is effective, which is well in agreement with the theoretical analysis.

8.4 Experimental Results

185

vo [100V/div]

iL [10A/div]

io [10A/div]

Time: [5ms/div] Fig. 8.9 Experimental waveforms of a single inverter module with stepped load current

vo [100V/div]

io1 [10A/div]

io2 [10A/div]

Time: [5ms/div] Fig. 8.10 Experimental waveforms of the DC–AC IPOP system with stepped load current

Figure 8.11 gives steady-state experimental waveforms of the DC–AC IPOP system with the improved method at full resistive load. As seen, the difference between the output currents of the two inverter modules is fully eliminated, which means OCS is achieved. Figure 8.12 shows the dynamic experimental waveforms of the DC–AC IPOP system when the load current is step changed between the full resistive load and no-load. It can be seen that the output currents of the two modules are equal during the transient showing that OCS is also achieved. Figure 8.13a shows the experimental waveforms of the system output voltage vo , output current of module 1, io1 , output current of module 2, io2 , when module 2 is plugged into the DC–AC IPOP system. As seen, before module 2 is plugged into

186

8 An Improved Average Current Control …

–

Fig. 8.11 Steady-state waveforms of the DC–AC IPOP system at full resistive load

Fig. 8.12 Dynamic waveforms of the DC–AC IPOP system

the system, module 1 provides all the load current; and when module 2 is plugged into the system, io1 declines, and io2 increases from zero, and finally, they provide the load current equally, and there is no obvious distortion in vo . Figure 8.13b gives the experimental waveforms when module 2 is plugged out from the DC–AC IPOP system. When the control circuit produces the plug-out command, power switches in module 2 are turned off, and module 2 is plugged out from the DC–AC IPOP system after about eight output cycles due to the delay of the output relay operation. So, during the period, module 1 supplies power to the load and the output filter capacitor of module 2, and io2 initially reduces to the output filter capacitor current and then reduces to zero after around 20 ms. It can be seen that there is no obvious distortion in vo in the above process. Figure 8.13 illustrates that the DC–AC IPOP system has excellent dynamic performance, and hot-swap is also achieved.

8.5 Summary

187

v o [100V/div]

i o1 [5A/div]

i o2 [5A/div]

Time: [2ms/div]

(a) Module 2 is plugged into the IPOP operation

v o [100V/div]

io1[5A/div] io2[5A/div] Time: [2ms/div]

(b) Module 2 is cut off from thedc-ac IPOP system Fig. 8.13 Hot-swap waveforms of the DC–AC IPOP system

8.5 Summary This chapter first introduces the main circuit topology and control method of the inverter module, and the operation principle of the current three-state hysteresis modulation is also presented. The load current feed-forward scheme is incorporated for making the output voltage of the inverter independent of load current. It

188

8 An Improved Average Current Control …

is worth noting that the load current feed-forward node should be located before the current-limiting unit, for not only improving the inverter output characteristics but also limiting the load current. This idea is also extended to the DC–AC IPOP systems. Second, this chapter puts forward the system architecture of average current distributed control strategy, which includes two critical signal buses, namely, synchronous voltage reference bus and average current reference bus. The voltage reference of each inverter module tracks the phase information of synchronous bus for the purpose of synchronization. In the meanwhile, the current references of all modules are averaged to obtain the common current reference signal for each module so as to realize output current sharing among the modules. With this average current distributed control, the DC–AC IPOP system achieves the hot-swap function and redundant ability. Finally, the prototype of two-module DC–AC IPOP systems is built in the lab, and the steady-state and dynamic experiments are provided. The experimental results are consistent with the theoretical analysis verifying the superiority of the proposed improved strategy and also realizing hot-swap of DC–AC IPOP system.

References 1. Bekiarov, S.B., Emadi, A.: Uninterruptible power supplies: classification operation dynamics and control. In: Proceedings of IEEE Applied Power Electronics Conference and Exposition, pp. 597–604 (2002) 2. Chen, X., Kang, Y., Chen, J.: Operation, control technique of parallel connected high power three-phase inverters. In: Proceedings of IEEE International Power Electronics and Motion Control Conference, pp. 956–959 (2004) 3. Chen, J., Chu, C.: Combination voltage-controlled and current-controlled PWM inverter for UPS parallel operation. IEEE Trans. Power Electron. 10(5), 547–558 (1995) 4. Lee, W.C., Lee, T.K., Lee, S.H., Kim, K.H., Hyun, D.S., Suh, I.Y.: A master and slave control strategy for parallel operation of three-phase UPS systems with different ratings. In: Proceedings of IEEE Applied Power Electronics Conference and Exposition, pp. 456–462 (2004) 5. Chiang, S.J., Lin, C.H., Yen, C.Y.: Current limitation control technique for parallel operation of UPS inverters. In: Proceedings of IEEE Power Electronics Specialists Conference, pp. 1922–1926 (2004) 6. Perez-Ladron, G., Cardenas, V., Espinosa, G.: Analysis and implementation of a master-slave control based on a passivity approach for parallel inverters operation. In: Proceedings of IEEE International Power Electronics Conference, pp. 1–5 (2006) 7. Xing, Y., Huang, L., Yan, Y.: Redundant parallel control for current regulated inverters with instantaneous current sharing. In: Proceedings of IEEE Power Electronics Specialists Conference, pp. 1438–1442 (2003) 8. Fang, T., Ruan, X., Xiao, L., Liu, A.: An improved distributed control strategy for parallel inverters. In: Proceedings of IEEE Power Electronics Specialists Conference, pp. 3500–3505 (2008) 9. Tuladhar, A.: Advanced control techniques for parallel inverter operation without control interconnections. Ph.D. dissertation, The University of British Columbia, Vancouver, Canada (2000) 10. Lin, X., Feng, F., Duan, S., Kang, Y., Chen, J.: The droop characteristic decoupling control of parallel connected UPS with no control interconnection. In: Proceedings of IEEE International Electric Machines and Drives Conference, pp. 1777–1780 (2003)

References

189

11. De Brabandere, K., Bolsens, B., Van den Keybus, J., Woyte, A., Driesen, J., Belmans, R.: A voltage and frequency droop control method for parallel inverters. IEEE Trans. Power Electron. 22(4), 1107–1115 (2007) 12. Yao, W., Chen, M., Gao, M., Qian, Z.: A wireless load sharing controller to improve the performance of parallel-connected inverters. In: Proceedings of IEEE Applied Power Electronics Conference and Exposition, pp. 1628–1631 (2008) 13. Guerrero, J., Vicuña, L., Matas, J., Castilla, M., Miret, J.: Output impedance design of parallelconnected UPS inverters with wireless load-sharing control. IEEE Trans. Ind. Electron. 52(4), 1126–1135 (2005) 14. Kheraluwala, M.H., Divan, D.M.: Delta modulation strategies for resonant link inverters. IEEE Trans. Power Electron. 5(2), 220–228 (1990) 15. Chen, L., Xiao, L., Gong, C., Yan, Y.: Circulating current’s characteristics analysis and the control strategy of parallel system based on double close-loop controlled VSI. In: Proceedings of IEEE Power Electronics Specialists Conference, pp. 4791–4797 (2004) 16. Fang, T., Ruan, X., Xiao, L., Liu, A.: An improved distributed control strategy for parallel inverters. In: Proceedings of IEEE Power Electronics Specialists Conference, pp. 3500–3505 (2008) 17. Fang, T.: Control strategies for input-parallel-output-parallel and input-series-output- series connected combined inverter systems (in Chinese). Ph.D. dissertation, Nanjing University of Aeronautics and Astronautics, Nanjing, China (2008)

Chapter 9

Input Voltage Sharing Control Strategy for ISOP Systems Under Extreme Load Conditions

Abstract Input-series and output-parallel (ISOP) systems are very suitable for high input voltage and high power applications. Input voltage sharing (IVS) and output current sharing (OCS) of the constituent modules in the ISOP systems should be ensured. This chapter reveals the failure mechanism of IVS of the ISOP systems under extreme load conditions. Based on the existed IVS control strategies, this chapter proposes two methods, including changing the weighted averages of IVS loops and adding an IVS auxiliary circuit, to achieve IVS of the modules under the extreme load conditions. The experimental results are presented to verify the effectiveness of the proposed methods. Keywords Input-series output-parallel system · Extreme load conditions Input voltage sharing · Weighted average · Auxiliary circuit For the ISOP systems, including the DC–DC ISOP system and DC–AC ISOP system, the control strategies proposed from Chaps. 2 to 7 realize IVS all by adjusting the module output power by means of regulating the duty cycle of the constituent modules. When the ISOP systems operate under extreme load conditions such as light load or short-circuit, the module output powers are very small, and the duty cycles are also very small. With the small duty cycle, the adjustment range of the duty cycle is limited, and the abovementioned control strategies cannot ensure IVS anymore, which may cause the module with higher input voltage damaged. This chapter will reveal the failure mechanism of IVS of the ISOP systems under extreme load conditions. Furthermore, this chapter will propose two methods, including changing the weighted averages of IVS loop and adding an auxiliary IVS circuit to achieve IVS for ISOP systems under extreme load conditions [1].

© Springer Nature Singapore Pte Ltd. and Science Press, Beijing, China 2019 X. Ruan et al., Control of Series-Parallel Conversion Systems, CPSS Power Electronics Series, https://doi.org/10.1007/978-981-13-2760-5_9

191

192

9 Input Voltage Sharing Control Strategy for ISOP Systems …

Iin

Iin1

+ Icd1 Vin1 _ Cd1

Io1

Io +

DC-DC (DC-AC) Module

Cf1

RLd

Vo _

Module 1

Vin

Iin2 + Icd2 Vin2 _ Cd2

Io2 DC-DC (DC-AC) Module

Cf2 Module 2

Fig. 9.1 Two-module ISOP system

9.1 Reason of Input Voltages Imbalance of ISOP Systems Under Extreme Load Conditions In Chaps. 2 and 5, the conversion efficiencies of the modules are assumed to be equal. However, this is impossible in practical circuits. This section will analyze the effect of mismatched conversion efficiencies on IVS under extreme load conditions. For simplification of analysis, a two-module ISOP system is taken as an example as shown in Fig. 9.1. Assuming the ISOP system operates at steady state and denoting the loss of module as Plossj (j = 1, 2), we have Pin1 Po1 + Ploss1 (9.1) Pin2 Po2 + Ploss2 where Pin1 and Po1 are the input power and output power of module 1, respectively and Pin2 and Po2 are the input power and output power of module 2, respectively. At steady state, the average currents of the input dividing capacitors are zero, i.e., I cd1 = I cd2 0, then the input currents of two modules are equal, i.e., Iin1 Iin2 Iin

(9.2)

With the help of IVS loops, the input voltages of two modules are equal, i.e., Vin1 Vin2 Vin /2

(9.3)

Combining (9.2) and (9.3), we have Pin1 Vin1 Iin1 Vin2 Iin2 Pin2

(9.4)

9.1 Reason of Input Voltages Imbalance of ISOP Systems …

193

This means that if IVS is achieved, the module input powers will be balanced. Combining (9.1) and (9.4), we have Po2 − Po1 Ploss1 − Ploss2

(9.5)

Assuming Ploss1 > Ploss2 , we have Po1 < Po2

(9.6)

From the above analysis, if the module losses are not identical, the IVS loops will adjust the module output power. The output power of the module with larger losses will be reduced and the output power of the module with lower losses will be increased. By doing so, the module input powers will be balanced and thus IVS between the modules is ensured. Obviously, the difference of the module output powers equals the difference of the module losses. Therefore, in order to realize IVS, the total output power of the ISOP system must be higher than the difference of the module losses. Only with this condition, it is possible for the IVS loops to regulate the module output power to make the module input powers be balanced. In the meanwhile, the IVS loops must have enough regulation ability to ensure that the difference of the module output powers can compensate the difference of the module losses. The output power of ISOP systems is close to zero under extreme load conditions, i.e., light load, no–load, or short-circuit, but the difference of the module losses may still exist and larger than the system output power, which means that the IVS loops cannot compensate for the difference of the module losses by regulating the module output power, leading to input voltages imbalance of the ISOP systems. According to (9.5), we have Io2 − Io1

Po2 − Po1 Ploss1 − Ploss2 Vo Vo

(9.7)

Equation (9.7) indicates that control strategies to ensure IVS for ISOP systems will cause output currents imbalance among the modules when the module conversion efficiencies are different. The output voltage of ISOP systems is close to zero under the short-circuit condition with current-limiting. It will cause large output current different among the modules even if the difference of the module losses is smaller when IVS is ensured. The module with larger output current may fail due to the overcurrent of the power devices. Therefore, we need to not only limit the system total output current but also limit the individual module output current under the condition of short-circuit current-limiting.

194

9 Input Voltage Sharing Control Strategy for ISOP Systems …

Fig. 9.2 Equivalent loss model under extreme load conditions

Cd1

Rloss1

Cd2

Rloss2

Cdn

Rlossn

Vin

9.2 Input Voltage Sharing Control Strategy for ISOP Systems Under Extreme Load Conditions In Sect. 9.1, it is pointed out that under extreme load conditions, the difference among the module losses is the main reason of the input voltages imbalance of ISOP systems with the control strategies mentioned in the previous chapters. Under extreme load conditions, the output power of each module is very small and approximately equal to its loss, therefore, the equivalent loss model of ISOP system under extreme load conditions can be illustrated in Fig. 9.2, where Rlossj (j = 1, 2, …, n) represents the module loss and it should be noted that Rlossj varies depending on the input voltage. At steady state, all Rlossj (j = 1, 2, …, n) have the same flowing current. Hence, the larger the Rlossj , the higher the module input voltage is, and the smaller Rlossj , the lower the module input voltage is. Thus, it is required that Rloss1 Rloss2 ··· Rlossn for achieving IVS. There are two methods to achieve IVS for ISOP systems under extreme load conditions, that is, (1) regulating Rlossj (j = 1, 2, …, n) to be equal and (2) connecting a parallel adjustable resistance to Rlossj (j = 1, 2, …, n) to make the all the equivalent resistances to be equal as shown in Fig. 9.3.

9.2.1 Changing the Weighted Values of the IVS Loops At light load, there is no need for all modules in the ISOP systems to work, and only a few of modules can be activated for supplying the needed load power. For the disabled modules, their equivalent resistances Rloss are infinite. If we can make the modules work intermittently and adjust their working time, then the equivalent resistances of modules can be adjusted to be equal to realize IVS under extreme load conditions.

9.2 Input Voltage Sharing Control Strategy for ISOP …

195

Fig. 9.3 Method of achieving equal equivalent resistance

Cd1

Rloss1

Cd2

Rloss2

Cdn

Rlossn

R1

R2

Vin

Rn

The DC–DC ISOP system is taken as an example to discuss the IVS control strategy under extreme load conditions, which is also suitable for the DC–AC ISOP system. For the convenience of illustration, the IVS control strategy for DC–DC ISOP system (see Fig. 3.19) is redrawn here as shown in Fig. 9.4. In order to explain the proposed control strategy with changeable weighted values of the IVS loops, the weighted value K vcd is added to the output of IVS regulators. When K vcd = 1, Fig. 9.4 is the same as Fig. 3.19. When the input voltage of module j (j = 1, 2,…, n) is lower than V in /n, the output of the IVS regulator vin_EAj increases and the current reference of the inner current loop declines, and when the current reference of inner current loop declines to be lower than zero, vEAj will be also lower than zero, which is lower than the minimum setting value of PWM controller, then the drive signals will be shut down and this module is disabled [2]. Then, the input current I inj 0 and the current of the dividing capacitor I cdj = I in , which charges the dividing capacitor and the module input voltage rises. When the module input voltage rises to be higher V in /n, vin_EAj declines, the current reference of inner current loop as well as the corresponding vEAj increases to be higher than the minimum setting value of PWM controller, then the drive signals will be turned on and the module works again. It can be seen that the modules will work intermittently if the weighted value K vcd is large enough to achieve IVS. To avoid severely regulating the module output current in transient, the weighted value K vcd cannot be too large. However, this cannot ensure the intermittent work of the modules under extreme load conditions. In other words, we hope that K vcd 1 at heavy load and K vcd > 1 at light load. So, two weighted values are set in the control strategy, i.e., 1 and K vcd_light , corresponding to the heavy load and light load conditions, respectively as shown in Fig. 9.5. The two weighted values are selected by a single-pole double-throw, which is controlled by a comparator. When the system output current io is larger than the threshold current I th , the output of comparator is low level and the weighted value of 1 is selected; and if io is smaller than I th , the output of comparator is high level and the weighted value K vcd_light is selected. Generally, the module loss is very small at light load, and the difference among the

196

9 Input Voltage Sharing Control Strategy for ISOP Systems …

+ Vin/n

+ Vin/n

+ Vin/n

– vin1

ivcdn

Gvcd vin_EAn

Kvcd

Gc

vEA1 Comparator 1 _ RS Flip-flop

Drive circuit 1

RS Flip-flop

Drive circuit 2

vEAn Comparator n _ RS Flip-flop + VRAMP n

Drive circuit n

VRAMP 1 – iLf2 +

–

Kvcd

vinn

–

+ ivcd2

Gvcd vin_EA2 IVS loops

– iLf1 +

–

Kvcd

– vin2

–

+

ivcd1 Gvcd vin_EA1

+

Gc Inner current loops

–

– iLfn +

+

Gc

+

vEA2 Comparator 2 _ VRAMP 2

+

vof

Gvo Voref vo_EA Output voltage loop

Fig. 9.4 IVS control strategy for ISOP system Con

+ Vin/n

+ Vin/n

– vin1 Gvcd vin_EA1 – vin2

Gvcd vin_EA2 IVS loops

+ Vin/n

+

–

–

vinn Gvcd vin_EAn

–

io

+

Ith

Comp

1

–

Kvcd_light

+

1

–

Kvcd_light

+

1

–

Kvcd_light

+

– iLf1 +

Gc

Vcl_mod

vEA1 Comparator1 _ VRAMP 1

+

RS Flip-flop

Drive circuit 1

RS Flip-flop

Drive circuit 2

RS Flip-flop

Drive circuit n

vEA2 Comparator 2 _

– iLf2 +

Gc Inner current + VRAMP 2 loops

Vcl_mod

– iLfn + Vcl_mod

Gc

vEAn Comparator n _ VRAMP n

+

vof

Gvo Voref vo_EA Output voltage loop

Vcl_total

Fig. 9.5 Control strategy with changeable weighted values of IVS loops

module losses can be considered to be less than 5%I o , where I o is the system rated output current. So, we set I th 5%I o . The output current of ISOP system has to be limited to protect the system. So, a Zener diode is introduced to the output of the output voltage regulator, and the Zener voltage V cl_total corresponds to the limited system output current as shown in Fig. 9.5. In the meanwhile, Zener diodes are introduced to the current references of the inner current loops to limit the individual module output current, and the Zener

9.2 Input Voltage Sharing Control Strategy for ISOP …

197

voltage V cl_mod corresponds to the limited module output current. V cl_mod must be higher than V cl_total for ensuring that the IVS loops could adjust the inner current loops during the system current-limiting.

9.2.2 Input Voltage Sharing Auxiliary Circuit Referring to the control strategy with changeable weighted values of IVS loops as shown in Fig. 9.5, when the output of ISOP system is shorted, the adjustment of the IVS loops to the current reference of inner current loops is limited by V cl_mod . This means that the IVS loops fail to adjust the current reference of the inner current loop, and cannot achieve IVS under short-circuit condition, and it is only effective at light load. Therefore, we have to introduce IVS auxiliary circuits to the input sides of the modules when short-circuit occurs to regulate the equivalent resistances of all the modules to be equal for IVS. Because the module losses are different and vary with the input voltage, the IVS auxiliary circuits should have the following features: (1) the equivalent resistance of the IVS auxiliary circuit is adjustable; (2) ease of modularization; (3) the IVS auxiliary circuits are only enabled when IVS is not achieved under extreme load conditions to avoid extra loss under normal load conditions. The IVS auxiliary circuits with adjustable resistance and the corresponding control circuits are shown in Fig. 9.6. Each IVS auxiliary circuit is composed of a resistor Rreg and a switch Qaj (j = 1, 2, …, n). Qaj is PWM controlled and the equivalent resistance of the IVS auxiliary circuit is Rreg /DQaj , where DQaj is the duty cycle of Qaj . When DQaj 1, the power dissipated by Rreg must be larger than the maximum difference among the module losses. When the ISOP system works at normal load, the IVS loops can realize IVS among modules, and Qaj is shut off. When the system total output power is lower than the difference among the module losses, the input voltage of the module with larger Rloss will increase from the average value. Until the module input voltage is higher than the allowed maximum input voltage V mod_max , the IVS auxiliary circuit starts to work and regulate DQaj to adjust the power dissipated by Rreg , limiting the module input voltage at V mod_max . The IVS auxiliary circuit can ensure the ISOP system working normally under extreme load conditions, and its structure is simple, but the loss under extreme load conditions is increased. To ensure the stable operation of ISOP system in the full load range, we can utilize the IVS auxiliary circuit or combine the IVS auxiliary circuit with the control strategy with changeable weighted values of the IVS loops. When the output current is lower than I th , we can use the control strategy with changeable weighted values of the IVS loops to realize IVS and the auxiliary circuit is disabled. When the ISOP system output is shorted with current-limiting, the output current is higher than I th , the IVS auxiliary circuit will be activated to limit the module input voltage at V mod_max to ensure the system work stably.

198

9 Input Voltage Sharing Control Strategy for ISOP Systems …

RLd Rreg Cd1

Cf

Qa1

Rreg Vin

Module 1

Cd2

Qa2

Rreg Cdn

Qan

+ Vo –

Module 2

Module n

(a) IVS auxiliary circuit VRAMPComparator _

_

Vmod_max Vinj

+

EA

+

Qaj Driver

(b) Control circuit of IVS auxiliary circuit Fig. 9.6 IVS auxiliary circuit and control circuit

9.3 Simulation and Experimental Verification A DC–DC ISOP system consisting of two phase-shifted full-bridge (PSFB) converter modules is taken as an example as shown in Fig. 9.7, to verify the effectiveness of the proposed methods to realize IVS under extreme load conditions. The specifications of the ISOP system are as follows: input voltage V in 540 Vdc, output voltage V o 48 Vdc, and rated system output current I o 104 A. The main parameters of the PS-FB converters are as follows: switching frequency f s 100 kHz, filter inductance L f1 L f2 10 µH, resonant inductance L r1 L r2 6 µH, and transformer turns ratio K = 14:3. To simulate the module loss, resistors with two different resistances are connected in series with the power switches, i.e., Rl1 0.5 and Rl2 2 , and it is readily to know that Rloss1 < Rloss2 when the input voltages of two modules are equal under extreme load conditions. The threshold current I th is set at 10%I o , i.e., I th 10.4 A. Rreg is chosen as 1 k, and its dissipated power is 73 W when Qaj (j = 1, 2) is always ON and the module input voltage is 270 Vdc. V mod_max is set at 110%V in /2, i.e., V mod_max 297 Vdc. Figure 9.8 shows the simulation waveforms of the ISOP system employing the IVS control strategy (see Fig. 9.4 and K vcd 1) under no-load condition. vin1 and vin2 are the input voltages of modules 1 and 2, respectively, vAB1 and vAB2 are the

9.3 Simulation and Experimental Verification

199

Module 1

Lf1 T1

Q11

Rreg

+ Vin1 –

Rl1 A1 Rl1

Q12 T1

Lr1 ip1

Rl1 B1 Rl1

*

* DR11

iLf1 Cf

RLd + Vo –

* DR12

Qa1 Q13

+ vrect1 –

Q14 Lf2

Vin T2 Rreg

Q21 + Vin2 –

Rl2 A2 Rl2

Q22 T2

Lr2 ip2

Rl2 B2 Rl2

*

Qa2 Q23

+ * DR21 v rect2 –

iLf2

* DR22

Q24 Module 2

Fig. 9.7 Circuit diagram of ISOP system consisting of two PSFB modules

(V)

400

vin2

300

vin1

200

(V)

0 −200 −400 400 200

(V)

100 400 200

0 −200 −400 9.500m

vAB1

vAB2

9.504m

9.508m

9.512m

9.516m

9.520m

9.524m

9.528m

t (s) Fig. 9.8 Simulation waveforms with the IVS control strategy (see Fig. 9.4 and K vcd 1)

voltages across the middle points of the phase-legs of the two modules. It can be seen that vin2 is much higher than vin1 due to Rloss1 < Rloss2 and the voltage difference reaches 220 Vdc, and IVS is not realized under no-load condition.

200

9 Input Voltage Sharing Control Strategy for ISOP Systems … 275

(V)

vin1 270

vin2 265

(V)

400 200

vAB1

0 −200 −400

(V)

400 200

vAB2

0 −200 −400 38.77m

38.78m

38.79m

38.80m

38.81m

38.82m

38.83m

38.84m

38.85m

t (s) Fig. 9.9 Simulation waveforms with the IVS control strategy (see Fig. 9.5 and K vcd_light 2) under no-load condition

Figure 9.9 shows the simulation waveforms of the ISOP system when employing the control strategy with changeable weighted values of the IVS loops (see Fig. 9.5 and K vcd_light 2) under no-load condition. It can be seen that module 1 works intermittently due to Rloss1 < Rloss2 and the average input voltages of two modules are approximately the same. Figure 9.10 shows the simulation waveforms of the ISOP system when employing the control strategy (see Fig. 9.4) and IVS auxiliary circuit under no-load condition. Figure 9.10a shows the waveforms under no-load condition at steady state, where Qa2 is the drive signal of power switch in the IVS auxiliary circuit of module 2. The IVS auxiliary circuit of module 2 works due to Rloss1 < Rloss2 and controls the input voltage of module 2 at 110%V in /2, i.e., 297 Vdc. Figure 9.10b shows the waveforms corresponding to load stepping between half load and no load. At half load, the system output power is larger than the difference between the module losses, the IVS loops can realize IVS and the auxiliary circuit is disabled. At no load, the input voltage of module 2 rises and when it reaches 297 Vdc, the IVS auxiliary circuit of module 2 works and controls the input voltage of module 2 at 297 Vdc, and consequently, the input voltage of module 1 is controlled at 243 Vdc, and the system can operate steadily. Figure 9.11a shows the simulation waveforms of the ISOP system with the control strategy (see Fig. 9.5) and without the IVS auxiliary circuit under short-circuit current-limiting condition. The system output current is limited at 110 A. It can be seen that vin2 is higher than vin1 due to Rloss1 < Rloss2 , the current reference of inner current loop of module 2 increases and when its IVS loop regulator is saturated,

9.3 Simulation and Experimental Verification

201

(V)

(V)

(V)

(V)

400

vin2

300

vin1

200 100 400 200 0 −200 −400 400 200 0 −200 −400 20

vAB1

vAB2

Qa2

10 0 21.662m

21.666m

21.670m

21.674m

21.678m

21.682m

21.686m

21.690m

t (s)

(V)

(A)

(a) Waveforms at steady state 60 40 20 0 −20 60

io

40

vo

20

(V)

0 350

vin2

300 250

vin1

(V)

200 20

Qa2

10 0

2.0m

4.0m

6.0m

8.0m

10.0m

12.0m

14.0m

t (s)

(b) Waveforms corresponding to stepped load Fig. 9.10 Simulation waveforms with the control strategy (see Fig. 9.4) and IVS auxiliary circuit under no-load condition

the output current of module 2 is 65 A, which is the sum of average limited output current 55 and 10 A corresponding to the regulation of IVS loop. The output current of module 1 is 45 A and the difference of the input voltages of two modules reaches 100 Vdc.

(V)

(V)

(V)

(A)

202

9 Input Voltage Sharing Control Strategy for ISOP Systems … 70 60 50 40 30 350

iLf2 iLf1

300

vin2

250

vin1

200 400 200 0 -200 -400 400 200 0 -200 -400 19.76m

vAB1

vAB2

19.78m

19.8m

19.82m

19.84m

19.86m

19.88m

17.34m

17.36m

t (s)

(V)

(V)

(V)

(V)

(A)

(a) Without IVS auxiliary circuit 70 60 50 40 30 350

iLf2 iLf1

vin2

300 250

vin1

200 400 200 0 -200 -400 400 200 0 -200 -400 20

vAB1

vAB2

Qa2

10 0

17.24m

17.26m

17.28m

17.30m t (s)

17.32m

(b) With IVS auxiliary circuit Fig. 9.11 Simulation waveforms of ISOP system with the control strategy (see Fig. 9.5) under short-circuit current-limiting condition

9.3 Simulation and Experimental Verification

203

Figure 9.11b shows the simulation waveforms of the ISOP system with both the control strategy (see Fig. 9.5) and the IVS auxiliary circuit under short-circuit currentlimiting condition. It can be seen that the input voltage of module 2 is controlled at 297 Vdc and its output current is 60 A. An ISOP system prototype is also built in our lab with the same specifications as that in the simulation. For each module with rated input voltage 270 Vdc and under no-load condition, the measured module losses are 48.7 and 33.4 W for modules 1 and 2, respectively, and the power difference is 15.3 W. Figure 9.12a shows the experimental waveforms of the ISOP system with the IVS control strategy (see Fig. 9.4 and K vcd 1) when the system output current is 1 A, i.e., the system output power is 48 W. It can be seen that two modules share the system input voltage well because the system output power is larger than the difference of two module losses. Figure 9.12b shows the experimental waveforms of the ISOP system with the IVS control strategy (see Fig. 9.4 and K vcd 1) under no-load condition. It can be seen that the input voltage of module 2 is much higher than that of module 1 because the no-load loss of module 2 is smaller than that of module 1 with the same input voltage, and the difference of the module input voltages reaches 180 Vdc. Therefore, the ISOP system must adopt special IVS control strategy under extreme load conditions. Figure 9.13 shows the experimental waveforms of ISOP system when employing the control strategy with changeable weighted values of the IVS loops under no-load condition. Figure 9.13a shows the waveforms at steady state. The waveforms of vAB are distorted because the frequency of vAB is 100 kHz but the time scale of the figure is 500 ms/div. The waveforms of vin1 and vin2 indicate that the two modules work intermittently because the two dividing capacitors charge and discharge alternately, leading to the input voltages rise and fall alternately. The average values of the module input voltages are approximately equal. Figure 9.13b shows the waveforms corresponding to load stepping between quarter load and no load. The experimental results shown in Fig. 9.13 verify that the control strategy with changeable weighted values of the IVS loops can ensure the IVS in the full load range. Figure 9.14 shows the experimental waveforms of the ISOP system with the control strategy (see Fig. 9.4 and K vcd 1) and IVS auxiliary circuit under no-load condition. Here, V mod_max is set at 110%V in /2, i.e., V mod_max 297 Vdc. Figure 9.14a, b show the waveforms of vin1 , vin2 , vAB1 , vAB2 , and the drive signal of the auxiliary circuit of module 2 at steady state. As seen, the input voltage of module 2 is controlled at 297 Vdc with the help of IVS auxiliary circuit, and the input voltage of module 1 is controlled at 243 Vdc. Figure 9.14c shows the waveforms corresponding to load stepping between quarter load (25 A) and no load. It can be seen that the IVS is achieved by the control strategy, and the IVS auxiliary circuit is disabled when the output power is larger than the difference between the module losses. Under noload condition, the IVS auxiliary circuit of module 2 starts to work when the input voltage of module 2 rises to 297 Vdc and regulates the input voltage module 2 at 297 Vdc, and consequently, the input voltage of module 1 is controlled at 243 Vdc. The experimental results shown in Fig. 9.14 verify that the IVS auxiliary circuit works well under no-load condition.

204

9 Input Voltage Sharing Control Strategy for ISOP Systems …

C1: vin2 [100V/div] C2: vin1 [100V/div]

C1/C2

C3: vAB1 [400V/div]

C3

C4: vAB2 [400V/div]

C4

Time: [2μs/div]

(a) Output current is 1 A C1: vin2 [100V/div] C2: vin1 [100V/div]

C1/C2 C3

C3: vAB1 [400V/div]

C4: vAB2 [400V/div]

C4

Time: [2μs/div]

(b) No-load Fig. 9.12 Experimental waveforms with the IVS control strategy (see Fig. 9.4 and K vcd 1)

Figure 9.15 shows the experimental waveforms of the ISOP system with the control strategy (see Fig. 9.4 and K vcd 1) and IVS auxiliary circuit under shortcircuit current-limiting condition. For the sake of safety, the system output current is limited at 50 A and output current of each module is limited at 27.5 A (110%I o /2). From Fig. 9.15a, it can be seen that the output current of module 2 is limited at 27.5 A because the no-load loss of module 2 is smaller than that of module 1 under the same input voltage, and output current of module 1 is 22.5 A. Figure 9.15b shows that the input voltage of module 2 is controlled at 297 Vdc with the help of IVS

9.3 Simulation and Experimental Verification

205

C1: vin2 [100V/div] C2: vin1 [100V/div]

C1/C2 C3

C3: vAB1 [400V/div] C4: vAB2 [400V/div]

C4

Time: [500ms/div] (a) Waveforms of steady state

C3

C3: io [20A/div]

C1: vin2 [100V/div] C2: vin1 [100V/div]

C1/C2

Time: [500ms/div] (b) Waveforms corresponding to stepped load Fig. 9.13 Experimental waveforms of ISOP system when employing the control strategy with changeable weighted values of the IVS loops under no-load condition

auxiliary circuit, and the input voltage of module 1 is 243 Vdc. Figure 9.15c shows the waveforms corresponding to load stepping between 25 A and limited current (50 A). It can be seen that the IVS is achieved by the control strategy and the IVS auxiliary circuit is disabled under normal load condition. The experimental results shown in Fig. 9.15 verify that the IVS auxiliary circuit works well under short-circuit current-limiting condition.

206

9 Input Voltage Sharing Control Strategy for ISOP Systems … C1: vin2 [100V/div] C2: vin1 [100V/div]

C1/C2 C3

C4

C3: vAB1 [400V/div]

C4: vAB2 [400V/div] Time: [2μs/div]

(a) Waveforms of vin1, vin2, vAB1, and vAB2 of steady state C1: vin2 [100V/div] C2: vin1 [100V/div]

C1/C2

C3 C3: Qa2 [10V/div] Time: [2μs/div]

(b) Waveforms of vin1, vin2 and drive signal of auxiliary circuit of module 2

C4

C4: io [20A/div]

C1: vin2 [100V/div]

C2: vin1 [100V/div] C3 C1/C2

C3: Qa2 [20V/div]

Time: [500ms/div]

(c) Waveforms corresponding to load stepping Fig. 9.14 Experimental waveforms of ISOP system with the control strategy (see Fig. 9.4) and IVS auxiliary circuit under no-load condition

9.3 Simulation and Experimental Verification

M1: iLf1 [10A/div]

207

C2: iLf2 [10A/div]

M1/C2

C3

C4

C3: vAB1 [400V/div]

C4: vAB2 [400V/div] Time: [2μs/div] (a) Waveforms of vAB1, vAB2, iLf1, and iLf2 of steady state

C2: vin2 [100V/div] C1: vin1 [100V/div]

C1/C2

C3 C3: Qa2 [10V/div] Time: [2μs/div]

(b) Waveforms of vin1, vin2 and drive signal of auxiliary circuit of module 2

C4: vo [40V/div]

C4

C2: vin2 [100V/div] C1: vin1 [100V/div] C3 C3: Qa2 [20V/div]

C1/C2 Time: [500 ms/div]

(c) Waveforms corresponding to load stepping

Fig. 9.15 Experimental waveforms of ISOP system with the control strategy (see Fig. 9.4) and IVS auxiliary circuit under short-circuit current-limiting condition

208

9 Input Voltage Sharing Control Strategy for ISOP Systems …

9.4 Summary This chapter reveals the failure mechanism of IVS of the ISOP systems employing the control strategies proposed from Chaps. 2 to 7 under extreme load conditions, and points out that when the output power of the ISOP systems is smaller than the difference among the module losses, these control strategies cannot regulate enough module output power to compensate for the loss difference, leading to the IVS failure. Moreover, output currents imbalance will occur under short-circuit current-limiting condition. To solve these problems, two methods of achieving IVS under extreme load conditions was proposed, namely the control strategy with changeable weighted values of the IVS loops and the IVS auxiliary circuit. It is recommended that the control strategy with changeable weighted values of the IVS loops should be used under light-load or no-load condition, and the IVS auxiliary circuit should be used under short-circuit current-limiting condition. The effectiveness of the two methods is verified by simulation and experimental results.

References 1. Yan, H., Ruan, X., Chen, W.: The input voltage sharing control strategy for input-series and output-parallel converter under extreme conditions. In: Proceedings of IEEE Energy Conversion Congress and Exposition, pp. 662–667 (2009) 2. UCC3895 BiCMOS advanced phase-shift PWM controller datasheet, http://www.ti.com/lit/ds/ symlink/ucc3895.pdf

Index

A Active current, 125, 126 Active power, 4, 21, 23, 124, 126, 131, 133, 150, 156, 157 Aeronautical static inverter, 16, 175 Alternating current (AC), 1, 107 Auxiliary circuit, 197, 200, 203, 208 Average current control, 12, 18, 24, 175, 180 B Bipolar junction transistor, 2 Bipolar modulation, 177 BJT, see Bipolar junction transistor Body diode, 56 Boost-derived converter, 116 Buck/boost-derived converter, 37 Buck-derived converter, 37, 40, 41 Bus voltage, 130–132, 164, 177 C Central control unit, 175 Centralized control, 16, 17, 24, 122, 148–151, 169, 171, 172, 175 Changeable weighted values, 195–197, 200, 203, 208 Circulating current, 19, 60, 126 Common current reference, 118, 171, 181, 183, 188 Common output voltage control loop, 10, 11, 67 Compensator, 71–73, 136, 137, 163, 164 Compound balanced control strategy, 21, 22, 113–115, 118, 119, 121, 124, 125, 127, 141, 154, 156, 164–166, 168, 172 Constant power source, 35, 114

Constituent inverter modules, 17, 19, 124 Controlled sources, 60 Converter circuit, 2 Corner frequency, 71 Cross-feedback control, 12, 22, 23 Cross-feedback for output inductor current, 22 Crossover frequency, 71–73 Crystal oscillator, 17 Current closed-loop controlled, 8 cross-feedback control strategy, 12, 14 deviation, 17 follower, 158, 178, 183 hysteresis, 158, 162, 176, 177, 187 imbalance, 193, 208 reference, 8, 12, 15, 18–21, 23, 76, 79, 117, 124, 150, 154, 161, 171, 172, 176, 181, 183, 195, 196, 200 ripple, 57, 60, 130, 177 sharing, 7, 8, 17–19, 51, 116, 118, 119, 175, 181, 188 sharing reference, 41 sharing regulating speed, 18 Current-limiting, 19, 24, 176, 178, 180, 183, 185, 188, 193, 197, 200, 202–204, 207, 208 Current sharing bus, 8 D DC, see Direct current DC–AC inverter, 7, 15, 16, 24, 25, 107, 108, 110, 114, 122, 153, 154 Decoupling relationship, 122, 151 Deﬁcit current, 36 Democratic current sharing, 7–9

© Springer Nature Singapore Pte Ltd. and Science Press, Beijing, China 2019 X. Ruan et al., Control of Series-Parallel Conversion Systems, CPSS Power Electronics Series, https://doi.org/10.1007/978-981-13-2760-5

209

210 Device, 3–5, 123, 132, 193 Digital signal processor, 149 Direct current, 110, 111 Discharging current, 101 Distributed control, 16, 18, 24, 148–151, 154, 169, 171, 172, 175, 180, 183, 188 Drive circuit, 3–5 signal, 15, 127, 195, 200, 203 DSP, see Digital signal processor Duty cycle, 9–13, 40–42, 45–47, 49, 56, 58, 59, 67, 74, 82, 85–87, 95, 191, 197 Dynamic response, 21, 47, 49, 51, 55, 74, 75, 82, 84 E Electric vehicle, 2 Electromagnetic compatibility, 3 Equilibrium point, 19, 37, 114 voltage, 36 Equivalent negative resistance, 37 Equivalent resistance, 79, 194, 197 Excessive current, 36 Extreme load condition, 191–195, 197, 198, 203, 208 F Fault tolerance, 8 Feedback signal, 11, 23, 79, 92, 117, 118, 124, 150, 154, 156, 157, 170 Filter capacitor, 15, 20, 22, 23, 30, 35, 36, 46, 56, 70, 94, 108, 114, 118, 125, 126, 130, 131, 138, 139, 146, 151, 154, 155, 157, 164, 166, 167, 170, 172, 177, 186 Filter inductor, 12, 15, 18, 20, 46, 56–59, 70, 76, 79, 80, 94, 100, 118, 124–126, 131, 132, 138, 139, 141, 142, 145, 150, 151, 154, 156, 157, 161, 164, 166, 167, 170, 172, 176, 177, 183 Freewheeling, 56, 176, 177 Full-bridge, 15, 40, 46, 55, 56, 123, 127, 130–132, 154, 161, 164, 176, 177, 198 Full-controlled device, 2 Full load, 47, 49, 51, 74, 97, 99–103, 164, 166, 197, 203 Full resistive load, 131, 141, 143, 147, 164, 166, 185, 186 Fundamental angular frequency, 126 Fuzzy control, 3

Index G Gallium nitride, 3 Galvanic isolation, 122 GaN, see Gallium nitride Gate turn-off thyristor, 2 Genetic control, 3 Gradient gain, 87, 91, 92, 94, 97 GTO, see Gate turn-off thyristor H Half-bridge, 40, 123 Half-controlled device, 2 Half load, 47, 49, 51, 99, 100, 103 Heavy load, 195 High voltage integrated circuit, 3 HIVC, see High voltage integrated circuit Hot-swap, 102, 175, 183, 187, 188 Hysteresis width, 176 I ICS, see Input current sharing IGBT, see Insulated gate bipolar transistor Inductive load, 141, 143, 144, 164, 165, 168 Inductor-current cross-feedback control, 22, 23 Inductor current feedback hysteretic control, 21 Inner current control loop, 11, 12 Inner output ﬁlter capacitor current loop, 22 Input current perturbation, 78, 80 current sharing, 9, 29, 107, 110, 119, 175 impedance, 3, 37–39, 77–79, 84, 89, 136 power, 31–33, 35, 79, 81, 87, 96, 108, 114, 119, 124, 128, 135, 156, 161, 192, 193 side, 5, 30, 34–37, 46, 48, 50, 61, 78, 114, 115, 122, 133, 153, 164, 197 terminal, 38, 136, 171 voltage, 6, 11, 12, 15, 21–23, 29, 30, 32–34, 36, 38–42, 46–51, 56, 58, 60, 61, 67–69, 74, 79, 81, 82, 85–87, 89, 91, 95, 97–99, 101, 102, 105, 108, 110, 111, 113, 117–119, 121, 123, 124, 128, 131, 136–139, 141, 143, 150, 153, 154, 158, 161, 164, 166, 169, 183, 192–195, 197, 198, 200, 203, 204 voltage perturbations, 38, 80, 97, 133 voltage sharing, 9, 11, 21–23, 29, 107, 121, 153 voltage sharing error, 91, 93 voltage sharing regulator, 21 Input-dividing capacitor voltage, 124

Index Input-parallel-connected systems, 30, 35, 36, 39, 41, 42, 51, 113, 115, 119 Input-parallel-output-parallel, 6, 107, 175 Input-parallel-output-series, 6, 9, 11, 30, 32, 39, 46, 48, 85, 107, 111, 113, 117, 119 Input-series-connected systems, 35, 36, 38–42, 51, 85, 86, 90, 97, 105, 113, 115, 119, 121 Input-series-output-parallel, 6, 107 Input-series-output-series, 6, 13, 22, 24, 30, 33, 34, 36–39, 42, 43, 46, 50, 52, 85, 92, 94, 102, 107, 112, 113, 115, 116, 118, 119, 153, 154, 156, 159, 161, 164–166, 172 Insulated gate bipolar transistor (IGBT), 3 Intelligent power module (IPM), 5 Interconnected control, 16 Interleaved switching, 13 Intrinsic capacitors, 56 IPM, see Intelligent power module IPOP, see Input-parallel-output-parallel IPOS, see Input-parallel-output-series ISOP, see Input-series-output-parallel ISOS, see Input-series-output-series IVS, see Input voltage sharing IVSR, see Input voltages sharing regulator L Light load, 191, 193–195, 197 Linearization, 40 Line-frequency transformer, 122 Load current, 7, 8, 19, 34, 74, 113, 135, 141, 157, 164, 167, 176, 178–180, 183, 185–187 current feed-forward control, 178, 180 regulation, 7 sharing, 100 M Magnetic core, 5, 123 Master–slave current sharing, 7, 8 Mercury arc rectiﬁer, 1 Microelectronic, 2, 4 Minority-carrier device, 3 Modularization architecture, 30, 43, 45, 51, 85, 86 Module efﬁciencies, 32–34, 81 Mutual cross-feedback, 13 N Natural rebalancing ability, 13 Negative resistance, 38, 39, 77, 81, 84, 133, 136 Negative resistance characteristics, 84

211 Neural network, 3 Neural network control, 3 O OCS, see Output current sharing Output characteristics, 19, 24, 175, 178, 180, 183, 184, 188 current feedback control, 126, 127, 129, 138, 141, 145 current sharing, 11, 16, 17, 19–21, 24, 29, 46, 97, 107, 110, 121, 175, 181 voltage, 6, 7, 9–13, 15, 17–24, 29, 30, 32–35, 40–42, 44–47, 49, 51, 55, 65, 69–71, 74, 82, 84–87, 89–92, 94, 97–99, 102, 103, 105, 108–111, 113, 114, 116–119, 122–124, 126–128, 131–133, 136–139, 141, 143, 149, 151, 153–155, 158, 161, 162, 164, 167, 169, 171, 172, 176–181, 183, 187, 196, 198 voltage control loop regulator, 86 voltage/current sharing control, 119 voltage regulator, 8–11, 18, 20–23, 42, 67, 127, 128, 132, 150, 158, 162, 170 voltage sharing, 20, 22–24, 29, 153 Output-voltage cross-feedback control, 23, 25 OVR, see Output voltage regulator OVS, see Output voltage sharing P Parallel control unit, 17 Parasitic resistance, 97 Peak-to-peak voltage, 70 Perturbation, 34–36, 38–42, 58, 78, 80, 81, 87, 95–97, 113, 114, 124, 133, 135 PFC, see Power factor correction PFM, see Pulse-frequency modulation Phase lag, 163 Phase-leg mid-points, 124 Phase-locked loop, 17 Phase-shift control, 123 Phase-shifted full-bridge converter, 123 Phase synchronization, 156, 158, 172 Phasotron Photovoltaic system, 6, 9 PI, see Proportional-integral PLL, see Phase-locked loop Plugged into, 186 Plugged out, 186 Positive feedback loop, 37 Positive gradient regulation characteristic, 87 Positive resistance, 77, 79, 80, 84, 89, 136, 164 Positive zero-crossing point, 181

212 Power balance, 25, 107, 116, 150, 157 Power electronics building block, 4 Power electronics system integration, 3, 4, 25 modularization, 3, 25 standardization, 3, 25 Power factor angles, 22, 109, 155, 172, 179 correction, 154 Power integrated device, 3 Power switch, 11, 15, 56, 94, 121, 123, 176, 186, 198, 200 Primary-and-secondary-windings ratio, 56 Primary current, 56–59, 132 Primary duty cycle, 56, 58, 60 Proportional coefﬁcient, 86 Proportional–integral, 71 Protection circuit, 3–5 PSFB, see Phase-shifted full-bridge converter Pulse-frequency modulation, 3 Pulse-width modulation, 3 Push-pull, 123 PWM comparator, 42 PWM, see Pulse width modulation R Rated output current, 46, 48, 49, 51, 97, 196 Reactive current, 125, 126, 146 Reactive power, 19, 107, 108, 110–115, 118, 126, 139, 153, 157, 165, 167 Real power, 108, 110–115, 118, 171 Rebalancing ability, 14 Rectiﬁer diodes, 46, 56, 94 Redundancy, 7, 19, 100, 102, 105, 149, 175 Regulator, 9–12, 15, 18, 20, 41, 71, 72, 76, 84, 86, 124, 150, 158, 171, 178, 181, 183, 195, 196, 200 Resonant inductance, 46, 198 Right-half-plane pole, 39 Ripple current, 130 Rising edge, 181 RMS, see Root-mean-square Root-mean-square, 109 S Sampled output voltage Sampling clock, 176, 177 factor, 45 frequency, 162 signal, 15 Saturation voltage, 3 Sawtooth carrier, 127, 131

Index signal, 10, 11, 67, 70, 86, 96 SCR, see Silicon controlled rectiﬁer Secondary rectiﬁed voltage, 58 Secondary side, 56 Second stage, 123, 130 Semiconductor devices, 1–4 Series–parallel combined systems, 15 Series–parallel power conversion system, 5–7, 29, 30, 42, 45, 51, 55, 67, 85, 86, 107–109, 113–119 Short-circuit, 193, 197, 202–205, 208 Short fault, 101 Short switch, 101, 102 SiC, see Silicon carbide Silicon carbide, 3 Silicon controlled rectiﬁer, 1 Single-phase full-bridge inverter, 15, 16 Sinusoidal pulse-width modulator, 18 Smart power integrated circuit, 3 Small-signal model, 15, 37–39, 55, 58, 60, 61, 72, 74, 78, 80, 83, 94, 133 Soft-switching, 123 SPIC, see Smart power integrated circuit SPWM, see Sinusoidal pulse-width modulator Static error, 178 Stepped input voltage, 46, 49, 51, 82, 102 Stepped load, 46, 49, 51, 82, 99, 102, 185 Switching frequency, 57, 58, 70, 98, 130, 131, 162, 178, 198 Switching loss, 14, 123 Synchronization of the output current phase, 124 Synchronization pulse signal, 17 Synchronizing bus, 181 T Three control loops, 21, 22, 124, 154, 168 Three-dimensional package technology, 5 Three-loop control strategy, 20, 80, 82, 116 Three-state hysteresis modulation, 158, 176, 177, 183 Three voltage level, 123 Threshold current, 195, 198 Thyratron, 1 Total harmonic distortion, 21 Total load current, 17 Turning frequency, 162 Turns ratio, 9, 40, 61, 70, 95, 131 Two-stage inverter, 123, 131 Two-transistor forward converter, 94, 97 U Unbalanced current, 37 Uncompensated loop gain, 70–72, 162, 163

Index Uninterruptible power supply, 16 Unipolar modulation, 177 UPS, see Uninterruptible power supply V Voltage amplitude, 17, 23, 114 disturbance, 12, 136, 147 diverge, 136 Voltage-controlled device, 2 Voltage–current double closed-loop, 8, 15, 75, 76, 80, 84

213 W Wire connection, 7 Wireless droop control, 16, 19, 20, 175, 181 Wireless IVS control strategy, 15, 86, 89, 92–94, 97, 100 Z Zero-crossing phase-lock, 181 Zero state, 176

*When life gives you a hundred reasons to cry, show life that you have a thousand reasons to smile*

© Copyright 2015 - 2022 AZPDF.TIPS - All rights reserved.