MULTIPLE-INPUT MULTIPLE-OUTPUT CONVERTERS FOR FUTURE LOW-VOLTAGE DC POWER DISTRIBUTION ARCHITECTURES by Yajian Tong A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in The Faculty of Graduate and Postdoctoral Studies (Electrical and Computer Engineering) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) April 2015 © Yajian Tong, 2015 ii ABSTRACT Multiple-input multiple-output (MIMO) converters have been identified as a cost-effective approach for energy harvesting and dispatching in hybrid power systems such as those envisioned in future smart homes and DC microgrids. Compared with relatively complex set-up of single-input single-output (SISO) converters linked at a common DC bus to exchange power, the MIMO converters possess promising features of fewer components, higher power density, and centralized control. This thesis addresses various issues regarding the development of MIMO converters. Both non-isolated and isolated MIMO converter topologies are proposed. Steady-state analysis and dynamic modeling of MIMO non-inverting buck–boost and flyback converters are introduced and presented in detail. Specific switching strategies are proposed and appropriate control algorithms are presented to enable power budgeting between diverse sources and loads in addition to regulating output voltages. Furthermore, a simple method is put forward for deriving the non-isolated MIMO converters with DC-link inductor (DLI) and DC-link capacitor (DLC). Based on a basic structure, a set of rules is listed for the synthesis of MIMO converters. Using the time-sharing concept, multiple sources provide energy in one period, and multiple loads draw energy in the subsequent period. In the end, general techniques are introduced for extending the SISO converters to their MIMO versions, where parts of the conventional SISO converters are replaced with multiport structures. It is envisioned that MIMO converters presented in this thesis will find their acceptance in the future in various applications with DC distribution, which are becoming increasingly accepted by industry. iii PREFACE Some of the research results presented in this thesis have been published in or submitted to several conference proceedings. In all publications, I was responsible for developing the topologies, deriving the mathematical formulations, implementing the models, conducting the simulations and compiling the results, as well as writing the drafts of manuscripts. My research advisor, Dr. J. Jatskevich, provided the overall supervisory comments and editing during the process of conducting the research and writing the manuscripts. The contributions of other co-authors are explained below as applies for each manuscript: A version of chapter 2 has been published. Y. Tong and J. Jatskevich, “A transformerless multiple-port DC-DC converter for energy harvesting and dispatching” in Proceedings of IEEE 15th Workshop on Control and Modeling for Power Electronics, Santander, Spain, June 22–25, 2014, pp. 1–9. A version of chapter 3 has been published. Y. Tong, Z. Shan, J. Jatskevich, and C. K. Tse, “A flyback converter with multiple ports for power management in DC distribution systems” in Proceedings of IEEE International Power Electronics and Applications Conference and Exposition, Shanghai, China, Nov. 5–8, 2014, pp. 1531–1536. Dr. Shan and Dr. Tse provided useful discussions, then revised and proofread the manuscript. A part of chapter 4 has been published. Y. Tong and J. Jatskevich, “A methodology to derive single-stage multiple-input multiple-output DC–DC converters,” in Proceedings of IEEE 36th International Telecommunication Energy Conference, Vancouver, Canada, Sept. 28–Oct. 2, 2014, pp. 1–7. Another part of chapter 4 has been published. Y. Tong, Z. Shan, J. Jatskevich, and A. Davoudi, “A nonisolated multiple-input multiple-output DC–DC converter for DC distribution of future energy efficient homes,” in Proceedings of the 40th Annual Conference of the IEEE Industrial Electronics Society, Dallas, TX, USA, Oct. 29–Nov. 1, 2014, pp. 4126–4132. Dr. Shan provided comments and corrections. Dr. Davoudi provided very useful feedback and revised the manuscript. iv A part of chapter 5 has been published. Y. Tong, J. Jatskevich, and A. Davoudi, “Topology design of isolated multiport converters for smart DC distribution systems,” in Proceedings of the 30th Annual IEEE Applied Power Electronics Conference and Exposition, Charlotte, NC, USA, Mar. 15–19, 2015, pp. 2678–2683. Dr. Davoudi provided useful discussions of results, comments, and helped to revise the manuscript. A part of chapter 5 has been submitted for peer review. Y. Tong, Z. Shan, N. M. Ho, and J. Jatskevich, “Concept of synthesizing modular power supply for interfacing diverse energy sources and loads.” Dr. Shan proofread the manuscript. Dr. Ho provided comments, suggestions, and constructive feedback. v TABLE OF CONTENTS ABSTRACT .................................................................................................................................... ii PREFACE ...................................................................................................................................... iii TABLE OF CONTENTS ................................................................................................................. v LIST OF TABLES ....................................................................................................................... viii LIST OF FIGURES ......................................................................................................................... ix LIST OF ABBREVIATIONS ....................................................................................................... xii ACKNOWLEDGEMENTS ......................................................................................................... xiii CHAPTER 1: INTRODUCTION ..................................................................................................... 1 1.1 Motivation ....................................................................................................................... 1 1.2 Literature Review ............................................................................................................ 3 1.2.1 MISO Converters ................................................................................................... 4 1.2.2 SIMO Converters ................................................................................................... 5 1.3 Research Objectives of This Thesis ................................................................................ 5 1.4 Composition of the Thesis .............................................................................................. 6 CHAPTER 2: MIMO BUCK–BOOST CONVERTER WITH INDEPENDENT OUTPUTS ........ 8 2.1 Circuit Configuration and Operation Principle ............................................................... 8 2.2 Static Characteristics ....................................................................................................... 9 2.2.1 Operational Analysis of the Circuit ........................................................................ 9 2.2.2 Steady-State Analysis ............................................................................................. 9 2.2.2.1 CCM ............................................................................................................. 10 2.2.2.2 DCM ............................................................................................................. 13 2.3 Average-Value Modeling .............................................................................................. 15 2.4 Power Management and Control Strategy .................................................................... 17 2.4.1 Power Budgeting .................................................................................................. 17 2.4.2 Controller Design ................................................................................................. 17 2.5 Computer Studies .......................................................................................................... 19 2.5.1 System Response to Input Voltage Variation ....................................................... 19 vi 2.5.2 System Response to Output Load Variation ........................................................ 19 2.5.3 System Response to Input Power Variation ......................................................... 20 CHAPTER 3: MIMO FLYBACK CONVERTER WITH STACKED CAPACITORS ................ 24 3.1 Circuit Configuration and Operation Principle ............................................................. 24 3.2 Static Characteristics ..................................................................................................... 25 3.2.1 Operational Analysis of the Circuit ...................................................................... 25 3.2.2 Steady-State Analysis ........................................................................................... 26 3.2.2.1 CCM ............................................................................................................. 26 3.2.2.2 DCM ............................................................................................................. 27 3.3 Average-Value Modeling .............................................................................................. 30 3.4 Power Management and Control Strategy .................................................................... 31 3.5 Computer Studies .......................................................................................................... 33 3.5.1 System Response to Input Voltage Variation ....................................................... 33 3.5.2 System Response to Output Load Variation ........................................................ 34 3.5.3 System Response to Input Power Variation ......................................................... 34 CHAPTER 4: DERIVATION OF MIMO CONVERTERS BASED ON DC LINK ..................... 38 4.1 Basic Ideas for Deriving DLI- or DLC-Coupled MIMO Converters ............................ 38 4.2 Non-Isolated PSCs and Their Connection Rules .......................................................... 38 4.2.1 Basic PSCs ........................................................................................................... 38 4.2.2 Hybrid PSCs ......................................................................................................... 40 4.2.3 Connection Rules of PSCs ................................................................................... 41 4.2.3.1 Connection Rules of PVSCs ........................................................................ 41 4.2.3.2 Connection Rules of PCSCs ......................................................................... 41 4.3 Central Energy Buffer Element—DLI and DLC .......................................................... 42 4.3.1 DLI Cell ................................................................................................................ 42 4.3.2 DLC Cell .............................................................................................................. 42 4.4 FCs and Their Connection Rules .................................................................................. 43 4.4.1 C-FCs and Their Connection Rules ...................................................................... 43 4.4.2 LC-FCs and Their Connection Rules ................................................................... 44 4.5 Synthesis of Non-Isolated MIMO Converters .............................................................. 45 4.5.1 DLI-Coupled MIMO Converters .......................................................................... 45 4.5.2 DLC-Coupled MIMO Converters ........................................................................ 46 4.6 Zeta-Derived DLI-Coupled MIMO Converter .............................................................. 48 vii 4.6.1 State-Space Averaging ......................................................................................... 49 4.6.2 Design Considerations .......................................................................................... 51 4.6.3 Power Flow Management ..................................................................................... 52 4.6.4 Control Scheme .................................................................................................... 52 4.6.5 Case Studies ......................................................................................................... 54 CHAPTER 5: EXTENSION OF SISO CONVERTERS TO THEIR MIMO VERSIONS ............ 59 5.1 Isolated PSCs and Their Connection Rules .................................................................. 59 5.2 Basic Configuration of Conventional SISO Converters ............................................... 59 5.3 Realization of Multiport Structure ................................................................................ 61 5.3.1 Multiple-Input Structure ....................................................................................... 61 5.3.2 Multiple-Output Structure .................................................................................... 61 5.4 Synthesis of MIMO Converters .................................................................................... 62 5.5 Transformer-Coupled MIMO Converters ..................................................................... 66 CHAPTER 6: CONCLUSIONS ..................................................................................................... 70 6.1 Contributions of the Thesis ........................................................................................... 70 6.2 Future Work .................................................................................................................. 70 6.2.1 Practical Implementation ...................................................................................... 70 6.2.2 Controller Optimization ....................................................................................... 71 6.2.3 Non-Ideal MIMO Converters ............................................................................... 71 6.2.4 Bidirectional Multiport Converters ...................................................................... 71 REFERENCES ............................................................................................................................... 72 viii LIST OF TABLES 5.1 Eligible blocks that can be replaced with multiport structure.............................................. 62 ix LIST OF FIGURES 1.1 Example of residential low-voltage DC power distribution system. ...................................... 2 1.2 Conventional DC distribution system with multiple SISO converters. .................................. 3 1.3 Proposed DC distribution system with MIMO converter. ...................................................... 4 2.1 MIMO non-inverting buck–boost converter. .......................................................................... 9 2.2 Example switching strategy of MIMO non-inverting buck–boost converter. ...................... 10 2.3 Two operating stages of MIMO non-inverting buck–boost converter. ................................ 11 2.4 Discontinuous inductor current example. ............................................................................. 14 2.5 Control block diagram of the MIMO non-inverting buck–boost converter. ........................ 17 2.6 Simulated system response of the MIMO non-inverting buck–boost converter to input voltage change. ............................................................................................................ 21 2.7 Simulated system response of the MIMO non-inverting buck–boost converter to load change. .................................................................................................................................. 22 2.8 Simulated system response of the MIMO non-inverting buck–boost converter to input power change. .............................................................................................................. 23 3.1 Example of MIMO flyback converter with a separate winding for each port. ..................... 25 3.2 Proposed MIMO flyback converter that minimizes the number of windings. ..................... 26 3.3 Gating signals of the MOSFETs for the proposed MIMO flyback converter. ..................... 27 3.4 Equivalent circuits of the MIMO flyback converter at two operating stages. ...................... 28 3.5 Block diagram of the proposed control structure for the MIMO flyback converter. ............ 33 3.6 Simulated system response of the MIMO flyback converter to input voltage variation. ..... 35 3.7 Simulated system response of the MIMO flyback converter to load variation. ................... 36 3.8 Simulated system response of the MIMO flyback converter to input power variation. ....... 37 4.1 Basic structure of MIMO converters based on DLI or DLC. ............................................... 39 4.2 Non-isolated basic PVSCs: (a) buck; (b) Ćuk; and (c) Zeta. ................................................ 39 4.3 Non-isolated basic PCSCs: (a) boost; (b) buck-boost; and (c) SEPIC. ................................ 39 4.4 Non-isolated hybrid PVSCs: (a) buck-Zeta; (b) Ćuk-Zeta; and (c) Zeta-Zeta. .................... 40 4.5 Non-isolated hybrid PVSCs: (a) boost-Ćuk; (b) buck–boost-Zeta; and (c) SEPIC-Ćuk. ...................................................................................................................................... 40 4.6 Non-isolated hybrid PCSCs: (a) buck-buck–boost; (b) Ćuk-Buck–boost; and (c) Zeta-buck–boost. ........................................................................................................................... 41 x 4.7 Non-isolated hybrid PCSCs: (a) boost-SEPIC; (b) buck-SEPIC; and (c) SEPIC-SEPIC. .................................................................................................................................. 41 4.8 Combination of PSCs: (a) PVSCs in series; (b) PVSCs in parallel; (c) PCSCs in series; and (d) PCSCs in parallel. ......................................................................................... 42 4.9 DC-link configurations: (a) DLI cell; and (b) DLC cell. ...................................................... 42 4.10 FCs: (a) C-FC; and (b) LC-FC.............................................................................................. 43 4.11 Combination of FCs: (a) C-FCs in series; (b) C-FCs in parallel; (c) LC-FCs in series; and (d) LC-FCs in parallel. ................................................................................................... 43 4.12 Outputs taken from C-FCs: (a) series-connected C-FCs and series outputs; (b) series-connected C-FCs and parallel outputs; and (c) parallel-connected C-FCs and parallel outputs. ................................................................................................................................. 44 4.13 Outputs taken from LC-FCs: (a) series-connected LC-FCs and outputs; and (b) parallel-connected LC-FCs and outputs. .............................................................................. 44 4.14 Buck-derived DLI-coupled MIMO Converters: (a) buck PVSCs in series and C-FCs in series with series outputs; (b) buck PVSCs in parallel and C-FCs in series with series outputs; (c) buck PVSCs in series and C-FCs in series with parallel outputs; (d) buck PVSCs in parallel and C-FCs in series with parallel outputs; (e) buck PVSCs in series and C-FCs in parallel with parallel outputs; and (f) buck PVSCs in parallel and C-FCs in parallel with parallel outputs. ................................................................................ 46 4.15 Boost-derived DLC-coupled MIMO Converters: (a) boost PCSCs in series and LC-FCs in series; (b) boost PCSCs in parallel and LC-FCs in series; (c) boost PCSCs in series and LC-FCs in parallel; and (d) boost PCSCs in parallel and C-FCs in parallel. ....... 47 4.16 Zeta-derived DLI-coupled MIMO converter example. ........................................................ 48 4.17 Switching pattern for the Zeta-derived DLI-coupled MIMO converter. .............................. 48 4.18 Block diagram of the closed-loop system. ............................................................................ 53 4.19 Simulated waveforms of the Zeta-derived DLI-coupled MIMO converter in response to step change in input voltage. ............................................................................................ 56 4.20 Simulated waveforms of the Zeta-derived DLI-coupled MIMO converter in response to step change in load. .......................................................................................................... 57 4.21 Simulated waveforms of the Zeta-derived DLI-coupled MIMO converter in response to one source missing. .......................................................................................................... 58 5.1 Isolated PVSCs: (a) full-bridge isolated buck; (b) push–pull isolated buck; and (c) forward.................................................................................................................................. 60 xi 5.2 Isolated PCSCs: (a) full-bridge isolated boost; (b) push–pull isolated boost; and (c) flyback. ................................................................................................................................. 60 5.3 General configuration of a SISO converter. ......................................................................... 61 5.4 Basic configuration of non-isolated SISO converter. ........................................................... 61 5.5 Basic configuration of isolated SISO converter. .................................................................. 61 5.6 Circuit configurations: (a) SISO buck converter; and (b) PVSC-source MIMO converter generated by several buck PVSC in series and LC-FCs in series. ........................ 63 5.7 Circuit configurations: (a) SISO flyback converter; and (b) PCSC-source MIMO converter generated by several flyback PCSCs in parallel and C-FCs in parallel. ............... 63 5.8 Circuit configurations: (a) SISO Zeta converter; and (b) MIMO Zeta converter with parallel-connected buck PVSCs as inputs and parallel-connected LC-FCs as outputs. ....... 64 5.9 Circuit configurations: (a) SISO push–pull isolated boost converter; and (b) MIMO push–pull isolated boost converter with parallel-connected boost PCSCs as inputs and series-connected C-FCs as outputs. ............................................................................... 65 5.10 Circuit configurations: (a) SISO forward converter; and (b) PVSC-source MIMO converter generated by several forward PVSCs in series and C-FCs in series. .................... 66 5.11 Circuit configurations: (a) SISO full-bridge isolated buck converter; and (b) MIMO full-bridge isolated buck converter with series-connected buck PVSCs as inputs and parallel-connected C-FCs as outputs. ................................................................................... 67 5.12 Basic structure of SPSSWTC MIMO converter. .................................................................. 68 5.13 Basic structure of SPMSWTC MIMO converter. ................................................................. 68 5.14 Basic structure of MPSSWTC MIMO converter. ................................................................. 68 5.15 Basic structure of MPMSWTC MIMO converter. ............................................................... 68 xii LIST OF ABBREVIATIONS AVM Average-Value Model CCM Continuous Conduction Mode C-FC Capacitor Filter Cell DCM Discontinuous Conduction Mode DLC DC-Link Capacitor DLI DC-Link Inductor DM Detailed Model LC-FC Inductor-Capacitor Filter Cell MIMO Multiple-Input Multiple-Output MISO Multiple-Input Single-Output MPMSWTC Multiple-Primary-Multiple-Secondary-Winding-Transformer-Coupled MPSSWTC Multiple-Primary-Single-Secondary-Winding-Transformer-Coupled FC Filter Cell PCSC Pulsating Current Source Cell PSC Pulsating Source Cell PVSC Pulsating Voltage Source Cell PWM Pulse-Width Modulation SIMO Single-Input Multiple-Output SISO Single-Input Single-Output SPMSWTC Single-Primary-Multiple-Secondary-Winding-Transformer-Coupled SPSSWTC Single-Primary-Single-Secondary-Winding-Transformer-Coupled xiii ACKNOWLEDGEMENTS I offer my enduring gratitude to my advisor, Dr. J. Jatskevich, for his invaluable guidance and strong support. He has provided me immense help and excellent advices. His kindness will always be remembered. The financial support for this research was provided through the Natural Science and Engineering Research Council (NSERC) of Canada, Collaborative Research and Development Grant entitled “People and Planet Friendly Home” led by Dr. P. Nasiopoulos, and the Institute for Computing, Information and Cognitive Systems (ICICS) of the University of British Columbia. I am also grateful to Dr. J. R. Martí and Dr. Y. Chen for dedicating their valuable time to serve on my examining committee, and for all their useful comments and feedback that have helped to improve the quality of this thesis. Many special thanks go to Dr. Z. Shan, Dr. A. Davoudi, Dr. N. M. Ho, and Dr. C. K. Tse for their helpful discussions and support. I would also like to thank all my peers in the Electric Power and Energy Systems Group for their technical support and wonderful friendship: Mr. H. Chang, Dr. M. Chapariha, Mr. C. L. Chaw, Mr. L. Dong, Mr. Q. Han, Mr. Y. Huang, Mr. Y. Lei, Mr. M. Liu, Ms. S. Ren, Mr. J. C. Shen, Mr. F. Therrien, Mr. Z. Wang, Ms. T. Xu, Mr. Y. Xu, Mr. B. Zhang, and Mr. K. Zhang. In addition, I would like to extend my gratitude to the many friends and people who cared: Dr. C. He, Ms. N. Rong, Ms. X. Sun, Ms. Y. Yang, and Mr. C. Zhu. I am also grateful to everyone who has ever helped me. Although I am not able to mention all your names here, I want you to know this thesis would not have been possible without your help. Finally, warmest thanks go to my parents and my brother, for their understanding, unconditional support, and endless love. 1 CHAPTER 1: INTRODUCTION This chapter provides brief discussions of multiport DC–DC converters, which are envisioned as key components that interact with other elements such as distributed generation and modern electronic loads in future residential buildings and commercial facilities. Also included in this chapter is a description of thesis objective and structure. 1.1 Motivation Utilization of renewable energy sources on user premises has attracted a significant interest for many commercial and industrial applications owing to their merits of non-pollution and rich reserves. Due to the intermittent nature of renewable energy, storage and standby sources are usually required to function as backup. A hybrid power system may lower environmental impacts and improve security of supply. Besides, a simultaneous combination of sources is available for optimal energy/economic dispatch. At the same time, many loads and appliances used in offices, commercial facilities, and residential buildings often dictate the need of power supply with different gains. Thus, the need of technology for distributing power to a variety of consumption loads whose voltage levels are different motivates the development of supply structure with multiple voltages. Many distributed energy resources include but are not limited to solar panels and fuel cells generate DC voltages, and a growing number of consumption loads and appliances are using DC, e.g. data centers, portable devices, LED lights, etc. Thus, DC distribution systems are envisioned to interact with different energy sources, modern electronic loads and storage units for simplicity and efficiency [1]–[3]. Figure 1.1 shows a conceptual DC power distribution architecture for future residential applications where wind and solar energy are interfaced, storage devices are installed, and different loads are powered. Usually, there are two approaches to form such a system with multiple ports. Conventionally, single-input single-output (SISO) converters are arranged in parallel at a common DC bus to exchange power (Figure 1.2). In this architecture, separate conversion stages are employed for individual sources and loads, and the converters would be controlled independently. Thus, a communication system may be included to exchange information and manage the power flow between different ports. Although such a configuration is prevalent in distribution systems today, complex configurations generally result in a large number of modules and high costs. In addition, the communication-based control system may cause software delays and data errors, which would also degrade the performance of the system [4]. As 2 a prominent alternative, multiple-input multiple-output (MIMO) converter can replace the complicated set-up, as pictured in Figure 1.3. In this integrated and single-stage conversion architecture, voltage regulation and power management can be carried out simultaneously. Additionally, compact packaging and relatively straightforward control become possible. Though less attention has been given to the development of MIMO converters so far [5]–[8], multiple-input single-output (MISO) converters and single-input multiple-output (SIMO) converters have been well studied. In the established literature, MISO converters are identified as a cost-effective and modular technology to incorporate more than one source. Applications of using MISO converters for integrating sources with complementary nature have been found in photovoltaic-utility systems [9]–[12], photovoltaic-wind systems [13]–[16], renewable generation systems with battery backup [17]–[22], and hybrid electric vehicles [23]–[30]. Meanwhile, SIMO converters are seen as an efficient power router to feed several loads. Many SIMO converters have been reported in the literature for various applications, such as portable and electronic devices [31]–[38], telecom and computer systems [39]–[41], fuel cell generation systems [42], diode-clamped multilevel inverters [43]–[45], and others [46], [47]. The motivation of this thesis is to design MIMO converters, which can combine the advantages of MISO and SIMO converters. The proposed MIMO converters can be a substitute of the conventional architecture consisting of SISO converters to simplify the conversion Figure 1.1. Example of residential low-voltage DC power distribution system. 3°CACDC3 structure and provide easy energy management. The reduced conversion stage may also improve the power density and eliminate complicated conventional communication-based control issues between individual conversion stages. With proper design, flexibility of source integration and power dispatching may be enhanced for the distribution systems, while savings in manufacturing cost and mass become achievable. Although the MIMO converter topologies discussed in this thesis may only allow for unidirectional inputs, bidirectional power flow can be done extrinsically by using additional converters (which is outside the scope of this thesis). 1.2 Literature Review In general, the multiport converters, including MISO, SIMO, and MIMO converters, can be classified into two categories: non-isolated topologies and isolated topologies. The non-isolated topologies possess advantages of compact structure, low cost, high power density and straightforward power flow control. A drawback of this kind of structure lies in the fact that a wild range of voltage transformation is not easily obtainable and galvanic isolation is not provided (even when it may be a requirement). Non-isolated MISO converters can be derived from a single converter, i.e. buck [48], boost [49], buck–boost [50]–[52], Ćuk [53] and SEPIC [54], [55], or a combination of several converters, such as buck/buck–boost [15], buck/SEPIC [56]. Non-isolated SIMO converters can be obtained by connecting multiple outputs in an independent manner [57], or piling them up [44]. Figure 1.2. Conventional DC distribution system with multiple SISO converters. ================Washing Machine Refrigerator LED LightingSolar Panels Wind Turbine Li-on BatteriesElectric Vehicle4 Isolated topologies include a transformer and adjustment of voltage levels through changing the transformer turns ratio could be beneficial to avoid the device handling high voltage and current. However, complex circuitry and control strategy may mitigate the converters’ performance. Isolated MISO converters can be derived from flyback [9], [10], and bridge [58]–[62] converters. Isolated SIMO converters may be generated from flyback [31], [32], [63], forward [39], push-pull [64], and bridge [40] topologies. Moreover, derivations of isolated SIMO converters can be based on a combination of flyback and forward converters [65], [66]. 1.2.1 MISO Converters Multiple input sub-circuits can be placed in parallel or series. In parallel configurations, MISO converters can be obtained by adding primary windings and primary side sub-circuits of a conventional flyback converter [9], [10], or paralleling the input sub-circuits of a buck–boost converter [12], [50]. As these converters are designed in a time shared operation mode, only one source is allowed to deliver power at a time. This limitation can be overcome by using current-source converters. For example, [58] presented a full-bridge isolated boost converter, but the number of required switches is four times the input ports. Reference [67] proposed a half-bridge isolate boost topology. Though the number of switches is reduced by half, the need of inductors is doubled. Therefore, it still makes the design inherently complex and costly. Other alternatives include connecting the multiple input sub-circuits at trivial points, such as linking the SISO converters by paralleling them at the output capacitor [24], [26]. Moreover, connection of Figure 1.3. Proposed DC distribution system with MIMO converter. Electric Vehicle Washing Machine Refrigerator LED LightingMultiple-Input Multiple-Output ConverterSolar Panels Wind Turbine Li-on Batteries5 multiple input sub-circuits may occur in a series way [13], [14] to achieve simultaneous power transfer. The authors of [68] and [69] also proposed a method in which the MISO converters are accomplished by means of series-connected H-bridge cells. 1.2.2 SIMO Converters SIMO converters can be obtained by placing sub-circuits in parallel or series at the output side. In parallel configurations, a straightforward method to provide multiple outputs is to use a transformer with multiple secondary windings. Based on this approach, two topologies are commonly used due to simplicity and effectiveness, i.e. SIMO flyback and forward converters. The SISO flyback converter topologies have fewer components, but face a cross-regulation problem [70]. In order to keep all the output voltages tightly regulated, various approaches have been proposed [71], [72]. Likely, SIMO forward converters also have a drawback of poor regulation, and several methods are readily taken to improve the converters’ performance [73]–[75]. Instead of using a transformer, the authors of [34] came up with a method to realize a controlled current source and distribute the current to the outputs on an interleaving basis. Authors of [57] introduced a class of topologies where only one inductor was implemented while several outputs were regulated. The inductor is sequentially connected in a parallel output arrangement with a number of loads via a switch-network. Some of the proposed topologies are also capable of producing bipolar output voltages. In series configurations, the authors of [44] and [46] presented a SIMO converter topology for applications in feeding multilevel inverters. 1.3 Research Objectives of This Thesis The aim of this research project is to explore feasible power electronic converters for energy harvesting and dispatching, which could be applied for the development of future low-voltage DC distribution systems. The work presented in this thesis is mostly theoretical and it is supported using the computer simulations wherever it is necessary. In view of the current situation in DC microgrids, characteristics of distributed energy resources and modern electronic loads, and the prospect in applications of multiport DC–DC converters, the proposed converters should be able to accommodate variable energy sources, and provide multiple outputs at different voltage levels. Specifically, this work shall include: 1) Topology design and detailed analysis of MIMO converters with/without a transformer. If there is no requirement of galvanic isolation, compact structure and manufacturing cost are the priority to be considered. Thus, topologies without incorporating transformers would be 6 preferred. Otherwise, isolated topologies should be implemented. Therefore, at least one non-isolated and one isolated topologies shall be proposed and demonstrated. 2) Coordination and management of generation units connected to the converter, and tight regulation of output voltages at different levels. Power flow control strategy needs to be formulated, and the controller design and tuning for multivariable systems shall be investigated. 3) Simulation and dynamic modeling of MIMO converters. Detailed models (DMs) shall be established to verify the converters’ feasibility. Analytical state-space averaging method and numerically-constructed average-value models (AVMs) shall be derived for studying the dynamic behavior of the new proposed converters in system-level analysis and transients. 4) Exploration and derivation of various feasible multiport DC power converter topologies. As each converter has its own advantages and disadvantages, a systematic approach for constructing MIMO converters for various applications and specifications is of significant value for the future designers. 1.4 Composition of the Thesis This thesis is organized in four parts. The first part comprises Chapter 1, which gives a literature review on the previous and current developments of multiport DC–DC converters. The motivation of pursuing research in this direction and objectives of this thesis are described. The second part of this thesis, consisting of Chapter 2 and 3, outlines the concepts, features, operating principles, control methods, and modeling of two typical MIMO converters. Specifically, Chapter 2 presents a MIMO non-inverting buck–boost converter, where multiple sources can supply power either individually or simultaneously. Also, positively referenced output voltages are obtained without using any transformer, and sources with equal and unequal voltages can be accommodated. Chapter 3 presents a MIMO flyback converter, which offers galvanic isolation and precise voltage regulation. This topology is flexible and configurable to provide double-polarity outputs and isolation, where it may be required. The third part of this thesis, comprising of Chapter 4 and 5, introduces some approaches for derivation of MIMO converters. Specifically, Chapter 4 presents the concept of pulsating source cells (PSCs) [76] and filter cell (FCs) [77], which are used for interfacing sources and loads, respectively. Various forms of PSCs and FCs are discussed. Derivation principles for non-isolated topologies based on DC-link inductor (DLI) and DC-link capacitor (DLC) are presented. A unified operation principle can be applied on these newly designed converters. The pursuit of other approaches for synthesizing general MIMO converters is continued in Chapter 5, which presents a set of rules to construct MIMO converters from existing basic SISO converters. 7 Finally, the forth part of this thesis, Chapter 6, gives conclusions by summarizing the major findings and contributions of this work. The various areas for future work are highlighted. 8 CHAPTER 2: MIMO BUCK–BOOST CONVERTER WITH INDEPENDENT OUTPUTS This chapter presents a non-isolated MIMO converter derived from non-inverting buck–boost converter. The operation principle, pulse-width modulation, feed-back control strategy, and dynamic modeling are presented and analyzed in detail. The AVM is demonstrated and compared with DM. 2.1 Circuit Configuration and Operation Principle The schematic diagram of the proposed MIMO non-inverting buck–boost converter is depicted in Figure 2.1. The input voltages are Vin,i (i = 1,…, m) and the output voltages are Vout, j ( j = 1,…, n). All the input switches Sin,i (i = 1,…, m) are bidirectional-carrying forward-blocking, and all the output switches Sout, j ( j = 2,…, n) are forward-conducting bidirectional-blocking, except for Sout,1. The bidirectional-carrying forward-blocking switch is realized by a MOSFET, and the forward-conducting bidirectional-blocking switch is realized by a series MOSFET and diode pair. The inputs can be arbitrarily ordered. For simplicity of analysis, the inputs are arranged in an descending order of duty ratios; that is Din,1 > Din,2 > ··· > Din,m. The outputs are assumed to be regulated such that Vout,1 > Vout,2 > ··· > Vout,n. All the MOSFETs operate at the same switching frequency, and the gating signals are depicted in Figure 2.2. The trailing edges of the input and output MOSFETs’ gating signals are synchronized, respectively. Switch S is on whenever any input switch Sin,i (i = 1,…, or m) is on. It is off only when all the input switches are off. The concept of the overlapping duty ratio Din, olp, i (i = 1,…, or m) of the input switch is defined as the portion of time when there are i inputs supplying power at a time , , 1, ,,, 1,..., 1, .in i in iin olp iin iD D i mDD i m+− = −= = (2.1) If two or more output switches conduct at a time, only the one connected to the lowest-voltage output is on, and the others are off. The concept of the effective duty ratio Dout,eff , j ( j = 1,…, or n) of the output switch is defined as the portion of time when the jth output switch Sout, j carries nonzero current , , 1, ,, , 1 , , 10, 1,..., 1, out j out jout eff jout j out j out j out jD DD j nD D D D++ +≤= = −− > (2.2) where Dout,1 = 1 Din,1 and Dout, eff, n = Dout, n. 9 2.2 Static Characteristics 2.2.1 Operational Analysis of the Circuit The basic principle of the MIMO non-inverting buck–boost converter is to charge the inductor L from Vin,i (i = 1,…, or m) or their combination in one period, and discharge it to the output capacitors C j and loads R j ( j = 1,…, n) in the subsequent period of a switching cycle Ts. Thus, the converter exhibits two operating stages depending on the state of the inductor (refer to Figure 2.3). Stage 1: The inductor is in charge-state. Switch S is on, at least one input switch Sin,i (i = 1,…, or m) is on, and all the output switches Sout, j ( j = 1,…, n) are off. The inductor L is energized, and power demands for the loads R j ( j = 1,…, n) are satisfied by discharging the output capacitors C j ( j = 1,…, n). There are m subintervals in this stage. If k input switches Sin,i (i = 1,…, k) are on, power is delivered from the k inputs Vin,i (i = 1,…, k) simultaneously, and the inductor voltage VL is Vin,1 + Vin,2 + ··· + Vin,k. Stage 2: The inductor is in discharge-state. Switch S is off, and all the input switches Sin, i (i = 1,…, m) are off. The inductor L is discharged. If several output switches are on at a time, the inductor voltage is equal to the lowest of the output voltages for which respective switch is on. In this stage, there are n subintervals. The energy storage in inductor L is released to the output capacitors C j and loads Rj ( j = 1,…, n) in a sequential manner. 2.2.2 Steady-State Analysis In the following derivations, the converter is assumed to be lossless. The lower-case variables represent the large-signal states, upper-case variables represent the equilibrium points, and the hatted variables denote the small-signal perturbations. Both continuous conduction mode (CCM) and discontinuous conduction mode (DCM) are analyzed. Figure 2.1. MIMO non-inverting buck–boost converter. S+−+−+−+−+−LVin, 1Vin, m S in, mS in, 1S out, 1 S out, 2 S out, nC1 C2 CnR1 R2 RnVout , 1 Vout , 2 Vout , n10 2.2.2.1 CCM If the inductor carries nonzero current in steady state, CCM results. Assuming the output capacitors are sufficiently large, averaging the inductor voltage over one switching cycle based on the volt-second balance principle yields , , , , , ,1 1 10.m i nL in olp i in k out eff j out ji k jV D V D V= = == − =∑ ∑ ∑ (2.3) Applying the amp-second balance theorem on the output capacitors C j ( j = 1,…, n), the average capacitor currents over one switching cycle are zero in steady state ,, ,0, 1,..., .jout jC out eff j LjVI D I j nR= − = = (2.4) The duty ratios usually vary with time, but for a given set of values one obtains , , ,1 12, ,1m iin olp i in ki kL nout eff j jjD VID R= ===∑∑∑ (2.5) , , ,1 1, , ,2, ,1, 1,..., .m iin olp i in ki kout j j out eff j nout eff j jjD VV R D j nD R= === =∑∑∑ (2.6) Figure 2.2. Example switching strategy of MIMO non-inverting buck–boost converter. Sin,1Sin,2Sin,mSout,nSout,2Dout,eff,1TsTsSDin,olp,1Ts Din,olp,mTs Dout,eff,nTs11 Figure 2.3. Two operating stages of MIMO non-inverting buck–boost converter. S+−+−+−+−+−+−+−+−LStage 1Stage 2Vin, 1Vin, m S in, mS in, 1+−S out, 1C1 R1Vout , 1+−S out, nCn RnVout , nSLVin, 1Vin, m S in, mS in, 1+−S out, 1C1 R1Vout , 1+−S out, nCn RnVout , nSLVin, 1Vin, m S in, mS in, 1+−S out, 1C1 R1Vout , 1+−Cn RnVout , nSLVin, 1Vin, m S in, mS in, 1+−S out, 1C1 R1Vout , 1+−S out, nCn RnVout , nS out, n12 Provided the inductor current is piecewise-linear, the change in inductor current during its ith charge subintervals Din,olp,iTs and its jth discharge subintervals Dout,eff, jTs are , ,, ,1, 1,...,iin olp i sL i in kkD Ti V i mL+=∆ = =∑ (2.7) , ,, ,, 1,..., .out eff j sL j out jD Ti V j nL−∆ = = (2.8) The total peak-to-peak inductor current ripple is the sum of (2.7) in the positive direction or (2.8) in the negative direction , , , ,1 1 1m m isL L i in olp i in ki i kTi i D VL+= = =∆ = ∆ =∑ ∑∑ (2.9) , , , ,1 1.n nsL L j out eff j out jj jTi i D VL−= =∆ = ∆ =∑ ∑ (2.10) There are several forms in which the peak inductor current IL, peak can be expressed. On simplification, it can be estimated by ,.2LL peak LiI I ∆= + (2.11) Alternatively, identifying the area under the iL curve yields ( ) ( ), ,max , ,min, ,max , ,min, , , ,1 12 2m nL j L jL i L iL s in olp i s out eff j si jI II II T D T D T− −+ += =++= +∑ ∑ (2.12) where IL+,i,max, IL+,i , min, IL , j, max and IL , j,min represent the maximum and minimum values of the inductor current during its respective charge and discharge subintervals , ,1, ,max,, 1,..., 1, mL peak L kk iL iL peakI i i mII i m+= ++− ∆ = −= =∑ (2.13) , ,min , , , 1,...,mL i L peak L kk iI I i i m+ +== − ∆ =∑ (2.14) ,1, ,max, ,1, 1, 2,...,L peakjL jL peak L kkI jII i j n−−−=== − ∆ = ∑ (2.15) , ,min , ,1, 1,..., .jL j L peak L kkI I i j n− −== − ∆ =∑ (2.16) 13 Substitute (2.13)–(2.16) into (2.12) and solving 1, , ,1 ,1 , , , , ,2 11, , , , , , , ,1 11 1( )2 21 1( ) .2 2m iL peak L in olp L in olp k in olp i L ii kn nout eff k out eff j L j out eff n L nj k jI I D i D D iD D i D i−+ += =−− −= = += + ∆ + + ∆ ++ ∆ + ∆∑ ∑∑ ∑ (2.17) Expressions of IL+,i,max, IL+,i,min, IL , j,max and IL , j,min can be found by substituting (2.7)–(2.8) and (2.17) into (2.13)–(2.16). Also, the average inductor currents over each charge and discharge subintervals satisfy , ,max , ,min,, 1,...,2L i L iL iI II i m+ +++= = (2.18) , ,max , ,min,, 1,...,2L j L jL jI II j n− −−+= = (2.19) resulting in different form of the output voltage equations , , , ,, 1,..., .out j j out eff j L jV R D I j n−= = (2.20) Again, to approximate the peak-to-peak voltage ripples on the output capacitors, the time constants are assumed to be relatively large compared to the switching cycle Ts. The discharge of the capacitor Cj occurs when the related output switch Sout, j is off. The linear-ripple approximation leads to , , ,,(1 ), 1,2,..., .out eff j s out jC jj jD T Vv j nC R−∆ = = (2.21) 2.2.2.2 DCM If the inductor current collapses to zero in its pth discharge subinterval (refer to Figure 2.4), the MIMO converter is defined to operate in the pth discontinuous mode. Similar to CCM, the inductor current is assumed to change linearly in each subinterval. Since the inductor current starts from zero each switching cycle in DCM, the peak inductor current is given by , , , , ,1 1 1m m isL peak L i in olp i in ki i kTI i D VL+= = == ∆ =∑ ∑∑ (2.22) and the energy stored in the inductor at the very beginning when the inductor gets discharged is 0.5LI ²L, peak. 14 Assuming the converter is in the pth discontinuous mode, the effective duty ratio of the last conducting output switch Sout,p may not be Dout,eff,p. The duration that takes the inductor to completely discharge after p 1 discharge subintervals is determined by , ,max,,L ppout pIt LV−−∆ = (2.23) which is also the length of time when the inductor discharges to the pth output. Thus, the actual effective duty ratio of Sout, p is , , ,max, ,,' .p L peff out ps out p st LIDT V T− −∆= = (2.24) Regarding Section 2.2.1, the inductor discharges to the first p outputs sequentially. The first discharge state is illustrated for example. It takes the inductor Dout,eff ,1Ts to discharge to the first output. At the end of this subinterval, the energy stored in the inductor is 0.5LI ²L ,1,min. Thus, the amount of energy passing to the capacitor C1 and load R1 during Dout,eff ,1Ts is 2 2 2,1 , , ,1 , ,1 ,11 1 1( )2 2 2L L peak L peak L L peak L LW LI L I i LI i L i− − − −∆ = − − ∆ = ∆ − ∆ (2.25) where ∆iL ,1 = Dout,eff,1TsVout,1 / L. The energy stored in the capacitor C1 increases by 1,1 ,12 21 ,1 1 ,1 1 ,1 ,11 1( ) ( )2 2 2 2out outC out out out outv vW C V C V C V v∆ ∆∆ = + − − = ∆ (2.26) where ∆vout,1 = (1 Dout, eff, 1) Ts Vout,1 / C1R1, and the energy delivering to the load R1 is 12,1, ,11.outR out eff sVW D TR∆ = (2.27) Figure 2.4. Discontinuous inductor current example. ∆iL_, j0 tiLt_, pIL, peakIL_, p, maxDin, olp, iTs ∆iL+, iTs∆Dout, eff, jTs 15 Applying the energy conservation law1 1,1L C RW W W−∆ = ∆ + ∆ , and performing several substitutions and manipulations, the following equation is obtained 1 , , ,1,1 21 , ,12.2L peak out effoutout eff sLR I DVL R D T=+ (2.28) The variation of the inductor current in Dout, eff, 1Ts is then expressed as 2, 1 , ,121 ,,1,1 2.2L peak out eff sout eff sLI R D TL R TiD− +∆ = (2.29) From (2.29), it can be observed that the inductor current decreases from IL ,2,max = IL, peak – ∆iL ,1 in the second discharge subinterval. To summarize, the output voltages can be calculated as follows: Step 1: Update the output effective duty ratios. The first p 1 output effective duty ratios remain at the commanded values, whereas the pth output effective duty ratio is modified as (2.24). Step 2: Find out the starting-point of the inductor current in its jth discharge subinterval, that is, IL , j,max, using (2.15). Step 3: Solve the jth output voltage Vout, j according to (2.30). Step 4: Determine ∆iL , j in terms of Vout, j (2.8). Step 5: Repeat Step 2 to 4 until j reaches p. , ,max , ,, 2, ,2.2j L j out eff jout jj out eff j sLR I DVL R D T−=+ (2.30) It is worth noting Vout, j ( j = p+1,…, n) are zero since the last n p outputs receive no power from the inductor. The first p 1 output voltages are in the form of (2.30), and the express ion of the pth output voltage can be further simplified as , , ,max .2pout p L psLRV IT−= (2.31) 2.3 Average-Value Modeling Average-value modeling is considered as an option to analyze nonlinear time-varying power electronic systems. Although a number of techniques have been reported to acquire AVMs, state-space averaging is a prevalent method. The resulting AVM is valid in a frequency range adequately below the switching frequency. The following equations in this chapter are expressed 16 with respect to the commanded input duty ratios Din, i (i = 1,…, m) and output duty ratios Dout, j ( j = 1,…, n), which can be alternatively constructed from the input overlapping duty ratios Din,olp,i (i = 1,…, m) and output effective duty ratios Dout,eff , j ( j = 1,…, n). If the inductor current and capacitor voltages are selected as state variables, the state equations can be derived as ( )1, , , , 1 , , ,1 1m nLin i in i out j out j C j out n C ni jdiL d v d d v d vdt−+= == − − −∑ ∑ (2.32) ( ) ,, , 1,,,, 1,..., 1, .C jout j out j LjC jjC jout j Ljvd d i j nRdvCvdtd i j nR+− − = −= − = (2.33) When the input powers and output voltages are taken as the outputs, the output equations are , , , , ,, 1,...,in i in i in i in i L in ip v i v i d i m= = = (2.34) , ,, 1,..., .out j C jv v j n= = (2.35) The input voltages vin,i (i = 1,…, m), input switch duty ratios din,i (i = 1,…, m), and output switch duty ratios dout, j ( j = 1,…, n) are considered time variant, which can be represented as , , ,ˆ , 1,...,in i in i in iv V v i m= + = (2.36) , , ,ˆ, 1,...,in i in i in id D d i m= + = (2.37) , ,,, ,ˆ, 1ˆ, 2,..., .out j in jout jout j out jD d jdD d j n− == + = (2.38) In response to these inputs, the average inductor current iL, the average capacitor voltages vC, j ( j = 1,…, n), the average input powers pin,i (i = 1,…, m), and the average output voltages vout, j (j = 1,…, n), can be expressed as equilibrium points plus small-signal perturbations as follows ˆL L Li I i= + (2.39) , , ,ˆ , 1,...,C j C j C jv V v j n= + = (2.40) , , ,ˆ , 1,...,in i in i in ip P p i m= + = (2.41) , , ,ˆ , 1,..., .out j out j out jv V v j n= + = (2.42) 17 2.4 Power Management and Control Strategy 2.4.1 Power Budgeting For fixed input voltages, there could be an infinite number of combinations of duty ratios to yield the same output voltages. Accordingly, different power flow can be realized. That means it is feasible to change the ratio of the amount of power supplied by each input without changing the total power delivered to the loads while keeping the output voltages at desired levels. In a practical system, the generated power from each input can be managed by regulating the current, voltage or power based on applicable specifications. However, the total power consumed by the outputs must be equal to the total power supplied by the inputs. In order to maintain the power balance, one of the inputs should behave as a slake source. For instance, the power coming from the first input Pin,1 is relaxed while the others Pin,i (i = 2,…,m) are regulated; that is, ,1 , ,1 2n min out j in ij iP P P= == −∑ ∑ (2.43) where , , ,, 2,...,in i in i L in iP V I D i m= = (2.44) 2,,, 1,..., .out jout jjVP j nR= = (2.45) 2.4.2 Controller Design For an m-input n-output converter, it is possible to regulate the n outputs at near-constant voltages and operate the m 1 inputs at near-constant powers; that is, m 1 input powers are selected as the control objectives as well as the n output voltages, i.e. Din,i (i = 1,…, m) and Dout, j ( j = 2,…, n) are the control variables for the controller. Figure 2.5. Control block diagram of the MIMO non-inverting buck–boost converter. + PIPIPI∆y1∆y2∆ym+n-1H J***∆y1∆y2∆ym+n-1_+_+_18 Assuming the proposed converter is operating in a given region, the output voltages and input powers can be alternatively expressed in terms of Din, i (i = 1,…, m) and Dout, j ( j = 1,…, n) as follows ( )( )( ), , 1 , ,11 2 2, , 1 ,1,, , ,11 2 2, , 1 ,1, 1,..., 1, mj out j out j in k in kknout k out k k out n nkout j mj out j in k in kknout k out k k out n nkR D D D Vj nD D R D RVR D D Vj nD D R D R+=−+==−+=− = −− += =− +∑∑∑∑ (2.46) ( ), , , ,1, 1 2 2, , 1 ,1, 2,..., .min i in i in k in kkin i nout k out k k out n nkV D D VP i mD D R D R=−+== =− +∑∑ (2.47) Replacing Dout,1 with 1 Din,1 and rewriting (2.46) and (2.47) in matrix form ( ) ( ) ( ) ( ) ( )( )T,1 ,2 , ,2 ,( ) , ,..., , ,...,out in in n out out mV P P V V=Y D D D D D D (2.48) where D = (Din,1,…, Din,m, Dout,2,…, Dout,n)T. Vector-function Y represents the proposed nonlinear system, and is intended to be linearized around an equilibrium point D* for a control-oriented model ( )* * * *( ) ( ) ( ) ( )ο= + − + −Y D Y DD DD D DJ (2.49) where J is the Jacobian matrix of Y with respect to D ,1 ,1 ,1 ,1,1 , ,2 ,,2 ,2 ,2 ,2,1 , ,2 ,, , , ,,1 , ,2 ,,2 ,2,1 ,out out out outin in m out out nin in in inin in m out out nin m in m in m in min in m out out nout out outin in mV V V VD D D DP P P PD D D DP P P PD D D DV V VD D∂ ∂ ∂ ∂∂ ∂ ∂ ∂∂ ∂ ∂ ∂∂ ∂ ∂ ∂∂ ∂ ∂ ∂= ∂ ∂ ∂ ∂∂ ∂ ∂∂ ∂J⋯ ⋯⋯ ⋯⋮ ⋱ ⋮ ⋮ ⋱ ⋮⋯ ⋯⋯ ,2 ,2,2 ,, , , ,,1 , ,2 ,.outout out nout n out n out n out nin in m out out nVD DV V V VD D D D            ∂  ∂ ∂   ∂ ∂ ∂ ∂  ∂ ∂ ∂ ∂ ⋯⋮ ⋱ ⋮ ⋮ ⋱ ⋮⋯ ⋯ (2.50) 19 Therefore, the relationship between the controls and the outputs around equilibrium operating point D* is ∆ = ∆Y J D (2.51) where J represents the gain matrix of the proposed MIMO non-inverting buck–boost converter. Figure 2.5 shows the control block diagram. The compensator matrix H can be described as follows. With (2.51) written as ∆Y = J∆D*, the term ∆D* is defined as a modified vector, i.e. ∆D* = H∆D. In this manner, the goal of the matrix H is to make JH a diagonal matrix. A straightforward design is to choose H as the inverse of the gain matrix, namely J 1 . Thereafter, separate PI controllers can be implemented in the overall system to control the outputs. 2.5 Computer Studies A triple-input triple-output non-inverting buck–boost converter is examined. The input voltage are Vin,1 = 120 V, Vin,2 = 96 V and Vin,3 = 90 V. The output voltages are regulated at Vout,1 = 190 V, Vout,2 = 24 V and Vout,3 = 12 V, whereas R1 = 5 Ω, R2 = 10 Ω and R3 = 20 Ω define the loads at the corresponding outputs. Whenever the power demand is higher than the generation from Vin,2 and Vin,3, the input Vin,1 supplies the deficit power. Whenever the loads require less power than the generation from Vin,2 and Vin,3, surplus power is stored separately or the power reference should be changed appropriately. Both Pin,2 and Pin,3 are regulated at 1 kW. 2.5.1 System Response to Input Voltage Variation To investigate the dynamic performance of the proposed converter, a change in Vin,3 from 90 V to 60 V is performed. Figure 2.6 shows the simulation results of the DM and AVM. As it can be observed, the output voltages are tightly controlled. As the reference value of Pin,3 does not change and the total power fed to the loads remains constant, the deficit power due to a drop in Vin,3 is compensated by increasing Iin,3. Thus, Pin,3 is regulated at the desired level. The output voltages undergo a small transient when Vin,3 steps down. The ratio of the power drawn from these three sources stays the same and the power is stably provided to all the loads. The simulation results of DM and AVM closely match in both steady state and transient. 2.5.2 System Response to Output Load Variation Figure 2.7 illustrates the transient response due to a change in load R1 from 5 Ω to 3 Ω. As can be seen from Figure 2.7, the output voltages return to the desired levels after transient when R1 is 20 reduced to 3 Ω. The Pin,2 and Pin,3 are regulated to keep track of the command values, whereas the increased power demand is automatically supplied by Vin,1. The output voltages are regulated by receiving adequate power from Vin,1 even if R1 changes. It is clearly shown that the desired power management is achieved, drawing constant powers from Pin,2 and Pin,3 while variation in the load demand takes place. Figure 2.7 demonstrates the behaviors of the DM and AVM, which match very well in both steady state and transient. This study confirms converter’s capability to autonomously match the load variation while the Pin,2 and Pin,3 are kept around the reference values and output voltages are held stable. 2.5.3 System Response to Input Power Variation Figure 2.8 shows the response of the controlled MIMO non-inverting buck–boost converter to a reference change in Pin,2 from 1 kW to 2 kW. As can be observed, the output voltages undergo a transient and return to specified levels. The reference for Pin,3 is fixed at 1 kW to deliver constant power. Therefore, Pin,3 returns to its specified level after transient. Since Pin,2 is taking a larger part in power demand of the loads, Pin,1 automatically decreases to maintain the power balance. The results of the DM and AVM are in good agreement, as expected, which verifies the derivations presented in this chapter. Also, it validates the proposed MIMO converter can alter the ratio of the amount of power supplied by each input in case of constant load. 21 Figure 2.6. Simulated system response of the MIMO non-inverting buck–boost converter to input voltage change. 4.555.5P in,1 (kW)0.911.1P in,2 (kW)0.40.60.81P in,3 (kW)170180190200V out,1 (V)a) DMb) AVM2025V out,2 (V)101214V out,3 (V)0.1 s22 Figure 2.7. Simulated system response of the MIMO non-inverting buck–boost converter to load change. 6810P in,1 (kW)a) DMb) AVM0.811.21.41.6P in,2 (kW)0.811.21.41.6P in,3 (kW)140160180200V out,1 (V)2025V out,2 (V)0.1 s68101214V out,3 (V)23 Figure 2.8. Simulated system response of the MIMO non-inverting buck–boost converter to input power change. 3.544.555.5P in,1 (kW)11.52P in,2 (kW)a) DMb) AVM0.911.1P in,3 (kW)185190195200V out,1 (V)222426V out,2 (V)1012141618V out,3 (V)0.1 s24 CHAPTER 3: MIMO FLYBACK CONVERTER WITH STACKED CAPACITORS In this chapter, an isolated MIMO converter is investigated. The operation principle is given, and a specific switching pattern is proposed. A multivariable control scheme is presented that enables budgeting the input powers coming from different sources in addition to regulating several output voltages. Both DM and AVM are provided to validate the operation of the proposed MIMO converter. 3.1 Circuit Configuration and Operation Principle Conventional flyback converter is a good topology to derive isolated MIMO converter. In fact, there exist two methods to expand the flyback converter to its MIMO version. One is to use a transformer with a separate winding for each input and output, as shown in Figure 3.1. In this case, all the ports are galvanically isolated, though the MIMO converter is kind of paralleling flyback converters on one core. The other is to connect the inputs on a single winding at the primary side of the transformer and the outputs on the other single winding at the secondary side, as pictured in Figure 3.2. The second topology is preferable due to its reduced parts count and compact structure. Figure 3.2 shows the proposed MIMO flyback converter. The primary side sub-circuit consists of m input legs in parallel, and the secondary side sub-circuit consists of n output legs in parallel. Each input and output leg contains a forward-conducting bidirectional-blocking switch except for the first output leg interfacing Vout,1. The forward-conducting bidirectional-blocking switch is realized by a series pair of a MOSFET and a diode. The output capacitors are shared between every two outputs. For instance, C1 is placed between Vout,1 and Vout,2. The transformer is modeled as a magnetizing inductance LM in parallel with an ideal transformer and the turns ratio is defined as 1/N. The inputs and outputs can be arranged such that Vin,1 > Vin,2 > ··· > Vin,m and Vout,1 > Vout,2 > ··· > Vout,n. The switching pattern is demonstrated in Figure 3.3. All the MOSFETs operate at the same frequency. The input MOSFETs are synchronized with the same turn-on transition but different turn-off moments, and the output MOSFETs are synchronized with the same turn-off transition but different turn-on moments. The duty ratios of the input switches Sin,i and output switches Sout, j are denoted by Din, i and Dout, j, respectively. Although Sout,1 is uncontrollable, abstract gating signal is assigned on it to simplify the converter analysis, i.e. Dout,1 = 1 – max{Din,i}. If two or more input switches conduct at a time, only the one connected to the highest input voltage could be on. Similarly, if two or more output switches conduct at a time, only the one connected to the lowest 25 output voltage could be on. Effective duty ratios Din, eff, i and Dout, eff, j are defined as the portion of time when the corresponding switches conduct , , 1, ,, , 1 , , 10, 2,...,, in i in iin eff iin i in i in i in iD Di mD D D DD −− −≤= =− > (3.1) , , 1, ,, , 1 , , 10 1,..., 1, out j out jout eff jout j out j out j out jD D j nD D DDD++ +≤= = −− > (3.2) where Din,eff,1 = Din,1 and Dout,eff, n = Dout, n. 3.2 Static Characteristics 3.2.1 Operational Analysis of the Circuit The operation principle of the proposed MIMO flyback converter is to charge the transformer in one period of a switching cycle Ts, and discharge it in the following period. The converter operation can be divided into two stages. The equivalent circuits of the converter in these two stages are shown in Figure 3.4. Stage 1: The transformer is in charge-state, and one of the input switches is on. There are m subintervals. The transformer is charged by the inputs from the lowest index to the highest. When Sin,p is on and the others are off, voltage applied on the transformer’s primary winding is Vin,p. Stage 2: The transformer is in discharge-state, all the input switches are off, and one of the output switches is on. Similarly, there are n subintervals. The transformer is discharged to outputs Figure 3.1. Example of MIMO flyback converter with a separate winding for each port. Vin, 1Vin, mS in, mS in, 1 S out, 1S out, nC1Cn+−NaNbNcNd+−+−Vout , 1R1Rn+−Vout , n26 in an increasing order of indices. When Sout,q is on and the rest are off, voltage applied on the transformer’s secondary winding is Vout,q. Referring it to the primary side is Vout,q / N. 3.2.2 Steady-State Analysis 3.2.2.1 CCM The volt-second balance principle implies the average voltage of the magnetizing inductance LM is zero in steady state. Considering an ideal converter, one obtain , , , , , ,1 11 0.m nL in eff i in i out eff j out ji jV D V D VN= == − =∑ ∑ (3.3) In addition, the capacitor charge balance implies ,, , ,1 11 0, 1,..., .j jout kC j out eff k Lk k kVI D I j nN R= == − = =∑ ∑ (3.4) Solving (3.3) and (3.4) 2, , ,12, ,1min eff i in iiL nout eff j jjN D VID R===∑∑ (3.5) , , ,1, , ,2, ,1, 1,..., .min eff i in iiout j out eff j j nout eff k kkN D VV D R j nD R=== =∑∑ (3.6) Figure 3.2. Proposed MIMO flyback converter that minimizes the number of windings. vL+_transformer model1:NVin, 1 +−Vin, 2 +−Vin, m +−S in, 1S in, 2S in, mS out, 2S out, 1S out, nC1CnLMiLVout , 1R1Vout , 2R2Vout , nRnC227 Assuming the switching frequency is faster than the inductor dynamics, the shape of the magnetizing current is a polygonal curve. Thus, the peak-to-peak magnetizing current ripple is , , ,1msL in eff i in iiMTi D VL=∆ = ∑ (3.7) or , , ,1.nsL out eff j out jjMTi D VNL=∆ = ∑ (3.8) The output voltage ripples are approximated with the same assumption as , ,1, ,1(1 ), 1,..., .jout eff k s jkC j out kkjD Tv I j nC==−∆ = =∑ ∑ (3.9) 3.2.2.2 DCM DCM occurs when the proposed MIMO flyback converter operates at light load. The converter is defined to operate in the hth discontinuous mode if the energy stored in the magnetizing inductance is released completely during the hth transformer-discharge subinterval. It is worth noting that the switch Sout,h may not conduct for the whole duration Dout,eff,hTs if the converter is in the hth discontinuous mode. The actual effective duty ratio Dout,eff,h of the hth output switch is then Figure 3.3. Gating signals of the MOSFETs for the proposed MIMO flyback converter. Sin,1Sin,m−1Sin,mSout,1Sout,n−1Sout,nDin,eff,1TsTsDin,mTsDout,eff,n−1TsDout,eff,nTs28 Figure 3.4. Equivalent circuits of the MIMO flyback converter at two operating stages. vL+_1:NVin, 1Vin, mS in, 1S in, mS out, 1S out, n CnLMiLVout , 1R1Vout , nRnC1vL+_1:NVin, 1Vin, mS in, 1S in, mS out, 1S out, n CnLMiLVout , 1R1Vout , nRnC1Stage 1Stage 2vL+_1:NVin, 1Vin, mS in, 1S in, mS out, 1S out, n CnLMiLVout , 1R1Vout , nRnC1vL+_1:NVin, 1Vin, mS in, 1S in, mS out, 1S out, n CnLMiLVout , 1R1Vout , nRnC1+−+−+−+−+−+−+−+−29 1, , , , , ,1 1, ,,.m hin eff i in i out eff j out ji jout eff hout hN D V D VDV−= =−=∑ ∑ (3.10) The effective duty ratios of the output switches before the hth transformer-discharge subinterval stage the same and those after it become zero. In the following derivations of this section, updated effective duty ratios of the output switches are used. The magnetizing current ripples in each transformer-charge and -discharge subinterval are , , , ,, 1,...,sL i in eff i in iMTi D V i mL+∆ = = (3.11) , , , ,, 1,..., .sL j out eff j out jMTi D V j nNL−∆ = = (3.12) Since in DCM the magnetizing current starts from zero each switching cycle, the magnitude of the magnetizing current is , , , , ,1 1m msL peak L i in eff i in ii iMTI i D VL+= == ∆ =∑ ∑ (3.13) and the initial value of the magnetizing current in the jth transformer-discharge subinterval is ,1, ,max, ,1, 1, 2,..., .L peakjL jL peak L kkI jII i j n−−−=== − ∆ = ∑ (3.14) The voltage ripple of the output capacitor can be expressed as , ,1, ,1(1 ), 1,..., .jout eff k s jkC j out kkjD Tv I j nC==−∆ = =∑ ∑ (3.15) Assuming the MIMO flyback converter is in the jth transformer-discharge subinterval, the amount of energy stored in the transformer released to the capacitors and loads during this period is 2 2, , ,max,2 2, 1,max , ,max( ), 12( ), 2,..., .2ML peak L jL jML j L jL I I jPL I I j n−− − −− =∆ = − = (3.16) All the energy passes to the capacitors and loads. The net energy transferred to the pth capacitor Cp during the jth transformer-discharge subinterval is 30 2 2, , ,max , , 1,max, ,( ), 20, pC p j C p jC p jCV V p jPp j+ + −− ≥∆ =  < (3.17) where VC+, p, j, max is the maximum value of the pth capacitor voltage in the jth transformer-discharge subinterval , , ,, 1, , ,max ,, ,1.2jC p out eff kC p kC p j C p pout eff kkv DvV VD=+=∆∆= − +∑∑ (3.18) Meanwhile, the energy passing to the pth load Rp during the jth transformer-discharge subinterval is , , , ,, ,, 0, .out eff j out p out pR p jD V I p jPp j≥∆ = < (3.19) This yields the energy balance equation for the jth transformer-discharge subinterval, which has the following form , , , , ,1( ).nL j C k j R k jkP P P=∆ = ∆ + ∆∑ (3.20) Substituting (3.16), (3.17) and (3.19) into (3.20), the output voltages might be solved. 3.3 Average-Value Modeling As explained in Section 3.2, the converter operation is divided into two stages: (1) the transformer is in the charge-state; (2) the transformer is in the discharge-state. To facilitate the explanation of the operation of the proposed MIMO flyback converter, the differential equations for these two stages are derived. Stage 1: When Sin, p is on, the differential equations describing the converter behavior are ,LM in pdiL vdt= (3.21) ,,, 1 ,1, 1, 2,...,nC kk j jC jj nC j C kjk j jv jRdvCdv vdtC j ndt R=−−=− == − =∑∑ (3.22) , ,, 1,...,nout j C kk jv v j n== =∑ (3.23) 31 ,,, 0, 1,..., 1, 1,..., .Lin iin ii i pii i p p m== = = − + (3.24) Stage 2: When Sout, q is on, the system are represented by ,1 nLM C kk qdiL vdt N== − ∑ (3.25) ,, 1 ,,1, 1 ,1, 1, 2,..., 1, 1,...,1, nC kk j jnC j C kC jjjk j jnC j C kL jk j jv jRdv vdvC j q q nCdt Rdtdv vi C j qN dt R=−−=−−=− =− = − +=  + − =∑∑∑ (3.26) , ,, 1,...,nout j C kk jv v j n== =∑ (3.27) ,0, 1,..., .in ii i m= = (3.28) Considering (3.21)–(3.28), the AVM of the MIMO converter is derived as follows , , , , , ,1 11m n nLM in eff i in i out eff j C ki j k jd ddiL v vdt N= = == −∑ ∑∑ (3.29) , ,, ,1 11, 1,...,j j nC j C kj in eff k Lk s k s sdv vC i j ndtdN R= = == − =∑ ∑∑ (3.30) , ,, 1,...,nout j C kk jv v j n== =∑ (3.31) , , ,, 1,..., .in i in eff i Li i i md= = (3.32) 3.4 Power Management and Control Strategy For an m-input n-output flyback converter, it has m+n MOSFETs, but only m+n 1 of them can be independently controlled. If the n output voltages are desired to keep at specific values, it is unable to regulate the m input powers simultaneously. Assuming a lossless converter, the following equation based on the power balance principle is obtained , ,1 1m nin i out ji jP P= ==∑ ∑ (3.33) where Pin,i and Pout, j denote the input and output powers, respectively. 32 2, , ,1, , , , , ,2, ,1, 1,...,min eff k in kkin i in i in i in eff i in i nk out eff kkN D VP V I D V i mR D=== = =∑∑ (3.34) 2,,, 1,..., .out jout jjVP j nR= = (3.35) According to the conservation of power principle, one of the inputs should not be directly regulated. For instance, the first input is relaxed while the rest are regulated to deliver commanded powers. The output formulas are composed of (3.6) and (3.34) except for the equation of Pin,1. They can be alternatively expressed in terms of commanded duty ratios Din, i and Dout, j instead of the effective duty ratios Din,eff,i and Dout,eff, j. In matrix form, one obtains ( )T,1 ,2 , ,2 ,( ) ( ), ( ), , ( ), ( ), , ( )out in in m out out nV P P V V=Y D D D D D D… … (3.36) where D = (Din,1, …, Din,m, Dout,2, …, Dout,n)T. The Taylor series of (3.36) around an equilibrium point D* can be expressed using Jacobian matrix J of Y with respective to D ( ) ( ) ( )( )∗ ∗ ∗−= +Y D Y D J D D D (3.37) where ,1 ,1 ,1 ,1,1 , ,2 ,,2 ,2 ,2 ,2,1 , ,2 ,, , , ,,1 , ,2 ,,2 ,2,1 ,out out out outin in m out out nin in in inin in m out out nin m in m in m in min in m out out nout out outin in mV V V VD D D DP P P PD D D DP P P PD D D DV V VD D∂ ∂ ∂ ∂∂ ∂ ∂ ∂∂ ∂ ∂ ∂∂ ∂ ∂ ∂∂ ∂ ∂ ∂= ∂ ∂ ∂ ∂∂ ∂ ∂∂ ∂J⋯ ⋯⋯ ⋯⋮ ⋱ ⋮ ⋮ ⋱ ⋮⋯ ⋯⋯ ,2 ,2,2 ,, , , ,,1 , ,2 ,.outout out nout n out n out n out nin in m out out nVD DV V V VD D D D           ∂ ∂ ∂   ∂ ∂ ∂ ∂  ∂ ∂ ∂ ∂ ⋯⋮ ⋱ ⋮ ⋮ ⋱ ⋮⋯ ⋯ (3.38) The control-to-output relationship can be derived using the linearization of the static characteristic of the proposed MIMO flyback converter around an operating point. It can be observed that interactions exist between different loops, which shall be mitigated by a compensation network H. The product JH needs to be diagonal, enabling the required decoupling. For simplicity, it can be chosen that JH = diag(1,…,1)T. Thus, multiple PI controllers 33 G1,…, Gm+n 1 , as depicted in Figure 3.5, can be implemented to regulate the input powers and output voltages. By tuning the coefficients of the PI controllers, one can obtain a desired dynamic performance of the overall system. 3.5 Computer Studies A double-input double-output flyback converter is examined here to verify the proposed concept. Two input DC voltages are defined as Vin,1 = 120 V and Vin,2 = 50 V. A transformer with turns ratio N = 1/2 is incorporated, and two resistive loads are assumed to be connected to the converter, which are represented by R1 = 1 Ω and R2 = 1 Ω, respectively. The input power from the second input Pin,2 is regulated at 1.5 kW, and the output voltages are adjusted to maintain at Vout,1 = 48 V and Vout,2 = 24 V, respectively. 3.5.1 System Response to Input Voltage Variation To demonstrate the converter performance under variability of input voltage, variation in Vin,2 from 50 V to 60 V is applied. As can be seen from Figure 3.6, the input power Pin,2, is regulated in addition to the output voltages. As the preset point of the second input power does not move, power supplied by the second source remains the same as before, which results in Pin,1 keeping constant by adjusting the duty ratios through the controller. It also indicates the closed-loop behavior matches in both steady state and transient between the DM and AVM. Figure 3.5. Block diagram of the proposed control structure for the MIMO flyback converter. J 11J x1J 1xJ xxΣΣH11Hx1H1xHxxΣΣ∆Y1∆YxG1Gx-+∆Y1*∆Yx*-+H Jx=m+n-134 3.5.2 System Response to Output Load Variation To demonstrate the converter performance under variability of load, a step-change in R1 from 1 Ω to 0.7 Ω is considered. The system response is depicted in Figure 3.7. As shown, the output voltages undergo a transient as the controller regulates the output voltages to the specified reference values. The power extracted from the first input increases to compensate for the deficit power, while the second input is regulated to deliver constant power. Thus, Pin,2 has a transient but returns to the same reference level of 1.5 kW. 3.5.3 System Response to Input Power Variation To demonstrate the converter performance under variability of input power, the preset point of the second input power Pin,2 is moved from 1.5 kW to 1 kW. The corresponding transient response of the converter is pictured in Figure 3.8. Obtainable from the figure, the decreased amount of power from the second source results in more power drawn from the first source; that is, the first source supplies the power difference between the required power of the loads and readily available power from the second source. Meanwhile, the output voltages undergo a transient and return to their specified reference values. 35 Figure 3.6. Simulated system response of the MIMO flyback converter to input voltage variation. 1.21.31.41.51.6 1.41.61.82a) DMb) AVM455055V out,1 (V)22242628V out,2 (V)P in,1P in,2 (kW) (kW)0.1 s36 Figure 3.7. Simulated system response of the MIMO flyback converter to load variation. 11.522.5 a) DMb) AVM1.21.41.61.82 404550V out,1 (V)202530V out,2 (V)0.1 sP in,1P in,2 (kW) (kW)37 Figure 3.8. Simulated system response of the MIMO flyback converter to input power variation. 1.21.41.61.82 0.811.21.41.6464850V out,1 (V)a) DMb) AVM232425V out,2 (V)0.1 sP in,1P in,2 (kW) (kW)38 CHAPTER 4: DERIVATION OF MIMO CONVERTERS BASED ON DC LINK This chapter provides a description of PSCs, FCs, DLI, and DLC that are used to make up non-isolated MIMO converters. Connection rules of PSCs and FCs are introduced. A uniform structure is put forward for generating MIMO converters based on DLI and DLC. The synthesis procedures are presented. 4.1 Basic Ideas for Deriving DLI- or DLC-Coupled MIMO Converters As can be seen from the MIMO non-inverting buck–boost and flyback converters, the converter is a cascade connection of multiple inputs, inductor/transformer, and multiple outputs. Within an appropriate switch network, the inductor or transformer accumulates energy from the inputs in one period, and releases energy to the outputs in the subsequent period. The inductor and transformer function as central energy buffer element, and multiple inputs and outputs are interconnected through them. Based on this concept, the non-isolated MIMO converters can be realized by coupling multiple PSCs and FCs to an intermediate DC link with necessary switches (Figure 4.1). The DC-link quantity can be expressed by an inductor or a capacitor. The PSCs and FCs can be categorized into two types. One is voltage type, which includes pulsating voltage source cells (PVSCs) [76] and capacitor filter cells (C-FCs) [77]. These type cells either source or sink voltage. The other is current type, which includes pulsating current source cells (PCSCs) [76] and inductor-capacitor filter cells (LC-FCs) [77]. These type cells either source or sink current. The voltage-type cells shall be interconnected through DLI and the current-type cells shall be interconnected through DLC. 4.2 Non-Isolated PSCs and Their Connection Rules 4.2.1 Basic PSCs The concept of non-isolated PSCs, including PVSCs and PCSCs, are presented here. Three PVSCs extracted from buck, Ćuk, and Zeta converters are presented in Figure 4.2, and three PCSCs extracted from boost, buck–boost, and SEPIC converters are demonstrated in Figure 4.3. The PVSC is composed of a DC voltage source and a switch network. It provides a high-frequency pulse-wave voltage. When the active switch is on, the terminal voltage of the PVSC has a nonzero value. When the active switch is off, the terminal voltage is zero. The supplied power can also be properly controlled through the active switch. 39 The PCSC consists of a DC current source and a switch network. It generates a pulsating current. When the active switch is on, the outgoing current is zero. When the active switch is off, the outgoing current has a nonzero value. Independent power flow control is also achieved by closing and opening the active switch. It should be mentioned the DC current source in PCSC is not ideal, but realized by a DC voltage source associated with an inductor in practice. Thus, if a PCSC is disconnected from the circuit for a long time to stop providing power and the active switch in the PCSC is kept on, the inductor would get overcharged resulting in destruction of devices. Figure 4.1. Basic structure of MIMO converters based on DLI or DLC. Figure 4.2. Non-isolated basic PVSCs: (a) buck; (b) Ćuk; and (c) Zeta. Figure 4.3. Non-isolated basic PCSCs: (a) boost; (b) buck-boost; and (c) SEPIC. Combination of Multple PSCsDLI/DLCwith SwitchesCombination of Multple FCsPSC 1PSC 2PSC mFC 1FC 2FC nVout , 1Vout , n+−+−+−+−+−+−(a) (b) (c)+− +−Vin Vin Vin+−+−+−(a) (b) (c)−+Vin Vin Vin40 4.2.2 Hybrid PSCs In non-isolated topologies, the inductor and capacitor play the role of energy buffer. Auxiliary PSCs can be inserted to provide additional energy to the inductor or/and capacitor of a primary PSC besides the original source. In fact, auxiliary PVSC can be placed in series with an inductor of a primary PSC to obtain hybrid PVSC. Auxiliary PCSC can be placed in parallel with a capacitor of a primary PSC to obtain hybrid PCSC. Figure 4.4 shows some hybrid PVSCs, where auxiliary PVSCs—buck, Ćuk, and Zeta are inserted in series with an inductor of Zeta primary PVSC, respectively, and Figure 4.5 shows some hybrid PVSCs, where auxiliary PCSCs—boost, buck–boost, and SEPIC, are placed in parallel with a capacitor of Ćuk primary PVSC, respectively. Similarly, hybrid PCSCs can be obtained by inserting auxiliary PVSC to the inductor branch and/or paralleling auxiliary PCSC with the capacitor. Figures 4.5 and 4.6 show some hybrid PCSCs based on buck–boost and SEPIC primary PCSCs. Although the above developed and discussed hybrid PSCs are limited to two inputs, hybrid PSCs with more than two inputs can also be synthesized by the same principle. Figure 4.4. Non-isolated hybrid PVSCs: (a) buck-Zeta; (b) Ćuk-Zeta; and (c) Zeta-Zeta. Figure 4.5. Non-isolated hybrid PVSCs: (a) boost-Ćuk; (b) buck–boost-Zeta; and (c) SEPIC-Ćuk. +−+−(a)+−Vin,1+−Vin,2+−+−(b)+−Vin,1+−+ −Vin,2+−+−(c)+−Vin,1+−+−Vin,2+−+−(a)+−Vin,1+−Vin,2+−+−(b)+−Vin,1+−Vin,2+−+−(c)+−Vin,1+−−+Vin,241 4.2.3 Connection Rules of PSCs 4.2.3.1 Connection Rules of PVSCs Multiple PVSCs can be connected in series or parallel, as demonstrated in Figure 4.8(a) and (b). In series configurations, multiple PVSCs supply power individually or simultaneously. In parallel configurations, since PVSCs may have different terminal voltages, appropriate switch arrangement is required to prevent them from being directly connected together. The apparent limitation of parallel configurations is that only one source is allowed to supply power at a time. 4.2.3.2 Connection Rules of PCSCs Multiple PCSCs can be connected in series or parallel, as shown in Figure 4.8(c) and (d). In series configurations, multiple PCSCs can be connected in a manner without violating Kirchhoff’s current law provided that only one PCSC delivers power at a time. Additional switches are Figure 4.6. Non-isolated hybrid PCSCs: (a) buck-buck–boost; (b) Ćuk-Buck–boost; and (c) Zeta-buck–boost. Figure 4.7. Non-isolated hybrid PCSCs: (a) boost-SEPIC; (b) buck-SEPIC; and (c) SEPIC-SEPIC. +−(a)Vin,1+−Vin,2+−(b)Vin,1+−+ −Vin,2+−(c)Vin,1+−+−Vin,2+−(a)−+Vin,1+−Vin,2+−(b)−+Vin,1+−Vin,2+−(c)−+Vin,1+−−+Vin,242 required to prevent direct series connection; that is, when a PSCS supplies power, it is essential to insulate other PCSCs. As the DC current source of PCSC is not ideal, series configurations may only work under some constraints. In parallel configurations, the multiple PCSCs supply power individually or simultaneously. 4.3 Central Energy Buffer Element—DLI and DLC 4.3.1 DLI Cell Figure 4.9(a) shows the DLI within a switch network. It is used to interconnect voltage-type cells. The active switch is on when power is deposited, and it is off when power is withdrawn. When the active switch is in on-state, DLI is connected to the PVSCs. Energy is stored temporarily in a magnetic field. When the active switch is off, the diode conducts to provide a bypass path for the inductor current and the DLI is connected to the C-FCs. Energy is transferred from the magnetic field through the C-FCs to the loads. 4.3.2 DLC Cell Figure 4.9(b) shows the DLC within a switch network. It is used to interconnect current-type cells. The active switch is off as long as the DLC accumulates energy, and it is off when the DLC Figure 4.8. Combination of PSCs: (a) PVSCs in series; (b) PVSCs in parallel; (c) PCSCs in series; and (d) PCSCs in parallel. Figure 4.9. DC-link configurations: (a) DLI cell; and (b) DLC cell. −+−+(a) (b) (c) (d)PVSC 1PVSC m PVSC 1 PVSC m PCSC 1PCSC 1PCSC n PCSC n−(a)++−(b)−+43 releases energy. When the active switch is in off-state, the DLC is connected to the PCSCs. Energy is transferred from the PCSCs and stored electrostatically in an electric field. Voltage of the DLC increases. When the PCSCs no longer supply energy to the DLC, the active switch conducts and the diode blocks reverse current. The DLC is connected to the LC-FCs, and the energy is released from the electric field through the LC-FCs to the loads. Also, voltage of the DLC drops. 4.4 FCs and Their Connection Rules 4.4.1 C-FCs and Their Connection Rules The C-FC, as shown in Figure 4.10(a), is a first-order low-pass filter or a voltage sink leaking constant voltage. When constructing multiple-output structure using C-FCs, necessary switches are needed to avoid shorting the capacitors. In other words, multiple C-FCs need to be placed within a switch network. Similar to the combination of multiple PVSCs, the C-FCs can be connected in series or parallel, as pictured in Figure 4.11(a) and (b). When multiple C-FCs are connected in series, the C-FCs are capable to draw power simultaneously from the DLI. There exist alternative ways to place the loads so that different output voltages can be obtained. Figure 4.12(a) and (b) show two examples. One is to place a load across each capacitor, and the other is to place a load across several capacitors. The case in Figure 4.12(a) renders equal-output-voltage Figure 4.10. FCs: (a) C-FC; and (b) LC-FC. Figure 4.11. Combination of FCs: (a) C-FCs in series; (b) C-FCs in parallel; (c) LC-FCs in series; and (d) LC-FCs in parallel. (a) (b)C-FC LC-FCPCSC PVSC VoutVout−+−+(a) (b) (c) (d)C-FC 1C-FC mLC-FC 1LC-FC n LC-FC 1 LC-FC nC-FC 1 C-FC m44 and the case in Figure 4.12(b) enables a common ground for the loads. When multiple C-FCs are connected in parallel, a common ground is obtained, but energy flows through the C-FCs one-by-one. Figure 4.12(c) depicts the circuit diagram including loads. 4.4.2 LC-FCs and Their Connection Rules The LC-FC, as shown in Figure 4.10(b), is a second-order low-pass filter or a current sink leaking constant current. Multiple LC-FCs can be connected in series or parallel, as shown in Figure 4.11(c) and (d). Similar to construct multiple-output using C-FCs, necessary switch arrangements are required for putting multiple LC-FCs together. When LC-FCs are connected in series, an Figure 4.12. Outputs taken from C-FCs: (a) series-connected C-FCs and series outputs; (b) series-connected C-FCs and parallel outputs; and (c) parallel-connected C-FCs and parallel outputs. Figure 4.13. Outputs taken from LC-FCs: (a) series-connected LC-FCs and outputs; and (b) parallel-connected LC-FCs and outputs. −+−+(a) (b) (c)−+−+−+−+−+−+−+Series Connection of C-FCs Series Connection of C-FCs Parallel Connection of C-FCs−+(a) (b)−+−+−+Series Connection of LC-FCs Parallel Connection of LC-FCs45 active switch is lumped together with each LC-FC. However, the LC-FCs can only be individually charged from the DLC. The diagram with a load placed at each capacitor is depicted in Figure 4.13(a). When multiple LC-FCs are connected in parallel, a diode is placed in parallel with each LC-FC. The diode is used for circulating current when the designated LC-FC is disconnected. An active switch is also used for insulating the LC-FC when it does not sink power. Thus, independent power flow control can be achieved. In parallel configurations, the LC-FCs can draw power individually or simultaneously, and a common ground is available for the loads. The resulting circuit with loads is shown in Figure 4.13(b). 4.5 Synthesis of Non-Isolated MIMO Converters As shown in Figure 4.1, the DLI or DLC is employed to create an interface between multiple PSCs and FCs. DLI-coupled and DLC-coupled MIMO converters are then generated, respectively. 4.5.1 DLI-Coupled MIMO Converters As the average current of the inductor shall not depend on the connected PSCs or FCs, DLI is used to link the multiple PVSCs and multiple C-FCs. The synthesis of the DLI-coupled MIMO converters is briefly described as follows: Step 1: Choose appropriate PVSCs and combine them according to the connection rules described in Section 4.2.3.1. Step 2: Construct the multiple-output structure by combining the C-FCs based on the connection rules described in Section 4.4.1. Step 3: Link the multiple PVSCs and multiple C-FCs with DLI. Figure 4.14 shows six DLI-coupled MIMO converter topologies generated by using buck PVSCs and C-FCs. The subfigures in the first and second columns show topologies synthesized by series-connected and parallel-connected multiple buck PVSCs, respectively. In series input configurations, as the freewheeling diodes associated with the DLI is redundant, it can be removed. Also, the PVSCs can supply power to the DLI either individually or simultaneously. In parallel input configurations, a diode is inserted in series with MOSFET of each PVSC to avoid direct parallel connection. Thus, only one PVSC is allowed to deliver power at a time. The subfigures in the first two rows of Figure 4.13 depict MIMO converters with series-connected multiple C-FCs, and the subfigures in the last row show topologies with parallel-connected C-FCs. In the series output configurations, the loads draw power from the DLI through C-FCs simultaneously. In parallel output configurations, the loads can only draw power individually. 46 4.5.2 DLC-Coupled MIMO Converters As the average voltage of the capacitor shall not depend on the connected PSCs or FCs, DLC is used to link the multiple PCSCs and multiple LC-FCs. The synthesis of DLC-coupled MIMO converters is described as follows: Figure 4.14. Buck-derived DLI-coupled MIMO Converters: (a) buck PVSCs in series and C-FCs in series with series outputs; (b) buck PVSCs in parallel and C-FCs in series with series outputs; (c) buck PVSCs in series and C-FCs in series with parallel outputs; (d) buck PVSCs in parallel and C-FCs in series with parallel outputs; (e) buck PVSCs in series and C-FCs in parallel with parallel outputs; and (f) buck PVSCs in parallel and C-FCs in parallel with parallel outputs. +−Vin,1+−Vin,m+−Vin,1+−Vin,m+−Vin,1+−Vin,m(a)(d)(f) +−Vin,1+−Vin,m(b)(c)+−Vin,1+−Vin,m(e)+−Vin,1+−Vin,m+−Vout , 1+−Vout , n+−Vout , 1+−Vout , n+−Vout , 1+−Vout , n+−Vout , 1+−Vout , n+−Vout , 1+−Vout , n+−Vout , 1+−Vout , n47 Step 1: Choose appropriate PCSCs and combine them according to the connection rules described in Section 4.2.3.2. Step 2: Construct the multiple-output structure by combining the LC-FCs based on the connection rules described in Section 4.4.2. Step 3: Link the multiple PCSCs and multiple LC-FCs with DLC. Following the synthesis procedure, four DLC-coupled MIMO converter topologies are generated by using boost PCSCs and LC-FCs, as shown in Figure 4.15. The subfigures in the first and second columns of Figure 4.15 show topologies synthesized by series-connected and parallel-connected multiple boost PCSCs, respectively. In series input configurations, the MOSFET of each PCSC also provides a freewheeling path for other PCSCs when it does not supply power externally. In this case, only one PCSC is allowed to supply power to the DLC at a time. In parallel input configurations, the diode associated with the DLC is removed due to redundancy. Also, power can be supplied simultaneously from the PCSCs. The subfigures in the first and second rows of Figure 4.15 depict MIMO converters with series-connected multiple LC-FCs and loads, and parallel-connected multiple LC-FCs and loads, respectively. In the series output Figure 4.15. Boost-derived DLC-coupled MIMO Converters: (a) boost PCSCs in series and LC-FCs in series; (b) boost PCSCs in parallel and LC-FCs in series; (c) boost PCSCs in series and LC-FCs in parallel; and (d) boost PCSCs in parallel and C-FCs in parallel. +−+−Vin,1Vin,m+−Vin,1+−Vin,m(c)(b)(d)+−+−Vin,1Vin,m+−Vin,1+−Vin,m(a)+−Vout , 1+−Vout , n+−Vout , 1+−Vout , n+−Vout , 1+−Vout , n+−Vout , 1+−Vout , n48 configurations, only one load is allowed to draw power from the DLC through the LC-FC because series connection of current sources should be avoided. In parallel output configurations, the loads can draw power through the LC-FCs either individually or simultaneously. 4.6 Zeta-Derived DLI-Coupled MIMO Converter The layout of a newly designed MIMO converter is shown in Figure 4.16, where m series-connected Zeta PVSCs and n parallel-connected C-FCs are linked by DLI Lc. The PVSCs are arranged in a descending order of the MOSFETs’ duty ratios din,1 > din,2 > ··· > din,m, and the C-FCs are ordered such that vout,1 > vout,2 > ··· > vout,n. The switching scheme is demonstrated in Figure 4.17 with switching cycle of Ts. The trailing edges of the gating signals of the input MOSFETs coincide, so as the output MOSFETs. The Figure 4.16. Zeta-derived DLI-coupled MIMO converter example. Figure 4.17. Switching pattern for the Zeta-derived DLI-coupled MIMO converter. Vout ,1Vout,nSout, 1Sout,nSCout,nCout,1LcRout,n+−+−Rout, 1Vin, 1Vin, mS in, 1S in, mD in, 1D in, mL in, 1C in, 1L in, mC in, mSin,1Sin,2Sin,mSout,nSout,2TsSout,1Dout,eff,1TsDout,eff,nTsDin,olp,mTs49 main switch S is on whenever there is any input switch Sin,i (i = 1,…, or m) on; it is off only when all the input switches Sin,i (i = 1,…, m) are off. The following equality is satisfied ,1.min iid d==∪ (4.1) Here, din,i (i = 1,…, m) and dout, j (j = 1,…, n) denote the commanded input and output duty ratios, and d denotes the duty ratio of S. As Sout,1 is simplified as a diode here, it does not need gating signal. However, to simplify the model formulation, virtual gating signal dout,1 = 1 d is assigned on it. The concepts of the overlapping duty ratio of the input switch din,olp,i and effective duty ratio of the output switch dout,eff, j are defined same as in Section 2.1 , , ,1, ,, , , , , ,1 10, 1,..., 1, min i in olp kk iin olp i m min i in olp k in i in olp kk i k id dd i md d d d= += + = + ≤= = −− >∑∑ ∑ (4.2) , , ,1, ,, , , , , ,1 10, 1,..., 1, nout j out eff kk jout eff j n nout j out eff k out j out eff kk j k jd dd j nd d d d= += + = +<= = −− ≥∑∑ ∑ (4.3) where din,olp,m = din,m and dout,eff,n = dout,n. There are m+n switching subintervals. In each subinterval, either different combinations of sources are utilized, or one load is powered. Thus, m+n states exist in the system. 4.6.1 State-Space Averaging For an m-input n-output converter, the state-space description of the converter in each subinterval can be expressed as 1,...,k kk kk m n= += += +Kx A x B uy C x E uɺ (4.4) where x is a state vector containing the inductor currents and capacitor voltages, y is an output vector containing m input powers and n output voltages, u is an input vector containing the input voltages. K is a matrix containing the inductances and capacitances. Matrices Ak, Bk, Ck, and Ek contain constants of proportionality. Given these state equations, one obtains 50 = += +Kx Ax Buy Cx Euɺ (4.5) where , , , ,1 1, , , ,1 1, , , ,1 1, , , ,1 1.m nin olp i i out eff j m ji jm nin olp i i out eff j m ji jm nin olp i i out eff j m ji jm nin olp i i out eff j m ji jd dd dd dd d+= =+= =+= =+= == += += += +∑ ∑∑ ∑∑ ∑∑ ∑A A AB B BC C CE E E (4.6) It is worth to note the first output’s effective duty ratio dout,eff,1 is not controllable. Therefore, it has to be expressed in terms of the other output effective duty ratios and input overlapping duty ratios as , ,1 , , , ,1 21 .m nout eff in olp i out eff ji jd d d= == − −∑ ∑ (4.7) Considering (4.7), (4.6) can be modified as follows , , 1 , , 1 11 2, , 1 , , 1 11 2, , 1 , , 1 11 2, , 11( ) ( )( ) ( )( ) ( )( )m nin olp i i m out eff j m j m mi jm nin olp i i m out eff j m j m mi jm nin olp i i m out eff j m j m mi jmin olp i i m ouid dd dd dd d+ + + += =+ + + += =+ + + += =+== − + − += − + − += − + − += − +∑ ∑∑ ∑∑ ∑∑A A A A A AB B B B B BC C C C C CE E E, , 1 12( ) .nt eff j m j m mj+ + +=− + ∑ E E E (4.8) Alternatively, one can obtain the following form of the state equations ,,, ,,,, , , , , ,, , , , ,1 1, , , , ,(1 ) , 1,...,( )(1 ) , 1,.in iin icin i out jin ic in im mLin i in olp k in i in olp k ck i k im m nLc in olp k in i c out eff j Ci k i jm mCin i in olp k L in olp k Lk i k idiL d v d v i mdtdiL d v v d vdtdvC d i d i idt= == = == == − − == + −= − + − =∑ ∑∑∑ ∑∑ ∑, ,, , ,..,, 1,...,out i out icC Cout j out eff j Ljmdv vC d i j ndt R= − = (4.9) 51 ,,, , , ,,( ), 1,...,, 1,..., .c in iout jmin i in olp k in i L Lk iout j cp d v i i i mv v j n== + = = =∑ (4.10) To obtain the steady-state input powers and output voltages, all time-derivative terms in (4.9) and (4.10) are set to zero, which yields , ,, , , , ,, ,1, ,2, ,1(1 ), 1,...,1out eff km m mmin olp p in k in olp pin olp kk p k p kk iin i in i m nin olp k kk i kD V DDP V i mD D R= = === =−= =−∑∑ ∑∑∑ ∑ (4.11) , ,, , , , ,1, , ,21(1 ), 1,..., .out eff km m min olp p in k in olp pk p k p kout j out eff j j nkkD V DV D R j nD R= = ==−= =∑∑ ∑∑ (4.12) 4.6.2 Design Considerations The parameters of the proposed converter’s components need to be properly selected. These include all the inductances and capacitances. Assuming time constants are relatively large with respect to the switching cycle, the inductor current ripples and capacitor voltage ripples can be calculated based on the linear-ripple approximation. The selection of input inductances Lin,i (i = 1,…, m) is similar as that in a conventional converter. When the input switch Sin,i is on, the voltage across inductor Lin,i is Vin,i. Then ,, ,, ,, 1,...,in iLin i in imin olp k sk iIL V i mD T=∆= =∑ (4.13) where ,in iLI∆ is the desired peak-to-peak current ripple of the inductor Lin,i. Thus, if ,in iLI∆ needs to be controlled within a certain value ,(max)in iLI∆ , the following inequality should be satisfied ,, , ,,(max), 1,..., .in imin olp k in i sk iin iLD V TL i mI=≥ =∆∑ (4.14) Similarly, if a maximum inductor current ripple (max)cLI∆ is required, the central inductance Lc would be determined as follows , , , , ,1(max)(1 ).cm m msc in olp k in i in olp ki k i k iLTL D V DI= = =≥ −∆ ∑∑ ∑ (4.15) 52 In steady state, when the ith input switch Sin,i is off, the current flowing through the ith input capacitor is the ith input inductor current ,,,, ,, 1,..., .(1 )in iin iCin i Lmin olp k sk iVC I i mD T=∆= =−∑ (4.16) If the input capacitors’ voltage ripple is limited within a specific value ,(max)in iCV∆ , the input capacitance should satisfy the following conditions: ,, ,, , , , ,, , , ,1,2(max), ,1(1 )(1 ), 1,..., .1in iout eff km m mm min olp p in k in olp pin olp k s in olp kk p k p kk i k iin i m nCin olp k kk i kD V DD T DC i mV D D R= = == == =−−≥ =∆−∑∑ ∑∑ ∑∑ ∑ (4.17) Also, if the output capacitance is selected to comply with the maximum allowable voltage ripples ,(max)out jCV∆ ,, , ,,(max), 1,..., .out jout eff j s out jout jC jD T VC j nV R≥ =∆ (4.18) 4.6.3 Power Flow Management The MIMO converter should be capable of controlling several input powers synergistically and coping with power mismatch among different inputs. As it is not possible to regulate all the input powers and output voltages at the same time, one of the inputs is relaxed. If power losses are neglected and a steady-state operating point is assumed, the energy conservation principle implies that the total power supplied by the inputs must be equal to the total power consumed by the outputs , ,1 1.m nin i out ji jP P= ==∑ ∑ (4.19) where the output powers are calculated as ,2,, 1,..., .Cout jout jjVP j nR= = (4.20) 4.6.4 Control Scheme A multivariable controller is proposed to regulate the input powers and output voltages simultaneously. For an m-input n-output converter, it has m+n 1 degrees of freedoms. Here, the 53 last m 1 input powers are regulated at command values in addition to the n output voltages. The controller design and stability assessment often require the control-to-output transfer function extracted from small-signal model using the state-space averaging technique. For this purpose, the average model described in (4.5) is linearized around an equilibrium operating point. The steady-state terms are removed from the perturbed equations, and only the small-signal linear terms and nonlinear terms are considered. The higher-order perturbation terms are negligible since they are very small. Also, the variations of the input voltages are neglected and the commanded duty ratios are alternatively constructed from the input overlapping and output effective duty ratios. In matrix form, the small-signal model of the open-loop system is ˆˆ ˆˆˆ ˆ = += +Kx Ax Bdy Cx Edɺ ɶ ɶɶ ɶ (4.21) where xˆ, yˆ , and ˆdare perturbations around equilibrium points ,1 , ,1 , ,1 ,TT,1 ,2 , ,2 ,T,1 ,2 , ,2 ,ˆ ˆ ˆˆ ˆ ˆ ˆ ˆ( ,..., , , ,..., , ,..., )ˆ ˆ ˆ ˆ ˆ ˆ( , ,..., , ,..., )ˆ ˆ ˆ ˆ ˆˆ ( , ,..., , ,..., )in in m c in in m out out nL L L C C C Cout in in m out out nin in in m out out ni i i v v v vv p p v vd d d d d ===xyd (4.22) and Aɶ , Bɶ , Cɶ , and Eɶ are coefficient matrices of open-loop system. A multivariable controller, as shown in Figure 4.18, can be implemented to regulate the input powers and output voltages, which can be described as Figure 4.18. Block diagram of the closed-loop system. -+ PI11PI1c-+sampling circuitMIMOCPIc1PIcc++M* c = m+n−1yˆ rˆ vˆ ˆ ˆ ˆˆ ˆ ˆ = + = +Kx Ax Bvy Cx Evɺ ɶ ɶɶ ɶ++54 IPs= +KM K (4.23) where KP is the proportional gain matrix and KI is the integral gain matrix. The state-space description of the closed-loop system is then derived as follows ˆ ˆˆˆˆˆˆˆˆ      = +           = +   xK 0 xA Br0 I ppxy C Erpɺɺ (4.24) where rˆ is the reference vector, pˆ is the defined as ˆ ˆ ˆ= −p r yɺ , and A ,B,C, and E are the state, input, output, and feedforward matrices of the closed-loop system ( )11 1221 22121 21( )P P−  =      =    == +A AAA ABBBC C CE I EK EKɶ ɶ (4.25) with ( )( )11 11112121111221( )( )P PI P PP IIP−− −−− −−−=== − +=− +− +− +AAA IK A K BK I EK CK BK K BK IK CA I KEK EKEEKEɶ ɶɶ ɶɶ ɶ ɶ ɶɶɶ ɶɶ (4.26) ( )( )1211 11P P P PP P−− −−− ++= = −K BK K BK I EK EKI I KBEB EKɶ ɶ ɶ ɶɶ ɶ (4.27) 1112( )( ) .PP I−− = += +C I K CC IEEK KEɶ ɶɶɶ (4.28) 4.6.5 Case Studies A double-input double-output Zeta-derived DLI-coupled converter is considered in this section. Two input voltages Vin,1 = 120 V and Vin,2 = 50 V are assumed. The outputs are regulated at Vout,1 = 55 48 V and Vout,2 = 24 V, while two 1 Ω resistors define the loads. Pin,2 is kept at 1.5 kW while Pin,1 is relaxed. To demonstrate the operation of the proposed MIMO converter, several transient studies are described below, in which the converter is assumed to initially operate in steady state defined by the nominal condition specified above. In the first study, a step-change in Vin,2 from 50 V to 60 V is applied. The converter response is depicted in Figure 4.19, which indicates the output voltages are regulated and return to their desired values after the transient. Since the demanded power stays the same, the input powers drained from both sources return to their preset values. In the second study, the converter response to a load change is considered. Figure 4.20 shows the system response to the first load R1 changing from 1 Ω to 0.7 Ω. The controller regulates the outputs to track the specified voltage values, while the increased power demand from R1 is compensated by the first input. The input power Pin,2 is regulated to remain constant at the reference level of 1.5 kW. In the final study, it is assumed that the second source Vin,2 is disconnected. Figure 4.21 illustrates the simulation result corresponding to this event. The power drop in the second input automatically results in the first input supplying the deficit power. The output voltages undergo a transient and return to their specified reference levels, as expected. 56 Figure 4.19. Simulated waveforms of the Zeta-derived DLI-coupled MIMO converter in response to step change in input voltage. 1.31.41.51.61.41.61.824648505254222426280.2 sP in,1 (kW)P in,2 (kW)V out,1 (V)V out,2 (V)57 Figure 4.20. Simulated waveforms of the Zeta-derived DLI-coupled MIMO converter in response to step change in load. 11.522.511.523540455055202530350.2 sP in,1 (kW)P in,2 (kW)V out,1 (V)V out,2 (V)58 Figure 4.21. Simulated waveforms of the Zeta-derived DLI-coupled MIMO converter in response to one source missing. 012300.511.51020304050510152025300.2 sP in,1 (kW)P in,2 (kW)V out,1 (V)V out,2 (V)59 CHAPTER 5: EXTENSION OF SISO CONVERTERS TO THEIR MIMO VERSIONS This chapter proposes some general synthesis techniques for the generation of both non-isolated and isolated MIMO converters. These techniques are simple in concept. Using PSCs and FCs, feasible MIMO converters are constructed by replacing the original PSC or DC source of SISO converters with combination of multiple PSCs, and the original FC or simply the C-FC with combination of multiple FCs or C-FCs. 5.1 Isolated PSCs and Their Connection Rules Apart from the six aforementioned non-isolated PSCs, six isolated PSCs are presented here. Figure 5.1 shows three isolated PVSCs, including full-bridge isolated buck, push–pull isolated buck, and forward. Figure 5.2 shows three isolated PCSCs, including full-bridge isolated boost, push–pull isolated boost, and flyback. Applying analog connection rules of non-isolated PVSCs, isolated PVSCs can be connected in series or parallel. In series configurations, multiple isolated PVSCs deliver power individually or simultaneously. In parallel configurations, only one PVSC is able to deliver power at a time. Same as the non-isolated PCSCs, multiple isolated PCSCs can be connected in series or parallel. In series configurations, direct connection of isolated PCSCs should be avoided as per Kirchhoff’ current law and appropriate adjustment of duty ratios is needed. Thus, only one PCSC delivers power at a time. In parallel configurations, multiple isolated PCSCs can deliver power individually or simultaneously. Therefore, the connection rules are uniform for both non-isolated and isolated PSCs. In the following derivations, they will not be distinguished. 5.2 Basic Configuration of Conventional SISO Converters For the SISO converters, they can be decomposed into PSC and FC (Figure 5.3). As can be seen, the output capacitors are employed in these converters to balance the instantaneous power difference between the input source and output load and minimize voltage variation. In other words, the C-FC is always included at the load while it is possible to decompose the LC-FC into an inductor and a C-FC. Figure 5.4 shows the basic structure of non-isolated SISO converter. It comprises an input source, an active switch, at least one intermediate storage element, a diode, and a filter (consisting of an optional inductor and a C-FC). The input can be a stable DC voltage source or a stable DC 60 current source. The active switch operates to chop the voltage source or current source to a high-frequency pulse-train voltage or current waveform. The intermediate storage element between the input and filter alters the high-frequency pulse-train voltage or current waveform to high-frequency pulse-train current or voltage waveform. The diode can be placed in two manners. The series diode blocks possible voltage difference, while the parallel diode is supplemented for circulating possible current. Finally, the high-frequency pulse-train voltage or current waveform is filtered out by the combination of optional inductor and C-FC. Figure 5.5 shows the basic structure of isolated SISO converter. It consists of a DC voltage source or a DC current source, a switch network, a high-frequency transformer, an output rectifier, and a filter (consisting of an optional inductor and a C-FC). The operational principle of isolated SISO converter is to convert the DC source to an AC pulsating source through the switching network, step-up or -down the AC source via the transformer, rectify the AC source to DC, and filter it to smooth and stable DC voltage. The SISO converters can be identified into two types. One is the voltage-source type, including buck, buck–boost, Zeta, full-bridge isolated buck, push–pull isolated buck, forward, and flyback converters, and the other is current-source type, including boost, Ćuk, SEPIC, full-bridge isolated boost, and push–pull isolated boost converters. Figure 5.1. Isolated PVSCs: (a) full-bridge isolated buck; (b) push–pull isolated buck; and (c) forward. Figure 5.2. Isolated PCSCs: (a) full-bridge isolated boost; (b) push–pull isolated boost; and (c) flyback. (c)+−Vin+−(b)+−Vin+−+−Vin(a)+−(c)+−Vin+−Vin(a)+−Vin(b)61 5.3 Realization of Multiport Structure 5.3.1 Multiple-Input Structure Multiple PSCs can be connected in series or parallel as shown in Figure 4.8, and then use them to replace the original single PSC or DC source of the SISO converter. Specifically, the combination of multiple PVSCs is used to replace the voltage-type blocks (including the original PVSC and DC voltage source), and the combination of multiple PCSCs is used to replace the current-type blocks (including the original PCSC and DC current source). 5.3.2 Multiple-Output Structure Multiple FCs can be connected in series or parallel as shown in Figure 4.10, and then use them to replace the original single FC of the SISO converter. In fact, original C-FC is replaced with multiple series- or parallel-connected C-FCs and original LC-FC is replaced with multiple series- or parallel-connected LC-FCs. In addition, as the LC-FC could be represented by an inductor followed by a C-FC, it is possible to only replace the separated C-FC from LC-FC with the combination of multiple C-FCs. Figure 5.3. General configuration of a SISO converter. Figure 5.4. Basic configuration of non-isolated SISO converter. Figure 5.5. Basic configuration of isolated SISO converter. PSC FC Vout+−orDC Voltage orCurrent SourceActiveSwitchEnergy IntermediateStorage ElementPowerDiodeOpitionalInductorOutputC-FCVout+−orDC Voltage orCurrent SourceSwitchNetworkHigh-FrequencyTransformerOutputRectifierOpitionalInductorOutputC-FCVout62 5.4 Synthesis of MIMO Converters Replacing different blocks, four combinatorial structures of MIMO converter can be derived from the PVSC-source SISO converter and two combinatorial structures of MIMO converter can be derived from the PCSC-source SISO converter. Table 5.1 shows the eligible blocks of a SISO converter that can be replaced with multiport structure. 1) The original PSC and integrated FC of the SISO converter are replaced. This method is applicable to extend both PVSC- and PCSC-source SISO converters to their MIMO versions. Figure 5.6 shows the circuit topology of MIMO converter generated by replacing the original PVSC of the buck converter with series-connected multiple buck PVSCs, and the original LC with series-connected multiple LC-FCs. Similarly, using series or parallel connection of multiple Ćuk, Zeta, full-bridge isolated buck, push–pull isolated buck, forward PVSCs, or their combination to replace the original PVSC can lead to several other MIMO converters. Also, the original single LC-FC can be replaced with alternative series or parallel connection of multiple LC-FCs. Figure 5.7 shows the circuit topology of MIMO converter generated by replacing the original PCSC of the flyback converter with parallel-connected multiple flyback PCSCs, and the original C-FC with parallel-connected multiple C-FCs. Similarly, using series or parallel connection of multiple boost, buck–boost, SEPIC, full-bridge isolated boost, push–pull isolated boost PCSCs, or their combination to replace the original PCSC can lead to several other MIMO converters. Table 5.1. Eligible blocks that can be replaced with multiport structure. DC voltage source DC current source PVSC PCSC C-FC LC-FC Buck     Ćuk     Zeta     Boost    Buck–boost    SEPIC    Full-bridge buck     Push–pull buck     Forward     Full-bridge boost    Push–pull boost    Flyback    63 Figure 5.6. Circuit configurations: (a) SISO buck converter; and (b) PVSC-source MIMO converter generated by several buck PVSC in series and LC-FCs in series. Figure 5.7. Circuit configurations: (a) SISO flyback converter; and (b) PCSC-source MIMO converter generated by several flyback PCSCs in parallel and C-FCs in parallel. +−Vin,1+−Vin,m+−VinOriginal Buck PVSC Original LC-FCSeries-Connected Buck PVSCsSeries-Connected LC-FCs(a) (b)Vout +−+−Vout , 1+−Vout , n+−Vin,1+−Vin,mParallel-Connected C-FCsParallel-Connected Flyback PCSCs+−Vin+−Vout Original Flyback PCSC Original C-FC(a)(b)+−Vout , 1+−Vout , n64 Series connection of multiple C-FCs can also be used to replace the original single C-FC, which may give different configurations. 2) The original DC source and FC of the SISO converter are replaced. This method is applicable to extend both PVSC- and PCSC-source SISO converters to their MIMO versions. Figure 5.8 shows a voltage-source MIMO converter based on Zeta converter. The original DC voltage source is replaced with parallel-connected multiple buck PVSCs and the original LC-FC is replaced with parallel-connected multiple LC-FCs. In this topology, the PVSCs supply power in interleaved mode, and the loads draw power individually. Figure 5.9 depicts a current-source MIMO converter, which is generated by integrating multiple boost PCSCs with the primary side sub-circuit of a push-pull isolated boost converter, and multiple series-connected FCs with the secondary side sub-circuit while providing a common ground for the loads. In this topology, all the PCSCs are allowed to deliver power either individually or simultaneously, but the loads can only draw power simultaneously. 3) The original PSC and separated C-FC from LC-FC of the SISO converter are replaced. This approach is only applicable for extension of PVSC-source SISO converter. Figure 5.8. Circuit configurations: (a) SISO Zeta converter; and (b) MIMO Zeta converter with parallel-connected buck PVSCs as inputs and parallel-connected LC-FCs as outputs. Original LC-FCParallel-Connected Buck PVSCs(a)(b)+−VinOriginal DC Voltage SourceParallel-Connected LC-FCs+−Vout , 1+−Vout , n+−Vin,1+−Vin,m+−Vout65 Figure 5.10 presents a PVSC-source MIMO converter based on forward converter topology. Derivation of this MIMO converter is achieved by putting multiple forward PVSCs in series to replace the original PVSC, and putting multiple C-FCs in parallel to replace the separated C-FC. Here, the PVSCs power the loads individually or simultaneously, but more than one load may be allowed to draw power at a time. 4) The original DC source and separated C-FC from LC-FC of the SISO converter are replaced. This method is only application for extension of PVSC-source SISO converter. Figure 5.11 shows an example for the synthesis of PVSC-source MIMO converter. The primary side sub-circuit is formed by replacing the original DC voltage source of the full-bridge converter with series-connected multiple buck PVSCs, and the secondary side sub-circuit is configured by using parallel-connected multiple C-FCs to replace the separated C-FC. In this topology, simultaneous power transfer from the multiple PVSCs is available, while only one load is allowed to draw power at a time. Figure 5.9. Circuit configurations: (a) SISO push–pull isolated boost converter; and (b) MIMO push–pull isolated boost converter with parallel-connected boost PCSCs as inputs and series-connected C-FCs as outputs. +−Vin,1+−Vin,1+−Vin+−Vout Original DC Current Source+−Vout , 1+−Vout , nOriginal C-FCParallel-Connected Boost PCSCs Series-Connected C-FCs(a)(b)66 5.5 Transformer-Coupled MIMO Converters There are two categories of isolated MIMO converters. One category of converters uses multiple transformers, and the other category of converters involves a single transformer with multiple windings wound on a core. Using multiple isolated PSCs is one option to implement isolation via multiple transformers, but isolated MIMO converters can also be derived from transformer-coupled SISO converter topology in an inexpensive manner, by adding multiple primary windings with primary side sub-circuits, or/and multiple secondary windings with secondary side sub-circuits. In general, the transformer-coupled MIMO converters can be divided into four categories: 1) single-primary-single-secondary-winding-transformer-coupled (SPSSWTC) topologies; 2) single-primary-multiple-secondary-winding-transformer-coupled (SPMSWTC) topologies; 3) multiple-primary-single-secondary-winding-transformer-coupled (MPSSWTC) topologies; and 4) multiple-primary-multiple-secondary-winding-transformer-coupled (MPMSWTC) topologies. Figure 5.10. Circuit configurations: (a) SISO forward converter; and (b) PVSC-source MIMO converter generated by several forward PVSCs in series and C-FCs in series. (a)+−VinOriginal Forward PVSCSeries-Connected Forward PVSCs+−Vout Separated C-FCSeries-Connected C-FCs+−Vout , 1+−Vout , n(b)+−Vin,1+−Vin,m67 1) SPSSWTC MIMO converters can be obtained by using the combinations of multiple non-isolated PSCs to replace the original DC source of the conventional transformer-isolated SISO converter and the combinations of multiple FCs to replace the original single FC. In these topologies, galvanic isolation only occurs between the primary side and secondary side. Basic topology structure is demonstrated in Figure 5.12. 2) SPMSWTC MIMO converters can be obtained replacing the original DC source of the conventional isolated SISO converter with multiple non-isolated PSCs and paralleling the secondary sub-circuits. This kind of converters provides isolation between multiple outputs, and turns ratios can be chosen to acquire the desired output voltages. The general configuration of SPMSWTC MIMO converter is shown in Figure 5.13. 3) MPSSWTC MIMO converters can be obtained by paralleling primary sub-circuits and replacing the original single FC with its multiport version. Therefore, galvanic isolation is provided between multiple inputs, and the voltage levels are adjustable through the transformer Figure 5.11. Circuit configurations: (a) SISO full-bridge isolated buck converter; and (b) MIMO full-bridge isolated buck converter with series-connected buck PVSCs as inputs and parallel-connected C-FCs as outputs. +−Vin,1+−Vin,m+−Vin+−Vout Series-Connected Buck PVSCs Parallel-Connected C-FCsOriginal DC Voltage Source Separated C-FC(a)(b)+−Vout , 1+−Vout , n68 Figure 5.12. Basic structure of SPSSWTC MIMO converter. Figure 5.13. Basic structure of SPMSWTC MIMO converter. Figure 5.14. Basic structure of MPSSWTC MIMO converter. Figure 5.15. Basic structure of MPMSWTC MIMO converter. SwitchNetworkOutputRectifierOpitionalInductorPSC 1PSC 2PSC mC-FC- Load 1C-FC- Load nVout , 1Vout , nOutputRectifier 1OpitionalInductor 1OutputC-FC 1OutputRectifier nOpitionalInductor nOutputC-FC nSwitchNetworkPSC 1PSC 2PSC mVout , 1Vout , nOutputRectifierOpitionalInductorSwitchNetwork 1SwitchNetwork mDCSource 1DCSource mOutputC-FC 1Output C-FC nVout , 1Vout , nSwitchNetwork 1DCSource 1SwitchNetwork mDCSource mOutputRectifier 1OpitionalInductor 1OutputC-FC 1OutputRectifier nOpitionalIndcutor nOutputC-FC nVout , 1Vout , n69 winding ratios, as each input provides power to the system through an individual winding. Typical structure of MPSSWTC MIMO converter is depicted in Figure 5.14. 4) MPMSWTC MIMO converters can be simply obtained by paralleling both primary and secondary sub-circuits. Therefore, all ports ate electrically isolated and voltage levels of each port could be flexible. Structure illustration is shown in Figure 5.15. 70 CHAPTER 6: CONCLUSIONS In this concluding chapter, the major contributions of this research project are summarized and possible future work is discussed. 6.1 Contributions of the Thesis This work introduces theoretical concepts and presents overall discussions to the field of MIMO converters. Two major contributions are reported. Firstly, two typical MIMO converters are introduced in detail, and a simple approach that can be applied in analyzing these converters is presented. The proposed two MIMO converters include a non-isolated topology and an isolated topology. Both topologies can be scaled to arbitrary numbers of inputs and outputs, and the output voltages could be regulated individually either greater than the maximum input voltage or less than the minimum. Closed-loop examples consisting of decoupler and controller are shown to provide actions for power management, voltage regulation and duty ratio adjustments. The DMs are presented to validate the operation of the proposed MIMO converters, and the AVMs are developed and compared with the DMs. AVMs and DMs are shown to be in good agreement. Secondly, general approaches for deriving MIMO converters are proposed. With basic building blocks (including PSCs and FCs), a basic structure based on DLI/DLC is proposed for the synthesis of a family of non-isolated MIMO converters. Connection rules for building blocks of PSCs and FCs are listed. Formalization of interface between multiple PSCs and FCs is realized by DLI or DLC with necessary switches. Following a uniform method, two types of DC-linked MIMO converters are obtained. In the end, a set of uniform rules for synthesizing general MIMO converters based on basic SISO converters are proposed. MIMO converters can be derived by replacing the PSC or original DC source of a conventional converter with series- or parallel-connected PSCs, and the FC or original DC load with series- or parallel-connected FCs. 6.2 Future Work 6.2.1 Practical Implementation Realization of the proposed MIMO converters in hardware is not considered in this thesis. The theoretical derivation of the MIMO non-inverting buck–boost and flyback converters have been validated by computer simulations, but it is still worthwhile to investigate if their practical 71 implementation would be effective for a specific application such as DC distribution system of a smart home with multiple DC sources and loads at different voltage levels. 6.2.2 Controller Optimization The controllers employed with the MIMO converters proposed in Chapter 2–4 of the thesis have been demonstrated to be capable of regulating the input powers along with the output voltages. However, since the proposed methodology is based on linearization, more development and extensive analysis is required for controller tuning and optimal performance over wide range of operating conditions. 6.2.3 Non-Ideal MIMO Converters The AVMs in the Chapter 2 and 3 are built based on ideal MIMO converters, which neglect parasitics or losses to simplify the derivations and modeling procedures, and the resulting models may not be sufficiently accurate for practical converters. It would be helpful to model the MIMO converters with parasitic elements. 6.2.4 Bidirectional Multiport Converters The converters presented in this thesis allows for only unidirectional power flow. However, to satisfy the application where an energy storage element is indispensable, bidirectional power flow is required. Instead of using additional converter for feeding the energy back, it would be desirable to analyze and synthesize bidirectional multiport converters. 72 REFERENCES [1] D. Nilsson and A. Sannino, “Efficiency analysis of low- and medium-voltage DC distribution systems,” in Proceedings of IEEE Power and Energy Society General Meeting, 2004, vol.2, pp. 2315-2321. [2] P.-W. Lee, Y.-S. Lee, and B.-T. Lin, “Power distribution systems for future homes,” in Proceedings of IEEE International Conference on Power Electronics and Drive Systems, vol. 2, 1999, pp. 1140–1146. [3] A. Sannino, G. Postiglione, and M. Bollen, “Feasibility of a DC network for commercial facilities,” IEEE Transactions on Industry Applications, vol. 39, no. 5, pp. 1499–1507, Sep./Oct. 2003. [4] W. Jiang and B. Fahimi, “Multiport power electronic interface—Concept, modeling and design,” IEEE Transactions on Power Electronics, vol. 26, no. 7, pp. 1890–1900, Jul. 2011. [5] H. Behjati and A. Davoudi, “Single-stage multi-port DC–DC converter topology,” IET Power Electronics, vol. 6, no. 2, pp. 392–403, Feb. 2013. [6] H. Behjati and A. Davoudi, “A multiple-input multiple-output DC–DC converter,” IEEE Transactions on Industry Applications, vol. 49, no. 3, pp. 1464–1479, May/Jun. 2013. [7] H. Behjati and A. Davoudi, “Power budgeting between diversified energy sources and loads using a multiple-input multiple-output DC–DC converter,” IEEE Transactions on Industry Applications, vol. 49, no. 6, pp. 2761–2772, Nov/Dec. 2013. [8] H. Keyhani and H. A. Toliyat, “A ZVS single-inductor multi-input multi-output DC-DC converter with the step up/down capability,” in Proceedings of IEEE Energy Conversion Congress and Exposition, 2013, pp. 5546–5552. [9] H. Matsuo, W. Lin, F. Kurokawa, T. Shigemizu, and N. Watanabe, “Characteristics of the multiple-input DC–DC converter,” IEEE Transactions on Industrial Electronics, vol. 51, no. 3, pp. 625–631, Jun. 2004. [10] K. Kobayashi, H. Matsuo, and Y. Sekine, “Novel solar-cell power supply system using a multiple-input DC–DC converter,” IEEE Transactions on Industrial Electronics, vol. 53, no. 1, pp. 281–286, Feb. 2006. [11] Y.-M. Chen, Y.-C. Liu, and F.-Y. Wu, “Multiinput converter with power factor correction, maximum power point tracking, and ripple-free input currents,” IEEE Transactions on Power Electronics, vol. 19, no. 3, pp. 631–639, May 2004. 73 [12] N. D. Benavides and P. L. Chapman, “Power budgeting of a multiple input buck–boost converter,” IEEE Transactions on Power Electronics, vol. 20, no. 6, pp. 1303–1309, Nov. 2005. [13] F. Caricchi, F. Crescimbini, O. Honorati, A. D. Napoli, and E. Santini, “Testing of a new DC/DC converter topology for integrated wind-photovoltaic generating systems,” in Proceedings of European Conference on Power Electronics, 1993, pp. 83–88. [14] L. Solero, F. Caricchi, F. Crescimbini, O. Honorati, and F. Mezzetti, “Performance of a 10 kW power electronic interface for combined wind/PV isolated generating systems,” in Proceedings of IEEE Power Electronics Specialists Conference, 1996, pp. 1027 - 1032. [15] Y.-M. Chen, Y.-C. Liu, and S.-H. Lin, “Double-input PWM DC/DC converter for high-/low-voltage sources,” IEEE Transactions on Industrial Electronics, vol. 53, no. 5, pp. 1538–1545, Oct. 2006. [16] J. Zeng, W. Qiao, L. Qu, and Y. Jiao, “An isolated multiport DC–DC converter for simultaneous power management of multiple different renewable energy sources,” IEEE Journal of Emerging and Selected Topics in Power Electronics, vol. 2, no. 1, pp. 70–78, Mar. 2014. [17] A. Hussam, F. Tian, and I. Batarseh, “Tri-modal half-bridge converter topology for three-port interface,” IEEE Transactions on Power Electronics, vol. 22, no. 1, pp. 341–345, Jan. 2007. [18] Z. Qian, O. Abdel-Rahman, H. Al-Atrash, and I. Batarseh, “Modeling and control of three-port DC/DC converter interface for satellite applications,” IEEE Transactions on Power Electronics, vol. 25, no. 3, pp. 637–649, Mar. 2010. [19] Z. Qian, O. Abdel-Rahman, and I. Batarseh, “An integrated four-port DC/DC converter for renewable energy applications,” IEEE Transactions on Power Electronics, vol. 25, no. 7, pp. 1877–1887, Jul. 2010. [20] S. Falcones, R. Ayyanar, and X. Mao, “A DC–DC multiport-converter-based solid-state transformer integrating distributed generation and storage,” IEEE Transactions on Power Electronics, vol. 28, no. 5, pp. 2192–2203, May 2013. [21] Z. Wang and H. Li, “An integrated three-port bidirectional DC–DC converter for PV application on a DC distribution system,” IEEE Transactions on Power Electronics, vol. 28, no. 10, pp. 4612–4624, Oct. 2013. [22] H. Wu, P. Xu, Z. Zhou, and Y. Xin, “Multiport Converters Based on Integration of Full-Bridge and Bidirectional DC–DC Topologies for Renewable Generation Systems,” IEEE Transactions on Industrial Electronics, vol. 61, no. 2, pp. 856–869, Feb. 2014. 74 [23] A. Di Napoli, F. Crescimbini, S. Rodo, and L. Solero, “Multiple input DC–DC power converter for fuel-cell powered hybrid vehicles,” in Proceedings of IEEE Power Electronics Specialists Conference, 2002, pp. 1685 - 1690. [24] L. Solero, A. Lidozzi, and J. A. Pomilio, “Design of multiple-input power converter for hybrid vehicles,” IEEE Transactions on Power Electronics, vol. 20, no. 5, pp. 1007–1016, Sep. 2005. [25] H. Tao, A. Kotsopoulos, J. L. Duarte, and M. A. M. Hendrix, “Multi-input bidirectional DC–DC converter combining DC-link and magnetic-coupling for fuel cell systems,” in Proceedings of IEEE Industry Applications Society Annual Meeting, 2005, pp. 2021–2028. [26] M. Marchesoni and C. Vacca, “New DC–DC converter for energy storage system interfacing in fuel cell hybrid electric vehicles,” IEEE Transactions on Power Electronics, vol. 22, no. 1, pp. 301–308, Jan. 2007. [27] Z J. Duarte, M. Hendrix, and M. Simoes, “Three-port bidirectional converter for hybrid fuel cell systems,” IEEE Transactions on Power Electronics, vol. 22, no. 2, pp. 480–487, Mar. 2007. [28] G.-J. Su and L. Tang, “A reduced-part, ripple-voltage DC–DC converter for EV/HEV power management,” IEEE Transactions on Power Electronics, vol. 24, no. 10, pp. 2406–2410, Oct. 2009. [29] L. K. Gummi and M. Ferdowsi, “Double-input DC–DC power electronic converters for electric-drive vehicles—topology exploration and synthesis using a single-pole triple-throw switch,” IEEE Transactions on Industry Applications, vol. 57, no. 2, pp. 617–623, Feb. 2010. [30] Z. Zhang, Z. Ouyang, O. C. Thomsen, and M. A. E. Andersen, “Analysis and Design of a Bidirectional Isolated DC–DC Converter for Fuel Cells and Supercapacitors Hybrid System,” IEEE Transactions on Power Electronics, vol. 27, no. 2, pp. 848–859, Feb. 2012. [31] H. Matsuo and K. Harada, “New energy-storage DC-DC converter with multiple outputs,” IEEE Transactions on Magnetics, vol. mag-14, no. 5, pp. 1005–1007, Sept. 1978. [32] F. Kurokawa, H. Matsuo, "A new multiple-output hybrid power supply", IEEE Transactions on Power Electronics, vol. 3, no. 4, pp. 412 – 419, Oct. 1988. [33] D. Ma, W.-H. Ki, C.-Y. Tsui, and P. K. T. Mok, “Single-inductor multiple-output switching converters with time-multiplexing control in discontinuous conduction mode,” IEEE Journal of Solid-State Circuits, vol. 38, no. 1, pp. 89–100, Jan. 2003. 75 [34] S. Chakraborty, A. K. Jain, and N. Mohan, “A novel converter topology for multiple individually regulated outputs,” IEEE Transactions on Power Electronics, vol. 21, no. 2, pp. 361–369, Mar. 2006. [35] J.-J. Lee and B.-H. Kwon, “DC–DC converter using a multiple-coupled inductor for low output voltages,” IEEE Transactions on Industrial Electronics, vol. 54, no. 1, pp. 467–478, Feb. 2007. [36] D. W. Kwon and G. A. Rincón-Mora, “Single-inductor-multiple-output switching DC-DC converters,” IEEE Transactions on Circuits Systems–II, Express Briefs, vol. 56, no. 8, pp. 614–618, Aug. 2009. [37] P. Patra, A. Patra, and N. Misra, “A single-inductor multiple-output switcher with simultaneous buck boost and inverted outputs,” IEEE Transactions on Power Electronics, vol. 27, no. 4, pp. 1936–1951, Apr. 2012. [38] V.-L. Tran and W. Choi, “Novel time division multiple control method for multiple output battery charger,” IEEE Transactions on Power Electronics, vol. 29, no. 10, pp. 5102–5105, Oct. 2014. [39] Y. Xi and P. K. Jain, “A forward converter topology with independently and precisely regulated multiple outputs,” IEEE Transactions on Power Electronics, vol. 18, no. 2, pp. 648–658, Mar. 2003. [40] Y. Gu, Y. Kong, S. Nie, and X. Pei, “A simple structure of LLC resonant DC–DC converter for multi-output applications” in Proceedings of IEEE Applied Power Electronics Conference and Exposition, 2005, pp. 1485–1490. [41] H.-S. Kim, J.-H. Jung, J.-W. Baek, and H.-J. Kim, “Analysis and design of a multioutput converter using asymmetrical PWM half-bridge flyback converter employing a parallel–series transformer,” IEEE Transactions on Industrial Electronics, vol. 60, no. 8, pp. 3115–3125, Aug. 2013. [42] R.-J. Wai, and K.-H. Jheng, “High-efficiency single-input multiple-output DC–DC converter,” IEEE Transactions on Power Electronics, vol. 28, no. 2, pp. 886–898, Feb. 2013. [43] A. A. Boora, F. Zare, G. Ledwich, and A. Ghosh, “A new DC–DC converter with multi output: topology and control strategies,” in Proceedings of Power Electronics and Motion Control Conference, 2008, pp. 468–474. [44] A. Nami, F. Zare, A. Ghosh, and F. Blaabjerg, “Multi-output DC–DC converters based on diode-clamped converters configuration: topology and control strategy,” IET Power Electronics, vol. 3, no. 2, pp. 197–208, Mar. 2010. 76 [45] A. Boora, F. Zare, and A. Ghosh, “Multi-output buck–boost converter with enhanced dynamic response to load and input voltage changes,” IET Power Electronics, vol. 4, no. 2, pp. 194–208, Feb. 2011. [46] Y. Zhang and D. Xu, “Design and implementation of an accurately regulated multiple output ZVS DC–DC converter,” IEEE Transactions on Power Electronics, vol. 22, no. 5, pp. 1731–1742, Sep. 2007. [47] Y. Chen, Y. Kang, S. Nie, and X. Pei, “The multiple-output DC–DC converter with shared ZCS lagging leg,” IEEE Transactions on Power Electronics, vol. 26, no. 8, pp. 2278–2294, Aug. 2011. [48] L. K. Gummi and M. Ferdowsi, “Double-input DC–DC power electronic converters for electric-drive vehicles—topology exploration and synthesis using a single-pole triple-throw switch,” IEEE Transactions Industry Applications, vol. 57, no. 2, pp. 617–623, Feb. 2010. [49] A. Di Napoli, F. Crescimbini, F. Guilii Capponi, and L. Solero, “Control strategy for multiple input DC–DC power converters devoted to hybrid vehicle propulsion systems,” in Proceedings of IEEE Power Electronics Specialists Conference, 2002, pp. 1685–1690. [50] B. G. Dobbs and P. L. Chapman, “A multiple-input DC–DC converter topology,” IEEE Power Electronics Letters, vol. 1, no. 1, pp. 6–9, Mar. 2003. [51] O. C. Onar, O. H. A. Shirazi, and A. Khaligh, “Grid interaction operation of a telecommunications power system with a novel topology for multiple-input buck–boost converter,” IEEE Transactions on Power Delivery, vol. 25, no. 4, pp. 2633–2645, Oct. 2010. [52] H. Wu, J. Zhang, and Y. Xing, “A family of multiport buck–boost converters based on DC-link-inductors (DLIs),” IEEE Transactions on Power Electronics, vol. 30, no. 2, pp. 735–746, Feb. 2015. [53] S. Bae and A. Kwasinski, “Maximum power point tracker for a multiple-input Ćuk DC–DC converter,” in Proceedings of International Telecommunications Energy Conference, 2009, pp. 1–5. [54] R. Zhao and A. Kwasinski, “Multiple-input single ended primary inductor converter (SEPIC) converter for distributed generation applications,” in Proceedings of Energy Conversion Congress and Exposition, Sep. 2009, pp. 1847–1854. [55] S.-Y. Yu and A. Kwasinski, “Analysis of Soft-Switching Isolated Time-Sharing Multiple-Input Converters for DC Distribution Systems,” IEEE Transactions on Power Electronics, vol. 28, no. 4, pp. 1783–1794, Apr. 2013. 77 [56] M. Veerachary, “Two-loop controlled buck–SEPIC converter for input source power management,” IEEE Transactions on Industrial Electronics, vol. 59, no. 11, pp. 4075–4087, Nov. 2012. [57] W. Ki and D. Ma, “Single-inductor multiple-output switching converters,” in Proceedings of IEEE Power Electronics Specialists Conference, 2001, vol. 1, pp. 226–231. [58] Y.-M. Chen, Y.-C. Liu, and F.-Y. Wu, “Multi-input DC/DC converter based on the multiwinding transformer for renewable energy application,” IEEE Transactions on Industry Applications, vol. 38, no. 4, pp. 1096–1104, Jul. 2002. [59] H. Wu, R. Chen, J. Zhang, Y. Xing, H. Hu, and H. Ge, “A family of three-port half-bridge converters for stand-alone renewable power system,” IEEE Transactions on Power Electronics, vol. 26, no. 9, pp. 2697–2706, Sep. 2011. [60] Z. Wang and H. Li, “An integrated three-port bidirectional DC–DC converter for PV application on a DC distribution system,” IEEE Transactions on Power Electronics, vol. 28, no. 10, pp. 4612–4624, Oct. 2013. [61] H. Wu, P. Xu, H. Hu, Z. Zhou, and Y. Xing, “Multiport converters based on integration of full-bridge and bidirectional DC–DC topologies for renewable generation systems,” IEEE Transactions on Industrial Electronics, vol. 61, no. 2, pp. 856–869, Feb. 2014. [62] F. Liu, Z. Wang, Y. Mao, and X. Ruan, “Asymmetrical half-bridge double-input DC/DC converters adopting pulsating voltage source cells for low power applications,” IEEE Transactions on Power Electronics, vol. 29, no. 9, pp. 4741–4751, Sep. 2014. [63] D. Maksimović and R. Erickson, “Modeling of cross-regulation in multiple-output flyback converters,” in Proceedings of IEEE Applied Power Electronics Conference and Exposition, 1999, pp. 1066–1072. [64] M. Goldman, A. F. Witulski, “Analysis and prediction of regulation in a multiple-output current-mode controlled DC-to-DC converter,” IEEE Transactions on Aerospace and Electronic Systems, vol. 31, no. 2, pp.626–633, Apr. 1995. [65] J. Sebastian, J. Uceda, M. Rico, M. A. Perez, and F. Aldana, “A complete study of the double forward-flyback converter,” in Proceedings of IEEE Power Electronics Specialists Conference, 1988, pp. 142–149. [66] H. E. Tacca, “Single-switch two-output flyback-forward converter operation,” IEEE Transactions on Power Electronics, vol. 13, no. 5, pp. 903–911, Sep. 1998. [67] H.-J. Chiu, H.-M. Huang, L.-W. Lin, and M.-H. Tseng, “A multiple-input DC/DC converter for renewable energy systems,” in Proceedings of IEEE International Conference on Industry Technology, 2005, pp. 1304–1308. 78 [68] R. Ahmadi and M. Ferdowsi, “Double-input converters based on H-bridge cells: Derivation, small-signal modeling, and power sharing analysis,” IEEE Transactions on Circuits Systems–I, Regular Papers, vol. 59, no. 4, pp. 875–888, Apr. 2012. [69] R. Ahmadi, H. Zargarzadeh, and M. Ferdowsi, “Nonlinear power sharing controller for a double-input h-bridge-based buckboost–buckboost converter,” IEEE Transactions on Power Electronics, vol. 28, no. 5, pp. 2402–2414, May 2013. [70] H. Matsuo, “Comparison of multiple-output DC–DC converters using cross regulation,” IEEE Transactions on Industrial Electronics and Control Instrumentation, vol. IECI-27, no. 3, pp. 176–189, Aug. 1980. [71] C. Ji, K. M. Smith, Jr., K. M. Smedley, and K. King, “Cross regulation in flyback converters: analytic model and solution,” IEEE Transactions on Power Electronics, vol. 16, no. 2, pp. 231–239, Mar. 2001. [72] C.-C. Wen, C.-L. Chen, W. Chen, and J. Jiang, “Magamp post regulation for flyback converter,” in Proceedings of IEEE Power Electronics Specialists Conference, 2001, pp. 333–338 [73] Y.-L. Lin and K. H. Liu, “A new synchronous-switch post regulator for multi-output forward converters,” in Proceedings of IEEE Applied Power Electronics Conference and Exposition, 1990, pp. 693–696. [74] Q. Chen, F. C. Lee, and M. M. Jovanović, “Analysis and design of weighted voltage-mode control for a multiple-output forward converter,” in Proceedings of IEEE Applied Power Electronics Conference and Exposition, 1993, pp. 449–455. [75] A. Ferreres, J. A. Carrasco, E. Sanchis, J.M. Espi, and E. Maset, “Application of a novel parallel regulation technique in a two output forward converter,” in Proceedings of IEEE Power Electronics Specialists Conference, 1999, pp. 914–919. [76] Y.-C. Liu and Y.-M. Chen, “A systematic approach to synthesizing multi-input DC–DC converters,” IEEE Transactions on Power Electronics, vol. 24, no. 1, pp. 116–127, Jan. 2009. [77] Y. Li, X. Ruan, D. Yang, F. Liu, and C. K. Tse, “Synthesis of multiple input DC/DC converters,” IEEE Transactions on Power Electronics, vol. 25, no. 9, pp. 2372–2385, Mar. 2010.