Topologies and Control of Single-Stage AC–DC and DC–AC Converters for Advanced Performance

Abstract

AC–DC and DC–AC converters are used in a wide range of applications. This thesis investigates advanced topologies and control methods for single-phase, single-stage, isolated AC–DC and DC–AC converters. The aim is to advance the performance of these systems in terms of efficiency, cost effectiveness, and power density in the context of real-world applications. Firstly, for low-power applications, such as consumer electronics, a new solution - the active power decoupling integrated active clamp flyback (iACF) converter - is investigated. The topology features a low component count (i.e., two active switches and one transformer) and reduced twice-line frequency power buffer volume (through an active-power decoupling control method), leading to low-cost, high-density, and high-efficiency design. These features make the iACF converter perfect for low-power applications, which are generally sensitive to cost and size. Under the proposed control strategy on a 100-W laboratory iACF prototype, the converter achieves a heavy-load (67 W–100 W) efficiency of approximately 93%, demonstrating its superiority. Secondly, to further improve the iACF converter’s light-load efficiency performance, a new modulation method is proposed. The modulation method combines Continuous-Conduction- Mode (CCM) and burst mode of operation, thereby effectively reducing the system’s operating frequency and switching losses while maintaining all the best features of CCM iACF mentioned above. The same iACF prototype is developed to verify the feasibility of the proposed light-load modulation method, showcasing a four-point average efficiency of 91.9%, which is superior to conventional two-stage solutions.

Thirdly, for medium-power applications, such as micro-inverters and bi-directional energy routers, a real-time control method is investigated for the Dual-Bridge Series Resonant Converter (DBSRC) that can dynamically maximize the system efficiency while keeping output power regulated. Here, the DBSRC-based DC–AC converter is selected due to its simple topology and potentially high efficiency at medium power level. Conventionally, real-time control for a DBSRC-based DC–AC converter is challenging. This is mainly because simultaneous optimisation for multiple control freedoms of a DBSRC-based DC–AC converter is needed, and that the instantaneous operating states of the converter are varying. These factors often require high computing power for optimal performance control. In this thesis, the optimal control condition is obtained analytically and expressed in a simple closed form, allowing us to achieve real-time control of the DBSRC-based DC–AC converter. The effectiveness of the proposed control strategy is validated through comparative analysis against Single-Phase-Shift (SPS) control based on a DBSRC-based DC–AC converter prototype built in the lab, demonstrating over 2% improvement in efficiency. Finally, for high-power applications, such as electric vehicle (EV) onboard chargers, a novel semi-single-stage bridgeless Star power-factor-correction (PFC) architecture is proposed. This new architecture hybridizes the traditional two-stage converter architecture in such a way that it operates alternatively between a single-stage converter and a traditional two-stage converter.

By intelligently 'borrowing' the inductor/transformer current from the second-stage DC–DC back- end, the converter achieves Zero Voltage Switching (ZVS) conditions for all switches in the PFC front end while maintaining CCM operation. This innovative approach minimizes both conduction and switching losses - a paradox in traditional two-stage topologies. The merits of the proposed architecture are validated through a scaled-down 240-W GaN-based laboratory prototype, achieving a full-load efficiency of 96.1% and a power density of 50 W/in³ uncased. Through these contributions, the thesis advances the topology and control of single-stage AC–DC and DC–AC converters, providing solutions that improve efficiency, reduce component size, and enhance performance across a wider range of operating conditions.