Search

Stochastic Computing (SC) offers a promising computing paradigm for low-power and cost-effective applications, with the added advantage of high error tolerance. In parallel, Coarse-Grained Reconfigurable Arrays (CGRA) prove to be a highly promising platform for domain-specific applications due to their combination of energy efficiency and flexibility. Intuitively, introducing SC to CGRA would significantly reinforce the strengths of both paradigms. However, existing SC-based architectures often encounter inherent computation errors, while the stochastic number generators employed in SC result in exponentially growing latency, which is deemed unacceptable in CGRA. In this work, we propose an SC-based CGRA by replacing the exact multiplication in traditional CGRA with an SC-based multiplication. To improve the accuracy of SC and shorten the latency of Stochastic Number Generators (SNG), we introduce the leading zero shifting and comparator truncation, while keeping the length of bitstream fixed. In addition, due to the flexible interconnections among PEs, we propose a quality scaling strategy that combines neighbor PEs to achieve high-accuracy operations without switching costs like power-gating. Compared to the state-of-the-art approximate computing design of CGRA, our proposed CGRA can averagely achieve a 65.3% reduction in output error while having a 21.2% reduction in energy consumption and a noteworthy 28.37% area savings.

Modular multi-active bridge (MMAB) converter with inherent advantages of galvanic isolation, modularity and power density, can provide multiple dc ports for multifarious power conversion through high-frequency links (HFLs). However, the traditional single phase-shift control, which is widely applied in the converter due to the complex coupling among ports, leads to unfavorable efficiency performance. To independently handle each port with optimal capabilities, a dual-loop optimal control is proposed for MMAB converters to achieve the decoupling of active and reactive powers. Conditioning circuits are designed to extract power-related information that can be used to eliminate the fundamental reactive power, thus optimizing conducting and switching characteristics. The proposed control method regards each port as a subsystem regardless of the coupling through the HFL, which can be easily extended to multiport systems, which have HFL. Finally, the effectiveness of the proposed reactive-power optimal control strategy is validated on a hardware-in-the-loop-based four-port MMAB converter.

In this article, a time-division multiplexing (TDM) modulation technology is proposed to enhance the efficiency of modular multiactive bridge (MMAB) converters in multiport power electronic transformers over a wide operating range. The TDM modulation retains the simplicity of single phase-shift (SPS) modulation while achieving efficiency improvement by adopting different structures (i.e., half-bridge and full-bridge) under different operating conditions with dc bias capacitors. However, determining the optimal structure under the same operating condition poses a significant challenge due to the strong coupling among MMAB ports. To address this issue, this article analyzes a dual-active bridge converter to investigate the steady-state characteristics of different structures, including the root-mean-square current and zero-voltage switching. By decomposing the MMAB converter into different transmission branches, the optimal combination structure of half-bridge and full-bridge are selected based on the branch powers, forming a simple and efficient TDM modulation technology. Finally, an experimental prototype of a four-port MMAB converter is established and tested. The results demonstrate that the proposed TDM can increase efficiency up to 19.09% compared to SPS, validating the correctness and effectiveness of the theoretical analysis.

With the rapid development of renewable energy, high-frequency bus electric energy router that provides multiple ports for different sources and loads is gaining attention, and the application of redundant modules for electric energy router is necessary to improve the reliability of the system. However, the strong coupling characteristics of the common high-frequency bus structure can cause transient current dc bias and current fluctuation for redundant modules during the entry stage, which may result in overcurrent, device damage, and failed entry of redundant modules. This article proposes a partial module high-frequency current model to analyze the causes of transient current dc bias of redundant modules when entering into normal operation and designs a seamless entry control strategy for redundant modules to eliminate the current dc bias and achieve rapid entry into steady-state operation, which can remove the high-frequency current sampling process to simplify the control system. The feasibility of the proposed control strategy is verified by offline simulation and hardware-in-loop real-time simulation platform.

With the increasing popularity of renewable energy and electric terminal equipment, the large-scale application of isolated bidirectional dc–dc converters (IBDCs) is growing. Among these converters, the dual active bridge (DAB) converter, as the core topology of IBDCs, has garnered significant attention due to its bidirectional power transmission capability, simple modularity, and wide soft-switching range. To enhance the efficiency of DAB, various multi-degree-of-freedom (M-DOF) modulation techniques and related solutions have been proposed. However, there is currently a lack of in-depth understanding and thorough comparison of different M-DOF modulation techniques, making it difficult to obtain the optimal solution for different application scenarios. Therefore, based on the latest research and contributions, this article provides an overview of M-DOF, including models, objectives, and algorithms. Furthermore, the existing typical M-DOF modulation techniques are compared through experiments, exploring the differences and advantages among these technologies. Through these analyses, this article offers robust support for the advancement and application of M-DOF modulation techniques in DAB converters.

In modular multiactive bridge (MMAB) converters, which are inherently coupled among dc ports, single phase shift (SPS) control has been widely adopted due to its simplicity in implementation and neglect of coupling characteristics in control algorithms. However, under SPS control, the converter may have unfavorable system efficiency performance, especially when voltage conversion ratios of ports mismatch. To improve the efficiency, this article proposes a common phase shift (CPS) control method for a generalized MMAB or multiactive bridge (MAB) converter. A comprehensive operation analysis of MMAB converters under CPS control with switching characterization of power semiconductors has been carried out. Besides, an overall power loss model is derived, with main electrical components and their power-loss behaviors taken into consideration. Moreover, an optimal strategy aimed at power loss optimization is proposed to improve the efficiency of the converter. Finally, a four-port experimental MMAB converter prototype has been constructed. The experimental results and tested efficiency curves have verified the feasibility and effectiveness of the proposed CPS control.

Based on the phase-shift modulation, the dual-active-bridge (DAB) dc–dc converter always features the natural self-volt-second balance performance even during the transient process. Besides, different phase-shift ratios usually result in different initial series-inductance currents. Therefore, when the phase-shift ratio is suddenly changed for dealing with the change of required transferred power, a dc offset in the series-inductance current will emerge. Especially for the high-power condition, this dc offset lasts for a long time because of the high-current value but low damping conduction resistor. Then, the magnetic saturation phenomenon may emerge in the transformer, resulting in the potential failure of the converter system. To address this issue, a simple dc-offset eliminating method is proposed for the DAB dc–dc converter. Different from most of the existing methods, the proposed one is independent from the phase-shift modulation method. Therefore, the proposed dc-offset eliminating method can be easily embedded in different phase-shift modulation methods, such as single-phase-shift, double-phase-shift, and triple-phase-shift methods. Finally, the experimental results are provided to verify the effectiveness of the proposed dc-offset eliminating method for the DAB dc–dc converter.