Globally Unified ZVS and Quasi-Optimal Minimum Conduction Loss Modulation of DAB Converters

Abstract

Due to the quick development of wide bandgap power devices, the power density of the converters is expected to be much higher. In order to increase the power density and efficiency of the DAB converter, especially the efficiency under light load, a quasi-optimal current root-mean-square (RMS) control strategy to achieve All-ZVS is proposed by taking the inductor RMS current optimization algorithm into account in this paper. First the constraint conditions for realizing ZVS of all switching devices under different operating modes are concluded by analyzing the switching mode and the working principle of the DAB converter under the triple phase shift (TPS) modulation. The ZVS domain for All-ZVS operation in each operating mode is further obtained. Then, the GOC strategy is combined with the ZVS condition to derive the novel modulation for all operating modes. Based on this scheme, the switching conditions of all switches can be improved, and the efficiency of the DAB system under the light load is optimized. With ZVS operation, the switching frequency can be raised. Finally, a GaN HEMT based prototype of DAB converter was built and the switching frequency is 200kHz, the effectiveness of the proposed modulation scheme was verified by the experimental results.

A Si IGBT/SiC MOSFET Hybrid Isolated Bidirectional DC–DC Converter for Reducing Losses and Costs of DC Solid State Transformer

“…To address the above problems, extensive research has been carried out on modulation and optimized strategies. By increasing the internal phase-shift angles within the primary or secondary H-bridges of DAB, the modulation strategies of extended phase shift (EPS) [6,7], dual phase shift (DPS) [8,9], and triple phase shift (TPS) [10][11][12][13][14] have been successively proposed, which reduce the conduction losses of the converter and expand the ZVS range with the suitable combination of phase-shift angles. On the other hand, Refs.…”

“…However, the modeling accuracy of the HCA method depends on the maximum order of the Fourier The optimization of the converter can be regarded as a problem of constrained nonlinear optimization, and the effect of optimization depends on the accuracy of the model of converter, the constraints, and the optimized algorithm [20]. The modeling methods for DAB converter are usually divided into two types: time domain analysis (TDA) method [6][7][8]12,14,16,17] and harmonic component analysis (HCA) method [9,15,18,19]. The TDA method can deduce detailed expressions for parameters such as voltage and current based on dividing the operating regions and the operating modes of each region.…”

“…However, the modeling accuracy of the HCA method depends on the maximum order of the Fourier series, and the model is usually not accurate enough under non-unit voltage conversion ratio or asymmetric modulation. The optimized objective of DAB is usually chosen as current stress [8,14,15], rms value of current [7,12,13,18,19], or reactive power [9], and the constraints are usually chosen as ZVS constraints. The choice of optimized algorithm depends on the complexity of the model.…”

“…By comparing existing optimized modulation strategies, it is found that the optimized modulation strategies are mostly optimized for the operating characteristics such as current stress, rms value of current, and switching characteristics in the low-power and mediumpower regions, and still use SPS modulation in the high-power region [12][13][14][15]18,19]. Therefore, the turn-off current is high in the high-power region.…”

“…As a result, SiC MOSFETs are mostly used to reduce deterioration of efficiency in the high-power region. However, the problem of efficiency degradation still exists [12][13][14][15]18,19] and increases the costs of the converter. Utilizing Si IGBT/SiC MOSFET hybrid circuit structure and making Si IGBTs realize low-current turnoff by means of suitable modulation strategy is an effective way to solve the above problems, while there is a lack of related research.…”