Modified Model Predictive Control for Three-Phase Dual-Active-Bridge DC-DC Converters with Recursive Least Square Estimation

In this paper, a modified moving discretized control set model predictive control (MDCS-MPC) for three-phase dual-active-bridge (3p-DAB) DC-DC converters is proposed to improve the dynamic performance, voltage tracking capacity, and robustness to parameter variation. Firstly, the operating principle of MDCS-MPC for 3p-DAB is introduced. Meanwhile, in this proposed modified MDCS-MPC, a cost function and an adaptive step are designed, which not only provide stiff voltage regulation but also promote the transition performance. Next, a recursive least square (RLS) method is utilized to estimate the leakage inductance of high-frequency transformer with attempt to improve the robustness to parameter variation. Finally, the comparison simulations and the comparison experiments with PI control based on phase shift modulation are carried out to verify the theory claims and the efficiency of the proposed solution.