Research on multi-objective optimization strategy of three-level dual active bridge converter based on model predictive control

Abstract

To address the problems of hard turn-on of power switches, poor dynamic performance and large backflow power in the three-level dual active bridge (3L-DAB) DC-DC converter, a multi-objective optimization strategy based on model predictive control (MPC) is proposed in this paper. First, the switching process of power switches under dual phase-shift (DPS) modulation is analyzed, and the mathematical model of power transmission for the three-level dual active bridge (3L-DAB) converter is established. Taking the stable tracking of output voltage as the core objective, the model predictive control (MPC) model of the converter is constructed based on state-space equations, and an error correction method is introduced as the feedback correction link to eliminate the steady-state error caused by parameter mismatch. Second, the power regions are divided by combining the zero-voltage switching (ZVS) turn-on conditions and the backflow power optimization criteria. The optimal phase-shift angles are solved according to the operating characteristics of different regions, and a comprehensive optimization control strategy that takes into account ZVS turn-on, backflow power suppression and dynamic performance improvement is formulated. Finally, the experimental verification results show that the proposed control strategy can achieve ZVS turn-on of power switches, effectively reduce backflow power and improve the dynamic performance of the system.