Adaptive Variable-Set Model Predictive Control with Error-Dynamics and Current-Stress Awareness for Dual-Active-Bridge Converters

Dual-Active-Bridge (DAB) DC-DC converters are widely used in modern power electronics applications due to their bidirectional power flow, high efficiency, and electrical isolation. However, traditional linear controllers are highly sensitive to parameter variations, which can degrade control performance under changing operating conditions. In addition, conventional Finite Control Set Model Predictive Control (FCS-MPC) methods may result in relatively high current and voltage ripple under steady-state conditions. Furthermore, existing Moving Discrete Control Set (MDCS-MPC) strategies can only partially address these issues because they have a fixed or limited control set. To address these problems, this paper proposes an Error-Dynamics and Current-Stress-Aware Adaptive Variable-Size Moving Control Set MPC (ED-AVCS-MPC) for DAB converters. In this article, using the example of a single-phase shift-operated dual-active-bridge converter, the proposed control design is explained in detail and analyzed thoroughly. The proposed strategy employs a composite activation index that continuously evaluates both the magnitude and the rate of change in the voltage error. Thus, the control set size (3, 5, or 7 elements) is dynamically adjusted via a hysteresis-based switching mechanism. A five-term dynamically weighted cost function is formulated to simultaneously penalize tracking deviation, voltage variation, excessive control effort, current stress, and boundary violations. Through comprehensive simulation studies, the superiority of the ED-AVCS-MPC over traditional PI and fixed-set MDCS-MPC approaches has been confirmed.