Model Predictive Control With Two-Parameter Identification for Input-Independent-Output-Parallel Dual-Active Bridge Converters

Due to the demands for dynamic performance, model predictive control (MPC) has garnered significant attention in input-independent-output-parallel (IIOP) dual-active-bridge (DAB) converters. However, traditional MPC relies heavily on model parameters, and inaccuracies in the model can lead to control failure. In DAB converters, inductance and capacitance are particularly prominent parameters. This paper proposes an improved MPC method based on recursive least squares (RLS) and least-squares analysis (LSA) for parameter identification of inductance and capacitance, enabling online identification and updating of these values. Experimental results validate that the proposed strategy can eliminate steady-state deviations and ensure effective current sharing.