Enhanced Inverse Model Predictive Control for EV Chargers: Solution for DC–DC Side

Abstract

This article presents an approach for enhancing the reliability and robustness of electric vehicle (EV) chargers, particularly the dc–dc side of the EV chargers, by using the inverse model predictive control (IMPC). IMPC, a recently introduced control method for power electronic converters, leverages the strengths of model predictive control (MPC), while minimizing its computational burden. IMPC excels in managing sophisticated and nonlinear systems, controlling multiple objectives, and adhering to various constraints. However, the effectiveness of conventional IMPC is heavily dependent on the accurate dynamic model of the power converter. This dependency makes IMPC susceptible to uncertainties and disturbances. To address this challenge, the proposed method employs an adaptive estimation strategy utilizing a recursive least square algorithm for online dynamic model estimation. This real-time estimated model enables IMPC to predict optimal switching states with improved reliability. The proposed control technique is designed to provide constant power, constant current, and constant voltage modes, with the ability to seamlessly transition between them. The efficacy of this technique is demonstrated through extensive simulations and experimental validation for a dual active bridge (DAB) converter. This adaptive method underscores the potential of IMPC for practical EV charging scenarios, ensuring reliable and high-performance charging.