Generalized Extended State Observer-Based Model Predictive Control of On-Board Electric Vehicle Charger

In this article, a finite control set model predictive controller (FCS-MPC) based on a generalized extended state observer (GESO) is proposed for a single-phase current source rectifier (CSR) used in electric vehicle (EV) chargers. The proposed approach involves creating the rectifier model considering parameter uncertainties, including grid inductance, filter capacitance, and unmodeled dynamics in the rectifier's output. To estimate these factors, a GESO is utilized. The FCS-MPC is then designed using the estimated states to generate the activation pulses for the converter switches. The proposed method compensates for calculation delay using control input prediction, and the system model estimation is employed in the next two steps to predict the control input. Obtaining the system model in the next two steps necessitates estimating the disturbances in the next step, which are the variables added to the system through GESO. The validity of the proposed method has been confirmed through MATLAB simulation and a laboratory prototype.