Performance Enhancement of EV Drive System under Open-Phase Fault Using Reinforcement Learning with MPC

Improving the operation of electric vehicles (EVs) under fault is a very important subject because it increases their ability to operate in case of an emergency. This paper proposes a fault-tolerant control methodology for an EV drive system under open phase fault (OPF). The 5-ph interior permanent magnet synchronous motor (IPMSM) is employed in the drive system because it has several merits, such as high efficiency and reliability. The proposed technique is based on utilizing a reinforcement learning (RL) control algorithm, based twin-delayed deep deterministic policy gradient (TD3) algorithm, to operate the motor at maximum torque per ampere under OPF. This technique is compared with the PI-online tuning method that employs fuzzy logic control (FLC). The torque ripples are minimized using the model predictive control with a constructed cost function. The European drive cycle (ECE-15) is used to test the performance of the proposed technique. This technique is verified by obtaining simulation results using the MATLAB Simulink package. The results of the simulation indicate that the proposed control strategy results in superior performance as it achieved lower values of mean square error, integral square error, and percentage overshoot compared with the classical PI online tuning using FLC.