Deep Reinforcement Learning Approach for EV Charger Control Under Load Variation

In this paper, we propose an electric vehicle (EV) charger controller using a deep reinforcement learning algorithm to address load variations. The gain curve of the EV charger is derived to analyze its electrical characteristics. Subsequently, the deep deterministic policy gradient (DDPG) method is employed to optimize the control gains of PI controller. Simulation results demonstrate that the proposed method improves ripple reduction performance compared to a conventional PI controller under load resistance variations.