Data-Driven Deadbeat Control for Energy Storage Systems

In recent years, Deadbeat Control (DBC) has gained recognition as an effective method for controlling Energy Storage Systems (ESS). However, traditional DBC often exhibits substantial steady-state errors attributed to inaccurate system modeling and losses. This paper introduces a novel approach employing Deep Reinforcement Learning (DRL) for model-based control to accurately generate current references by compensating for power losses. The proposed method is evaluated through simulation results, highlighting its effectiveness in addressing the limitations of traditional DBC.