An Energy Management Strategy for Hybrid Energy Storage System Based on Reinforcement Learning

Due to the continuous high traction power impact on energy storage medium, it is easy cause many safety risks during driving process, such as triggering aging mechanism, causing rapid deterioration of battery performance process and even thermal runaway. Hybrid an effective way solve this problem. The ultracapacitor device that has density, which can withstand instantaneous currents be charged discharged quickly. By combining batteries ultracapacitors in a hybrid system, sources with different characteristics combined take advantage their respective strengths increase efficiency lifetime system. management strategy plays important role systems. Traditional optimization algorithms have difficulty improving flexibility practicality applications. In paper, based reinforcement learning proposed. results indicate proposed method effectively distribute charging discharging conditions supply maintain SOC and, at same time, meet demand working cost less loss realize goal optimizing overall strategy.