Mobile battery energy storage system control with knowledge‐assisted deep reinforcement learning

Most mobile battery energy storage systems (MBESSs) are designed to enhance power system resilience and provide ancillary service for the operator using storage. As penetration of renewable fluctuation electricity price increase in system, demand-side commercial entities can be more profitable utilizing mobility flexibility MBESSs compared stational system. The profit is closely related spatiotemporal decision model influenced by environmental uncertainties, such as traffic conditions. However, solving real-time control problem considering long-term uncertainties time-consuming. To address this problem, paper proposes a deep reinforcement learning framework maximize through market arbitrage. A knowledge-assisted double Q network (KA-DDQN) algorithm proposed based on learn optimal policy efficiency. Moreover, two criteria action generation methods integer actions scheduling short-term programming results. Simulation results show that method achieve result, KA-DDQN accelerate process original approximately 30%.