Deep reinforcement learning based research on low‐carbon scheduling with distribution network schedulable resources

Reducing carbon emissions is a crucial way to achieve the goal of green and sustainable development. To accomplish this goal, electric vehicles (EVs) are considered system-schedulable energy storage devices, suppressing negative impact randomness fluctuation renewable on system's operation. In paper, coordination control strategy aimed at minimising distribution network between EVs, static var compensators (SVCs) proposed. A model-free deep reinforcement learning (DRL)-based approach developed learn optimal with constraint avoiding system overload caused by random EV access. The twin-delayed deterministic policy gradient (TD3) framework applied design method. After model completed, neural can quickly generate real-time low-carbon scheduling according operating situation. Finally, simulation IEEE 33-bus verifies effectiveness robustness On premise meeting charging demand vehicles, method optimise operation controlling charge-discharge process effectively absorbing in reducing