Multi-objective DDPG Optimal Dispatch for Low-voltage Distribution Station Area Flexible Interconnection System
Aiming at strong uncertainty of source,load and equipment and the power mutual aid characteristics between substation areas in distribution station area flexible interconnection system(DSAFIS),a coordination and optimization dispatch method based on deep deterministic policy gradient(DDPG)for the operation cost,new energy consumption,load balancing is proposed.A deep reinforcement learning day-ahead optimal dispatch decision framework with automatic linkage between system model and physical system is constructed.The optimal dispatch DDPG model considering multi-objective reward and operation constraint reward is designed,DDPG adopts an online-learning mode,the day-ahead dispatch plan is output to the actual DSAFIS after the algorithm converges.An example is given to verify that the proposed method can automatically adapt to the strong system uncertainty,and can reduce the operating cost while taking into account the new energy consumption and the load balance of substation area.
distribution station area flexible interconnection systemday-ahead optimal dispatchDDPGmulti-objectiveload balancing
万方数据
维普
