双碳背景下GCN和DRL的含高比例可再生能源消纳约束的主动配电网经济调度策略OA

In order to achieve the dual carbon goals and improve the consumption rate of renewable resources in ac-tive distribution network(ADN),while enhancing the economic operation of distribution system scheduling,this paper proposes a GCN-DDPG strategy that integrates graph convolutional network(GCN)and deep deterministic policy gradient(DDPG)algorithms.The strategy constructs an economic dispatching model for photovoltaic,load,and energy storage systems with the objective of minimizing the daily operating cost of ADN.The ADN scheduling problem is formulated as a Markov decision process(MDP)in deep reinforcement learning(DRL),where the state space,action space,and reward function of the system are defined.Furthermore,the DDPG algorithm is applied in DRL.The GCN is incorporated to enhance the ability of agent to represent the graph data of distribution network.The effectiveness and robustness of the model are verified in the modified IEEE 33-bus simulation system.Com-pared with the DDPG algorithm,the scheduling cost is reduced by 29%.The scheduling costs of DDPG increase by 27%,45%,and 51%,respectively in comparison with different topologies.