Bilateral cooperative energy management of park integrated energy system based on improved deep deterministic policy gradient algorithm

In order to fully take advantage of the supply-demand bilateral flexibility of the park integrated energy system (PIES) and achieve economical and reliable operation of the system in uncertain environment, a PIES bilateral cooperative energy management method based on improved deep deterministic policy gradient (DDPG) algorithm is proposed. Firstly, a mathematical model is established for the bilateral cooperative energy management problem of PIES that considers both supply and demand dispatchable resources. Secondly, it is formulated as a Markov decision process with continuous state and action space. Then, an improved DDPG algorithm combining priority experience replay mechanism andL2 regularisation is proposed for the bilateral cooperative energy management decision-making. The proposed method can adaptively manage the dispatchable resources on supply side and controllable loads on demand side, while avoiding the influence of uncertain factor s on energy optimisation management decisions. Case studies show that the proposed method can cooperatively manage "source-storage-load" multi-link flexible resources in uncertain environment, effectively improving the economy and flexibility of system operation.