Cascade Optimal Operation of Community Integrated Energy System Based on Improved Soft Actor-Critic

Integrated energy system realizes multi-energy complementarity through coupling mechanism between multiple energies. It can significantly improve the efficiency of energy utilization, thereby contributing to low-carbon transformation of global energy system. However, existing research has not fully considered the field of optimal operation of integrated energy system from the perspective of energy supply structure, which is likely to cause energy loss in whole system. In order to solve the above problem, a cascade optimization operation strategy of community integrated energy system based on improved Soft Actor-Critic algorithm is proposed. Firstly, considering diversified demands of community integrated energy system, load is divided according to energy grade, and a cascade optimization model of community integrated energy system is established. Moreover, the cascade optimization operation problem is transformed into a reinforcement learning framework. The corresponding Markov decision process model is constructed and the rationality of action selection is guaranteed by combining domain knowledge rules. Finally, the effectiveness and superiority of the proposed method in improving energy efficiency and reducing operating cost of the system is verified by a case.