A model-free optimal operation strategy of diversified demands-park integrated energy system considering energy cascade utilization

Abstract

In order to improve the energy efficiency of park integrated energy system, this work proposes a model-free optimal operation strategy of diversified demands-park integrated energy system considering energy cascade utilization. First, by considering diversified demands of various users for energy, an energy cascade utilization mode and its corresponding multi-dimensional energy supply and demand balance equation is developed. Moreover, in view of the above structure, a model-free DD-PIES optimal operation strategy is further proposed. Among them, the model-free approach is realized on the basis of the interaction between agent and environment, combined with sufficient exploration and consequences learning of its behavior through experience. Compared to the previous works, this proposed strategy has good self-learning ability for multiple uncertainties of renewable energy and energy demand, and cuts down on inefficient exploration during the training process, while guarantee the security during the execution of actions by the agent. Finally, the rationality and validity of the proposed strategy in reducing operating cost and improving energy efficiency of the system are verified by a case study.