Dynamic Economic Dispatch for Integrated Energy System Based on Deep Reinforcement Learning

The optimal dispatch of integrated energy systems is of great significance for the realization of multi-energy complementary and economic operation of the system. However, the intermittence of renewable energy and the uncertainty of users' energy demands in the system cause the random fluctuation on both the supply and demand sides in the system. Traditional dispatch methods are difficult to adapt to the dynamic changes of the actual environment accurately. In accordance to this problem, a dynamic economic dispatch method for integrated energy systems considering the time-varying characteristics of renewable energy and heterogeneous loads is proposed. Firstly, the dynamic economic dispatch problem for integrated energy systems is described mathematically. Secondly, the dispatch decision problem is formulated as a reinforcement learning framework, in which the observation state, dispatch action and reward function of the system are defined. Then, deep deterministic policy gradient (DDPG) algorithm is used to make dynamic dispatch decisions in continuous state and action spaces. The proposed method does not need to predict or model the uncertainty, and can dynamically respond to the random fluctuations of the source and loads. Finally, simulation is carried out to demonstrate the effectiveness of the proposed method. © 2021 Automation of Electric Power Systems Press.