Multi-objective combined dispatching optimization of pumped storage and thermal power

Abstract To realize precise and stable power grid dispatching, it is urgent to realize the combined scheduling of Pumped Storage and Thermal Power (PSTP). Considering some factors in the power system, a multi-objective combined optimization model of PSTP is proposed in this paper. Then, this paper proposes an improved multi-objective optimization algorithm base on deep reinforcement learning to solve the above model. Finally, this paper analyzes the advantages and disadvantages of the proposed multi-objective optimization model and finds that the model proposed in this paper sacrifices certain economic benefits but can reduce the system’s fluctuation and improve the power system’s reliability.