Online Reactive Power Optimization of Distribution Network with Soft Open Point Based on Hierarchical Reinforcement Learning

Abstract

Aiming at the problem of large voltage fluctuation and high network loss caused by high proportion of distributed energy grid connection, combined with the coordinated control of soft open point (SOP) and reactive power compensation equipment, a reactive power optimization operation scheme of distribution network based on task hierarchical reinforcement learning is proposed. Based on the MaxQ value function decomposition method, this scheme constructs the online reactive power optimization framework of two-level tasks for reactive power optimization of distribution network, and realizes the real-time operation optimization of traditional reactive power compensation equipment with SOP. Finally, the improved IEEE 33-bus distribution system is taken as an example to verify that the proposed method has better performance.