eduzhai > Applied Sciences > Engineering >

Model-Free Voltage Regulation of Unbalanced Distribution Network Based on Surrogate Model and Deep Reinforcement Learning

  • Save

... pages left unread,continue reading

Document pages: 8 pages

Abstract: Accurate knowledge of the distribution system topology and parameters isrequired to achieve good voltage controls, but this is difficult to obtain inpractice. This paper develops a model-free approach based on the surrogatemodel and deep reinforcement learning (DRL). We have also extended it to dealwith unbalanced three-phase scenarios. The key idea is to learn a surrogatemodel to capture the relationship between the power injections and voltagefluctuation of each node from historical data instead of using the originalinaccurate model affected by errors and uncertainties. This allows us tointegrate the DRL with the learned surrogate model. In particular, DRL isapplied to learn the optimal control strategy from the experiences obtained bycontinuous interactions with the surrogate model. The integrated frameworkcontains training three networks, i.e., surrogate model, actor, and criticnetworks, which fully leverage the strong nonlinear fitting ability of deeplearning and DRL for online decision making. Several single-phase approacheshave also been extended to deal with three-phase unbalance scenarios and thesimulation results on the IEEE 123-bus system show that our proposed method canachieve similar performance as those that use accurate physical models.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×