Reactive Voltage Control Method Based on Depth Map Reinforcement Learning

With the development of large scale grid connection and large scale operation of UHVDC lines, AC-DC hybrid grid has been formed in China’s power system. Traditional methods based on physical model are not suitable for the complicated power grid, and their interpretation is not good, so they can not be generalized. Graph deep learning provides a new approach to the problem of voltage control and reactive power. DGRL is A Deep Graph Reinforcement Learning. DGRL approach is presented in this paper. The method can be used to analyze the operational environment of the power system, and obtain the useful information for the decision-making of the reactive voltage stability. It can adapt well to topology change scene.