Graph convolution deep network-based rapid calculation method for reactive power reserve demand of power grid

The invention discloses a rapid calculation method for a reactive power reserve demand of a power grid based on a graph convolution deep network, and the method comprises the steps: collecting power flow section data of the power grid, and obtaining bus feature information and topological information; calculating a section comprehensive distance based on dual-scale similarity; reducing the redundant power flow section data sample; extracting section feature data to obtain final sample data; and training and testing to complete graph convolution deep network learning, obtaining a reactive power reserve demand calculation model, and carrying out rapid calculation on the reactive power reserve demand of the power grid. The method can improve the calculation speed of the reactive power reserve demand of the power grid, and solves the problem of high calculation complexity and time consumption of the reactive power reserve demand of the power grid. 本发明公开了基于图卷积深度网络的电网无功储备需求快速计算方法，包括：采集电网潮流断面数据，获取母线特征信息和拓扑信息；计算基于双尺度