Multi device fault association recognition in intelligent substations based on spatiotemporal graph convolutional network model

Given the poor accuracy of existing recognition methods in fault localization and fault propagation path recognition, this study focuses on intelligent substation multi device fault association recognition based on spatiotemporal graph convolutional network model. Regard substation equipment as nodes in the graph, treat the connection relationships between equipment as edges, and fuse the time series data of equipment operation status to construct a spatiotemporal graph model. By utilizing graph convolution operations to capture spatial dependencies and combining them with temporal models to extract temporal dynamic features, accurate prediction of fault propagation paths and device association recognition can ultimately be achieved. The test results showed that all four groups of tests had positioning accuracy exceeding 90 times. Based on these data calculations, it can be concluded that the accuracy of fault localization is at a relatively high level; The accuracy of fault propagation path recognition for all four groups reached 100%, achieving high-precision fault propagation path recognition. Enhance the ability of substations to respond to faults and ensure the stable operation of the power system.