Fault Identification Method for the Whole Map of Station Area Considering Node Correlation

With the introduction of digital grid, transparent grid, and smart grid concepts, a large number of sensors are used in power systems. As the "last mile" of the power grid, the safe operation of the low-voltage distribution network is directly related to the personal safety of power users. In the traditional low-voltage distribution network, fault identification is reported by customers combined with manual on-site inspection, which is time-consuming and laborious. At the same time, the danger has increased. For this reason, this paper proposes a short-circuit identification method for low-voltage distribution network based on the classification and identification of the whole map, combined with the operation data of the low-voltage distribution network. Firstly, users in the station area are used as nodes to automatically generate the topological adjacency matrix of the station area; secondly, the intelligent equipment of the terminal in the low-voltage station area is used to collect and transmit the characteristic data when a fault occurs to the main station, and jointly form graph data in the adjacency matrix; Finally, input the data of all nodes in the station area to the graph convolutional neural network, obtain the feature representation of the station area through average graph pooling, and identify the short circuit type of the station area. Finally, it is verified by simulation that this method has an accuracy rate of 97% in the research and judgment of short-circuit fault types in low-voltage stations, which is effective.