Research On Fault Diagnosis And Reconfiguration Optimization Of Distribution Network Based On Graph Neural Network

With the increasing demand for electrical energy in people’s daily production and life,the density and complexity of the distribution network are also increasing.A line failure in the distribution network will affect the normal power supply of the distribution network system,thereby affecting the normal production and life of the people and causing huge economic losses.How to quickly locate the fault location in the distribution network and how to dynamically adjust the network topology of the distribution network according to the system status after the fault are of great significance to the safety and reliability of the distribution network.In order to solve the problem of fault location in the distribution network,a fault location model GAE-GCN based on graph convolutional neural network is proposed.The model extracts the physical properties of the lines in the distribution network as feature information by considering the current and voltage of the lines in the distribution network.And then reduce the noise and dimensionality of the feature through the Graph AutoEncoder(GAE),and then input into the Graph Convolutional Network(GCN)for failure point classification.This model overcomes the shortcomings of traditional fault location methods that do not fully consider the distribution network topology,and can better improve the reliability and accuracy of the model.In order to solve the problem of dynamically adjusting the network topology after a fault in the distribution network,a model LSTMA3 C based on the combination of Long Short Term Memory(LSTM)and A3C(Asynchronous advantage actor-critic)algorithm is proposed.The algorithm overcomes the problems of "combination explosion" of the switch state of the traditional distribution network reconfiguration method.In order to verify the effectiveness of the proposed algorithm,tests were conducted on the IEEE 123 bus and IEEE 14-node system.The test results show that the fault location model of the distribution network based on the GAE-GCN algorithm and the distribution network reconfiguration model based on the LSTM-A3 C are superior to other models.