A graph neural network-based topology identification and reconfiguration method for electrical and electronic power distribution systems in a big data environment

Accurate distribution system topology is of great significance for distribution network planning operation and analysis.This project constructs a distribution system network model, applies graph convolutional network and graph attention network in graph neural network, and designs the topology identification method of distribution system.On this basis, a reconfiguration model of the distribution system is given, and the network structure after topology identification is used for trend calculation, and the model reconfiguration is realized by using the extensive learning quantum evolutionary algorithm.Through experimental analysis of several test systems, it is found that the topology identification F1 values of this paper's method are all above 0.9, which are 5.64% to 29.64% higher than other methods, confirming the good accuracy and robustness of the GNN topology identification model.In addition, the CLQIEA method can give the correct distribution system reconfiguration optimization scheme, which reduces the network loss to a larger extent and improves most of the node voltage values, and the network loss decreases by 31.91% and 56.11%, and the voltage values are improved by an average of 1.95% and 1.23% in the two test systems, which makes the power supply of the distribution system of a higher quality, and the operation of the power supply system is more economical, which is important for the distribution automation and the power supply department's optimal scheduling is of great significance.