基于改进图卷积神经网络的配网线路故障诊断

随着电网的不断扩容，系统结构越来越复杂，多故障频发，而多故障是故障诊断的关键和难点。为解决故障处理数据量大，需要快速、准确地诊断电网故障的问题，本文提出了一种基于模糊优化图卷积神经网络的配网故障诊断模型。首先处理采集的配网故障线路的特征数据；其次，搭建基于图卷积神经网络的故障诊断模型，利用模糊理论建立配电网故障的隶属函数；最后利用训练好的模型进行配网故障诊断。仿真结果表明，模糊优化图卷积神经网络对多故障诊断的准确率高于卷积神经网络以及其他方法，本文方法做出的诊断结果更加精确，综合诊断效果最好。

Distribution Network Line Fault Diagnosis Based on Improved Graph Convolutional Neural Network

With the continuous expansion of power grid, the system structure becomes more and more complex, and multiple faults occur frequently, which is the key and difficult point of fault diagnosis. In order to solve the problem of the large amount of fault processing data and the need for fast and accurate fault diagnosis of power grid, a fault diagnosis model of distribution network based on fuzzy optimized graph convolution neural network is proposed in this paper. Firstly, the collected characteristic data of distribution network fault lines are processed. Secondly, a fault diagnosis model based on graph convolutional neural network is built, and the membership function of distribution network faults is established by fuzzy theory. Finally, the trained model is used to diagnose the distribution network faults. The simulation results show that the accuracy of fuzzy optimized graph convolutional neural network for multi-fault diagnosis is higher than that of convolutional neural network and other methods. The diagnosis results of the proposed method are more accurate and the comprehensive diagnosis effect is the best.