Power Transformer Fault Diagnosis Based on the Graph Convolutional Networks
To improve the accuracy of power trans-former fault diagnosis,a method based on the GCN is pro-posed.This method utilizes the adjacency matrix of GCN to fully represent the similarity measure between unknown and labeled samples,using graph convolutional layers as classifiers to find complex nonlinear relationships between dissolved gases and fault types,employing back propaga-tion algorithms to complete the training process of GCN.Comparative analysis results show that GCN outperforms other existing models under different input features,signifi-cantly enhancing the accuracy of the fault diagnosis.
万方数据
维普
