GIS Partial Discharge Type Identification Based on Optimized Support Vector Machine

The fault diagnosis of partial discharge in GIS is of great significance to the evaluation of the operation status of GIS. The traditional pattern recognition algorithms include neural networks and support vector machines (SVM). However, the isolated data points in the data samples will cause over-fitting during classifier training because of the existence of interference signals. The training error was continuously reduced and the complexity of the model was continuously improved, which eventually leads to an increase in generalization error and a decrease in recognition rate. In order to solve these problems, four partial discharge models were established and detected by UHF method to extract characteristic parameters. Besides, SVM was optimized by choosing appropriate kernel function, adopting L2 regularization and principal component analysis (PCA). The recognition accuracy of 96.5% can be achieved. The recognition speed was improved by applying PCA. The results show that the optimized SVM algorithm can improve the efficiency of pattern recognition.