The online diagnosis of internal insulation fault of capacitance voltage transformers based on machine learning

Abstract

Abstract In order to diagnose internal insulation faults of CVT in time, an online diagnosis method for the internal insulation performance of CVT is proposed in this paper. The relationship between CVT insulation performance and measured data is obtained by analyzing CVT internal insulation structure and transfer function. After constructing an appropriate eigenvector by the inter-group correlation amplitude parameter of CVT, three methods of SVM, Bayes, and KNN are used as classifiers for insulation fault diagnosis. The experiment result shows that KNN method can effectively diagnose internal insulation faults with an accuracy rate of 100%.