Dissolved gases in oil diagnosis based on support vector machine

Abstract

A kind of analysis method in which SVM is used for power transformer DGA is proposed in this paper. This method uses the SVM algorithm to classify the composition of DGA in transformer and diagnoses the fault of the transformer. At the same time, it introduces the fuzzy membership function, and it can eliminate unable diagnosis area when the discrete decision function is used. Then by using a example to test this method, it shows the SVM play excellent performance in the fault diagnosis of power transformer. © the authors.