Fault diagnosis of power transformer based on clustering binary tree SVMs

Abstract

Conventional transformer fault diagnosis model is based on the principle of empiric risk minimization which will result in a fall of generalization and low accuracy of fault diagnosis. Support vector machine which is based on the principle of structural risk minimization and cluster technique have been introduced into transformer fault diagnosis. A SVM based multilevel binary tree transformer fault diagnosis model has been established. Adaptive k-means clustering algorithm is put forward to resolve multi-class problem. With the completion of sub-SVM training, the structure of the model is achieved. A great deal of transformer fault diagnosis tests have been done to compare the diagnosis accuracy of the model with different kernel functions and obtain the appropriate kernel function.