Power transformer fault diagnosis based on improved gray clustering analysis
WU Bao-chun · Journal of Jilin University · 2008
A gray clustering algorithm based on Gauss whitening weighting function and a method to choose the parameter σ of the function were proposed.The values of the parameter σ were defined preliminarily after studying the DGA data of 34 groups of the fault power transformers.The experimental analysis demonstrates that the proposed fault diagnosis method leads to the results in good agreement with those by practical detections.This method seems effective and practical,and is charaterized by fast analysis and good real-timeness.