Fault diagnosis model of power transformer based on improved grey relation method with combined weight

Abstract

Dissolved gas analysis is worldwide accepted as an effective method for identifying incipient fault of power transformer. In order to enhance its accuracy and reliability, fault diagnosis model has been developed based on improved grey relation method with combined weight in this paper. Experts determined weight of dissolved gases for fault diagnosis with analytic hierarchical process and improved grey relation method is used to obviate non-uniform weight distribution caused by expert's subjective factors. Besides, Objective weight is calculated with entropy method. The minimum relative information entropy principle is applied to obtain the combined weight of indices. Then, the combined weights are applied to the proposed method for transformer fault diagnosis. The proposed model has been applied to real database and the results presented in this paper clearly indicate relative high degree accuracy over conventional methods of transformer fault diagnosis using DGA.