Diagnostic Research for the Failure of Electrical Transformer Winding Based on Digital Twin Technology

Abstract

Transformer is a key equipment for constructing power system, and its safe and stable operation is the basis of power system. In order to solve the problems of difficult identification and low detection accuracy of transformer winding at operating state, in this paper, a diagnosis method based on digital twinning technique for transformer winding failure is proposed. Firstly a digital twin transformer model with high simulation degree based on transformer entities was built, and then multiple physical field simulation were used to deduce the information change of oil tank surface vibration information of digital twin transformer under different working conditions, various winding failures. Here, the vibration signal on oil tank is effectively decomposed to extract the feature vector by independent component analysis method and wavelet packet transform, which in turn is based on neural network algorithm to learn and diagnose the failure signal of winding. In order to test the accuracy and feasibility of the method, the diagnostic test of oil tank vibration signal detection of health and fault transformer is carried out on an experimental prototype of transformer with 110 kV, and the diagnostic results show that this method can diagnose the problems occurred in transformer windings with high efficiency and accuracy. © 2022 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC.