Study on Discrimination between Inrush and Fault in Transformer: ANN Approach

Abstract

Transformer protection is a critical issue in power systems because it involves accurately and quickly distinguishing magnetising inrush current from internal fault current. An artificial neural network has been proposed and demonstrated its ability to solve the transformer monitoring and fault detection problem using a low-cost, dependable, and noninvasive procedure. This paper presents an algorithm in which statistical parameters of detailed d1 level wavelet coefficients of signal are used as an input to the artificial neural network (ANN), which develops into a novel approach for online detection method to discriminate the magnetising inrush current and inter-turn fault, as well as the location of fault, i.e. whether the interturn fault lies in primary or secondary winding, using discrete wavelet transform and artificial neural network (ANNs). In the laboratory, information from controlled experiments was collected using a custom-built single-phase transformer. Following feature extraction with the discrete wavelet transform (DWT), a neural network model MLP was designed and rigorously trained. It is also discussed the proposed on-line detection scheme.