Power transformer winding fault analysis using transfer function

Abstract

Condition monitoring of power transformers is essential to maintain the reliability of electricity grid. Frequency response analysis (FRA) is a useful tool for the detection of incipient mechanical faults within power transformers. There are various methods of evaluating the frequency spectrum for fault classification and quantification. This paper presents a useful method to detect and to identify various winding faults of a power transformer based on its FRA signature. Simulation results show that location of resonance frequencies which are the key parameters in identifying various winding deformations are uniquely altered due to different faults.