Winding Fault Detection in Power Transformers Using Multiconductor Transmission Line-Based Model of the Winding

Abstract

Power transformer windings can be damaged by internal faults, such as short-circuit currents and the forces they generate, which can lead to, e.g., internal shorts due to damaged insulation. If these faults are not detected early, they can lead to catastrophic failures. In general, distributed-parameter models can simulate the performance of a transformer in both time and frequency domains more accurately than conventional lumpedparameter models, especially at higher frequency range (> 1 MHz). In this work, a multiconductor transmission line (MTL) model is used to simulate transformer windings in the frequency domain. Two typical winding configurations, continuous disk and interleaved disk, are studied. Internal shorts and buckling faults are applied to both models, and their frequency responses are compared using the frequency response analysis (FRA). The results show that each fault causes a noticeable change in the resonant frequency and admittance curve, except in the case of buckling in the interleaved-disk winding. The results demonstrate that combining MTL simulations with FRA can effectively identify the type of internal faults in transformer windings, offering a more comprehensive diagnostic tool than conventional frequency-domain analyses using lumped-element models.