Real-Time Diagnosis of Winding Axial Displacement in Power Transformers Using Time and Frequency Responses

Abstract

The importance of a power transformer lies not only in its capital cost but also in the financial losses incurred due to the lack of supplied energy during its failure. Fault diagnosis in power transformers through online monitoring has gained significant attention over the past decade. The transfer function method has been increasingly employed for fault diagnosis in power equipment, particularly for detecting winding deformations in transformers. This paper investigates the online diagnosis of winding axial displacement in power transformers using time and frequency domain responses. The sensitivity of each method for detecting axial displacement is compared. Among the various criteria, the peak amplitude in the step response at the first extremum was selected as the most effective criterion, although other criteria showed slightly higher sensitivity. The monotonic variation of this criterion with respect to axial displacement played a significant role in this selection.