Transfer function analysis using STFT for improvement of the fault detection sensitivity in transformer impulse test

Insulation of transformer windings may shift as a result of short circuit current or impact during transportation. The shift modifies the dielectric space between the layers of the windings and may cause an insulation breakdown, leading to a complete transformer failure. As transformers are very costly to replace, it is important that their condition determined accurately without having to dismantle the apparatus to visually inspect it. Testing of winding insulation is performed by using the standard impulse test, via applying fast Fourier transform (FFT) to analyze the transformer state (healthy or faulty) in the frequency domain, as a transfer function (TF). Nonetheless, one of the shortcomings of the FFT is that it cannot be used with non-stationary signals. Voltage and current waveforms in the transformer are treated as nonstationary signals, especially when there is a fault. In addition, FFT does not give any information on the time at which a frequency component occurs. To obtain better signature analysis and to increase the detection sensitivity, this paper suggests a new method using short time Fourier transform (STFT) in the transfer function analysis. It is hoped that this high resolution method will help to reduce the subjective judgments of technicians when making decisions about changes in the winding structure of the transformer.