Characterization of partial discharge signals using wavelet and statistical techniques

Study of partial discharge (PD) behavior in electrical apparatus is important to know the degradation of insulating materials. The characterization of these pulses in the form of cavity discharge, corona discharge and surface discharge is important to identify the faulty location and to quantify the degree of deterioration. A laboratory study is done by making the models of these discharges. Both time and frequency domain measurements were done. The necessary interfacing electronics to minimize the 50 Hz and harmonics from the laboratory power supply is developed. Wavelet signal processing is used to recover the PD signal by eliminating the noises of many natures. Furthermore, different wavelet filters and windowing techniques are used to improve the efficiency of PD signal extraction. With the fabricated models to create only a type of discharge, the statistical characteristic of that type of discharge is identified. A significant number of indicators are got to identify the type of discharge.