SPECTRUM ANALYSIS OF PARTIAL DISCHARGE SIGNALS

Partial Discharge (PD) is the major cause for insulation degradation. These discharges occur due to presence of voids or cracks or impurities inside the insulation. There are different types of partial discharges occurring in the insulation system depending on the nature and location of source of discharge. Commonly observed discharge types in rotating machines are delamination, void discharge, end winding discharge, surface discharge etc. Different discharge sources produce different Phase Resolved Partial Discharge (PRPD) patterns. Physical processes taking place at different locations with different intensities during occurrence of partial discharges are responsible for such varying behaviour. The PD detector gives the PRPD data. The values of number of PD pulses (N) and Apparent Charge (Q) from this data is used to calculate effective charge. This effective charge is used to obtain the spectrum plots for delamination type of discharge with increasing applied voltage. Simple Digital Signal Processing (DSP) based tools are used to obtain the frequency spectrum from the PD data. It is observed from these results that the even though the amplitude of the effective charge increases for increase in applied voltage, the spectrum of the PD remains same. This paper proposes spectrum analysis as one of the method for PD source identification.