EVALUATION OF DIGITAL-FILTERS FOR REJECTING DISCRETE SPECTRAL INTERFERENCE IN ON-SITE PD MEASUREMENTS

While Partial Discharge (PD) measurements are widely used in testing power apparatus after manufacture, there is now a trend to extend them to on-site measurements. The major problem encountered in the latter measurements is the strong coupling of external noises particularly from discrete spectral interferences (DSI) e.g. broadcasting stations as well as impulsive disturbances. A critical study of the performance of several digital filters for rejecting DSI is reported. The filters are evaluated with particular reference to distortion introduced on successive PD pulses. A new digital filtering method based on a cascaded 2nd order IIR lattice notch filter has been proposed. The study showed that, for similar performance, methods based on linear prediction required longer filter orders when compared to cascaded IIR notch filters. The non-adaptive methods were found to be stable against impulsive disturbances while adaptive methods were unstable. The new method proposed in this paper gave the best performance, low distortion to a sequence of PD pulses, good stability and fast filtering time. The performance at higher sampling rates, even with PD pulses of long widths was also found to be superior. Further, the method is implementable using fixed-point arithmetic leading to faster processing. Hence, the proposed method can be suggested for on-site PD measurements.