Optimum threshold estimator for de-noising partial discharge signal using wavelet transform technique

Condition monitoring of electrical power equipment is a vital step in extending the lifetime of inservice equipment which are quite aged. Among the various condition assessment techniques partial discharge (PD) monitoring has emerged as one of the useful technique for condition assessment of high voltage equipment. However the major challenge during PD measurement at site is that PD signals are severely affected by external noises and disturbances like white noise, random noise, Discrete Spectral Interferences (DSI), which are generated due to broadcasting stations, stochastic noise and pulses from power electronics at site conditions. Extracting PD signals from these noises is a challenging task. Several methods have been proposed by researchers for de-noising PD signals. This paper proposes a new method of threshold estimation for de-noising PD signals using wavelet transform (WT) technique. The threshold estimation is done by considering the entire range of interferences thereby eliminating most of the noise without affecting the original signal. This method was used to de-noise simulated PD signals and the results show that the new threshold estimator produces better results when compared to the existing methods. The performance of the proposed method is evaluated using four different parameters namely, Signal to Noise Ratio, Cross Correlation Coefficient, Pulse Amplitude Distortion and Mean Square Error where it consistently outperformed the existing methods. The results are presented and discussed.