De-noising of partial discharge signals using second generation wavelet transform

Partial discharge (PD) measurement is used for condition monitoring of insulation in high voltage equipments. Extraction of PD pulses buried in noise and interferences is a challenging task encountered in PD measurement. Recent research shows that second generation wavelet transform (SGWT) can potentially be employed for extracting PD pulses from severe noise and interferences. In this work, an SGWT based de-noising scheme is proposed for extracting PD pulses from severe noise and interferences. The thresholding rule adopted in this de-noising scheme is an extension of the reconstructed time-domain components based thresholding method. To verify its effectiveness, the proposed de-noising method is applied on two numerically simulated PD signals having different properties and the de-noising results are evaluated using various performance indices.