De-noising of GIS partial discharge signal corrupted with narrow-band noise using Shannon wavelet

Ultra high frequency(UHF) method has been widely used in GIS partial discharge monitoring for its outstanding ability of anti-interference nowadays. However, suffused narrowband noise with frequencies lying in the band of the sensor may still intrude into the monitor system, thus distorting the partial discharge monitoring signal. The acquired monitoring signal is used to identify the existence of GIS internal insulation defects, so it is of great significance to de-noise the signal. In this paper, Shannon wavelet is selected for its rectangular spectral characteristic in frequency domain, and sub-band Shannon entropy is regarded as a criterion to locate the narrow-band noise frequency. The de-noised signal is obtained by setting to zero the coefficients of sub-bands with over-low Shannon entropy. Numeric simulation experiments have been set to verify the feasibility and correctness of the proposed method.