Suppression method of partial-discharge interference signal based on the combination of morphological filters and complex wavelet transform

Solving the periodic narrowband interference and suppression of white noise during the actual detection of partial discharge (PD) are difficult. In this paper, a method of suppressing the interference signal method based on the combination of morphological filters and complex wavelet transform is proposed. Generalized morphological filters based on the principle of mathematical morphology were structured and set as a pre-filter unit to realize pre-treatment of the original PD signals, and complex wavelet transform was then employed to process the PD signals. Finally, denoised PD signals were obtained. Applying this method to deal with the noise in the simulation and acquisition of PD signals, the results showed that the method can restrain effectively the PD periodic narrowband interference and white noise. Compared with the same wavelet base of the wavelet and the complex wavelet denoising methods, this method reduces energy loss and retains well the PD signals characteristics. Streszczenie: W artykule opisano metode tlumienia zaklocen (zaklocenia wąskopasmowe, bialy szum) w sygnalach, w przypadku wykrycia wyladowan niezupelnych. W rozwiązaniu wykorzystano kombinacje filtrow morfologicznych, dzialających jako pre-filtr i zlozoną transformate falkową do dalszej obrobki sygnalow. Wykonano badania symulacyjne, potwierdzające skutecznośc filtracji. (Tlumienie zaklocen od pochodzących od wyladowan niezupelnych - filtry morfologiczne i zlozona transformata falkowa).