Wavelet de-noise on gas PD signals of coaxial cylinder electrodes

The in-service failures of high voltage apparatus with gaseous-insulation can be predicted according to the behavior of the gas under the applied electric field. It has practical significance for engineers to deduce the failure location, property, and danger level from the partial discharge (PD) signals occurring in insulating gas dielectrics. Thus the signal detected and de-noise technique are obviously very important. The paper presents the basic principle and method for dealing with PD signals employing multiple wavelet transform methods. Practically, demonstrations show that the multiple wavelet transform is very useful and effective for extracting and recognizing PD pulse signals from the defects of high voltage apparatus on-line or off-line. It also makes certain that failures located in the GIS, GIC and GIT, can be precisely measured, located, predicted and monitored.