Suppressing periodic narrowband noise in partial discharge signal using EEMD and KICA

This paper puts forward a new method based on Ensemble Empirical Mode Decomposition (EEMD) and Kernel Independent Component Analysis (KICA) for de-noising to extract the Partial Discharge (PD) signals against narrow-band noise. At first, this method used EEMD to decompose the detection signal to get Intrinsic Mode Function (IMF) components which would be the inputs of KICA algorithm, then KICA could extract PD from the detection signal. The method is introduced to separate the PD signals against the narrow-band interferences of multifrequency and multiamplitude. The method can alleviate the mixing mode phenomenon in detection of PD and have a better detection result. It also has advantages such as the antinoise ability and small waveform distortion. The proceeding result of simulation model and signal from field test pro---+ves its effectiveness.