Research on wavelet transform extraction partial discharge signal in power capacitor

Yan Hongyan,Huang Zhichao,Yang Shengzhen,Zhang Kewei,Fan Xingming,Shen Shuangjiang
DOI: https://doi.org/10.3969/j.issn.1673-808X.2012.05.009
2012-01-01
Abstract:In order to eliminate the noise of partial discharge signal in power capacitor and improve the effectiveness of on-line detection of partial discharge signal in power capacitor,a extraction method of power capacitor partial discharge signal based on the wavelet transform threshold de-noising method was proposed.The suitable wavelet base function and decomposition levels were chosen to disintegrate the signal,the right threshold rule and threshold estimation method were determined to make threshold value quantization processing for the wavelet coefficients to get the de-noising partial discharge signal.Root mean square error RMSE and signal-to-noise ratio RSN were chosen as the criterion to analyze and compare the de-noising effects of different wavelet functions,decomposition levels,threshold selection rules and threshold functions.The results show that when the db2 wavelet is chosen to be the optimum wavelet function and decomposition layers is 6,the electric power capacitor partial discharge signal can be extracted effectively by using minimaxi threshold value estimation under soft threshold rule,and the de-noising effect is the best.
What problem does this paper attempt to address?