小波变换提取电力电容器局部放电信号研究

Research on wavelet transform extraction partial discharge signal in power capacitor

为了消除电力电容器局部放电信号的噪声,提高电力电容器局部放电信号在线检测的有效性,提出基于小波变换阈值消噪方法的电力电容器局部放电信号提取方法。选择合适的小波基函数和分解层数对信号进行小波分解,确定合适的阈值规则以及阈值估计方法对小波分解系数进行阈值量化处理,从而得到消噪后的局部放电信号。以均方根误差RMSE和信噪比RSN为衡量指标,分析了小波基函数、分解层数、阈值规则和阈值估计方法等对消噪效果的影响。结果表明,当选择db2小波作为最优小波基函数,分解层数定为6层时,利用最大最小准则阈值估计法估计的阈值在软阈值规则下能有效提取电力电容器局部放电信号,达到了最优去噪效果。

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.