VMD和小波阈值重构的电力电缆局部放电信号去噪法

Denoising method of cable partial discharge signals based on VMD and wavelet threshold reconstruction

矿用电力电缆的绝缘在劣化初期会产生局部放电信号,因此通过测量局放信号能够判断电缆的绝缘状态。然而由于局放信号微弱,受干扰信号影响严重,导致其提取难度大,从而降低了根据局放量判断设备绝缘劣化的准确性。提出以一种基于变分模态分解和小波阈值重构的矿用电力电缆局部放电信号去噪方法,将局放信号从干扰中提取。利用变分模态分解法首先将含噪的局放信号分为多个本征模态分量,其次采用小波阈值法对每个本征模态分量进行处理,最后将处理后的信号进行重构,得到去噪后的局放信号。对实验室6 kV电力电缆施加2.5 kV直流电压获得的局放电压信号进行去噪处理,与已有的去噪方法对比去噪效果。经对比分析,该方法的去噪效果优于现有的其他方法,尤其是对于信噪比极低的信号,采用该方法后,重构信号的信噪比、相关系数、均方根误差、噪声抑制比均得到了显著提升。

The insulation of mine power cable will produce partial discharge signals at the initial stage of deterioration with its insulation state judged by measuring the partial discharge signals.However,it is difficult to extract the partial discharge signals due to weak signals and strong interference signals,which reduces the accuracy of judging the equipment insulation deterioration.This paper proposes a denoising method of cable partial discharge signal based on variational mode decomposition and wavelet threshold reconstruction.By means of variational modal decomposition,the noise-containing partial discharge signals are divided into several intrinsic mode functions,each being processed by wavelet threshold,and the processed ones are reconstructed to obtain the de-noised partial discharge signals.The partial discharge voltage signal obtained by applying 2.5 kV DC voltage to 6 kV power cable in laboratory is denoised,and the denoising effect is compared with that the existing denoising methods,which shows that the de-noising effects of the method presented is better than that of other existing methods,especially for signals with extremely low snr,the signal-to-noise ratio(SNR),correlation coefficient,root-mean-square error(RMS).And noise suppression ratio(NCR)of the reconstructed signal are significantly improved.