基于改进VMD的矿用电缆局放信号降噪方法  

A denoising method of mining cable partial discharge signals based on improved VMD

作　　者：李春锋 马星河[2] 刘广朋 Li Chunfeng;Ma Xinghe;Liu Guangpeng(Henan Tiantong Electric Power Co.,Ltd.,Pingdingshan 467000,China;Henan Polytechnic University,Jiaozuo 454003,China)

机构地区：[1]河南天通电力有限公司,河南平顶山467000 [2]河南理工大学,河南焦作454003

出　　处：《能源与环保》2023年第12期268-274,共7页CHINA ENERGY AND ENVIRONMENTAL PROTECTION

基　　金：河南省科技攻关项目(182102310936)。

摘　　要：提出一种基于改进变分模态分解的局放信号去噪方法,去除矿用电缆局放信号中夹杂着白噪声和窄带干扰。首先通过Spearman相关系数对VMD算法中的K进行优化;含噪局放信号被分解成K个本征模态分量(IMF);IMF分量中的噪声主导分量可采用峭度准则找出,并对该分量应用改进小波阈值法去噪;最后将去噪后的分量构造为去噪后的PD信号。通过仿真和实测信号表明:这种方法可以有效去除PD信号中的噪声,并与完全集合经验模态分解(CEEMD)算法相比,信噪比提高了46.29%,均方误差降低了32.61%,具有优良的去噪性能。We proposed a method for denoising partial discharge(PD)signal of cable based on improved variational modal decomposition in this paper,which removes white noise and narrowband interference mixed in the partial discharge signals of mining cables.Firstly,the K in VMD algorithm was optimized by Spearman correlation coefficient,the noisy PD signal was decomposed into K intrinsic mode components(IMF).The noise dominant component in the IMF could be identified using the kurtosis criterion,and a improved wavelet threshold method was applied to denoise it.Finally,the denoised component was constructed as the denoised PD signal.The simulation and measured signal results show that this method can effectively denoise the PD signal.Compared with the CEEMD algorithm,the SNR is increased by 46.29% and the MSE is reduced by 32.61%,and this method has excellent noise reduction performance.

关 键 词：变分模态分解 局部放电 降噪 改进小波阈值 Spearman相关系数 

分 类 号：TM855[电气工程—高电压与绝缘技术]