基于优化VMD与噪声估计的间隔阈值局部放电去噪方法

为了抑制局部放电(PD)信号中含有的窄带周期干扰和白噪声,提出一种基于优化的变分模态分解(VMD)阈值去噪方法。首先针对VMD算法可能造成染噪信号欠分解或过分解的问题,提出一种基于频谱分析和四分位数的模态分解数K值优化方法,并结合模态的峭度特征去除窄带周期干扰和高频白噪声;针对PD信号主导模态中残留的白噪声,利用文中研究中发现的VMD分解白噪声所得模态的两个统计特性,提出一种噪声标准差估计方法来确定阈值,最后引入间隔阈值函数对PD信号主导模态进一步去噪。采用该方法对仿真和实测信号进行去噪处理,并将其与传统方法进行对比,结果表明,所提方法不仅可以更有效地抑制噪声,同时也能更好地保留PD信号的特征。

Partial Discharge Signal Denoising Method Based on Optimized VMD and Noise Estimation for Interval Threshold

In order to suppress the periodic narrow-band interference and white noise in the partial discharge(PD)signal,a denoising method based on the optimized variational mode decomposition(VMD) and threshold of the modes is proposed.Firstly,aiming at the problem that the VMD algorithm may under-decompose or over-decompose the noise signal,a method of K value determination based on spectrum analysis and quantile is proposed,and the narrow-band interference and high frequency white noise are removed by using kurtosis;Secondly,aiming at the residual white noise in the PD-dominant mode,a noise standard deviation estimation method based on two statistical characteristics of decomposing white noise with VMD is proposed to determine the threshold,and the interval threshold function is introduced to further denoise the PD-dominant mode.The denoising method was applied on the simulative and field measured signal,compared with the traditional methods,the results show that the proposed method can preserve the characteristics of the PD signal better while suppressing the noise more effectively.