基于小波阈值的局部放电去噪新方法

New de-noising method for partial discharge signals based on wavelet threshold

针对局部放电信号去噪,传统的小波阈值法因小波基、阈值和分解层数这三个因素的影响,会使去噪后的波形发生畸变,产生较大误差。为了减小这些因素的干扰,本文提出了基于小波阈值去噪的新方法。首先利用波形相似法选取最优小波基,其次通过对理想局部放电信号和高斯白噪声进行每个尺度的小波分解与重构,并结合统计学知识确定局部放电信号去噪的阈值,最后对高频信号和低频信号进行能量分析,确定最优的分解层数。利用该方法和传统的小波阈值法对仿真放电信号去噪,去噪结果表明新方法在信噪比、均方根误差、相关系数和波形畸变率四个不同的指标上都得到了有效的提升,定性和定量的分析验证了该方法的有效性,实测的去噪结果表明新方法去噪效果令人满意,为局部放电信号去噪提供了一种新思路。

For partial discharge signals de-noising,the denoised waveforms will be distorted and result in large errors due to the impacts of wavelet base,threshold and decomposition levels for the traditional wavelet threshold algorithm. In order to reduce the impacts of these factors,this paper proposes a new method based on the wavelet threshold de-noising. Firstly,this paper selects the optimal wavelet base through the waveform similar method. Secondly,through each scale wavelet decomposition and reconstruction for the partial discharge signals and Gaussian white noise,this paper determines the de-noising threshold of partial discharge signals combining with statisticalknowledge. At last,it can determine the number of optimal decomposition level through the energy analysis for high frequency signals and low frequency signals. Making use of the new de-noising method and the traditional wavelet threshold de-noising algorithm to denoise for the discharge simulation signals,the de-noising results show that the new method is more effective than the traditional wavelet threshold algorithm in signal to noise ratio,root mean squared error,correlation coefficient and waveform distortion. This paper tests and verifies the effectiveness of the new method. The measured de-noising results show that the de-noising effect of new algorithm achieves the desired expectations,which provides a new way for the partial discharge signals to denoise.