针对局部放电信号中受到的平稳随机型干扰和窄带周期型干扰,提出了一种基于谱峭度的局部放电信号提取方法。首先求取出含噪局部放电信号的谱峭度,然后根据谱峭度设计自适应带通Wiener滤波器。对经过滤波器滤波后的信号进行小波平滑去噪,提取出更加精确的局部放电信号。分别采用所提方法和小波阈值去噪方法提取局部放电信号,结果表明,所提方法能够很好地抑制平稳随机型干扰和窄带周期型干扰,且各个性能指标都优于小波阈值去噪方法。
To remove the smooth random interferences and narrow-band periodic interferences in partial discharge signals,a method of partial discharge signal extraction based on spectral kurtosis is proposed. The spectral kurtosis of partial discharge signals containing noises is calculated and the corresponding adaptive band-pass Wiener filter is designed,by which,the narrow-band periodic interferences and white noises are filtered from the partial discharge signals. The wavelet smooth denoising is then carried out to extract the more accurate partial discharge signals. Test results show that,compared with the wavelet threshold denoising method,the proposed method,with superior performance indexes,suppresses the smooth random interferences and narrow-band oeriod interferences much better.