基于广义S变换模时频矩阵的局部放电特高频信号去噪方法

为有效抑制局部放电特高频信号中的噪声干扰,提出一种基于广义S变换模时频矩阵的去噪方法。基于二维模时频矩阵,采用区域最大能量法提取周期性窄带干扰的特征量,并通过矩阵逆向分离将其去除;采用奇异值分解去噪方法抑制信号中的高斯白噪声。使用该方法对仿真信号和实验室实测信号进行去噪处理,并与传统方法去噪结果进行对比。结果表明,所提方法能有效抑制局部放电信号特高频信号中的噪声,同时更好地保留了原始局部放电信号特征。对现场实测信号进行去噪处理,与传统方法相比,该方法具有较高的噪声抑制比和较低的幅值衰减比,可以有效提取局部放电超高频信号。

Partial Discharge Ultrahigh Frequency Signal Denoising Method Based on Generalized S-Transform Modular Time-Frequency Matrix

In order to suppress the noise in the partial discharge (PD) ultrahigh frequency (UHF) signal,a denoising method based on generalized S-transform modular time-frequency matrix is proposed.Maximum local energy method was used to extract the characteristics of periodic narrow-band interference and cancel the interference through the matrix reverse separation based on two-dimension modular tine-frequency matrix.In addition,singular value decomposition was used to suppress of Gaussian white noise in the PD signal.The denoising method presented in this paper was applied on the simulation and laboratory measured signals.Compared the denoising results with other four traditional denoising methods,the results show that the method presented in this paper suppressed the noise in the PD UHF signal effectively and retained more characteristics of PD signal.Compared with traditional methods,the denoising results of filed test signal validated the effectiveness of extracting PD signal with higher noise reduction ratio and lower amplitude reduction ratio.