基于短时傅里叶变换的局部放电窄带干扰抑制

Narrowband Interference Suppression of Partial Discharge Based on Short-time Fourier Transform

摘要:

为有效抑制局部放电(PD)信号中周期性窄带干扰，文中提出短时傅里叶变换和矩阵束相结合的局部放电窄带干扰抑制方法。该方法利用短时傅里叶变换分析染噪PD信号，得到窄带干扰数目，同时分离出染噪PD信号中信号帧和噪声帧；采用矩阵束算法在噪声帧中对窄带干扰参数进行估计，重构出全时段的窄带干扰，实现局部放电的窄带干扰抑制。仿真及实测染噪PD信号的去噪效果表明，与传统的傅里叶级数法和局部能量比法相比，所提方法去噪后的残余噪声更小，对染噪PD信号中窄带干扰有很好的抑制效果。

Abstract:

In order to effectively suppress the periodic narrowband interference in partial discharge (PD) signal, a method combining shorttime Fourier transform and matrix pencil is proposed. The short-time Fourier transform algorithm is used to analyze the noisy PD signal, and to obtain the number of narrowband interference and separate noise frame and signal frame from noisy PD signal. The matrix pencil algorithm is used to estimate the narrowband interference parameters in the noise frame, and to reconstruct the narrowband interference of the entire period and suppress the narrowband interference in partial discharge. The denoising results of the simulated and measured noisy PD signals show that compared with the traditional Fourier series method and local energy ratio method, the proposed method can get the smaller residual noise and produce a good effect of narrowband interference suppression.