用于提取局放脉冲的波形匹配识别方法初探

Research on Waveform Matching Recognition Algorithm in Partial Discharge Pulse Extraction

针对现场采集的局部放电信号中经常混杂白噪声和窄带周期干扰现象,本文首次提出使用Laplace小波按照波形匹配的原则来提取局部放电脉冲信号的方法。利用该小波波形与局部放电脉冲波形相吻合程度较好的特点,根据波形匹配的原则对含噪信号在一个由该小波构成的小波基函数空间中进行模式识别,进而去除噪声,提取局部放电脉冲信号。该方法不仅能够用于去噪,而且可以确定脉冲的放电主频率、衰减度、放电时刻及放电个数,从而为故障识别和定位奠定良好的基础。仿真和实验结果显示该算法具有良好的去噪性,同时可以获得更多的局放脉冲信息。

A new extraction method of partial discharge (PD) is proposed in this paper, which utilized Laplace wavelet based on waveform matching recognition algorithm. Laplace wavelet has the similar waveform as PD signals, so it can be picked up from those signals under noisy condition by way of pattern recognition of matching principle within the space consisted by Laplace wavelet atoms. The proposed method has some advantages. It can determine main-frequency, attenuation factor, occurrence time and the number of PD accurately. These information can be used for further research on classification and location of PDS. Simulations and experiments results demonstrate the correctness and effectiveness of the method proposed in this paper.