基于小波集合的局部放电信息提取算法

对于已知波形特征的局部放电信号，选取最优母小波进行多分辨分析，经阂值处理后可以从强干扰中成功地检取出局部放电信息。但对于机理复杂、类型繁多的局部放电而言，单一的最优小波分析显然是较为局限的。文中提出了基于小波集合的局部放电信息提取算法，利用各小波在时、频域的不同特性，在较强的窄带和白噪声干扰条件下，能够更完整地提取多种形态的局部放电波形信息，在仿真计算和实测数据的处理中部取得了良好的效果。

ALGORITHM FOR EXTRACTING PD SIGNALS BASED ON A WAVELET-SET

As to the known shapes of the PD (partial discharge) pulses, the PD pulse buried in excessive noise can be extracted successfully by multi-resolution signal decomposition that based on the optimal wavelet, and then reconstructed via a proper threshold. However, the actual PDs are always of complex mechanism and different shapes, and a certain single optimal wavelet can often be proven quite limited in a practical case. The paper hereafter proposes an algorithm based on a wavelet-set, which makes use cf the distinguished characteristics of each wavelet in time and frequency domains, so that the diversiform PDs would be extracted more completely, even under excessive discrete spectral interference and white noise. The algorithm is applied both to simulation and practical data and the preferable results are obtained.