用谐波小波包变换法提取GIS局部放电信号多尺度特征参数

Multi-Scale Feature Parameters Extraction of GIS Partial Discharge Signal with Harmonic Wavelet Packet Transform

超高频(UHF)法在GIS局部放电(PD)检测中已得到了广泛应用,UHF PD信号的特征提取对准确识别GIS内部绝缘缺陷类型和指导检修工作具有重要意义,但目前仍然缺乏有效的特征提取方法。为此,本文利用谐波小波具有严格盒形频谱的优点,提出一种提取UHF PD特征信息的谐波小波包变换(HWPT)方法,对实验室获取的4种典型放电模型产生的UHF PD信号,采用HWPT进行多尺度分解,以克服实小波包分解子带间存在频谱混叠和能量泄漏的缺陷,利用UHF PD信号在不同尺度能量和复杂度的差异,提取多尺度能量和多尺度样本熵参数作为模式识别的特征量,更加精确地描述了UHF PD信号的时频域信息。最后利用支持向量机分类识别的结果表明,该方法可以取得比实小波包更好的识别效果,多尺度能量和多尺度样本熵特征参数均能有效识别4种绝缘缺陷。

Ultra-high frequency(UHF) method has been widely used in GIS partial discharge(PD) detection, and the feature extraction of UHF PD signal is important for the accurate identification of the GIS internal insulation defect types and guiding the repair work, but it's still lack of effective feature extraction methods. Therefore, this paper takes advantage of the strict box-shaped spectral of harmonic wavelet and proposes a harmonic wavelet packet transform(HWPT) method for UHF PD feature information extraction. Multi-scale decomposition with HWPT is adopted to the UHF PD signal produced by the four typical discharge models in the laboratory, and the sub-band spectrum aliasing and energy leaks of real wavelet packet decomposition is overcome. Using the energy and complexity difference of the UHF PD signal in different scales, the multi-scale energy and multi-scale sample entropy are extracted as the parameters for pattern recognition, so the time-frequency domain information of the UHF PD signal could be described more accurately. The classification and recognition results with support vector machine shows that this method has better recognition result than the real wavelet packet, and the multi-scale energy and multi-scale sample entropy feature parameters are both able to identify the four kinds of insulation defects effectively.