时频匹配滤波法用于变压器局部放电模式识别的实验研究

UHF partial discharge classification based on time-frequency matched filter algorithm

局部放电模式识别对于电力变压器绝缘状况诊断具有重要意义。分析了典型变压器局部放电缺陷,建立4种放电模型,采用小波变换获取了超高频局放脉冲的三维时频谱图,该三维谱图综合反映了局放信号的3个基本特征:时间分量、频率分量和放电能量的分布。根据时频谱构造了4种局放模型的三维匹配滤波器,待测局放信号与同类型滤波器相匹配。结果表明,该方法可以有效提取出局部放电信号的主要特征和趋势。

For insulation condition assessment of HV transformer, partial discharge （PD） monitoring is one of the most effective techniques. It was analyzed the characteristic of UHF matched signal at time and frequency domain. On the basis of wavelet analysis, a three-dimensional time-frequency pattern to characterize UHF PD signal was upbuih which comprehensively reflected three basic features of PD impulses： time component, frequency component and distribution of discharging energy. The matched filters of three-dimension were constructed by mathematic model according to PD time-frequency patterns. Based on the matched three-dimensional pattern, the major features and tendency of discharge were characterized. Then measuring PD signals of three-dimension passed four parallel matched filters simultaneously and the maximum coefficients could be extracted for pattern recognition. The result shows that the method could effectively distinguish the type of PD.