基于改进希尔伯特–黄变换的泄漏电流特性分析

Research on Characteristics of Leakage Current Based on Improved HHT Algorithm

绝缘子污秽状态在线监测,对防止污秽闪络具有重要意义。该文提出了绝缘子电弧放电的评判标准和一种基于泄漏电流特性的改进希尔伯特–黄变换:包括基于泄漏电流稳态特性提出的经验模态分解停止准则,基于泄漏电流频率特性提出的最优筛分次数以及基于泄漏电流脉冲特性提出的改善模态混叠新方法。通过与传统时频分析的对比研究证明了该方法更适合处理泄漏电流这类同时具有稳态特性和非稳态特性的时变信号,并用该方法对不同阶段的泄漏电流进行了分解,将分解得到的固有模态函数(intrinsic mode function,IMF)的标准差(standard deviation,STD)作为区域判断特征值。此外,还开发了一套泄漏电流在线监测装置并给出了整个算法的实现过程。仿真及实验结果表明,Hilbert谱中基波分量不连续、高次谐波瞬间变大反映了绝缘子表面的局部放电,基波IMF的STD与总IMF的STD之比可以作为泄漏电流区域判断的标准。

Online monitoring of insulator contamination plays an important role to avoid the pollution flashover. This paper proposed a quantified standard for arcing evaluation and an improved Hilbert-Huang transform （HHT） based on the features of leakage current （LC）. Novel method included a new empirical mode decomposition （EMD） stopping criterion, an optimal sifting iteration and a new masking signal technique to alleviate the mode mixing phenomenon. The paper investigated into the LC waveforms by comparing the improved HHT with traditional time-frequency （TF） methods, which indicated that the advanced approach is more suitable to deal with LC. The standard deviation （STD） of each intrinsic mode functions （IMFs） were calculated in order to provide a characteristic value for region discrimination. Besides, this paper designed an online LC monitoring system and given an entire process of flashover predicting algorithm. Result shows that the discontinuous fundamental component and the instantly increased low frequency harmonics indicate the presence of partial discharges, moreover, the ratio between STD of fundamental harmonic IMF and STD of the whole IMFs can be used to determine which region LC are in.