基于EEMD阈值的MOA避雷器泄漏电流去噪的研究

发布时间：2018-03-01 08:43

本文关键词： MOA避雷器泄漏电流阈值原则均方根值三次倍频幅值比总体经验模态分解　出处：《科学技术与工程》2014年31期 　论文类型：期刊论文

【摘要】：MOA避雷器泄漏电流中的阻性分量是反映MOA运行状态的有效特征量,但MOA通常运行在强电场和强磁场环境中,MOA避雷器在线监测系统采集的泄漏电流信号必然混叠很多噪声,不经有效的信号去噪即采用,会影响监测系统的信号处理、分析和MOA避雷器故障诊断的准确性。针对MOA避雷器泄漏电流的特点,结合通用阈值(sqtwolog)原则、启发式阈值(heursure)原则、无偏似然估计阈值(rigsure)原则和极值阈值(minimax)原则,提出了总体经验模态分解(简称EEMD)阈值去噪方法。采用均方根值、三次倍频幅值比和去噪后的波形这三个评价标准来判断EEMD阈值去噪的优劣,通过对信号仿真和现场采集的实测信号进行分析,得出基于无偏似然估计阈值原则是最佳的EEMD阈值去噪方法。
[Abstract]:The resistive component in the leakage current of MOA arrester is an effective characteristic to reflect the running state of MOA. However, the leakage current signals collected by the on-line monitoring system of MOA arrester usually run in the environment of strong electric field and strong magnetic field. Without effective signal de-noising, it will affect the signal processing, analysis and the accuracy of MOA arrester fault diagnosis. According to the characteristics of leakage current of MOA arrester, combining with the general threshold value and heuristic threshold heuristics principle, Based on the unbiased likelihood estimation threshold and extremum threshold minimax principle, a threshold denoising method based on global empirical mode decomposition (EEMDM) is proposed. The root-mean-square (RMS) value is used. The three evaluation criteria, namely, the ratio of amplitude to value of cubic frequency doubling and the waveform after denoising, are used to judge the merits and demerits of EEMD threshold denoising. Through the analysis of the signal simulation and the measured signals collected in the field, It is concluded that the unbiased likelihood estimation threshold principle is the best EEMD threshold denoising method.
【作者单位】：四川理工学院人工智能四川省重点实验室;
【基金】：人工智能四川省实验室重点项目(2011RZ005)资助
【分类号】：TM862

【参考文献】