基于EEMD阈值的MOA避雷器泄漏电流去噪的研究  

Research on De-noising of MOA Leakage Current Based on EEMD Threshold

作　　者：赵腾[1] 姚毅[1] 周芯竹 

机构地区：[1]四川理工学院人工智能四川省重点实验室,自贡643000

出　　处：《科学技术与工程》2014年第31期231-236,共6页Science Technology and Engineering

基　　金：人工智能四川省实验室重点项目(2011RZ005)资助

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。The resistive component in leakage current of MOA arrester could reflect the operation state of MOA effectively, but MOA usually operate in the strong electromagnetic environment, so the leakage current must mixed with much of noise. Using it without effective signal de-noising will affect the accuracy of monitoring system for ana- lyzing signal and MOA fault diagnosis. Aiming at the characteristics of leakage current of MOA, combined with the universal threshold, the heuristic threshold, the unbiased risk estimate threshold and the minimax threshold, put forward the Ensemble Empirical Mode Decomposition threshold de-noising methods. Though the three evaluation criteria of root-mean-square value, three harmonic amplitude ratio and de-noised waveform, judging the result of EEMD threshold de-noising. Analyzing simulated signal and measured signal acquisition from field to get the con- clusion that the method EEMD based the unbiased risk estimate threshold is the optimal EEMD threshold de-noising method.

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 

分 类 号：TM863[电气工程—高电压与绝缘技术]