一种EEMD阈值降噪方法在故障诊断中的应用  

A de-noising approach by EEMD threshold for the application of bearing fault diagnosis

作　　者：常勇[1] 何婷 CHANG Yong;HE Ting(Lanzhou Petrochemical College of Vocational Technology,Lanzhou 730060,China;Gansu Academy of Science Institute of Natural Energy,Lanzhou 730030,China)

机构地区：[1]兰州石化职业技术学院电子电气工程系,兰州730060 [2]甘肃省科学院自然能源研究所,兰州730030

出　　处：《工业仪表与自动化装置》2018年第4期3-7,共5页Industrial Instrumentation & Automation

基　　金：兰州石化职业技术学院基金项目(KJ2015-20)

摘　　要：针对机械故障信号分析中随机噪声严重影响和误导人们对故障位置和类型的判断分析,很难准确地提取故障特征信息,该文通过研究集合经验模态分解(EEMD)和小波变换的原理,将二者相融合提出了新的信号降噪方法。该方法就是首先对含有噪声的故障信号进行EEMD分解,对分解后的固有模态函数(Intrinsic Mode Functions,IMFs)采用自适应阈值降噪的办法,对降噪后的故障信号进行重构,再进行故障类型分析。实验分析研究表明,该方法有效地解决了EMD在降噪过程中出现的模态混叠、端点效应问题,降噪效果更好。通过仿真和实测故障信号验证分析,结果表明,所提出的方法不但降噪效果良好,而且能更有效地进行故障特征提取和故障类型判断。In view of the random noise signal causing interference in the signal analysis of fault diagnosis,and mislead people for fault location and type of judgement. In this paper,introduced the ensemble empirical mode decomposition（EEMD） and wavelet threshold combining the signal de-noising method. The approach is firstly to EEMD for actual noise fault signal,then according to the decomposition after each intrinsic mode functions（IMFs） using adaptive threshold de-noising method,then the fault signal of de-noising after reconstruction,and then to fault type of analysis. Simulation studies show that the proposed method is effective to solve the EEMD in modal aliasing problems appeared in the process of noise reduction,better de-noising effect,to the actual fault signal effectively extract the fault characteristic frequency,judge fault type for more effective.

关 键 词：集合经验模态分解(EEMD) 小波阈值 固有模态函数(IMF) 

分 类 号：TM315[电气工程—电机]