基于改进EMD的齿轮箱故障特征提取方法研究  

Study on the Method of Extracting Fault Feature in Gearbox based on Improved EMD

作　　者：谌继南 崔彦平[1] 王磊[1] 朱庆芹[1] 

机构地区：[1]河北科技大学机械工程学院,河北石家庄050018

出　　处：《机械传动》2014年第5期102-106,共5页Journal of Mechanical Transmission

基　　金：河北省科技支撑计划项目(10212112D);河北省高等学校科学技术研究重点项目(ZD2013081)

摘　　要：针对经验模式分解(Empirical mode decomposition,简称EMD)方法在信号处理时普遍出现的端点效应问题,结合齿轮箱早期故障非稳态信号特点,提出了基于分段三次Hermit插值的极值截断法与相关筛选法相结合的改进EMD方法。该方法将信号用分段三次Hermit插值法求取包络线,在距离端点最近的极值点处对包络线及信号进行截断,取中间部分继续处理,将分解后得到的基本模式分量(Intrinsic mode function,简称IMF)与原信号计算相关系数,筛选主要IMF作为进一步分析对象。实验结果表明,本改进EMD方法可以有效提取齿轮箱故障特征。According to Empirical mode decomposition （EMD） generally appearing to end effects in signal processing, combining features of non--stationary vibration signal from gearbox early faults, an improved EMD method is proposed. The method is based on the combination of piecewise Hermit interpolation of order three with extreme truncation and correlation coefficient screening method. Firstly, the piecewise Hermit interpolation of order three is used to calculate envelopes. Secondly, the signal and envelopes are truncated at the extreme the shortest distance to endpoints, and the middie of them is left for further processing. Lastly, after decomposition all of intrinsic mode functions （IMFs） are used to calculate correlation coefficients with the original signal, so as to find out main IMFs to be made further analysis. The results show that the improved EMD method can extract faults feature in gearbox effectively.

关 键 词：EMD 端点效应 插值 截断 相关筛选 

分 类 号：TH165.3[机械工程—机械制造及自动化]