基于 AEEMD 和峭度－相关系数联合准则的轴承故障诊断

摘要

针对经验模态分解（EMD）和总体集成经验模态分解（EEMD）算法在轴承故障诊断中的缺陷，提出了一种基于自适应总体集成经验模态分解（AEEMD）与峭度和相关系数联合准则的滚动轴承故障诊断方法。首先，通过 AEEMD 将振动信号分解成具有不同特征时间尺度的本征模态分量，然后利用峭度和相关系数联合准则选取包含故障信息最丰富的 IMF，最后对选取的特征 IMF 做包络解调，进行故障诊断。并将 AEEMD 与 EMD 算法进行对比，证明了 AEEMD 算法的有效性和自适应性。%Aiming at shortages of EMD and EEMD algorithm in fault diagnosis for bearings,a fault diagnosis method for rolling bearings is proposed based on AEEMD and kurtosis -correlation coefficients joint criterion.Firstly,the AEEMD is used to decompose vibration signal into several IMFs of different characteristic time scales .Then the IMF which con-tain the most abundant fault information is selected according to kurtosis -correlation coefficients criterion.Finally,an envelope demodulation is adopted with selected IMF to diagnose fault.The AEEMD and EMD algorithm are compared to verify effectiveness and adaptivity of AEEMD algorithm.