改进局部均值分解的齿轮箱复合故障特征提取  

Compound Fault Feature Extraction of Gearbox with Improved Local Mean Decomposition

作　　者：柴慧理[1] 叶美桃[1] Chai Huili;Ye Meitao(Department of Vehicle Engineering,Shanxi Traffic Vocational And Technical College,Taiyuan 030031,China)

机构地区：[1]山西交通职业技术学院车辆工程系

出　　处：《机械传动》2019年第8期130-134,共5页Journal of Mechanical Transmission

基　　金：国家自然科学基金(59975064);山西省基础研究项目(2015011063)

摘　　要：在强噪声环境下,针对局部均值分解(Local Mean Decomposition,LMD)出现的模态混叠现象,提出了总体局部均值分解方法(Ensemble Local Mean Decomposition,ELMD),但ELMD中所添加的白噪声不能完全被中和,这会导致PF分量受到所加白噪声的影响,导致重构误差增大。因此,提出基于PE-CELMD(Permutation Entropy-Complementary Ensemble Local Mean Decomposition)的齿轮箱复合故障诊断方法,该思路是在ELMD的基础上通过添加成对白噪声再结合排列熵(PermutationEntropy,PE)的方法优化LMD。将该方法应用于仿真信号和实测信号,并通过与LMD、CELMD对比,结果表明,PE-CELMD方法是一种有效的复合故障特征提取方法。In the case of strong noise, Ensemble local mean decomposition(ELMD) is proposed for themodal aliasing phenomenon of local mean decomposition(LMD). However, the white noise added in ELMD can-not be completely neutralized, which will result in the reconstruction error increases due to the Product functions(PF)components to be affected by the added white noise. Therefore, a compound fault feature extraction methodfor gearbox based on PE-CELMD(Permutation Entropy-Complementary Ensemble local mean decomposition) isproposed. The idea is to optimize ELMD by adding pairwise white noise in combination with Permutation Entro-py(PE) method based on ELMD. The method is applied to the simulated signal and the measured signal, andcompared with LMD and CELMD, the results show that the PE-CELMD method is an effective compound faultfeature extraction method.

关 键 词：局部均值分解 排列熵 复合故障 

分 类 号：TH132.41[机械工程—机械制造及自动化]