基于VMD与FastICA的滚动轴承故障诊断  

Fault diagnosis method of rolling bearing based on VMD and FastICA

作　　者：张雪英 刘秀丽[1] 栾忠权[1] ZHANG Xueying;LIU Xiuli;LUAN Zhongquan(The Ministry of Education Key Laboratory of Modem Measurement and Control Technology,Beijing Information Science ＆ Technology University,Beijing 100192,China)

机构地区：[1]北京信息科技大学现代测控技术教育部重点实验室,北京100192

出　　处：《北京信息科技大学学报（自然科学版）》2018年第5期28-33,87,共7页Journal of Beijing Information Science and Technology University

基　　金：国家自然科学基金资助项目(51275052);国家高技术发展研究计划资助项目(2015AA043702);北京市教委科技计划一般项目(KM201811232023)

摘　　要：针对滚动轴承故障成分易被强背景噪声淹没,造成故障特征提取困难的问题,提出了一种采用变分模态分解(variational mode decomposition, VMD)与基于负熵的快速独立成分分析(fast independent component analysis, FastICA)相结合的故障诊断方法。利用经验模态分解(empirical mode decomposition, EMD)确定VMD模态数n,采用VMD将目标信号分解为n个模态分量;以连续的3个分量为一序列组合进行FastICA分析,从每组结果中选取一个最优分量,共n-2个;重构故障信号并进行Hilbert包络谱分析。通过试验分析并与EMD-FastICA方法、单一VMD方法比较,结果表明该方法能够更加清晰准确地提取故障特征信息,可用于轴承故障诊断。The fault signal of rolling bearing is often covered by the strong noise, making it difficult to extract fault information. To solve this problem, a method of fault diagnosis is proposed which combines variational mode decomposition （VMD） with fast independent component analysis based on negative entropy（FastICA）. EMD method is used to estimate the modal number n of VMD, while VMD is used to decompose the target signal into the sum of n components. The obtained components which are arranged into one sequence combination per constant third-order components are analyzed by FastICA. The n -2 components which are the best ones of each group are received. The fault signals which are decomposed by n - 2 components are analyzed after Hilbert envelope to extract the fault features for identification. Through experimental analysis, the method shows its good ability of distinctly and accurately extracting the fault information than EMD-FastICA method and VMD method.

关 键 词：变分模态分解 快速独立成分分析 故障诊断 滚动轴承 

分 类 号：TH133.33[机械工程—机械制造及自动化]