LCD、k-means与ICA相结合的滚动轴承故障诊断方法  

Fault Diagnosis of Rolling Bearing using LCD,k-means and ICA

作　　者：孟凡磊[1] 崔伟成[1] 李伟[1] 刘林密[1] Meng Fanlei Cui Weicheng Li Wei Liu Linmi(Department of Aircraft Engineering, Naval Aeronautical and Astronautical University, Shandong Yantai 264001, China)

机构地区：[1]海军航空工程学院飞行器工程系,山东烟台264001

出　　处：《机械科学与技术》2017年第9期1402-1407,共6页Mechanical Science and Technology for Aerospace Engineering

基　　金：国家部委预研基金项目(9140A27020214JB1446)资助

摘　　要：为了准确地进行滚动轴承故障诊断,针对故障振动信号的低信噪比特征,提出了局部特征尺度分解、k均值聚类分析和独立分量分析相结合的故障诊断方法。首先,应用局部特征尺度分解对振动信号进行分解,得到若干个内禀尺度分量;然后,依据分量与原始信号的互相关系数及峭度值,应用k均值聚类方法选取有效的分量组成新的观测信号;最后,对观测信号进行独立分量分析处理,实现信噪分离,依据峭度值选取信号分量,对信号应用希尔伯特包络谱技术实现故障诊断。通过轴承内圈故障数据分析,验证了方法的有效性。The rolling bearing fault vibration signals have low signal-to-noise ratio. Aiming at diagnosing the fault of rolling bearing accurately, a method based on local characteristic-scale decomposition （LCD）, k-means cluster analysis and independent component analysis （ICA） was proposed. Firstly, the vibration signal was decomposed into some intrinsic mode components （ISC） by LCD. Then the correlation coefficients of every ISC and the original signal and the kurtosis value of every ISC were calculated, the efficient components were selected by means of k- means cluster analysis. The efficient components were processed by ICA to separate the signal from the noise, and the signal components were selected according to the kurtosis values. Finally, the Hilbert envelope aptitude envelope spectrum was used for fault diagnosis. The analysis of the bearing fault data shows that the method can realize weak signal detection and fault diagnosis effectively.

关 键 词：局部特征尺度分解 聚类分析 独立分量分析 互相关系数 峭度 轴承故障诊断 

分 类 号：TN911.23[电子电信—通信与信息系统] TP206.3[电子电信—信息与通信工程]