基于ITD与ICA的滚动轴承故障特征提取方法  

Fault diagnosis method for rolling bearings based on ITD and ICA

作　　者：柏林[1] 陆超[1] 赵鑫 

机构地区：[1]重庆大学机械传动国家重点实验室,重庆400044 [2]武汉锐科光纤激光器有限责任公司,武汉430000

出　　处：《振动与冲击》2015年第14期153-156,共4页Journal of Vibration and Shock

基　　金：国家自然科学基金项目(51475052)

摘　　要：针对滚动轴承故障信号因受背景噪声、信号传递途径、轴承各部件间相互作用及其它能量较大振源信号干扰,限制传统方法提取故障特征信息的准确性问题,提出结合固有时间尺度分解(ITD)及独立分量分析(ICA)的信号分析方法,将单通道振动信号进行ITD分解,得到若干固有旋转分量及一个趋势项,基于互相关准则对分解信号进行重组作为ICA的输入矩阵,采用Fast ICA算法解混,实现故障特征信号与噪声信号分离,从而提取故障特征信息。通过滚动轴承故障诊断实验结果分析表明该方法有效可行,具有一定工程应用价值。Rolling bearing fault signals appear mostly in the form of modulation,and are vulnerable to be influenced by other strong energy source signals,causing the great limitation of traditional method to be used in information extraction.According to these characteristics,the signal analysis method of intrinsic time scale decomposition （ITD） combined with independent component analysis （ICA）was proposed.The signal was decomposed into several proper rotation components and a trend component by the ITD method,and then they were reconstructed as an input matrix of ICA based on mutual correlation criterion.The FastICA algorithm was adopted to dissplve the mixedness so as to realize the separation of fault signal and noise signal.The method was applied to the fault diagnosis of rolling bearings.The analysis using the field data show that the method is effective and feasible,and has certain engineering application value.

关 键 词：固有时间尺度分析 独立分量分析 滚动轴承 故障诊断 

分 类 号：TP206.3[自动化与计算机技术—检测技术与自动化装置] TH132[自动化与计算机技术—控制科学与工程]