基于SALSA优化算法的变转速轴承故障特征稀疏表示方法研究  

The sparse representation of feature extraction of bearing fault based on SALSA algorithm at varying speed condition

作　　者：吴楠 石娟娟[1] 江星星[1] 朱忠奎[1] WU Nan;SHI Juanjuan;JIANG Xingxing;ZHU Zhongkui(School of Urban Rail Transportation,Soochow University,Jiangsu Suzhou,215137,China)

机构地区：[1]苏州大学城市轨道交通学院,江苏苏州215137

出　　处：《机械设计与制造工程》2018年第6期17-21,共5页Machine Design and Manufacturing Engineering

基　　金：江苏省自然科学青年基金项目(BK20160318)

摘　　要：变转速工况下轴承等设备的关键部件出现故障后对设备的危害十分严重,而且此类工况下振动信号更加复杂。为了克服此类问题,引入分裂增广拉格朗日收缩算法建立变转速轴承故障特征稀疏表示方法,实现变转速下轴承故障冲击特征的准确提取。首先,基于变转速轴承故障振动响应模型,分析不同转速下轴承故障振动响应形态变化规律,构造Laplace小波基底过完备字典;然后运用分裂增广拉格朗日收缩算法实现故障信号的稀疏表示与重构,通过提取重构信号的特征阶次实现轴承的故障诊断。轴承故障诊断实例验证了所提方法在变转速工况下轴承故障诊断的有效性。It proposes method based on sparse representation for bearing fault feature extraction at varying speed.Based on the analysis of fault-induced transient morphology at varying speed,it establishes the dictionary,uses the split augmented Lagrangian shrinkage algorithm to represent the signal sparsely. Via analyzing the characteristic order extracted from the reconstructed signal,it realizes the fault diagnosis. The experimental signals validates effectiveness of the proposed method.

关 键 词：变转速 稀疏字典 振动模型 故障诊断 

分 类 号：TH113.1[机械工程—机械设计及理论]