基于LCD和稀疏表示的滚动轴承故障诊断  

Rolling bearing fault diagnosis based on LCD and sparse representation

作　　者：张思郁 丁锋[1] 翟丹宁 ZHANG Siyu;DING Feng;ZHAI Danning(School of Mechanical and Electrical Engineering,Xi’an Technological University,Xi’an 710021,Shaanxi,China)

机构地区：[1]西安工业大学机电工程学院,陕西西安710021

出　　处：《中国工程机械学报》2024年第3期404-409,共6页Chinese Journal of Construction Machinery

基　　金：陕西省自然科学基础研究计划资助项目(2023-JC-YB-347)。

摘　　要：针对局部特征尺度分解(LCD)重构信号存在虚假成分、正交匹配追踪(OMP)算法迭代终止条件较难确定,以及相似原子难以区分的问题,将LCD和改进OMP算法相结合用于滚动轴承特征提取,用LCD对故障信号进行降噪与分解处理。根据相关峭度值,选择分解后的分量进行重构,利用改进OMP算法进行稀疏表示和特征提取,取得较好的效果。使用仿真信号和试验信号对比分析验证了该方法的有效性。In view of problems existing in local characteristic-scale decomposition(LCD) reconstruction signals false components,difficult to determine the iterative termination conditions of the orthogonal matching pursuit(OMP) algorithm and distinguish similar atoms.LCD and improved OMP algorithm are combined for rolling bearing feature extraction,LCD is used to reduce noise and decompose the fault signal,the improved OMP algorithm is used to sparsely represent and extract features of the decomposed signal.According to the relevant steepness value,the decomposed components are selected for reconstruction,and the improved OMP algorithm is used for sparse representation and feature extraction.The results are good.The comparison and analysis of simulated experimental signals verifies the effectiveness of the proposed method.

关 键 词：局部特征尺度分解 正交匹配追踪算法 稀疏表示 故障诊断 

分 类 号：TH133.33[机械工程—机械制造及自动化]