基于谐波小波包和DAG-RVM的滚动轴承故障诊断  

FAULT DIAGNOSIS OF ROLLING BEARING BASED ON HARMONIC WAVELET PACKET AND DAG-RVM

作　　者：齐磊[1] 王海瑞[1] 李宇芳[1] 李英[1] 任玉卿 

机构地区：[1]昆明理工大学信息工程与自动化学院,云南昆明650500

出　　处：《计算机应用与软件》2017年第5期61-67,103,共8页Computer Applications and Software

基　　金：国家自然科学基金项目(61263023)

摘　　要：针对传统滚动轴承故障诊断方法受人为因素影响较为严重,故障成因相对复杂等问题,在现有的研究基础上提出一种基于小波包分析和有向无环图相关向量机相结合的故障诊断方法。将滚动轴承在不同的故障条件下的振动信号进行谐波小波包分解与重构,提取频带能量作为特征向量,应用有向无环图相关向量机建立从特征向量到故障模式之间的映射,最终做到对滚动轴承的故障诊断。结果表明,该方法能够快速准确地诊断出滚动轴承故障,验证了该方法的有效性和稳定性。此外,通过与支持向量机(SVM)的对比分析,显示了RVM在智能故障诊断应用中的优越性。In view of the traditional rolling bearing fault diagnosis methods is affected by human factors,and the cause of the fault is relatively complex. Based on the existing research,a fault diagnosis method based on wavelet packet analysis and acyclic graph relevance vector machine is proposed in this paper. The vibration signals of the rolling bearing under different fault conditions are decomposed and reconstructed by harmonic wavelet packet,and the frequency band energy is extracted as feature vector. The mapping from feature vector to fault mode is established by using acyclic graph relevance vector machine,finally the fault diagnosis of rolling bearing is solved. The results show that this method can quickly and accurately diagnose rolling bearing faults,and verify the effectiveness and stability of the method. In addition,compared with SVM,it shows the superiority of RVM in intelligent fault diagnosis application.

关 键 词：谐波小波包 有向无环图 相关向量机 故障诊断 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]