滚动轴承故障诊断中的Volterra核-HMM识别方法  

Volterra Kernel-HMM Recognition Method in Fault Recognition of Rolling Bearings

作　　者：蒋静[1] 郑国彦 李志农[3] 

机构地区：[1]南昌航空大学无损检测技术教育部重点实验室,南昌330063 [2]河南电力工业学校,郑州450007 [3]郑州大学机械工程学院．郑州450001

出　　处：《煤矿机械》2010年第6期230-233,共4页Coal Mine Machinery

基　　金：国家自然科学基金(50775208);河南省教育厅自然科学基金(2008C460003)

摘　　要：结合Volterra级数和隐Markov模型,提出了一种基于Volterra核特征提取的HMM故障识别方法。在该方法中,利用子空间法从正常、滚动体故障、内圈故障和外圈故障4种不同的轴承中提取Volterra核作为特征向量,然后,输入到各种故障模式的HMM中进行识别。提出方法利用电机转轴末端滚动轴承采集的实验数据得到了验证。Combining voherra series with Hidden Markov Model （HMM）, a new bearing fault recognition method named Volterra kernel-HMM is proposed. vectors are extracted from from normal, ball, inner In the proposed method, the Volterra kernel feature and outer fault by the subspace method. Then these feature vectors are input the each fault HMM to recognize. The experiment result shows that the proposed method is very effective. The proposed method is tested with the experiment data sampled from drive end ball bearing of an induction motor driven mechanical system.

关 键 词：VOLTERRA核 隐Markov模型(HMM) 子空间 故障诊断 模式识别 

分 类 号：TH133.33[机械工程—机械制造及自动化] TP306[自动化与计算机技术—计算机系统结构]