基于球结构支持向量机的故障劣化趋势诊断  

Fault Diagnosis of Deterioration Trend of Based on Sphere-structured SVM

作　　者：田英[1] 杨洁明[2] 申永强[3] 

机构地区：[1]山西机电职业技术学院,山西长治046011 [2]太原理工大学,太原030024 [3]西南石油大学,成都610500

出　　处：《机械工程师》2010年第4期88-89,共2页Mechanical Engineer

摘　　要：提出了一种球结构支持向量机的故障劣化趋势检测方法。该方法是以EMD的奇异值分量作为特征向量,输入球结构支持向量机分类器,对滚动轴承滚珠的故障劣化趋势的三种状态进行诊断,并且采用主成份分析的可视化方法进行验证,结果表明,该方法能够准确地诊断故障劣化趋势。Deterioration trend of the fault diagnosis method based on sphere-structured SVM is put forward. The method regards EMD SVD method as characteristic vector input to establish sphere-structured SVM and identifies roller bearing in rolling body fault three deteriorating state. Every clustering is visualized for verifying it by principal component analysis. It shows the method can identify roller bearing inside track fault deteriorating state accurately .

关 键 词：球结构支持向量机 故障诊断 劣化趋势 

分 类 号：TP206.3[自动化与计算机技术—检测技术与自动化装置] O235[自动化与计算机技术—控制科学与工程]