基于RS-LSSVM的高速列车走行部滚动轴承故障诊断研究  

Research on fault diagnosis method of high-speed train running gear rolling bearing based on RS and LSSVM

作　　者：贺德强[1] 陈二恒 李笑梅[1] 刘旗扬[3] 

机构地区：[1]广西大学机械工程学院,广西南宁530004 [2]重庆大学机械工程学院,重庆400044 [3]广西大学科技处,广西南宁530004

出　　处：《广西大学学报（自然科学版）》2017年第2期403-408,共6页Journal of Guangxi University（Natural Science Edition）

基　　金：国家自然科学基金资助项目(51165001);广西科技攻关项目(桂科攻1598009-6);南宁市科技攻关项目(20151021)

摘　　要：针对高速列车走行部滚动轴承故障诊断模型构建时间较长、诊断准确率不高的问题,提出一种基于粗糙集(RS)和最小二乘支持向量机(LSSVM)的方法。该方法利用小波包变换构造能量特征集,使用粗糙集属性约简算法对离散后的能量特征集处理,得到最小约简,将其输入到基于最小二乘支持向量机的故障诊断模型中进行状态识别。测试实例证明了粗糙集属性约简算法不仅保留了能量特征集的重要属性,缩短了后期故障诊断模型构建时间,而且保证了故障诊断的准确率,其模型构建时间为0.071 s,故障诊断准确率为100%。因此,RS和LSSVM相结合是一种优秀的故障诊断方法,可以作为高速列车走行部滚动轴承故障诊断的新思路。Aiming at the problem that the fault diagnosis model building needs a long time and the diagnostic accuracy is not high for the high speed train running gear rolling bearing,the method based on rough sets（ RS） and least square support vector machine（ LSSVM） is proposed. The wavelet packet transform is used to construct energy feature set,then in order to get the minimal reduction,the discrete energy feature set is dealt with rough sets attribution reduction algorithm.Finally,the minimal reduction is sent to fault diagnosis model based on least square support vector machine to identify the status. The test examples show that the rough sets attribution reductionalgorithm not only preserves the important properties of the energy feature set,but also reduces the time of the late fault diagnosis model and guarantees the accuracy of fault diagnosis. The modeling time is 0. 071 s,the accuracy of fault diagnosis is 100%. Therefore,the RS and LSSVM is a kind of excellent fault diagnosis method,which can be used as a new method for high speed train running gear rolling bearing fault diagnosis.

关 键 词：高速列车走行部 滚动轴承 故障诊断 粗糙集 最小二乘支持向量机 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术] U260[自动化与计算机技术—计算机科学与技术]