RS-CS-SVM在电液伺服系统故障诊断中的应用  

The Application of RS-CS-SVM in Fault Diagnosis of Electro-hydraulic Servo System

作　　者：孙海霞[1] 木合塔尔.克力木 王晨[1] 李卉[1] SUN Hai-xia;MUHETAER Ke-li-mu;WANG Chen;LI Hui(School of Mechanical Engineering,Xinjiang University,Urlumuqi 830047,China)

机构地区：[1]新疆大学机械工程学院,乌鲁木齐830047

出　　处：《组合机床与自动化加工技术》2018年第6期47-50,55,共5页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：国家自然科学基金项目(51365052)

摘　　要：针对电液伺服系统故障数据冗杂、非线性以及多样性等问题,提出了一种基于Rough Set(RS)和Cuckoo Search(CS)算法优化的Support Vector Machine(SVM)的故障诊断方法。该方法通过AMESim仿真软件对穿戴式康复训练机器人电液伺服系统进行建模,并提取故障特征量;利用粗糙集把故障特征量的冗余信息剔除,再利用布谷鸟算法优化进行向量机参数的选取,将优化处理后的故障数据作为样本输入支持向量机,实现故障诊断和分类。通过将该方法与其他几种优化支持向量机方法相比较,这种方法对于电液伺服系统故障数据冗杂、非线性及较差的故障分类具有很好的诊断功能,且其诊断正确率较高以及诊断时间大大缩短。According to tedious,nonlinear of electro-hydraulic servo system failure data and diversity of failure forms,this paper proposes a method based on rough set（RS） and the cuckoo search（CS） algorithm of support vector machine（SVM） for fault diagnosis. The wearable rehabilitation training robot electro-hydraulic servo system is modeled in simulation software of AMEsim,and the fault characteristic quantities are extracted; redundant information of the fault characteristic quantities are eliminated by CS,vector machine parameters are selected by CS. Failure data optimized are inputed support vector machine（SVM） as sample,then fault diagnosis and classification are implemented in the analysis and research of the electro-hydraulic servo system. Compared this method with several other optimization support vector machine（SVM） method,this method for electro-hydraulic servo system fault data＇s mad,nonlinear and less fault classification has the good diagnostic function,and the diagnostic accuracy is higher,and time is greatly shortened.

关 键 词：粗糙集 布谷鸟算法搜索 支持向量机 电液伺服系统 故障诊断 

分 类 号：TH137.9[机械工程—机械制造及自动化] TG506[金属学及工艺—金属切削加工及机床]