KPCA和改进SVM在滚动轴承剩余寿命预测中的应用研究  

Research on Application of KPCA and Improved SVM in Residual Life Prediction of Rolling Bearings

作　　者：者娜 杨剑锋[1] 刘文彬[1] 陈良超 ZHE Na;YANG Jian-feng;LIU Wen-bin;CHEN Liang-chao(Chemical Safety Engineering Research Center of the Ministry of Education,Beijing University of Chemical Technology,Beijing 100029,China)

机构地区：[1]北京化工大学化工安全教育部工程研究中心

出　　处：《机械设计与制造》2019年第11期1-4,8,共5页Machinery Design & Manufacture

基　　金：国家科技支撑计划资助项目(2011BAK06B03)

摘　　要：为解决支持向量机模型在预测滚动轴承剩余寿命时准确率不高的问题,对核主成分分析(Kernel Principal Com-ponent Analysis,KPCA)和最小二乘支持向量机(Least Square Support Vector Machine,LSSVM)在剩余寿命预测中的应用进行了研究。采用核主成分分析方法融合轴承振动信号时域、频域特征指标并提取第一主成分评估轴承性能退化情况,并将满足要求的多个KPCA主成分作为输入,建立KPCA-LSSVM模型来对轴承剩余寿命进行预测。采用轴承全寿命试验数据对该方法的有效性进行验证,结果表明,该方法提取的轴承性能退化评估指标能够更为全面地表征轴承性能退化情况,建立的KPCA-LSSVM模型可在滚动轴承剩余寿命预测工作中获得良好的预测效果。Application ofkernel principal component analysis(KPCA)and least square support vector machine(LSSVM)in residual life prediction of rolling bearings was studied toimproving the prediction accuracy of support vector machine model. With feature indexesin time domain and frequency domain of bearing vibration signal integrated by kernel principal component analysis,the firstprincipal component was extractedto evaluate bearing performance degradation. The KPCA-LSSVM modelto predict the residual life of bearingswas established by introducing the KPCA principal components satisfying the requirements into the model. The bearing run-to-failure life test was carried out to verify the effectiveness of the method,and the results showed that the evaluation index could more perfectly characterize the bearing performance degradation. Meanwhile,accurate predictive value of rolling bearing residual life could be achieved by the KPCA-LSSVM model.

关 键 词：滚动轴承 剩余寿命预测 评估指标 核主成分分析 最小二乘支持向量机 

分 类 号：TH16[机械工程—机械制造及自动化] TH17