摘要
将安全域的思想引入滚动轴承的状态监测中,综合利用经验模式分解(EMD)、主成分分析(PCA)和最小二乘支持向量机(LSSVM),进行了滚动轴承运行状态的安全域估计以及正常和各种故障状态的辨识。首先,按一定的时间间隔将采集的振动数据分段,每段数据进行EMD后获得各本征模函数(IMF)分量;其次,基于各段数据的本征模函数分量,利用主成分分析方法提取出每段数据的T^2统计量和平方预估误差(SPE)统计量控制限值作为滚动轴承的状态特征量;最后,利用二分类的LSSVM进行滚动轴承运行状态的安全域估计,利用多分类的LSSVM进行滚动轴承的正常以及滚动体故障、内圈故障、外圈故障四种状态的辨识。试验结果显示安全域估计准确率和多种状态辨识正确率均大于95%,验证了上述方法的有效性。
The idea of safety region was introduced into the condition monitoring of rolling bearings, and the research on estimation of a rolling bearing＇ s safe operating region and identification of a rolling bearing＇ s operating state （ normal or at fault） was performed by combinative use of empirical mode decomposition （EMD）, principal component analysis （PCA） and least square support vector machine （LSSVM）. Firstly, the vibration data of a rolling bearing was collected and it was segmented at regular intervals, and intrinsic mode functions （IMFs） of each segment＇ s data were obtained by using EMD. Then two statistical variables＇ control limits as the state characteristics of the roiling bearing were calculated based on PCA. At last, the boundary of the safety region was estimated by a twoclassification LSSVM, and the normal condition and three fault conditions were identified by a multi-classification LSSVM. The experimental results indicated that the accuracy of safety region estimation and that of state identification were both more than 95%, so the effectiveness of the above method was verified.
出处
《高技术通讯》
CAS
CSCD
北大核心
2013年第5期525-532,共8页
Chinese High Technology Letters
基金
863计划(2011AA110501)
国家科技支撑计划(2011BAG01B05)
国家重点实验室自主课题(RCS2010ZZ002)资助项目
