Using Support Vector Machine for Prediction of Machine Degradation Trend Based on Vibration Data
A condition based maintenance system needs to perform the following functions: analyzing collected data and identifying the components that have deteriorated significantly, predicting the degradation of these components, and making optimal maintenance decisions in order to minimize the total equipment operation cost. In this paper, we use Support Vector Machines (SVM) regression for prediction of machines degradation. The selection of SVM model parameters is investigated. A new rule is proposed for selection of the error zone value, one of the SVM model parameters. The proposed rule is also compared with CMas method and the results show that applying the proposed rule, SVM regression can provide better prediction than CMas method.
support vector machine regression parameter selection error zone value input target value optimization method vibration data
Hui Lin Ming J. Zuo
Department of Mechanical Engineering, University of Alberta, Canada
国际会议
The First International Conference on Maintenance Engineering(首届维修工程国际学术会议)
成都
英文
283-291
2006-10-15（万方平台首次上网日期，不代表论文的发表时间）