Forecasting Machine Vibration Trends Using Support Vector Machines
Volume 4, Number 2, April 2008 - Paper 6 - pp. 169 - 181
FAN JIANG1 and MING J ZUO21Petro-Canada Edmonton Refinery, Edmonton, AB, T5J 2G9, Canada
2University of Alberta, Edmonton, Alberta, T6G 2G8, Canada
(Received on May 30, 2007)
Equipment deteriorates as it is used and its vibration level usually increases accordingly. As a result, vibration has been used as an indicator of equipment health condition. Effective prediction of equipment deterioration trend can help prevention of equipment breakdown. However, little work has been done in machine degradation forecasting based on vibration data. Support Vector Machines (SVM) is a new tool for solving regression and classification problems. Successful applications of SVM for time series predictions reported in the literature motivates this study to use SVM for vibration trend prediction. This paper describes applications of SVM in vibration trend prediction. We compare the results obtained from time series forecasting methods and SVM.
Click here to download the paper.
Please note : You will need Adobe Acrobat viewer to view the full articles.