Title :
SVM-PHM: A Novel Method for Remaining Useful Life Prediction
Author :
Tianle, Feng ; Jianmin, Zhao
Author_Institution :
Dept. of Equip. Command & Manage., Mech. Eng. Coll., Shijiazhuang, China
Abstract :
The Remaining Useful Life(RUL) prediction of the equipment plays a significant role in maintenance management. The accurate RUL prediction based on the current and previous health condition of the equipment is essential to make a timely maintenance decision for failure avoidance. In this paper, we presented a novel RUL forecasting method of Proportional Hazards Model (PHM) assembled with Support Vector Machine (SVM). In this method, we employed SVM to identify abnormal data and regress raw data. A case study is presented, and the performances of RUL prediction of PHM and SVN-PHM are examined.
Keywords :
data analysis; decision making; fault diagnosis; forecasting theory; hazards; life testing; maintenance engineering; regression analysis; support vector machines; SVM; failure avoidance; forecasting method; maintenance management; proportional hazards model; remaining useful life prediction; support vector machine; Data models; Engines; Hazards; Iron; Predictive models; Prognostics and health management; Support vector machines; outlier detection; prediction; proportional hazards model; remaining useful life; sdupport vector machine;
Conference_Titel :
Information Science and Management Engineering (ISME), 2010 International Conference of
Conference_Location :
Xi´an
Print_ISBN :
978-1-4244-7669-5
Electronic_ISBN :
978-1-4244-7670-1
DOI :
10.1109/ISME.2010.214
