Title :
SVM-based dynamic modeling for machine life stage identification and machine wear estimation
Author :
Zhou, Jun-Hong ; Pang, Chee Khiang ; Wang, Xiaoyun
Author_Institution :
Manuf. Execution & Control Group, A*STAR Singapore Inst. of Manuf. Technol., Singapore, Singapore
Abstract :
Identification of machine condition is crucial to reduce machine downtime and scrap parts in the manufacturing industries. In this paper, we propose a novel methodology to identify life stage and estimate machine wear. The proposed framework is based on stage identification using Support Vector Machines (SVMs) and machine wear estimation using AutoRegressive Moving Average with eXogenous inputs (ARMAX) models. Our proposed framework is evaluated on a high-speed industrial milling machine, and the effectiveness of the proposed methodology in tool wear stage identification and tool wear estimation is verified with experimental results.
Keywords :
autoregressive moving average processes; condition monitoring; milling machines; pattern classification; production engineering computing; support vector machines; wear; ARMAX model; SVM-based dynamic modeling; autoregressive moving average with exogenous input; high-speed industrial milling machine; machine condition identification; machine downtime; machine life stage identification; machine wear estimation; manufacturing industry; support vector machines; tool wear estimation; tool wear stage identification; wear stage classification; Computational modeling; Data models; Estimation; Feature extraction; Force; Linear regression; Support vector machines; ARMAX models; Support Vector Machine (SVM); Tool Condition Monitoring (TCM); dynamic models; stage classification;
Conference_Titel :
Intelligent Engineering Systems (INES), 2012 IEEE 16th International Conference on
Conference_Location :
Lisbon
Print_ISBN :
978-1-4673-2694-0
Electronic_ISBN :
978-1-4673-2693-3
DOI :
10.1109/INES.2012.6249835
