SVM-based dynamic modeling for machine life stage identification and machine wear estimation

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. © 2012 IEEE.