Chee Khiang Pang, Jun-Hong Zhou, and Xiaoyun Wang
ARMAX model, classification, condition-based maintenance, pre-cognitive maintenance, support vector machine, tool conditionmonitoring.
Forecasting of process outages for reducing downtime and machine scrap parts have been actively pursued in the manufacturing industries to ensure that maintenance is carried out only when required. In this paper, we propose a precognitive maintenance framework based on mixed time and condition-based models to predict both wear and degradation stage of realistic engineering systems. The decision-making framework performs stage classification using support vector machines and wear estimation using time-based autoregressive moving average with exogenous inputs models. Our proposed framework is supported by mathematical rigour, and its effectiveness is evaluated with experimental results on a high-speed industrial computer numerical control milling machine for tool wear stage identification and wear estimation.
Important Links:
Go Back
