Application of stochastic modelling to support predictive maintenance for industrial environments

Abstract

This paper presents the work that is being done by the UNINOVA's Intelligent Robotics Group, on a computerised numerical control (CNC) monitoring and prognosis system, using stochastic autoregressive integrated moving average (ARIMA) models. The experiments are based on an integrated hardware/software environment including CNC lathe and mill machines. The machining process is monitored using sensors for vibrations, sound and power consumption. The results obtained using real data, captured in real time using these sensors on a CNC machine, and modelled using stochastic ARIMA models, are presented. The authors' point of view about quality of conformity, related with the supervision of process control in manufacturing during machining tasks, and its implications in the enhancement of the system's efficiency, are also discussed.