This paper presents the work done by the Intelligent Robotics Group on CNC Machining Process Supervision, in the context of its participation in the B-LEARN II ESPRIT Ill project. An integrated hardware/software environment including CNC lathe and mill machines was developed. The machining process was monitored in real time using sensors for vibrations, sound and power consumption. Models for machine process characterisation suitable for the application of machine learning techniques related to semi-automatic training were also developed and assessed. The paper concentrates on the results of an Expert Machine Work Monitor based on sensor features extracted in real time and the NC program under execution. The developed concept of Specialised Monitors, as welI as techniques for sensor data processing, feature extraction / selection and class clustering are presented. An approach to sensor signal modelling using frequency analysis based on FFT and auto-regressive stochastic models, namely the ARMA(p,q) for characterisation of noisy sensorial data about process monitoring and evolution tracking, applied to the results obtained with real data, is presented.
Monitoring, Prognosis, CNC, Sensor Integration, Expert Supervision, Stochastic Process.
All articles published by Studies in Informatics and Control (SIC) are published in accordance with the terms and conditions of the Creative Commons Attribution License (CC BY-NC-SA 4.0).