Condition monitoring of manufacturing processes

Novelty Detection is an approach to classification that offers advantages over conventional techniques for monitoring high-integrity systems and manufacturing processes: it is not necessary to provide fault data to the system during development. Instead, providing a sufficiently comprehensive model of the system‟s normal behaviour has been formed, deviations from normal working conditions can be identified automatically. As information is discovered regarding the causes of these deviations it is then possible to move from novelty detection to diagnosis, but the ability to identify previously unseen abnormalities is retained at all stages. This paper describes the results of an investigation for condition monitoring of a manufacturing process using techniques from novelty detection. We show that this approach allows the automatic identification of abnormal processes, providing identification of potential tool wear, giving advance notification of failure for the avoidance of bad processes.