Responding to the Forecast

Abstract

Machines become increasingly complex. At the same time, more and more sensors are installed and information is gathered in order to enable a close to real-time prediction of a machine's state. Compa-nies try to implement Predictive Maintenance strategies to avoid machine downtimes on a large scale. For this purpose, artificial neural networks are applied more and more often. However, the classifica-tion of machine states with artificial neural networks is still not accurate enough. This is partially due to a lack of standards in data processing and in the harmonization of data from different sensor types. We aim to contribute to close these research gaps by developing a standard PM concept for machine and plant manufactures.