Predictive Maintenance for Existing Process Manufacturing Plants

© Fraunhofer IFF, Dirk Mahler

Project VIPro

 

A methodology was developed in this project, which enables process plant operators to predictively maintain (smart manufacturing/Industrie 4.0) plant components without increased dependence on component manufacturers. This entailed ascertaining which plant operating data are relevant to the performance of particular components. Suitable mathematical methods, e.g. neural networks, were used to analyze data and then interpret (big data and Industrie 4.0) to predict which plant components must be maintained when. This makes it possible to extend unnecessarily short maintenance intervals, increase plant availability, and improve plants’ efficiency.