Dynamic scheduling of maintenance measures in complex production systems

Kategorien Zeitschriften/Aufsätze (reviewed)
Jahr 2013
Autoren Denkena, B., Kröning, S.:
Veröffentlicht in CIRP Journal of Manufacturing Science and Technology, Volume 6, Issue 4, 2013, S. 292–300.
Beschreibung

In complex production systems, involving multiple, mutually dependent process steps, the unplanned failure of even a single machine can have significant consequences (e.g. loss of production) for upstream and downstream process steps and equipment. In order to detect maintenance needs before the machine components fail, condition monitoring systems are increasingly applied in practice. However, the main requirement for short-term scheduling of planned maintenance measures is to predict the impact of the ensuing downtime of the machine on the productivity of the whole production system. This is hardly possible for the production planner due to the dynamic and stochastic interactions (e.g. variation of buffer levels over time) of production processes. As a consequence, high production and maintenance costs tend to occur. Hence, a model-based planning approach which enables the dynamic short-term scheduling of planned maintenance measures has been developed. Instead of focusing separately on production and maintenance, this integrated approach allows a forecasting of the most cost-effective period of necessary maintenance measures for machines which are part of a complex production system. In contrast to existing, static approaches which fail at accurately representing the dynamic of production states and information, this approach aims to determine the various failure-related costs and to evaluate the impact of the maintenance measures on the production processes by means of event-driven simulation. The prediction accuracy is ensured by integrating the dynamic and stochastic correlations characteristic of real production systems into the simulation model. A concluding evaluation of the approach based on real application scenarios of an industrial company illustrates the need for a dynamic planning to lastingly reduce the production and maintenance costs of production systems compared to a purely static evaluation and planning based on average values (e.g. for transport time).