Maintenance, repair and overhaul processes (MRO processes) are elaborate and complex. Rising demands on these after sales services require reliable production planning and control methods particularly for maintaining
valuable capital goods. Downtimes lead to high costs and an inability to meet delivery due dates results in severe contract penalties. Predicting the required capacities for maintenance orders in advance is often difficult due to unknown part conditions unless the goods are actually inspected. This planning uncertainty results in extensive capital tie-up by rising stock levels within the whole MRO network.
This paper outlines an approach to planning capacities when maintenance data forecasting is volatile. It focuses on the development of prerequisites for a reliable capacity planning method. This is achieved by deriving order probabilities from Bayesian networks. In addition the residual wear margin is
used to determine an appropriate part-dependent work content.