Condition-based maintenance via a targeted bayesian network meta-model

This chapter suggests a condition-based maintenance (CBM)-based policy that combines both a simulation model of the system and a predictive meta-model. The simulation model of the system is based on expert knowledge and historical data. The chapter overviews the challenges of preventive maintenance optimization in reliability-availability-maintainability models and refers to several methods for addressing these challenges. The chapter also describes targeted Bayesian Network learning and the motivation for using a Bayesian Network as a meta-model for prediction. The chapter then provides a schematic framework for the implementation of the proposed CBM approach. It demonstrates the implementation of the proposed approach on a freight rail fleet based on a real case study of a European operator. The chapter further analyzes some key features of the proposed approach. It concludes with a short discussion on potential future directions.