Managing infrastructure asset: Bayesian networks for inspection and maintenance decisions reasoning and planning

Models of maintenance problems must handle complex assumptions, allowing, for example, the condition some assets to be rated directly using multiple states while in others rating is inferred from that components which they are assembled. The overall inferred, informs decisions, requires evidential reasoning under uncertainty. This paper uses Bayesian networks address these challenges with real case studies. We apply binary factorisation technique allow inference multi-state prediction, and further extend it predict an asset components. These models used recommend inspection decisions such as inspect when them. also developed evaluate effectiveness repair interventions use this suggest actions. show how model within life cycle considering both deterioration. modelling allows us plan activities over its whole cycle.