Inspection and maintenance planning for offshore wind structural components: integrating fatigue failure criteria with Bayesian networks and Markov decision processes

Exposed to the cyclic action of wind and waves, offshore structures are subject fatigue deterioration processes throughout their operational life, therefore constituting a structural failure risk. In order control risk adverse events, physics-based models, which often contain significant uncertainties, can be updated with information collected from inspections, thus enabling decision-makers dictate more optimal informed maintenance interventions. The identified decision rules are, however, influenced by model criterion specified in formulation pre-posterior decision-making problem. this paper, criteria integrated Bayesian networks Markov processes. proposed methodology is implemented numerical experiments, various crack growth models criteria, for management an detail under deterioration. Within propagation, reliability estimates, policies derived through heuristics partially observable (POMDPs) thoroughly analysed, demonstrating capability assessment diagram redundancy substructures, as well adaptability POMDP policies.