Probabilistic optimization framework for inspection/repair planning of fatigue-critical details using dynamic Bayesian networks

Bridges, ships, and other civil and marine structures are subjected to fatigue damage due to repeated load fluctuations. Fatigue damage is likely to jeopardize the functionality and even structural safety of these structures. Therefore, inspections and timely repair actions are needed to ensure adequate structural performance throughout their lifetime. Nevertheless, inspection/repair actions involve additional life-cycle costs. Therefore, efficient planning of inspection and repair actions of fatigue-critical details are not only essential to ensure structural functionality and safety, but also important to control the total life-cycle cost. In this paper, a novel framework for optimizing inspection/repair planning is developed by using efficient Bayesian updating with dynamic Bayesian network. Specifically, inspection plans, including inspection schedules and inspection techniques, are optimized using pre-posterior analysis. Decisions of repair actions are made based on inspection results following an evidence-informed and cost-driven repair strategy. This strategy allows for time-dependent and adaptive repair actions considering fatigue damage development, available inspection results, and previous repair actions. Optimal inspection/repair plans with the lowest expected life-cycle cost are then obtained using both single- and multi-objective optimizations.