Digital Twin-driven framework for fatigue lifecycle management of steel bridges

This paper presents a Digital Twin-driven framework for fatigue lifecycle management of steel bridges. A probabilistic multi-scale fatigue deterioration model is proposed to predict the entire fatigue process of steel bridges. Bayesian inference of the deterioration parameters realizes the real-time updating of the predicted lifecycle fatigue evolution process, which provides a good basis for lifecycle optimization. To avoid an empirically predefined repair crack size for maintenance, an optimization process for maintenance strategies is included. The relationship of the extended lifetime and the design repair crack size is constructed by numerical experimental design and surrogate modeling. The solution for optimum repair crack size is obtained while maximizing the extended fatigue life and minimizing the maintenance costs. Based on the occurrence time distribution of the optimum repair crack size, the inspection/monitoring planning is determined from a probabilistic optimization process based on the minimization of the expected damage detection delay and the lifecycle costs. The uncertainties associated with the damage occurrence and detection ability are considered during the formulation of the expected damage detection delay by decision tree analysis. Based on Digital Twin concept, the predicted deterioration process, derived maintenance, and inspection/monitoring planning are timely updated until a defined stopping rule is met.