Buffeting induced fatigue damage assessment of long-span bridge decks under uncertain turbulence conditions

Fatigue in wind-sensitive structures is influenced by the mean wind speed and the turbulence characteristics. This study investigates the importance of uncertainty in the turbulence characteristics of buffeting-induced fatigue damage of flexible wind-sensitive structures, focusing on long-span bridges. We address the importance of the variability of uncertainty in the wind field by using a probabilistic wind field model, including uncertain turbulence parameters for calculating the accumulated fatigue damage throughout the structure’s design life. However, it has been demonstrated that a high computational load accompanies this approach. A new method is proposed that significantly reduces the computational effort through surrogate modelling. This is achieved through a sequentially updating Gaussian Process surrogate modelling approach, which integrates new simulation points into the training dataset and reduces the number of buffeting calculations required for the surrogate to converge to the true solution, thereby reducing the computational cost by almost 95%. The efficiency of the algorithm is demonstrated by using it to compute the fatigue damage accumulation through the design life for a long-span bridge girder, resulting in the increase of damage by several orders of magnitude in comparison to the conventional method of treating the turbulence parameters as deterministic and uncorrelated.