Probabilistic Fatigue Life Estimation of Steel Bridges by Using a Bilinear $\mathsf{S}\mathrm{\text{-}}\mathsf{N}$ Approach

This paper focuses on estimating the fatigue life below the constant amplitude fatigue threshold (CAFT) of steel bridges by using a probabilistic approach on the basis of a bilinear stress life (i.e., the $S\mathrm{\text{-}}N$ approach). The current AASHTO $S\mathrm{\text{-}}N$ approach uses a single $S\mathrm{\text{-}}N$ line for predicting the fatigue life. However, because of the variation of actual applied live-load stress cycles, this approach very often results in a severe underestimation of the useful life of structures. It implies that fatigue damage in respective structural steel details may be overestimated. To improve fatigue life estimation, a bilinear $S\mathrm{\text{-}}N$ approach is integrated into a probabilistic framework that can model the uncertainties associated with the fatigue deterioration process. In this approach, the equivalent stress range is computed by using two $S\mathrm{\text{-}}N$ slopes and several probability density functions associated with stress ranges. These probabilistic functions are determined on the basis of stress-range bin histograms from long-term monitoring. An existing bridge that is expected to experience finite fatigue life is used to illustrate the application of the proposed approach.