Probabilistic Model for Remaining Fatigue Life Estimation of Bridge Components

Structural health monitoring provides a measurement-informed quantitative and systematic framework to better assess the state of fatigue in aging infrastructure. In this paper, a general methodology for probabilistic modeling of fatigue damage accumulation is presented from a structural health monitoring perspective, which allows for estimating the remaining fatigue life from strain measurements. The method presented uses a damage-sensitive feature derived from the Palmgren–Miner rule—a commonly used measure in fatigue analysis and design—as a surrogate for degradation and probabilistically models the process of degradation. A Bayesian approach is employed to estimate the parameters of this degradation model, and the remaining fatigue life is predicted using Markov chain Monte Carlo (MCMC) simulations. In the simplified case of constant stress amplitude, an analytical solution for the fatigue life has been derived. Moreover, techniques are presented to account for various challenges such as the lack of structural health monitoring data (SHM) for the initial unmonitored period and the absence of the continuous monitoring program. The main contribution of the paper is to develop a probabilistic degradation modeling framework using a damage-sensitive feature derived from the Miner rule for the prediction of the remaining fatigue life of critical bridge components. A numerical case study is presented to demonstrate the proposed methodology, and the major challenges of its implementation in the field are demonstrated through illustrations.