Probabilistic modelling of fatigue damage accumulation in steel bridge decks under stochastic and dynamic traffic load

Abstract

A realistic traffic simulation is a premise of the fatigue damage estimation. This study presents a novel computational framework for probabilistic modelling of fatigue damage accumulation in steel bridge decks under stochastic traffic load. The stochastic traffic load is simulated based on bridge weigh-in-motion measurements, and the dynamic effects are solved in a vehicle-bridge coupled vibration system. A response surface method is utilised to replace the finite-element analysis. Probability models of the fatigue stress ranges are finally used for fatigue reliability evaluation of an orthotropic steel bridge deck. Numerical results show that the stress spectrum under the actual traffic loads follows a bimodal distribution. Both the fatigue reliability index and the fatigue life decrease rapidly under a continuous traffic growth. A strict threshold overload ratio reduces the decreasing range of the reliability index caused by traffic growth.