Probabilistic aerostability capacity models and fragility estimates for cable-stayed bridge decks based on wind tunnel test data

Probabilistic capacity models are developed for cable-stayed bridges considering flutter and aerostatic instability failure. Additional probabilistic models for implicit design variables are also developed to make the capacity model in explicit expression of only basic design variables and model parameters. All the probabilistic models are constructed to incorporate the understanding of the physics/mechanics of the phenomena by considering existing deterministic models and to include information from experimental wind test data via correction terms. The developed models are constructed to give unbiased estimates of the capacities of interest and properly account of the relevant uncertainties. The measured capacity values from wind tunnel tests are used to construct posterior statistics of unknown model parameters through a Bayesian approach. Using the developed capacity models, fragility estimates for flutter instability failure, aerostatic instability failure and system instability failure are obtained for three example bridges. Fragility is defined in this paper as the conditional probability of failure for given wind speed demand.