Collapse reliability of wind-excited reinforced concrete structures by stratified sampling and nonlinear dynamic analysis

Performance-based wind engineering (PBWE) is gaining consensus as an alternative to current wind design procedures for its potential to allow the explicit evaluation of system performance over a full range of performance objectives including collapse. To achieve the maximum benefits from PBWE, advanced computational approaches that enable the efficient prediction of the inelastic performance, as well as probabilities of failure under extreme winds, are essential. In this paper, a stochastic simulation-based framework is proposed to efficiently estimate the reliabilities/probabilities of failure of reinforced concrete (RC) structures subject to a full range of wind intensities. The framework is based on integrating a high-fidelity nonlinear modeling environment with a wind-tunnel-informed stochastic wind load model. An optimal stratified sampling scheme is adopted to propagate structural and load uncertainties and subsequently estimate the small probabilities of failure characterizing collapse with limited computational efforts. Through the illustration on a 45-story archetype RC building with a hypothetical location in New York City, the inelastic behavior and the subsequent reliabilities associated with various collapse limit states of interest are investigated. In particular, typical collapse mechanisms for the RC structures under extreme winds are explicitly captured by evaluating the progression of the localized damage, including reinforcing bar yielding, fracture, low-cycle fatigue failure, and concrete crushing.