Inelastic responses of wind-excited tall buildings: Improved estimation and understanding by statistical linearization approaches

Two statistical linearization approaches are used to determine the inelastic response statistics of wind-excited tall buildings with bilinear hysteretic restoring force character. Their accuracy and effectiveness are illustrated though the comparison with the predictions from response time history simulations. The statistical linearization approaches give good estimations of response standard deviation (STD), extreme value distribution and fatigue damage, while discrepancies are observed when the inelastic behaviour is significant that leads to hardening non-Gaussian response distribution. The results demonstrate that the inelastic response with zero mean component is lower than that of corresponding elastic system attributed to the increase in system damping resulted from the hysteretic restoring force. The ductility factor and reduction in response increase with the decrease in yield displacement, especially, for flexible tall buildings. On the other hand, when response has non-zero mean component under static load, the yielding results in displacement drift and the inelastic displacement is featured by stochastic drift and fluctuating components. The time-varying mean response can be calculated from state-space equation of motion, and its steady-state value is determined by the mean load and second stiffness. While the hysteretic damping results in reduction of the fluctuating response, the increase in the mean component can lead to total inelastic response noticeably higher than that of the corresponding elastic system. The results of this study help in developing improved understanding of inelastic building response under ultimate wind loads, contributing to achieve safer and more economical performance-based design of buildings beyond the current linear elastic framework.