Estimation of Inelastic Crosswind Response of Base-Isolated Tall Buildings: Performance of Statistical Linearization Approaches

The base-isolation system is increasingly adopted for tall buildings for which the wind-induced response is of concern. Previous studies have shown that the wind-induced response can be reduced as compared to a fixed-base building when the hysteretic damping generated by yielding of base-isolation system is noticeable. The yielding also causes the responses of the base-isolation system and structure to have non-Gaussian probability distributions. This study examines the accuracy and effectiveness of statistical linearization approaches for estimating stochastic inelastic crosswind response of base-isolated tall buildings via comparison with response-history analysis for a wide range of structural parameters. The hysteretic base-restoring force is represented in a Bouc-Wen model in terms of hysteretic displacement, base displacement, and velocity. The linearization is to represent the hysteretic velocity as a linear function of hysteretic displacement, base displacement, and velocity. Both Gaussian and non-Gaussian statistical linearization approaches are investigated, where these responses are assumed to follow Gaussian or non-Gaussian probability distributions. To capture the higher-mode contribution to building acceleration, the building motion state variables are also included in the linearization of hysteretic velocity. The results show that the Gaussian linearization approach is able to give quite accurate estimations of building top displacement and base shear force in a wide range of wind speeds or ductility factors of base displacement. However, it underestimates the base displacement and building top acceleration at higher wind speeds. Consideration of a non-Gaussian distribution of hysteretic displacement and inclusion of additional building-motion state variables in the linearization lead to much improved estimation of building acceleration. A further consideration of the non-Gaussian distribution of base velocity results in improved estimation of base displacement.