Probabilistic assessment of vibration exceedance for a full-scale tall building under typhoon conditions

Full-scale wind and vibration measurements were conducted on a 270-m tall building in Hong Kong. For the modal identification, a refined fast Bayesian fast Fourier transform method was newly developed by considering the exponential form of the modal wind force spectrum. Then, the modal parameters of the monitored building and their uncertainties were identified from the full-scale vibration data under typhoon conditions by the proposed refined fast Bayesian fast Fourier transform method. For assessing the vibration exceedance of the monitored building under typhoon, a stochastic wind force field generation technique was developed by combining the spectral representation method and the computational fluid dynamics method. Wind-induced vibration of the tall building under Typhoon Kammuri was reanalyzed with generated wind force time history data. Finally, a time-domain Monte Carlo simulation framework was proposed by integrating the modal identification and the stochastic wind force field generation technique. In this framework, the Metropolis-Hastings sampling algorithm was adopted to estimate the vibration exceedance probability, which was defined through a limit state performance function associated with the ISO vibration assessment criteria.