Seismic response analysis and modal identification of a full-scale five-story base-isolated building tested on the NEES@UCSD shake table

Abstract

This paper presents the analysis of the seismic response and the identification of the dynamic properties of an equivalent linear time-invariant model of a full-scale five-story base-isolated reinforced concrete building tested on the unidirectional NEES@UCSD shake table. A sequence of seven scaled historic earthquake records was applied by the shake table to the building, progressively increasing the seismic demand on the structure and its nonstructural components and systems. The effects of the isolation system on elongating the predominant period of the building, concentrating the displacement in the isolation layer, and augmenting the energy dissipation capacity of the building system are investigated. Low-amplitude white noise (WN) base excitation tests were conducted before and after each seismic test and ambient vibration (AV) data were recorded continuously for a period of approximately sixteen days containing the sequence of seven seismic tests performed. Because of the low intensity of the white noise base and ambient excitations, a quasi-linear response of the base-isolated building is assumed, allowing the modal parameters of an equivalent linear viscously damped time-invariant model to be estimated from vibration data recorded during the WN and AV tests. Different state-of-the-art system identification methods, including output-only and input-output methods, are used to estimate the modal properties of the base-isolated building. Results show that the identified modal parameters obtained from different methods are in good agreement and that the assumption of quasi-linear response for each low-amplitude test is appropriate. The effects of the amplitude of the excitation, earthquake input motions, and environmental conditions are clearly evidenced by the changes induced in the estimated modal parameters of the building.