A full-scale, seven-story reinforced concrete building slice was tested on the UCSD-NEES shake table. Six different state-of-the-art system identification algorithms including three input-output and three output-only methods were used to estimate the modal parameters (natural frequencies, damping ratios, and mode shapes) based on the measured response of the building subject to ambient as well as white noise base excitations at different damage states.
The identified modal parameters obtained using different methods were compared to study the performance of these system identification methods, and also to investigate the sensitivity of the estimated modal parameters to actual structural damage. The results obtained in this study were then used to identify damage in the building based on a sensitivity-based finite element (FE) model updating algorithm.
The damage identification results were verified through comparison with the actual damage observed in the test structure. Furthermore, the performances of the three output-only system identification methods as well as the FE model updating for damage identification were (numerically) investigated due to variability of different input factors such as amplitude of input excitation, spatial density of measurements, measurement noise, length of response data used in the identification process, and the modeling error.