Nonlinear Identification of a Seven-Story Shear Wall Building Based on Numerically Simulated Seismic Data

Abstract

This study is focused on evaluating the performance of different approaches for nonlinear finite element (FE) model calibration of a seven-story shear wall structure based on its response during an earthquake. The seismic response of the structure is simulated numerically using a refined FE model of a specimen that was tested on the UCSD-NEES shake table in 2006. A simplified FE model of the structure with fewer modeling parameters is calibrated to match the ‘measured’ data. In this model, nonlinearity is defined by hysteretic models at fiber elements to represent the behavior of reinforced concrete walls. Calibration is performed by minimizing different types of residuals between measured/identified and model predicted response features (correlation features). Three types of correlation features used in this study consist of (1) time-varying modal parameters identified at different amplitudes of response, (2) response time-histories such as displacement and acceleration, and (3) a combination of response time histories and modal parameters. Accuracy of the updated models is studied by comparing the frequency and time responses of the calibrated models and the exact ones. The updated models are further simplified in the form of state-space nonlinear models suitable for real-time system identification. The parameters of the state-space models are finally identified by means of the Unscented Kalman filter (UKF).