Finite Element Model Updating of a Skewed Highway Bridge Using a Multi-Variable Sensitivity-Based Optimization Approach

A Finite Element (FE) model which can precisely predict the actual dynamic responses of a bridge is necessary for accurate condition assessment and health monitoring purposes. However, most of the bridge FE models initially developed for design purposes may not be able to accurately simulate the current response of the bridges due to various reasons. Some of these reasons may be attributed to high level of model idealization, uncertainties involved in modeling complex structural components and changes in the boundary conditions or material properties from the time of construction. Therefore, the initial FE model of the bridge needs to be calibrated. This paper presents the implementation of a sensitivity-based model calibration procedure to minimize the error between the numerical and experimental modal properties of the Carquinez Bridge (officially named the Alfred Zampa Memorial Bridge); a 1.06 km long suspension bridge located in Vallejo, California. Twenty two Narada wireless sensing units were installed on the bridge to measure the ambient-induced vibration responses of the bridge. The modal frequencies and modal shapes were extracted from the measured vibration responses. Sensitivity analyses were performed to identify the most sensitive structural parameters of the FE model. A multi-variable sensitivity-based objective function was developed, and an optimization problem was solved to update the identified structural parameters through an iterative procedure. The updated FE model of the bridge provides improved modal properties which closely match with those from the measurements. The calibrated FE model is expected to generate much more accurate results for the post-earthquake evaluation of the structural performance.