Updating structural properties using modal parameters considering measurement errors

Evaluating up-to-date structural properties of an existing bridge is essential to improve the accuracy of the assessment for the structural performance. In this study, we propose a methodology to update the structural properties of a bridge using modal parameters considering measurement error. Modal parameters are usually obtained from a vibration testing conducted on a bridge. Time domain decomposition (TDD) is used in this paper to extract modal parameters from time history acceleration responses. The measurement error from the extraction process and from the noises in the acceleration sensors are estimated and considered in estimating the structural properties. There are five steps in the proposed framework: 1) extract modal parameters from a vibration test; 2) build a preliminary finite element model (FEM) based on the bridge design drawings; 3) estimate the measurement error by simulating vibration tests; 4) update the preliminary FEM to a baseline model using a Bayesian model updating approach with mode shapes; and 5) identify local structural stiffness damage using the damage index method with model shapes of the existing bridge and of the baseline structure found in the previous step. The proposed framework is constructed using the Bayesian model updating and the damage index method. The proposed framework is illustrated with a simulated finite element model of the Lavic Road Overcrossing bridge, a typical box-girder RC highway bridge in California.