Structural health monitoring and condition assessment of bridge structures

The work described in this thesis is concerned with monitoring-based condition assessment of bridge structures, including vibration-based damage detection and structural reliability evaluation, modelling of temperature-frequency correlation, and condition assessment of bridge expansion joints.; A comparative study is made of two regularization methods and three regularization-parameter-optimization approaches for treatment of the ill-conditioning in finite element (FE) model updating. Minimum product criterion is shown effective and robust in selecting appropriate regularization parameters for the two regularization methods.; A two-stage stochastic model updating method is proposed to accommodate the uncertainties in the measured modal parameters. An improved perturbation method and the Monte Carlo Simulation (MCS) method are used to perform the first-stage updating. At the second-stage updating, the first-stage updating results are combined with the prior information about updating parameters via Bayesian theory to achieve the posterior estimator. The performance of perturbation and MCS methods for stochastic model updating under a variety of uncertainties is compared in some detail.; Using the stochastically updated model, structural reliability theory is applied to determine the reliability index for the predefined limit state. With the obtained reliability index, rational maintenance strategies can be laid down according to the correspondence between reliability index and required maintenance action. Following this approach, structural health monitoring is able to provide quantitative information regarding bridge inspection and maintenance.; A combined method of principal component analysis (PCA) and support vector regression (SVR) is proposed to characterize the temperature-frequency correlation using long-term monitoring data. The performance of the formulated SVR models with the SVR hyper-parameters determined by a grid search method and a heuristic method, respectively, is examined. Both the 'dynamic' SVR model accounting for thermal inertia effect and the 'static' SVR model without considering the effect are compared. The proposed PCA-SVR model is compared with the conventional SVR model and multivariate linear regression model.; A procedure for the assessment of bridge expansion joints by using long-term monitoring data is developed. The developed procedure enables maintenance and replacement of expansion joints to be made based on their actual condition.