ABSTRACT
Ensuring the structural integrity of bridges is of paramount importance for public safety and the longevity of these critical infrastructures. Early detection and assessment of damage are essential to prevent catastrophic failures and optimize maintenance efforts. Vibration-based methods have emerged as effective tools for structural health monitoring (SHM) in bridge structures due to their non-intrusive nature and sensitivity to subtle changes in structural behavior. In this study, we present a comprehensive comparative analysis between two prominent vibration-based (VB) approaches: modal parametric and nonparametric methods, for damage characterization in bridges. The modal parametric approaches rely on interpreting the variation of the dynamic response of the structure based on identified modal parameters like natural frequencies and damping ratios sensitive to damage. While modal parametric approaches require system identification for damage characterization, these non-parametric methods involve extracting features from raw signals that are sensitive to damage, utilizing time series modeling techniques. After conducting a brief and comprehensive literature review, data collected from tests conducted on bridge cables were analyzed using both modal parametric and data-driven VB damage identification techniques. We then evaluate their performance in terms of sensitivity to damage characterization.
