Fusing Modal Parameters and Curvature Influence Lines for Damage Localization Under Vehicle Excitation

Transport infrastructure serves as a vital component for facilitating the movement of people and goods, contributing significantly to sustainable development and territorial cohesion. However, bridges, which are crucial elements of transportation networks, are increasingly susceptible to degradation caused by growing traffic volumes and severe weather events. Recent research efforts have focused on developing damage-sensitive features specifically designed for bridges, with curvature being one of the most used indicators. Traditionally, curvature estimation requires multiple instrumented locations. However, dense networks introduce challenges related to data management, synchronization, and battery replacement. This study presents an algorithm that enables the identification of dense curvature estimates for civil infrastructure using a limited set of accelerometers. The algorithm combines sparse curvature estimates derived from modal parameters with curvature influence lines obtained from the vibration response recorded during vehicle passage. A Kalman filter is employed to fuse these two features, mitigating the error resulting from traffic variations. The effectiveness of the proposed method is demonstrated using data collected from a real steel truss bridge with artificially induced damage.