Damage detection in bridges based on patterns of dynamic amplification

The pattern of Dynamic Amplification Factor (DAF) of the bridge strain response to a moving vehicle versus vehicle velocity is used to develop a level I damage technique. The challenge is to detect damage that causes only a small and difficult to detect frequency change with respect to the healthy condition. For this purpose, a damage index is defined based on subtracting the DAF-velocity pattern for the bridge in a prior healthy state from the DAF-velocity pattern corresponding to the damaged bridge. Simulations from a 3D vehicle-bridge interaction model are employed to show how the index increases with damage. The influence of the location of the strain sensors, the location, and severity of the damage, the road roughness, the corruption of measurements by noise, and the velocity range on the robustness of the technique are analysed. The relative changes in the proposed index as a result of damage are shown to clearly outperform the associated relative changes in frequencies, even for measurement locations far apart from the damage.