Probability of detecting damage in simply supported beams from corrupted midspan displacement due to a crossing quarter-car

This paper presents a comprehensive numerical investigation into the probability of detecting damage in simply supported bridge beam models utilizing how the mid-span displacement responds to a crossing quarter-car vehicle model. Some forms of damage can be modelled as a stiffness loss. However, the displacement can change not only due to stiffness loss, but also because of factors external to the bridge, such as vehicle, road, and environmental parameters, and noise associated with the response measurement. Changes in external factors between two different instants in time have the potential to lead engineers astray by suggesting the existence of damage that may not be present or, more concerning, by causing them to overlook actual damage. Furthermore, the probability of detecting damage is significantly influenced by its characteristics, i.e., location and severity, and by bridge parameters including dimensions, section properties, material characteristics, stiffness and mass distribution, and boundary conditions. This research quantifies the impact of one of these factors, i.e., noise, on detecting a global stiffness loss, via probability of detection curves, offering a graphical representation of the efficacy in damage detection as a result of the interplay between a specific vehicle-road-bridge interaction scenario and three noise levels. The detection rates are calculated based on the raw simulated response, without an attempt to mitigate the negative influence of noise prior to the analysis. These curves serve as a reference to assess the extent to which damage detection methodologies can improve the reliability of results when external factors are addressed