Developing Digital Twins to Characterize Bridge Behavior Using Measurements Taken under Random Traffic

This paper presents a method of developing digital twins (DTs) of road bridges directly from field measurements taken under random traffic loading. In a physics-based approach, the full three-dimensional behavior of the bridge is represented using response functions and distribution factors. In contrast to conventional finite-element analysis, this approach focuses on the relationship between the applied loads and the measured responses, given the limitations on the information about the applied loads due to random passing traffic. At the same time, it takes advantage of some key features of bridge traffic loading that are consistent, regardless of the weights of the passing vehicles. The nature of traffic loading is that axles travel from one end of a bridge to the other and the response is a linear combination of axle weights and ordinates of the same influence line function, adjusted for relative axle locations. Small/medium span concrete slab-girder decks are the target structures of the study. The three-dimensional nature of such structures is a particular challenge, especially in the case of multiple vehicle presence. While these bridges are strongly orthotropic, there is a significant degree of load distribution between the girders immediately under the passing vehicle and girders under adjacent lanes. This is addressed using an iterative approach that uses transverse distribution factors. The proposed DT model is verified using both numerical simulation and field tests.