Using bridge weigh-in-motion system for structural health monitoring of bridges

Ageing transport infrastructure and increasing traffic loads increase the demand for more accurate and detailed structural assessments supported by more and more data. One advanced method addressing this need is structural health monitoring (SHM), as a complement to classic bridge inspections. SHM is often associated with finite element model updating (FEMU), a key to improving finite element (FE) models by aligning their performance with measured structural responses. Traditionally, FEMU relies on structural dynamic properties, such as natural frequencies and mode shapes obtained from acceleration data. In contrast, static data sets, such as traffic-induced displacements and strains of the structure, are used less frequently due to logistical and cost-related challenges. This thesis investigates the potential of employing strain measurements and strain influence lines (ILs), derived from bridge weigh-in-motion (B-WIM) measurements and originally designed for traffic load monitoring, for FEMU purposes. A series of studies was conducted on two multi-span concrete roadway bridges equipped with B-WIM and SHM systems. Various FEMU strategies, including residual minimisation (RM) and error-domain model falsification (EDMF), were applied, using both structural dynamic properties and static data sets, to update FE model parameters such as stiffness and boundary conditions of structural elements. The thesis provides new insight into the role of manual FEMU, the limitations regarding the number of parameters to be updated, the influence of non-structural elements, the challenges of combining static data sets and structural dynamic properties for FEMU and the validation of updated FE models with independently measured displacements. Importantly, it demonstrates the integration of B-WIM-measured strains and B-WIM-derived strain ILs into FEMU workflows. The results suggest that the measurements on bridges, which a B-WIM traditionally uses to calculate vehicle loads, can be effectively repurposed for FEMU. The two investigated parameters were the maximum measured traffic-induced strains and B-WIM-derived strain ILs. Combining structural dynamic properties and static data sets proved challenging and sometimes yielded contradictory FEMU results. This is a critical reminder for future FEMU studies that rely solely on structural dynamic properties. Finally, the important conclusion of the studies was that the updated stiffness of both superstructures was more than 20% higher than in the original design documentation, demonstrating a clear reserve in their performance. Additionally, at the considered load levels, the concrete safety barriers contributed not only to the additional mass but also to the stiffness of the structure.