PBSHM - guidance for ensuring quality when creating an IE model for a bridge

Population-Based Structural Health Monitoring (PBSHM) is a new approach in the Structural Health Monitoring (SHM) community that looks to open up the opportunity of data sharing between structures. Many SHM practices that use machine learning approaches require substantial amounts of training data to be available, which can be challenging for bridge structures. PBSHM is being newly applied to bridges, and this paper presents guidance on additional measures that can be taken to ensure accurate models of bridges and other large structures. The PBSHM approach works by comparing Attributed Graphs (AGs) of structures, using graph theory, to assess levels of similarity, and these comparisons are sensitive to changes in structural topology. The AGs are generated from Irreducible Element models of the structures, where it is important to capture each structure's topology accurately for reliable comparisons to be achieved. Therefore, this paper proposes a naming convention for individual elements of the Irreducible Element model that facilitates the implementation of quality checks on the mapped joints, and methods for calculating the 3D spatial coordinates of the joints to accommodate aspects of bridge geometry such as slope, skew or curvature in the elements.