PSI - Issue 78

Alessia Furiosi et al. / Procedia Structural Integrity 78 (2026) 753–760

757

; s c k E d k G d = = n c

(1)

The interaction between the masonry and the backfill was modeled using contact interfaces with stiffness values set to 20% of those of the backing and spandrel walls. Moreover, the contacts between masonry rigid blocks and backfill were assumed to have zero tension, zero cohesion, and a friction angle of 35° (Furiosi et al. 2025). For the contacts between the structure and the rigid blocks representing the boundary conditions (Fig. 3), the same mechanical properties as those of the adjacent component were assigned. The contact stiffness values assigned to the different block interfaces are summarized in Table 2. Note that only the values related to the main contact types are listed, as all other values can be directly derived from these. Further details on the values of d c adopted for the different backfill idealizations are provided in Section 4.1.

Table 2. Normal and shear stiffness values assigned to the block contact interfaces. Type of contact k n [MPa/m] k s [MPa/m] Arch vaults, spandrel walls, backing 2740 876 Backfill material (i.e., rigid or Voronoi blocks) 320 130

4. Sensitivity analysis This research work is part of the 2022–2023 and 2024–2026 DPC-ReLUIS and DPC-EUCENTRE projects, within which a discrete element model of the entire bridge structure was employed to perform multi-stripe nonlinear time history analyses aimed at generating case-specific seismic fragility curves. In these simulations, the backfill was modeled as a single deformable block to achieve an optimal balance between computational efficiency and numerical accuracy. Based on the results of these analyses, limit state threshold values were proposed for masonry arch bridges with comparable construction characteristics and mechanical properties (Furiosi et al. 2025, DPC-ReLUIS 2023). 4.1. Modeling strategies for backfill material In the present study, substructural models incorporating alternative backfill representations are developed primarily to evaluate the sensitivity of the bridge response to the numerical modeling of the backfill, an aspect that remains critical yet not fully explored, and subsequently to assess the consistency of the proposed limit state thresholds. This modeling strategy enables parametric variations while maintaining a reasonable computational cost, thus allowing for a more efficient investigation of the influence of backfill representation on structural performance.

Fig. 5. Three-dimensional discrete model of SUB-D, SUB-R, and SUB-V.