PSI - Issue 78

Maria Eleonora Pipistrelli et al. / Procedia Structural Integrity 78 (2026) 1911–1918

1916

3. Numerical modelling The unit was analyzed by evaluating its seismic response under three different configurations: a first configuration in which the building is considered isolated and subjected to an increased seismic action, to simulate the influence of the adjacent portions of the aggregate; a second configuration in which boundary conditions are modeled using constraints that represent the stiffness of the walls of the surrounding structural units in contact with its perimeter, according to the procedure shown in C. Valotto et al. (2016); a third configuration in which the unit is modeled together with the two adjacent portions of the aggregate. A length of 10 meters is considered as representative of the adjacent cells. EFM approach has been used and the configurations have been implemented in 3Muri software, introducing the following simplifications to reduce computational efforts. Wall segments belonging to the same wall in the actual building have varying thicknesses, which were unified in the model and represented using an average thickness, with openings inserted to simulate the presence of niches. Wall alignments were simplified to orthogonal directions. Openings on the north wall were realigned while maintaining the original configuration as much as possible, in order to avoid meshing issues. Interior partition walls were not explicitly modeled, but their effect was included as additional permanent structural loads applied to the floor slabs at each level, except for those masonry elements on the ground floor, which were modeled to account for the variety of floor systems present. In the EFM models, vaults were not modelled. Instead, floors with equivalent stiffness have been used. The three developed configurations are shown in Fig. 5, with the different colors corresponding to the different materials.

Figure 5: S.U. numerical model configurations: (a) isolated; (b) with constraints; (c) with portions of the adjacent S.U.s.

4. Results

4.1 Capacity curves Pushover analysis was performed on the three configurations, using load profiles consistent with Italian building standards. The resulting weighted average drifts were recorded at a single, centrally located control node on the top slab. This node was chosen to average drifts across multiple roof nodes and to minimize additional noise during comparisons. Fig. 6 presents the capacity curves derived from the pushover analysis that led to the maximum shear capacity for each configuration in both directions. Capacity and drift values are normalized by the total mass of the model and the height of the control node, respectively. A significant increase in X-direction capacity is observed from Configuration 1 to Configuration 3. Taken together, these three pushover curves quantitatively illustrate how progressively more realistic modeling of adjacent units systematically increases base shear capacity. These findings align with other similar studies, which observe that partial models increase ultimate load-carrying capacity compared to completely isolated models. An opposite trend is observed for ultimate displacement.