PSI - Issue 64

1970 E. Michelini et al. / Procedia Structural Integrity 64 (2024) 1967–1974 4 Elena Michelini, S ł awomir Dudziak, Simone Ravasini, Beatrice Belletti / Structural Integrity Procedia 00 (2019) 000–000

A general view of the two numerical models (in ABAQUS and Seismostruct) with the assumed boundary conditions and mesh discretisation is shown in Figure 2(a), whereas the force – displacement curves obtained from FE analyses are compared to the experimental results in Figure 2(b). Numerical models properly predict the initial stiffness, the reinforcement yielding (i.e., flexural resistance), and the ductility of the tested frame (the latter is slightly overestimated by the ABAQUS model, probably due to the simplified material relation assumed for concrete). The biggest discrepancies between numerical and experimental outcomes can be observed in the phase governed by the arch action, while the catenary action is correctly reproduced. Points associated with the reaching of the LS listed in Table 1 (which are defined for low-code buildings) are indicated on the numerical curve obtained with Seismostruct for informative purposes, even if the frame was designed according to modern design standards. LS2 and LS3 are governed by the reaching of limit values for steel strains. As can be seen, the reaching of LS4 – based on sectional analysis – neglects the beneficial effects related to the development of the catenary action, so the assumed threshold value for steel strain, which is taken from the literature for low-code buildings, appears to be very conservative.

(a) (b) Fig. 2. Validation of the modelling methodology: (a) FE models (from the top: ABAQUS and Seismostruct), (b) comparison of force vs. displacement curves for experimental data and outcomes of FE analysis 4. Damage assessment of an existing Italian school building subjected to large column settlements Numerical analyses have subsequently focused on the “A. De Gasperi – R. Battaglia” school in Norcia. The latter, which is characterized by a RC framed structure, has been chosen as case-study building due to the large number of information available (Lima et al., 2020; Belletti et al., 2022). The building is formed by three blocks, which are separated by two technical gaps along the shorter dimension; only the left part is analyzed in this work. As depicted in Figure 3, its in-plan layout and vertical distribution can be considered as representative of a large number of Italian RC school dating back to the same construction period (beginning of the ‘60s). Some details regarding the cross section dimensions and the reinforcement layout of columns and beams are reported in the same Figure 3. The floor slabs are made of RC joists with hollow clay blocks and are arranged perpendicular to the transverse frames. Further details on the building, here omitted for brevity, can be found in the cited literature. Numerical analyses are focused on the internal transverse frame along alignment #4 (see Fig. 3), which is subjected to the two different settlement scenarios sketched in Figures 4(c), (d), respectively involving the central column, and the lateral one. Two models of the frame are preliminarily analysed; i.e., considering the presence of inclined roof beams and neglecting them, as depicted in Figures 4(a), (b), which also show the adopted FE mesh. In the case of roof beams, internal hinges are assumed, whose position is shown in Figures 4(c), (d). The parameters adopted for material models are based on the available technical documentation, as follows; reinforcement: E s = 200 GPa, f y = 375 MPa, f t = 450 MPa, ε u = 0.045; concrete: f cm = 25.2 MPa, f ctm = 2.0 MPa, E ci = 22 GPa, ε c1 = 0.0022. The following loading values are assumed for a typical floor: dead load 5.1 kN/m 2 , live load 3 kN/m 2 . The accidental load combination according to Eurocode 0 (1.0 dead load + 0.7 live load) is used to determine the values of uniformly distributed loads applied to the main beams. In the case of the model without roof, concentrated forces are applied to the top nodes, considering the total load acting on roof beams and their own weight. Fixed supports are introduced at all columns’ bases. During the analyses, the total inclination of the side column ( α ) and chord rotation of some selected beams ( θ 1