PSI - Issue 78

Eleonora Bruschi et al. / Procedia Structural Integrity 78 (2026) 49–56

52

3. OpenSees numerical analyses 3D numerical models of the structures are developed using the OpenSees framework (McKenna et al., 2000; OpenSeesWiki). Beams and columns are modeled with the forceBeamColumn element in its beamWithHinges formulation (Scott and Fenves, 2006). This approach assigns linear elastic behavior to the interior sub-element and non-linear behavior to the two external sub-elements, whose length corresponds to the plastic hinge length calculated according to EN 1998 – 1. The plastic hinge regions are modeled with fiber sections, where each reinforcing bar is represented by an individual fiber. Steel behavior follows the Giuffré – Menegotto – Pinto model - Steel02 in OpenSees – (Menegotto and Pinto, 1973; Filippou et al., 1983) with isotropic strain hardening. Concrete behavior is modeled using the Concrete04 material, based on Popovics' model (1973). Section properties for the confined core are evaluated in accordance with EN 1998-1 neglecting the tensile strength of concrete in both core and cover regions (Bruschi et al., 2021; Barbagallo et al., 2020). Material properties are assigned disregarding the confidence factors typically used for existing buildings (D.M. 2018; EN 1998-1). To represent the effect of cracking in the elastic sub element, the effective moment of inertia is set to 50% of the gross section inertia I , consistent with current European and Italian codes (D.M. 2018; EN 1998-1). This modeling approach has been validated against the seismic response of RC members dominated by flexural behavior, showing sufficient accuracy for performance assessment (Bruschi et al., 2021). Masses are concentrated at the floor center of mass, considering the contribution of beams, columns, and slabs. Dead and live loads are distributed based on the tributary area method. P-Delta effects are included, while bond-slip and low-cycle fatigue are neglected. Ground-floor columns are modeled with fixed-base supports, simulating rigid foundations. Rayleigh damping is applied based on the tangent stiffness matrix, assuming a 5% viscous damping ratio to account for energy dissipation from non-structural components such as infills (Faleschini et al., 2019; Bruschi et al., 2021). Floor diaphragms are modeled as rigid, imposing equal in-plane displacements to all nodes at each story level. To prevent the development of artificial axial forces due to the interaction between fiber-based beam elements and the rigid diaphragms, “axial buffers” are introduced using zeroLength elements (Barbagallo et al., 2020). These elements have negligible axial stiffness and very high stiffness in shear and bending, effectively acting as axial releases. Braces equipped with hysteretic dampers are modeled as truss elements associated with an elastic-perfectly plastic material model (OpenSeesWiki). Seismic actions are defined according to the D.M. 2018 provisions for the sites of Benevento (soil type B, category T1) and Tramutola (soil type C, category T1). Three seismic limit states are considered: 1. LLS (life safety requirement): 10% probability of exceedance in 50 years ( =475 years) 2. CLS1 (non-collapse requirement 1): 5% probability of exceedance ( =975 years) 3. CLS2 (non-collapse requirement 2): 2.5% probability of exceedance ( =1950 years), aligned with ASCE/SEI 41-17 For each site and limit state, seven pairs of bidirectional natural ground motion records are selected from the European Ground Motion Database (Ambraseys et al., 2002) using the Rexel tool (Iervolino et al., 2010). The records are scaled to match the D.M. 2018 target spectrum for a nominal building life of 50 years (functional class II) with a damping ratio of 5%, within the period range 0.15 – 2.00 s and tolerances of −10%/+30%. Selected events have magnitudes between 5 and 7 and epicentral distances R ep between 0 and 30 km. Each bidirectional ground motion is applied in two orthogonal configurations, switching the components relative to the building’s principal axes to avoid any possible directionality effect. 4. Discussions of the results: structural reliability and damper reliability The increased displacement demand for the dampers under ground motions of higher intensity than those considered in the design is evaluated by means of the amplification index β , defined as the ratio between the damper displacement Δ TR975years or Δ TR1950years calculated under the seismic action effect at either CLS1 or CLS2, respectively, and the displacement Δ TR475years under the seismic action effect at LLS, that is assumed as reference. In order to distinguish between the two return periods, the labels β 1 and β 2 are introduced, which refer to the ratios Δ TR975years /