PSI - Issue 78

Paolo Ielpo et al. / Procedia Structural Integrity 78 (2026) 1024–1031

1027

distributed longitudinal bars. The column is reinforced with three Ø14 mm bars per side, while the beam includes a total of eight longitudinal bars symmetrically arranged in the top and bottom sections: two Ø14 mm and two Ø12 mm bars per side (Fig. 2). Material mechanical properties were determined through laboratory tests: the concrete exhibited an average compressive strength of fc = 21.5 N/mm², and the improved-bond B450C steel reinforcement showed an average yield strength fy,med = 480 N/mm². The specimen was subjected to a quasi-static, displacement- controlled cyclic test. An axial load of 580 kN was applied at the top of the column, while an incremental horizontal cyclic load was imposed at a height of 3 meters from the base, with three cycles per predefined drift level. The test was carried out until structural collapse occurred. For specimen T4, failure was governed by a purely flexural mechanism, characterized by cracking at the beam-column interface and local instability of the bottom longitudinal reinforcement, observed at approximately 3% drift. For more details, refer to the full experimental documentation in Masi et al. (2013).

Fig. 2. schematic representation of the test setup along with technical details of specimen T4

4. 3D nonlinear finite element numerical model Following the experimental test, a nonlinear finite element model was developed using the Atena 3D software (version 5.4.1). The adopted workflow initially involved the definition of the geometry, with the beam and column modelled using solid (volumetric) elements, while the longitudinal reinforcements and stirrups were represented by one-dimensional elements. Subsequently, the material properties were characterized, and appropriate constitutive laws were defined. Specifically, for the concrete, a nonlinear constitutive model under plane stress conditions was employed. Particular attention was given to modelling concrete damage through a smeared crack approach, with fracture energy Gf defined according to the formulation provided by the Model Code 2010. This parameter is crucial as it directly affects the accuracy of the model, representing the area under the exponential curve that describes crack propagation after exceeding the maximum tensile strength. The latter was not measured experimentally but calculated based on the characteristic compressive strength of the concrete. Additionally, the bond – slip behaviour between concrete and reinforcement within the joint was modelled using the approach outlined in the Model Code 2010, which allows representation of bond effects and the loss of interaction between steel and concrete under advanced damage conditions. The modelling phase involved defining three loading steps to replicate the experimental conditions see Fig. 3. Due to numerical convergence issues, it was not possible to implement a cyclic loading function.