PSI - Issue 44

Michele Angiolilli et al. / Procedia Structural Integrity 44 (2023) 870–877 M. Angiolilli et al./ Structural Integrity Procedia 00 (2022) 000 – 000

875

6

Moreover, based on empirical data, Vollum and Newman (1999) proposed a formulation that provides a shear strength of 729 kN. Finally, the formulation proposed by Bakir and Boduroglu (2002) led to a shear strength of 535 kN. In light of the experimental data presented in this paper, it can be concluded that all formulations given in the literature have a small to significant overestimation of the shear strength. The NTC (2018) result came the closest to matching the experimental outcome. 4. Preliminary numerical investigation Extensive research on numerical modelling and analysis focused on RC joints has been carried out (e.g. Grande et al. 2021, Hakuto et al (2000), Hassan and Elmorsy (2022), Hu and Schnobrich (1990), Pampanin et al. (2003)). However, reliable prediction of the structural performance of unreinforced beam-column joints, especially under cyclic load, is still a challenge, which justifies the various methods proposed in the literature (see Nicoletti et al. (2022), Lima et al. (2012a, 2012b)). In this study, a numerical model was developed using the concrete smeared crack (CSC) model implemented in the commercial software MIDAS FEA NX v1.1. Note that the CSC was already successfully simulated RC members under seismic loads (e.g. Asgarpoor et al. (2021), Di Carlo et al. (2017), Earij et al. (2017)). With respect to the discrete models, in which the cracks are explicitly modelled merely through the separation of particles when tensile strength is reached, in the CSC approach, the extension of cracks is predicted by using the concept of fracture mechanics and studying the stress concentrations at the crack tip. Constitutive calculations are performed independently at each integration point of the finite element model. In particular, the concrete structure is modelled with eight-node solid brick elements, while the steel rebars and ties are defined as two-node one-dimensional elements. Fig. 5 shows the numerical model geometry and steel reinforcements. The steel plates at the ends of the beams and at the bottom of the column were simulated by rigid links.

Fig. 5. Numerical model with the indication of the geometry and steel rebar details.

Figure 6 shows the numerical curve obtained by applying a monotonic displacement, separately in the two directions. First, one can see a good prediction of V max (equal to 231 kN; 6% overestimation). The main difference with respect to the experimental response regards the slope of the numerical curve, which is higher as compared to the experimental one. This is mainly due to the test type analysis performed for the experimental or numerical tests (i.e. cyclic vs. monotonic) as well as a possible bar-slip phenomenon. The first important stiffness degradation observed numerically is associated with the attainment of V =67 kN (see Fig. 6) that is investigated below.