PSI - Issue 78

Marco Gaetani d’Aragona et al. / Procedia Structural Integrity 78 (2026) 968–975

972

On the other hand, fibers pertaining to the effective concrete core, adopts stress-strain relationships from Mander et al. (1988). The steel rebars employs a bilinear stress-strain relationship with modified strength and ductility. This way, the model is able to explicitly incorporate the effect of corrosion on the strength and deformation capacity of both concrete and steel reinforcement, neglecting other phenomena that can affect or be magnified in corroded RC member (e.g., bond deterioration, magnified buckling phenomena). The fiber-based model for each bent column is constructed via the OpenSees® software adopting a forceBeamColumn element to simulate the flexural behavior of the nonlinear beam-column element. In this element the plasticity is spread along the entire element via 7 integration points. At the cross-sectional level, the area of the concrete part is generated by means of rectangular patches. The fibers of the concrete cover are simulated assigning to the patch uniaxialMaterial Concrete02, while for both the ineffective and the effective confined concrete uniaxialMaterial Concrete04 is assigned to the fibers. The fibers of reinforcement bars are generated by using a layer command to generate a number of fibers along a line. To these fibers the stress-strain relationship is obtained via uniaxialMaterial Steel02. To account for the reduction in ductility of corroded materials a MinMax material has been assigned to both corroded concrete and steel, to limit the maximum strain according to the formulations. Each cross-section has been discretized into several fibers to ensure a realistic representation of the section behavior. For the effective confined core, a total of 360 fibers, for the ineffective confined core and the cover of 640 fibers. The mesh has been calibrated in order to not have significant changing in final results when changing iteratively the number of fibers, however, further refinements are required to allow a better trade-off between accuracy and computational efficiency. 5. First results The multi-column bent has been analyzed via Nonlinear Static analysis procedure. The top displacement and the base shear have been recorded to obtain the corresponding pushover curve. Fig. 3 depicts the Pushover curve obtained for the original uncorroded condition for the multi-column bents. The black curve represents the condition of pure flexural behavior in which the shear springs of the capbeams and shear and axial springs of the columns have been substituted by rigid elements to avoid the development of brittle failures. In this case the maximum base shear is equal to 3.21*10 6 N with a peak corresponding to 0.128m. After the peak, the lateral bent capacity softly decreases up to a displacement of about 0.145m after which is possible to evidence small sudden drops of capacity related to the progressive crushing of concrete. Considering as a collapse criteria the point of the curve corresponding to a capacity drop of 80% with respect to the maximum base shear, the ultimate bent capacity corresponds to a lateral displacement of 0.205m, and an interstory drift ratio of about 3.4%. When the possible triggering of brittle failure mechanism is activated, a shear failure after flexural yielding is expected since for each column of the bent the column shear for pure flexural behavior (V p ) is lower than the column shear capacity evaluated according to Sezen (V n ) (Sezen & Moehle, 2004), thus shear-critical behavior is not expected. However, since the shear capacity of RC members decreases with the plastic ductility demand to 0.7V n , and V p <0.7V n , a flexure-shear critical failure is expected. The modeling approach proposed in the previous section allows to account for this type of failure, and the corresponding Pushover curve is depicted in Fig. 3 with a continuous red line. In this case, is possible to evidence a more deformable behavior, related to the shear deformation component. In particular, before reaching the peak lateral strength, the activation of brittle failure of capbeams increase the deformability of the system. Then a peak value in base shear of 3.19*10^6N, corresponding to a top displacement of 0.119m, is reached. After this peak, the triggering of first brittle failures makes the lateral capacity drop suddenly. After this drop, the decrease in shear firstly follows the original trend of the pure flexure behavior, but starting with a lower value, then follows a steeper pattern dominated by the triggering of the shear failure of the remaining columns. Adopting the same collapse criterion as before, the 80% of peak strength is reached for a displacement of 0.162m. Thus, in case the brittle failure of bents is explicitly simulated, a reduction of about 21% in ductility capacity with respect to the case in which only flexural behavior is simulated is expected. In case the effect of corrosion are accounted for, a further decrease in lateral capacity in terms of strength and ductility is expected. In particular, according to the formulations in §3, two possible level of mass loss, namely 10%