PSI - Issue 62

Giulia Rossini et al. / Procedia Structural Integrity 62 (2024) 347–354

352

6

Giulia Rossini et al. / Structural Integrity Procedia 00 (2019) 000 – 000

4. Numerical analysis The release test performed using the Saw-Cut -Web method, which experimentally proved to be the most reliable, was simulated with a finite element (FE) model using Midas Gen 2022 software (version 9.2.5). A 3D model and two simplified 2D plane stress models of the beam were generated as elastic models. To thoroughly investigate stress diffusion after the cuts on the web a 3D model of an I-beam with flanges 50 cm wide, 20 cm thick and web 18 cm wide (y-direction), 80 cm high (z-direction) and 5 m long (x-direction) was developed, which closely resembles the real beam tested in the laboratory (Fig. 3- left). Although the real laboratory beam is 10 m long, the considered 5 m model was considered sufficient to dissipate the disturbance zones at the ends. 8-node brick elements were used in this model. The mesh is not uniform: starting from 40x40x40 mm elements, the mesh gets finer approaching the area where saw-cuts were modeled. In particular, in the region close to the saw cuts, brick elements of size 0.5 cm, 1 cm, 1 cm (x,y,z) were used to represent the cuts made by the angle grinder, whose blade generates a cut in the concrete of width 0.4 cm and length 26 cm. Saw cuts were simulated by removing 0.5 cm wide elements for a length of 26 cm. All four different geometric configurations (GCs) of the cuts applied in the experimental tests were considered. As experimentally observed, the models of the two configurations with pairs of cuts 10 cm apart (GC_10_2 and GC_10_3) do not allow complete tension release. The tests indicate that the region between the two cuts remains in compression in both the xz and xy planes following the execution of the cuts. Conversely, the models of the two configurations with pairs of cuts 6 cm apart (GC_6_2 and GC_6_3) exhibit tensile stresses residing in the area between the two cuts. Specifically, in the model with cuts 6 cm apart and 2 cm deep, a tensile stress of 0.5 MPa is found, whereas in the model with cuts 6 cm apart and 3 cm deep, a tensile stress of 1.5 MPa is calculated (Fig. 3- right). This longitudinal tensile stress results field from the formation of an arc-shaped stress distribution, a phenomenon not appreciated in experimental tests.

Fig. 3 (left) 3D model with Midas Gen 2022 software. (Right) Zoom of the stresses in the 3D model in the xz plane for the SCw_GC_6_3 configuration after applying the saw cuts. Bagge et al. (2017) and Kraľovanec (2021) , who simulated saw cuts 12 cm apart and 5 cm and 3 cm deep, respectively, in a 2D FE model, also obtained tensile stresses in the region between saw cuts. Also two 2D models representing the xz plane and the xy plane were considered in this study. By comparing the 3D model with the simplified 2D models, it is evident that for the pairs of cuts made in the web, the 2D models lack reliability. They fail to accurately capture the correct stress state in the region between the two cuts. Specifically, a stress of 0 MPa is obtained in the xz-plane model, while a compressive stress of -1 MPa is calculated in the xy-plane model. This emphasizes the importance of using a 3D model to correctly capture the stress distribution in the region between the saw cuts.