PSI - Issue 54

6

T. Fekete, D. Antók, L. Tatár, P. Bereczki Structural Integrity Procedia 00 (2019) 000–000

T. Fekete et al. / Procedia Structural Integrity 54 (2024) 314–321

319

3.3. Digital Twins for the tensile tests

Hexagon Marc/Mentat 2022.3 FE (2022) system was used to perform simulations of the tensile tests. Large deformation theory was applied. The 3D models of the specimens were constructed of 20 -node hexahedron elements with 0.5 mm edge length along the gauge length, which matches the typical length scale of the original marking grid applied to the test specimens. The models consist of around 120 000 nodes. The Young's modulus was set to be 210 GPa and the Poisson's ratio was 0.3 . The von Mises plasticity theory was used as a basis for monitoring plastic state. The flow curve of material S460 was evaluated by the help of Choung Cho method (2008) from the data measured during the tensile tests: force and strain data from the Gleeble 3800 thermomechanical simulator and contours of the specimens determined by a photo-processing algorithm. Since the simulations were designed to demonstrate the relevance of incorporating geometric imperfections into the DT , other imperfections, such as material inhomogeneity and anisotropy were excluded from the study. Therefore, the material was considered homogeneous and isotropic. In this paper 3 FE models are introduced. A model called FN is a geometrically perfect model (i.e., the model implementing the idealized design shape), with nominal dimensions. F1 and F2 models are the DT s of the specimens described in Section 3.2. The evolution of specimen shapes and the strain fields is illustrated in Figure 6. As can be clearly seen, the FN model produced a homogeneous deformation field along the gauge length in the early stages of the simulation; the strain field remained symmetric for larger displacements and necking emerged in the mid-section of the gauge length. In contrast, the strain field of the F1 and F2 models was already inhomogeneous from the very beginning, and the necking location was strongly shifted towards one of the ends of the gauge section. Simulation results show that geometric imperfections can significantly affect the time evolution of the specimen strain and stress field, as well as its shape, including the location of necking and failure.

Fig. 6. Strain evolution of the geometrically perfect FN model and the geometrically imperfect F1 and F2 models

Figures 7 and 8 show how the time evolution of the F1 and F2 shapes observed during the measurements –with the optical data acquisition system– fits to the corresponding results of the DT simulations. They present some typical results, observations at the top, simulations at the bottom. The plots labelled ' a ' show initial state of the specimens, ' b ' the state near the end of uniform elongation, ' c ' a moment during diffusive necking, and ' d ' the specimen's final state.