PSI - Issue 61

Bekir Kaçmaz et al. / Procedia Structural Integrity 61 (2024) 130–137

135

6

Author name / Structural Integrity Procedia 00 (2024) 000–000

between the points B and K shown in Figure 2 is calculated for every increment of the analysis. Comparison of experiment and FE results for orthogonal CMOD can be seen in Figure 2. It is worthy to note that FE results are based on mesh designated as #3 that uses 40184 elements for the beam. As far as peak force prediction is concerned, the di ff erence between LIGD based FE prediction and average experimental result is around 2.5 %. In the post peak regime, FE analysis result falls in between the two experimental curves almost throughout the whole loading history. At the end of the experiment, a major non-planar crack forms reflecting the complex stress state developing during the loading. Since ‘cracking’ is smeared to a finite volume, it is not possible to get a sharp crack from continuum damage mechanics based models. However, a crack ‘surface’ can be constructed by using the damage variables at integration points.

1400

Current Study (LIGD) Experiment #1 Experiment #2

1200

1000

800

600

400 Reaction Force (N)

200

0

0

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

1

Orthogonal CMOD Values (mm)

Fig. 2. Comparison of analysis and experiment results for orthogonal CMOD for prismatic skew notched concrete beam

To this end, the coordinates of the integration points where the damage variable D is equal or larger than 0.99 at the end of the analysis, are collected. These coordinates and the boundary of the specimen are plotted to create a visual representation of the crack surface. Additionally, the same model is re-analyzed with CIGD model which is simply recovered from LIGD by setting g = 1. In Figure 3, constructed fracture surfaces from CIGD and LIGD results are shown side by side. When compared with the experimental crack surface given in Je ff erson et al. (2004), it can be clearly stated that LIGD model captures the non-planar crack surface successfully. On the contrary, the crack surface prediction of CIGD is thicker and closer to a planar geometry. Failure prediction under complex three-dimensional stress states is a delicate task. In this paper, failure of skew notched concrete beam with non-planar failure surface is modeled by LIGD formulation. Resulting crack surfaces and force-displacement curves are compared with experimental findings reported in the literature. Failure surface predictions obtained by LIGD seems to be much better than the CIGD formulation. Furthermore, predicted peak load for the skew notched beam with square cross-section is very close to the experimental peak value. Although not reported here, the di ff erence reaches to approximately 14 % in case of cylindrical specimen, please see Kac¸maz (2022). Ignoring the complexities associated with the notches, a plausible explanation could be as follows: 4. Conclusion and Outlook