PSI - Issue 52

8

Satrio Wicaksono et al. / Structural Integrity Procedia 00 (2023) 000 – 000

Satrio Wicaksono et al. / Procedia Structural Integrity 52 (2024) 438–454

445

In order to make the model running as it should be in real experimental test, the stack direction for all components in the T-joint model had to be assigned. For the solid geometries, this stack direction would not give significant effect to the result. However, the assignment of stack directions gave substantial influence to the behaviour of the materials upon impact. Moreover, Mesh Stack Direction gave effect to the deflection direction on the materials when they are given certain loadings. Therefore, for the materials with extreme thinness, deflection direction would significantly affect the deformation performance. Aside from defining element types and mesh stack directions, mesh generation was performed by adjusting the size and shape of the elements, so the model relative error is less than 1%. With that being said, the failure analysis was expected to come near to the actual condition. Afterwards, the convergence test was also taken into action. The test result shows that the analysis result started to converge at element size of 3 mm. In addition, in order to maximize the analysis results, the mesh size at critical region of the model (vertical sandwich structure) was set to be 1 mm, while the other regions had element size of 2 mm. 3. Results and Discussions In this study, the validation process was carried out by comparing the numerical analysis results with the experiment conducted by Caliskan et al. [14], which was a low-speed impact on T-joint with a kinetic energy of 60 J. Subsequently, parametric study by varying the value of several mechanical properties and parameters of PVC 70.75 and AL11050-H14 was performed. As shown in Figure 5, it can be concluded that the numerical model has obtained the same main failure mode as the experimental test result, namely crippling at the end of the vertical sandwich structure. However, in numerical analysis result, the crippling area on the vertical sandwich is greater than the test results. This is due to the assumption on the material properties used in the model. Table 4 shows that at a kinetic energy of 40 J, the test results show that there has not been any crippling in the vertical sandwich structure of T-joint, but in the numerical result, the crippling has occurred slightly. This concludes that the material properties used in the numerical model have lower absorption energy and stiffness than the actual one. Therefore, a deeper study on the properties that should be used in the model is very important to reduce the errors that occur. 3.1. Validation of numerical simulation result to experimental test result

Figure 5 Comparison of failure visualization between experimental model and numerical model.

Table 4 Comparison of failure visualization at kinetic energy of 20 joules, 40 joules, and 60 joules.

Experimental Result

Numerical Result

Kinetic