PSI - Issue 42

Harry O. Psihoyos et al. / Procedia Structural Integrity 42 (2022) 299–306 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

305

7

computational time of the model with uniform cartesian mesh. The computational time of the model of the layered tetrahedralmesh strategy is about 10 times the respective of uniformcartesianmesh, making the mesh scheme highly inefficient for this element size. Regarding thequality of thepresentedmesh strategies, it can be characterised by the element quality index and the average structural error. The element quality is an index in ANSYS tha t characterises the aspect ratio, skewness andminimumedge angle of the elements and it can be varied from0 to 1, with 1 being the best quality for elements. Structural error is a measureof the discontinuityof thestress field fromelement to element. Cartesian mesh schemes (the uniformprimarily) present thebest quality contrary to layered tetrahedralmeshingdue to the genera tion of many skewed elements in the part in the la tter. However, the average structural erro r of layer tetrahedralmodel is significant smaller contrary to cartesian models due to the better geometrical discretization that led to better continuity of stress fields.

Table 1. Characteristics of models of the examined meshing strategies (based on the work of Weber et al.(2020)). Mesh Strategy

Cartesian with Projection Factor

Uniform Cartesian Mesh

Layered Tetrahedral Mesh

Mesh Size [mm] Number of nodes Number of elements Computational Time [s]

0.4

0.4

0.4

0.8

1.2

103,573 66,076 3,227

103,263 65,776 3,203

980,829 618,634 92,097

140,027 72,350 5,114

70,356 29,326 1,825 0.781 2.471

Average Quality

0.994 5.425

0.970 3.479

0.818 0.693

0.784 1.230

Average Structural Error [mJ]

4.2. Assessment of theelement size in layered tetrahedral meshing To further investiga te the efficiency of layered tetrahedral mesh strategy in the simulation of the SLM process a parametric analysis for the element size was performed. Successive simulations for element sizes of 0.4, 0.8 and 1.2 mm were performed. The comparison between the experimental measurements and the predicted results for each model is presented in theFigure 6. The layered tetrahedralmeshingmodels of different element sizes present thesame accuracy. The convergence canbe acquired for a relatively coarse mesh of 1.2 mm indicating that evena modelwith coarse mesh could lead to accurate predictions in low computational times. Although, the mesh quality of layered tetrahedral models became inferior as the element size was increased, the average structural error was less than of cartesian models, denoting the capability of layered tetrahedral coarse mesh for accurate SLM simulation for the prediction of residual stresses. 5. Conclusions In the present work, cartesian and layer tetrahedral mesh stra tegies were compared in the context of the thermomechanical simula tion of the SLM process, with emphasis on the eva luation of layered tetrahedral mesh scheme. Bothuniformandgeometric adaptive cartesianmesh scheme could lead to accurate simulations with the best element characteristics for sma ll element size. Layered tetrahedral mesh scheme models could lead to efficient simula tionwith relative coarsemeshwith a continuous distributionof stresses and strains in thepart domain, the mesh qua lity should be carefully examined though. This denotes the potential of the utilisa tion of layered tetrahedralmesh stra tegy for the part scale simulation of the SLM process. Moreover, this highlights the prospective for subsequent finite element fracture mechanics analysis of SLM parts with layered tetrahedral mesh models due to the inherent capabilities of thesemeshes to model the crackpropagation. References Ba iges, J., Chiumenti,M.,Moreira, C. A., Cervera,M., &Codina, R. (2021). An adaptive Finite Element strategy for the numerical simulationof additive manufacturingprocesses. Additive Manufacturing , 37 , 101650.