PSI - Issue 53

Benjamin Möller et al. / Procedia Structural Integrity 53 (2024) 190–202 Author name / Structural Integrity Procedia 00 (2023) 000–000

197

8

(a)

(b)

Fig. 9. Local notch shape parameters of the fatigue-relevant notch (a) specimen HB_2D; (b) specimen HB_4U.

The procedure to investigate the size of the element dimension followed the steps: 1. The model of the hollow structure specimen is cut 10 mm below and 10 mm above the notch radius in order to reduce the computing effort and increase the simulation speed using a sub-model. 2. The number of elements is reduced using the Blender-function “mesh clean-up” that reduces the number of triangles needed to define a surface by combining adjacent ones into a bigger one. This Blender function reduces the number of the elements from the zones where the surface is more regular (smaller curvature) in order to harness the potential of the elements. 3. The sub-model of the hollow structure specimen is then imported in Abaqus and meshed. 4. The definition of constraints and loads in agreement to the fatigue testing forms the basis for the load simulation. From the simulation, von Mises peak stresses are used as the criterion for the convergence. The procedure is repeated changing the value of the “reduction factor” for the number of elements of the “mesh clean-up” function (in Blender). This is the value that defines the ratio in between the number of elements after and before the reduction process, e. g. if the parameter is set to 0.5 the number of elements after the reduction process will be half of the original number of elements. The “reduction factors” investigated range from 0.1 to 1.0 in steps of 0.1. The results of this analysis are shown in Fig.10 including a diagram for von Mises peak stress over the reduction factor for the number of elements, representations of two FE sub-models and the resulting element dimensions. The von Mises peak stress shows that the relative difference in between peak stresses is always below or equal 8 % as long as the value chosen is bigger than 0.2. This led to two consequences:  The reduction factor (for the number of elements of the “mesh clean-up” function) chosen for the final simulation of test conditions for the hollow structure specimens was set to 0.25.  The effect of the size (dimensions) of the elements was not considered in the following work.