PSI - Issue 38

W. Radlof et al. / Procedia Structural Integrity 38 (2022) 50–59 W. Radlof et al. / Structural Integrity Procedia 00 (2021) 000 – 000

53

4

2.3. Micrographs Micrograph preparations were made on specimens from fatigue tests, which have been stopped prior to complete failure. Therefore, the measured potential difference was used as a termination criterion. After termination of the experiment, the lattice structure was embedded with EpoxiCure and afterwards grinded as well as polished using diamond suspensions until cracks were visible on the struts. The micrographs were analyzed using the Hirox KH 7700 digital microscope. Fig. 2: Experimental set-ups of (a) bending and (b) torsion tests. Thermography camera and cameras for digital image correlation (DIC) are included into the test rigs. The magnification of the samples additionally shows the attachments for the potential drop method.

3. Results and discussion 3.1. Fatigue behavior

To characterize the fatigue behavior of the porous structures, the stiffness degradation of the sample was determined. Therefore, the relative stiffness of each sample was calculated as the relation of ∆ ⁄∆ ( : bending moment; : displacement) for bending fatigue tests and ∆ ⁄∆ ( : torsion moment; : torsion angle) for torsion fatigue tests. Subsequently, the stiffness is normalized to the stiffness at approx. 1000 cycles, when the initial steady state is reached. The progression of the normalized stiffness over the fatigue life is exemplarily shown in Fig. 3 for one representative of each porous-design. The progression of the measured potential difference is additionally plotted in the diagrams. It becomes obvious that the stiffness of the bending samples was nearly constant over a long lifetime. From approx. 80% of the fatigue life, the stiffness decreases rapidly until complete failure of the structure occurs. Whereby, the rapid stiffness decrease takes place in stages (see Fig. 3a). For torsion samples, the stiffness decreases uniformly up to approx. 80-90% of the fatigue life. With further increase in fatigue life, the stiffness drops rapidly, but in stages, too (see Fig. 3b). While the loss of structural stiffness is a well-known method to represent the damage of the sample, the potential drop method seems to be an equally good method. Due to the fact that the progression of the potential difference shows a similar behavior with an opposite curve slope as the stiffness curves. Especially the stage-like behavior prior to the whole structure failure is similar. Moreover, the measured potential seems to be somewhat more sensitive than the stiffness behavior, which will be additionally discussed in the next sections. Furthermore, no significant design-dependence is evident in the stiffness loss for both, bending and torsion loaded samples. However, a design-dependence is visible in the potential progressions. Where, the 70% porosity