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

54

5

design additionally shows a load level dependence in individual cases, which can be seen in Fig. 3a and b from the yellow dashed curves. Further investigations will be carried out to determine the exact reasons for this.

Fig. 3: Stiffness degradation and potential increase versus fatigue lifetime exemplarily shown for one porosity representative under (a) bending and (b) torsion loading.

3.2. Digital image correlation technique The strain distribution on the lattice structures of both, bending and torsion samples, was determined from the images using the DIC technique. Therefore, the software VIC-2D (version 6.0.6) was utilized with an exhaustive search by the normalized squared-differences criterion for the correlation. The optimized eight-tap method was used to interpolate between the discrete gray values of the pixels. The subset sizes varied between 29 × 29 and 33 × 33 pixels with a step size of seven to ensure four ranges over a strut. Moreover, the Hencky (logarithmic) strain was computed with a filter size of five, which enables comparability with subsequently planned FE simulations. From the DIC-analyzed images of the lattice structures, local strains on individual struts were determined. The progressions of these strains were plotted together with the measured potential difference as well as stiffness loss of the examined lattice structure versus the fatigue life. The results are shown in Fig. 4 and Fig. 6 exemplarily for a 50% porosity bending and torsion sample, respectively. 3.2.1. Bending loading From the recorded as well as DIC-analyzed images in Fig. 4, it becomes obvious that struts at the tensile loaded side (bottom – mirror, Fig. 4c) of the specimen fail first as expected. Additionally, the major increase of local strains is determined at a fatigue life of 45% as shown in Fig. 4b. At this lifetime no change in the stiffness curve is visible, whereas an increase in the potential difference is recognizable. Furthermore, at each change of the potential difference an increase of local strains of different struts are detectable (Fig. 4b, c). At a fatigue life of approx. 90% and higher, strut failure on the front side of the lattice structure is visible. At this point the stiffness curve is already decreased and a rapid increase in potential difference is also obvious (Fig. 4a, d). These results indicate a sensitivity of the measured potential to the damage behavior of the tested sample, which is also independent of the porosity design. To examine also the inner struts of the lattice structures, a bending test was stopped at a potential difference of 0.1 mV. The micrograph of the tensile loaded side of the structure is shown in Fig. 5. Multiple cracks and strut failure were visible on inner struts fabricated not only in build direction (BD), but also perpendicular to the build direction, as shown in Fig. 5 b and c. With DIC only strut failures were visible on struts fabricated in build direction at 0.1 mV potential difference. Moreover, the microstructural analysis indicates that cracks initiate from the surface roughness.