PSI - Issue 68

J. Köckritz et al. / Procedia Structural Integrity 68 (2025) 962–968

965

4

J. Köckritz et al. / Structural Integrity Procedia 00 (2025) 000–000

2.2. Applied material, process, roughness measurement and fractography The investigated components are manufactured of Al2139-AM with a general particle size distribution of 20-65 µm (MDS EOS (2024)) in an EOS M290. The material properties with YS , ultimate tensile strength UTS and the elongation at fracture A* are displayed in Table 1 alongside the cyclic material data of the two-section S-N curve experimentally derived from polished PBF-LB/M samples of the same producer that manufactured the FWA, kindly provided by Härtel (2024). The data includes the fatigue exponents k1 and k2 of the first and second section, the transition stress σ tr , transition life N tr and assumed fatigue limit σ D . Table 1: Quasistatic and cyclic material data of the applied Al2139-AM manufactured by PBF-LB/M, target producer properties after T4 heat treatment (MDS EOS (2024)) and vertical build direction for polished specimen (kindly provided by Härtel (2024)). Data type YS [MPa] UTS [MPa] A* [%] k1 k2 σ tr [MPa] N tr [-] σ D [MPa] Target values producer 460 520 4 - - - - - Polished specimen 503 550 6 5.69 17.51 175.6 3.5·10 5 145 The components were scaled in a relation of 75 % and printed in sets of four components per build job with a build direction that allows minimal supports, see Fig. 2 (b). After printing, supports were manually removed and a T4 heat treatment applied. For fatigue testing, nine components are left in the as-built (AB) surface condition, whereas the remaining nine undergo an industry standard trowalising (TW) surface treatment. There, the components are placed in a vibratory bowl with ceramic (Otec DBS 6/6) and porcelain (Otec ZSP 3/5) abrasives in a slightly acidic media for several hours. The surfaces of all components were characterized by non-destructive measurements using a confocal microscope of type MarSurf CM Explorer with a 20x M1065 objective lens on an area of 2.3x2.3 mm and located in the area of highest tension stresses visible in Fig. 5. This area is located in downskin with an overhang angle of up to 48°. Line roughness values R a and R z as well as the area roughness values S a and S v were evaluated. After damage, all crack surfaces were extracted and assessed with a Keyence VHX-900F light microscope. Detected crack initiating defects were measured for size and position in an 150x enlargement. Additionally, selected fracture surfaces were assessed with a scanning electron microscope (SEM). Energy dispersive X-ray spectroscopy (EDX) was carried out around the initiating defects to detect varying element composition. 2.3. Fatigue experiments Fatigue testing was performed for the centric bump impact with a load ratio of R = 0. A simplified test rig was designed, where the load is applied on a single rather than two mirrored FWA, leading to a minor change in the stress distribution. The FWA is mounted on an angled adapter which represents the chassis. The load is applied by an Hänchen hydraulic cylinder series 320 with a maximum force of 500 kN at a lever with a ball end, to allow the expectable degrees of freedom in the load case. Tests are run force-controlled with an MTS® control system at a frequency of 8 Hz, displacement is measured by the cylinder-integrated distance sensor. Tests are stopped at an absolute boundary of displacement signaling significant crack growth. In data post-processing, the displacement over cycles is evaluated for changes in stiffness which correspond to the observed crack growth. The evaluated lifetimes for the components are derived at the position of this sudden increase in displacement. The S-N curve and scatter bands were derived according to DIN 50100 (2022) for the pearl string method. 3. Results and discussion The surface roughness measurements at the area of critical stresses show a notable effect of the TW on the line roughness values R a and R z . The AB components had a mean of R a,AB = 9.5 µm and R z,AB = 57.2 µm, at the lower boundary usual for AM aluminium alloys (Martucci et al. (2023)), which was reduced by TW to R a,TW = 3 µm and R z,TW = 18 µm. These measurements correspond to the visual and haptic improvement of the surface. Similarly, the mean area roughness S a showed an improvement from S a,AB = 13.1 µm to S a,TW = 6 µm. However, the deepest valleys