PSI - Issue 83

Ahmad Issmail et al. / Procedia Structural Integrity 83 (2026) 229–238

231

2. Materials and Methods The specimens of the geometries shown in Fig. 1 were manufactured via PBF-LB using the EOS M290-2 machine, from EOS Aluminum AlSi10Mg|30µm. The adopted process parameters were those recommended by the manufacturer; specifically, a layer thickness of 30 µm, a laser spot size of 85 µm, and for the infill hatching a laser power of 370 W, a laser speed of 1300 mm/s, and a hatch distance of 0.13 mm. Five different combinations of specimen size (standard (S) or miniature (m)) and build orientations (vertical (V), 45°, or horizontal (H)) were adopted, resulting in the five specimen categories: S|V, S|45°, S|H, m|45°, and m|H. The specimens were kept in their as-built surface condition and underwent the EOS T6 heat treatment specified by the manufacturer [17]. The heat treatment consisted of a solution annealing in a preheated oven at 530°C for 30min followed by water quenching, before artificial aging at 165°C for 6h, followed by cooling in open air. This heat treatment is referred to as T6 HT within this manuscript. The profile surface roughness parameters of the downface (for H and 45°) of the gauge portion of the specimens were calculated using Alicona Infinite Focus Microscope (IFM) using a ×5 magnification and a lateral and vertical resolutions of 2.2 and 1.5 µm. Tensile tests were conducted on one specimen of each build orientation of the standard specimens (S|V, S|45°, and S|H) using the MTS landmark servo-hydraulic test machine (Minnesota, USA) with a load cell capacity of 50 kN and a displacement rate of 5 mm/min. Uniaxial fatigue tests were performed in tension-tension with a load ratio R=0.1 and a frequency of 20 Hz on S|V, S|45°, m|45°, and m|H. The same machine used for the tensile tests was used for the fatigue tests of the standard specimens, while for the miniature specimens, the apparatus was the Instron ElectroPuls E10000 all-electric test system (High Wycombe, UK) with a maximum load cell capacity of 10 kN. After failure, the fracture surfaces were analyzed using SEM, and the effective stress was corrected with the average effective load-bearing cross-section calculated using image processing tools (Inkscape). Runouts have been fixed at 2×10 6 cycles, and the S-N data have been statistically elaborated with ASTM E739 as well as with the Maximum Likelihood Estimation (MLE) method for a more representative estimation of the fatigue limit considered as the stress withstood for 2×10 6 cycles.

Fig. 1: Standard and Miniature specimen geometries

3. Results and Discussions 3.1. Surface Roughness

The surface roughness profiles are shown in Fig. 2, and the corresponding surface profile parameters are reported in Table 1. Considering, for instance, the 45° orientation, m|45° shows significantly reduced surface profile irregularity compared to S|45° (Fig. 2(a)) and lower surface profile parameters (Table 1). This observed reduction in surface roughness for miniature specimens is less pronounced for m|H due to support remnants and partially fused powder particles on the downface of the gage area. As explained by Peng et al. [10], the overall decrease in roughness with size can be attributed to the limited volume of the underlying solidified material in thinner specimens that compromise the heat conduction to the substrate, which results in higher heat accumulation and shifts the dominant heat conduction towards the surface, promoting more