PSI - Issue 75

Kalle Lipiäinen et al. / Procedia Structural Integrity 75 (2025) 19–28 Lipiäinen et. al. / Structural Integrity Procedia (2025)

22 4

support free print. A lot of small material removals, pockets and holes, had to be done to achieve lower mass. The final geometry weighted 206 g which was 4 g more than the original part. In general, it would not be sensible to create holes, down skins and small notches in the corners of pockets for saving mass. However, in this work, they were seen as added value. The design process is illustrated in Fig 3a. 2.3. Additive manufacturing configuration AM for the component was carried by L-PBF with five components per a batch. The components were removed by wire cutting and the bolt holes were machined after printing. Otherwise, the surfaces were left non-machined. All specimens were heat treated 120 min at 800°C (in vacuum). Specimen coding is given in Table 1. Glass peening was applied for the EOS G and SLM B test series after building. Separate workshop was used to perform shot blasting with ceramic media Z425 (Almen strip A intensity 0.005 inch) Table 1. Specimen coding and characteristics. Specimen code Printer Layer height Surface after printing Separate workshop Finish Note G EOS 40 Glass peening No Powder EOS Titanium Ti64 Grade 5 S EOS 40 Powder removal Shot peened SLMB SLM 60 Glass peening No Powder SLM Ti64ELI SLMW SLM 60 Powder removal Shot peened 2.4. Fatigue testing Fatigue tests (Fig. 3b) for components were carried out with stress ratio R = -0.66 with 150 kN load-controlled servo-hydraulic test rig. The attachment of AM component was manufactured from t = 6 mm steel plates with stiffeners at backside. The configuration could be considered rigid. Fatigue tests were continued until final rupture.

Fig. 3. (a) Design process and component and (b) fatigue test setup.

2.5. Numerical analysis Once the design was completed, a detailed FE-model was prepared at Patria Aviation with MSC Patran and analyzed with linear analysis (SOL101) and geometrically and materially non-linear analysis (SOL400) by the MSC Nastran solvers. These FE-analysis results were compared to the actual physical test strain gauge measurements. The component FE-model consists of TET10 solid elements and the fasteners and centre bolt are modelled with BAR2 beam elements attached to the structure by RBE2 rigid elements. Simply supported constraints from fastener element