PSI - Issue 75

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

24 6

3. Results 3.1. Build quality

Surface quality measured by optical profilometer was similar in all four cases (Table 1) The measured R a values varied between 6-8 µm and R z values from 40 to 50 µm. However, surface characteristics obtained by the SEM analysis varied significantly. Surface characteristics are illustrated in Fig. 6. The G-series was peened with high coverage from both inside and outside the web members. In other test series, the inside coverage was low or non-existent. The EOS G and SLM B series were treated in a similar way, it is possible that EOS surface quality was better before treatment which leads the quality difference after processing. The notch area supported the best quality of the EOS G series.

SLM W-series

EOS G -series

SLM B -series

EOS S -series

25 μ m

Outside

2 mm

-25 μ m

Inside

Outside Inside

Notch

Outside

Notch surface

Fig. 6. Surface quality and characterises.

The residual stresses were measured from middle of outer surface at front side. Two specimens were selected for the measurement. The glass-peened EOS build component with the highest fatigue performance amongst tested specimens was found to have high compressive residual stresses (avg. -675 MPa) at the surface whereas shot peening with ceramic media resulted significantly lower compressive residual stresses (avg. -268 MPa). Also, the full width at half maximum (FWHM) value indicated higher impact by the surface peening.

Table 2. Residual stress measurements.

Specimen

Normal stress 0° ± FWHM 0.0° Normal stress 90° ± FWHM 90.0°

Glass peened (EOS) -712 Shot peened (SLM) -332

32 2.82 31 2.35

-637 -203

16 2.83 26 2.31

Polished section was curved top side section was prepared to study density and imperfections of the AM material. The microstructural features of all four samples are presented in Fig.7; these features are captured via back-scatter electron (BSE) channeling carried out on the polished surface of the specimens using scanning electron microscopy (SEM). According to Fig. 7, all specimens had mainly similar microstructures comprising a Widmanstätten morphology of α phase (hexagonal) laths separated by comparatively narrower β phase (body -centered cubic) features. Considering the lath thickness of the α phase being in the range of multiple micrometers and L -PBF as the AM process utilized in this study, the specimens, especially SLM B and EOS G due to their relatively coarser α blades, seem to have been subjected to annealing heat treatment after the AM procedure. Hence, no dominant presence of the martensitic α’ phase is expected in these specimens. However, EOS S and SLM W might have experienced different