PSI - Issue 53

S. Senol et al. / Procedia Structural Integrity 53 (2024) 12–28 Author name / Structural Integrity Procedia 00 (2019) 000–000

22

11

The fractographs in Fig. 7 reveal the crack initiation and propagation zones followed by the final fracture zone for all conditions. Multiple crack initiation points are recorded in the cases of AB (Fig. 7(a)) and EDM (Fig. 7(c)) conditions. Zoomed in crack initiation points show surface initiation for AB (Fig. 7(e)), R (Fig. 7(f)), and EDM (Fig. 7(g)) surfaces, while a sub-surface defect is noted for M condition (Fig. 7(h)). While the transition from propagation zone (Fig. 7(j)) to final fracture zone (Fig. 7(k)) can be seen, the dimples (Fig. 7(l)) indicate ductile failure. 4. Discussion There are several fatigue-influencing factors that affect the crack initiation, propagation and/or final fracture steps, hence the overall fatigue performance of a part. This study is focused on how different surface treatments, namely, dL-PBF, EDM and milling, influence (1) the fatigue-influencing factors (i.e. surface roughness, concomitant stress concentration factor, surface residual stress, hardness) and (2) the resultant fatigue performance, focusing mainly on the crack initiation. Firstly, the highest surface roughness, as well as the highest critical stress concentration factor ( k t ) are recorded in the case of the AB condition (Table 1). This can attributed to the intrinsic properties of the L-PBF process, characterized by the so-called stair-case effect on inclined surfaces, melt-pool break-up due to balling effects, and the adhesion of partially-melted powder particles on the part surfaces (Nasab et al., 2018; Strano et al., 2013). Especially for inclined surfaces, such as the curved region of the 3PBF samples in this study, the stair-case effect is considered to be the predominant contributor to the increased surface roughness (Strano et al., 2013). On the other hand, a remarkably lower surface roughness and, consequently, k t are achieved for the R condition, with reductions of 73% in Ra , 67% in Rv , and 63% in k t . This improvement is made possible thanks to the application of the dL-PBF process by (1) removal of the excess powder in the curved region of the 3PBF samples utilizing shock waves induced by a PW laser, and (2) smoothening the powder-free, ablated surface by re-melting (laser-polishing) with a CW laser as discussed in (Metelkova, Vanmunster, et al., 2021). During EDM and milling, the surface asperities are removed thermally (Cheng et al., 1994; Ho & Newman, 2003) and mechanically (Masoudi et al., 2015), respectively. The surface roughness improvement and stress concentration factor reduction after EDM and milling (M), in comparison to the AB condition, are 72% and 85% in Ra , 68% and 78% in Rv , 69% and 71% in k t respectively, pointing out that the lowest surface roughness is achieved by milling. More interestingly, the surface roughness and k t values that are achieved for samples with the R, EDM, and M surface conditions are not significantly different, indicating similar and lower amount of critical notches at these surfaces as compared to the AB surface. Therefore, it can be argued that dL-PBF is an effective alternative to produce inclined surfaces with improved surface quality on parts made of high strength, ceramic particle reinforced aluminium-based metal matrix composites. Regarding the fatigue performance, high surface roughness is known to be one of the most detrimental factors as it results in several stress concentration points at the surface, facilitating early crack initiation (du Plessis & Beretta, 2020). This study also shows that surface roughness plays a significant role, as samples with the AB surface condition (i.e. with highest surface roughness and k t ) exhibit the worst fatigue performance (Fig. 6), with several crack initiation points at the surface (Fig. 7(a)). Whereas fatigue life is significantly improved with the decreased surface roughness for EDM, R, and M conditions, as compared to AB. More specifically, as other fatigue-influencing factors, discussed in the following paragraphs, are similar for AB and R conditions, it can be concluded that the lower surface roughness and the concomitant stress concentration factor are the predominant parameters governing the improved fatigue performance for R samples as compared to AB samples. Secondly, it is observed that the surface residual stress states vary for different surface conditions. While AB and R conditions have comparable and tensile residual stresses, EDM and M conditions display significantly different and compressive stresses on the surface. This divergence can be attributed to several mechanisms. On one hand, the high temperature gradients, the rapid solidification, and the cyclic heating and cooling involved in laser melting induce macroscopic tensile stresses measured in the cases of AB and R conditions (Mercelis & Kruth, 2006). Moreover, it is demonstrated in this study that the tensile stresses in the AB condition are partially relaxed during the first step of dL PBF, due to material removal by PW laser ablation and laser induced shock waves (LISW), which propagate into the material, inducing plastic deformation and compressive stresses (Ding & Ye, 2006; Fabbro et al., 1998; Peyre et al., 1996; Sun et al., 2018). However, subsequent to the re-melting (second) step of dL-PBF, the tensile stresses are re introduced, thus the residual stress states of AB and R become comparable. On the other hand, in the case of samples