PSI - Issue 57

David Mellé et al. / Procedia Structural Integrity 57 (2024) 61–72 David Melle´ / Structural Integrity Procedia 00 (2023) 000–000

64

4

Fig. 1. Fatigue bending tests flat coupon geometry.

size ( √ area Murakami’s parameter Murakami and Endo (1986)), depth and its stress concentration. To calculate the stress concentration for each defect position, the stress concentration distribution was calculated for the whole surface using a FFT based methodology proposed by Kim et al. (Kim (2020)).

3. Results

As already mentionned, two surface states (net-shape and 30 min etched) were tested using a locati methodology. Several coupons were tested in each condition but only coupons for which the cracks initiated on surface are consid ered to be valid in this study. The others are internal initiations or invalid results (for example crack initiation from the corner of the gauge length). On both surface state conditions, 16 specimens with crack initiated on surface were considered. The fatigue failure surfaces were also observed using a Zeiss MA25 EVO scanning electron microscope (SEM). The crack initiation mechanisms were identified from these SEM images. Three initiation mechanisms were observed, all linked to surface micro-geometric features. The first one is the layer-by-layer manufacturing induced surface val leys, corresponding to the intersection between the layers at the surface. This initiation mechanism is illustrated in Figure 2. The two other initiation mechanisms are due to surface connected porosities: gas pores and lack-of-fusion. An illustration of these mechanisms is given in Figure 3 and Figure 4. On net-shape surfaces, the surface valley mechanism dominates with 12 crack initiations out of 16 coupons. For the other 4 coupons, cracks initiated at lack-of fusion pores. On chemically etched coupons, after 30 min of treatment, the criticity of surface valleys has significantly decreased with the polishing e ff ect of the chemical etching (with only 2 valleys induced crack initiations out of 16 specimens). The other coupons failed in similar proportions of lack-of-fusion and gas pores with respectively 6 and 8 initiations for these mechanisms. The surface porosities or valleys responsible for the fatigue crack initiations were measured using the ImageJ software according to the measurement methodologies proposed by Murakami et al. (Murakami and Endo (1994); Murakami (2002)). The sizes relative to each initiation mechanism are summarized in Table 2 and Table 3.

Table 2. Sizes of critical surface features for each initiation mechanism on net-shape surfaces.

Min size µ m

Mean size µ m

Max size µ m

Surface valleys Lack-of-fusion

306 311

437 399

654 489

On net-shape surfaces, the critical lack-of-fusion pores and the critical surface valleys are of similar sizes (there is 1 large valley measured to be 654 µ m, the next smallest one was 503 µ m which makes the minimal, mean and maximal