PSI - Issue 39

522 8 Lucia Morales-Rivas et al. / Procedia Structural Integrity 39 (2022) 515–527 Author name / Structural Integrity Procedia 00 (2019) 000–000 (2) assuming the ratio between the defect length and depth equal the desired selected values ( = 3 8 0 0 = 0.375), and the actual component geometry (red line in Fig. 4). Additionally, points A and B in Fig. 4 indicate the Anderson shape factor values under the selected FIB defect depth, =30 μm. Eq. 6 was now re-written in such a way to be comparable to Anderson´s equation (Eq. 8), using as the area of a semi-circle with radius a. It results in a shape factor denoted as α adapted (Fig . 4): α adapted = 0.65 �� 2 = 0.728

Fig. 4. Shape factor versus crack depth generated based on Anderson´s solution, and Murakami´s solution adapted to Atzori´s equation for long cracks ( α adapted ). As a final stage in the FIB defect design procedure, three curves (see Fig. 5) analogous to that proposed by Atzori, et al.(2003, Atzori, et al.(2005) were plotted based on Eq. 3, where the black curve was generated using the value α adapted as the shape factor, and the red and orange curves were plotted using the Anderson´s solution for = 1 and = 3 8 0 0 = 0.375, respectively. It can be observed that for crack length up to 1 µm, all graphs perfectly converge. However, with increasing crack length, a deviation appears in the trend of the curve corresponding to = 0.375 from the other two. The orange graph ( = 3 8 0 0 = 0.375) was employed for the estimation of Δσ th of the specimens weakened by the FIB defect. The solid black spot in the diagram (Fig. 5) specifies the normalized stress range selected for the interrupted fatigue test of Specimen 3, implying a stress amplitude S a =250 MPa.