PSI - Issue 71

Shohei Matsuda et al. / Procedia Structural Integrity 71 (2025) 4–9

7

changing the direction. Also, multiple parallel cracks propagating closely were frequently observed at the edge of defects. These observational results suggest that the crystallographic orientation of grains would significantly affect crack propagation. The entire fracture surface of the specimens was macroscopically normal to the maximum principal stress plane. However, microscopically, the fracture surface near the artificial defects was never a flat plane but composed of many smaller flat fracture surfaces, hereafter referred to as facets, with different inclinations, as shown in Fig. 4. These facets that incline from the plane perpendicular to the tensile axis are subjected to shear stress as well as normal stress. Also, the facet size was comparable to the average grain size (0.5 mm). Those fracture surface features imply that the facets occur by transgranular slip-plane cracking, activating shear mode, or mixed-mode crack propagation mechanism. Since natural defects, such as porosities, were not observed near the artificial defects that had become the fracture origin, interference between artificial and natural defects is not the cause of the scatter in the fatigue strength. It is supposed that artificial defects became a primal crack initiation site due to stress concentration, and the size and crystallographic orientation of the grains around the defects significantly influenced the conditions for crack initiation and subsequent propagation. The size of the artificial defect is √ = 287 µm, which is smaller than the average grain size of 0.5 mm. Therefore, if grains surrounding the artificial defect have a size larger than 1 mm, for instance, and their crystallographic orientations happen to be favorable for crack initiation and growth, the effective defect size that affects the fatigue strength is larger than the original size √ = 287 µm. Thus, it is thought that the presence of grains comparable to or larger than the artificial defects causes the scatter in the fatigue life and the fatigue limit. Furthermore, the effective defect size would be larger than the original defect size. This misestimation of √ may be another cause for the experimental fatigue limit results lower than the predicted value by equation (1). Further quantitative studies on the role of large grains in determining fatigue strength is necessary to propose a reasonable fatigue strength prediction method from a practical perspective.

Fig. 2. S-N data obtained by rotating bending fatigue tests of specimens containing identical 4-hole defects with √ = 287µ