PSI - Issue 19

Masahiro Takanashi et al. / Procedia Structural Integrity 19 (2019) 275–283 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

279

5

replica and the depth was estimated from the aspect ratio of the smallest beach mark observed on the fracture surface. Compared with the crack depth, the length has wider scatter due to crack coalescence. The mean values of crack depth, half crack length, and aspect ratio ( a / c ) at N bfc were 3.8 mm, 13.8 mm, and 0.27, respectively.

(a) Crack depth

(b) Half crack depth

(c) Aspect ratio

Fig. 3. Crack size of large-scale specimen at fatigue life determined by best-fit curve of DFC subcommittee.

As an example, a fracture surface and beach marks are presented in Fig. 4. As can be seen from this figure (presented together with a scale), it is unreasonable and uneconomical to define the fatigue life of a large structure with a small crack as shown in Fig. 3. A large-scale structure is likely to be able to be designed or maintained in consideration of the crack growth life. In the next chapter, crack growth analysis is performed based on the beach marks read on the fracture surface. In the mean stress effect test, however, the upper strain at the notch root greatly exceeded the yield strain, and the small-scale yielding condition may not be satisfied. Therefore, the analysis was conducted on three tests, CS1, LAS1, and LAS2 to which the mean stress was not applied.

Fig. 4. Example of fracture surface and beach marks (CS1 specimen).

4. Crack growth analysis

4.1. Postulation of initial crack size and specimen dimension

The crack was postulated to be a semi-elliptic surface one. The crack growth analysis was started from the crack length and depth determined from the beach mark first observed on the fracture surface. Although a shallow dull notch with a depth of 5 mm was installed in a 90 mm thick specimen, the thickness was postulated to be 90 mm. The plate width was 150 mm.