PSI - Issue 19

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

281

7

In LAS2 tested with smaller strain amplitude, the beach marks were introduced only three times. Despite its macroscopic coalescence, the crack growths in both surface and depth directions agreed with the analysis result by JSME S NA1. 4.5. Fatigue life prediction for actual components As presented above, approximately 3 – 4-mm-deep crack was found at N bfc , which was derived from the fatigue test results of the small-sized specimens. If a designer tries to design a component or structure that is much larger than the small-size specimen, the designer may consider prolonging its design life expecting the crack growth life. The failure life of the large-scale specimen, therefore, is discussed in this section. A fatigue life can be expressed as the total life of the crack initiation and growth lives. Although the definition of fatigue crack initiation is not clear technically, as far as the DFC best-fit curve is concerned, a series of large-scale tests revealed that the best-fit curve provided approximately 3 – 4-mm-deep crack initiation life. Nakane et al. (2019) also performed the fatigue tests using the small-sized specimens of carbon, low-alloy, and austenitic stainless steels and reported that approximately 3 – 4-mm-deep cracks were observed on the fracture surface of small-sized specimens. To predict the failure life, the crack growth life should be calculated from approximately 3 – 4-mm-deep crack. Regarding the aspect ratio, it was assumed to be approximately 0.3 from the experimental results. The aspect ratio of the fatigue crack strongly depends on the thickness of the components and the ratio of membrane and bending stresses. In a large-scale actual component where the thickness is large and the stress gradient is gentle, the geometric and mechanical conditions are considered to be almost similar to those of the large-scale fatigue tests conducted in this study. The crack growth analysis was conducted from the 3-mm-deep initial crack in the same manner as described in the previous section until the crack depth exceeded half the plate thickness. The predicted failure life was obtained as the sum of N bfc and the crack growth life. Figure 8 shows the relation between the experimental and the predicted failure lives. The solid line represents the experimental failure life that is equal to the predicted one; the dashed lines indicate variations by a factor of 2. Regardless of the crack growth laws, all data were plotted just below the solid line. Even under the assumption of 3 mm-deep initial crack, which gives longer crack propagation life than that of 4-mm-deep crack, the experimental failure life was predicted accurately and safely. The fatigue life of the actual component can be predicted in a similar manner. Reversed four-point bending fatigue tests were conducted on large-scale specimens of carbon steel and low-alloy steel plates. The test results were compared with the best-fit curve for carbon and low-alloy steels proposed by the DFC subcommittee in Japan. The fatigue lives by the best-fit curve corresponded to approximately 3 – 4-mm-deep crack initiation lives in large-scale specimens. During the fatigue test, beach marks were introduced to each specimen to trace the crack growth behavior. Then fatigue crack growth analysis was performed on three large-scale specimens to which no mean strain was applied. It was confirmed that although the amounts of crack growth were scattered, the crack growth behavior in the large-scale specimen could be simulated by the crack growth laws by JSME S NA1 (2016) and Kobayashi et al. (1989). In designing a large-scale machine or structure based on the fatigue data of small-sized specimens, the predicted crack depth is of the order of several millimeters at the fatigue life given by the best-fit curve. If this crack depth is relatively small compared to the component size designed, a designer can prolong the design fatigue life taking the crack growth life into consideration. When analyzing the crack growth, the postulation of initial crack size is important. This study suggested that after reaching the fatigue life by the best-fit curve, it would be sufficient to consider the crack growth from the 3-mm-deep crack with the aspect ratio of 0.3. 5. Concluding remarks