Fig. 2 Schematic illustration of surface cracked specimen.
A rod 200 mm in length was cut out of the supplied wire specimen, and its surface was first polished with ＃02 sandpaper, and thereafter, by discharge machining, a surface notch either 150 μm or 130 μm deep and perpendicular to the wire length was introduced. The notch was U-shaped, 70 μm wide, and a tip curvature radius of 35 μm. Figure 1 gives the specimen dimensions.
2.2 Test seawater
Clean natural seawater collected near Izu-Oshima was used. The level of C1 in the seawater was 2 %, its pH was 8.2, and its specific weight was 1.025.
2.3 Corrosion fatigue test
120 ml of the seawater solution was introduced into a glass cell enveloped in a jacket. A 40 mm length of a test specimen was immersed into this solution. Then, cyclic loading stress was applied to the notched part of the specimen. The cyclic frequency (f) was varied in the range of 15〜0.1 Hz. Maximum nominal tensile stress applied to the notch was 332 MN/m2, cyclic stress amplitude was 226 MN/m2, and stress ratio (R) was 0.32. Seawater temperature in the test cell was held constant at level in the range of 308K〜288K by circulating water in the surrounding jacket, equipped with a thermostat. Seawater was replaced every 6 days. Reference fatigue testing in the atmosphere was conducted with f = 10 Hz, while other mechanical parameters were identical to those in corrosion fatigue test runs conducted in seawater.
The fatigue test machine used was a vertical electrical oil-hydraulic servo machine with a 5-ton capacity, and the applied stress wave pattem was sinusoidal.
2.4 Measurement for crack growth
After a set number of cycles N, the specimen was broken by bending at room temperature and the fracture surface was examined in the optical microscope with a magnification of about 200 to determine the parameters A, B, l and 2c shown in Figure 2. Examined relationships were I vs. N and l/c (aspect ratio) vs. l. From these relationships, the relationship between dl/dN (crack growth rate) and Δ K (stress intensity factor range) was derived where desirable. The following approximation for the semi-elliptical surface crack (eccentricity angle φ ) in a semi-infinite solid was used to derive Δ K [3].
where E (k) is defined by
and σ refers to nominal tensile stress.
2.5 Fracture surface examination
The specimen was broken by bending at room temperature after a special N, and the fracture surface was dried and observed by Scanning Electron Microscope (SEM).
3. Results and Discussion
Figure 3 summarizes the fatigue crack growth curves observed for the SWRM10 steel specimen with notch 130 μm deep and for the SWRH42A steel specimen with notch 150 μ m deep at f = 10 Hz in air. As can be seen, cracks developed steadily with increasing N in both specimens. While not shown in the figure, cracks did not develop in the SWRH42A steel specimen when notched 130 μm deep up to 800 kc.
Fig. 3 Fatigue crack depth vs. number of fatigue cycles in air at 10Hz.
Figure 4 summarizes the crack growth curves for both specimens with 130 μm notches in seawater at 308K and 298K. At 308K, cracks grew steadily with increased N at any f in the range of 15〜0.7 Hz for SWRM10, and 10-0.3 Hz for SWRH42A. For SWRM10, growth rate tended to increase when f decreased from 15 Hz to 10 Hz, but tended to decrease when f was further decreased from 10 Hz to 5 HZ. When f was decreased to as low as 1〜0.7 Hz, growth rate decreased to a level lower than that at 15 Hz. For SWRH42A, however, the growth rate was not depended on f, and there was no difference in growth rate at any f examined.
