PSI - Issue 14

Takashi Nakamura et al. / Procedia Structural Integrity 14 (2019) 978–985 Author name / Structural Integrity Procedia 00 (2018) 000–000

980

3

The crack growth tests were interrupted periodically, and the surface crack behavior was monitored by using a digital microscope (Keyence: VHX-2000). The microscope enables a high spatial resolution of submicron order. During the observation, a tensile load equivalent to 80% of the maximum loading was applied to open the crack. As shown in Fig. 4, the crack length 2 a was measured as the distance between each of the two crack tips projected to the plane perpendicular to the loading direction. Following the fatigue tests, fracture surface observations were carried out by using an SEM (Keyence: VE-9800) to investigate the details of the crack propagation behavior in each environment especially focusing on the granular region. The more detailed information of the experimental procedures are available from the literature (Yoshinaka, et. al, 2016).

Table 1. Chemical composition.

Al H Ti 6.12 4.27 0.16 0.002 0.02 0.15 0.0029 Bal. V O N C Fe

Table 2. Mechanical properties.

0.2% proof stress [MPa]

Tensile strength [MPa]

Reduction of area (%)

Elongation (%)

Hardness (Hv)

943

860

17

40

316

Fig. 2. ( α + β ) dual-phase microstructure.

Fig. 4. An example of the fatigue cracks initiated from the artificial small defect. The crack length 2 a was defined as the distance between each of the two crack tips projected to the plane perpendicular to the loading direction.

Fig. 3. Shape and dimensions of fatigue specimen after polishing.

3. Results and discussion

3.1. Crack growth behaviors

The crack initiation was detected at less than 1000 cycles regardless of the test environment. As mentioned in Chapter 2, the observation results over the crack length 2 a of 90 µm were discussed to ignore the effect of stress concentration due to the small defect. Fig. 5 shows the crack length from 2 a = 90 µm with respect to the number of cycles N . The number of cycles for crack length to reach 90 µm were 2.2 × 10 4 and 4.2 × 10 4 in air, and 6.5 × 10 5 and 1.4 × 10 6 in vacuum. Fatigue life spent in the early stage of crack propagation in the regime below 2 a = 90 µm was much larger in vacuum than in air. In addition, the shape of 2 a - N curve in vacuum had a plateau around 1–3 × 10 6 cycles, and showed a significant slow propagation from 2 a = 90 µm to 200 µm. The number of cycles to failure were 4.6 × 10 4 and 7.2 × 10 4 in air, and 5.5 × 10 6 and 6.9 × 10 6 in vacuum. The fatigue life in vacuum was significantly improved, and was two orders of magnitude larger than that in air. The effect of vacuum on crack growth was much more profound in the small crack regime below around 2 a = 200 µm in the present study. By using the data in Fig. 5, the relationships between crack propagation rate d a /d N and the stress intensity factor range ∆ K were investigated. d a /d N was defined as the average crack propagation rate Δ a / Δ N , which was calculated by dividing the crack propagation length Δ a by the increment of number of cycles Δ N between two adjacent measuring points. Δ K was calculated by using Eq. 1 (Nishitani and Chen, 1984). ∆ = I ∆ √ (1)