PSI - Issue 19

Masanori Nakatani et al. / Procedia Structural Integrity 19 (2019) 312–319 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

316

5

the crack-growth threshold of a small crack under a large-scale yielding condition, for which the value of the threshold stress intensity factor range is expressed in the √ area parameter model as: 3 1/ 3 th 3.3 10 ( 120)( ) K HV area −  =  + (2) where HV is the Vickers hardness of the material. The unit of Δ Κ th is MPa √ m and that of √ area is µm. As the value of √ area approached the large-crack regime ( √ area ≈ 300 µm), the small-scale yielding condition was fulfilled and the LEFM became valid. Therefore, in this regime, the fatigue limit was determined by the Δ Κ th value obtained for a large crack, Δ Κ th ,lc . Considering all of the test results obtained from materials with no H-charging, the fatigue limit was classified according to three phases, based on defect size: i) harmless-defect regime, ii) small-crack regime and iii) large-crack regime.

Fig. 3. Relationship between the fatigue limit and defect size in Alloy 718. The fatigue limit was classified according to three stages, depending on individual defect size.

3.2. Influence of hydrogen on the fatigue limit of Alloy 718 with defects

As discussed in the previous section, small-crack regime defects were more likely to degrade the FG fatigue limit, rather than that of CG, owing to its lower defect-size tolerance ( cf. Fig. 3). Moreover, the difference in the fatigue limit of FG and CG specimens with drill holes of √ area = 463 µm was not significant ( σ w = 250 MPa in FG, σ w = 240 MPa in CG). Consequently, the Δ Κ th ,lc values of both materials were presumed to be nearly equivalent. In consideration of the afore-mentioned results, H-charged FG specimens were used to study the effect of hydrogen on the fatigue limit of defects as small-crack thresholds. On the other hand, H-charged CG specimens were preferred for the investigations into defects as large-crack thresholds. Figure 4 presents the S - N diagram of H-charged and non-charged FG specimens with small defects. The fatigue test results for defect-free or smooth specimens were plotted together in this figure. The artificial defects introduced into the FG specimens were the drill hole ( √ area = 93 µm), as shown in Fig. 2 (a), and the 200-µm, semi-circular EDM notch ( √ area = 125 µm), as shown in Fig. 2 (d). Solute hydrogen degraded the fatigue limit of the drill-holed FG specimen by 20%, in the case of H-charging at 11 MPa ( C s = 26.3 mass ppm), and by 29%, in the H-charging condition of 100 MPa ( C s = 91.0 mass ppm). Similarly, in the EDM-notched specimens subjected to both H-charging conditions, hydrogen was seen to have negatively affected the fatigue limit by 26%. To understand the effect of hydrogen on the crack-growth behavior of the small cracks, replica observations were undertaken during the fatigue tests, with a comparison of the behavior in the non-charged versus the H-charged specimens. The results which stemmed from examination of the EDM-notched specimens are provided in Fig. 5. As recorded during the test, the progressive increase in surface crack-length, 2 a , has been plotted in Fig. 5 (a). It is to be noted that the stress amplitude, σ a , of each type of mark is different and duly indicated in the plot legend. The σ a of the H-charged specimens (solid colored marks) is lower than that of the non-charged specimens (open marks).