PSI - Issue 19

Jean-Gabriel SEZGIN et al. / Procedia Structural Integrity 19 (2019) 249–258 Jean-Gabriel Sezgin et al./ Structural Integrity Procedia 00 (2019) 000 – 000

254

6

3.4. Test frequency-dependent failure mechanism

It has been established that the fatigue-life of the present specimen is closely related to the FCG properties. Thus, the H-assisted FCG acceleration ratio, defined as the ratio of fatigue-life of non-charged to H-charged specimen N air / N H2 , was analysed. Figure 5 shows the test-frequency dependence of the H-assisted FCG acceleration ratio measured for stress amplitudes between 150 MPa and 400 MPa. For all the testing conditions, the FCG acceleration ratio appeared to be bounded. The literature suggested that in case of the HISCG model, an upper bound equal to 30 exists for the BCC structure, as illustrated by the example of low-alloy steels plotted with a dashed line (Matsuoka et al. 2017). In addition, the fracture surface within the framework of the HISCG model was covered with only QC. In the present case, in high-cycle regime, the upper bound of the H-assisted FCG acceleration ratio was around 30 and the fracture surface was covered with QC; thus, the observed FCG acceleration could be justified by the HISCG model. However, in low-cycle regime ( e.g. stress amplitude of 300 MPa), the upper bound exceeded the value of 30, up to around 100 and, additionally, the fracture surface was covered with QC and IG. This fact translated that the IG fracture provoked an FCG acceleration produced by a mechanism different from the HISCG model, leading to an increased value of the upper bound. The graph below clearly shows an increase of the FCG acceleration ratio with the stress amplitude and the decrease of the testing frequency. For this reason, it was relevant to quantify the amount of IG facets covering the fracture surface in order to verify the occurrence of this additional failure mechanism.

Figure 5 – Test-frequency dependence of the H-assisted FCG acceleration ratio illustrated for four stress amplitudes. The acceleration ratio appears to be bounded at different levels depending on the stress amplitude. The data related to SCM435 have been taken from (Matsuoka et al. 2017) Percentage of IG surfaces was measured on the dataset corresponding to a stress amplitude of 300 MPa, the SEM observations being available in a wide range of test frequencies. These measurements were conducted by observing fracture surface located at 1 mm from the notch root by SEM. Figure 6 illustrates the measurement results by highlighting in red the IG facets in the 10 Hz (a) and the 10 -3 Hz (b) cases. The measurements were realized in an area judged representative of the fracture mode and extended enough to limit localized measurements. The micrographs showed that the fraction of IG surface was 3.6% at 10 Hz (a) and 15.3% at 10 -3 Hz (b). The results of the measurements in the whole frequency range were represented in Figure 7. This figure clearly shows a correlation between the percentage of IG facets covering the fracture surface and the testing frequency. The data were then fitted and the results were given in the frame and represented by the dashed line. 3.5. Prediction of the test frequency-dependent fatigue crack growth rate The experimental facts suggested that in the H1150 steel, the underlying mechanism of H-assisted FCG acceleration presented some similitude with the HISCG mechanism. The HISCG mechanism has been identified in low-alloy steels with the UTS of ≤ 900 MPa (e.g. JIS-SCM435, JIS-SCM439) (Matsuoka et al. 2017). However, the experimental facts have concluded that at the low-cycle regime, the upper bound of the FCG acceleration ratio exceeded 30 and IG facets were observed on the fracture surface. In this case, the fracture surface has somehow