PSI - Issue 57

Stéphan Courtin et al. / Procedia Structural Integrity 57 (2024) 4–13 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

10

7

According to the previous assumptions, the effective stress intensity factors ΔK eff (see Eq.1) obtained with the 3 options are as follows (see Table 1): Table 1. Sensibility analyses about the cyclic elastic-plastic stress-strain curve. ‘RCC -MRx- E’ option ‘RCC -MRx- Sy’ option ‘shifted RCC - MRx’ option ΔK eff (MPa.m ½ ) 95.66 87.98 98.25 The numerical results obtained with the 3 options on the cyclic elastic-plastic stress-strain curve highlight, for this case, a difference reaching about 12 % between the lowest ΔK eff value (for the ‘RCC -MRx- Sy’ option) and the highest ΔK eff value (for the ‘shifted RCC - MRx’ option). This is not negligible and may significantly impact fatigue crack growth assessments because the n parameter in Paris law (see Eq.2) is about 4. Thus a 12 % difference in ΔK eff values leads approximately to a 55% difference in the evaluation of the crack length increase in Eq.2. As a preliminary conclusion, the issue of the choice of the cyclic elastic-plastic stress-strain curve must be kept in mind when using the ΔJ approach considered in this study. 5. Comparisons between numerical results and experimental data for the thermal loading #1 Then the overall fatigue crack growth evaluation scheme detailed above has been carried out. This paragraph presents the results for the thermal loading #1 which generates fatigue cracks propagating on both faces of the mock up, with quite similar crack growth rates (see Fig. 3). In a first simplified approach, only straight crack fronts have been numerically taken into account, before considering free crack fronts to try to better represent experimental results.

5.1. Straight crack fronts

Fig. 6 shows the comparisons between the numerical results and the experimental data for the thermal loading #1, for the so-called F3 crack (see Fig. 3), and for the 3 options on the cyclic elastic-plastic stress-strain curve. Four crack growth Paris laws from the French RSE-M code (2014) have been considered: mean or conservative, and in air or in PWR water environment. The comparisons are performed in the middle of the thickness of the mock-up, that is including some uncertainties since test data are only available at top and bottom faces.

Fig. 6. Numerical and experimental crack growths for the so-called F3 crack with the thermal loading #1, with straight crack fronts.