PSI - Issue 33

Naoya Oie et al. / Procedia Structural Integrity 33 (2021) 586–597 Oie, N. / Structural Integrity Procedia 00 (2019) 000–000

596

11 �17�

� �

� ��� � �� �

Fig. 11. Optical microscopy image of a microcrack.

Fig. 12. Distribution of P ( X ) of the original and modified Yoshizu model.

The value of γ for each model was optimized so that the average of P ( X ) was maximized. Table 4 shows the results of optimizing γ in each f based on the Bordet and Yoshizu models and � in all local approach models mentioned in this study. In Table 4, � shows that the fracture probabilities of the optimized Bordet and Yoshizu models are higher than those of the original models. The distributions of P ( X ) of the original Yoshizu model and the addition of with the optimized γ exponent are shown in Fig. 12. Whether this functional form is appropriate is highly debatable, but at least it suggests that not only and � but also are involved in the brittle fracture of steel. Table 4. � for each local approach model. model γ f � Beremin - � �� 5.83% Bordet - � � �� 6.91% Yoshizu - �� � � � �� 6.54% Opt.-Bordet 3.61 7.24% Opt.-Yoshizu 5.93 7.48% 5. Conclusion In this study, we concluded the following points:  The critical quasi-CTOD was affected by the specimen thickness and notch depth because of plastic constraints.  For the critical Weibull stress, the modified Weibull stresses (Bordet or Yoshizu model) allow for the evaluation of brittle fracture initiation properties more accurately than the original Beremin model and are more consistent with the fracture initiation position.  From the example of a microcrack, not only � but also is considered to contribute to microcrack nucleation. Based on the maximization of the prediction accuracy of the initiation position, a new modified Weibull stress