PSI - Issue 13

Sergiy Kotrechko / Procedia Structural Integrity 13 (2018) 11–21 Sergiy Kotrechko / Structural Integrity Procedia 00 (2018) 000–000

15

5

ahead of a macrocrack / notch. Actually, this fact is the basis of a LA to fracture, where the fracture initiation criterion is formulated in a statistical statement. In classical version of LA [BEREMIN (1983)] and it modifications [Bordet et al. (2005)] probability of fracture initiation is described two- or three-parameter Weibull distribution. In these models, the number of crack nuclei is determined not explicitly, but as a ratio of PZ volume to the elementary volume 0 V . The latter is nothing but the volume per one crack nucleus, i.e. inverse density ρ of the crack nuclei. Thus, an increase in the intensity of crack nuclei formation in the irradiated metal, ρ , should lead to an increase in the probability of cleavage initiation and, correspondingly, to a decrease in the mean value of local cleavage stress f σ . Thus, radiation hardening Y ∆σ and radiation-induced reduction of the local cleavage stress f ∆σ are two main factors governing the service time of RPV steels. Unfortunately, all these ideas about the physical nature of radiation embrittlement are not taken into account in the life time prediction of RPV. 3. Lifetime prediction of RPV Idea of the conventional RPV life time prediction method is illustrated in Fig. 2. Determination of the critical value of fluence с Φ assumes the solution of two main tasks, namely: ( i ) determination of the critical shift a K Т of the fracture toughness Ic K curve of RPV steel (Fig. 2a); ( ii ) building of the trend dependence of shift of the critical embrittlement temperature F T ∆ on the fluence value (Fig. 2b) [VERLIFE (2008)]. The solution of the first problem is to build the curves of fracture toughness of RPV steel in the initial (unirradiated) state and after irradiation and also to construct the temperature dependence of the value of stress intensity factor I K for emergency cooling (pressurized thermal shock) of a reactor pressure vessel. The trend curve of radiation embrittlement (Fig. 2b) is usually built based on the results of tests for impact bending of irradiated Charpy surveillance specimens. The significant limitation of this approach is the use of a critical temperature shift F T ∆ of the Charpy impact test to predict the shift in the fracture toughness Ic K curve. From a physical point of view, the comparison of these shifts is not correct. Results of experimental studies carried out by Wallin et al. (1995), have shown that relation between F T ∆ and a shift of the reference temperature 0 T ∆ for pre-cracked specimens depends on the magnitude of fluence and may give a non-conservative error at large fluences, i.e. for long-term operation of power units.

a

b

Fig. 2. Determination of the critical temperature a

K Т ( a ) and the critical value of fluence

с Φ ( b ) (scheme):

unirr Ic K is the lower envelope curve of the temperature dependence of fracture toughness of RPV steel in unirradiated state; ∗ unirr Ic K is the same dependence shifted by 0 K T ( 0 K T is the initial critical temperature); irr Ic K is the temperature dependence of fracture toughness at the critical shift a K Т ( a K Т is the critical emergency temperature) ( a ); the trend curve for the shift of critical temperature, obtained by the results of testing of impact surveillance specimens ( b ).