ASSESSMENT OF DYNAMIC FRACTURE TOUGHNESS VALUES K_(Jc) AND REFERENCE TEMPERATURES T_(0,x) DETERMINED FOR A GERMAN RPV STEEL AT ELEVATED LOADING RATES ACCORDING TO ASTM E1921

摘要

Ferritic steels usually show significant embrittlement with increasing loading rates in the lower shelf and transition region, i.e. the fracture toughness versus temperature curve K_(Jc)(T) is shifted to higher temperatures. However, this embrittlement is limited and the ASME-reference curve K_(IR) is commonly accepted as a lower bound curve. It was one aim of this project, to verify this lower bound curve for the German reactor pressure vessel steel 22 NiMoCr 3 7 (≈ A 508 C1.2) in particular for elevated loading rates. The investigations with high speed fracture mechanics tests were performed within a joint project of Fraunhofer IWM and MPA. Details are reported in Boehme et al. (2012). The results of compact tension and three point bend specimens were evaluated according to ASTM E1921 (2011) following the Master Curve approach. All determined dynamic K_(Jc)-values were above the ASME reference curve K_(IR). The applied Master Curve evaluation with an assumed Weibull distribution according to ASTM E1921 (2011), results for some test series in 5%-fractile lower bound curves, which are below the K_(IR)-curve. However, a lognormal-distribution, recently proposed for crack arrest in ASTM E1221 (2010), leads to a more meaningful agreement of the calculated 5%-fractile curves with the K_(IR)-curve. Statistical analysis of the data confirmed, that the lognormal-distribution is a best fit of the experimental high rate data. Finally, the data distribution tends to show a second peak, thus giving a hint, that two mixed collectives of data may exist, a more brittle and a more ductile one. At the highest loading rates the more brittle one seems to dominate, resulting in comparable T_(0,x)-values, independent of specimen type, size and loading rate, and thus possibly determining a physically caused lower limit of embrittlement, which confirms the engineering lower bound K_(IR)-curve. This final conclusion would be subject to further investigations.