In the ductile-to-brittle transition range of ferritic steels, fracture toughness exhibits a size effect. Up to now, in the safety demonstration of the French Reactor Pressure Vessel (RPV) integrity, a size effect correction has been considered by the operator to take into account fracture toughness variation of ferritic steels with crack length. The correction consists in increasing the toughness estimated on the RCC-M curve by a factor which depends on a reference length and on the crack length considered.
IRSN has already examined the relevance of this correction through statistical analysis of toughness results coming from two ferritic steel databases. To complete its evaluation on French RPV steel, IRSN has supported a large experimental campaign on 16MND5 steel at different temperatures in the ductile-to-brittle transition (from −150°C to −50°C), including tests on various Compact Tension (CT) specimen geometries. Specimens with semi-elliptical crack have been also considered.
The results confirm the observations made in its previous study: a size effect exists on mean or median toughness, for the latter more or less in accordance with Beremin theory. Nevertheless, the minimum toughness appears to be independent of the specimen geometry. This indicates that the use of a size effect correction on minimum toughness is not relevant.