Analysis of the Difference Between the Reference Temperature Values Derived From Conventional and Mini-C(T) Specimens: Effect of the Number of Valid Fracture Tests

Abstract The Long-Term-Operation (LTO) of nuclear power plants (NPPs) requires, among other technical and scientific challenges, a precise knowledge of the fracture behavior of the steels (both base material and welds) that constitute the reactor pressure vessel (RPV). This material property evolves (decreases) with the irradiation level, so surveillance programs are defined and completed to measure it throughout the NPP operational life and to ensure that the structural integrity of the RPV is not jeopardized. The surveillance programs are additionally performed by testing materials placed inside the RPV (surveillance capsules) which were initially designed to cover the initial lifespan (i.e., 40 years). This implies that the amount of available material to be tested during the life extension of NPPs may be scarce and it is not possible to perform fracture toughness tests using conventional specimens (e.g., 1T or 25.4 mm thick). In this context, mini-C(T) specimens (0.16T or 4 mm thick) appear as a key technology to perform the fracture characterization during the LTO, as long as 8 mini-C(T) specimens may be extracted from previously tested Charpy specimens or 48 mini-C(T) specimens may be obtained from a previously tested 1T CT specimen. This provides actual values of the material fracture toughness (unlike in the case of Charpy tests), multiplies the knowledge (i.e., the experimental results) about the material fracture behavior and allows irradiated tested material to be reused. In other words, mini-C(T) specimens optimize the use of the finite remaining stock of surveillance material from operating NPPs. Constituting such a promising technology, the use of mini-C(T) specimens to characterize the reference temperature (T0) of ferritic steels (i.e., the fracture behavior within the ductile-to-brittle transition zone, DBTZ) has been extensively analyzed in the last decade. Generally, mini-C(T) specimens provide similar results to those generated by using conventional (e.g., 1T) specimens in both irradiated and non-irradiated conditions. However, experimental results found in literature reveal situations where the deviation between the T0 predictions of conventional and mini-C(T) specimens are basically negligible, and situations where the deviation between such predictions achieve up to 40 °C. This situation must be clarified so that regulatory bodies accept the use of mini-C(T) specimens for the fracture characterization of RPV steels. This paper provides insights into how the mentioned deviations depend on the extension of the experimental program performed for the characterization of T0, revealing that larger experimental programs (i.e., more fracture toughness results) provide much smaller discrepancies and also suggesting that for very large programs the two types of predictions tend to converge.