Abstract
The standard master curve (MC) approach has a major limitation in that it is only applicable to homogeneous datasets. In nature, steels are macroscopically inhomogeneous. Reactor pressure vessel (RPV) steel has different fracture toughness with varying distance from the inner surface of the wall due to the higher cooling rate at the surface (deterministic material inhomogeneity). On the other hand, the T0 value itself behaves like a random parameter when the datasets have large scatter because the datasets are for several different materials (random inhomogeneity). In this paper, four regions, the surface, 1/8 T, 1/4 T, and 1/2 T, were considered for fracture toughness specimens of Korean Standard Nuclear Plant (KSNP) SA508 Gr. 3 steel to provide information on deterministic material inhomogeneity and random inhomogeneity effects. Fracture toughness tests were carried out for the four regions at three test temperatures in the transition region and the microstructure of each region was analyzed. The amount of upper bainite increased toward the center, which has a lower cooling rate; therefore, the center has lower fracture toughness than the surface so reference temperature (T0) is higher. The fracture toughness was evaluated using the bimodal master curve (BMC) approach. The results of the BMC analyses were compared with those obtained via a conventional master curve analyses. The results indicate that the bimodal master approach considering inhomogeneous materials provides a better description of scatter in the fracture toughness data than a conventional master curve analysis does.
