Structural integrity of reactor pressure vessels (RPVs) in nuclear power plants (NPP) is essential for safe and long-term operation, as RPVs are typically non-replaceable components in harsh environments, including neutron radiation, high temperatures, and mechanical stresses. Neutron irradiation causes embrittlement in RPV steels, typically a reduction in fracture toughness and a shift in the ductile-to-brittle transition temperature (DBTT) to higher levels. Assessing the fracture toughness of RPV steels is therefore critical for predicting the NPPs' lifetime and ensuring reactor safety. The Master-Curve method is a powerful approach for the determination of fracture toughness by identifying the reference temperature T_0 associated with the DBT. However, applying the Master-Curve method to sub-sized, neutron-irradiated specimens presents two primary challenges: small irradiated specimens may not meet the minimum size requirements for valid T_0 determination; intrinsic inhomogeneity within the steel can introduce additional scatter in toughness data, impacting T_0 calculations if not properly accounted for.

This research aims to understand and quantify effects of specimen size and material inhomogeneity on fracture behavior in the DBT regime by evaluating low-alloy RPV reference steel JRQ. A matrix of elastic-plastic fracture tests were conducted with compact tension (C(T)) specimens of different sizes, namely 0.5T, 0.18T and 0.09T C(T) of a notable level of inhomogeneity difference manifested by different depths of extraction locations. Results indicate significant scatter in fracture toughness data, attributed to both specimen size and material inhomogeneity. Notably, the middle plate, with its lower yield strength, exhibited a larger constraint loss effect than the surface plate, resulting in a lower K_Jc,limit, and more rapidly constraint loss in specimen with a lower T_0 in 0.18T C(T) specimens.

To address the limitations of the 'size-adjustment' recommended in the ASTM-E1921-24 standard, which assumes fully constrained specimens and only consider the statistical effects of the crack length, a two-step correction procedure was developed based on the calibration of the local approach to fracture criterion involving the attainment of a critical stress over a stressed volume near the crack tip using a series of finite element simulations. This procedure involves applying a constraint loss correction to maintain a stressed area across specimen sizes, namely between sub-sized and full-sized specimens, ensuring the comparable "driving force" for fracture initiation. The second step applied the recommended crack front length adjustment, bringing 0.18T C(T) into closer agreement with 1T C(T) specimen and ultimately achieving a more consistent T_0 determination across specimen sizes.

This work contributes to the development of methodologies for analyzing fracture behavior in the DBT regime of sub-sized inhomogeneous RPV steels. By refining the size adjustment procedures and distinguishing the effects of material inhomogeneity from constraint loss, this research reveals how to interpret and transfer data from small fracture specimens to real structure and enables more accurate embrittlement assessments using miniaturized specimens. These findings support the ongoing effort to safely manage RPV degradation in nuclear reactors, offering a practical approach to predict structural integrity with irradiation-induced embrittment.