Reactor pressure vessel (RPV) multilayer welding seams show an inhomogeneous structure. It raises concerns that the evaluation of non-uniform material might not be amenable to the statistical analysis methods on which the Master Curve approach is based. In particular with regard to weld metals, it can be expected that the cleavage fracture toughness is strongly influenced by the orientation of the Charpy size SE(B) specimen. The T-L oriented SE(B) specimen (axis axial and crack propagation in circumferential direction) comprises of various welding beads along the crack front whereas in a L-S specimen (axis axial and crack propagation through the thickness) the crack tip is located in one welding bead with an approximately uniform structure.
The paper summarises fracture toughness results measured on welding seams of decommissioned and non-commissioned RPVs of WWER type nuclear reactors and the non-commissioned Biblis-C RPV. Specimens of T-L and T-S orientation were tested. The results show, that in general the cleavage fracture toughness values, KJc-1T, follow the Master Curve description. However, the number of KJc-1T data outside the 2% and 98% tolerance bounds is larger than predicted by the underlying model, which indicates non-uniform material.
There is a large variation in the evaluated through thickness T0 values of the investigated multilayer beltline welding seams. Within the sampling range of the surveillance specimens, T0 values vary with a span of 30 to about 60 K depending on the applied welding technology. The fracture toughness strongly depends on the intrinsic weld bead structure. Hence, the position of the fatigue crack tip of the pre-cracked SE(B) specimen at the multilayer welding seam is crucial and defines the cleavage fracture toughness. Modified Master Curve based evaluation procedures like the MC based approach of the SINTAP procedure were applied to get fracture toughness values which are representative for the most brittle fraction the test series.
Despite of the pronounced non-homogeneity of the micro-structure along the crack front of T-L specimens, crack initiation sites are randomly distributed along the crack front. This means that one of the basic assumptions in ASTM E1921, i.e. the uniform distribution of initiation sites, is fulfilled also for the T-L specimens from the multilayer weld metal.