Tests on SENT and CT Specimens to Study Geometry Effects in the Brittle to Ductile Transition

Abstract

The accurate prediction of ductile fracture behaviour plays an important role in structural integrity assessments of critical engineering structures under fully plastic regime, including nuclear reactors and piping systems. Many structural steels and aluminium alloys generally exhibit significant increases in fracture toughness, characterized by the J-integral, over the first few mm of stable crack extension (Δa), often accompanied by large increases in background plastic deformation. Conventional testing programs to measure crack growth resistance (J–Δa) curves employ three-point bend, SEN(B), or compact, CT. However, laboratory testing of fracture specimens to measure resistance curves (J–Δa) consistently reveals a marked effect of absolute specimen size, geometry, relative crack size (a/W ratio) and loading mode (tension vs. bending) on R-curves. These effects observed in R-curves have enormous practical implications in defect assessments and repair decisions of in-service structures under low constraint conditions. Structural components falling into this category include pressurized piping systems with surface flaws that form during fabrication or during in-service operation. A research program was launched by EDF R&D to study geometry effects (e.g. triaxiality effects) in the brittle to ductile transition of carbon-manganese steels using Single-edge notch tension (SENT) specimens, by comparing the results obtained on these specimens with the results obtained on CT specimens. This paper presents the results of the tests conducted between −40°C and −100°C on a large number of specimens of both types. The toughness values of the SENT specimens appear to be included in the scatter of the CT12.5 ones, so the geometry effect between CT and SENT specimens in the brittle to ductile region is not significant. Moreover, the results of the CT12.5 cut in the L-S direction are not very different of those of the specimens cut in the T-S direction. The Master Curve methodology fits rather well the CT12.5 results, whereas the SENT results are not well covered by this methodology. The energetic approach called GP has been applied to the analysis of some tests. This approach shows that the geometry effect between both types of specimens is limited, in agreement with the experimental observations.