Modelling of size effects on fracture in the brittle-to-ductile transition regime

Measurements of toughness in the ductile-to-brittle fracture regime exhibit an effect of testpiece size. For a given level of toughness (KJC), this may be considered as a positive shift in the transition temperature (∆T) with increasing specimen thickness for a particular geometry. This Paper describes a numerical programme undertaken to investigate the influence of specimen size on the fracture toughness behaviour of submerged-arc weld material in the ductile-to-brittle transition regime. The influences of sampling volume and constraint on ∆T have been assessed for 10 mm thick CT (CT-10) and 20 % side-grooved pre-cracked Charpy specimens (PC-CVN) relative to 25 mm thick standard plane-sided compact-tension (CT-25) specimens. Three-dimensional, elastic-plastic finite element analyses have been used to assess the constraint behaviour of these specimens, and the BEREMIN model has been applied to predict ∆T. The analyses have indicated a loss of constraint over a significant proportion of the crack front for both CT-10 and PC-CVN specimens at toughness levels greater than approximately 100 MPa√m. For CT-10 specimens, constraint-loss is most significant towards the outer edges of the specimen whilst high constraint conditions are initially maintained over the central region of the crack-front. This high constraint region may be viewed as an effective crack-front width that reduces with increasing deformation. For side-grooved PCCVN specimens, constraint loss occurs uniformly across the crack front. Statistical models, based on empirical data, give a value of ∆T = 18 oC for this material at a mean reference toughness of 100 MPa√m. The BEREMIN model tends to over-predict this value by ~ 35 oC for CT-10 specimens and ~ 55 oC for side-grooved PC-CVN specimens.