TWI Industrial Member Report Summary 713/2000
A D Karstensen and M R Goldthorpe
Background
Assessments of the integrity of structures containing defects require values of fracture toughness measured using deeply notched bend specimens. These specimens provide lower bound results for fracture toughness, thus ensuring conservative assessments of structural integrity. However, the apparent fracture toughness of defects in structures which impose low levels of plastic constraint on a defect, for example shallow cracks or embedded defects in tensile stress fields, can be much higher than the recommended values measured on highly constrained test specimens. This can lead to unduly pessimistic assessments of defective structures that are actually fit for service. Furthermore, due to the scatter in fracture toughness of ferritic steel and weld metals in the ductile-to-brittle transition region, statistical consideration is needed to provide a reliable lower bound estimate of fracture toughness.
A methodology is thus required to account for both enhanced structural toughness and the scatter in fracture toughness results.
To achieve this, TWI has developed micromechanical models that can predict the stable growth of a flaw by ductile tearing or failure by cleavage fracture. The two elements of this coupled model use material-specific fracture parameters in finite element analyses to predict the behaviour of steel structures, including welded joints. Ductile crack growth is modelled deterministically using a finite element program. This is achieved by using material constitutive equations accounting for ductile damage in conjunction with a specially developed Crack Growth User Element along the crack path. Cleavage fracture is predicted statistically by post-processing the finite element analyses using the Beremin model; a method of predicting brittle fracture based on the weakest link concept. The coupled model can be used to predict structural behaviour in the ductile-to-brittle transition region and can take account of the enhanced fracture toughness of real structures.
Objectives
- To modify the existing Crack Growth User Element software to allow different advance criteria along the crack front and so provide more accurate modelling of ductile crack growth, in particular the breaking of shear lips.
- To tune the parameters of the ductile crack growth model against the results of ductile tearing tests carried out on SENB specimens of a steel used in offshore applications, and to examine the effect of using a different elementsize at the crack tip.
- To use the ductile crack growth model to predict ductile tearing in a more structurally representative surface notched specimen loaded in tension.
- To use the coupled model to predict cleavage failure of SENB specimens in the ductile-to-brittle transition region and make comparisons with test results.