Geometric dependence of fracture criteria in AL2024

The failure by tensile fracture of a 2024 aluminium alloy was investigated using a number of fracture criteria. The J integral method was shown to exhibit a significant dependency on specimen geometry and was therefore deemed unsuitable for use across a broad range of specimen types. Three 'local approach' criteria were also employed, the standard Beremin type model based on maximum principal stress and hydrostatic stress, as well as an alternative method based on dilatational strain energy density. All three models were found to be capable of characterizing the failure behaviour of the material in question, although it was noted that the results of the Beremin method were highly sensitive to the data used to calibrate the model. The energy type method was found to be more robust, but under predicted real failure probability at low loads. Introduction of a global failure threshold to account for the propagation dependency of failure was found to greatly improve the quality of failure prediction and the stability of the calibration process.