The tentative test procedure for determining J-R curves (crack-growth resistance curves based on the J-integral) on compact and three-point bend specimens showed a discrepancy between specimen types: the bend specimen gave a higher J-R curve than the compact specimen. The purpose of this paper is to investigate this discrepancy, numerically, by simulating the fracture process on compact and bend specimens using a two-dimensional finite-element analysis and a critical crack-tip opening displacement (CTOD) criterion.

An elastic-plastic (incremental and small strain) finite-element analysis was used with the critical CTOD criterion to model stable crack growth in compact specimens made of HY-130 steel. A combined analysis between plane-stress and plane-strain conditions was found to simulate the fracture process quite well. In the combined analysis, plane-strain conditions were imposed around the crack tip while plane-stress conditions were imposed elsewhere. The same analysis was then used to predict stable crack growth behavior of bend specimens. J-R curves were then calculated from the numerical results for each specimen type using several different methods. In general, the J-R curves obtained from the bend specimens were higher than those obtained from the compact specimens, especially beyond maximum load. Up to maximum load, however, the modified deformation theory of plasticity J*_D and the contour-integral J_Γ methods gave essentially the same J-R curves for both specimen types.