Ductile fracture in thin sheet metals: a FEM study of the Sandia fracture challenge problem based on the Gurson-Tvergaard-Needleman fracture model

In this paper, the Gurson-Tvergaard-Needleman (GTN) fracture model has been implemented in a 3D nonlinear finite element framework to investigate ductile failure. The simulation consists of both model parameter calibrations and predictions based on the experimental configurations that were described in the Sandia fracture challenge in 2012. The goal is to test the robustness of the GTN model in predicting crack initiation and propagation in a ductile structural stainless steel (15-5 PH). It has been observed that the predicted load drops at crack initiation agree well with the experimental data. Both the simulation and experiments reveal that cracks always initiate in the subsurface. The observed crack path in the experiments, however, is not unique and is strongly influenced by the geometric variations in the specimen introduced in the fabrication process. The crack path is quite sensitive to the relative location of pre-existing features (notch and holes) in this challenge problem, and a very small variation can potentially lead to different path, hence affecting the ductile failure pattern significantly. Variations in the crack path are also observed in the simulation with different choices of GTN model parameters that control the void nucleation and growth behavior.