Skip over navigation

 

Creating a failure criteria for aluminum alloys from 3D finite element modeling of void growth and coalescence

Geoffrey Bomarito (Cornell University), Derek Warner (Cornell University)

Crack initiation and growth: methods, applications, and challenges

Wed 9:00 - 10:30

Barus-Holley 161

Ductile fracture is characterized by the nucleation, growth, and coalescence of microvoids, all of which are significantly influenced by the stress state in which they occur. In this study, the ductile fracture process of an AL 5083 alloy is simulated through use of computational cell models. In order to elucidate the effect of stress state on ductility, an array of cell models are each subjected to a given stress state, defined by a stress triaxiality and Lode parameter. The cells are loaded until failure in two loading orientations that correspond to the two expected failure mechanisms: ligament necking and shear localization. The minimum failure point of the two failure mechanisms is taken from each stress state and are combined to develop a failure surface, without assumption of its functional form. The result is a description of the failure of this material that is derived from basic principles and requires no fitting parameters. Subsequently, a larger scale Finite Element (FE) model, which employs the failure surface, is compared to experimental data. The larger scale FE model is shown to over predict ductility when compared to experiments. The effects of load path, initial void volume fraction, and heterogeneity on the accuracy of the larger scale model are also investigated. Through this investigation, void nucleation and heterogeneity in the microstructure are found to be responsible for the over prediction in ductility, emphasizing that their incorporation into cell models is critical to developing quantitatively accurate physics based ductile failure models.