Background and motivation: The capabilities of predicting failure in ductile metallic materials have significance in additive manufacturing, metal forming, fatigue analysis, and numerous defense applications. The development of predictive numerical tools relies on: (1). constitutive relations accurately describing the influence of poly-crystalline microscopic structures and defects, such as grain orientations and voids, on material macroscopic properties; (2). numerical methods effectively bridging plastic deformation and damage mechanisms at the microscale, such as dislocation and twinning (Fig. 1), with material macroscale responses, such as shear localization and fracture.

Collaborated with researchers from the Los Alamos National Laboratory, we developed an explicit finite element formulation using embedded weak discontinuity to treat shear localization under severe dynamic loading conditions. Particularly, our team adopted a global tracking strategy to ensure the continuous propagation of shear band surfaces through the connectivity graph of general three-dimensional finite element meshes (Fig. 2). This approach can avoid explicitly resolving shear bands during the spatial discretization, reduce the mesh dependence of localization problems, and improve the overall computational efficiency. In hexagonal close packed materials such as titanium, crystallographic slip and deformation twinning are plastic deformation mechanisms with a significant disparity of length-scales. Using the embedded weak discontinuity, we proposed a novel computational approach to combine both mechanisms in the crystal plasticity finite element method. This approach was successfully applied to simulate the whole process of twin initiation, propagation and growth in multi-grain structures.