Identification of dislocation reaction kinetics in complex dislocation networks for continuum modelling using data-driven methods

Plastic deformation of metals involves the complex evolution of dislocations forming strongly connected dislocation networks. These dislocation networks are based on dislocation reactions, which can form junctions during the interactions of different slip systems. Extracting the fundamentals of the network behaviour during plastic deformation by adequate physically based theories is essential for crystal plasticity models. In this work, we demonstrate how knowledge from discrete dislocation dynamics simulations to continuum-based formulations can be transferred by applying a physically based dislocation network evolution theory. By using data-driven methods, we validate a slip system dependent rate formulation of network evolution. We analyse different discrete dislocation dynamics simulation data sets of face -centred cubic single-crystals in high symmetric and non-high symmetric orientations under uniaxial tensile loading. Here, we focus on the reaction evolution during stage II plastic deformation. Our physically based model for network evolution depends on the plastic shear rate and the dislocation travel distance described by the dislocation density. We reveal a dependence of the reaction kinetics on the crystal orientation and the activity of the interacting slip systems, which can be described by the Schmid factor. It has been found, that the generation of new reaction density is mainly driven by active slip systems. However, the deposition of generated reaction density is not necessarily dependent on the slip system activity of the considered slip system, i.e. we observe a deposition of reaction density on inactive slip systems especially for glissile and coplanar reactions.

Automated summary

Plastic deformation of metals involves the formation and evolution of complex dislocation networks, which are crucial for the development of crystal plasticity models. This study demonstrates the transfer of knowledge from discrete dislocation dynamics simulations to continuum-based models through a physically based dislocation network evolution theory. The results show that the evolution of dislocation networks is influenced by crystal orientation and the activity of slip systems.