The mechanical properties and behaviour of metallic materials mainly depend on the content and structure of dislocations network, this points out the need to incorporate microstructure concepts into the numerical models.
The goal is to correctly describe the main physical mechanisms occurring in metals during thermomechanical processes i.e. work-hardening, recovery, grain boundary migration, nucleation and grain growth related to dynamic, static or metadynamic recrystallization, phase transformation, ductile damage mechanisms...
Macroscopic and homogenized models, the so-called mean-field models, are widely used in the industry, mainly due to their low computational cost. If this mean field framework is quite convenient, it can be synonymous, for a given material, with a large amount of experiments with advanced laboratory devices.
Moreover, the homogenization of the microstructure does not usually enable to capture local events. Over the last decades, mesoscopic numerical methods (called full field models) have been developed in order to simulate explicitly the microstructure evolutions and to build a bridge with lower-scale approaches like dislocation dynamics or molecular dynamics. The idea behind these mesoscale simulations and the concept of digital material is that the morphology and the topology of the microstructure characteristics must be taken into account to develop accurate and reliable simulations.
Some trends in terms of development of these numerical methods will be illustrated before focusing on our recent efforts, concentrated on the DIGIMU industrial consortium and the homonym industrial ANR Chair, to shift from laboratory-centred modelling activities to helping industry equip itself with advanced modelling tools and skills concerning this domain.