Comparison of semi-empirical and dislocation density based material equations for fast modeling of multistage hot working of steel

Work hardening in metal forming can be modeled using different material model types. To capture the material response to plastic deformation semi-empirical models developed by e.g. Sellars and others compete with the physically-based - or internal variable - models developed by e.g. Bergstrøm or Kocks and Mecking. The more physical nature of the internal variable models means that they typically consist of complex systems of differential equations build upon a multitude of parameters (often > 20) to describe the various specific phenomena involved. The question thus open to debate is whether those internal variable models are more an advantage or a burden from an application point of view. In order to shed light on this question a direct comparison between semi-empirical and internal variable material models is drawn. Both model types are assessed using the following categories: model complexity, effort of model calibration, performance in compression tests, and applicability to hot rolling. The general quality of the model fitted to the same high manganese steel is demonstrated by a double hit compression tests. Additionally the material models are used to predict flow stress, recrystallized fractions and roll forces in a typical hot strip rolling schedule i.e. a complex multistage hot working operation. A difference in the best approach and necessary effort was exposed during model calibration. Validation trials reveal a good agreement between both model types. When modeling a complex rolling operation, however, profound differences in the microstructure predictions become apparent