In his dissertation, Kostas Giannis develops and validates a novel elastic-plastic multi-contact model for the discrete element method (DEM). The aim is to significantly improve the predictive accuracy of DEM simulations – especially when compressing pharmaceutical powders into tablets under high loads.
The model expands on existing approaches by realistically depicting both elastic and plastic deformations of particles. This allows the mechanical properties of tablets to be predicted more accurately and digital twins of pharmaceutical manufacturing processes to be created. These digital representations open up new possibilities for process monitoring, optimization, and the targeted development of formulations, thereby increasing quality and efficiency in tablet production.
Another focus of the work is on more realistic modeling of particle geometries. While DEM simulations have mostly assumed spherical particles up to now, the new approach also integrates non-spherical shapes using modern AI methods. This increases the applicability of DEM in materials science and beyond in many industries where particle behavior plays a central role.