Powder compaction with polygonal particles built from radially extending one-dimensional frictional devices

Powder compaction is a major ingredient in a wide range of production techniques for objects of every day importance. Diverse applications – from pharmaceutical tablets to metallic and ceramic parts, where compaction is usually a part in the sintering process – are covered by current technology. The aim of cold compaction is to increase the relative density of the part. Due to the granular nature of the powder material compaction is random process which requires careful mastering and comprehensive understanding. The experimental access to compaction processes, even as simple ones as uniaxial compaction, are limited. Therefore simulation of compaction processes offers the opportunity to improve understanding of powder compaction. The Direct Element Method (DEM) simulates the powder as individual particles. These particles are distinct from each other and the forces applied to the entire powder are equilibrated by the contacting force between the particles. Usually, an explicit time integration, with or without considering dynamic effects, allows the particles to move and the powder to be compacted. Especially for metallic powders the plastic deformation of individual particles plays an important role and has a perceptible influence up to the macroscopic scale of the whole part. This leads to efforts to formulate a DE method in which individual particles are discretized as distinct FE models. Yet, such approach is deemed too costly for the simulation entire parts. Therefore a new approach for plastic particle deformation has been devised. The particles are simulated as ’hedgehogs’ of one-dimensional frictional devices. The frictional devices form spikes that extend in a radial way from the center of the particle. The tips of the spikes are connected and the connections form the edges of the polygonal particles. The contacts between the particles are found by geometric means as intersections of the spikes of one particle with the edges of the particle’s contact partner. This indents the spike and its frictional device generates forces that act on the two particles. Multiple contacts between two particles are allowed and concave particles can be treated intrinsically. Preliminary results indicate that this new approach to model plastic particles might bridge the gap between sufficiently realistic behaviour of the plastic particles and low computational effort. However, for now, the results are based on two-dimensional proofof-concept simulations and, for the sake of simplicity, the springs in the spikes are linear and the friction is rate-independent and perfectly plastic. These model assumption already offer considerable liberties to tune the plastic behaviour of the particles to experimental results