ResearchList of all research projects
Non-konvexe Partikel und parallelisierte Berechnung

Non-convex particle shape and parallelization

Led by:  P. Wriggers
Team:  M. Hothan
Year:  2014
Funding:  DFG (Project: IRTG 1627)

Simulation of a steel ball colliding with a wall of brick elements. The wall consists of 72 brick elements made of sphere clusters (162 spheres).



Many materials found in nature or technical processes have a granulated structure. Examples are sand and ores, fruits and grain, (dry) pharmaceutical and chemical products. Compared to other materials, granular materials are difficult to handle: Different particle shapes result in different material behaviour.


Non-convex particle

The macroscopic behaviour of particles is based on the behaviour of every single particle. Most simulations are based on purely convex-particles like spheres of ellipsoids due to the low computational costs. A cluster of spheres as a particle model combines the benefits of spherical particles (low memory consumption and low computational costs) with the needs for more complex shapes.



A more realistic particle model leads to a more complex particle description which means a higher computational effort to calculate the particle's behaviour. The memory consumption is also higher compared to spheres or ellipsoids. By simulating a large number of particles the resources of a single computer are to small. By using parallelization techniques like MPI theses limit can be avoided.