Filtern
Erscheinungsjahr
- 2015 (1) (entfernen)
Dokumenttyp
- Vortrag (1)
Sprache
- Englisch (1)
Referierte Publikation
- nein (1)
Schlagworte
- Concrete (1) (entfernen)
Eingeladener Vortrag
- nein (1) (entfernen)
Simulations of concrete on the mesoscale require complex geometry models that resolve single aggregates. Since they fill up to 60-80% of the virtual specimen, efficient algorithms for their placement are necessary.
Event-driven molecular dynamics (EDMD) algorithms using growing spheres are proven to create dense polydisperse sphere packings. In this paper, a modified EDMD algorithm is presented and compared to the widely used random sequential addition algorithm. It reaches denser aggregate packings and saves computation time at high volume fractions.
When meshing the geometry, a minimal distance between the aggregates ensures undistorted elements. It is obtained by increasing the aggregate size during the placement process, which makes the packing problem more challenging. For a given aggregate volume fraction, the presented algorithm maximizes this distance and meshable mesoscale geometries for concrete specimens with more than 70% aggregate content are produced.