Multi-Level Coarse-Grain Model in DEM and CFD-DEM Simulations
Sprache des Vortragstitels:
Englisch
Original Tagungtitel:
The 2nd International Symposium on Computational Particle Technology and 13th International Conference on CFD in the Minerals and Process Industries
Sprache des Tagungstitel:
Englisch
Original Kurzfassung:
The discrete element method (DEM), often used in combination with computational fluid dynamics (CFD), has proven to be a viable tool for the analysis of granular flows. In a broad range of industries, DEM and CFD-DEM simulations are successfully used to support process design and optimization. On the downside, the DEM is a computationally demanding method, due to actually resolving the contact between particles, and thus making it difficult to apply to large-scale systems.
The coarse-grain (CG) model of the DEM relaxes this computational restriction by replacing multiple equal particles by a single coarser (pseudo) particle, thus significantly decreasing the number of particles involved in the computations. However, due to the violation of geometric similarity, this approach fails to capture effects that intrinsically depend on particle size. Particularly, this becomes an issue in multi-scale systems typically found in large industrial facilities.
Seeking for a computationally feasible description of such large-scale systems, we have developed a method which concurrently couples multiple coarse-grain levels to adjust the resolution of the simulation as required. Spatially confined sub-domains of finer scale are embedded into coarser representations of the system and coupled by exchanging volumetric properties of the granular flow, such as mass flow rate, particle velocity and size distribution. On one side the coarse-grained data enables us to impose proper boundary conditions in each sub-region and at the same time the fine-scale information may be used to enhance the more inexact coarse-grain simulation. By this means it is possible to preserve the particulars of the granular system in spatially confined regions while keeping the benefit of the speedup of the coarse-grain model, where a lower level of detail is sufficient.