461156 Numerical Simulations of Fluidized Beds in Fine-Grid Geometries Using Source Smoothing with Cell Clustering Method for DEM-CFD Coupling

Thursday, November 17, 2016: 12:49 PM
Golden Gate (Hotel Nikko San Francisco)
Oleh Baran1, Doran Greening2 and Petr Kodl2, (1)CD-adapco New Hampshire, Lebanon, NH, (2)CD-adapco, Lebanon, NH

Discrete Element Method (DEM) proves to be often useful and accurate for modeling flow of particles in fluidized bed [1,2]. In typical industrial cases the number of particles being large imposes computational cost constraint on the choice of the 2-way coupling method between fluid (Eulerian) and particle (DEM) phases. This constraint singles out the coupling approach which requires particle size being smaller than cell size in order to maintain the smoothness in particle solid fraction distribution which is used to calculate the momentum sources in Eulerian phase. Unfortunately this requirement for cell size with respect to particle size is not easy to realize for typical fluidized bed geometries which often include high curvature boundaries and fluidized bed internals, resulting in inaccuracy or instability of the solution.

In this work we introduce the source smoothing approach for 2-way DEM-CFD coupling which allows to run efficient, accurate and stable 2-way coupled simulation without requirement of cell size being larger than particle size. Within this approach the second coarser grid is generated from original fine grid by clustering neighboring cells. The fluid solution is solved on the original fine grid. The particle volume fraction is calculated on the coarser grid of clustered cells and mapped onto the fine grid to be used by fluid solver. Compared to the two-grid formulation for fluid-particle systems proposed by Deb and Tafti [3] our approach has an advantage of being applicable for complex geometries that resolve fine details. This is because of adopting the novel algorithm for clustering cells which minimizes the difference in volume and shape between cells of resulting coarse grid. In addition, our approach is implemented within commercial software STAR-CCM+ version 11.06 by CD-adapco and allows running DEM coupled CFD models in parallel on many cores, as well as running coarse grain particle model [1] at the same time.

In this work we first validate the source smoothing method for 2-way coupling by reproducing the fluidized bed results obtained in [1]. We then show that modifying geometry to include new small scale details (by including the tube bundle or other internals into the model) does not render the model unstable or inaccurate even though the coarse grain particles are larger than cells throughout the volume of geometry.

[1] O. Baran, A. Newale, D. Greening, DEM simulations of industrial size fluidized bed using coarse grain model with particle size distribution, AIChE 2014
[2] R. H. Aglave, O. Baran, M. Tandon, A. Karnik and . Lo, Numerical Simulation of Dense Gas-Solid Fluidized Beds: Comparison between Eulerian Multiphase and Discrete Element Methods, AIChE 2015
[3] S. Deb, D. Tafti, A novel two-grid formulation for fluid–particle systems using the discrete element method, Powder Technology 246 (2013) 601–616

Extended Abstract: File Not Uploaded