Coarse-Grained DEM for Large-Scale Fluidized Bed Applications

Fluidized beds are used in a wide range of applications including those in the chemical and process industry. To ensure the efficient and safe operation of these unit operations, Computational Fluid Dynamics (CFD) is often employed during the design and optimization of the units. Yet, when it comes to numerical simulation of industrial scale units, classical methods like the Kinetic Theory of Granular Flow (RANS-KTGF or KTGF) and the Discrete Element Method (DEM) are limited by their hardware requirements. In the case of KTGF it is the high number of cells needed to resolve all important flow structures (i.e., grid resolution), whereas DEM is limited by the particle count that can be tracked efficiently. To avoid this, in the KTGF framework, sub-grid or filtering models like the EMMS-drag model can be used. However, this increases the complexity of the model by adding several unknown closures.

A promising DEM framework-based approach is the so called Coarse-Grained DEM (CG-DEM). This approach tries to circumvent the hardware-related limitations of DEM by combining several individual particles into one representative parcel or cloud. As in DEM, the parcels trajectories and collisions are tracked individually and interaction forces (i.e., gravity, drag) are scaled to account for the artificially reduced number of interactions.

This current study will show the capabilities of CG-DEM in the field of large-scale fluidized bed applications. High fidelity experimental data, taken from a twelve-inch semi-circular pilot-scale unit, are compared with numerical results obtained by CG-DEM. Particles of different Geldart Group classification are investigated. The impact of drag, collision force scaling, and the effect of increasing the number of particles per parcel on the accuracy of the numerical results will be compared and discussed.