Large scale simulation of particle coating using coupled CFD-DEM

Abstract

Spouted fluidized beds are widely used for particle coating because of its excellent mixing rates and favorable heat and mass transfer characteristics. The combination of Computational Fluid Dynamics (CFD) and the Discrete Element Method (DEM) has previously been employed for simulation of the complex phenomena of such processes. However, simulating a fluidized bed on a large scale with DEM requires exceptional computational power as all the interactions between the particles are fully resolved. In addition, simulating the spray droplets further increases the computational demand. Accordingly, coupled DEM and CFD simulations with a well resolved spray have typically been limited to system sizes from a few thousand up to much less than a million particles. The goal of the current thesis was to perform spray coating simulations on systems with more than 1 million particles, including a Lagrangian spray phase and a well resolved fluid. The thesis is carried out using the DEM-CFD solver IPS FluidizationTM developed at Fraunhofer Chalmers Centre. The solver is based on an in-house DEM code and the in-house immersed boundary CFD code IBOFlowR . Due to heavy use of the Graphical Processing Unit (GPU), the code allows simulating a large number of particles and a well resolved fluid on a standard desktop computer. In the first part of the thesis, single spout simulations are carried out to validate the coupled solver. The simulations show excellent agreement with the experimental data available in the open literature. Further, 1D studies are conducted for verifying the heat transfer model. The numerical predictions are shown to be accurate based on comparisons with analytical 1D models. Finally, large scale simulations including the spray and drying are conducted on a Wurster bed system with both the particles and the spray considered in a Lagrangian sense. The simulations show the versatility of the tool and the possibility to e.g. characterize the particle coating thickness in terms of the original particle size, as well as it proves applicability to cases with more than 1 million particles.