Performance improvement of existing drag models in two-fluid modeling of gas–solid flows using a PR-DNS based drag model

• We proposed a hybrid standard/PR-DNS drag model for two fluid modeling simulations.
• Constituent model failed to produce results close to the experimental results.
• We implemented switching between models based on interparticle cohesive forces.
• Modeling is significantly sensitive to the threshold of the cohesive index.
• The proposed hybrid modeling could improve the results by 60%.

This paper investigates a new drag model for the simulation of the fluidization of fluid catalytic cracking (FCC) particles with air in a fluidized bed using the two-fluid model (TFM) within the Multiphase Flow with Interphase Exchanges (MFIX) code. A cohesion index parameter based on the interparticle cohesive forces has been implemented in the MFIX-TFM code. This index is used as a switching criterion between a particle resolved drag model developed by Tenneti et al. (2011), and some of the drag models available in the MFIX for homogeneous particles, namely the Gidaspow, Syam–O'brien, and Wen–Yu models. The proposed drag correlation in this paper implements an indirect method of introducing interparticle cohesive forces to our TFM simulations. Significant improvement in the solid volume fraction profile along the riser was obtained for all of the drag law combinations, depending on the conditions set in the switching procedures. In the best case, the utilization of the Gidaspow and TGS models resulted in a 60% improvement in maximum deviation of numerical results from the available experimental data. The proposed model can be used in simulations of fluidized beds, where standard models fail to produce accurate results even on extremely refined computational grid, especially for Geldart A type particles that may exhibit strong clustering behavior.

Gas–solid flow simulations in a fluidized bed using the MFIX code show improved accuracy for the solid volume fraction along the riser, when a cohesion force indicator function, Ha, is used to switch the drag law in the two-fluid model between existing drag laws and a drag law that is based on direct numerical simulations.Figure optionsDownload as PowerPoint slide