AbstractIn this work the influences of solid viscosity and the way to scale-down traditional drag models on the predicted hydrodynamics of Geldart A particles in a lab-scale gas-solid bubbling fluidized bed are investigated. To evaluate the effects of drag models, the modified Gibilaro et al. drag model (constant correction factor) and the EMMS drag model (non-constant correction factor) are tested. And the influences of solid viscosity are assessed by considering the empirical model proposed by Gidaspow et al. (1997, Turbulence, Viscosity and Numerical Simulation of FCC Particles in CFB. Fluidization and Fluid-particle Systems, AIChE Annual Meeting, Los Angeles, 58–62) and the models based on kinetic theory of granular flow (KTGF) with or without frictional stress. The resulting hydrodynamics by incorporating the different combinations of the drag model and solid viscosity model into two-fluid model (TFM) simulations are compared with the experimental data of Zhu et al. (2008, Detailed Measurements of Flow Structure inside a Dense Gas-Solids Fluidized Bed.”Powder Technological180:339–349). The simulation results show that the predicted hydrodynamics closely depends on the setting of solid viscosity. When solid viscosity is calculated from the empirical model of Gidaspow et al., both drag models can reasonably predict the radial solid concentration profiles and particle velocity profiles. When the KTGF viscosity model without frictional stress is adopted, the EMMS drag model significantly over-estimates the bed expansion, whereas the modified Gibilaro et al. drag model can still give acceptable radial solid concentration profiles but over-estimate particle upwards and downwards velocity. When KTGF viscosity model with frictional stress is chosen, both drag models predict the occurrence of slugging. At this time, the particle velocity profiles predicted by EMMS drag model are still in well agreement with the experimental data, but the bed expansion is under-estimated.
Performance improvement of existing drag models in two-fluid modeling of gas–solid flows using a PR-DNS based drag model