Comparative CFD modeling of a bubbling bed using a Eulerian-Eulerian two-fluid model (TFM) and a Eulerian-Lagrangian dense discrete phase model (DDPM)

Eulerian-Eulerian and Eulerian-Lagrangian numerical approaches are both widely used to investigate hydrody-namic behavior in dense gas-solid fluidized bed reactors, yet there has been a lack of comparative investigations involving the two. Therefore, the present work compares the numerical performance of a two-fluid model (TFM) anda dense discrete phase model (DDPM) in describing the hydrodynamic behavior of a pilot-scale bubbling bed reactor. The effects of several different parameters - drag force, grid size, fluid time-step, time-averaging interval, particle-wall specularity coefficient, particle-particle restitution coefficient, particle-wall reflection coefficient, and numbers of parcels (the latter two parameters only for the DDPM) - are investigated and compared for the two approaches using two-dimensional (2D) simulations. The numerical results for both models indicate that sub-grid drag correction based on the energy minimization and multiscale (EMMS) theory is the most essen-tial modeling parameter to account for the multiscale structures (i.e., bubbles and void spaces) and to resolve the axial and radial solid distributions. Both the TFM and DDPM, coupled with the EMMS/bubbling drag, predict sim-ilar hydrodynamics behavior; however, the better accuracy in terms of axial and radial solids concentration pro -files is achieved from the TFM approach. Further, our grid size analysis results indicate that the DDPM generates a better grid-independent solution than the TFM. This important advantage of the DDPM over TFM makes it a suit-able candidate for large-scale industrial applications by employing it on much coarser grids. Meanwhile, the re-sults from the time-step, time-averaging interval, specularity coefficient, restitution coefficient, reflection coefficient, and numbers of parcels represent a minor effect on the overall hydrodynamics behavior of the pilot-scale bubbling bed reactor.

Automated summary

The study compared the performance of Eulerian-Eulerian and Eulerian-Lagrangian numerical approaches in describing the hydrodynamic behavior of a pilot-scale bubbling bed reactor. The sub-grid drag correction based on the energy minimization and multiscale theory was found to be the key modeling parameter for both models. Both models predicted similar hydrodynamics behavior, with TFM achieving better accuracy in terms of axial and radial solids concentration profiles. The grid size analysis showed that DDPM generated a better grid-independent solution than TFM, making it a suitable candidate for large-scale industrial applications on coarser grids. Other parameters like time-step, time-averaging interval, specularity coefficient, restitution coefficient, reflection coefficient, and numbers of parcels had minor effects on the overall hydrodynamics behavior of the reactor.