A Novel Approach to Simulate Large-Scale Fluidized Beds By CFD-DEM

Coupling of Computational Fluid Dynamics and Discrete Element Method (CFD-DEM) is a recently developed approach for the modelling of particulate flows, such as fluidized bed, pneumatic conveying and cyclones. However, this approach suffers from several shortcomings, including the need for highly calibrated drag laws of particles (and swarms), the modeling of turbulence dampening/generation by particles and the need for collisional models that account accurately for enduring and short contacts. Moreover, the number of simulated particles is restricted due to the extremely high computational time. Consequently, in this study, we attempt to analyze a methodology to reduce the computational cost by replacing the original particles with larger particles (decreasing the particle number) while the characteristic dimensionless groups are kept constant for both scaled and base cases. In other words, replacing original particles with larger particles enables us to decrease the computational cost while having the similar hydrodynamic behavior for both base and scaled cases. The dimensionless numbers (groups) which are kept similar in both base and scaled simulations are particle Reynolds number (Re), particle Froude number (Fr), particle to fluid density ratio, bed-to-particle diameter ratio and bed-width to initial bed-height ratio. The dynamic behavior of a fluidized bed for both, base case and scaled case, is also analyzed to determine the significance of each dimensionless number.

  

In the current study, our high-performance GPU code XPS coupled with an CFD code is used. The adequacy of the proposed methodology is tested on a pseudo 2D single-spout fluidized bed. The setup under investigation was previously used to validate the coupled CFD-DEM code [1]. To check the hydrodynamic similarity the instantaneous solid volume fraction, instantaneous gas phase velocity and time-averaged particle velocity at different locations of the fluidized beds are compared between the base and scaled cases. A typical comparison of the instantaneous gas velocities between base and scaled cases is presented in Figure 1.

The simulation results showed that similar dynamics can be obtained for both base and scaled cases, proving that the proposed method may be applied to simulate the large-scale fluidized beds with a reasonable computational time.

Figure 1: Comparison of the instantaneous gas phase velocities between base case and scaled case (a) y=3D/4, (b) y=D, (c) x=D/2, (d) x=D/4.

Reference:

[1] D. Jajcevic, E. Siegmann, C. Radeke, J.G. Khinast, Large-scale CFD–DEM simulations of fluidized granular systems, Chemical Engineering Science, 98 (2013) 298-310.