The MP-PIC Method for CFD-Simulation of Fluidized Beds - Comparison of Two Different Implementations

Fluidized beds are very common in the industry, for example in combustion or gasification processes. Due to the large dimensions, the high number of particles and the reaction kinetics of industrial-scale combustors the simulation is commonly too computationally expensive to resolve every particle interaction. To allow the simulation of industrial plants with reasonable computational time the multi-phase – particle in cell (MP-PIC) method is applied. This method results from a modification of the Euler-Lagrange approach, where the fluid is described by the Navier-Stokes equations on a Eulerian grid in combination with Lagrangian particles. To reduce the computational costs a defined number of particles with the same properties is represented by so-called parcels. This facilitates the consideration of various effects like particle size distributions as well as shrinkage effects and reduces the computational costs compared to the discrete particle model (DPM).

In this study two different implementation of the MP-PIC method in Barracuda VR® and OpenFOAM® are compared with experimental results of a lab-scale fluidized bed reactor to validate the applicability of the MP-PIC method. The fluidization velocities are set in the range from 0.2 – 1.3 m/s with a reactor diameter of 0.1 m. The bed height of quartz sand with a sauter mean diameter of 220 µm is 0.1 m. For the comparison between simulation results and experimental data video recordings as well as frequency analysis are used. To evaluate the simulation results radial and axial profiles of the particle volume fraction are compared to literature.

Preliminary results indicated that both implementations of the MP-PIC method could predict the fluidization behavior at lower fluidization velocities with reasonable agreement, while the simulations fail to predict the impact of higher fluidization velocities. This is presumably due to the application of the homogeneous Gidaspow drag model, which neglects the reduced drag force by the formation of clusters and therefore overestimates the momentum exchange between the fluid and the particle phase. To allow for more accurate simulations a drag model based on the promising energy minimization multi-scale (EMMS) theory will be applied.

We gratefully acknowledge for the financial support the German Research Foundation (DFG) (Germany). Project number HE 4526/21-1.