Investigation of Flow Regimes in
Trickle Bed Reactors using Volume of Fluid and Lattice Boltzmann Methods

Humair Nadeem¹, Amina Islam¹, Imen Ben Salem¹, Sylvie Chevalier¹, Titly
Farhana Faisal¹, Mohamed Sassi¹

¹Masdar Institute of
Science and Technology, Abu Dhabi, UAE

Trickle
Bed Reactors (TBRs) are a group of three-phase reactors that allow the flow of
liquid and gas phase over a bed of solid catalyst particles. Due to their
significance in industrial applications, modeling and simulation of TBRs are
very important to researches as it would allow the use of rapidly growing
computational facilities to optimize processes.

Since many of the studies in
literature have focused on traditional CFD techniques such Finite Volume
Discretization, in this work we attempt to run the simulations using both the
Finite Volume Discretization method and the Lattice Boltzmann Method (LBM) and
then compare the results to study the different flow regimes of a TBR; trickle
flow, pulse flow, bubble flow and spray flow. The LBM is a meshless
technique that has received a lot of attention during the past decade due to
its image-based input, which allows for the accurate representation of porous
media through which flow occurs. Since one of the main limitations of traditional
CFD techniques is how the porosity value of the catalyst bed are distributed
within the model, an accurate multiphase LBM model would allow for hydrodynamic
studies to be done directly on images of the solid catalyst bed since transport
processes are controlled by particle size, shape and packing method among other
parameters. In addition to that, the CFD method becomes very computationally
intensive for 3D simulations as a result of using the mesh. Validating LBM
results with CFD for 2D simulations would help run 3D simulations for TBR as
the LBM is meshless.  

In the finite volume
computational study, the VoF method is chosen as it is a surface-tracking
method applied to a fixed Eulerian mesh when the locus of the interface between
two or more immiscible fluids of interest. In this method, a single set of
momentum equation is shared by the fluids, and the volume fraction of each
fluid in each computational cell is tracked throughout the domain.

Gas and liquid are injected
from the concentric injectors placed at the top of TBR as shown in the
figure. The computational domains are created, meshed and labeled in
Gambit 2.3.16. The governing equations together with the boundary conditions
are solved using the finite volume CFD code, Fluent 13.0.0. The equations were
solved with the Semi-Implicit Pressure-Linked Equation (SIMPLE) algorithm.
Discretization scheme are PRESTO for pressure, modified HRIC for volume
fraction and first order upwind for momentum and turbulence model.

In addition to that, the LBM method is used to simulate multiphase
flow in a similar 2D model in order to compare with the Finite Volume model. In
the LBM technique, fluid particles are confined to the nodes of the lattice and
the particle velocities are restricted to a finite set of orientations. Fluid
flow is described by a set of distribution functions and it undergoes two basic
steps; streaming and collision. The Bhatnagar-Gross-Krook collision operator is
used and it leads the system to the local Maxwellian equilibrium and then in
order to introduce the nonlocal interaction among particles, the multiphase
Shan Chen model is adopted. Local velocity
and density are obtained by appropriate summation of the distribution
functions. 

A comparison is done between the results from the
two computational methods in order to study the different flow regimes of the
TBR.

1.