DPM Analysis of a gas-solid fluidised bed


The purpose of this numerical simulation is to validate the following parameters of incompressible laminar Discrete Phase Model (DPM) for a ‘Gas-Solid Fluidised Bed’:

  • Mixing distribution: by single bubble injection in a monodisperse fluidised bed.

The numerical simulation were carried out using the Discrete Phase Model (DPM) with Multiphase Particle-in-Cell method (MPPIC) [1] for collision of particles. Two cases were studied with different number of particles, ‘Case (b)’ with 30,000 prticles and ‘Case (c)’ with 57,000 particles. The results of simulations performed by SimScale were compared to the experimental results shown in [2].

Import validation project into workspace


The geometry of the study is a rectangular channel (Bed). The pulse jet is injected from the bottom face patch highlighted in red color (see Fig.1.). A brief description of the dimensions is provided by the table below.


Fig.1. Geometry of the Fluidised Bed

  Width (x) Height (z) Depth (y) Jet width
Value [m] \(0.15\) \(0.45\) \(0.015\) \(0.01\)

Domain and Analysis type

The domain is the internal region of the fluidised bed with the domain extents same as the geometrical dimensions. For this study a hexahedral mesh was created using the opensource “block Mesh” tool (see Fig.2.). The grid points are equally spaced in each direction using the simple grading value of 1. The details of the mesh are listed in the following table:

Mesh and Element types :

Number of cells in X Number of cells in Y Number of cells in Z Type
30 1 45 hexahedral

Fig.2. Fluidised Bed mesh used for the SimScale case

The numerical analysis performed is detailed as follows:

Tool Type : OPENFOAM®

Analysis Type : Transient Laminar Discrete Phase Model

Solver : Multiphase Particle-in-Cell (MP-PIC) method for collisional exchange

Particle Drag Model : Ergun-Wen-Yu

Turbulence Model : Laminar

Simulation Setup

Fluid (gas): Air
Kinematic viscosity (\(\nu\)) \(= 1.56e{-5} m^2/s\)
Density (\(\rho\)) \(= 1.2 kg/m^3\)

Solid (particle): Glass
Density (\(\rho\)) \(= 2526 kg/m^3\)
Diameter \(= 2.5 mm\)
Number of particles [-] \(= 30,000\ (case\ b)\ - \ 57,000\ (case\ c)\)

Boundary Conditions:

For the bottom face boundary, an inlet background velocity was applied along with a pulse jet inlet at the mid patch, while a pressure boundary condition was applied at the outlet. While for the side faces No-Slip wall condition was applied. The following table provides the further details.

Boundary type Velocity Pressure
Inlet Interstitial Inlet velocity: \(1.7\ ms^{-1}\) Fixed flux pressure
Pulse Jet Inlet Fixed value velocity via Table input: \(20\ ms^{-1}\ (for 150 ms)\) Fixed flux pressure
Wall no-slip Fixed Value: \(0.0\ ms^{-1}\) Fixed flux pressure
Outlet Pressure-inlet-outlet-velocity Fixed value: \(0\ Pa\)


The numerical simulation results for the particle mixing are compared with experimental results provided by G.A. Bokkers, M. van Sint Annaland, J.A.M. Kuipers [2].

A comparison of the particle mixing distribution obtained with SimScale and experimental results is given in Fig.3A-C. The figures show the distribution at three different times of 150, 200, 800 ms after the injection of jet pulse*. The particles were initially ordered in two seperate colored layers for visualization of the extent of mixing due to the jet (bubble). For the case (c) (with 57,000 particles), 2 runs were performed, one in parallel other in serial. This was done to get the right post-processing output files from OPENFOAM® which were from the serial run.

From Fig. 3A-B it is observable that the current utilized Ergun-Wen-Yu drag model predicts the particle distribution for the bubble phase sufficiently well, while for a later time of 800 ms the model over-predicts the particle mixing as shown in Fig. 3C. Better results may be attained by the use of other drag models, such as the ‘Koch & Hill’ drag model which has been demonstrated by G.A. Bokkers [2].


Fig.3. Comparison of experiment [2] with Simscale results for single bubble injection in fluidised bed (monodispersed)

The comparison of velocity vectors between experiment and simulation results at 150ms after injection is shown in the Fig.4 below.


Fig.4. Comparison of experiment [2] with Simscale results for velocity vectors at 150ms after injection. The legend vector corresponds to a particle velocity of 1m/s.

A visualization of the particle mixing evolution and alpha-continuous phase for simulation results is depicted by Fig.5A-B below.


Fig.5. Discrete phase model simulation results for single bubble injection in fluidised bed for Ergun-Wen-Yu drag model, A) particle mixing, B) alpha-continuous phase.


[1]A model for collisional exchange in gas/liquid/solid fluidized beds. P. J. O’Rourkea, P. Zhaob, D. Sniderb
[2](1, 2, 3, 4, 5) Mixing and segregation in a bidisperse gas–solid fluidised bed: A numerical and experimental study. G.A. Bokkers, M. van Sint Annaland, J.A.M. Kuipers


This offering is not approved or endorsed by OpenCFD Limited, producer and distributor of the OpenFOAM software and owner of the OPENFOAM® and OpenCFD® trade marks. OPENFOAM® is a registered trade mark of OpenCFD Limited, producer and distributor of the OpenFOAM software.