[FabianB]

Dear Foamers,

I am trying to simulate a gas-solid fluidized bed based on the paper by Müller et al. with twoPhaseEulerFoam.

http://ac.els-cdn.com/S0032591007006...c76b438be521b3

I've used the twoPhaseEulerFoam/laminar/fluidizedbed tutorial as basis of the simulation and changed all the following parameters according to this paper:

• Bed size (pseudo two-dimensional bed, 44x 10 x 120 mm )
• Particle-diameter and density ( =1.2 mm =1000 kg m)
• Temperature of air and particles to uniform 298.15 K
• Air viscosity to 1.8e-5 Pa s

As boundary condition for the particles I started with a JohnsonJacksonParticleSlip ( specularityCoefficient 0.9 ) for U.particles and JohnsonJacksonParticleTheta (specularityCoefficient 0.9, restitutionCoefficient 0.97) for Theta.particles. At the inlet of the air I used a fixed Value with 0.6 m/s and changend the internalField to uniform 0. With the setFieldsDict I set the alpha.particles-values to 0.6 to a height of 32 mm (calculated from the given number of particles) and alphamax to 0.62 . The initial drag-coefficient was GidaspowErgunWenYu. And because I'm only interested in an isothermal solution I've set minIter and maxIter to 0 at (h|e) in fvSolution.

Now, the problem is, that the results from this simulation differ a lot from the experimental and simulated results as reported in the paper. When comparing the solids volume fraction for an air velocity of 0.9 m/s you can clearly see the formation of bubbles in Müller's simulation (with DEM) whereas with twoPhaseEulerFoam you can only see homogeneous expansion ( I've attached a snapshot). Also, when looking at the Particle mean-velocity in y-direction for an air-velocity of 0.6 m/s you can see that the particle velocity with twoPhaseEulerFoam is underestimated by the factor 1000.
In order to get better results I've tried the follwoing things:

• Changed the Particle Drag-Law (Syamal O'Brien, Schiller Naumann)
• Changed the simulation-type for air from lamiar to RAS
• Made simulations with different number of cells (increased the fineness)
• Changed the values for the friction and specularity coefficient

But, with all these changes the outcome is in general the same as described.
In the paper, the Poisson's ratio, coefficient of normal restitution and coefficient of friction is given for the particles. I was thinking that these values for the particles might have a large influence on the results, but couldn't find anything where to change these values within twoPhaseEulerFoam.
Therefore, my question is the following:
Is it possible the define these particle parameters ? (apart from diameter and density) And is there a possibility to define particle-particle interaction?

The "particles" in Müller's simulation are poppy-seeds. So far, I've found many simulations (with Euler-Euler approach) where the simulations give reasonable results, but in all of them glass beads are the "particles". To me, it seems that this approach does not work for poppy seeds without changing the source-code.
Maybe some of you have tried something similar and could give me a hint if I've forgotten some important changes. Thanks in advance!