twoPhaseEulerFoam (very) high pressure when using kinetic theory
 

Hello everybody,

I'm basically simulating a slurry flow with solid particles in a liquid carrier. I managed to run simulation with different turbulent models for the liquid carrier (basically k-w SST and std. k-epsilon) getting reasonable velocity profiles and pressure drops.

In most of the papers I'm trying to emulate (i.e. CFD modeling for pipeline flow of fine particles at high concentration - Kaushal et al.) they use kinetic-theory for granular flow.

I tried to enable it but I get a practically laminar profile for the solid phase.

Here an image that compares the two velocity profiles with same volume fraction (13%), inlet velocity (1.5), particle diameter (1.3mm), k-w for carrier and in the red profile I used the kinetic-theory for granular flows with equilibrium hypothesis.

GEOMETRY DETAILS
http://i.imgur.com/NZAcxWXl.png
Pipe Diameter: 15e-3 [m]
Pipe Length: 0.6 [m]
Gravity: [0 0 -9.81]
Position of the sample line: 0.4 [m] from the inlet
Boundary Conditions:

• inlet:
  • U.carrier: 1.5 [m/s]
  • U.dispersed: 1.5 [m/s]
  • k: Intensity 0.005 U.carrier (turbulentIntensityKineticEnergyInlet)
  • omega.carrier: FixedValue 1
  • alpha.carrier = FixedValue 0.13
  • p = calculated
  • p_rgh = zeroGradient
  • Theta.dispersed = FixedValue 0
  • nut for both: calculated

• outlet:
  • U.carrier: zeroGradient
  • U.dispersed: zeroGradient
  • k: zeroGradient
  • omega.carrier: FixedValue 1
  • alpha.carrier = FixedValue 0.13
  • p = calculated
  • p_rgh = zeroGradient
  • Theta.dispersed = FixedValue 0
  • nut for both: calculated
• walls
  • U.* = no-slip
  • k = LowReWallFunction/Std Wall Function (depends from the mesh)
  • omega = Wall Function
  • p = calculated
  • p_rgh = zeroGradient
  • Theta = zeroGradient
  • nut: calculated
• symmetryPlane:
  • symmetry for all

PHASES INFORMATIONS
Liquid Carrier Density: 870 [Kg/m³]
Liquid Carrier Viscosity: 0.292E-3 [Pa s]
Dispersed Phase Density: 1026 [Kg/m³]
Dispersed Phase Viscosity: 0.292E-3 [Pa s]

(Case naming system: 15_13_kw_keq:

• 15 = 1.5 m/s inlet velocity
• 13 = 1.3mm particle diameter
• kw = k-omega sst turbulence for carrier
• keq = kinetic theory with equilibrium )

Seems to be related to viscosity provided by the model: it's basically a laminar profile so I would say that there is not enough viscosity.



An help would be very appreciated, thanks...

OpenFOAM 3.0.1



Here is my case too:

Attachment 47417

http://i.imgur.com/8pnXjoKl.png

Hello Ruben (tidusuper91)



I am simulating slurry flow by twophaseeulerfoam, and I study your case. I am surprised because of your refine mesh (very small). I read that "The cell size is chosen to be roughly about 30 times the particle diameter of the solid phase, and If a finer mesh should be used, it is important that the size of the control volumes is not smaller than the diameter of the particle. In a control volume smaller than the particle diameter, the particle would fill the entire control volume (αp = 1) and not allow for e.g., particle-particle interactions.", but the minimum volume size of your mesh is smaller than particle volume. What is your idea about the size of the mesh when we use twophaseeulerfoam for slurry flow?
I know that your question was in 2016, but could you tell me how did your problem solve?


Sincerely yours

Hello Masoud,



it was indeed long time ago. Where did you read that? Can you provide a reference? :)

Hello Ruben
Thank you for your kind response . I read a tutorial about twophaseeulerfoam at Chalmers university was written by Busch:

https://www.google.com/url?sa=t&rct=...j7gVliLiYwiVsf

It has been a while that I haven't read a well prepared question such as this one in this forum.

Thanks.

Well, as I said, it was long time ago. But I don't think there's a limitation on that ratio numerically speaking. The question would then be if the results are physically sound. I was able, at the end, to extract useful profiles from this configuration.

This question was related to my M.Sc. thesis that you can find here. I hope that would help you! :)

Ahahah, thanks. It was long time ago and... yet nobody answered. :)

Hello Ruben
Thank you for your response.

Hi Masoud, did you simulate slurry flow without problem? I am also doing this research.

Thanks for your post. I found in Fvsolution file, you used under relaxation factor. However, during the simulation, it was neglected no matter how you changed the coefficient. Also, the issue of time step is significant.