CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Multiphase Convergence Issues (http://www.cfd-online.com/Forums/fluent/103187-multiphase-convergence-issues.html)

RodriguezFatz June 13, 2012 03:37

Multiphase Convergence Issues
 
Hi all,

I'm currently trying to simulate a multiphase gas flow through a pipe (elbow shape with some pipe in front of and some pipe behind the elbow).
I use the mixture model with air (primary phase) and 5% granular secondary phase of some heavy material. Turbulence model is DES with SA-model, but my problems occur with every turbulence model, even in steady state calculations.

When I create an "erroneous" grid with much too large cells in the boundary layer (y+ up to a few hundret), the simulation converges (both transient and steady). Now, when I decrease the cell sizes in the boundary, to get a correct boundary behavior by the turbulence model, all simulations diverge very quickly within the first timesteps. For very small under-realxation factors or time steps, I can get some more stable time steps, but at some point the residuals increase.

Basically grids with false boundary mesh work and those with correct bondary meshes don't.
Same grid with just one phase works all the time!

Can anyone help me with this?

RodriguezFatz June 13, 2012 07:51

Ok... I kept trying, now I found out that it seems like the granular particle size did the trick. When I change that size to much smaller values (e.g. 1.0e-7) it works for fine grids. Seems like the model / equations become nonsense, when certain cells are smaller than the particle size? I can't really see a problem here, can anyone explain?

teethfish March 11, 2013 01:00

I met the same problem but it's DES+cavitation. But opposite to yours, when I increase the bubble diameters from 1e-6 to 1e-5, it converges better. Still not very well, thus I think think that when use DES+multiphase models, should be some special consideration.


Quote:

Originally Posted by RodriguezFatz (Post 366219)
Ok... I kept trying, now I found out that it seems like the granular particle size did the trick. When I change that size to much smaller values (e.g. 1.0e-7) it works for fine grids. Seems like the model / equations become nonsense, when certain cells are smaller than the particle size? I can't really see a problem here, can anyone explain?



All times are GMT -4. The time now is 15:40.