|
[Sponsors] |
March 27, 2006, 11:25 |
Hi,
i am trying to run a ca
|
#1 |
New Member
Join Date: Mar 2009
Location: Wuppertal, Germany
Posts: 7
Rep Power: 17 |
Hi,
i am trying to run a calculation with buoyantFoam related to a simple plume simulation for an axisymmetric problem. The boundary conditions that i have selected are empty(Symmetry-Axis), Inlet (fixed Values for Temperature and velocity) and pressure outlet (for the side and the top of the computational domain). The initial velocity Fields for the internal Mesh are set to be zero. The calculation does not converge (nan for the Courandt Numbers and the convergence parameters). Now i would to ask you: 1.) handles the buoyantFoam solver compressible Flows (variabel density case and low mach number) or the bousinesq approximation? 2.) Because i have trying various combinations of discretization schemes, could anybody gives to me informations abbout suggested initial Field values? |
|
March 27, 2006, 12:03 |
One thing I can tell you about
|
#2 |
Senior Member
Eugene de Villiers
Join Date: Mar 2009
Posts: 725
Rep Power: 21 |
One thing I can tell you about buoyant calculations from painfull personal experience: make 100% sure all your wall and inlet pressure boundaries are defined as "wallBuoyantPressure" and not "zeroGradient" otherwise you will never converge.
|
|
March 27, 2006, 12:41 |
i have now select the fixedTem
|
#3 |
New Member
Join Date: Mar 2009
Location: Wuppertal, Germany
Posts: 7
Rep Power: 17 |
i have now select the fixedTemperatureWall option with fixed velocity (as this is the case in the inlet boundary) for the previously as inlet defined region. This allows for to define a "wallBuoyantPressure" but the code writes an errormessage out with the suggestion to use "zeroGradient" for kinetic Energy k in the same boundary.
What can i do? |
|
March 27, 2006, 12:44 |
Submit a bug report and then f
|
#4 |
Senior Member
Eugene de Villiers
Join Date: Mar 2009
Posts: 725
Rep Power: 21 |
Submit a bug report and then fire up your favourite text editor and change the entries manually.
|
|
March 27, 2006, 12:48 |
Thanks Eugene
i will follow
|
#5 |
New Member
Join Date: Mar 2009
Location: Wuppertal, Germany
Posts: 7
Rep Power: 17 |
Thanks Eugene
i will follow your suggestions. |
|
March 27, 2006, 13:07 |
i have edit manually the k-sou
|
#6 |
New Member
Join Date: Mar 2009
Location: Wuppertal, Germany
Posts: 7
Rep Power: 17 |
i have edit manually the k-source-file in the 0-directory but now the calculation does not converge.
Nothing is so simple as like it looks before. |
|
March 27, 2006, 13:35 |
why did you edit the k file? B
|
#7 |
Senior Member
Eugene de Villiers
Join Date: Mar 2009
Posts: 725
Rep Power: 21 |
why did you edit the k file? Buoyancy has no effect on k and epsilon BCs. You should use fixedValue for k at the inlet.
You need wallBuoyantPressure for p at all boundaries where it is usually zeroGradient. |
|
March 28, 2006, 06:11 |
now i have edit the pressure b
|
#8 |
New Member
Join Date: Mar 2009
Location: Wuppertal, Germany
Posts: 7
Rep Power: 17 |
now i have edit the pressure boundaries in the p-file (Directory 0), so the zeroGradient condition becomes to wallBuoyantPressure and the fixed Value conditions (outlets at top and side) remains as is.
But even the Courandt Numbers are inacceptable and the flow behaves like inviscid Fluid where the inlet velocity remains constant throughout the computational domain and no turbulence effects affects the flow in the sense of developed eddy dissipation and velocity componets coupling. The FluidMixture is air as prescribed in the code, so i dont thing that the thermophysical properties are poorely chosen. |
|
March 28, 2006, 08:16 |
So what is your inlet value fo
|
#9 |
Senior Member
Eugene de Villiers
Join Date: Mar 2009
Posts: 725
Rep Power: 21 |
So what is your inlet value for k and epsilon?
Is the turbulence solver doing anything? |
|
March 28, 2006, 09:19 |
i make various gues for k like
|
#10 |
New Member
Join Date: Mar 2009
Location: Wuppertal, Germany
Posts: 7
Rep Power: 17 |
i make various gues for k like:
k=3/2 (u_fluct)**2 last value k=0.024 but these makes the convergence to be more defecault. The convergence monitor shows only details for the velocity components and the variable pb. |
|
March 28, 2006, 13:21 |
Well if it isnt showing conver
|
#11 |
Senior Member
Eugene de Villiers
Join Date: Mar 2009
Posts: 725
Rep Power: 21 |
Well if it isnt showing convergence info for k and epsilon, you are not solving for turbulence. Check your constant/turbulenceProperties dictionary. turbulence should be "on" and the turbulence model should be set to something other than laminar.
|
|
March 29, 2006, 06:52 |
The turbulence modell is set t
|
#12 |
New Member
Join Date: Mar 2009
Location: Wuppertal, Germany
Posts: 7
Rep Power: 17 |
The turbulence modell is set to kEpsilon and the turbulence is on.
In the solution outputs (t-Directories) are files with information about these properties but the values shows not changes. In FoamX are the related inputs also selected! |
|
Thread Tools | Search this Thread |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Buoyant plumes | Annette Nilsson | FLUENT | 0 | May 28, 2001 09:18 |
buoyant jet | appleyy | FLUENT | 3 | March 29, 2001 01:49 |
buoyant jet | Qi Yuan | FLUENT | 0 | February 13, 2001 06:49 |
LES for buoyant flows? | George Bergantz | Phoenics | 0 | December 18, 2000 11:41 |
Rocket Plumes | Kurt Motekew | Main CFD Forum | 2 | April 22, 1999 09:44 |