CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   OpenFOAM Running, Solving & CFD (https://www.cfd-online.com/Forums/openfoam-solving/)
-   -   compressible flow calculation error using rhoSimpleFoam solver (https://www.cfd-online.com/Forums/openfoam-solving/71898-compressible-flow-calculation-error-using-rhosimplefoam-solver.html)

student4326 January 19, 2010 09:19

compressible flow calculation error using rhoSimpleFoam solver
 
4 Attachment(s)
Hi,

I'm currently experimenting with the rhoSimpleFoam solver (OpenFOAM Version 1.6) for steady-state calculations. The case is a bended tube (diameter 0.047 m) with the following boundary conditions:
Inlet: flowRateInletVelocity (mass flow rate) 0.1 kg/s, T = 874 K
Walls: T = 300 K
Outlet: atmospheric pressure (p = 101325 Pa)
kEpsilon tubulence-model activated.

Please see attached files for how I set up the boundary conditions with OpenFoam.

Calculations always stop after 5-8 timesteps. I monitor an instant rise of pressure/rho.

I've tried the following modifications though without any success:

Underrelaxing rho 0.05 -> 1.0
Non Ortogonal Corrector: 0 -> 5
nCellsInCoarsestLevel 10 -> 20
Smoother: GaussSeidel -> DICGaussSeidel
Turn off of turbulent flow (kEpsilon: laminar)

When investigating the ultimate timestep I found some cells with a maximum velocity of over 15000 m/s. This would indicate a supersonic flow for which other users recommend the "Sonic" solvers.
More oddly I discouvered cells with a pressure of 1.0e07 - 1.0e09 Pa situated at the pressure outlet. It almost looks like the flow can't exit the tube.

In the next step I've reduced the mass flow by factor 10 to 0.01 kg/s and the inlet temperatur to T = 300 K (these seem to be standard values in the OpenFOAM tutorials were the usage of BC "flowRateInletVelocity" is explained -> solvers rhoPorousFoam and rhoPimpleFoam). By doing this I received a reasonably result which compares quite well with a Fluent calculation.

Next I increased the inlet temperatur. This caused higher internal flow rates (detected during postprocessing):
T_in = 473 K -> m_dot = 0.0166 kg/s
T_in = 673 K -> m_dot = 0.017 kg/s
T_in = 873 K -> m_dot = 0.0307 kg/s
To me this behaviour appears to be highly questionable as I suppose the flow rate to be constant!

Finally I played around with other compressible solver examples from the tutorial (rhoPimpleFoam and rhoPorousFoam solvers). In both solvers, an increase of the default mass flow rate from 0.01 kg/s to 0.1 kg/s resulted in an calculation error.

I would be glad to know if the boundary conditions are set up correctly.

Has anyone experience in handling compressible flow solvers, perhaps with similar BC? I would very much appreciate comments/suggestions regarding this case or/and solver.

Thanks for your help,
Sven

student4326 February 1, 2010 10:07

2 Attachment(s)
Hey,

by modifying the underrel. factors

rho 0.05 -> 0.9
p 0.3 -> 0.01

I received a convergent solution.

The odd behaviour of the internal flow rates (different inlet temp. -> different flow rates) I've detected during postprocessing with Fluent is solved as well. The internal flow rates are in fact constant! The files "phi" in the time directories returned correct flow rates on inlet and outlet. Unfortunately the Fluent report function (report -> result reports... -> surface integral -> Report type: massflow rate) didn't work correctly.

When comparing the OF with a Fluent calculation the temperature and velocity mag. results show a rather big discrepance between the Fluent and OF values. Therefore I've implemented an adiabatic wall (temperature BC for wall: zeroGradient) in both calculations in order to investigate the source of the velocity discrepancy.

Right now Fluent calculates velocities from 0 to 78 m/s. OF results range from 0 to 240 m/s (please see attachments). I hope someone can give me a hint to get to the source of this problem.

student4326 February 3, 2010 03:55

Problem solved! Discrepancy due to incorrect material setup Fluent.

rock.senthilkumar December 28, 2010 01:50

Hi student ,
With reference to your post on "January 19, 2010, 18:49 " in which you quoted as ".. indicate a supersonic flow for which other users recommend the "Sonic" solvers." Kindly clear whether for supersonic flow rhoSimpleFoam (according to me it is best) is suitable or not?. Can you sugeest any "Sonic" solver for steady state supersonic flow?.Furthermore, If possible upload your constant directory with polymesh subdirectory for the above case for which uploaded 0,system,constant directory or any other case file for rhoSimpleFoam.

Quote:

Originally Posted by student4326 (Post 243071)
Hi,

I'm currently experimenting with the rhoSimpleFoam solver (OpenFOAM Version 1.6) for steady-state calculations. The case is a bended tube (diameter 0.047 m) with the following boundary conditions:
Inlet: flowRateInletVelocity (mass flow rate) 0.1 kg/s, T = 874 K
Walls: T = 300 K
Outlet: atmospheric pressure (p = 101325 Pa)
kEpsilon tubulence-model activated.

Please see attached files for how I set up the boundary conditions with OpenFoam.

Calculations always stop after 5-8 timesteps. I monitor an instant rise of pressure/rho.

I've tried the following modifications though without any success:

Underrelaxing rho 0.05 -> 1.0
Non Ortogonal Corrector: 0 -> 5
nCellsInCoarsestLevel 10 -> 20
Smoother: GaussSeidel -> DICGaussSeidel
Turn off of turbulent flow (kEpsilon: laminar)

When investigating the ultimate timestep I found some cells with a maximum velocity of over 15000 m/s. This would indicate a supersonic flow for which other users recommend the "Sonic" solvers.
More oddly I discouvered cells with a pressure of 1.0e07 - 1.0e09 Pa situated at the pressure outlet. It almost looks like the flow can't exit the tube.

In the next step I've reduced the mass flow by factor 10 to 0.01 kg/s and the inlet temperatur to T = 300 K (these seem to be standard values in the OpenFOAM tutorials were the usage of BC "flowRateInletVelocity" is explained -> solvers rhoPorousFoam and rhoPimpleFoam). By doing this I received a reasonably result which compares quite well with a Fluent calculation.

Next I increased the inlet temperatur. This caused higher internal flow rates (detected during postprocessing):
T_in = 473 K -> m_dot = 0.0166 kg/s
T_in = 673 K -> m_dot = 0.017 kg/s
T_in = 873 K -> m_dot = 0.0307 kg/s
To me this behaviour appears to be highly questionable as I suppose the flow rate to be constant!

Finally I played around with other compressible solver examples from the tutorial (rhoPimpleFoam and rhoPorousFoam solvers). In both solvers, an increase of the default mass flow rate from 0.01 kg/s to 0.1 kg/s resulted in an calculation error.

I would be glad to know if the boundary conditions are set up correctly.

Has anyone experience in handling compressible flow solvers, perhaps with similar BC? I would very much appreciate comments/suggestions regarding this case or/and solver.

Thanks for your help,
Sven


msshah February 10, 2011 12:13

rhoSimpleFoam
 
Hey, I am also trying to solve steady, turbulent, compressible flow through a pipe using rhoSimpleFoam in OpenFOAM 1.6

I would appreciate if some body can send me a complete working tutorial.

Thanks,

Manish,
Mumbai.

Abhinay Kulkarni February 9, 2012 13:31

Hi Manish,

Did you get the tutorial??or did you manage to solve the case??

Can you please reply if you have solved the case?? I am also trying to solve a similar case using rhoSimplecfoam.

Would be great if i could get some help.

Regards
Abhinay

Abhinay Kulkarni February 10, 2012 10:36

Hi Sven,

I am also trying out a similar case using rhoSimplecFoam. My geometry is not a tube exactly as the inlet and outlet dia are not the same as the tube dia. I am trying to simulate a steady, compressible flow through this geometry and would be great if i get some help.

For a start Can you kindly attach your boundary file so that i get a better understanding of your case??

Regards
Abhinay

Garfield November 2, 2015 12:34

Hi, Abhinay
How about your rhoSimplecFoam calculation, I am currently working on the simulation using rhoSimplecFoam, however, I keep meeting problems. Could you kindly attach one of your successful case so that I can study that?
Thanks a lot!
Best
Garfield


All times are GMT -4. The time now is 08:19.