CFD Online Discussion Forums

CFD Online Discussion Forums (
-   OpenFOAM Running, Solving & CFD (
-   -   Energy based steady state solver. (

atareen64 April 24, 2011 22:39

Energy based steady state solver.
Hello world!

I was using rhoSimpleFoam to simulate fluid (air) through a rectangular tube: inlet on one side and outlet on the other side, higher pressure on the inlet and lower pressure on the outlet. After running rhoSimpleFoam, my velocity, pressure and temperature profile look great!

The problem however is with the magnitude of U and Pressure: I am getting velocities of 1000+ m/s and pressures of 65 thousand pascals, clearly non physical results for my simple geometry (pressure at boundary conditions isn't high enough to reach these velocities).

The problem is with the thermoPhysicalProperties dictionary, the default thermo package for rhoSimpleFoam for air was

thermoType hPsiThermo<pureMixture<constTransport<specieThermo <hConstThermo<perfectGas>>>>>;

mixture air 1 28.966 1006.43 1.41e5 17.894e-06 0.720;

This works, the profiles look right but the numbers are completely are incorrect.

So I changed the rhoSimpleFoam solver so it would be able to use the following package

thermoType ePsiThermo<pureMixture<constTransport<specieThermo <eConstThermo<perfectGas>>>>>;

mixture air 1 28.9 717.5 10 1.82e-05 0.73;

The solver compiled successfully but now it blows up after 3 iterations.


how can I use rhoSimpleFoam to simulate air without getting crazy numbers OR
why does replacing enthalpy with energy in rhoSimpleFoam make it unstable? How can I fix this.

Thank you so much for reading this or trying to help!


atareen64 April 25, 2011 12:04

3 Attachment(s)
I am attaching my test case, which is set up to be run with 'rhoSimpleFoam'. Like I said I am getting unusual numbers for U and P. With a little bit of research, I found out that chagning the relaxation factors in the fvSolution file seem to affect the results quite a lot. I keep changing them and getting different results with no way of knowing what values are correct. Can somebody please run this case and suggest what the relaxation factors should be? or point out any other mistakes I am making?

Thank you!

atareen64 April 26, 2011 11:02

Still having trouble...
I'm still having a lot of trouble with rhoSimpleFoam...
really any help would be good!

rflats May 3, 2011 15:28

Ammar, try changing the relaxation factors to:

p 0.2;
rho 0.2;
U 0.8;
h 0.8;


atareen64 May 3, 2011 15:37

Ricardo, just tried these values: the solution blew up after 17 iterations.

Also my goal was to run the rhoSimpleFoam solver on a laval (convergent-divergent) nozzle. My nozzle has approximate dimensions of 20 mm by 4 mm by 4 mm, and it has roughly 100,000 mesh cells. Also I noticed my program reaches a few thousand iterations if set the under-relaxation factors to really small values like 0.005, but still the results look unphysical: I get negative pressures!

I am totally confused about why this is happening. I'll go through the boundary conditions once again and perhaps post the case here so you can look at it may be?

Thanks for helping out!

rflats May 3, 2011 16:05

What OF version are you using? With 1.6-ext these parameters where ok.

If you are using 1.7, try to change the linear solver for h as below:

solver PBiCG;
preconditioner DILU;
tolerance 1e-10;
relTol 0.1;
solver smoothSolver;
smoother DILUGaussSeidel;
nSweeps 2;
tolerance 1e-10;
relTol 0.01;

Also, modify the relaxation parameters to:

p 0.2;
rho 0.2;
U 0.8;
h 0.6;

Good luck!

atareen64 May 6, 2011 11:21

I am using OF 1.7.1. The simulation works now: on a whim I set my under-relaxation factors much to small, something like ~ 1e-05. The solution converged and I got good results that I could compare to results that I got from another software.

Thanks for the help though!

All times are GMT -4. The time now is 21:00.