CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   Solver fails with compressible flow (http://www.cfd-online.com/Forums/cfx/22083-solver-fails-compressible-flow.html)

Eric January 25, 2006 09:13

Solver fails with compressible flow
 
Hi! I use CFX 10 with the option of compressible flow. So I set my density as a function of the pressure. My first question is: can I set the density with something else than "pabs" (p, solver pressure...?). And then, the solver fails when I use a formula of the form rho=f(pabs^2,pabs) and doesn't with a formula of the form rho=f(pabs). Doesn't someone knows why?? Thanks,

Eric

Glenn Horrocks January 29, 2006 18:26

Re: Solver fails with compressible flow
 
Hi,

Does it diverge or some other error? Getting a converged solution with an exotic equation of state (EOS) is difficult, you will need very small timesteps.

Glenn Horrocks

Eric January 30, 2006 11:02

Re: Solver fails with compressible flow
 
Hi Glenn,

I got a convergence when I set a linear law for the density. But when I set a parabolic law, it diverges first with those kind of error message:

| ****** Notice ****** | | While evaluating Static Enthalpy, | | Static Pressure on boundary upwall | | went outside of its upper limit. Its maximum value was | | 8.6234E+20. The bounds error was handled by clipping. | | If this situation persists, consider increasing the table range.

And then it fails with this error message:

| ERROR #004100018 has occurred in subroutine FINMES. | | Message: | | Fatal overflow in linear solver.

I tried to increase the table range but it's still the same... Do you (or someone alse) have some advise for me to avoid that? Regards,

Eric

Glenn Horrocks January 30, 2006 18:05

Re: Solver fails with compressible flow
 
Hi,

The root of the problem is almost certainly to do with the enthalpy/pressure error first mentioned. Sounds like the simulation is diverging. Do a run and stop it before it crashes and have a look in CFX-Post. You will probably find a region of flow which is producing rediculous results and that will be source of the problem.

Glenn Horrocks

Neale February 4, 2006 09:28

Re: Solver fails with compressible flow
 
Eric,

If you are using a CEL or User FORTRAN for the equation of state the flow solver internally generates tables for h(T,p) and s(T,p). One thing to make sure is that the table generation limits match what you expect to happen for your case. The default limits are 0.01 bar to 10 bar and 100 to 5000 K.

Neale


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