CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM Running, Solving & CFD (http://www.cfd-online.com/Forums/openfoam-solving/)
-   -   High pressure residual simpleFoam (http://www.cfd-online.com/Forums/openfoam-solving/90758-high-pressure-residual-simplefoam.html)

CiccioPasticcio July 20, 2011 04:30

High pressure residual simpleFoam
 
Hi all,
I am running a simulation for external aerodynamics with simpleFoam. I am quite confident with my "standard" setup, never had big troubles. But there is one case (same bc, just different geometry) where I have some problem. The solution doesn't look to bad, but I have high residuals mainly for pressure. Can this suggest that there is a problem somewhere?
The residuals go incredibly low in the first 50-100 iteration, pressure to 0.0005 (usually these simulation converged at 0.008) but then suddenly start increasing. I am trying also with first order and 1 nonOrthoCorrector but it looks similar. I'll paste the output and some setting:

without corrector:

smoothSolver: Solving for Ux, Initial residual = 9.53135923113133e-05, Final residual = 6.40125609233725e-06, No Iterations 3
smoothSolver: Solving for Uy, Initial residual = 0.00583292437866579, Final residual = 0.000408635173001429, No Iterations 3
smoothSolver: Solving for Uz, Initial residual = 0.00107172584371381, Final residual = 7.33804021586605e-05, No Iterations 3
GAMG: Solving for p, Initial residual = 0.370892890676336, Final residual = 0.00805233261805146, No Iterations 2
time step continuity errors : sum local = 8.93727966506124e-05, global = -7.38809614603939e-07, cumulative = 0.00600977288580794
smoothSolver: Solving for omega, Initial residual = 5.71563590189836e-05, Final residual = 5.20189909513911e-06, No Iterations 4
smoothSolver: Solving for k, Initial residual = 0.00296034798651617, Final residual = 0.000165450839972275, No Iterations 3
ExecutionTime = 68868.42 s ClockTime = 69068 s

with corrector:


Time = 1238

smoothSolver: Solving for Ux, Initial residual = 0.000235167399322179, Final residual = 1.49861216273521e-05, No Iterations 3
smoothSolver: Solving for Uy, Initial residual = 0.0167873770137804, Final residual = 0.00104310120382418, No Iterations 3
smoothSolver: Solving for Uz, Initial residual = 0.00393616314129379, Final residual = 0.000237925232432727, No Iterations 3
GAMG: Solving for p, Initial residual = 0.479556064779533, Final residual = 0.00355207736688959, No Iterations 3
GAMG: Solving for p, Initial residual = 0.171937712915105, Final residual = 0.00129806088756816, No Iterations 4
time step continuity errors : sum local = 7.17828577284389e-05, global = -6.80223237496025e-07, cumulative = 0.00594026418728066
smoothSolver: Solving for omega, Initial residual = 8.49925768943358e-05, Final residual = 5.65726497999539e-06, No Iterations 4
smoothSolver: Solving for k, Initial residual = 0.00579911848397957, Final residual = 0.000283318812986071, No Iterations 3
ExecutionTime = 69734.74 s ClockTime = 69934 s


Settings:


divSchemes

{
default Gauss linear;
//div(phi, U) Gauss linearUpwindV cellLimited Gauss linear 1;
//div(phi,U) Gauss linearUpwindV Gauss linear;
div(phi,U) Gauss upwind;
div(phi,Ualter) Gauss linearUpwindV cellMDLimited Gauss linear 1;
div(phi,k) Gauss upwind;
div(phi,kl) Gauss limitedLinear 1;
div(phi,epsilon) Gauss limitedLinear 1;
div(phi,omega) Gauss upwind;
div(phi,R) Gauss upwind;
div(R) Gauss linear;
div(phi,nuTilda) Gauss limitedLinear 1;
div((nuEff*dev(grad(U).T()))) Gauss linear;
}

SIMPLE
{
nNonOrthogonalCorrectors 1;
pressureImplicitPorousity false;
pRefCell 0;
pRefValue 0;
pdRefCell 0;
pdRefValue 0;
}

solvers
{
//p GAMG { agglomerator faceAreaPair ; mergeLevels 1 ; cacheAgglomeration true ; nCellsInCoarsestLevel 200 ; tolerance 1e-07 ; relTol 0.05 ; smoother GaussSeidel ; nPreSweeps 0 ; nPostSweeps 2 ; nFinestSweeps 2 ; };
p GAMG { agglomerator faceAreaPair ; mergeLevels 1 ; cacheAgglomeration true ; nCellsInCoarsestLevel 200 ; tolerance 1e-07 ; relTol 0.01 ; smoother GaussSeidel ; nPreSweeps 0 ; nPostSweeps 2 ; nFinestSweeps 2 ; };
U smoothSolver { smoother GaussSeidel ; tolerance 1e-06 ; relTol 0.1 ; };
k smoothSolver { smoother GaussSeidel ; tolerance 1e-06 ; relTol 0.1 ; };
kl smoothSolver { smoother GaussSeidel ; tolerance 1e-06 ; relTol 0.1 ; };
epsilon smoothSolver { smoother GaussSeidel ; tolerance 1e-06 ; relTol 0.1 ; };
R smoothSolver { smoother GaussSeidel ; tolerance 1e-06 ; relTol 0.1 ; };
nuTilda smoothSolver { smoother GaussSeidel ; tolerance 1e-06 ; relTol 0.1 ; };
omega smoothSolver { smoother GaussSeidel ; tolerance 1e-06 ; relTol 0.1 ; };
h smoothSolver { smoother GaussSeidel ; tolerance 1e-06 ; relTol 0.1 ; };
T smoothSolver { smoother GaussSeidel ; tolerance 1e-06 ; relTol 0.1 ; };
rho PCG { preconditioner DIC ; tolerance 0 ; relTol 0 ; };
}


relaxationFactors
{
p 0.2;
pd 0.3;
U 0.5;
k 0.5;
kl 0.7;
epsilon 0.7;
R 0.7;
nuTilda 0.7;
omega 0.7;
h 0.7;
rho 1;
}


I have tried different under relaxation but if didn't help.
I hope someone can help me or suggest somethings!


thanks in advance

Ciao!

Ciccio

alberto July 21, 2011 03:41

Since the set-up works in other cases, I would check the mesh.

P.S. It looks like you are using a non-standard version of simpleFOAM.

best,

CiccioPasticcio July 25, 2011 02:56

Hi Alberto,

Thanks for your advice, actually the mesh was not that bad. Btw I remeshed the geometry with some different parameter and now it works fine.

Thanks again, best regards

Luca

alberto July 25, 2011 03:03

Good to hear it is working. Sometimes a few cells cause the problem, and the solution is not really affected ;-)

hz283 December 27, 2012 16:07

Hi All,

I use rhoPimpleFoam (LES) for my simulations. I also use GAMG linear solver to update the pressure, but I found that the the amount of iteration step for pressure is much higher than other quantities. I copy some output here:

diagonal: Solving for rho, Initial residual = 0, Final residual = 0, No Iterations 0
PIMPLE: iteration 1
DILUPBiCG: Solving for Ux, Initial residual = 0.000161779, Final residual = 1.72098e-09, No Iterations 1
DILUPBiCG: Solving for Uy, Initial residual = 0.000254413, Final residual = 2.84039e-09, No Iterations 1
GAMG: Solving for p, Initial residual = 0.140828, Final residual = 9.19848e-07, No Iterations 97
diagonal: Solving for rho, Initial residual = 0, Final residual = 0, No Iterations 0
time step continuity errors (Low Ma): sum local = 1.24031e-11, global = -8.09049e-14, cumulative = -6.05813e-11
rho max/min : 1.8221 1.1966

The setting for the pressure linear solver is as follows:

"(p)"
{
solver GAMG;
preconditioner FDIC;
tolerance 1e-6;
relTol 0;
nCellsInCoarsestLevel 100;
agglomerator faceAreaPair;
mergeLevels 1;
smoother GaussSeidel;
nPreSweeps 0;
nPostSweeps 1;
nFinestSweeps 1;
}

The question I would like to ask is: why the pressure should be iterate so many times compared to others? How can it be reduced to a reassonable level without loss of accuracy? Does anybody know something about this? I really appreciate it if some can give me some help.

regards,
H

hawkeye321 January 8, 2013 17:38

Hi hz283
Please let me know if you figures it out!

nikesh March 5, 2013 22:20

Hii there,
I have a similar problem. I am running RANS simulation for a 2Dflat plate case.
I had also posted this problem earlier but unable to hear from anybody about it.
If you have managed to figure it out, would anybody here plz tell me about it?

http://www.cfd-online.com/Forums/ope...-too-high.html

This is the link I had posted to earlier.
My initial p-residuals are higher than the others Ux,Uy,k,omega, etc..


All times are GMT -4. The time now is 19:34.