k-wSST convergence problem
 

Hello everyone,

ı tried obtain drag and lift forces for an airfoil. ı couldnt decrease k-residual under 1e-3 with linearUpwind scheme and (i guess) it is not enough to calculate accurate drag force(ı get drag force with %20 error). When ı use Gauss upwind for k, all residuals decrease under 10E-5, but, i get very inaccurate drag force. thanks for your advice about k and other parameters.

setup;
ddtSchemes
{
default steadyState;
}
gradSchemes
{
default Gauss linear; grad(U) leastSquares; grad(p) leastSquares;
}
divSchemes
{
default none;
div(phi,U) Gauss linearUpwindV grad(U);
div(phi,k) Gauss linearUpwind grad(k);
div(phi,omega) Gauss upwind;
div((nuEff*dev2(T(grad(U))))) Gauss linear;
}
laplacianSchemes
{
default Gauss linear corrected;
}
interpolationSchemes
{
default linear;
}
snGradSchemes
{
default corrected;
}
wallDist
{
method meshWave;
}
// ************************************************** *********************** //
solvers
{
p
{
solver GAMG;
tolerance 1e-6;
relTol 0;
smoother DICGaussSeidel;
cacheAgglomeration no;
maxIter 50;
}
"(U|k|epsilon|omega)*"
{
solver PBiCGStab;
preconditioner DILU;
tolerance 1e-6;
relTol 0.0;
}
Phi
{
solver GAMG;
tolerance 1e-6;
relTol 0;
smoother DICGaussSeidel;
cacheAgglomeration no;
maxIter 50;
}
}

potentialFlow
{
nonOrthogonalCorrector 10;
}
SIMPLE
{
nNonOrthogonalCorrectors 3;

momentumPredictor on;
consistent yes;

pRefCell 0;
pRefValue 0;
}
relaxationFactors
{
"(U|k|epsilon|omega).*" 0.7;
pressure 0.3;
}

cache
{
grad(U);
}

Have you check your mesh quality? The relaxation factors are too high. Try to use less than 0.5

****************
Max cell openness = 3.70842e-16 OK.
Max aspect ratio = 10.3301 OK.
Minimum face area = 2.61443e-10. Maximum face area = 0.00306011. Face area magnitudes OK.
Min volume = 1.09866e-11. Max volume = 7.5217e-05. Total volume = 0.0855509. Cell volumes OK.
Mesh non-orthogonality Max: 40.2866 average: 5.66878
Non-orthogonality check OK.
Face pyramids OK.
Max skewness = 0.844117 OK.
******************************

My university professor said that to get good drag coefficient I should use Gamma-Re-theta turbulence model.

Hello again,

sorry for the late reply. I have read your answers now because of some reasons. I tried small relaxation factor(for all, 0.3) and now convergence of all variables is satisfied~2e-5-I guess :). However, cd error still too high. can it be related to unstructured mesh or flow conditions( Re=5e+5) ? By the way max. y+ =4, i wonder if ı should decrease first cell height?

it is very logical, in conclusion the flow around a airfoil (c=0.5 m) is not fully developed turbulent flow everywhere on the wall. However, as far as ı read in the forum the people generally can obtain accurate cd with conventional RANS models. Is there any tutorial with gamma-Re-theta turbulence in the web?

Maybe, at least ı should try to get cd for a longer ı should try to get cd for a longer airfoil(maybe, c=1m.) which causes the more fully develop turbulent flow

What turbulence model are you using now? k-omega SST should have done the job well. Have you use the kLowReWallFunction?
I think you should refine more at the region fore and aft of the airfoil