[Scabbard]

Hi guys,

Recently, I am simulating flow past a stationary cylinder by using komega-sst. However, I found that there are 3 omega values on inlet from different calculation method/tools. And the values are totally different.

1.omega=k^0.5/(l*Cmu^0.25) where k is the turbulence kinetic energy on inlet, l is the turbulence length scale and Cmu=0.09
with this condition, at re=10000 my omega on inlet boundary is 0.6388

2.Turbulence variables (k, ε, ω) from turbulence intensity (Tu), eddy viscosity ratio (μt/μ), freestream velocity (U∞) and kinematic viscosity (ν)
nut/nu = k/ (nu*omega) which from the eddy viscosity=0.1, at this condition, my omega is 60

3. Turbulence variables (k, ε, ω) from turbulence intensity (Tu), length-scale (TuL) and freestream velocity (U∞) by using the tools on http://www.cfd-online.com/Tools/turbulence.php
that result shows my omega is 0.3499

Why are those omega value totally different, and by using different omega value at inlet boundary I got different Drag force coefficient

Could you give me some recommendations about which formula I should use for calculating omega?
reference value Freestream velocity U=1m/s,Turbulence intensity/level Tu=2%, Turbulence length scale TuL=0.07D=0.07m, Kinematic viscosity=1e-04, Diameter of the cylinder=1m, Eddy viscosity ratio=0.1

Many thx,
Scabbard

[ArathoN]

I'm interested too in this problem, same thing for the calculation of the epsilon that you can define based on the eddy viscosity ratio or kinetic energy and turbulence intensity.

check this page too LINK.

I can't find the omega file in the sources, that's where we can find the solution.

Ok i searched teh sources and the only reference with a law is this and can be found in the omegawallfunction file:

  231     // Set omega and G
  232     forAll(nutw, faceI)
  233     {
  234         label cellI = patch.faceCells()[faceI];
  235 
  236         scalar w = cornerWeights[faceI];
  237 
  238         scalar omegaVis = 6.0*nuw[faceI]/(beta1_*sqr(y[faceI]));
  239 
  240         scalar omegaLog = sqrt(k[cellI])/(Cmu25*kappa_*y[faceI]);
  241 
  242         omega[cellI] += w*sqrt(sqr(omegaVis) + sqr(omegaLog));
  243 
  244         G[cellI] +=
  245             w
  246            *(nutw[faceI] + nuw[faceI])
  247            *magGradUw[faceI]
  248            *Cmu25*sqrt(k[cellI])
  249            /(kappa_*y[faceI]);
  250     }
  251 }

dunno if this is valid only for the cells near the wall or it is global. I searched the models files but they all retun to the constant omega or omega_ (don't know still the differences) that is not defined by any relation only as a scalar field.

[ArathoN]

ok i've found the relation in the kOmegaSST.H file from line 251 where it is defining epsilon and there you'll see that:
epsilon_ = betaStar_*k_*omega_
From here you can explicit omega. There is another expression that is valid only for compressible flows (there should be teh rho variable but i didn't see it inn the source file) that is:
nut=k/omega.

This is valid only away from the wall otherwise you have to apply the relation that a posted in the last post.