[alc96]

Hi guys,

I'm trying to validate a turbulent flow around NACA0012 at Reynolds number 2e+6. My mesh is structured around the airfoil with y+ < 1 and unstructured outside the airfoil. The mesh is generated with Gmsh. I am using the Spalart-Allmaras turbulence model.

Now the thing is that the lift coefficient follow the experimental curve, but the drag coefficient is in agree at low angle of attack (0-4 degree) but from angle of attack = 5-6 increase a lot with relative error of 200% !!!

At the inlet I have set nut = nuTilda = 0.12 calculated from the turbulent kinetic energy k and the dissipation rate epsilon:

k = 3/2*(U*I)^2 with I = 0.02 and U = 50 m/s (M = 0.15);

epsilon = cmu*(k^1.5)/l with cmu = 0.09 and l/c = 0.1 being l the turbulence length scale and c = 1 m the airfoil chord.

nut = nuTilda = cmu*k^2/epsilon


I have also read that at the inlet vt can be computed as vt = 0.1*v (v = 2.5e-5 in my case) but in this case the simulation crashes because I have the error bounding nuTilda after few iteration.


How can I solve this problem ?

Thank you very much !!