[recnice]

Hi,

i'm trying to simulate the flow over a symmetric airfoil and computing the lift and drag coefficients. The Reynolds Number is 1e6 and i'm using the Spalart Allmaras turbulance modell. The angle of attack is 0


Here is the function to compute the coefficients:

functions
{
    forces
    {
        type            forceCoeffs;
        functionObjectLibs ( "libforces.so" );
        outputControl   timeStep;
        outputInterval  1;

        patches
        (
           Finne-Wall
        );

        pName       p;
        UName       U;
        log         true;
        rhoName     rhoInf;
        rhoInf      1;
        CofR        ( 0 0 0 );
        liftDir     (0 1 0);
	dragDir     (1 0 0);
        pitchAxis   ( 0 0 0 );
        magUInf     22.22;          // [m/s] 
        lRef        0.11;                // [m]
        Aref        0.00036385;     // [m^2]

And this is the solution

SIMPLE solution converged in 319 iterations

forceCoeffs output:
    Cd = 789148
    Cl = -12015.9
    Cm = 0

The lift coefficient has to be 0, because it is an symmetric airfoil. Does anyone has an idea, what could be wrong ??


Denis

[daveatstyacht]

Denis,
Your Aref and LRef are likely the reason you are getting such huge lift and drag coefficients. It is not unusual for a very small lift coefficient at 0 degrees (your lift is 65 times smaller than the drag) due to not quite perfect symmetry of the mesh surface. It is pretty difficult to get it exactly 0 with most meshers.

Regards,
Dave

[recnice]

Hi Dave,

thank you for your reply!

Ok I understand that the lift-coefficient has not to be exactly zero.

The main problem is the extreme high value.
I attached a picture of the model. It looks like a NACA airfoil, but it is actually a fin. So I expected a lift coefficients between 0 and 4. So do you think that this is a "phyiscal" solution?

Denis

[daveatstyacht]

Denis,
As I said in my prior post, I think the issue is your Aref. You could manually calculate the coefficients from your force output: "type forces" rather than "type forcecoeff" to confirm that your simulation is producing physically the right results.
Dave