NACA 0012 lift and drag coefficient.

I am running ANSYS fluent simulation flow over naca 0012 airfoil.
So far the lift coefficient is similar with the published papers.
The problem is drag coefficient is different so much in term of value compared to published papers.
I have check my mesh quality and boundary condition, it shows no problem.
Any idea why drag coefficient is different ???
 PhysOrg.com science news on PhysOrg.com >> New language discovery reveals linguistic insights>> US official: Solar plane to help ground energy use (Update)>> Four microphones, computer algorithm enough to produce 3-D model of simple, convex room
 I don't know if it's the case, but I do know that calculating drag has always been the bane of CFD.
 To get the drag correctly, you need to resolve the sub-layer. Your y+ value needs to be around y+=1.0. Also, drag calculations are sensitive to the downstream vortices and the pressure profile in the far field. You need to accurately capture the wake of the airfoil for at least 10 cord lengths, and the distance to the boundary of the far field needs to be be about the same length. So basically, there is at least 10 cord lengths of mesh in all directions around the airfoil. Also, the predictions depend a lot on the turbulence model used. You can still be off by 10-20% though because of the bad prediction of the transition point where the laminar boundary layer becomes turbulent.

NACA 0012 lift and drag coefficient.

From the literature I've seen on ANSYS, at least 25 chord lengths were suggested.

I've done around 10 and had problems before.
 I just have done the experiment with NACA0012 finding drag and lift. I had same problem about the amount of drag force. but I think there are some interference. for example... pitot's effect, boundary effect... sorry for weak English.

 Similar discussions for: NACA 0012 lift and drag coefficient. Thread Forum Replies Aerospace Engineering 1 Aerospace Engineering 3 Aerospace Engineering 1 Mechanical Engineering 3 Mechanical Engineering 1