1 Reply Latest reply on May 7, 2009 11:21 PM by Anthony Botting

    2d airfoil meshing and results

    Ben Dankongkakul
      I've been trying to get accurate lift and drag coefficients of a naca 2412 for days now and have finally started making some progress. I've took someones advice here on increasing the boundary to 200*c and the lift coefficient is now pretty close. However, the drag coefficient is still about double the experimental value. I've tried changing the turbulence boundary and testing at different aoa and reynolds number. Any ideas what I could be doing wrong? I heard that drag is difficult to get accurate results with, is that true for even low aoa and moderate flight conditions?

      Also, what is the best way to create a mesh for this simulation? It seems like I can only refine the mesh is horizontal and vertical bands or very closely around the airfoil. When i refine the mesh around the airfoil using to local mesh tool, it creates many extra cells along the z-dir instead of leaving it at 1 cell for 2d simulation. Thank you and sorry for themany questions
        • 2d airfoil meshing and results
          Anthony Botting
          I tried this too and resorted to low alpha's and used the Cp equation (as applicable at low alpha's), integrating the pressure distribution around the airfoil surface using a spreadsheet. It works pretty good but I only was interested in lift coefficient. I too, have also read that it is very, very difficult to simulate drag due to surface effects parallel to the flow (or nearly so), I did find these recommendations from a help file (copied and pasted): It is advisable that airfoil chord lies on the grid line; The back edge should be placed in the cell node; When calculating the airfoil at the angle of attack it is advisable to turn the flow but not the airfoil; It is advisable that leading plane of the computation domain is on ~2 chord length from the airfoil nose, the back plane of the computational domain is on ~3 chord length from the back airfoil edge, the bottom and the top planes of the computation domain are on ~2 chord length each from the airfoil chord (this is the minimal computation domain size)