[playboy]

Does anyone know if freestream velocity of 15m/s over an airfoil needs a spalart almaras type model or a k-epsilon?? Im assuming k-epsilon due to slower freestream velocity speeds. Also with respect to the model done in Gambit, Ive put the wing in the centre of a large circle 10 chords in radius-is this feasible, if not should i use a velocity inlet(change geometry to a square)??

Thanks

[gerritgroot]

I don't know about the turbulence model. Just apply trial and error if you happen to have wind tunnel data. I know that k-w doesn't work well at high alfas. If you use fluent, make sure y+>30 so fluent uses wall functions. (It's not worth the effort to really compute the viscous sub-layer, certainly if you only look at integral data like cl and cd)

About the farfield, the shape shouldn't matter but can make life easier if you're forced to give inflow and outflow data, in fluent that depends in the solver you use. If possible, use pressure far field. 10 chords seems very near to me, I'd use 25 chords. I know there are solvers that somehow place a vortex at infinity in order to be able to place infinity closer to the geometry, but I don't know whether fluent has that (suppose not, as it's a general solver), so to be sure I'd use 25 chords.

If it's 2D flow, you might as well forget about gambit and use some elliptic grid generator, they're not difficult to make, maybe you can find them for free on the web as well.

[playboy]

Thanks, I will try the eliptic grid generator mayb write my own source code. I guess it must be composed of Jacobians not really sur but will read up. My airfoil is a Jalbert airfoil or RAM airfoil used in paragliders I have a source for this but the paper is 25$ for AIAA menbers. I have "The theory of wing sections" but it doesnt have these type of airfoil. My CD values are quite off. Test speeds are between 11-20m/s. Do you mind me emailing you my case file perhaps?? If you not too busy and comment accordingly.


Thank you