help with fluent please

[Atilla]

Ya I know spot the undergrade. but I am in dire need of help. I am trying to recreate the Pavon et al 2004, 'Hydrodynamic study of a canting keel based appendage configuation'. Not in its entirity of course just the eppler 392 section bit. I have modeled the foil in gambit and I've imported it into fluent6. I am trying to obtain the Cl and Cd for varying angles of attack. I am modeling it at 8knots (4m/s ish). My trouble is that I calculated the Re for the foil to be 1.78e6. So I can no longer use laminar flow viscous model. Can anyone tell me the correct turbulent model. Can I use a single equation Spalart-Allmaras or do I have to go the two equations model route
Standard k-e
RNG k-e
Realizable k-e
Standard k-w
SST k-w
I've tried going through the fluent documentation but to be honest it just made my head spin. any help would be very much appreciated. I am a fluent novice and have been unable to find any good guidelines on how to deal with the situation.
Rgds
Atilla

[Karsten]

I never used fluent and I'm a bit rusty on the old fluid dynamics but I'll give you my 5 cents worth anyway.

I think you can use a single equation turbulence model as long as there is only one turbulence area above your boundary layer. I think the way they work is they look for the first turbulence maximum from the surface, assume that that is the end of the laminar sub layer and calculate the speed distribution in the boundary layer from that. Everything else is assumed laminar. This doesn't work if there is for example a turbulent vortex above a turbulent boundary layer. The laminar friction model would be used for the vortex. I used Baldwin - Lomax once and it also had problems with boundary layer separation.

The other model I used was k-w and it got the turbulent vortex above a turbulent boundary layer right becaue it calculates the turbulence at every mesh point. The problem is that it obviously takes longer to crunch the numbers and it is also less stabel than Baldwin - Lomax.

Be also carefull with the mesh. You need a very fine mesh in the boundary layer because that is where you have big speed gradients. The problem is more severe for turbulent flow.

If you really want to know how the models work and which one is good and why I'm affraid you have to do some head spinning. I found it helpfull to figure out what all the bits in those page long equations actually do.

Good luck!

Karsten

[Atilla]

Thanks Karsten.