lift coefficient of an airfoil
 

hi

I have a problem with fluent, I have to modelize a flow around a simple 3d wing and after give lift coefficient(Cl) with different angle of attack.

I used K-epsilon model and I don't know why my (Cl)increase so slowly.

angle of attack:0 Cl: 0.024 (normal)

6 0.04 (wrong)

15 0.053 (wrong) normaly:1.5

I am sure my mesh is good because the drag coefficient is good.I certainly forgot a thing in fluent, Help me please.

Re: lift coefficient of an airfoil
 

A non-zero angle of attack causes flow to separate at the the suction side. And one should be little careful while chossing a turbulence model for the expected separation of the flow. if the chosen k-epsilon is standard one then I guess it will not treat the separation properly. And to choose the other variants of the k-epsilon the mesh requirement has to be fulfilled.

Best wishes, Rakesh Jha

Re: lift coefficient of an airfoil
 

In continuation to my previous response, In case of flow over an aerofoil, prediction of transition becomes important to have a better acurate pressure distribution in both the sides. k-epsilon is not a good model to handle transition of the flow. One may like to test the turbulence model for prediction of transition with simple flow over a flat plate. CFX has SST-Transition turbulence model (use Gamma-Theta variation,, which may predict better. Jha

Re: lift coefficient of an airfoil
 

some suggestion

- try with Spalart-Allmaras ( 1 eq. model in place of 2 eq. model as K-eps is, then less expensive computationally) - check the main features of the flow before judging your analisys on tha basis of cl only (for example cp at stagnation point, behaviour of the pressure around the wing, total pressure values, y+ values coherent with the turb model chosen..to name a few); - if your wing is "simple", try before with a 2d analysis of a section, chosen in area not much affected by 3d behaviour of the flow ( tip or root areas..).So you can take confidence with method, turbulence model, mesh refinement and so on....

Hope it can help you

Re: lift coefficient of an airfoil
 

Hi, a 3d wing is not an airfoil. ;-) Anyways, you might want to try the SST k-w turbulence model as it's supposedly better for adverse pressure gradient flows since it combines the functionality of the k-w and the k-e models. Good luck.

Re: lift coefficient of an airfoil
 

Of course an airfoil is not a 3D wing !!

But a correct analysis made on a simple airfoil will give more confidence in what you will do in a simple wing.

A lot of experimental results made in wind tunnel on simple wing models are compared with 2d analysis before running 3d analysis....and the numerical results are close enough to the experiments in terms to lift coefficient ( not for drag...but you were talking about cl ) to justify a "look " in what happens in two dimension.

A lot of publication show how the aerospace community is facing the development of CFD in industry; turbulence is only one point among many points...and among many turbulence models.

Bye

Re: lift coefficient of an airfoil
 

hi all,

have a very similar problem. wanted to model a wing and the results were no good. so moved to basic and modelled an airfoil and the results were as bad as with the wing. can someone please tell me how good is the comparason of fluent airfoil results with experimental?

jon

Re: lift coefficient of an airfoil
 

hi,

can anyone please tell me how is the comparason of the fluent airfoil results with experimental?

joh

Re: lift coefficient of an airfoil
 

Surely it will be never so close to reality as every would like..!! But this is true for every cfd code.

A simple way is to take the Abbott and Doenhoff text, choose an airfoil and compare your analysis with the results reported. Try to "tune" the model in terms of mesh densities,turbulence, convergence and see how close are you going to the exp. results. Only so you will have the feeling for understanding if the software you are using does well its job or not.

Bye

Re: lift coefficient of an airfoil
 

hi, thanx for the rsponse. with my cfd model i am getting a lift curve slope of .055 per degree as opposed to 0.1 per degree to experimetal and the drag values are as high as 500%. i am sure fluent can do better than this. any comments please?

cheers joh