2D NACA 0012 simulation in CFX
 

I ran the steady-state simulations with SST model. The Reynolds number is 360,000. In the light of the experimental data, the stall angle should be around 10~11 degree. However, I got the angle around 16. Can anybody give some help? Thanks in advance.

Re: 2D NACA 0012 simulation in CFX
 

The flow space is big enough? I mean, that in the flow area, the wing must have relative very little dimensions. I've seen a problem like that, with Naca0012 also, and it gives a stall angle about 18degrees. But the wing was "too big" in the comparison with the "air space". Maybe an other turbulence modell gives better values, closer to the reality, but i'm not competent in turbulence modells too much.

If you have results please write it!

Attesz

Re: 2D NACA 0012 simulation in CFX
 

The chord length is 0.1 m while the dimension of the whole flow area is 3 x 2 m. Is it big enough?

Re: 2D NACA 0012 simulation in CFX
 

Hi,

Do a sensitivity analysis to establish if it is big enough. While you are at it, check your convergence tolerance also. predicting stall angles is tricky stuff and you will need a very accurate numerical scheme to get it.

Turbulence transition may also be important.

Glenn Horrocks

Re: 2D NACA 0012 simulation in CFX
 

an example:

http://kepfeltoltes.hu/view/090303/i...toltes.hu_.jpg

attesz

Re: 2D NACA 0012 simulation in CFX
 

Hi, I checked the picture. So, how many cells do you get in all? The distance of my first cell adjcent to the airfoil surface is 0.01 mm. Is it small enough? Cheers.

Re: 2D NACA 0012 simulation in CFX
 

Hi,

For airfoils the best way to discuss mesh resolution is by y+. What y+ does your mesh work out to be? How many elements along the chord and span?

Glenn Horrocks

Re: 2D NACA 0012 simulation in CFX
 

I got 240 elements along the chord, and 240 elements along span, respectively. The maximum y+ is around 0.8.

Re: 2D NACA 0012 simulation in CFX
 

Hi,

That mesh should be OK but it is problem dependent so do a sensitivity check to confirm. As you are trying to get the stall angle you will need to be careful that everything is accurate.

So do a sensitivity check on the mesh size and size of the external domain.

Also make sure you don't expand the mesh too quickly away from the airfoil. For delicate stuff like this I recommend 1.05 maximum and more like 1.02 or 1.03 if possible (again a sensitivity check is best here).

Glenn Horrocks

Re: 2D NACA 0012 simulation in CFX
 

Hi, Glenn.How do you do a sensitivity check? Change the mesh slightly and compare the results? Is there any other more efficient ways? Cheers.

Re: 2D NACA 0012 simulation in CFX
 

Some people mentioned "transition model". Does it help?

Re: 2D NACA 0012 simulation in CFX
 

Hello Hao,

If this a 2D simulation then I don't understand why must you divide the cells along the span of the wing? What are the boundary conditions at the end of wing? Symmetry?

Roland

Re: 2D NACA 0012 simulation in CFX
 

Hao,

By the way I think, it is not sure that the error is anyway in your calculation and settings. If your calculation is a 2D case then you are going to get different results as in the case of the measuring because there it can not be eliminate the 3D effects (for example the effect off the walls at the end of the wing)...

Re: 2D NACA 0012 simulation in CFX
 

Hi, Roland. You are right. The boundary conditions are both "symmetry" at both the ends of the wing. I mistook "span" for "width". Yes, It's a 2D simulation and there is only one cell in the span direction. So, you think I probably cannot get very accurate result of stall angle in 2D regime, don't you?

Re: 2D NACA 0012 simulation in CFX
 

Hi,

Doa significant change in mesh resolution, maybe double and half the mesh density (ie 1mm elements become 0.5mm elements for the fine case and 2mm elements in the coarse case). Then compare a parameter of interest - in your case stall angle. Work out the stall angle for the different meshes.

A simple approach is to keep refining the mesh until the difference between meshes shrinks to an acceptable tolerance. A more sophisticated approach uses the grid convergence index (GCI) concept and Richardson Extrapolation to get even more accurate solutions (see http://www.grc.nasa.gov/WWW/wind/val.../spatconv.html - the verify program at the bottom of the page is an excellent tool for analysing grid convergence.)

Glenn Horrocks

Re: 2D NACA 0012 simulation in CFX
 

I calculated a lot of NACA and Joukowsky airfoils recently around Re = 100000-300000. I would like to plan a new wing for my model aircraft, therefore I study some airfoil in 2D with SST turb. model. I experienced in case of the most airfoil that the stall angle is ~12-13 degree in this Re number range (as in the reality in the case of measurings). I think that your stall angle really too small...

Re: 2D NACA 0012 simulation in CFX
 

Sorry, I mistook my last comment; your stall angle is too large...

Re: 2D NACA 0012 simulation in CFX
 

Hi, Roland. I also think that my results are not good. So far, I have tried a structured grid with total cells of 90,000 (one cell in Z direction, created in ICEM). And the stall angle is still too large. Would you mind sending me some of your res files? I am really stuck with this for a long time and the due day is soon. Thanks in advance.

Re: 2D NACA 0012 simulation in CFX
 

Hello Hao,

I studied the NACA 0012 airfoil yesterday, the stall angle is 14 degree in my simulation. In the case of 15-16 degree the lift coefficient decreases clearly. Where found you the results about the measuring? I think that you send me your *.cfx file, so I can see your settings and I can help you perhabs. (rakosroland@gmail.com)

Re: 2D NACA 0012 simulation in CFX
 

Hi, Roland. I sent my file to you. Have you got it? As for the measuring data you asked, I found it at http://www.cyberiad.net/library/airf...ta/n0012cl.htm