Discussion Groups

Simulation CFD

Reply
Contributor
adai8636
Posts: 15
Registered: ‎08-11-2011

Attempting to Understand Basic Wing Results

146 Views, 4 Replies
05-15-2013 07:38 AM

This may be more academic than related to the software but I am having a lot of difficulty understanding how the results I am seeing are possible. I have run a 2D analysis on an Eppler E423 and the results seemed reasonable. I didn't calculate Cl but the lift force was approximately 13 lbf coming from the wing scaled by a factor of 10 at an 8 degree angle of attack. Now, when I extend this to a 3-dimensional analysis, I am only getting ~4lbf of lift coming from the wing all paramters held the same. The ends of the wing touch the bounding wind tunnel box and I'm using Advection Scheme 5. The wing length is approximately 30 inches wide with it's chord length approaching ~10 inches, at the same 8 degree angle of attack yet I appear to be making no lift.  What am I doing wrong?

Please use plain text.
*Expert Elite*
OmkarJ
Posts: 422
Registered: ‎10-02-2012

Re: Attempting to Understand Basic Wing Results

05-15-2013 08:37 AM in reply to: adai8636

If you are not getting the exact results, it is related to the setup (mesh size/quality/turbulence model/convergence). But as for why 2D and 3D may generate the different results, this can be theoretically explained. 

 

The 2D simulation of airfoil assumes that the width of airfoil wing in thierd direction is infinite, and the flow in the third direction is zero. While, in 3D (and reality) the width is finite and there is a finite flow in 3rd direction. Hence, for same inlet velocities, the lift for 2D will be always higher than that  from 3D and hence for same angle of attack, 2D typically gives larger lifts.

There is some explanation provided here to estimate how these different values of angle of attack are related. 

http://www.mh-aerotools.de/airfoils/hdi_aoawing.htm

OJ

Please use plain text.
Product Support
Royce_adsk
Posts: 568
Registered: ‎08-24-2011

Re: Attempting to Understand Basic Wing Results

05-15-2013 12:45 PM in reply to: OmkarJ
When running 2D models in Simulation CFD the 'depth' is based on the unit system you are using, which might explain the difference. If you are in meters then the length of the 2D slide is actually 1 meter, if you are in inches then it is 1 in.

What unit system are you using for the 2D model?


Royce Abel
Please use plain text.
Contributor
adai8636
Posts: 15
Registered: ‎08-11-2011

Re: Attempting to Understand Basic Wing Results

05-15-2013 03:44 PM in reply to: adai8636

Royce, I suppose it was in inches which would correspond to a depth of 1 in. If this were true though, wouldn't my downforce have increased when I did my run where the airfoil was increased in depth by over 10X that?

 

I am tending to believe Omkar better. I have a hard time understanding an infinitely deep wing though. At what Aspect Ratio do we classify a wing as "infinitely deep". I assume a symmetric infinitely deep wing would not be the same as the cambered wing either, which the eppler 423 is.

Please use plain text.
*Expert Elite*
OmkarJ
Posts: 422
Registered: ‎10-02-2012

Re: Attempting to Understand Basic Wing Results

05-16-2013 01:01 AM in reply to: adai8636

For the sake of coordinate system, let's say your 2D domain is on XY plane, Z being the third direction along the width.  By infinitely long width (in Z direction), I mean that the ends of wings in +/-Z  direction do not influence the flow structure. In contrast in actual 3D wing, you have a finite width and ends in +/-Z direction influence the flow structure and small perturbations may cause the flow to marginally deviate towards Z direction, resulting in loss of flow in the direction of the air, ultimately resulting in loss of lift.

 

As for example, it is documented that angle of attack of 2 deg in 3D gives same lift as angle of attack of 1.2 deg in 2D.

Little dig and you can find a lot about this in the literature.

 

Buf of course, if the difference in unreasonably large, you may have other obvious problems in modelling.

 

OJ

Please use plain text.