CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   Main CFD Forum (https://www.cfd-online.com/Forums/main/)

 jd88 August 30, 2010 17:39

Velocity Vectors in VLM Tornado Code

Hello,

I would like to use the vortex lattice code Tornado to do an analysis over a swept wing. I was wondering if anyone knows if and how I can modify the source code so that instead of getting just the pressure distribution I can also get the velocity vectors.

Where should I start? What needs to be added?

Thank you

 CAVT September 3, 2010 14:44

It's impossible to do so not only in Tornado but also in any other VLM or doublet lattice method. That's because these methods calculate DeltaCp but not Cp on upper or lower surface (consequence of assuming a thin surface), in which case you would perfectly be able to calculate V from the formula:

Cp=1-(V/Vinf)^2

You might be able to do a trick by using a Weissinger method which uses actual airfoil data for chordwise load distribution, or examining section by section using airfoil softs like Xfoil, Javafoil or PABLO (this one's available for MatLab) and setting the section Cl you obtain in Tornado as an objective to reach and then you will have a guess. However, this is not going to estimate (IMHO) crossflow, which in swept wings is important.

My suggestion would be using a panel method, which still uses potential flow, needs low computer resources and accounts for airfoil thickness (you can even estimate some boundary layer properties). Try with these two, which are open source:

http://3dpanelmethod.com/index.php (there's a legacy version in MatLab/SciLab language).
http://xflr5.sourceforge.net/xflr5.htm

Hope it helps :)

 jd88 September 3, 2010 16:11

Somewhere in Tornado or any VLM method has to calculate the circulation of each vortex segment. And once these circulation are known can't I just use the circulations and potential flow to determine the velocity induced at any point by all of the vortex segments. Then add this to the free stream velocity and obtain the velocity at that point.

That is what I am currently trying to do. I thought I was doing it correctly but the answer doesn't make any sense to me currently. Is there something wrong with this idea?

 CAVT September 3, 2010 16:22

Could be if you think it in that way, but honestly I never tried it and never seen anyone doing so. The variable you want to check then is, I think, results.gamma, which is solved in the solver9.m file.
You may want to look at the Treftz plane calculation file, the code there might be of interest for you.
Success with it :)

 CAVT September 3, 2010 16:40

TAke into account that even if you calculate the velocities on points which would correspond to the actual surface, you cannot guarantee they are good values, since there wasn't any tangency condition indicated. I reccomend you to validate the results at least for one or two cases with a panel method. But you made me realize of a new possibility, thank you for your doubt, sometimes we need to hit the wall to see it ;).

 jd88 September 3, 2010 16:42

Thanks! I also thought there would be a problem when calculating the velocity on the surface. I was thinking of just sticking to points that are just above or below the surface. Hopefully that works a little better. I have been working on getting the code working for awhile. I will let you know how it goes.

 All times are GMT -4. The time now is 04:04.