Wind Tunnel Simulation: How to measure the downforce?
 

I think I have exhausted all variables in trying to set up a wind tunnel simulation to measure the downforce generated by a couple of race car rear wings that I have made in CATIA so I am turning to you guys for help.......




I have followed various tutorials provided with the software and they always work just fine but whenever I try to move those setting across to run a sim on my models they never converge or give any noticeable downforce result. Here's a list of stuff I've tried but with no success:

• various angles of attack for the wings
• both K-Omega and K-Epsilon solvers
• grid sequencing (as shown in the 'adjoint wing' tutorial)
• ran the sims to 1000 iterations
• various sizes of meshing

I can't think of everything I've tried but I have attempted at least 30 different methods. Now I am a first timer at both CFD and Star CCM+ so this does put me at a disadvantage but I was hoping that by following some tutorials I would get it, but unfortunately I haven't.

So I now turn to anyone on here that would like to help me out and take my models and set up a tunnel that gets some results and the show me how.
I'm trying to do this for a uni project to measure the DF created at a variety of yaw angles so I have been running sims using full models not half with a symmetry plane to gain the full result.

Any advice is much appreciated......

Regards Rob

Hello Rob,

Here is what I would do.

Create a box around your wing. make it atleast 10-12 times the wing size in all directions.

Create a mesh fine enough to capture the flow well. Use some prism layer mesh on the wing.

Assign the walls of the boundary as velocity inlet, pressure outlet and symmetry.

Create a force report for the wing with the force direction set in the downward direction. Create a monitor and plot from this.

Run the simulation.

Hope this helps.

Hey thanks for the reply........

unfortunately other than making the box (wind tunnel) as big as you have suggested that is what I have done by following various tutorials. I made the box (tunnel) considerably coarser than the wing (to save meshing time)and even put a smaller box around the wing set up under 'volumetric controls' to keep good definition within the air close to the wing.

Do you really think making the tunnel that big will help? Only in all the tutorials I've seen it is never THAT big.

So I was kinda hoping someone else would have a go with my model as much to prove that I am doing something wrong and that there is nothing wrong with the model.

Hi Rob, checked your model, cant see where youve set up any wind tunnel or regions, but like has been said above, you need to create a wind tunnel around it and they do need to be THAT large...Check out the following tutorial in the help guide if you havnt alraedy done so, should be the exact same...
"Adjoint Flow Solver: External Flow over a Dual Element Wing" its under Incompressible flow

Thanks for taking a look at my model I really appreciate it......the reason there is no tunnel set up on it is because I was hoping someone might set one up how they decide is the most appropriate way to do it rather than look at my ways which haven't yet worked.

Unfortunately I have already followed that tutorial (although I did have to make a minor couple of changes) and I got a poor result. The downforce measurement starts at around 0.6N fluctuates for a few iterations and then settles back to approx 0.6N and it would do this regardless of the angle of attack!

The tunnels I made have easily been more than 12 times the size of my model in length and height but width has only been approx 4 times after having seen how the other tutorials were set up. Does this sound right? Do I really need to make it 12 times wider too?

Hi Rob,

if you are using symmetry boundary conditions on all 4 sides (top, bottom, left and right) then 4-5 times might be acceptable.

But I don't think that using a small tunnel alone will give you very bad lift force.

just for information, did you do any comparison with test data?

This may show my naivety of the subject but I was unaware of there being test data to compare against.

I based my models on actual wings that do have CFD data on the manufacturers website but unfortunately no amazingly accurate airfoil profile data is available for them. So I based my airfoil profile on a 'cutaway' photo of the actual spoiler which showed the profile quite well. I realize this isn't an accurate way of doing things but it is the best I can do. However surely even if I put a flat spoiler with no airfoil shape to it I should still get some downforce at certain angles of attack, right?

From what you said, I believe something was wrong in the setup. If you can, send the complete sim file and may be I can take a look.

if the file size is too large, just delete the volume mesh (Representations->Volume Mesh-> right click and delete).

Hey thanks for the offer......unfortunately I can't manage to compress the file size enough to upload on here :(

Is there some other method of getting you the file?

Here you go Kris

Hi Rob,

I looked at the summary file and here are my comments:

1. The wind velocity is set as 1m/s. You specified it as 120kph in the initial conditions. Read about initial conditions in CCM help for more details. The actual inlet velocity (the wind velocity) should be specified under Regions->air->Boundaries->Block.inlet->physics values->Velocity Magnitude.

This difference is also what is probably causing difficult convergence.

2. I noticed is that you are using a coupled solver. I think you should be able to get good results with segregated solver. There is a section on segregated v/s coupled in CCM help. That should guide you better. Either way, I would suggest start out with segregated solver for your first run.

Try these and let me know how it goes.

-Kris.

Thanks Kris.....I'll get on it right away and report back.

Ok so I've run the simulation again but changed it to Segregated Flow and adjusted the wind velocity within the Region like you said and have gotten a much more positive result! Thank you very much for your help!

However although I have gotten a nice steady Downforce result, my Residuals still didn't converge?!?! Is this normal?

Please see the attached screenshot

That looks absolutely normal. I would accept that as a converged solution.

Thanks very much for all your help Kris, much appreciated:)