Way cool CFD picture

[Wright D]

I was sifting through the “my pictures” folder on my computer and rediscovered this picture. It saved it back when I was developing the F1k.08 aero, I had forgotten about it. I thought that it looked cool, so I just wanted to share the picture with you guys.

[Frank C]

Why are the contours asymmetric across the centerline?
- Frank C

[Wright D]

That is an excellent question.

I can think of three reasons why the flow lines are asymmetric. I let the CFD programe add the insertion points automatically. It placed the points asymmetrically. Also the flow through the side pods is not fully symmetric; more air flows through the radiator then the oil cooler. I have also added the hot exhaust exiting the right side.

[Keith Carter]

Awesome pic! Any other angles?

[Rick Ross]

Hi Dustin,

Nice pic. Which CFD application do you use? Also, I was wondering if you modeled different wheel sizes (not sure what the FB rules allow)?

[bdsmith21]

Looks like Solidworks add-in "Floworks"

[Rick Ross]

Is that a Windows-based application? When I graduated from engineering school over 20 years ago, you almost needed a super-computer for any decent CFD applications !

[Rennie Clayton]

Is that a Windows-based application? When I graduated from engineering school over 20 years ago, you almost needed a super-computer for any decent CFD applications !

20 years ago huh? Did it output the results on punch cards?

If I'm not mistaken, FB requires 13" wheels, maximum width of 10"


Cheers,
Rennie

[Billy Wight]

`

That is an excellent question.

I can think of three reasons why the flow lines are asymmetric. I let the CFD programe add the insertion points automatically. It placed the points asymmetrically. Also the flow through the side pods is not fully symmetric; more air flows through the radiator then the oil cooler. I have also added the hot exhaust exiting the right side.

Most likely the assymetry is from an assymetrical distribution of pathline initiation points. I had the same issues on some internal flow problems using Cosmos - the initiation points auto generated were not symmetricly distributed and manually inserting them (with symmetry) proved to be more trouble than it was worth...

What kind of computer were you running this on, and what were the run times? I was seeing about 10 hours of computational time for a rather small internal flow problem with a fairly advanced desktop system, though the cell sizes were rather small and it was compressible flow. Looks good though, how many iterations did you guys go through between the 07 and 08 areo?

[Wright D]

The CFD program I use is in fact Flo-Works.

I have built a special propose CAD, FEA, and CFD computer. It is a windows 64 bit. I don’t use it for any thing else, just engineering. The computer is nothing to great; I built it when I was in school, three years ago. It is not hooked up to the internet (virus software slows down machines), and only has a single DVD burner, and a few USB ports. Solid Works runs best when you use one of the graphics cards that they recommend. The special graphics card and a large flat panel monitor are the only changes I have made to the computer since it first stated humming.

The run times are typically 16 hours. I suspend the calculation during the day, so I can shut down the computer.

I have modeled different wheel/tire size combinations, and don’t want to share my results just yet, but thanks for asking.

There have been more iterations then I can even count. Lots of small aero tweaks are still being investigated.

My plan is to develop an aero map of the car before we go testing. The idea being that it might make aero changes at the track a bit easier, and help us to get the car balanced faster.

The Picture I have attached is of the air flow though the cooling system. There have been about 20 different side pods, chimneys, louvers, and wheel flip ups each. The trick is to get them all to work together. The first chimneys we put on had the air flow going the wrong way! The air was flowing into the chimney and out the back of the side pod, with little to no air going though the radiator. Looks like the aero solutions developed by F1 and F3 are not as simple to implement as I thought they would be. The cut and paste idea just didn’t work, but it is fun to develop an idea till it does work! Well, at least in the computer…Can't wait to get it onto the track!

[Rick Ross]

Thanks, Dustin. Very cool stuff. I'm always amazed at what can be done with a good desktop PC set-up these days.

Rennie -- I narrowly missed the slide-rule era, but I do seem to recall a punch card or two! When I was in school the wind tunnel was still king and CFD was the up-and-coming gee whiz stuff. I was able to obtain an aero engineering degree with only one computer science class (fortran!), and I probably only wrote 2 or 3 programs, none of which worked very well. But I ran plenty of wind tunnel tests and wrote lots of lab reports. I guess things in the aero engineering world have changed a little bit in the past 20 years .

Rick

[ghickman]

Dustin
Are those Barge Boards I see in the top view of your CFD screen shot? Planning on running those?

Gary

[bdsmith21]

Verry Cool! Thanks for sharing these.

At my work we do a lot of CFD work using CFDesign (on windows, no punchcards). I have worked on two full car simulations since being employed after graduation in May 07'. The more complex bodied car took about 5 hours to converge. A nice thing about the software was that it lets you set up multiple designs and run them sequentially at night. This way I could do about 3 design itterations per day. I did an interesting rear wing slot gap optimization that I could actually share if anyone is interested.

Were you able to predict the center of pressure with floworks? If so, have you tried changing it by changing wing settings ect.?

[Wright D]

Yes those are bargeboards. Our aero development would not be complete if we did not try them in the computer first. I think we will still need to run some more testing before we diced to run them or not.

You can indeed locate the center of pressure, and track its migration due to pitch, heave, roll, wing settings, ext. all using Flo-Work. To do this I set up to run multiple iterations. Instead of moving just the slot gap, as bdsmith21 did, I move the whole sprung mass of the car. I run, three different ride heights, each with three different pitch angles. This can all be done in a one-shot process using the “Batch Solver” feature of Flo-Works. When it is needed I can run extra calculations manually.

[bdsmith21]

Here is a ink to pictures of the wing optimization. It was purely a flap position study. I was looking for trends in lift and drag for a wing only. The tip vortex picture is the same wing with no endplates.

http://www.facebook.com/album.php?ai...36&id=52701181

[RobLav]

Nice graphs Brian, but can you label the x and y axis?

[bdsmith21]

The axis refer to displacement of the flap element from the original design. The X axis is longitudinal displacement and Y is vertical, in inches. The angle of attack was constant. The numbers are really only usefull to show the scale. The Y axis should probably have negative numbers because the the bottom of the graph coresponds to the flap almost touching the main element.

[RobLav]

Thanks - I understand it now.