My last two years in college I had the pleasure of working/playing with the UW FSAE team (visit Formula SAE for more info).  Entering this program was probably a major turnpoint in my life.  Whether for better or worse I don't know yet.  But I do know that it was an interesting mix of headache and fun.  It was an endless struggle trying to balance an ambitious project, schoolwork and some kind of social life.  But that's how it goes; work hard - play hard.  There will be times I'll remember the long nights we've spent laying up composite body panels, sanding away at molds and throwing chips in the shop.  All the staring at the computer screen waiting for SolidWorks to cough up data or crash.  The times I slept on the couch...  And all those progress reports... 
    But then there were the nights we spent playing gran turismo (turns out there really is an optimal level of intoxication for racing)!  I learned that getting boxed in by cops is not as fun as drag racing in the parking lot...  Then there was the time we were stealing burgers from the lounge and David spilled a whole bottle of ketchup in the elevator, only to drag an incriminating trail of red ooze all the way back to the pit.  Never satisified with our mayhem we would brutalize the Think bike and the waxed floors in the ME corridors.  There's a certain camaraderie in being labeled a 'degenerate' for damages done the previous night.  And then there was the car, the 4 cylinder beast.  Here we were, a bunch of college kids, with a car that raced to 60 in 4 way before any of us could afford such a thing.  And our job was to make this doozy go faster yet.
   
    Aside from laboring on composites my field of interest in FSAE was aerodynamics and downforce.  Typical of me I picked a rather difficult task.  Aerodynamics in SAE is not a very well understood or widely practiced field.  It is hard to achieve success with aerodynamics in FSAE for various reasons.  At the competition the cars usually corner at low speeds.  The rules and regulations that have to be obeyed are very restrictive and render aero packages even less effective (for example, the minimum front radius of the wings must be 1/2").  This craft is not very well understood because all the information on aerodynamics one can find is useless because of these rules, ie, the geometry is very different.  By 2004 there were maybe 4 different teams that were trying aerodynamic devices.  I think only two of them had any kind of success (UTA and Monash from Australia).

    I give serious props to University of Texas at Arlingon for this aero package.

    Basically,for aerodynamic devices to be worthwhile on the car they have to be completely optimal and as light as possible.  This means a lot of design, testing and composite construction.  For design there were two factors that helped:  FloWorks flow simulation package and my background in software.  To run a single iteration in FloWorks takes time and there are a lot of wing geometry variations to test.  Start putting multiple elements (ie, flaps) together and the exact positioning adds even more complexity.  And what about differences between the front and the rear wing? 
    So, instead of manually running these simulations I decided to put in a little effort early on and work up an automatic system.  I defined the wing Geometry in SolidWorks with easy to adjust parameters.  I then created an Excel spreadsheet that tied into SolidWorks using VB macros and could alter the wing geometry through a given range of parameters.  Then came the extra special DLL (Dynamic Link Library) that I wrote in MVisualC++ to control FloWorks (because unlike the rest of SolidWorks, FloWorks had no programmers interface).  This flochart shows how it all came together:



    Some of the results for single wing geometry.

    So, for a couple of weeks following that I would go into the ME lab before heading home and start the 30 computers on some iterations.  The software automatically loaded the geometry into SW, ran the FloWorks test, gathered the data, plotted it onto the chart, loaded new geometry and so on.  The next morning I would come in and collect the data, analyze it and pick the best geometry.  I estimated that I ran over 1000 computer hours of CFD.  Good thing I didn't have to baby sit it...  What I finally picked as the shape of the devices theoretically is the best possible.

The rear wing.  The main element has a 12" chord.  This combination achieves perfect balance between aggressive camber and flow separation management.	Front wing.  Note the cutout in the endplate.  This relieves the pressure in front of the wheel reducing unwanted turbulence and also keeps the flow on the underside attached.
Side view of the complete aero package.	Iso view showing the suspension links.  This was to be a suspension mounted package like UTA's.
Front view of the '04 car with the package.

    Before launching into manufacture full time I also tested the wing geometry in a UW windtunnel:

The model of the wing undergoing simulation.	Actual test wing in the wind tunnel.
At the controls.  I know, I know, it looks like I got a boner.  I assure you that is NOT the case.  This is cool, but I don't get excited about it that way.	Matt at the controls.

    Making the wings was no peice of cake.  I wish I could have had female molds machined, but the more economical way to go turned out to be hot-wiring foam blanks and laying up carbon fiber on top.  I covered the layup in mylar sheet coated with a gelcoat before throwing it all into the vacuum bag.  After the mylar is peeled off the result is a smooth, hard gel coated carbon fiber wing that weighs almost nothing and could easily support my weight.

The foam blank with the metal wing profile ready to be cut.	Hotwiring the wing blank.
CosmosWorks FEA on the undertray suspension bracket.	Ansys FEA on the rear wing rib.  This helped me decide on how many layers I needed there.