Suspension is king in FSAE, but there are some simple changes that could be made to benefit aerodynamic performance
As I've mentioned, we don't yet know with enough confidence what specific vehicle dynamics behaviour to target or optimise for when designing aerodynamic components. I won't even claim we are entirely confident when designing suspension, but we do have a better idea as we've been making suspension for just as long as we've been making cars (which isn't that long either mind you). Combined with the fact that poor suspension will doom an FSAE car more than poor aero, we decided between the aero and suspension teams that suspension would be designed entirely to suit the vehicle concept while ignoring all aerodynamic forces other than pure steady-state downforce. Setting a low downforce goal allows us to do this confidently; it's is the beauty of aiming small to start with, as I've also mentioned.
The optimal suspension setup initially offered to the aero team already had a bit of progression to keep roll rates in check while avoiding having to design and manufacture anti roll bars. This was a good start, but it meant the chassis was expected to drop 11mm near Vmax; nearly half its available bump travel. Our test track is pretty bumpy, and a few quick tests in our transient/impulse suspension simulation showed it getting worryingly close to bottom-out when our heaviest driver experienced a moderate bump. The progression was increased at the front and rear (by simply adjusting the rocker design) in proportion to the expected aero balance until the chassis only dropped 8mm (equally front and rear), which gave us a slightly more reassuring clearance in the same simulation case (Figure 1).
Figure 1: Front and rear air spring rates based on a compromise between optimal suspension performance and safe ride-heights under aero loading, with the red line indicating front and rear ride heights (wheel travel) at Vmax
Pitch rates looked reasonable and I had no reason to request a change. Positive longitudinal acceleration wasn't a concern, and under 2g of braking the pitch was expected to be 1.4 degrees. If we had a front wing we may benefit from a stiffer pitch rate, particularly for maintaining aero balance, but with just an undertray I didn't expect any significant issues. Roll rates also looked ok, especially after increasing the progression further. At 2g lateral the prediction was 1.9 degrees of roll. Perhaps more than is ideal for an undertray, but I can't really say having not yet run any CFD.
After going through the initial suspension parameters, there was still an outstanding issue with the roll of the car. As it was, a hypothetical undertray filling the entire regularity box between the front and rear axles and sitting flush with the chassis floor plane would impact the ground at its lateral extremities under 1.9 degrees of roll. One solution was to further increase the roll rate but I was told this wasn't really possible at this point without starting to see some significant sacrifices elsewhere (the drawbacks of a non-decoupled suspension setup). A goal of the suspension team this year was to reduce jacking, (when the roll centre is significantly to the outboard of the chassis, such that the car lifts as it rolls). Instead of increasing roll rate, I requested the jacking goal be lifted slightly and worked to evaluate the trade-off between this and achieving an additional 15% usable plan area for the undertray (without having to deviate upwards from the chassis ground plane). In the end we decided to go down this route; it also meant a smaller change in suspension since the previous year, which seemed safer. The end result was 5mm of ground clearance for the outboard tip of the hypothetical undertray with the heaviest driver pulling a theoretical 2.1g through a corner. Figure 2 shows the effect of the jacking.
Figure 2: Example rear wheel displacement under various longitudinal and lateral accelerations, where the red line approximately shows the contour for 0g longitudinal. It shows that full suspension compression occurs at >2.1g lateral, while full extension occurs at only 1.6g lateral, resulting in a net lift between left and right sides.
The final change we requested for suspension was a slight increase in anti-dive geometry. Our goal was to maintain the ground clearance of the region at and immediately forwards of the expected location of the undertray's "throat". This would significantly reduce downforce pitch sensitivity, and potentially also decrease balance sensitivity. The anti geometry can be changed small amounts without significantly affecting other kinematics, and the chassis up until this point was designed to allow room for suspension points to shift within a reasonable tolerance without putting any tubes outside of rules, so this change was easy to make. The result of increasing the anti-dive more is that the centre of the car now lifts slightly under braking (Figure 3), but without increasing ground clearance. The only issue is the effective position of the "throat" shifts forwards, but this would happen anyway and is unavoidable without active suspension or some seriously wacky wishbone angles. I expect this detriment will be partially counter-balanced by the raised diffuser increasing drag and applying a moment back onto the rear axle.
I'm not going to claim all of these changes were worth it for just an undertray (although they might be...time will tell), but these are ongoing designs and things will only get bigger and better.
Figure 3: Illustration of the effect of increasing anti-dive geometry under forward pitch (braking), showing the expected approximate longitudinal location of the undertray's "throat" sitting forward of the geometric centre