Defining the scope

The scope sets the limitations on how we should go about achieving our goals, and how much our goals can or can't be stretched or changed. This requires that we have goals to apply such limitations to. In the previous article I covered how we went about selecting specific optimisation and test target cases, but I did not mention any specific or measurable performance goals so I'll do that on this page, interwoven with the scope outline. 

With our small team and lack of sufficient previous data, it was obvious that designing, manufacturing, and testing a full aerodynamic package (front, mid, and rear of car) within 12 months was not going to be achievable. The team had done that before (twice), but that team was larger and even then the end results were lacking performance to justify the effort. Even worse was that testing time was stretched too thin and the team never understood the performance that was there, nor did they effectively record and document any info they did have. As such, there was no choice but to start design from scratch the following year, until ultimately the plug was pulled on aerodynamics entirely in 2018 for the next 6 years. 

That's where this year's goals and scope come in. This time, we're looking beyond the 1-year cycle. We want justified and worthwhile performance, but most importantly we want to set up a foundation to build upon each year. This means that for this year, development of testing methods and simulation software fall directly under the main overarching aerodynamic goals, and dictate every choice made. It's from this idea that our individual SMART goals and KPIs for the aero team naturally spawned, rather than forcing uninformed or arbitrary goals to meet within a restrictive 12 months before starting again. The build-up to this sustainable ongoing development was organised into a 5-year plan for the aerodynamics team:

Goals:

• Have the aerodynamics team big enough and experienced enough to be able to develop full aerodynamics packages on a 24-month optimisation cycle, with justified performance vs resources and development time trade-off, and sufficient data and documentation to confidently feed continual improvement 

• Solid understanding of both steady-state and dynamic suspension behaviour to enable informed decisions on maximising performance beyond just downforce and drag, including performance and platform response during transient suspension scenarios

• All significant performance-altering aspects of the CFD model should be either semi-permanently validated (pending "large" changes to the car), or have a defined recurring validation test plan 

  • Necessarily, be able to quickly determine, record, and understand aspects of the CFD model (both aero design and sim setup) that have significant effect on estimated performance whenever significant design changes are made

General timeline and methods to achieve the goals:

• Spend two years (first 24-month design cycle) focussing on just a simple ground-effect side-floor design to provide experience in CAD/CFD based surface design/optimisation, as well as experimental testing, kept within a manageable scope. During this time, develop and continuously improve the CFD setups for precision, accuracy, and relevance/usefulness. Ideally, the second year will be purely optimisation with no concept changes to maximise the benefit from the process. 

• Spend two years (second 24-month design cycle) adding a rear diffuser and side-mounted wing design to add an additional 50%-100% downforce. These increased forces will allow a similar learning experience to occur, but this time focussing on vehicle dynamics and suspension interaction as well as larger-scale aerofoil manufacture/mounting. Once again, by keeping the aero surfaces concentrated toward the centre of the car, adverse effects of a poorly set or sensitive aero balance are minimised while the aero team continues to learn about on-track sensitivities and general aero performance. While the forces are increased, they are still low enough that significant changes to suspension are not likely to be needed beyond the simple adjustments outlined in Suspension design. 

• Spend the second half of the fourth year (end of second 24-month design cycle) setting the goals and scope for a full aero package, while considering manufacture and mounting planning as these will now be significant processes in their own right. Begin surface design, continuing in the fifth year, and have a simple but efficient and insensitive full-aero package by the end of the fifth year that can be easily optimised and built upon with reasonable confidence over subsequent years. Consideration will be needed at this point as to whether the team chooses a sprung or unsprung aero mounting method.

Setting the scope to define the first two years was quite simple, as a centrally-mounted floor was really the only option if we were to have only one main aerodynamic component. Why? Well if we stuck on just a rear wing, we'd end up with the centre of pressure effectively acting several metres behind the car, thanks to not only the rearward position of a "rear" wing, but also the enormous drag moment they create. Similarly, if we made just a front wing, we'd have to limit its total downforce to avoid the opposite problem. Providing there isn't too much drag up high, making a decent amount of downforce with a central floor will hopefully bring the centre of pressure into the ballpark of our goal of 75% rear. It will also avoid too much complex mounting, as there is plenty of chassis real-estate and no rules in this area other than the already unrestrictive regularity zones. 

That's a rather open scope for now, but it will be refined once we start looking at actual surface design and what we think we can achieve after a bit of initial CFD analysis. We have enough to set some performance goals for this year though; some quick research and reconnaissance of other teams suggest a well-designed but simple floor should be capable of around CLA = -1.0 to -1.2. We set a target of -1.2 accordingly. For reference, if our car attained this during the skidpan driving scenario, we would be making 6-7 kg of downforce. At our Vmax of ~120 kph, it would be 75-80 kg. We expect that a CDA of no higher than 0.7 should also be achievable, and therefore targeted. These downforce figures are low in the world of FSAE, but with a worst-case design leaning on the high end of both weight and standardised cost event price, we are still making an estimated 14 point gain (2.8%) over all competition events, albeit assuming a consistent downforce in all driving scenarios. 

As I mentioned in the previous post, I'm not focussing this year on setting numerical goals for any aerodynamic forces other than simple downforce and drag figures, given we have no data yet to suggest what good targets would be for our vehicle, nor do I expect any other forces to be significant with just a centrally-mounted floor. This is another advantage of aiming small for this first year; total force magnitudes will be kept to manageable levels, so there won't be any significant unexpected issues with general vehicle dynamics or drivability should the aerodynamics perform wildly different to predictions. What's important though is that there is enough magnitude that we can measure and test these effects to find out how they interact with the suspension, general cornering performance, and handling behaviour to give us an idea of how we could improve those areas in future years.