In our previous articles we discussed the role and importance of determining sound performance metrics which define what success means within an aerodynamic development cycle. In this article we will shift focus to the modelling aspect.
The three more traditional approaches for modelling the aerodynamic performance are wind tunnel testing, computational simulations, and field testing. A fourth approach, more recent and still in consolidation, uses AI-based techniques. The role of AI will have its own separate article in this series, we will therefore leave it out from this discussion for now.
Wind tunnel testing has been used for aerodynamic development for many decades. It is a well-established and well understood, stable, controlled, and repeatable environment for engineers to incrementally test, understand and improve the aerodynamic performance of their product or application. Subject to the application, operating conditions, and scale, it can represent the real-life application very accurately. In the automotive and motorsport environments it tends to represent relatively mild operating conditions such as straight-line with mild crosswind very well, however the quality of its representation deteriorates with increased crosswind and/or lower cornering radius.
Computational simulations of aerodynamic performance have progressed substantially over the last 10-15 years. The increased availability and lower cost of high-performance computing hardware means we can now afford to simulate aerodynamic flows with much higher fidelity levels. Like the wind tunnel, when the process is adequately designed it also provides a stable, controlled, and repeatable environment for development.
While CFD still has limitations when comparing to wind tunnel testing, it also has many advantages in its ability to better model the real-life operating conditions. It can much more adequately model cornering conditions, as well as geometry shape changes (i.e. tyre deflection during cornering, bodywork aeroelasticity), ground and tyres roughness and others.
Field testing has the advantage of being the closest thing to reality. It is typically done using a prototype representing the final product (car, aircraft), thus removing any geometrical modelling approximations which might be required in the wind tunnel and CFD. On the other hand, the real-life environment tends to be less stable and repeatable, making the judgement of the aerodynamic performance significantly more challenging.
In motorsport straight-line testing is a widely used strategy as it strikes a compromise between testing the real car while retaining some level of repeatability. Full track testing can be successfully used to determine performance deltas between aerodynamic configurations but is very complex and requires a holistic approach in terms of using several flow, load and other car measurements, lap time and driver feedback to unpick the various variables affecting the test.
Our philosophy at Sabe is that all three approaches should, when possible, be used in a complementary manner. Each of them has their own strength and weaknesses. We often compare the task of judging aerodynamic performance to of a detective investigating a crime: we don’t simply have “faith” in the results coming from one tool or another, instead we collect pieces of performance evidence of different types and from multiple sources, we combine those with our experience and engineering judgement to make an assessment of the expected performance of a given application in real life.
In the next article we will discuss methodologies to measure aerodynamic performance.