In the last two articles, we introduced wind tunnels and CFD as predictive tools used in motorsport. In this article, we will go deeper into the discussion about their strengths and limitations and their role in aerodynamic development.
In the first decade of this century, wind tunnels were considered by many as the ground truth for motorsport aerodynamic performance. However, CFD as a whole (software and hardware availability) was not up to the level required for the application, and the aerodynamic performance comparisons between track and experiments were somehow limited, which in turn reinforced this blind faith in the wind tunnel as the tool which provides the correct answers.
Correlation exercises were limited by push-rod/pull-rod load measurements in straight-line constant speed track testing, a few pressure measurements, lap time and driver’s feedback. These last two were highly affected by many other variables.
Over the last decade, we have seen the development of new techniques for aerodynamic performance measurements on track which, combined with improvements in CFD techniques and hardware availability, started to paint a very different picture about the quality of the data from a wind tunnel.
While a wind tunnel is a powerful and repeatable tool for development, as discussed in a previous article, it is ultimately an approximation of reality. The data generated is subjected to a different model scale, different Reynolds number, geometrical approximations (i.e. parts with thicker trailing edges due to manufacturing limitations at the scale required), parts deforming differently under load, different tyre shapes, ground roughness, lack of flow curvature, approximations regarding internal blockages and many others.
CFD, on the other hand, can more easily address discrepancies due to scale, geometrical representation, surface deformation (if known), tyre deformation (if known), ground roughness and flow curvature; however, it introduces other modelling errors in the process.
Clearly, wind tunnel and CFD are complementary tools with different strengths and weaknesses, which we advocate should be used in combination for aerodynamic development. Although the higher-end motorsport categories moved away from considering wind tunnels as their primary development tool, nowadays, CFD has a much more active role in all development steps.
CFD is still heavily used as a preliminary conceptual analysis tool and a “filter” for wind tunnel development. But its role has expanded considerably; it provides a different perspective on the aerodynamic performance of a component compared to the wind tunnel and is a factor in the performance metrics.
It also serves as a very powerful “correlation/comparison bridge”, acting as an effective tool to decouple variables affecting a given comparison. For example, CFD can be used to subtract the effect of a wind change on the aerodynamic performance across the comparison between two different front wings. Or as another example, it can be used to isolate the difference between the tarmac and the wind tunnel belt roughness.
Another more recent utilization of CFD is for the generation of training data for AI-based methods, an area Sabe is actively researching and will be the subject of our next article.
The bottom line is that both wind tunnels and CFD are approximations of reality in their own ways; each has strengths and weaknesses, and both should be simply seen as tools for aerodynamic development.