UNDER THE HOOD
Pat Symonds on the end for windtunnels
THE END IS COMING FOR WINDTUNNELS
The windtunnel has been an essential tool in the development of F1 cars for around 50 years, and it is said that Brabham used the then-new full scale windtunnel at MIRA as early as 1964. Since the 1980s it has been the primary tool in adding performance to cars. The level of sophistication has developed enormously since those early days.
It might come as a surprise, therefore, that eight of the 10 teams recently agreed that a notional target of eliminating the use of tunnels 10 years from now should be pursued by the FIA. Before discussing what might have led to such a strategy, it is perhaps worth examining how windtunnel testing has developed over the years.
One of my early tasks on joining my first F1 team, Toleman, was to set up and undertake a windtunnel programme. Previously, the car’s design had not benefitted from any aerodynamic research. With the guidance of recruits from other teams we put together a quarter-scale model of the existing car which ran in the tunnels at Southampton University and Imperial College.
These were the only tunnels with a moving ground plane, a device like a large conveyor belt that moved underneath the stationary model and, together with a blown airflow, represented conditions encountered by a car moving over a stationary road and through the air.
The models were unsophisticated. They were made from aluminium and Jelutong: a natural wood from Malaysia that was relatively easy to work but had exceptional dimensional stability. The wheels were machined from nylon and ran on external axles attached to the side of the tunnel, so as not to interfere with the forces that were generated by the model. Forces were measured by a balance device that suspended the model from the roof.
The balance itself was a mechanical device not unlike a weighing machine, able to measure six forces and moments which it did by moving balance weights on a beam until equilibrium was achieved. Pressures were measured in the model by small tappings in the underbody and wings, which were connected to a water manometer. Data acquisition in those days consisted of writing down the forces and photographing the manometer.
Today’s windtunnel is fundamentally the same, but the level of sophistication is beyond recognition. The advent of complicated electronics and computer control has allowed much more numerous and precise measurements to be made. The models themselves have grown from the unsophisticated quarter-scale devices of my early Toleman days to 60% items of precision engineering.
To build a model today from scratch would cost around half a million pounds, such is the level of instrumentation and exactitude. Tyres have evolved from machined simple shapes to actual-scale pneumatic tyres that deflect under load in exactly the way their full-size cousins do, and the load on them is controlled by suspension loading devices.
The tunnels themselves have grown in size, with working sections typically around 15 square meters, plus fans requiring up to 3,000kw to drive the air at full speed. Naturally, this doesn’t come cheaply. Not only would a tunnel probably cost around £50m to build today but running costs, including staff and model production, would add £6m or £7m before you have employed any aerodynamicists. Even the electricity bill for the tunnel would be over £1m a year.
With this level of investment, and the undoubted returns that the experimental process provides to car performance, why would the teams be looking at banning their use? The answer lies perhaps in two areas. Firstly, the costs are unquestionably high. A new team unable to raise the capital required could rent time in a tunnel, but this comes at a cost of around £100,000 a day.
Secondly, it is only a matter of time before someone builds a test facility that overcomes some of the shortcomings of current tunnels. While regulations may limit what can be
done, an ultimate tunnel would encompass the curved flow that a car experiences when rounding a corner on a track. Or it may even be that the model is moved in a large 3D space over a scale replica of the race track. One can imagine that either solution would make the already enormous costs of windtunnel testing minute by comparison.
So, what are the alternatives if tunnels are banned? The answer is Computational Fluid Dynamics (CFD), a science teams are already heavily invested in. Over the last few years, the fidelity of CFD has improved enormously through better computer codes and more powerful hardware, allowing ever more detail to be simulated. The move to cloud computing should allow teams always to be running on the latest hardware, with the only limit on computing power being financial or regulatory.
CFD has advantages and disadvantages. To get accuracy, the virtual models need to be extremely complex. CFD works by solving a large number of simultaneous equations, over 300 million in some cases, therefore giving a large number of solutions which allow the airflow to be inspected in detail at any point on or around the car. This level of insight is simply not available in a tunnel, even with advanced flow visualisation techniques.
The downside is that only one attitude can be examined in one run, making it difficult to produce a full aerodynamic map of the car. Advances in machine learning could transform this, however,
“THE FIDELITY OF CFD HAS IMPROVED ENORMOUSLY THROUGH BETTER COMPUTER CODES AND MORE POWERFUL HARDWARE”
with the ability to deduce a full map from a relatively small number of points a realistic goal.
CFD can take the place of windtunnels, the question being not if but when. Is 10 years optimistic? Time will tell.