The most commonly known method in testing the behaviour of air around an object is trough wind tunnel testing. The first wind tunnel was already build in 1871 by Frank H. Wenham, a council member of the Aeronautical Society of Great Britain. It was not the first device for measuring the aerodynamic forces this was the whirling arm. However this device had several disadvantages because the object was moving in it’s own wake due to the centrifugal forces.
During world war two the largest wind tunnel of that time was built at Wrigth Field in Ohio. The wind tunnel was 14m long and had a diameter of 6.1m. The air was produced due to two 12m large fans driven by a 40000 hp electric. It was a huge progression because large scale aircrafts could be tested at speeds of 640 km/h.
Nowadays wind tunnels are more integrated in technological research centres. There are several companies who have their own wind tunnel for example Mercedes, but because of the high cost it is not affordable for every company.
This leads us to the domain of CFD (Computational Fluid Dynamics). This method is a branch of fluid mechanics. It uses numerical methods in solving fluid equations like the Navier Stokes equations. It started in 1910 where the calculations were done by hand which led to 2000 iterations per week.
But along with the technological evolution of computers, CFD technology improved a lot. Because of more computational power better results could be achieved and at a higher rate.
In 1985 CFD found his way to the well known “aero” industries like Boeing.
In 1995 CFD was also used for “non-aero” applications like GM, Ford, …
Today CFD is integrated in every company which does aerodynamic research.
It is a relative quick method for having a first impression on what will happen with the body due to the aerodynamic forces. It gives a full domain analyses, easy alternative analyses, it is way cheaper and has a better visualisation of results. The biggest disadvantage of CFD is the fact that it can be erroneous in certain situations, but those situations are mostly well known phenomena. For this reason it is also advised that the software is used by experienced modellers. On the other hand it is not an accepted industry standard. It can only be used on projects which aren’t too complex.
The big question is: Would CFD completely wash away wind tunnel testing in the feature?
Are there other domains where practical experiments are on the edge of being defeated by software?
I think CFD will win the battle in the end. It can defeat the wind tunnel in a few years. But for most of the objects tested in wind tunnels there is large list of safety rules like an aeroplane for example. For this reason I think real modelling with a wind tunnel will always exist, because people won’t take the risk in not simulating the real world.