Dear All,

I was having a discussion recently about whether CFD has removed the need for experimental testing??

I think not, but my colleague thinks otherwise. Your opinions would be very much appreciated - obviously they need to be sunbstantiated so that I can produce a valid arguement!!

Many thanks

CB

CFD has not removed the need for experimental testing. However, that does not mean that CFD cannot serve as a design tool. In the area I work in (external aerodynamics of aircraft) CFD has progressed to a high level of reliability in certain areas, such as the ability to perform inverse optimization to design a wing from a given lift profile. But prediction of drag is still a questionable area for some people. It is the combination of CFD and experiment that provides the most return on a given investment. We use CFD to help optimize test matrices, so that the dollars spent on testing can be directed at the most important areas that CFD is still weak in. The Integrated Test and Evaluation Association (ITEA) is one source of information concerning such efforts.

I would tend to agree with Greg on this one. Certainly, in external aerodynamics, CFD and testing are used complimentary. Characterising something in CFD is very expensive, compared to wind-tunnel testing. Problem-solving, on the other hand, is where CFD becomes extremely valuable. The thing to work on is to increase the contribution of CFD early in the design cycle. If one could achieve that effectively, you could look at maybe knocking one wind tunnel test cycle out of the design process (at best). That would still leave two or more wind tunnel test cycles.

CFD can can help to design more appropriate experiments, pointing out the critical areas in a design! It will however not replace experiment completely. Experiments can be used to "calibrate" CFD simulations. Its a mutual dependence. Neither can replace the other but together great things can be achieved! The use of experimentation for validation of CFD models cannot be over emphasised.

I agree with all of the above statements, however there are cases (few and far between) that prove that there is no other reliable alternative. An example of this may be looking at Thermal buonacy driven ventilation within a building. To model such an effect would be very difficult in a wind tunnel. Does CFD provide the correct solution to this problem ?? well that is another question all together. Does is providde the best attempt at a correct solution - probably !! This area of CFD or wind tunnel is very problem specific. But validation should always be sought by the CFD engineer.