Hello everybody, as an aerospace engineer what would you do if somebody asked to describe all available methods used for evaluating a new design of airplane in terms of aerodynamic performance? some comparisons between those are more than welcome.. Is there the best method or it is a combination of different ones? why? please, give me some hints and illuminations! anyway, I wish you all a great weekend!

;-)

best

I know of three general methods - ground test, flight test, and analytical/computational test. Of the three, flight test yields the best quality of information (no surprise), but is also the costliest and riskiest. Ground test (e.g. wind tunnel testing) can yield substantial high-quality information, but scaling the data up to flight conditions can be an issue. Ground testing is also expensive (though not as bad as flight testing). Computational/analytical testing yields data whose quality may be all over the map, but it is relatively cheap. In the right hands a good CFD analysis can shave hundreds of thousands of dollars off the cost of a ground test, because multiple scenarios can be run for the cost of acquiring a single data point in a wind tunnel. The qualified use of semi-empirical analytical tools can also provide significant data for the skilled user. Generally speaking, the ideal method combines all three - computational techniques early in the life cycle to establish the rough design parameters, some ground-testing to validate early design decisions and validate the computational analysis, use of the validated computational tools to further refine the design, some more ground testing as required to address issues that the computational techniqes can't resolve, and finally flight testing to work out the bugs in a real-world setting. Try searching for "integrated test and evaluation". That's the catch phrase that sums up what I've just described. Basically it boils down to using the right combination to minimize cost without compromising the end result.

Ag has written a very good explanation, but the statement "because multiple scenarios can be run for the cost of acquiring a single data point in a wind tunnel" is only partially true. In most aircraft development programs there will be wind tunnel testing anyway, and adding an extra polar to the test matrix costs a fraction of what CFD would cost. Nevertheless, using all these methods in complementary fashion is what it is all about, and CFD can do much to illuminate experimental results, and under the right conditions, can save a lot of experimental work.