Engine overheat

[Viper2000]

CFD is getting cheaper all the time. GIGO applies of course, but it's pretty good if you use it properly, especially in conjunction with known aircraft performance data.

The biggest limitation at present is the aeroelastic side, because that's much more computationally expensive. But it's very important; for example, a totally rigid CFD model of the Fw-190 would have a much more benign accelerated stall than the real aeroplane because the real thing's wing gets washed in slightly under load, promoting its well documented tendency to flick out.

If Moore's law continues to hold then ever-increasing fidelity will be possible. However, it's not obvious that there will be much of a consumer market for it, given the huge amounts of work (and cost) inherent in the generation of such complex models.