[mirko]

Hello,

I am looking for a critical assessment of CFD accuracy and utility in aircraft and automotive design (and other industries). For example, how accurate are CFD prediction vs. actual flight/driving performance?

Thus, when aircraft and car designers engineers are designing their product, what error bars are expected in simulations, and at what point do they transfer from CFD to wind tunnel to actual tests?

I am trying to understand how other industries deal with uncertain results from CFD in their product development.

The reason for my question is that in my field (plasma aided manufacture), the predictive capability of simulation is rather poor. Our simulation problems include reactive flow, electromagnetic, and heat transfer effects. Not all the physics is known, and some of what is known cannot be easily modeled. Also, the boundary conditions are often only guessed at: The state of the surface is unknown and changes with time, and the surface chemistry is sometimes poorly known. Finally, not all of the reaction rates (volumetric, surface), are known.

I am trying to understand our limitations in a larger context. It will help me in presenting our simulation limitations to my colleagues and managers.


Thank you,

Mirko