CFD Computational Fluid Mechanics CFD has been developed under the spell of Prandtl’s no-slip boundary condition coming with thin turbulent boundary layers at solid walls impossible to resolve computationally and thus requiring walls and turbulence models.
The state-of-the art is represented by Reynolds Average Navier-Stokes RANS and Large Eddy Simulation LES combined with wall models for turbulent no-slip boundary layers. But the design of such models meets severe difficulties and even if they could be overcome computational resolution thin boundary layers is needed.
The result is that high Reynolds number flow beyond the drag crisis (Reynolds number larger than 500.000) is not reached by RANS-LES with foreseeable computer power, with Direct Numerical Simulation DNS without models way beyond the horizon. Since flight concerns high Reynolds number flow beyond the drag crisis with Reynolds number of size for a jumbojet, flight simulation by RANS-LES thus will remain impossible for a long time. To see that this is so, listen to Spalart as one of the chief developers of turbulence models used in RANS-LES.
In this grim perspective DFS opens surprising new possibilities of realistic flight simulation here and now.