Simulating supersonic flight in game So this question about how to implement a good looking simulation of super-hypersonic flight of a rocket from launch point to orbit.
I have an idea how to implement a good looking aerodynamic on sub-sonic speeds. I want to implement a virtual aerotube. I want to create some simulation (based on papers which implementing  incompressive fluid solvers) which takes pressures on surface of rigid body in that simulation to calculate:


*

*Pressure center.

*Drag and lift coeficients.


For each possible angle of attack for the object. Store this data in object properties and, then calculate forces depending in this data, when the simulation will be played.
But.
When I implemented first incompressive solver, I started to collect data, how to implement compressive solver. And this staff, surprisingly became even more complex, than I could handle.
Today, I still thinking of two questions:


*

*Is there's some sophisticated scientific document, which could help me to get some idea of how to implement that staff(compressive flow), which I maybe lost in tons of docs, I've read already not-that-attentively-as-needed.


Or.


*Is there some method, which helps to get not-the-same-as-real-but-very-realistic-picture-in-some-conditions, when I able to use coeficients and pressure centers, that i got from my sub-sonic simulation? 

 A: As I mentioned in my comments, there are many approaches you can take here, and full-on Navier-Stokes is not something you want to attempt. Either incompressible or compressible. Fully solving the Navier-Stokes is very, very far from real-time capable. Even for the most basic problems I can think of. 
You said you have incompressible figured out. That's great! What method are you using? How does it work? Odds are good, whatever you are using in incompressible has an analogue for compressible. And it is much easier to solve compressible problems than incompressible ones. If you have low-Mach figured out, then larger Mach numbers are much easier to actually solve.
Your best bet is to look at various potential flow methods. You can leave them linear, or use the non-linear version. You can keep them steady or unsteady. There are forms for subsonic, trans-sonic and supersonic. The main drawback is that they cannot predict viscous forces. So no turbulence, no skin friction drag, no heat conduction to the surface. For trans-sonic and supersonic though, they can predict wave-drag due to shock waves... so you at least get something there. 
The best and likely fastest approach is panel methods. Take a look at this NASA review to get an idea of how they work. You would cover your body with panels and solve the linearized potential equations. This is helpful because your body is already covered with panels! The polygons used by your 3D engine to render the body are suitable for use as the panels! You would then just need to measure angle of attack (orientation of the panel relative to the body's velocity vector) and you're off running.
This will give you $C_p$ and $C_l$. If you also need drag coefficients, you are better off using some other approximate model for drag. You could define lift-over-drag for all your parts up front, then with $C_l$ you can get a $C_d$. This can be a function of angle of attack. Or you can look for other drag models. You'd have to be more specific on what you want to do with that.
