next »

[diegzumillo]

Hiall

I already made in the past a super simple flight sim but now I want to revisit that project with a new mindset.

Here's the thing. Aerodynamics is complicated and a believable implementation flexible enough to handle all kinds of flying vehicles is computationally demanding. BUT, what if I want a believable implementation of a single flying vehicle? all I would really need is a set of curves that determine lift and drag for all possible angles and velocities.

This sounds too good to be true. Can anyone spot the mistake in my logic? would all this data take much more space than I think it will?

If that is that good and true, well, that sounds like the kind of thing that already exists somewhere, but I can't find anything like it, just simple plots from flying courses and stuff like that. If not I guess I could code a tunnel wind kind of thing and just calculate all the parameters for all the cases I want. Since this is a game I can fill in the blanks with lazy interpolations and extrapolations, as long as it feels real.

Appreciate any thoughts and comments!

[diegzumillo]

First thing I noticed after I posted. A set of values for a full aircraft can't be of just lift and drag, but also torques. I can simplify that by having a table for parts (lift and drag are still vectors though), then make the vehicle out of parts and let the physics handle the natural torques.

It does seem more complicated than I thought.

[diegzumillo]

I did a little estimation here and things look bad. There are 9 parameters (3 angles, 3 components of flow velocity u, density rho, mass m) and for each set we want a force and torque (6 numbers). Now it's a matter of finding a good number of interpolations of these parameters that make this lookup table feasible.
size of table: (N angles)^3 x (N u)^3 x (N rho) x (N m) x 6
It makes sense to get more details for the angles so N angles =50, and let's make the rest 10. Size of table = 75 billion numbers. 

I need a different method.