So, I’ve been experimenting around with basic fluid dynamics and wanted to see if I could make a wind tunnel that produces at least semi-realistic results. This is what I got:
The spoiler isn’t always higher than the roof, but the shape of the car is smooth enough so the airflow gets pulled along the surface. In this case, there’s just a sudden drop and the spoiler is stuck in turbulent (aerodynamically dead) air.
It’s a several pass process. Every step, I simulate a set of particles moving relative to the car. The position of the particle is saved in an array. The amount of saved positions in a given volume will then give me a voxelized “pressure map”. I then use that pressure map to apply forces to the particles (force from high to low pressure, high pressure “pulls” particles along with it). I then use the results to do this another 4 times. Afterwards, I simply simulate another set of particles and display lines where they travel. That’s what you see in the screenshots.
Depends on the size of the model. The ones above each took about 10 seconds to simulate.
Most people have the ability to more or less predict where air will flow, without knowing all the formulas. (just like you know how to throw a ball (under the influence of gravity) to land more or less where you want)
