Yes and no.scarbs wrote:Rotating wheels have been ignored as they are too complex to model accurately, as I understand the flow around a rotating wheel is produces an unsteady wake, thus the flow pulses in and out. Conventional time averaged CFD could not model this properly.
Conventional Reynolds-Averaged Navier Stokes (RANS) CFD using turbulence models (typically eddy viscosity) can relatively accurately simulate steady and unsteady flows around a car, including rotating wheels.
Steady solutions are preferred (to reduce the computing costs), but if the physics are unsteady (say due to vortex shedding from wheels) you are unlikely to get a converged (valid) solution. So I guess in most cases there isn't vortex shedding from wheels on the ground otherwise we wouldn't see so many CFD solutions for F1.
If there are unsteady flow features then RANS CFD should be adequate to simulate it. It will mean running the solver in time accurate mode and collating the (massive amount of) data from each time step.
The weakness of RANS CFD is the turbulence model and its ability to predict separation especially from smooth surfaces (such as a wheel). The alternative approach is to use Large Eddy Simulation (LES), which has to run in unsteady (time dependent) mode and requires probably an order of magnitude more cells around the car to resolve 'large eddies' (they are actually quite small). For a full car this type of simulation is probably unfeasible even with the massive iron deployed in F1.
Hence yes and no.
Remember though most F1 teams have at least 1 wind tunnel - I'd be surprised if they aren't used to delve into wheel aerodynamics. CFD vs wind tunnel isn't a zero sum game.
