Quote:
Originally Posted by Hersbird
I believe you can get close with a model but you also have to scale the speed down the same. It sill will give a .05-.1 higher Cd even then. Better would simply use the real height and width as it's not a real wind tunnel but a computer simulation.
If it's already over .3 then adding wipers and mirrors is not going to get any better. Only if by "extreme" they mean ditch the fender flairs, keep the tires inside the body and maybe even flatten the roof. The roof may not change the CD much but dropping 3-4 square feet off the frontal aera will be even better.
|
The CFD sofware Tesla uses is full-scale,Full Navier-Stokes equation,large domain spherical vectors,millions of cells analyzed simultaneously,with over 99% accuracy compared to full-scale tunnel testing.A few years ago,with the fastest supercomputers,it took almost 2-days to run a single shape iteration.
I don't know about desktop programs.
One of the CFD online presentations demonstrated full-scale dimensions for the Cybertruck,at 39.34 meters/second velocity.They were trying for accuracy.
Wipers can be hidden below a valence,which shaves 0.005 off the Cd (DARKO)
Mirrors are an issue.My SCCA units cut drag from 0.024 down to 0.010,and frontal area,on the longer 1983 Subaru XT struts.
Again,testing velocity will be a function of dynamic similarity (varisimilitude).
It appears that Tesla is using intentional separation bubbles to sculpt inviscid flow,even captured longitudinal vortices,acting like virtual flying buttresses to actually kill attached vortices,otherwise shedding off the aeroshell portion of the body,as Space Shuttle's, Concorde's delta wings,and the NASA-Vought fly-by-wire, supercritical wing of the early 1970s; tuned with angles,tumblehome, and body rake.