|
This is a fascinating thread but I've been reading it and thinking most of the way along...
hey!, they're missing something.
Now maybe someone's dismissed this point already and I've missed it, maybe not, but my point seems to be illustrated by several car designs which appear to have steeper tails but claim good drag numbers.
The curves discussed in this thread are typically derived from NACA wing sections, but we're not dealing with a wing. A wing is very long compared to any other dimension, so it can to all intents and purposes be analysed as a 2D structure, each front to back section taken anywhere from the root to the tip of the wing can be thought of as being indistinguishable from the section adjacent to it.
But a car is inherently a 3D structure, the air over the centre line is very different to the air directly over the driver... (the best aero cars taper not just top to bottom, but side to side)
If we take a small but finite section of wing, all the air molecules at it's top edge will need to get back to the very tip of the trailing edge as the wing passes.
However in the simplest 3D case, a tapered dowl, not all the molecules around the fattest section will need to get back to the trailing point, only one need do this, the rest can fill a small tube around this, repeat this for each layer of molecules and the vast majority will not have to rebound terribly far.... in effect the air needs to rebound less in the 3D case than in the 2D thus one would expect higher angles in the idealised curve for a 3D object
A: have I got this all wrong?
B: has someone else already pointed this out?
Clearly if one is generating a simple curved kammback with flat sides then the wing section approach makes sense, but what if we're using complex compound curves.. can we get better performance in a shorter space?
Derek
|