[chillsworld]

Originally Posted by aerohead

*The same year Song/Zhang/Wang/Hu published,Martin Olsson's Masters Thesis was also published at Chalmers University,Goteberg,Sweden.
Olsson stated that at the time,only Direct Numerical Simulation could accurately predict 3-D flow of an automobile,and that it was "not today possible (lack of computer power)."
Olsson spoke of "incertainties,deviations,underestimations,overest imations,suspicious results,rerun (results) totally different.
He summed up,"due to the uncertainties in the CFD simulations of a few drag counts which make it impossible to trust all findings."
*'actual ' results did not correlate with CFD results.
FLUENT solves the Reynolds-averaged Navier-Stokes (RANS) which cannot predict all turbulence values.
Olsson was getting 45-drag count deviations (13%) which exceed the drag claims of 10.3% for the non-uniform surface.
Olsson recommended full-scale wind tunnel investigations so as to actually learn something.
-------------------------------------------------------------------------
Nathan A. Williams basically came to the same conclusion about CFD in his Masters Thesis of 2003.'desktop' CFD gives one insight and graphical tools to think about to steer the actual wind tunnel work.
--------------------------------------------------------------------------
I don't think Song et all knew what they didn't know and were carried along with presumptions of accuracy.They didn't even have the presence of mind to investigate the viability of the software they were using.
Personaly,I'm gonna wait for bigger dogs to weigh in before I can embrace what's being passed off for knowledge.

Well crap...



Maybe the plane people will have good results and inspire real world testing on automobiles some day.

~C