|
 |
To: 9. Discussion |
The purpose of the computations in the above sections
was to verify the unprecedented combination of
accuracy, reliability and flexibility of the
vorticity redistribution method. The mesh-free property
is the most dramatic advantage of our method for the computation of
separated flows around complex geometries.
The random walk method has this property too, but it can clearly not compete
with our accuracy.
And despite their reduced flexibility,
the particle methods cannot match the
accuracy and reliability of our method for the considered flows.
Surprisingly, our results for the impulsively translated cylinder at
also seem to be better than the finite difference results, and
as good or better than the spectral element results. Yet, this is still
a very simple geometry, and the separated vorticity remains relatively
close to the wall.
In the next chapter, we will discuss some advantages of our method from a purely theoretical point of view.