Consider a molecule of diameter that moves over one free path
length
. During that motion it will hit another molecule
if the center of the other molecule is within a radius
from the
path of the molecule. In other words, the center of the other
molecule must be inside a cylinder of radius
around the path
of the first molecule. There should be about one
collision in a free path, so there should be about one other
molecule within the cylinder. So the free path can be ballparked
from setting the volume of the cylinder equal to the average volume
per particle:
pointwisedensity and velocity at a given time? (Here
pointwisemeans using small volumes much smaller than
pointwisedensity and velocity that would be enough info, say, to compute the further flow development? In particular, take sea-level air. Then for what body size
pointsaround the body? For what size body can you no longer find a meaningful continuum velocity for the
pointsaround the body even if you used molecular dynamics?