Hi
I am still looking for a solution described in my previous email.
Thanks in advance.
Dundar.Hi ,
I would like to simulate a constant flow of molecules through to
simulation box.
Suppose that molecules are entering the simulation box at xz plane
located at y=0 and want to
assign drag force/velocity in y direction. Molecules which move through
the xz plane located at
y=yhi should be deleted. Is there any way to do this with Lammps?
Thanks.
Dundar Yilmaz
Hello,
What I meant is that the standard way to simulate confined flows is to
impose PBC in the direction of the flow, so that the molecules going
out from the y=yhi plane will be the very same entering the y=0 plane.
You can / should have a look at the huge literature on this kind of
problems and check that your way has never been used before. I can't
think of any potential advantage of your method, but I can see some
drawbacks: namely it is as if you were connecting the y=yhi plane to a
large volume of void, and your fluid will quickly evaporate, so I
don't see how you can get a stationary flow. Can you explain why you
want to delete the atoms crossing the y=yhi plane?
But anyway, if you really want to try, there is nothing simpler: just
use boundary f, so that any atom crossing the boundary will be lost,
and thermo_modify lost ignore so that lammps won't complain about it.
Best,
Laurent
2010/8/3 Dundar Yilmaz <[email protected]...>:
You can delete atoms when they exit a box
with a fixed boundary (f in boundary command),
but you can't insert atoms as you ask at another
boundary. That would require a new fix that
was smart about when/where to insert them. Something
like the fix pour command. If you want to use
periodic boundaries then atoms that exit will
reappear on the other end. That is how many people
model continuous flow. See examples/flow/in.flow
for a simple 2d example.
Steve