Respected Lammps-Users,
I have a system of granular atoms to which i have set a dipole moment and mass. The atoms are attracted towards one another (Due to the dipole moments) and would collide.
My query is as follows:
-
How would I assign an ‘event’ which is defined as the collision of two atoms such that i could note the time and coordinates of atoms during the collision.
-
I would want the two atoms that collided to be treated as a single granular sphere following the collision with the mass and dipole moments of the new sphere equal to the summation of the masses and dipole moments of smaller colliding spheres.
Help is deeply appreciated. And I would be grateful.
Sincerely,
Ashwin Padmanabhan
IIT Bombay, India
It sounds like you are saying you want to replace two touching
small spheres (just collided) with one large sphere. To do that
you’d have to write new code in LAMMPS that replaced 2 particles with
one large particle. If the large particle is a sphere, it takes up more
volume than the original two small particles. So there wouldn’t be
room for it in a dense granular packing.
So to me, your model does not make much sense. Regardless, to
implement it, you would have to do some careful new coding within
LAMMPS.
Steve
Respected Dr. Plimpton,
I deeply appreciate the help.
It seems to me, considering your reply that such a system doesnt make much sense.
What could be an alternative solution to this problem such that the collided system has a higher force due to gravity (because of the combined masses of the two colliding spheres) and modified dipole moment so that it responds differently to other dipoles in the system?
Thanking you,
Regards.
Ashwin Padmanabhan
Respected Sir,
I think this would answer my all my queries- Is coalescence of two particles possible in Lammps under the Granular model?
Im sorry if these questions seem very basic.
I deeply apologize if I have offended you in anyway.
Regards.
There are a couple options, none of which may
be exactly what you want:
a) there are granular potentials that include an attractive
force (like JKR potential), so that particles will “stick”
together when they collide. We have one of these
internally, but it hasn’t yet been released in LAMMPS.
There may be one in LIGGGHTS which is built on top
of LAMMPS, and you can google for it.
b) LAMMPS can model rigid bodies composed of
multiple granular particles, e.g. dimers. But you
need to start with a set of rigid bodies, or pour them
into a box, etc. See the fix rigid/small and fix pour commands.
There is no option currently in LAMMPS to convert colliding
particles into rigid bodies on the fly.
I deeply apologize if I have offended you in anyway.
you haven’t. You’re just asking Qs and getting advice.
Now you have some work to do, to figure out an appropriate model.
Steve
There are a couple options, none of which may
be exactly what you want:
a) there are granular potentials that include an attractive
force (like JKR potential), so that particles will "stick"
together when they collide. We have one of these
internally, but it hasn't yet been released in LAMMPS.
There may be one in LIGGGHTS which is built on top
of LAMMPS, and you can google for it.
b) LAMMPS can model rigid bodies composed of
multiple granular particles, e.g. dimers. But you
need to start with a set of rigid bodies, or pour them
into a box, etc. See the fix rigid/small and fix pour commands.
There is no option currently in LAMMPS to convert colliding
particles into rigid bodies on the fly.
perhaps not rigid, but to have particles stick together and not come apart
could be modeled with adding (harmonic) bonds using fix bond/create.
without source code modifications, that would only work for a monodisperse
system, and it would make the input quite complex and tricky (use atom
style hybrid with bond added; reserve space for a sufficient number of bond
types, bonds per atom, specials per atom; use special_bonds lj/coul 1.0 1.0
1.0 and possibly more tweaks).
i have a vague recollection, that something like this was discussed on this
mailing list before (w/o dipoles), so perhaps a careful dig into the
lammps-users mailing list archives might help.
axel.
Respected Dr. Plimpton and Dr. Kohlmeyer,
Thank you so much for the help. I deeply appreciate it.
Warmest Regards,
Ashwin Padmanabhan.