Hi,
This is my first time I’m attempting to run a molecular dynamics
simulation, so please excuse me if I ask questions which might appear
trivial to you.
don't do such a complex simulation as your first foray into MD. start with
simpler, already published, problems and learn how to reproduce them. thus
you can build your skills and gather required practical experience before
starting a simulation on a new and complex problem. otherwise you are
setting yourself up to needlessly repeat many beginner mistakes and wasting
a lot of time, since you'll never know whether unexpected behavior is
caused by the physics of your problems or mistakes of your setup.
I’m trying to simulate the squeeze film effect between two disks whose
diameter is 1.9mm. The disks are separated by a distance of 30-500 nm. The
gas in the gap between the disks is Argon, and the ambient (=initial)
pressure is approximately 20mbar.
While the bottom disk is fixed, I’d like to apply a force/displacement to
the top disk and measure its displacement/force acting on it, accordingly.
LAMMPS has options for each, you can use, e.g. fix move for displacement,
or fix addforce or fix aveforce for adding a force. similarly, you can
define groups of atoms and access the center of mass location or forces via
the corresponding variable functions or for other properties, set up a
suitable compute reduce.
- I did notice that in most cases the boundary conditions that are used in
LAMMPS are periodic bcs. In my case since there is flow in the radial
direction, I can’t use these boundary conditions, and thus I was wondering
how am I supposed to prescribe the ambient pressure on the circumference
(diameter=1.9mm) of the control volume which represents the gap?
if you don't have periodic boundaries _and_ want to contain atoms, you'll
have to have some kind of wall. these can be realized with wall potentials
via a fix wall variant, or with explicit, immobilized particles.
- How accurate is it to model the problem as a 2D problem, and does LAMMPS
support axisymmetric boundary conditions?
sufficient accuracy is in the eye of the beholder. LAMMPS supports the
boundary conditions mentioned on the boundary command documentation. no
more, no less.
please note, that most of your questions are not really LAMMPS questions,
but questions about how to realize your simulation and thus could be
applied to any MD code. therefore, you should keep in mind, that it is
_your_ job to ensure, that everything you are doing is justifiable and
meaningful. people here may give an opinion, if it is easy, but they
cannot fill in for having competent personal advice and supervision, as
required if you are new to a field.
axel.