Hi
I want to run a simulation for gas flow around a second phase of spherical particles.
atom_style sphere requires that I designate a mass and diameter, but the gas molecules and convecrted particles are not of the same order of magnitude.
Any idea how I can work around this problem?
Best regards,
Ro'ee
Hi
I want to run a simulation for gas flow around a second phase of
spherical particles.
atom_style sphere requires that I designate a mass and diameter, but
the gas molecules and convecrted particles are not of the same order
of magnitude.
Any idea how I can work around this problem?
sorry, nobody here is good at mind reading. your description is far
too vague to give any meaningful advice.
can you at least provide some example input? and then point out where
you see problems?
also, worrying about atom style is only a small part of your setup,
what would be more important to know is the kind of pair_style you
intend to use.
if the sizes of particles are very far apart, it may also be worth
considering to modeling using a hybrid model that uses a continuum
description for the gas, and discrete particles for the reads. check
out the code at CFDEM that couples OpenFOAM with LIGGGHTS, (a LAMMPS
derived code).
axel.
Some more details (though I suspect some of the members do read minds and you're selling them short):
I'm trying to run a simulation of solid particles in an air stream. I want to use the sphere model because of the solid particles, so I have 3 spherical particles (the air is made up of O2 and N2 molecules, the particle is larger).
For reasons that I can't get into at the moment, using a continuum approximation would not be an option.
As the gas is ideal, I would also simply use infinite potential at skin, zero beyond.
So is there an option to do what I'm trying to? use the sphere model for 3 sizes of particles, and if not, could you suggest an alternative?
All the best,
Ro'ee
Quoting Axel Kohlmeyer <[email protected]>:
Some more details (though I suspect some of the members do read minds and
you're selling them short):
that is just guessing based on previous experiences. the problem is
that guesses can be wrong and it is not needed to guess if the person
that asks for help does a better job in describing the problem. i
think that is just a fair request. after all, you will get an answer
for free and if you provide more information, the answer is likely
more accurate and finally, you cannot fault people for wanting to give
good answers instead of half-assed guesses.
I'm trying to run a simulation of solid particles in an air stream. I want
to use the sphere model because of the solid particles, so I have 3
spherical particles (the air is made up of O2 and N2 molecules, the particle
is larger).
that is not a convincing argument. you could do that also using the
rigid particle integrator and build the large particle from atoms.
For reasons that I can't get into at the moment, using a continuum
approximation would not be an option.
As the gas is ideal, I would also simply use infinite potential at skin,
zero beyond.
LAMMPS uses verlet like integrators and those are not a good choice
for integrating hard spheres. you either have to change your model or
pick a different MD code.
So is there an option to do what I'm trying to? use the sphere model for 3
sizes of particles, and if not, could you suggest an alternative?
please let me re-iterate that using an atom style sphere doesn't
automatically result in those particles being treated as extended
particles. if you use an interaction model that uses point particles,
the sphere radius will be simply ignored. only granular particles and
a few special cases use the diameter property of atom style sphere.
you also may need to use an adapted integrator.
so, you either need to spend some more time with the LAMMPS manual and
get a better sense for what can be done with it and how, or look for a
different code that can do what you want (or write one).
since you choose to not go into details, i will now choose not to
respond any further.
good luck,
axel.
From your response it appears I've offended you in some way (whih is pretty perplexing for me). That was not my intention.
Thank you for the information you did see fit to convey.
Quoting Axel Kohlmeyer <[email protected]>:
From your response it appears I've offended you in some way (whih is pretty
perplexing for me). That was not my intention.
i am not offended, it is just my personal choice.
please note that this is not a personal conversation and you didn't
hire me as your personal consultant. so the rules of the game are
different. since this mailing list is archived and people often search
the archive for solutions to problems similar to ones that are were
already discussed (and we frequently recommend people to do that, so
we don't have to answer the same questions over and over again), i
care about giving the best possible answer and for that i request a
full disclosure. while it may be adequate for you personally, if you
get a vague advice on a vague description, it may be misleading for
others and then this will reflect badly on me and not you.
i understand that sometimes full disclosure is not possible, but since
i volunteer my time here (as does everybody else), i expect that
people respect my choice to not respond beyond a certain point.
that is it. no hard feelings. just a statement of fact.
axel.
Atom_style sphere lets you specify a size for every
individual particle (diameter). So the multiple sizes
is not a problem. However, as Axel has said, you
need to use a pair style that actually uses the size
of the particle as specified by atom_style sphere.
The granular and colloidal pair styles
do, but most don’t. E.g. the LJ pair styles use the
sigma of the particle types, which are specified
by the pair_coeff command.
Steve