pressure differece

Dear lammps users,
I simulate two reservoirs of water that connected with a nanotube and I want to induce a flow by using pressure difference (pressure-driven flow).

can I use addforce for this puppose? Is it correct? any other suggestion?

I use fix nve in the simulation. with the pressure difference we will have a non-equilibrium system, do we change our ensamble or not?
please suggest some way out…
Thanks in advance
Aisan

Dear lammps users,
I simulate two reservoirs of water that connected with a nanotube and I
want to induce a flow by using pressure difference (pressure-driven flow).
can I use addforce for this puppose?

yes, ​you can induce a flow with fix addforce.

Is it correct? any other suggestion?

​this not a question that can be answered in this general way. what does
"correct" apply to? there are many ways to set up a syntactically correct
simulation and also physically meaningful simulations. but even more ways
to not do that. what is a good setup depends very much on what kind of
information (and how) you want to get from the simulation. whether kind of
boundary conditions you want to use, where and how you want to determine
the pressure differential and what the simulation is supposed to correspond
to in a macroscopic scale. the best way to go about this is to study the
relevant literature and make a survey of what others have done and then try
to figure out what is a meaningful approach and how you can implement it in
LAMMPS.

I use fix nve in the simulation. with the pressure difference we will have

a non-equilibrium system, do we change our ensamble or not?

​it has been stated on this very mailing list, that using fix nve (or nvt
or npt) as such does *not* determine​ what ensemble your simulation is in.
a quick look into a text book on statistical thermodynamics will confirm
that. so in the case of fix nve, *can* get a simulation in NVE ensemble,
but *only* if your system is in equilibrium, has full periodic boundary
conditions and there are *no other* manipulations that would affect volume,
number of particles and potential+kinetic energy.

please suggest some way out...

​the solution out of this dilemma is to study more about statistical
mechanics and thermodynamics.

axel.​

Dear Alex,
thank you for your aswering to my questions. I agree with you about using fix addforce for creating a pressure-driven flow, but my problem is that how can I understand the value of pressure( or what is the relation between pressure drop and the force that I use in addforce command?)
Best Regards for the Best theachers
Aisan

Dear Alex,
thank you for your aswering to my questions. I agree with you about using
fix addforce for creating a pressure-driven flow, but my problem is that
how can I understand the value of pressure( or what is the relation between
pressure drop and the force that I use in addforce command?)

​that is not a question about LAMMPS but about the statistical mechanical
interpretation of your model. as i was hinting at in my previous reply, you
are doing things backwards: rather then worrying about details of how to
realize a particular model, you need to first study and understand ​the
underlying physics and how you can interpret any simulation that you might
want to set up. *this* understanding will then inform you about how to best
model what you want to study and not the other way around.

axel.