Flow through Pressure Difference

Hello
I am a student and recently started working on Molecular Dynamics and LAMMPS. I want to know if there is a way I can simulate(2d) flow through a pipe or nozzle having pressure of 10 bar on one side and 2 bar on other. I tried joining two barostats using NPT and press/berendsen (two different files…not in same simluation) around the pipe or the nozzle. But I am not sure if this is working.
Please suggest some way out…
Thanks

Tapan K Mankodi
IIT B / SVNIT

There are many ways to induce a flow in molecular dynamics, and before
running any simulation you should search thoroughly the literature and
think of the best method in your case.

Here are a very few examples among many others:

A pressure difference between two reservoirs can be established using
a reflecting particle membrane:
http://prl.aps.org/abstract/PRL/v102/i18/e184502

Osmotic pressure can also be used:
http://www.pnas.org/content/100/18/10175.abstract

You can also have a look at the fluidized piston model.

But finally, the easier way to induce a flow is not through a pressure
difference but rather with a gravity-like force (fix addforce), which
is formally equivalent to the effect of a pressure gradient, and much
easier to implement with periodic boundary conditions.

Best,
Laurent

Hello
I am a student and recently started working on Molecular Dynamics and
LAMMPS. I want to know if there is a way I can simulate(2d) flow through a
pipe or nozzle having pressure of 10 bar on one side and 2 bar on other. I
tried joining two barostats using NPT and press/berendsen (two different
files..not in same simluation) around the pipe or the nozzle. But I am not
sure if this is working.

please note that fix npt is meant to simulate a system in equilibrium,
but you are describing a non-equilibrium system with (i presume)
a continuous flow through the nozzle.

cheers,
    axel.

I did simulate the fluid flow through nozzle using fix addforce statement.
But i have one problem…
once the molecule or atom passes through a nozzle geometry and crosses the boundary conditions of the simulation box. When it encounters the nozzle again, the simulation would have a different trajectory when compared to trajectory on atoms which has not yet encountered a nozzle…in short what i need is a continuous supply of flow atoms from one side and then go through the nozzle and then go to a sink…

Is it possible to simulate that way?? and that was the reason i thought of having npt fixes at the start and sink which will bring the flow atoms back to the initial conditions(as in conditions before they come across a nozzle)

Tapan K Mankodi
IITB SVNIT

what i
need is a continuous supply of flow atoms from one side and then go through
the nozzle and then go to a sink...

Is it possible to simulate that way??

Not unless you write a fix to continuously create atoms at one side. You could
look at fix pour or fix deposit and see if they do something similar to what you
want. Or you could write a fix that suitably randomizes some attributes of
a particle (e.g. it's velocity vector) when it crosses the periodic
boundary. That
way you could re-use particles.

Steve