How do I model Couette Flow in a NVT system with a void space

Dear Lammps users

The simulation system contains a rigid surface and a water film of 5-7 layers molecules sitting on the surface. To avoid unnecessary water-surface interactions caused by PBCs, a void space above the water is used. I am looking a way to model the Couette flow of water in this system. I know that the linear velocity profile of the fluid can be obtained by dragging a wall or used fix nvt/slood and fix deform commands. But I don’t want my box be deformed and I also don’t want the water confined in a channel. I tried to add a velocity paralleling to the surface to water molecules in the upper-most layers. However, the volume of the water is changed due to the void space. Thus, it’s difficult to define the water molecules in the upper most layer. How can I do to get the couette flow in this system?

Any help will be highly appreciated

Thank you all very much

Luo

If I understand correctly, you are modeling an air/water interface, where one face of the water film is exposed to the vacuum. The water film will inevitably deform to minimize its total surface energy.

If that’s not the correct interpretation, can you explain more clearly what you’re trying to model?

Giacomo

Dear Giacomo,
Thanks for your apply
Pls see the attached PNG file. In my system, one face of the water system is exposed to the vacuum, the other side is exposed to a rigid surface. A periodic boundary condition is used at three directions. The dimension of the simulation cell in the direction perpendicular to the surface is suffcientlly large to avoid unnecessary interactions due to the PBC.
system.png
LUO

Dear Giacomo,
Thanks for your apply
Pls see the attached PNG file. In my system, one face of the water system
is exposed to the vacuum, the other side is exposed to a rigid surface.

​please note, that the correct term is "immobile" not "rigid".​

A periodic boundary condition is used at three directions. The
dimension of the simulation cell in the direction perpendicular to the
surface is suffcientlly large to avoid unnecessary interactions due to the
PBC.

please explain, what is supposed to be inducing the flow here?

axel.​

system.png

Dear Axel,

After some thinking, I realized that this system was a little unrealistic. I decided to obtain the linear velocity profile of the fluid by dragging a wall (similar to the case in the Flow subfile). Meanwhile, I want to keep water molecules at constant temperature. Will the fix NVT command still work for a flowing fluid? Do I need to modify the temperature-controlling command?
Thank you for your reply.

LUO

Dear Axel,

After some thinking, I realized that this system was a little unrealistic.
I decided to obtain the linear velocity profile of the fluid by dragging a
wall (similar to the case in the Flow subfile). Meanwhile, I want to keep
water molecules at constant temperature. Will the fix NVT command still
work for a flowing fluid? Do I need to modify the temperature-controlling
command?

you need to subtract out the bias from the induced flow. start reading
here: http://lammps.sandia.gov/doc/Section_howto.html#howto-16

axel.