Moving a group of atoms at constant velocity

LAMMPS LAMMPS Mailing List Mirror
1

Dear All,

I want to move a group of atoms at one end of my structure at a constant velocity. By following the Lammps documentation, I set the velocity of that atom group to the desired velocity (“velocity set v” command) and turned off their forces by “fix setforce =0”. However, I noticed ( and now understand) that this method only works if “Fix NVE” time integration is used. When I use fix NVT integration the assigned velocity eventually dies off. So, about that I am looking for some advice/insights:

  1. Is there any similar way to achieve and maintain constant velocity using fix NVT for example?

  2. When I assign velocity (along with zeroing the force) and use fix NVT, there is always a sharp drop off the velocity about at the same time-step. No matter what structure size or velocity magnitude I use. What might be the rationale behind this?

Thanks for your kind help.

Best
Shoieb

2

Dear All,

I want to move a group of atoms at one end of my structure at a constant velocity. By following the Lammps documentation, I set the velocity of that atom group to the desired velocity (“velocity set v” command) and turned off their forces by “fix setforce =0”. However, I noticed ( and now understand) that this method only works if “Fix NVE” time integration is used. When I use fix NVT integration the assigned velocity eventually dies off. So, about that I am looking for some advice/insights:

  1. Is there any similar way to achieve and maintain constant velocity using fix NVT for example?

it does not make much sense to include atoms moving at a prescribed velocity into a thermostatted group.
so create a group that does not include those atoms and apply fix nvt only to those and fix nve to others.
e.g. something like this (added to the melt example):

create groups

group move id 1
velocity move set 0.0 0.0 0.01 units box
group temp subtract all move

time integration

fix 1 temp nvt temp 1.5 1.5 1.0
fix 2 move setforce 0.0 0.0 0.0
fix 3 move nve

or - even simpler - use fix move to propagate the particles at fixed velocity. this ignores forces completely, so no fix setforce is needed.

time integration

fix 1 temp nvt temp 1.5 1.5 1.0
fix 2 move move linear 0.0 0.0 0.01 units box

  1. When I assign velocity (along with zeroing the force) and use fix NVT, there is always a sharp drop off the velocity about at the same time-step. No matter what structure size or velocity magnitude I use. What might be the rationale behind this?

impossible to say without knowing more details.

axel.