# Fix move command

Dear all

My problem is, i would like to update position of a group (wall) in one direction, in each time step, independent of its own motion due to force interactions in all directions. It means they are in the time integration of specified ensemble for all directions, but i want to also add a constant position change in addition to their normal displacement.
Is there any easy way of doing this?
I started to use fix move command with style of variable.
Could someone can make it more clear the explanation of move command in manual as,

“If only the velocity variable for a component is specified as NULL,
then the displacement variable will be used to set the position of the atom,
and its velocity component will not be changed.
If only the displacement variable for a component is specified as NULL,
then the velocity variable will be used to set the velocity of the atom,
and the position of the atom will be time integrated using that velocity.”

At the first statement, the new position of atom depends on only the variable i set, then what is the point of unchanged velocity if it is not used to update position.
At the second statement, time integration only uses the velocity variable i assigned, then what is the point of setting the displacement as NULL. It becomes same with fix move command with style of linear.

Thanks

Murat Barisik
Research Assistant
Mechanical & Aerospace Engineering
Old Dominion University

Dear all

My problem is, i would like to update position of a group (wall) in one
direction, in each time step, independent of its own motion due to force
interactions in all directions. It means they are in the time integration of
specified ensemble for all directions, but i want to also add a constant
position change in addition to their normal displacement.
Is there any easy way of doing this?

try using "fix smd". that will add a force constraint to the group
and the tethering point for that constraint is moving with a fixed velocity.
(don't get confused by r0, just set it to 0.0).

cheers,
axel.

Either you want the x coord of a wall atom to be exactly where
you specify it on each timestep (in which case use fix move),
or you want it to move according to forces it feels from
other atoms (in which case just integrate if with fix nve).
You can't have both.

If you want to add a force so that the total force on the
wall atom is a combination of both (and then integrate
it with fix nve), then do what Axel says, or use fix addforce
or fix aveforce or the like.

Steve

Could you please explain me following statements from LAMMPS manual.

1. If only the velocity variable for a component is specified as NULL,
then the displacement variable will be used to set the position of the atom,
and its velocity component will not be changed.

What is the point of unchanged velocity? Is it equal to zero?

If only the displacement variable for a component is specified as NULL,
then the velocity variable will be used to set the velocity of the atom,
and the position of the atom will be time integrated using that velocity.

What is the point of setting the displacement as NULL since It becomes same with fix move command with style of linear?

Thanks

Murat Barisik

Steve

Could you please explain me following statements from LAMMPS manual.

1) If only the velocity variable for a component is specified as NULL,
then the displacement variable will be used to set the position of the atom,
and its velocity component will not be changed.

What is the point of unchanged velocity? Is it equal to zero?

No, it is "unchanged". So if it is non-zero, it remains non-zero.
The position is set to what you asked it to be. The velocity is
not used.

If only the displacement variable for a component is specified as NULL,
then the velocity variable will be used to set the velocity of the atom,
and the position of the atom will be time integrated using that velocity.

What is the point of setting the displacement as NULL since It becomes same
with fix move command with style of linear?

No it doesn't. If the velocity is a variable, then it can have some funky
time dependence, and the position will be integrated from that time-dependent
velocity.