Why can't we apply load along a non-periodic dimension?

Hello Lammps !
I was wondering why it is not possible in LAMMPS to apply a load in along a non-periodic boundary condition when there is no reason for it not to be physically/theoretically plausible. For example, none of the commands like fix deform, displace_atoms, displace_box, fix npt, fix box/relax work on non-perioidc boundaries.
This proves of immense disadvantage when I have to deal with the loading of specimens with non-periodic boundary conditions. Please let me know if there is indeed some method that I am unaware of.

Thanks very much,
Sankar

Hello Lammps !
I was wondering why it is not possible in LAMMPS to apply a load in along a
non-periodic boundary condition when there is no reason for it not to be
physically/theoretically plausible. For example, none of the commands like

sorry, i strongly disagree. this has been discussed
many times on this mailing list. i suggest you read
those discussions in the mailing list archive.

fix deform, displace_atoms, displace_box, fix npt, fix box/relax work on
non-perioidc boundaries.

for a good reason.

This proves of immense disadvantage when I have to deal with the loading of
specimens with non-periodic boundary conditions. Please let me know if there
is indeed some method that I am unaware of.

yes, just use fix move, fix indent, or fix wall.

axel.

How are you imagining "load" would be applied
in a non-periodic dimension? What is pushing the
atoms together, and why aren't they free to simply
relax back apart since there are free surfaces?

Steve

Hi Steve,
I think I should describe the details.
I am trying to shear-load a metal specimen and I want to do it through CG energy minimization without involving any dynamics (I need a rate and temperature independent result). The sample is periodic along its thickness and non-periodic along its lateral boundaries (which are constraints and I cannot change that). The best way is to shear the simulation box (shearing along the thickness is what I want) using displace_box which I cannot but do because LAMMPS requires that not only the thickness but also the other lateral boundary (involved in the shear) be periodic. This condition is actually not a physically necessary one, which is what I actually meant in my first post.

Since the above method is not possible I can resort to using ‘fix move’ or displace_atoms to move an outer boundary layer and thus apply a shear load. But then the problem is how to hold the outer layer static in the new position (fix move is not invoked during minimization); because it will spring back to its original position when I do energy minimization. I think there could be some intelligent way to do this in LAMMPS, but I haven’t been able to figure it out yet.

One method I thought would be to apply force on the boundary layer through say fix aveforce which in fact can work in concurrence with energy minimization. Hence it would help hold the boundary layer in the new position and then I can do an energy minimization. But it is more cumbersome than the ideal method.

I would really appreciate any idea in solving this.

Thank you !
Sankar

The best way is to shear the simulation box (shearing along the thickness is what I want) using displace_box which I cannot but do >because LAMMPS requires that not only the thickness but also the other lateral boundary (involved in the shear) be periodic.

This sounds like you want a triclinic box and to shear it via fix
deform. In the current LAMMPS
version (as of a month ago or so), you can have it be non-periodic in
the lateral dimension.

Steve