[lammps-users] displace_box/fix deform error

LAMMPS LAMMPS Mailing List Mirror
1

Dear Lammps users.
I have a equilibrated cubic system and I want to build a new system with twice of its size (volume) and cubic. So I thought of replicated the original system in one of the dimension and then using fix_deform/displace_box command to make it cubic (almost).

While I tried to use both fix_deform and displace_box command, it doesn’t seem to work out.

With displace_box
I used

…
fix 1 all npt 300 300 100 aniso 1 1 1 1 1 1 100 drag 0.5
displace_box all z scale 0.63 x volume y volume units box
…

Here, it seems that in the first step only, it reduces the box to final dimensions and then bring them back to original (initial) one in the later steps.

Should I use displace_box with another fix? I did not find any information (for what fix I can use) in the manual.

With fix deform
With fix_deform, I tried using all fixes (nve, npt, nvt) but the simulation is crashing everytime with following error messages

npt

…
fix 1 all npt 300 300 100 aniso 1 1 1 1 1 1 100 drag 0.5
fix 2 all deform 10 z scale 0.63 x volume y volume units box
…

Here, I get after 10 steps, volume = 0.0 and compute of out of range error : PPPM

nvt/nve

…
fix 1 all (nve/nvt fix)
fix 2 all deform 10 z scale 0.63 x volume y volume units box
…

Here, I get after 10 steps, volume = 0.0 and Error: Cannot use neighbor bins - box size << cutoff.

I am not sure where I am doing things wrong.

In this case also, what fix should I use to calculate forces/positions/velocities of the atoms.

Any help would be highly appreciated.

2

Too many steps at once for me to understand what is going wrong.
Is this a solid or liqud? Is it reasonable to deform it that much?
Did the initial replicate build a good system that you can run
dynamics on? Did you use the replicate command or replicate
tool? Are there bonds in the system?

I suggest you go thru your procedure one step at a time and
see what is breaking down. Try smaller deformations, etc.

Steve

3

Hi Steve, Here are my answers.

Is this a solid or liqud?
It is a liquid of mixture of two types of molecules. The molecular weights are both of them are 178 and 312.

Is it reasonable to deform it that much?
I basically want to change the shape of my system, slowly from a cuboid to a cube (the size distrubtion is 1, 1, 2) and I want it to be (cuberoot(2), cuberoot(2), cuberoot(2)) finally.

Did the initial replicate build a good system that you can rundynamics on?
Yes, If I dont use fix deform or displace box command, the system runs fine

Did you use the replicate command or replicate tool?
I used replicate command in the input script to replicate the system.

Are there bonds in the system?
Yes.

Thanks and regards,
Vikas

4

ok - can you run with a very slow fix deform, or does it crash
immediately? If slow is ok, then it may be problems with
deforming a system with bonds more quickly than it can respond

Steve

5

Hi Steve,
The I tried running as you suggested but it didn’t work. I put timestep to be 0.1 and put fix deform as
fix 2 all deform 10 z scale 0.99 x volume y volume units box
but with so less of a deformation it is giving the same error.

I am attaching the inputfile and start datafile seperately in a different email only to you.

Thanks and regards,
Vikas