Structure relaxation and energy minimization

LAMMPS LAMMPS Mailing List Mirror
1

I am trying to do structure relaxation and energy minimization by using Reax, eam and meam poetntials for nanorods (you can see input and data files in attachment).

Before I did both relaxation and minimization for very similar kind structures but for this I am not getting any results, its still running even after 30 hours.

units real

atom_style charge
read_data Nanorod5171 # the input structure

The force field

pair_style reax/c lammps_control
pair_coeff * * ffield Au #number of elements inside the file
fix 9 all qeq/reax 1 0.0 10.0 1.0e-6 reax/c #charge equilibration

optimize box size

fix 1 all box/relax iso 0.0 vmax 0.001

output of structures

dump 1 all xyz 100000 dump.pdef

frequency of energy… output + kind of output

thermo_style custom step pe density press lx ly lz
thermo 10

Minimize

min_style cg # (check it in the manual)

minimize 0.0 1.0e-8 100000 10000000

What could be wrong?

I would really apperciate your comments/suggestions.

Thanks!

input-EAM (517 Bytes)

Nanorod5171 (180 KB)

2

I am trying to do structure relaxation and energy minimization by using
Reax, eam and meam poetntials for nanorods (you can see input and data files
in attachment).

Before I did both relaxation and minimization for very similar kind
structures but for this I am not getting any results, its still running even
after 30 hours.
units real

[...]

What could be wrong?

have you discussed this with your adviser? that should be the first
person to address. after all, this mostly concerns your research and
not any technical issue with LAMMPS. if you believe LAMMPS is at
fault, you need to produce some more convincing evidence.

axel.

3

Hi Axel,

Thanks for the reply!

Yes, I already discussed with my supervisor and he suggested to use different ftol (force units) range i.e. 10^-4 or 10^-6 instead of 10^-8 but the problem remains.

Why I asked here because I did (relaxation and minimization) for very similar kind structures with the very same input description but this time its not working. I think there is problem with criteria selction for minimization so someone can suggest with his experience, it would be great.

Thanks!

4

Does it converge for a very loose tolerance?

Steve

5

Hi Steve,

Thanks a lot for reply!

I used 10^-8 previously and it worked for all the structures but now it is not working. One question how loose tolerance I can use, is there any limit?

Thanks!

6

It’s just an input, you can use any value you choose.

Whether it makes sense for what you want to do
is up to you. But it will generally be easier
for the minimizer to converge with a low tolerance.

Steve

7

It’s just an input, you can use any value you choose.
Whether it makes sense for what you want to do
is up to you. But it will generally be easier
for the minimizer to converge with a low tolerance.

And that on turn may be an indication that something else is out of whack. However, that is impossible to say in general. You will have to investigate and try to understand.

8

Thanks!

I have another question, when I run another structure with the similar input it runs OK in start but after 870 iterations it shows following warning:

WARNING: Fix qeq/reax CG convergence failed after 200 iterations at 881 step (…/fix_qeq_reax.cpp:648)

and then very strange values for PE fmax press and density.

What could be reason of this warning?

Thanks!

9

This usually indicates you have a bad dynamics due to bad input structure or bad settings (e.g. T too high).

Try visualize the system right before the convergence failure to see what really happened, and try minimize the system before dynamics.

Ray