Atom ID is changing

LAMMPS LAMMPS Mailing List Mirror
1

Hi there,

I’m trying to perform crack propagation simulation in Ti based binary alloys.

My goal is to monitor the force, pair-wise energy and distance of two adjacent atoms residing just ahead of the crack tip. In order to choose that particular pair of atoms, I’m picking up the atom IDs from the equilibrated data file. Interestingly, the particular atoms I choose in the data file (wrote using write_data) are different from the atoms in dump.0 file (generated during the simulation), although the atom IDs are same. Also, during the simulation the atoms are moving randomly within the box, even though it is a perfect crystal. It seems that atom ID is changing or I’m doing wrong to keep monitoring that pair of atoms.

I would like to monitor different properties of two atoms throughout the simulation. How to do that accurately within lammps?

Thanks,
R

2

By the way, I forgot mention that I’m writing the dump file using the command below,

dump 1 all cfg 500 dump.TiAl_*.cfg mass type xs ys zs c_energy_pe

Thanks for your help.

3

By the way, I forgot mention that I'm writing the dump file using the
command below,

*dump 1 all cfg 500 dump.TiAl_*.cfg mass type xs ys zs
c_energy_pe
*
*
*

The atom ID cannot change during a simulation. The problem you have, is
that the cfg format do not contain any atom ID. So, the order of the
atoms in the file determine it while reading the file with, for example,
atomEye. This order is not enforced by Lammps (it can be done actually,
but better not for maintaining the efficiency).

What you should do, is using a format that store atom ID (atom, custom,
etc.), or store the atom ID in you CFG file as supplementary column.

*Thanks for your help.*

    Hi there,

    I'm trying to perform crack propagation simulation in Ti based
    binary alloys.
    My goal is to monitor the force, pair-wise energy and distance of
    two adjacent atoms residing just ahead of the crack tip. In order
    to choose that particular pair of atoms, I'm picking up the atom
    IDs from the equilibrated data file. Interestingly, the particular
    atoms I choose in the data file (wrote using write_data) are
    different from the atoms in dump.0 file (generated during the
    simulation), although the atom IDs are same. Also, during the
    simulation the atoms are moving randomly within the box, even
    though it is a perfect crystal. It seems that atom ID is changing
    or I'm doing wrong to keep monitoring that pair of atoms.

    I would like to monitor different properties of two atoms
    throughout the simulation. How to do that accurately within lammps?

To keep the atom ID unchanged while using read_data command, you have to
use these 2 keywords "add merge".
The idea is first to create an empty simulation box, with a default
size, and then to merge it with the data file.

Julien

4

Thanks Julien for your input. Now I see that the atom ID is actually unchanged in the data file and the dump file. The only thing that I did is dumping custom files with atom ID.

Thanks,
R