Angle atoms missing right after being created

LAMMPS LAMMPS Beginners
,
1

Hello. I’ve been trying to run a script for gas permeabilitiy, and the way it works is by generating a methane molecule, having it run for a while, and then deleting that molecule and recreating it in a new location.

This works for a while, but after around 15 thousand steps, creating the atoms suddenly creates this error:

Deleted 5 atoms, new total = 2300

Created 5 atoms

using lattice units in orthogonal box = (0 0 -15.176166) to (19.7 15.2 91.056996)

create_atoms CPU = 0.000 seconds
5 atoms in group penetrant

Generated 1 of 1 mixed pair_coeff terms from arithmetic mixing rule

Setting up cg style minimization …
Unit style : real
Current step : 15618

WARNING: Inconsistent image flags (src/domain.cpp:1051)

ERROR on proc 0: Angle atoms 2302 2301 2304 missing on proc 0 at step 15618 (src/ntopo_angle_all.cpp:62)

Last command: minimize 0.0 1.0e-4 1000 10000003

the area of my code that is causing this error:

minimize 0.0 1.0e-4 1000 10000004

delete_atoms group penetrant

create_atoms 0 random 1 $s mobileR mol CH4mol 55 overlap 0.3 maxtry 5000

group penetrant id > 2300

fix 3 penetrant nvt temp 300.0 300.0 100.0

minimize 0.0 1.0e-4 1000 10000003
“”"

I don’t think it’s an overlap thing, since there wasn’t a warning about a failure to create the molecule. The minimize command just before I delete also didn’t cause any issues. What’s happening here, and how can I fix it?

2

Why delete and recreate? Why not just displace?

3

I’ve tried that. Seems to cause massive issues with kinetic energy, I presume due to overlap problems.

4

Shouldn’t you need to add “bond yes” and “mol yes” here?

5

There aren’t any bonds between atoms in this group and the rest of the system. The only bonds in the group penetrant are 4 CH bonds, and all are contatined in this group.

6

I suggest you then need to figure out which atoms exactly are missing.
That is best done by (over)writing out a data file after adding the molecule and inspecting it after LAMMPS crashes to identify the location causing the problem.

7

I know which atoms are missing. All the atoms in the error message (2302, 2301, and 2304), are molecules recently created by the “create_atoms” command.

8

Nevertheless, if you create the data file, can you restart from the data file without a problem or does it fail with the “Angle atoms missing” error. We’ve covered all “simple” issues that could be causing problems, from here on you will have to debug this yourself until your can produce a simple input deck that reproduces the issue quickly.

I don’t see anywhere which LAMMPS version you are using. Please let us know. Also the platform that you are running on.

9

Alright, I’ll give that a go. Thanks for your help.

My lammps version is LAMMPS (29 Aug 2024), and I’m running this on Linux.