Hello everyone, I hope you are having a good time. I am using both the April 8, 2021 version and the February 7, 2024 version. I want to combine two data files that have periodic boundary conditions for both. One of these data files includes a saltwater system and graphene walls, while the other contains a crystalline membrane structure. It is worth mentioning that the dimensions of both systems are not the same. The problem arises when combining the data files for the mentioned systems; the crystalline membrane, which has smaller dimensions, triggers an inconsistent image flag, causing issues with my simulation. I even used the reset_atoms and change_box commands as advised by Dr. Kohlmey( @akohlmey) , but the problem persists. Any guidance in this area would be extremely valuable to me. The relevant files will be attached.
Unfortunately, I was unable to upload the data file related to saltwater and graphene. I have uploaded the input file for generating this system. With endless gratitude
Is there any reasonable reason for you to do so? Why work with such an old version when you have a much more recent one available?
Inconsistent image flag is a recurrent warning with periodic solids. It can happen even in perfectly fine system because of bonds. See the explanation in the dedicated page.
What kind of issues?
You can use a google drive link or other external tools you trust to share files.
That said, only looking at your UIO-66-A.data file using Ovito, I can see that you have very long bonds (CC and CH) crossing your membrane along the z axis. They do not go though the simulation box, but the membrane itself, without crossing any simulation box face. That indicates that the cut from the periodic system has not been done correctly which might be, the cause of your simulation issues.
It is very difficult to get this to work out with multiple read_data commands in LAMMPS.
Both having different box dimensions and having non-zero image flags can cause problems.
I prefer to use external tools that use unwrapped coordinates instead, e.g. VMD /w TopoTools.
That will strip off the coeffs, but you can save them beforehand with write_coeffs. They will likely need some editing anyway.