Boundary issue, how to solve it?

6.26ls.txt (839 Bytes)
6.28.data (1.1 MB)
When I perform a deform operation to stretch my model, there is an issue with the boundary settings, causing the gap between the box and the model to grow larger during the stretching process. How should I set the boundary for the model in the x direction to address this issue?

image
image

Hi @Larry,

there is an issue with the boundary settings,

Which issue?

the gap between the box and the model to grow larger during the stretching process.

I see no bonds between atoms through periodic boundaries and pair interactions is set to none. How is that an issue? Why do you even set boundary conditions as periodic?

I’m sorry, I meant that I encountered a problem with my setup. I’ve tried different boundary conditions in the x and y directions, but the gap issue still persists. I don’t know why this problem occurs. The bonds during stretching are not visible in OVITO. I only want to validate the bonds without using pair, so I set it to “none.”

This does not answer my questions. :confused:

Boundary conditions are not directly related to box size. This does not tell us why this is an issue for you. What is the problem? A gap between atoms and boundary is not an issue in itself. If you want bonds to cross box boundaries, it would me more useful to know how you created your system.

This is because you are not using OVITO correctly. The bond informations are contained in the data file, not in the dump file. You have to load the initial datafile and then use the Load trajectory modifier to load the dump file. Anyway this is out of the point.

This is very risky. None does not build neighbour lists on which many commands and controls can rely. Use pair style zero instead.

If you want to create bonds that go through boundary conditions, set the boundary condition to p before creating the bonds. Else I see no issue with your simulation.

Hello Germain
I used a particle conversion package. By recognizing the image, it generates atomic coordinate information and box size. Then, I added bonds in this initial coordinate using create_bonds. Finally, I want to study bond breaking by stretching the system using deform. I did not use pair interactions, and I want to use fix bond/break to define a threshold for each bond, allowing them to break when the threshold is met.In the past few days, I tried setting parts of both ends as rigid bodies, and this issue still occurs during stretching. I also set the pair interaction to zero. However, the problem of the box stretching faster than the model still persists.

Img2Particle.py (8.1 KB)

I think I understand what you consider to be an issue now.

However, the problem of the box stretching faster than the model still persists.

The atoms on the side of the box are not bounded to it in any way. That means that when you deform the box, atoms are displaced by the deformation but pulled back to some extent by the bonds. If you want them to stick to the box wall (something they are not doing with your initial configuration), you should either exclude them from the group of atoms you integrate the equation of motion or make some bounds go through the periodic boundary conditions.

In you case this is easily done using:

change_box all x delta 0 -15 units box
group TY1 type 1
create_bonds many TY1 TY1 1 9 11

But you might need to double check your configuration generation procedure for a general solution.