Tensile loading

LAMMPS LAMMPS Mailing List Mirror
1

Dear Users,

I am simulating a system consisting of Kremer Grest chains of dimensions X= 700a, Y=40a, Z=100a. I am applying tensile load along Z dimension by giving velocity to upper and lower layers in opposite directions. These upper and lower layers are at a distance of 25a from centre of box (in Z direction). Each layer has thickness of 5a. I am thermostating thermal velocities in X and Y directions and computing the stress in Z direction. I have setup this system according to the published work and I want to reproduce results reported in this work. I am getting similar behaviour but smaller values of stress in Z direction. The main part of the code is as follow and diagram for stress is in attachment.

2

Two comments:

  1. if you are new to LAMMPS, you should start with simpler simulations and try to reproduce published results before moving to more complex systems.
  2. there should be plenty of discussions on inputs for tensile loading in the mailing lists archives. Some also will mention publications to look at.

Axel

3

Dear Muhammad,

Hello Muhammad. You can use the fix deform to deform your simulation box to simulate tensile loading. Please see the link that discusses the necessary input arguments for fix deform: https://lammps.sandia.gov/doc/fix_deform.html

If you are applying a tensile load to amorphous chains (that does not follow affine deformation) in the z-direction and elongating your box every 10 timesteps, for example, then your fix command should look something like this example: fix 1 all deform 10 x volume y volume z erate (strain rate value) remap v

I apply fix deform subject to NVT ensemble.

To compute stress for these tensile tests for every 100 timesteps, for example, output the pzz (pressure in the z-direction) and also lz using the thermo_style command

thermo 100
thermo_style custom step temp pe etotal epair pxx pyy pzz lx ly lz

Take the negative of pzz which will give you the stress. Divide lz by its original box dimension in the z-direction and you can plot the stress as a function of stretch in the z-direction.

Thank you.

Sincerely,

Masato Koizumi

4

Dear Kohlmeyer and Koizumi,

Thank you very much for your reply.

In last few weeks, I performed simulation for Kremer Grest chains and Polyethylene chains using Lammps in equilibrium, under shear using nvt/sllod+deform and uniaxial tension using deform command as suggested by Koizumi and reproduced published results accurately. Now, I am trying to apply tensile load by fixing upper and lower layers rigid and displacing them at constant velocity. The reason to do so is to run simulation for interfaces later on, in which we select region around the interface and then perform tensile test like this to determine the weld/interface thickness. When I compare this system, I am able to reproduce similar behaviour but with some difference in stress values (fig is attached). I am working first time on a system, in which holding layers fixed and displacing them with constant velocity. Thats why, I want to ask if I am doing something wrong or I have missed something in the script? Or there is any other way compute stress?

Thanks for your help and time.

Best Regards,
Anwar

stressZ.pdf (290 KB)