Dear all,
I am trying to do a straining simulation of a Ni structure with Helium and an originally rigid inclusion particle (using nickel atoms also) in it. With the aim of keeping the particle rigid at all times by using fix setforce to 0s and velocities set to 0s for the atoms making up the inclusion, as well as exclude atoms inside the particle from all the ensemble- and deformation-related fixes, the atoms inside the inclusion still “moves” and atoms outside of the inclusion also cut through it at some point. The velocities and forces on the inclusion atoms did maintain 0s throughout the simulation though.
By some inspections I suspect the problem arises from the followings but I am not sure what to change in order to properly proceed:
- The straining is done by fix deform with the setting of remap x. Since the inclusion atoms are excluded from the fix, I suspect some mobile atoms are remapped to a location inside the inclusion, affecting the rigidity of the inclusion.
- The inclusion is at a location crossing the periodic boundaries. As the (triclinic)box deforms, I am actually not sure how the location of the inclusion atoms change relatively as they theoretically are not “remapped” to the new box.
My intention is to make the inclusion rigid at all times while the straining takes place. I attached part of my script in the following. I am using lammps version 17Nov2016. Any help or comments are appreciated.
it works for me using the “remap v” option and applying time integration only to the mobile atoms.
axel.
Hi Axel,
I tried remap v and the same problem remains. When you say applying time integration only to the mobile atoms, do you mean excluding the rigid atoms from all the fixes? Or, is there a way for the “run” command to exclude the rigid atoms?
I found that after the first 5000 steps run under NVT ensemble the locations of the atoms seems fine, and since the fix deform started that’s where the problem starts. The location of the rigid atoms changes every time when a strain increment is employed (but during relaxation in between they didn’t). Do you have any recommendation on what I can check or test on in this case?
Thank you very much,
Esther
Hi Axel,
I tried remap v and the same problem remains. When you say applying time integration only to the mobile atoms, do you mean excluding the rigid atoms from all the fixes? Or, is there a way for the “run” command to exclude the rigid atoms?
please (re-)read the documentation on fixes, fix groups and time integration fixes very carefully. it is all explained in the manual.
I found that after the first 5000 steps run under NVT ensemble the locations of the atoms seems fine, and since the fix deform started that’s where the problem starts. The location of the rigid atoms changes every time when a strain increment is employed (but during relaxation in between they didn’t). Do you have any recommendation on what I can check or test on in this case?
i have no time to do your science for you, and it is impossible to do it on such a vague description, anyway. most likely you didn’t completely understand what i was suggesting and have to revisit my response more carefully.
i have shown, that the thing that you claim does not work, can be made to work. that is all i can do.
axel.