Dear Lammps Users and Developers
I should divided the simulation box to diffrent groups and compute some local properties like stress, volume and displacement locally. I need to create groups more than 32.
But there is a limitation for it in Lammps.
I studied the literatures that researchers calculate and averaged the stress and other properties for local groups by LAMMPS. as I understand they should use more than 100 groups (or maybe other thing) to do that.
is there any valuable advise and suggestion to calculate stress and other properties for more than 50 groups ?
Thanks in advance
Dear Lammps Users and Developers
I should divided the simulation box to diffrent groups and compute some
local properties like stress, volume and displacement locally. I need to
create groups more than 32.
But there is a limitation for it in Lammps.
I studied the literatures that researchers calculate and averaged the stress
and other properties for local groups by LAMMPS. as I understand they should
use more than 100 groups (or maybe other thing) to do that.
is there any valuable advise and suggestion to calculate stress and other
properties for more than 50 groups ?
have you looked through the mailing list archives?
this is a frequently recurring (and answered) question.
axel.
The limit of 32 groups is fairly fundamental (it would require big changes in the source code to alter).
Have a look into the ave/spatial fix to work out local properties.
Dear Axel
perhaps my searching was insufficient, though I didn’t see
something that fit my need. My situation is incredibly simple: I need define some region locally then calculate and averaging the stress and displacement of the atoms in regions in every 50 time steps. I found in mailing list about use of molecule ID but I couldnt fit it for my simualtion
Thanks
Dear Axel
perhaps my searching was insufficient, though I didn't see
something that fit my need. My situation is incredibly simple: I need define
some region locally then calculate and averaging the stress and
what have regions got to do with groups?
displacement of the atoms in regions in every 50 time steps. I found in
mailing list about use of molecule ID but I couldnt fit it for my simualtion
regions you can have as many as you want, but if they are regular, you
could use ave spatial,.without ever having to define a group.
but then again, for spatially averaged stress, you may be better off
using the AtC package. try searching for that in the archives.
i suspect, you have limited yourself by looking for the wrong keywords
and - perhaps - discarding useful messages too quickly.
axel.
Dear Axel
Thanks for your reply. I used region just to define specific groups and after that I dont need the regions.
as I know, ave spatial averaged the atoms that are exist in region at current time. but I need to calculate the average of displacement and stress of a specific group of atoms that move slightly during tensile and shearing simulation. Thanks
Dear Axel
Thanks for your reply. I used region just to define specific groups and
after that I dont need the regions.
as I know, ave spatial averaged the atoms that are exist in region at
current time. but I need to calculate the average of displacement and stress
of a specific group of atoms that move slightly during tensile and shearing
simulation. Thanks
so do your simulation once and then use "rerun". for each rerun you
can define 30 groups.
or assign molecule to each group and then use molecule based computes.
as i said, this has been discussed *many* times. i am getting tired of
sounding like a broken record.
axel.
Dear Axel
I am so sorry for bothering you.
I tried to understand the use molecule method instead of groups but it is not
clear for me (maybe I couldn’t understand the relation of molecule and my simulation).
I tried to find a sample in mailing list to clear this relation, but I couldn’t.
the following is a real sample of my simulation and the properties that I
want to compute for one group. can you please help me how
can I use molecule based computes to get same result as the following commands.
Thanks in advance
#create group
region center block INF INF 7 27 30 47 units box
group center region center
#compute properties
compute vol all voronoi/atom
compute stress all stress/atom
compute zdis all property/atom z
#reduce
compute cent center reduce sum c_vol[1]
compute centstr center reduce sum c_stress[3]
variable centstress equal c_centstr/c_cent
compute centdis center reduce ave c_zdis
thermo_style custom step vol lz press temp c_centdis c_centstress c_cent c_centstr
If you want me to do your thinking for you you have to hire me as your consultant. I gave you multiple workable suggestions, but you have to make them work. I have little patience for people with a one track mind that only see what they want to see.
Axel.