Yeap, I understand that the ‘atom volume’ is hard to define. However, people still want to calculate the pressure(or stress) in partial of the system. I noticed that there are so many similar questions at the mail lists.
I’ve tried " compute voronoi/atom " yesterday. I found that the volume of the atoms on the free surfaces(both inside and ouside) are obvious greater than others, which may relate to the define of voronoi volume. So, maybe it’s IMPOSSILE to figure out the pressure in a hollow wire precise, don’t you think so ?
Yeap, I understand that the 'atom volume' is hard to define. However,
people still want to calculate the pressure(or stress) in partial of the
system. I noticed that there are so many similar questions at the mail
lists.
then why don't you "harvest" those discussions for possible solutions
or improve your understanding of the issue?
particularly gain some understanding that pressure is not a meaningful
defined property to look at, in the first place.
I've tried " compute voronoi/atom " yesterday. I found that the volume of
the atoms on the free surfaces(both inside and ouside) are obvious greater
than others, which may relate to the define of voronoi volume. So, maybe
it's IMPOSSILE to figure out the pressure in a hollow wire precise, don't
you think so ?
as steve said, you really have not considered properly what pressure
means. let me restate: pressure is the force per area that a system
effects on its boundary or a "sample" area that you place at a
location. so it is a property that is independent of the number of
atoms and resolved per surface (not volume).
stress, however is readily available as a per-atom property (in the
form of virial stress). for other spatially resolved stress
computation you can refer to the options of the atoms-to-continuum
package (e.g. hardy stress).
in short, stress has well defined ways to get per atoms or spatially
resolved data. pressure doesn't.
I think I know what both of you mean, the reason why I put the pressure and stress together is because in continuum mechanics,
pressure and stress have a same unit(Pa) . but in LAMMPS , stress is in the form of virial stress, in units of pressure*volume.
what i want to do is : compute the X direction component of stress(in continuum mechanics form) ,or what you call pressure of a hollow pillar
the method i tried to used is: sum up all the stress( in the form of virial stress) of atom ,and divided by region volume
what still make me confusing are: 1. Is the method I mention above make sense ?
2. how to compute the " region volume " , which is hard to define
I think I know what both of you mean, the reason why I put the pressure and
stress together is because in continuum mechanics,
pressure and stress have a same unit(Pa) . but in LAMMPS , stress is in
the form of virial stress, in units of pressure*volume.
what i want to do is : compute the X direction component of stress(in
continuum mechanics form) ,or what you call pressure of a hollow pillar
the method i tried to used is: sum up all the stress( in the form of virial
stress) of atom ,and divided by region volume
what still make me confusing are: 1. Is the method I mention above make
sense ?
2. how to compute the "
region volume " , which is hard to define
for the third (and last time as far as i am concerned): you are
looking at this in the wrong way. you *cannot* transfer atomistic
properties into continuum properties through juggling units around. i
recommend - again - to look a the atoms to continuum package, which
provides better ways to relate properties between the continuum and
atomistic simulation worlds. why use an ad hoc (and know to be
inferior) approach, if better alternatives exist?