substract com-speed for stress/atom computation


How can I substract center of mass velocity of group atoms from kinetic energy contribution in stress/atom computation?

There's no current way to do that. You can subtract it
from a global temp, e.g. compute temp/com.


what about that way:
(don't check yet)
# compute com velocity (one dimensional)
compute vy all reduce/region verx ave vy
thermo_style custom c_vy
run 0

# subtract com velocity from atoms
variable v equal -c_vy
velocity verx set 0 v 0 sum yes units box

# compute without com velocity
compute p verx stress/atom
run 0

#return com velocity to atoms
variable v equal c_vy
velocity verx set 0 v 0 sum yes units box

There is a variable option for vcm() of a group.
I don't see how that helps compute stress/atom
calcalate a velocity-squared term with the
COM velocity subtracted out.


I substract c.o.m-velocity from system and for result system(equal first but without vcm speed) calculate stress/atom,

then i try back the speed of the system - but this is don't work - variable rewrite with new "wrong" velocity.

I'm not clear on what you are asking now.
I told you before that LAMMPS doesn't subtract
the COM velocity in compute stress/atom. I don't
think you can make it do this form the input script,
if would require changing the source code.