Decomposition of virial contributions for calculation of heat flux

Dear all,

I plan to calculate the contributions of different pairwise interactions (vdW and short-range Coulomb) to the total heat flux. In compute heat/flux command, by defining compute stress/atom pair, I can calculate the total pair contribution to the heat flux. However, compute stress/atom (as well as compute pe/atom) cannot decompose the contribution from vdW and short-range Coulomb. I am just wondering is there any command or method can help to decompose the contributions from different pairwise interactions?

Kind regards,

MC

Dear all,

I plan to calculate the contributions of different pairwise interactions
(vdW and short-range Coulomb) to the total heat flux. In compute heat/flux
command, by defining compute stress/atom pair, I can calculate the total
pair contribution to the heat flux. However, compute stress/atom (as well
as compute pe/atom) cannot decompose the contribution from vdW and
short-range Coulomb. I am just wondering is there any command or method can
help to decompose the contributions from different pairwise interactions?

​if you are using a kspace style, then i don't think there is much meaning
to determining the contributions from the short-range coulomb​, as that is
arbitrary based on the cutoff. another question is: how are you going to
assign the kinetic energy contributions?

probably your best shot to get anything meaningful out of this would be to
use the rerun command on a previously computed trajectory and then do
multiple analysis runs where you selectively turn off contributions to the
total heat flux.

another options could be to take the code from compute heat/flux/tally in
the USER-TALLY package and adapt it to your needs, but that could be
complex, as you would need to have a couple of additional pair style
instances (e.g .for lj/cut and coul/cut) around and then use their
Pair::single() method to recompute the force and energy contribution for
each interaction and each pair separately.

axel.

You can do it after the fact if you generate dump files then

rerun them via the “rerun” command and defining a potential(s)

that only have the components you want.

Steve