# pressure distribution

Dear all
I want to have a pressure distribution (pressure profile) of a group of atoms(here ‘flow’ group). I
use the following way to do this.

units real

compute peratom flow stress/atom

variable press1 atom -(c_peratom[1]+c_peratom[2]+c_peratom[3])/3

fix profile flow ave/spatial 1 100000 1000000 x lower 1 y lower 1 z lower 1 v_press1 units box ave running file MyProfile.txt

In the post processing procedure, i divide the v_press1 value by the volume of the corresponding Bin and the average Number of the atoms in that Bin.I.e. v_press1/(N_count * Bin_volume) [here the Bin_volume is 1].
So i have an average pressure for every Bin and a pressure distribution.

please tell me that if the above described way is true or not?
mahdi

mahdi sahebi
phd student
Isfahan university of technology

Dear all
I want to have a pressure distribution (pressure profile) of a group of
atoms(here 'flow' group). I
use the following way to do this.

units real

compute peratom flow stress/atom
variable press1 atom -(c_peratom[1]+c_peratom[2]+c_peratom[3])/3
fix profile flow ave/spatial 1 100000 1000000 x lower 1 y lower 1 z lower 1
v_press1 units box ave running file MyProfile.txt

In the post processing procedure, i divide the v_press1 value by the volume
of the corresponding Bin and the average Number of the atoms in that
Bin.I.e. v_press1/(N_count * Bin_volume) [here the Bin_volume is 1].
So i have an average pressure for every Bin and a pressure distribution.

please tell me that if the above described way is true or not?

please note that this mailing list is not an input file approval
service. would you be willing to debug other people's inputs. and by
showing just a small part, how can we know that the rest is done
correctly? so you have to come up with a strategy to validate your
approach by yourself. that is just good practice anyway. what does it
help you, if somebody here says "yes, it is correct", when it would
turn out to be wrong, because that person didn't know better or
overlooked a mistake? how can you know what and whom to trust (but
yourself)?

so why don't you just test it on a simulation of a homogeneous system
without a pressure gradient, where each bin should have (on average)
the same pressure as the global pressure? and start with one bin (i.e.
all of the system) and gradually increase the number of bins/slices to
the desired resolution to see what kind of averaging and simulation
length is required to get well converged results and then work in
steps toward your real target system.

axel.