[lammps-users] determination of average cell size from NPT simulation

LAMMPS LAMMPS Mailing List Mirror
1

I’d like to obtain average cell size (width,depth, and length) using NPT simulation.
If box is orthogonal, I think that I should calculate the average of width, depth and length with its standard daviation.

What variables are used for width, depth, and length of box and what command (e.g., those concerning to average, or trap() function etc) should I use to get their respective averages for NPT simulation?

Best regards,

Luke

2

There are several thermo keywords, like vol, lx, ly, lz, that
can be used with thermo_style custom to output
box info. They can also be used in variables, see
the variable doc page. The variable can be passed to
commands like fix ave/time to do further averaging and output.

Steve

2010/8/25 JhonY. I. <[email protected]...>:

3

Dear Steve:

First many thanks for your anwer really.

  1. I’ve performed NPT simulation as follows based on your advice.

fix 1 all npt temp 298. 298. 0.5 y 1. 1. 0.5 z 1. 1. 0.5
fix 2 all ave/time 10 100 1000 Lx Ly Lz ave running file lattice.log

However it showed an error message of “ERROR: Illegal fix ave/time command”.
The syntax of this av/time command was very common and I don’t know what’s wrong with it.
Would you let me know what’s the problem with it?

  1. Incidentally asking, what size (wha multiple of time step) should we choose for Pdamp(time unit) in NPT command generally?
    Is it 1000 times of time step (I did it in this way)?

Best regards,

Luke

4

The fix ave/time cannot use Lx, Ly, etc directly.
They need to be put in a variable, e.g.
variable myLx equal lx
and then referenced as v_myLx in the ave/time
command. Lx, Ly, etc are keywords for
the thermo_style custom command and
variables only. The doc pages explain all this.
A pressure damp factor of about 1000 timesteps
is typical. Note that isn't 1000 time units,
since 1 time unit may be 1000 timesteps in some
unit systems.

Steve

2010/8/31 JhonY. I. <[email protected]...>: