[lammps-users] ERROR: Non-numeric pressure - simulation unstable

LAMMPS LAMMPS Mailing List Mirror
1

Dear all LAMMPS users,
I am trying to simulate the compression test of the Al75Pd25 alloy. I face this problem. I am reading and studying other problems like this but can’t solve it properly or won’t understand the npt or nve issues. How can I solve this problem?

------------------------ INITIALIZATION ----------------------------

units metal
dimension 3
boundary p p p
atom_style atomic
variable latparam equal 4.05

----------------------- ATOM DEFINITION ----------------------------

read_data Al75Pd25.dat
Reading data file …
orthogonal box = (0.0000000 0.0000000 0.0000000) to (23.681459 23.681459 23.681459)
1 by 1 by 1 MPI processor grid
reading atoms …
1099 atoms
read_data CPU = 0.004 seconds

------------------------ FORCE FIELDS ------------------------------

pair_style meam/c
pair_coeff * * library.meam Al Pd PdAl.meam Al Pd

------------------------- SETTINGS ---------------------------------

compute 1 all stress/atom NULL
compute 2 all reduce sum c_1[1] c_1[2] c_1[3]

2

you need to investigate systematically what is wrong by turning off any unneeded tasks.

the minimum that should be possibly is a “run 0” command with “fix nve” and no other fixes.
the values you get for energies and pressure will inform you whether your initial geometry is reasonable or invalid.
if energy/pressure are too high or infinity or NaN, then you probably have overlapping atom positions or incorrect box dimensions or both.

once that is resolved, it should be possible to run for 1000 steps with fix nvt. from that you can see whether your system geometry is close to equilibrium or far away and whether you should first run for a bit with fix nvt or can switch to fix npt right away.

after sorting that out, you can properly equilibrate your system and after it is suitably equilibrated (and not before!) you can start your production calculation.

axel.

1 Like