[lammps-users] vel distb'n


I am using the following code to equilibrate my system( triangular sheet of carbon atoms) to 300K.

dimension 3
boundary s s s

atom_style atomic

pair_style airebo 3.0
read_data coordlmps.d (A file with all the coordintes of atoms)

pair_coeff * * CH.airebo C
#*** It’s fixed at both ends in x direction. The low region has only five atoms and middle and high regions have more atoms comparatively.
region 3 block INF 0.2 INF INF INF INF units box
group left region 3

region 4 block 1.0 2.5 INF INF INF INF units box
group low region 4

region 5 block 2.5 49.97 INF INF INF INF units box
group middle region 5

region 6 block 49.97 51.5 INF INF INF INF units box
group high region 6

region 7 block 51.5 INF INF INF INF INF units box
group right region 7

fix 1 left setforce 0.0 0.0 0.0
fix 2 right setforce 0.0 0.0 0.0
velocity all create 300.0 873586443 dist gaussian
velocity left set 0.0 0.0 0.0 units box
velocity right set 0.0 0.0 0.0 units box

fix 14 all nvt 300.0 300.0 0.25 drag 0.4
timestep 0.0005

run 100000
unfix 14
fix 5 all nve
compute 1 low temp
compute 2 high temp
compute 3 middle temp
timestep 0.0005
thermo_style custom temp pe ke etotal c_1 c_2 c_3
thermo 1
run 500000

I don't know. I would start with a simpler geometry and no
thermostats and insure you can run with NVE in a stable way.
I would not thermostat 5 atoms. I don't see a units command.
The default is lj. The CH.airebo file is parameterized for
metal units (as the doc page says), so if you are running in LJ
units, that's a serious error.