[lammps-users] NVT Temp not match the desired one

LAMMPS LAMMPS Mailing List Mirror
1

Dear lammps users,

     I used fix nvt command to set up the desired temperature. However,
the equilibrate temp of the system is ~350K much lower than my target
500K, and it is not a constant. Can anyone point out what's wrong in
my input file below? thanks.

Haibin Chen Ph.D.
Mechanical Engineering Dept
Carnegie Mellon University
Pittsburgh, Pa, 15213

# study CnT with REBO potential
# This input file is used in serial job
# 03/19/2010

# initialization Section
units metal
atom_style atomic
pair_style airebo 3.0 1 1

boundary f f p

read_data ${datafile}
pair_coeff * * /home/haibin/lammps/lammps-8Feb10/potentials/CH.airebo C

neighbor 2.0 bin
neigh_modify delay 5 check yes
timestep 0.0005

thermo_style custom step temp etotal pe ke
group cnt type = 1
thermo 100

#velocity cnt create 1000 87287
#fix 1 cnt nvt 1000.0 1000.0 100.0

dump 1 cnt atom 100 dump.cnt
dump 2 cnt xyz 10 dump.xyz

#minimize 1.0e-7 0.0001 10000 100000
velocity cnt create 500 87287
fix 1 cnt nvt 500.0 500.0 10.0

run 2000

Temp, TotE, output
       0 500 -1739.3394 -1754.786 15.446587
     100 386.0122 -1739.3356 -1751.2608 11.925142
     200 368.58633 -1739.3266 -1750.7134 11.386802
     300 357.97876 -1739.3276 -1750.3867 11.0591
     400 358.36626 -1739.3299 -1750.4009 11.071071
     500 286.28365 -1739.3187 -1748.1629 8.8442107
     600 336.21285 -1739.3237 -1749.7104 10.386682
     700 331.43811 -1739.3198 -1749.5589 10.239176
     800 350.27195 -1739.3236 -1750.1446 10.821013
     900 361.61857 -1739.3216 -1750.4931 11.171546
    1000 328.71831 -1739.3161 -1749.4713 10.155152

2

I don't think your system has equilibrated in so few timesteps.

I would try running the simulation for a longer duration and/or using a
smaller value for Tdamp.

John

3

I don't think your system has equilibrated in so few timesteps.

I would try running the simulation for a longer duration and/or using a
smaller value for Tdamp.

also setting the initial temperature to a higher value,
e.g. 800K will make it easier on the thermostat, as it
does not have to transfer so much energy into the system.

cheers,
   axel.

4

More specifically, you set the damping constant to
10 psec in metal units which is 20,000 timesteps
and you've only run for 1000 steps.

Steve