I want to calculate cohesive energy of fcc system at room temperature and 0 pressure. So I used following code for NPT.
CODE:
units metal
velocity all create 100 87287 mom yes rot yes
neighbor 2.0 nsq
neigh_modify once no every 1 delay 0 check yes
thermo_style custom step press vol temp etotal pe ke
timestep 0.01
fix 3 all npt temp 100 298 100 iso 0 0 100
variable eng atom "etotal"
dump 1 all custom 100 dump.temp v_eng
run 50000
Results after 50000 runs:
pressure is varying from -1300 to 1400
temperature oscillating about 90
why is pressure having negative value? Is number of runs insufficient? How can I minimize total time?
Thanks you for your response.
Surya Bijjala
Where / What is boundary conditions?
Where / what is potential between atoms?
What interaction model are you using? When I used COMB3 for a while, I learned that it does not properly conserve energy. So in an NVE calculation, temperature kept slowly decreasing and in an NVT simulation the temperature oscillated a little below the target temperature. Are you using COMB3 or another interaction model that leaks away energy?
Hi,
In this line you define the damping parameters, expressed in time units:
fix 3 all npt temp 100 298 100 iso 0 0 100
Since your timestep is 0.01, your T damping is actually 10000 timesteps, which may be too large.
I'd try something like:
fix 3 all npt temp 100 298 1 iso 0 0 10
The lammps manual explain all this very nicely, check the 'note' boxes:
https://lammps.sandia.gov/doc/fix_nh.html
Hope this helps,
Veniero.
- Badges
- Science topic
- Similar topics
- Biological Science
- Kinesiology
- Running