Pressure fluctuation

LAMMPS LAMMPS Mailing List Mirror
1

Dear Lammps users,

Greetings!

I read all of the old mailing list archive about pressure fluctuation during NVT MD simulation. As Dr. Axel suggested in the mail list, I have changed my small system to big system (~65000 atoms), even tried langevin command and I tried keyword drag to control temperature. I believe that during fix/deform (compressing volume) pressure should increase. but it is fluctuating a lot

what i’m doing here is:

  1. minimize system with cg

  2. use fix NVT to equilibrate untill temperature is stable

you can see negative pressures in OUTPUT values given below

###INPUT SCRIPT (few lines) ################################################

#MOLECULAR DYNAMICS OF only CNT /NANO-INDENTATION OF CNT REINFORCED POLYMER by a virtual spherical indenter with Mishin Potential

echo both
dimension 3
boundary p p p
units real
atom_style full
neighbor 2.0 bin
neigh_modify delay 1

#potentials
pair_style lj/class2 10.0
bond_style class2
angle_style class2
dihedral_style class2
pair_modify mix sixthpower
special_bonds lj/coul 0 1 1 extra 65

read_data OnlyPolymer.txt

########Shrinking the box before curing polymer##########

compute new all temp/com
velocity all create 600.0 862075 temp new
thermo 1000
thermo_modify temp new

###Seperate group for polymer molecules during box change

timestep 0.05
min_style cg
min_modify dmax 0.05 line quadratic
minimize 1.0e-8 1.0e-10 5000 50000

dump 1 all cfg 100 Dump_*.txt mass type xs ys zs

group 1 molecule 1:1668

fix 7 all nvt temp 300.0 300.0 100 drag 1.5

#fix 8 all langevin 300.0 300.0 100 48279 (tried this before)

###Setting up force field to keep polymer from running into CNT
fix 11 1 indent 2000.0 cylinder z 0.0 0.0 11.0 side out
timestep 0.01

#reset_timestep 0
fix 2 all deform 1 x final -140.0 140.0 y final -140.0 140.0 units box
run 667000
min_modify dmax 0.05 line quadratic
minimize 1.0e-5 1.0e-7 1000 1000
unfix 2

###OUTPOUT (few lines)###############################################

minimize 1.0e-8 1.0e-10 5000 50000
WARNING: Resetting reneighboring criteria during minimization (…/min.cpp:168)
Setting up minimization …
Memory usage per processor = 39.4468 Mbytes
Step Temp E_pair E_mol TotEng Press
0 600 3883913.2 6.5760607e+08 6.6160633e+08 -5585055.3
680 600 326616.49 113895.95 556855.25 425.55711
Loop time of 63.9065 on 5 procs (5 MPI x 1 OpenMP) for 680 steps with 65052 atoms

Minimization stats:
Stopping criterion = energy tolerance
Energy initial, next-to-last, final =
661489985.227 440512.44704 440512.443394
Force two-norm initial, final = 2.55781e+07 315.439
Force max component initial, final = 720454 80.1616
Final line search alpha, max atom move = 0.000427236 0.0342479
Iterations, force evaluations = 680 785

Pair time () = 23.9871 (37.5347) Bond time () = 26.2141 (41.0196)
Neigh time () = 0.617614 (0.966435) Comm time () = 12.65 (19.7945)
Outpt time () = 0 (0) Other time () = 0.437642 (0.684816)

Nlocal: 13010.4 ave 15423 max 9728 min
Histogram: 1 0 0 1 0 0 0 2 0 1
Nghost: 10429.6 ave 12455 max 7432 min
Histogram: 1 1 0 0 0 0 0 0 1 2
Neighs: 1.31146e+06 ave 1.72587e+06 max 906273 min
Histogram: 1 0 0 1 1 0 1 0 0 1

Total # of neighbors = 6557284
Ave neighs/atom = 100.801
Ave special neighs/atom = 2.10256
Neighbor list builds = 14
Dangerous builds = 0

fix 7 all nvt temp 300.0 300.0 100 drag 1.5

#Setting up force field to keep polymer from running into CNT
fix 11 1 indent 2000.0 cylinder z 0.0 0.0 11.0 side out
timestep 0.01

#reset_timestep 0
fix 2 all deform 1 x final -140.0 140.0 y final -140.0 140.0 units box
run 667000
Setting up run …
Memory usage per processor = 37.1968 Mbytes
Step Temp E_pair E_mol TotEng Press Volume
680 600 326616.49 113895.95 556855.25 425.55711 13500000
1000 311.62228 331542.77 166360.77 558328.55 227.04549 13499136
2000 355.41032 331629.34 157769.29 558314.36 109.78094 13496438
3000 338.63085 332283.39 160226.56 558172.06 282.86508 13493740
4000 315.1032 331413.37 165551.4 558064.74 163.16831 13491042
5000 341.99649 327120.38 164517.47 557952.57 -72.065164 13488344
6000 330.03312 325922.72 167910.33 557828.02 -168.9411 13485647
7000 338.43446 326823 165249.1 557696.12 -100.0788 13482950
8000 317.23944 329533.55 166520.12 557567.87 37.773201 13480253
9000 328.6269 327641.62 166082.54 557446.45 -64.349315 13477557
10000 329.94962 329109.39 164226.81 557314.97 57.070046 13474860

Thanks in advance

Bevara

2

Your energy is very positive indicating a much more serious problem. Until you fix this problem you should not worry about pressure fluctuation.

Ray

3

Mr. Ray,

very positive energy means, it is very high?. how can we decide how much that energy should be? Is there any limit?

As I know energy must be conservative for isolated system but not necessarily so in a simulation. Energy conservation may be violated for several reasons: the time step chosen for integration may be too large, the cutoff method chosen may not be sufficiently good, numerical limitations of the computer.

Thanks

Bevara

4

This means your system is not stable.

Ray

5

your initial configuration (atomic positions) is probably very wrong.

luca