brownian dynamics in lammps

LAMMPS LAMMPS Mailing List Mirror
1

hi,everyone!
I have done some brownian dynamics in lammps. The following is my input and data files.
#input
units lj
atom_style sphere
boundary p p p
dimension 3
neighbor 2.0 bin
neigh_modify every 6 delay 0 check no
read_data BD.data
group all type 1

thermo_style custom step temp pe etotal press vol
thermo_modify norm yes
thermo 1

pair_style lj/cut 5.0
pair_coeff * * 1.0 1.0

fix 1 all langevin 0.069 0.069 0.02848 48279
fix 2 all nve
fix 3 all gravity 1.0 spherical 0.0 0.0
timestep 0.0054

dump 1 all xyz 1000 dump.brownian
run 1000
#data
LAMMPS 2005 data file for brownian dynamic
10000 atoms

1 atom types

0 135 xlo xhi
0 135 ylo yhi
0 135 zlo zhi

Atoms
1 1 1 3.36E-20 5 5 5
2 1 1 3.36E-20 5 10 5
3 1 1 3.36E-20 5 15 5
4 1 1 3.36E-20 5 20 5
5 1 1 3.36E-20 5 25 5
6 1 1 3.36E-20 5 30 5
7 1 1 3.36E-20 5 35 5
8 1 1 3.36E-20 5 40 5
9 1 1 3.36E-20 5 45 5
10 1 1 3.36E-20 5 50 5
11 1 1 3.36E-20 5 55 5
12 1 1 3.36E-20 5 60 5
13 1 1 3.36E-20 5 65 5
14 1 1 3.36E-20 5 70 5
15 1 1 3.36E-20 5 75 5
16 1 1 3.36E-20 5 80 5
…
but errors happened,

dump 1 all xyz 1000 dump.brownian
run 1000
Memory usage per processor = 5.25631 Mbytes
Step Temp PotEng TotEng Press Volume
0 0 0 0 0 2460375
1 0.0001296591 0 0.00019446921 5.269365e-007 2460375
2 0.00038958856 0 0.0005843244 1.5832936e-006 2460375
3 0.00064876079 0 0.00097304387 2.6365734e-006 2460375
4 0.00090689418 0 0.0013602052 3.6856312e-006 2460375
5 0.0011608081 0 0.001741038 4.7175411e-006 2460375
ERROR: Lost atoms: original 10000 current 0 (…/thermo.cpp:389)
I am not sure what’s wrong with my simulation files, if you know, please help me! Thanks very much!

good luck!

2

Hey,

It looks like bad dynamics. Your atom is flying away. Either the initial structure isn’t ideal or it’s the physics of interactions that is wrong.

With best wishes,
Anna

3

Is there any particular reason you want particles with a density of 3.36E-20? This sounds like a very bad idea…

4

fix 1 all langevin 0.069 0.069 0.02848 48279
timestep 0.0054

I don't think this is the main source of your problems, but I see that
your "tdamp" parameter (0.02848) is only 5 times larger than your
timestep (0.0054).

http://lammps.sandia.gov/doc/fix_langevin.html

  To maintain temperature at this high damping rate would force LAMMPS
to use large random thermal forces to be applied to the atoms at every
timestep. I worry this is more likely to cause atoms to jump quickly.
(...perhaps leading to bad dynamics and "lost atoms").
    Again, I suspect this is not causing your current problem.
However, if you fix everything else, and you are still having bad
dynamics, try increasing "tdamp" to at least 50 timesteps or higher,
(and/or decreasing your timestep).

Cheers
Andrew

   I like to pontificate about this topic:
http://sourceforge.net/p/lammps/mailman/message/32765047/

5

Hi, Andrew, thanks very much for your suggestions. Good luck for you!