[lammps-users] FENE bond simulation

LAMMPS LAMMPS Mailing List Mirror
1

Hi,
I am trying to run a LAMMPS (svn) simulation on a single chain with
FENE bonds. I start with the chain in a zigzag configuration in the
x=0 plane and would like to let it relax before collecting statistics.
However, the simulation fails after the first step. If I set the
epsilon parameter in the bond_coeff line to be zero, it goes a little
further but still fails:

../../src/lmp_rc < in.chain
LAMMPS (11 Aug 2010)
Scanning data file ...
  1 = max bonds/atom
Reading data file ...
  orthogonal box = (-100 -100 -100) to (100 100 100)
  1 by 1 by 1 processor grid
  11 atoms
  10 bonds
Finding 1-2 1-3 1-4 neighbors ...
  2 = max # of 1-2 neighbors
  2 = max # of special neighbors
Setting up run ...
Memory usage per processor = 76.0118 Mbytes
Step Temp E_pair E_mol TotEng Press
       0 0 39.554059 2.2654102 41.819469 0.00024648068
      10 26.449181 0 5.8636366 41.930702 2.7063774e-05
      20 21.474684 0 12.622873 41.906532 1.1783188e-05
      30 13.583303 0 23.456362 41.979048 -2.0116461e-05
      40 4.3574789 0 35.891324 41.833341 -7.040495e-05
      50 3.68068 0 36.786571 41.80568 -6.1832851e-05
      60 10.664968 0 27.322138 41.865276 -2.9132045e-05
      70 18.610394 0 16.389581 41.76739 1.3915096e-06
      80 23.485878 0 9.5259131 41.55211 1.8256388e-05
      90 38.544439 24.993943 5.4637726 83.018314 0.00019026373
WARNING: FENE bond too long: 93 2 3 5.44244
ERROR on proc 0: Bad FENE bond

2

Did you try minimizing the structure before running MD?

3

Dear RC,

try to decrease your damp constant in
"fix 2 all langevin 1.0 1.0 100.0 301033"
from 100 to 1.0

Jan-Michael

4

Hi,
I calculated the bead damping coefficient from damp=D m_b N/k_B T
where D is the chain diffusivity, m_b the bead mass, and N the number
of beads, assuming a Rouse chain. Using values for my system, I got
damp to be approximately 100.

I did fix the problem by moving the initial bead locations further apart.
Thanks for taking a look at my problem.

RC

5

The damp constant for Langevin is in units of time (tau for LJ).
Since tau is typically about 200 steps, a setting of 100
is 20,000 steps which is way too long. A setting of 1 is
more typicaly. But your blow-up is likely due to a bad
initial config, not the damping constant issue.

Steve