[lammps-users] Question of bad bond computation

Dear LAMMPS users,

  I have a question about bad bond.
  Sometimes, because of large vander walls pairwise potential, the LAMMPS showed a waning "too long bond", or even aborted running with error "bad bond".
  However, sometimes lammps still run computation even if there is a too long bond. For example, in my simulation, average bond length is 1.1. Here is questioned data
  ITEM: BOX BOUNDS
  -5.02295 5.02295
  -5.8 5.8
  -10 10
  ITEM: ATOMS
  id type mol x y z ix iy iz vx
  811 1 25 -2.23361 3.36142 3.22614 45 -2 0 1.0029
  812 1 25 -1.39866 3.43565 2.49885 45 -1 0 1.61169
  The bond length between atom(id=811) to id(812) is
   811 812 11.7266
  Absolutely, this is too long bond. But lammps continued runs to the end.
  Could some one give the reason? And how to monitor this happenings and how to avoid it ?

Best wishes,
Yangpeng Ou

What kind of bond style are you using. Some styles
like FENE will throw an error when this happens.
Others like harmonic will just compute a large
force and continue, assuming the 2nd atom
in the bond is not so far away that a processor
can't find it. Either way, it's your responsibility
to insure you are running a reasonble model.
LAMMPS doesn't check.

Steve