Hi all,
I am trying to simulate a system, where the particles are bonded to
their neighbors via harmonic potential. Now, after the run is over I
found rest of the parts of my system is giving me desired result
except a smaller percolated cluster part. Then I found that for this
smaller part, in my script I have unfortunately typed the bonds (for
this smaller section only) twice where the rest of the part is left
correct.
My sincere asking whether declaring a bond twice will add up bond
energy twice (for this smaller section only)for those cases where it
has been declared twice or LAMMPS overlook it (as it has been declared
before)?
I think so, but I don't think this is likely to be the source of the
uncanny behavior in your simulation.
In any case, it is easy to check if this is so.
Try fixing the DATA file, and running LAMMPS for one time step to see
if the bond-energy changes. In other words, try:
thermo_style custom step pe epair ebond eangle edihed
thermo 1 # <--you could put any number here
run 1
Then fix the problem with your DATA file and try "run 1" again.
LAMMPS should print out the bond-energy at time t=0 (4th column). If
this number is different, then fixing your DATA file (removing the
extra bonds) will reduce the bond energy. You will see the
difference. You can also compare the other contributions to the
energy (pair, angle, dihedral) to see if they were modified by the
changes you made to the DATA file.
Otherwise I couldn't find any other reason to justify the
uncanny behavior.
You suggested that if there are two bonds between the same atoms, that
the energies and forces between the atoms would add together. I
think this might be true. This would cause the bond to be stiffer
(stronger). But even if this is true, I doubt this would cause
problems. Usually when a harmonic bond is too stiff, the worst thing
which can happen is a numerical explosion. If that occurred, you
would know about it. (Your simulation would die. LAMMPS would print
out messages about missing atoms or bonds, or bonds being too long and
then die.)
I suspect there may be other problems you have not found yet. But try
fixing the problem and see if it helps.
Cheers
Andrew