There are a couple options, none of which may
be exactly what you want:
a) there are granular potentials that include an attractive
force (like JKR potential), so that particles will "stick"
together when they collide. We have one of these
internally, but it hasn't yet been released in LAMMPS.
There may be one in LIGGGHTS which is built on top
of LAMMPS, and you can google for it.
b) LAMMPS can model rigid bodies composed of
multiple granular particles, e.g. dimers. But you
need to start with a set of rigid bodies, or pour them
into a box, etc. See the fix rigid/small and fix pour commands.
There is no option currently in LAMMPS to convert colliding
particles into rigid bodies on the fly.
> I deeply apologize if I have offended you in anyway.
you haven't. You're just asking Qs and getting advice.
Now you have some work to do, to figure out an appropriate model.