# solve the overlapping of a set of spheres initially distributed randomly in the simulation box

Hello everyone,

I need help to properly solve the overlapping of a set of atom_style spheres with a radius 2.5sigma each and are initially distributed randomly in the simulation box.

For spheres of radius 0 (particles) could consider lj/cut repulsive potential , defining a cutoff 2^(1/6)sigma and then apply an integration : fix nve/limit. However I not understand how to properly apply this idea of particles with a 2.5sigma radio, considering that initially there will be a shared volume when they overlap.

It could be considered a potential displaced LJ the surface of each sphere , however How to fix the overlap in the initial case where shared volume will exist between the spheres?

How to accomplish this in lammps ?

Regards

Hi,

I need help to properly solve the overlapping of a set of atom_style
spheres with a radius 2.5sigma each and are initially distributed randomly
in the simulation box.

For spheres of radius 0 (particles) could consider lj/cut repulsive
potential , defining a cutoff 2^(1/6)sigma and then apply an integration :
fix nve/limit. However I not understand how to properly apply this idea of
particles with a 2.5sigma radio, considering that initially there will be a
shared volume when they overlap.

It could be considered a potential displaced LJ the surface of each sphere
, however How to fix the overlap in the initial case where shared volume
will exist between the spheres?

How to accomplish this in lammps ?

You should first choose a strategy, as lammps is flexible in this respect.

As an example, you can start with a simulation box that is too large to reduce
strongly the number of overlaps, use low epsilon value for a repulsive potential
and use fix nve/limit to equilibrate then (possibly in combination) fix deform
to reach the target box size. You should also perform equilibration after the
final box size has been reached.

Regards,

Pierre

Hello everyone,

I need help to properly solve the overlapping of a set of atom_style
spheres with a radius 2.5sigma each and are initially distributed randomly
in the simulation box.

For spheres of radius 0 (particles) could consider lj/cut repulsive
potential , defining a cutoff 2^(1/6)sigma and then apply an integration :
fix nve/limit. However I not understand how to properly apply this idea of
particles with a 2.5sigma radio, considering that initially there will be a
shared volume when they overlap.

It could be considered a potential displaced LJ the surface of each sphere
, however How to fix the overlap in the initial case where shared volume
will exist between the spheres?

How to accomplish this in lammps ?

​there are two commonly applied strategies available in LAMMPS to deal with
overlapping particles.

1) use the overlap option of the delete_atoms​ command to delete particles
that are too close to others.

2) start a simulation with a soft core potential (pair_style soft) and use
fix adapt to slowly unoverlap the particles through raising the height of
the soft core and then switch to the pair_style of your desire. see the
micelle example for a demonstration.

​axel.​