DMT and van der Waals adhesion between macroscopic granules

Greetings,

I was wondering if anyone knew how to simply implement the DMT contact
model into Lammps, this simulation would be for micron sized granules
which still experience the van der Waals force. Currently,
granular/hertz/history potential allows for the simulation of the
hertzian contact model. However, it is not possible to use the
colloid potential in conjunction, as when particles come to contact
the force grows without bound. Here I am only interested in the van
der Waals adhesive potential between macroscopic spheres, where

d_0. d_0 is the theoretical interatomic distance, this distance is

what fully defines the van der Waals potential at contact, so that it
is not allowed to grow unbounded as particles come into contact.

In addition, I would like to add a constant adhesive force at contact,
as in agreement with DMT.

If some one knows of an easy fix, to say the colloid and granular
potentials, I would greatly appreciate it.

Thank you,
Eric

eric,

have you checked out the capabilities of LIGGGHTS?
http://www.cfdem.com/
it is a branch of LAMMPS with many enhancements
for granular material simulations.

cheers,
    axel.

I have done some simulations with van der walls along with
granular/hertz/history, when the particles are closer than d_0 one could
assume that the adhesion stays same as at d_0. That way one could avoid
the singularity in van der walls at d_0. For my application it worked
fine.

Deena

You can also ask Jeremy Lechman at Sandia (CCd)
who has developed granular models with adhesion.

Steve

Hi Eric,

   In fact, I think you can use the colloid pair potential with the granular/hertz/history if you use the pair hybrid command. Is this not so? If you are not interested in friction and, therefore, rotation of the particles you can just use pair_colloid as there is a short-ranged repulsion in that model.

   In addition, I (and a postdoc) have implemented a JKR model in granular lammps (change to DMT - trivial), but it is not in the general release as yet. We were planning on that by the end of the year.

   Finally, just as a anecdotal note, I have found it non-trivial to get the van der Waals logn-range attraction model to match smoothly with JKR/DMT adhesion + Hertzian repulsion. I am aware with a good deal of work on this, but I think the details are far from settled.

   Would be happy to chat with you more on this issue.

Best,
Jeremy

Thank you Jeremy. The application I was hoping to use lammps for was
the homogeneous shearing of dry adhesive particles. I am interested
in the friction resulting from rotation in these flow. Whats more is
that I will likely be using the Hookean model rather than the hertzian
in order to model small dissipation due to thermoelastic waves, which
should be possible assuming the DMT contact force model. So, I would
like to facilitate the deformation in granules.

I have tried the pair-hybrid approach, but for my weakly adhesive
particles, overlap tended to halt my simulations as the force
singularity was reached. Note I did try this with the repulsive part
activated, though in my simulations I will need set this to 0. What I
would like to do is simply truncate the attractive part of the colloid
potential and then keep it constant throughout the collision.

Could you clarify on that last point? It may speak to something I
have noticed, the nonsmooth transition from vdW potential to
DMT+Contact force tends to produce large errors (which converge very
slow) in terms of conservation of energy.

Thank you Jeremy. The application I was hoping to use lammps for was
the homogeneous shearing of dry adhesive particles. I am interested
in the friction resulting from rotation in these flow. Whats more is
that I will likely be using the Hookean model rather than the hertzian
in order to model small dissipation due to thermoelastic waves, which
should be possible assuming the DMT contact force model. So, I would
like to facilitate the deformation in granules.

I have tried the pair-hybrid approach, but for my weakly adhesive
particles, overlap tended to halt my simulations as the force
singularity was reached. Note I did try this with the repulsive part
activated, though in my simulations I will need set this to 0. What I
would like to do is simply truncate the attractive part of the colloid
potential and then keep it constant throughout the collision.

Could you clarify on that last point? It may speak to something I
have noticed, the nonsmooth transition from vdW potential to
DMT+Contact force tends to produce large errors (which converge very
slow) in terms of conservation of energy.