2d cohesion

Submitted by rbj on Mon, 08/20/2012 - 17:41

Dear all,

For a CFDEM project I tried to model interacting particles in a converging channel, to predict clogging behavior. To realise this I suspect to need cohesion behavior of the suspended phase.
In a 2d situation, also for a pure DEM calculation I get triangle neighbor list problems, even for very small time steps (1e-12) and particle 'explosion'; particles disappear when colliding.

Is cohesion in 2d unobtainable or is something else wrong? Attached my case and stl files.

Particles are quite large in comparison to channel size, just to check the behavior.

edit: tarred the attachments

RBJ

Forums:

rbj | Tue, 08/21/2012 - 18:07

version 1.5.3, may be I should update to version 2 indeed.

I should remark that the cohesion example works without problems.

By the way I'm calculating it 'quasi 2d': on a 3d system with all particles inserted in a plane, and the {y} << {x},{z}.

This to accomodate for the single cell depth of OpenFOAM. (so using a single cell sized geometry, but inserting particles on the y-plane, so that they interact 2d).

I just noticed the bugfix for version 1.1.6:
"[Fixed two issues with the cohesion model - there was a bug in calculating the contact area. Also added a check to ensure the model is only used for 3d simulation. Thanks to George Marketos (Imperial College London) for calling attention to this]"

so this suggests still only 3d should be done for cohesion?

rbj | Wed, 08/22/2012 - 14:22

continuing on this subject:

What happens in 3d if 2 particles accidently collide in plane? (so position and velocity vector for 1 dimension is equal for both particles)