Yade

High Young Modulus

Asked by Daniel Enrique Morales Molina

Hi all,
I have a problem with high Young modulus E. When I reach de GPa order, the hexagonal pack just explodes!(cohesion lost) . I've tried to decrease the timestep and increase the normalCohesion but still the packages explodes.
How can I get a stable pack material with high Young Modulus?
Thanks!

Question information

Language:
English Edit question
Status:
Solved
For:
Yade Edit question
Assignee:
No assignee Edit question
Solved by:
Luc Scholtès
Solved:
Last query:
Last reply:
Revision history for this message
Daniel Enrique Morales Molina (dmorales-ing-uchile) said : 		#1

None had this problem before?

Revision history for this message
Anton Gladky (gladky-anton) said : 		#2

Hi.
Usually it is caused by large timestep. Try to decrease it.

Anton

Revision history for this message
Daniel Enrique Morales Molina (dmorales-ing-uchile) said : 		#3

Thanks Anton,
I've tried that, i've reduced the timestep to O.dt=0.0001*utils.PWaveTimeStep(), but still have the problem. Some spheres even go through the facets!.

Revision history for this message
Anton Gladky (gladky-anton) said : 		#4

Did you try to reduce a strain rate ("velocity" of loading)?

Anton

Revision history for this message
Daniel Enrique Morales Molina (dmorales-ing-uchile) said : 		#5

Hi Anton,
What do you mean with velocity of loading?,
I'm not applying external forces,
 I'm just tryng to make an arrangement of spheres, with properties similar to a rock. In this case a cube of spheres (trying to model a rock with a block shape).
The spheres are inside of facets (a cube of facets), and when a reach the Gpa order, the pack explodes. No big problem with de Mpa.

Daniel

Revision history for this message
Best Luc Scholtès (luc) said : 		#6

Hi Daniel,

again, not sure to get to the point, but, from what I remember you work with regular packings generated with a certain overlapping (?). If the ratio between the overlapping and the "Young modulus" is high, it could explain your problem. Your packing is not stress free initially and the repulsive interparticle forces could produce the explosion. Have you tried with low values of the Young modulus? Have you experienced the same problem?

Please let us know.

From my part, I also work to simulate rock like material and I use the CohesiveFrictionalPM in order to deal with initially stress-free assemblies. For instance, CFpm utilizes the initial overlapping as the equilibrium distance to compute the normal force Fn=k*(D-Deq), whereas the CohesiveFrictionalContactLaw uses the "real" overlapping between the spheres Fn=kD.

Cheers

  Luc

Revision history for this message
Daniel Enrique Morales Molina (dmorales-ing-uchile) said : 		#7

Hi Luc, and thanks for your answer,
I'm working with a geometric and regular packing (pack.regularHexa). Based on the documentation (until now bzr2899): "spheres touch exactly, there are no overlaps (what some people call “stress-free” packing)".
But, how can I be sure?

I've tried with low values of Young modulus (3e7 Pa) and works: the material keeps the cohesion.

I will try as you says (use the CohesiveFrictionalPM)

Thanks,
Daniel

Revision history for this message
Daniel Enrique Morales Molina (dmorales-ing-uchile) said : 		#8

Hi all,
I've changed to CohesiveFrictionalPM (as Luc said) and now works fine.
Thanks!

Revision history for this message
Daniel Enrique Morales Molina (dmorales-ing-uchile) said : 		#9

Thanks Luc Scholtès, that solved my question.

Revision history for this message
Bruno Chareyre (bruno-chareyre) said : 		#10

Should we conclude that this statement is wrong?

"spheres touch exactly, there are no overlaps (what some people call “stress-free” packing)".

Revision history for this message
Daniel Enrique Morales Molina (dmorales-ing-uchile) said : 		#11

Hi Bruno,
The only explanation that I can find is that the the hexagonal packing produces an overlap . So, yes I think the statement is wrong.
Maybe, there is another method to conclude that, by measuring the overlap. I don't know how to do that.

Regards,
Daniel.