# How can I model a sand with rolling resistance

As somebody has pointed out, if we want to model sand dilatancy using sphere type assembladge, rolling resistance should be included.

I have read Plassiard's paper, but moment type contact law in Yade seems useless to model rolling resistance.

1, How can I model a rolling resistance with linear normal and tangential contact law?

2, As pointed by Christian Jakob, it seems he succeed in model sand with hertz contact law with rolling resistance,
but how can he modify the micromechanical parameters duiring the deviatoric compression process after isotropic compression process ? As I knew, only particle friction angle could be modified, how can I modify kn ks and kr?

## Question information

Language:
English Edit question
Status:
For:
Assignee:
No assignee Edit question
Last query:
2012-07-11
2012-07-11
 Bruno Chareyre (bruno-chareyre) said on 2012-07-11: #1

I see that the documentation of this law does not reflect the current state of the code. It will have to be updated. This sentence is no longer true:
"There is no maximum value of moments in the current implementation, though they could be added in the future."

Thanks to Chiara, we are now in the future. A max moment is defined as a rolling friction threshold.

 Jérôme Duriez (jduriez) said on 2012-07-11: #2

Hello,

I would not say that rolling resistance is required to simulate dilatancy (you will achieve this through the density of your sample). Rolling resistance is used generally to increase the macroscopic friction angle. It could be considered that it is a way approaching real geomaterials, despite the spherical shape of the particles.

2. About changing micromechanical parameters, the problem is that the stiffnesses of the interactions are computed (from the material properties) only once, when the interaction is created.

Then, if you change "during" a simulation the properties of the materials, it will not have any consequence. Looping on all interactions to take into account the changes is required.

 wangxiaoliang (wangxiaoliang) said on 2012-07-11: #3

Thanks, Chareyre and Jduriez, in which version yade, you have included plastic moment for rolling resistance?

another question, if we just add a resistance of rolling, I think 3 dof is enough, 6dof may cause high computation time?

 Jérôme Duriez (jduriez) said on 2012-07-11: #4

> another question, if we just add a resistance of rolling, I think 3 dof is enough, 6dof may cause high computation time?

If you want indeed to use Law2_ScGeom6D_CohFrictPhys_CohesionMoment, it uses the 6 dof (it's written in the name of the law, see "ScGeom6D")