yade-dev team mailing list archive

[Christian Jakob <jakob@xxxxxxxxxxxxxxxxxxx>]

Zitat von Bruno Chareyre <bruno.chareyre@xxxxxxxxxxx>:

> > I tried to understand how to calculate torsional moments for clumps.
> > Please have a look at the pictures in the attachment (see link above).
> > There two clumps are shown and I do not understand how their torsional
> > moments can be calculated. In the first picture there is box clumped
> > with a "half sphere". In the second picture you can see a box inside a
> > sphere, like I want to model my sand grains in future. How is
> > torsional moment calculated in this cases?
> I don't understand the question, sorry. Torsional/bending moments are
> defined for interactions, not for individual bodies (even if the bodies
> are clumps, it makes no difference). Your pictures have only one body,
> how could you define torsional moments?

Back to theory:
In DEM one gets a contact force F_c from the contact law. From F_c one  
gets angular moment M of the body (cross product with r_c and sum over  
all contacts c). The angular moment is given by multiplication of  
torsional moment J and angular acceleration a_a.

M = J*a_a     (1)

For spheres J = 2/5 * m * R^2, where m is mass and R is radius. Right?
Then one gets angular velocity by rearranging and time integration of (1).

a_a = int(5M/(2mR^2))dt (2)

In the case of a box J is a tensor, not a scalar. Calculation of  
angular velocities would be also no problem, just a little bit more  
complicated. (I hoped this is done by RotationEngine, but now I think  
its not.)

My question is, how can one obtain J (and I think one needs it to  
calculate angular velocities and displacements) in case of clumped  
box-sphere systems?

Or another question also comes in my mind: How is it done with clumped  
spheres?

> > Can this be simulated with RotationEngine already?
> RotationEngine is not related to your problem, it is only prescribing
> motion.
> > > For box-box interactions, there is a serious background in 3-DEC from
> > > itasca, where arbitray polyhedral shapes are modelized. It should apply
> > > to boxes.
> >
> > Hm, I think in the case of cubids (boxes) it should be much easier,
> > than a general polyhedral shape.
> Sure, but you will find interesting ideas in 3DEC regarding contact
> formulation.

Ok, I will have a look at it.

> > > My only question is if it is really necessary to modelize angular
> > > vertices instead of rounded shapes.
> >
> > If I want to model rounded shapes, I could clump spheres,
> Not really, a rounded box has flat surfaces that you will never have by
> clumping spheres. The contact between two flat surfaces will transmit
> moments.

Right, but isnt it too expensive in terms of CPU time? I mean a box  
has eight vertices and every vertex will be replaced by numerous  
triangles with a lot of vertices in the rounded case ...

> Bruno
>
>
> _______________________________________________
> Mailing list: https://launchpad.net/~yade-dev
> Post to     : yade-dev@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~yade-dev
> More help   : https://help.launchpad.net/ListHelp