yade-dev team mailing list archive

[Bruno Chareyre <bruno.chareyre@xxxxxxxxxxx>]

Vincent Richefeu a écrit :
> Hi everybody,
>
> I am very surprised that no viscosity is taken into account at contact 
> level in Yade.
> Maybe I'm wrong, but I have not seen anything like it!
> I do not usually use a global damping ('NonViscousDamping' in Yade) 
> and I prefer to
> use local (at the contact level) damping that can be related to a 
> 'coefficient of restitution'.
Global damping is the original Cundall's model damping, probably the 
most common. Its a numerical trick more than a physical ingredient, but 
I would say the same about the restitution coefficient (let the troll 
start ;-)).
Global damping is more efficient because it does not affect the critical 
timestep.

> Without going into sterile discussions, I would like to know if 
> something like that exists in Yade.
> Otherwise, I will need to add the class 'NormalViscousInteraction' 
> (and/or 'NormalViscousShearInteraction')
> that store the damping parameter. I will need also to add a relative 
> normal velocity parameter in the
> very complet class SpheresContactGeometry. Note that it will be the 
> possible to put aside the local
> vicosity by setting the damping parameter to zero.
No need to add relative velocity, you can compute vel(1)-vel(2) when you 
need it, and save memory.
Viscous creep as been implemented by janek, but the type of viscosity 
you want is not implemented yet.
You could just add it in ElasticFrictionalContactLaw, or even better, 
make a contact law inheriting from EFCL so people who don't want 
viscosity will use the current law without the need to set it to 0 and 
without useless if(viscosity!=0).

Bruno

>


--

_______________
Chareyre Bruno
Maitre de conference

Grenoble INP
Laboratoire 3SR - bureau E145
BP 53 - 38041, Grenoble cedex 9 - France
Tél : 33 4 56 52 86 21
Fax : 33 4 76 82 70 43
________________