yade-dev team mailing list archive

[Vincent Richefeu <richefeu@xxxxxxxxxxxxxxx>]

Hi everybody,

Just a few comments (as usual, do not feel obliged to agree with me!)

I understand (I think) what Chiara means. This is something that bothers me a lot (mentally) and, unfortunately, it leads me to continue using my poor bits of code instead of yade (please, do not be offended by that. I love yade and I'm not being sarcastic). We have already spoken when Vaclav came to a seminar in Grenoble. 

I said:
A parameter (physical or not) like mu (friction coefficient) has nothing to do within the bodies. This is an "inter-bodies" parameter and the bodies are rigid.

The answer was:
Right, but with this approach you can do what you want. For instance, if you want mu=0.4 between A-type bodies, and mu = 0.2 between A-type and B-type bodies you can do: mu_A = 0.4, mu_B = 0.0 and then mu will be defined as the mean value.

I agree with that. It works fine. But I see two disadvantages:

1 - mu_A and mu_B have lost their physical meaning (if they had)
2 - I will have problem if someday (after bzr update) the mean rule if changed for a max rule by someone well-intentioned

I understand that my comments are boring. Do not blame me please. The goal is to help (maybe) but not to criticize.

Vincent

Le 8 déc. 2010 à 15:59, Bruno Chareyre a écrit :

> 
>> For completeness I should have included also Ks/Kn. They are all
>> parameters of the interaction, however do they have a specific
>> physical-basis? I am not saying that it is wrong to employ them, that is
>> indeed a common assumption in a dem model.
> 
> I'm too sure what is a "physical" parameter (for some physicists, Young modulus and
> internal friction angle are not physical parameters but engineers tricks).
> 
> We have constitutive relations (the mathematical form) and constitutive parameters (the
> constants in there). It is logical to define constitutive parameters at the bodies level.
> Just giving a very practical reason here : I have particles of type A and B, and I want
> different values of friction for interactions A-A, A-B, and B-B (don't ask why there
> should be different, it is my constitutive assumption). How could I achieve that if
> friction was defined in the Ip functor directly?
> Be it friction, kn, ks, ktw, kroll or adhesion, it makes no difference.
> 
>> As we all know, those
>> parameters are generally quantified to obtain realistic behaviour at the
>> macro-scale.
> 
> Generally, not always. You, for instance, are not doing that.
> 
> 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