yade-users team mailing list archive

[Václav Šmilauer <eudoxos@xxxxxxxx>]

OK, I get it now, thanks for clarification. The code in
SimpleElasticRelationships is a little confusing, though correct; I was
veryfying those formulas accidentally this morning. It needs some
comment in the code.

v

> It is used in that sense, in the code I thougt the initial question was 
> reffering to.
> Sorry if the joke was obscure, I meant this naming was misleading.
> 
> One example in SimpleElasticRelationShip l.38 :
> 
> Real Va     = mat1->poisson;
> Real Vb     = mat2->poisson;
> 
> Real Kn = 2*Ea*Da*Eb*Db/(Ea*Da+Eb*Db);
> Real Ks = 2*Ea*Da*Va*Eb*Db*Vb/(Ea*Da*Va+Eb*Db*Va)
> 
> Which means :
> kna = Ea/Da, knb = Eb/Db
> ksa = poisson*kna, ksb = poisson*knb // ks = kn*nu
> 
> In short, if E, nu, and sizes are the same, it gives :
> 
> Kn = E/D, Ks=nu*Kn.
> 
> So, in the default triaxial test, what you put in the variable 
> "spherePoissonRatio", is used to define ks/kn.
> There is nothing wrong in the equations, for me at least, but I think 
> I'll rename this parameter "alpha" in the TT interface, so we don't have 
> confusion. I hope it clarifies the previous mail when I wrote this :
> 
> "For Poisson, the name is only due to the fact that Poisson's ratio 
> depends mostly on ks/kn. But there is no constant - not even linear - 
> relation between the micro and the macro in that case."
> 
> It means I named ks/kn "poisson" because it is the only variable (in the 
> contact law) that can modify nu. There is no proportionality between 
> ks/kn and nu, however.
> 
> I hope it clarifies. Sorry if I type too fast sometimes (I know I do it 
> a lot). I'm not asking anybody to change this name, I'm saying I could 
> change it.
> 
> Bruno
>