yade-users team mailing list archive

[Bruno Chareyre <bruno.chareyre@xxxxxxxxxxx>]

> > I'm a bit 
> > surprised, I though the plastic part was removed from total displacement 
> > (or ref point moved, which is the same).
> >     
> That is exactly what happens. Of course the result will be different,
> since (∫dus will be shear deformation at previous step, be it computed
> in any way, us is the current shear deformation us):
>
> Fs_tot=ks₁us=ks₁∫dus+ks₁Δus
> Fs_incr     =ks₀∫dus+ks₁Δus
>
>   
Heh... There is no friction in those equations, but I still see an effect!!
If I understand correctly now, u_s is total displacement 
(elastic+plastic, from contact creation to now), a "candidate" force is 
computed from it, and corrected according to the plastic rule. But if I 
change the plasticity criterion, the plastic part is computed 
differently, and the elastic part is apparently changing. Correct?

Initially, I thought the plastic part was really removed forever, 
meaning u_s from the previous step would have been only the elastic 
displacement. In that case, changing friction should not affect current 
forces.
In the incremental form, we keep fs, so that us=fs/ks is obviously the 
elastic part of the total displacement (*).

> (why do you need to modify friction coefficient? Isn't that supposed to
> be somehow constant?)
>
>   
No, its not supposed to be constant, because I want to change it! ;)

In my case, the point is to generate dense (5°) or loose (30°) packings 
depending on initial friction, while the end of the test is always done 
with the same 30°. Which means I need to change friction between phases 
of the simulation.

You can also imagine friction (or plastic limit) as a function of 
temperature, water content, weathering, etc.

Bruno