yade-users team mailing list archive

["Hadda Nejib" <nejib.hadda@xxxxxxxxxxx>]

All right Chiara  ! Thank you for these clarifications.

N


-------- Message d'origine--------
De: yade-users-bounces+nejib.hadda=cemagref.fr@xxxxxxxxxxxxxxxxxxx de la part de chiara modenese
Date: mer. 18/08/2010 16:37
À: yade-users@xxxxxxxxxxxxxxxxxxx
Objet : Re: [Yade-users] relative normal and shear displacements andrelative rotations
 
Hi Nejiib,

I give you my answers though Bruno could correct me if I say something wrong
or could add whatever I am missing.

>
> 1- Yade uses 3 entities to define the contact normal, "normal",
> "prevNormal" and "currentNormal". in what does currentNormal differ from
> "normal".
>
In fact, if you refer to ElasticContactLaw, we do not need prevNormal
anymore. We use only normal, which is updated into precompute function only.


> I think also that "prevNormal" is overwritten by "normal" or inversely
> when displaying them because there is no differences between the 2
> vectors (you can see that in an .xml file for example). so if it is the
> case I think we should fix it.
>
prevNormal is still begin used in few parts of the code (e.g.
CohesionlessMomentRotation.hpp/cpp) and I do not think we can get rid of it
atm.

>
> 2- While computing the shear increment extracting these componnents
> (vel,angVel,contactpoint,Pos1,Pos2,normal, O.dt), computing vectors c1x
> and c2x as follow
>
You should get the same result. Consider that the bool
avoidGranularRatcheting in ScGeom is True by default. If that is your case
(because you did change it) then the two results are rightly different. See
doc in ScGeom for some explanations about granular ratcheting.

cheers, Chiara


>
> c1x= ContactPoint-pos1
> c2x= ContactPoint-pos2
>
>
> and using this code (Same one used in ScGeom) :
>
>
> relVel=vel2+(angVel2).cross(c2x)-(vel1+(angVel1).cross(c1x))
>
> delta_un=(relVel*normal)*O.dt
>
> shearVel=relVel-(relVel*normal)*normal
>
> ShearInc=shearVel*O.dt
>
>
> I found that ShearInc computed as above differs from the ShearInc given
> by Yade bzr (Last version).
> Is there any projection, or operation on coordinate local/global systems
> left for me to do so that I find the same value as Yade.
>
> Many thanks !
>
>
>
>
>
>
>
>
>
>
>
> Le lundi 16 août 2010 à 15:28 +0200, chiara modenese a écrit :
> > Hi Nejib,
> >
> > sorry for long delay in replying.
> >
> >         'shearIncrement' seems not to be an attribute to the object
> >         'ScGeom', Do
> >         you agree ? is there any alternative to call directly
> >         otherwise?
> > May I guess what release are you using? In last release shearIncrement
> > is accessible via python in read-only mode (see line 54 of ScGeom.hpp)
> > which overrides shearInc private attribute of ScGeom (updated into
> > precompute function, line 35 of ScGeom.cpp).
> >
> > cheers, Chiara
> >
> >
> > _______________________________________________
> > Mailing list: https://launchpad.net/~yade-users<https://launchpad.net/%7Eyade-users>
> > Post to     : yade-users@xxxxxxxxxxxxxxxxxxx
> > Unsubscribe : https://launchpad.net/~yade-users<https://launchpad.net/%7Eyade-users>
> > More help   : https://help.launchpad.net/ListHelp
>
>
>