yade-users team mailing list archive

[Luc Scholtes <scholtes@xxxxxxxxxxxxxxx>]

Hi Janek,

I have try something which is a little different from your suggestion in 
order to include my "capillary cohesion". It's in the 
"InteractingSphere2InteractingSphere4SphereContactGeometry" file. From 
my point of view, the problem was that the ContactGeometry (normal, 
Radius) was only define if "penetrationDepth>0" (l. 53 of this file). 
So, I have added few lines to define it even if "penetrationDepth!>0":

if (penetrationDepth>0)
   {
       scm->contactPoint = 
se31.position+(s1->radius-0.5*penetrationDepth)*normal;//0.5*(pt1+pt2);
       scm->normal = normal;
       scm->penetrationDepth = penetrationDepth;
       scm->radius1 = s1->radius;
       scm->radius2 = s2->radius;
              
       if (!c->interactionGeometry)
           c->interactionGeometry = scm;
  
       return true;
   }
   else    // add
       scm->normal = normal;
       scm->radius1 = s1->radius;
       scm->radius2 = s2->radius;
      
       if (!c->interactionGeometry)
           c->interactionGeometry = scm;
      
       return true;
       //return false;  // before add


I don't know if this is the good way, but it seems working, Spheres 
interact without contact. *Unfortunately,* *there is still a problem* (...):

The condition for them to interact (in "CapillaryCohesiveLaw" file):

// Intergranular distance :         
           Real D            =             
(de2->se3.position-de1->se3.position).length()-(currentContactGeometry->radius1+currentContactGeometry->radius2);


if (D<Dcreation)

{
       Vector3r Fcap = 100000*currentContactGeometry->normal;

                   
   static_cast<Force*>   ( ncb->actionParameters->find( id1 ,
   actionForce   ->getClassIndex() ).get() )->force    += Fcap;
   static_cast<Force*>   ( ncb->actionParameters->find( id2 ,
   actionForce   ->getClassIndex() ).get() )->force    -= Fcap;
              
      currentContactPhysics->prevNormal = currentContactGeometry->normal;

}

is not active, and the elements always interact when their 
boundingvolumes interpenetrate, even if "Dcreation" is smaller than the 
"boxSize" ("aabb->halfSize = 
Vector3r(2*sphere->radius,2*sphere->radius,2*sphere->radius). I don't 
understand the cause of this: "Is there, somewhere, a condition which 
cancel mine?", or, "Is my condition wrong?".

I hope I am clear enough (...), thanks in advance.

         So long, Luc.

_______________________________________________
Yade-users mailing list
Yade-users@xxxxxxxxxxxxxxxx
http://lists.berlios.de/mailman/listinfo/yade-users