yade-users team mailing list archive

[Janek Kozicki <janek_listy@xxxxx>]

luc.scholtes said:     (by the date of Fri, 08 Sep 2006 12:30:09 +0200)

> It works well for the case of 2 spheres in interaction, but, of course, 
> there is a problem when a sphere and a box are concerned (no radius for 
> a box!).

Just a quick note: when sphere collides with a box, a "mirror" sphere of
double radius is assumed for the box. You can find a code for this in
InteractingBox2InteractingSphere4SpheresContactGeometry.cpp, lines:

  scm->radius1 = s->radius*2;

Still this file will be hugely modified after planned changes. All those
FIXME are indicators of design problems, that forthcoming changes will
solve.

Next week I should find more time to help you. So now you can try on
your own, if you have no success I'll try to help you, just when I find
some more time.... Sorry about that.

Please email later about your success or failure with this problem.

Your work here is very important because you are adding new kind of
interaction - redesign is done to solve this exact problem (we were
foreseeing this problem but we had only examples of gravity or magnetic
interactions, so it wasn't too convicing to make necessary design
changes...). But now when new design is finished you can be sure that
your capillary interaction will work :)

> Can I define (more properly) this condition of contact for every kind of 
> element (It is probably define somewhere, but where?) or do I have to 
> create "others" BoundingBoxes, specially for "without contact 
> interactions" (Is it possible and suitable?)?

solving this problem currently requires some kind of hack. Many options
are possible (all of them are a hack). You can put your checking whether
penetration depth is positive (overlapping occurs) in all necessary
places. You can add a new bool variable to Interaction class, that will
store information whether the interaction is a contact or on distance.
And perhaps few other solutions... Adding new class is also possible, so
you can even try that if you want.

I really can't wait when new design will allow doing that in clean (and
almost automatic!) way....

-- 
Janek Kozicki                                                         |
_______________________________________________
Yade-users mailing list
Yade-users@xxxxxxxxxxxxxxxx
https://lists.berlios.de/mailman/listinfo/yade-users