yade-users team mailing list archive

[Václav Šmilauer <eudoxos@xxxxxxxx>]

> so if I still want to do some simulation about boxes interaction, can i
> write the codes as RotatingBox?
> do i just need to add the following contents in my codes?
>
>          shared_ptr<InteractionGeometryMetaEngine>
> interactionGeometryDispatcher(new InteractionGeometryMetaEngine);
> interactionGeometryDispatcher->add("InteractingSphere2InteractingSphere4ClosestFeatures");
> interactionGeometryDispatcher->add("InteractingBox2InteractingSphere4ClosestFeatures");
> interactionGeometryDispatcher->add("InteractingBox2InteractingBox4ClosestFeatures");
>
>    
You can of course, but you also have to use constitutive law that uses 
ClosestFeatures. Make sure that you read code of 
InteractingBox2InteractingBox, but the code is quite hairy and I 
wouldn't trust it as-is: as far as I remember I found myself (sorry) 
fixing bugs in Olivier's code quite often; I wasn't at yade crew back 
then but it seems to me that the point was to prove how flexible yade is 
(or can be) by implementing many algorithms, but without necessarily 
scrutinizing them for their correctness.

Besides, you have to use FrictionLessContactLaw, which takes only 
penetration depth of bodies into account; therefore you gen no shear, 
for instance.

Yet another point may be the speed of box collisions; maybe replacing 
box by a few clumped spheres would be good enough for your purposes.

HTH, Vaclav