← Back to team overview

yade-users team mailing list archive

Re: [Question #242137]: Periodic boundaries on given surfaces of the cell

 

Question #242137 on Yade changed:
https://answers.launchpad.net/yade/+question/242137

    Status: Open => Answered

Kneib François proposed the following answer:
Hi,

It is possible to generate a parallelepipedic cell by introducing non-
diagonal terms in O.cell.hSize :

Example :
O.cell.hSize=Matrix3(1.,1.,0.,1.,1.,1.,0.,1.,1.)

If you don't want to do that, you can make a rectangular cell oversized
in the  non-periodic directions. Then add some very large walls or boxes
to contain spheres for this two directions. The trick is to set the
"allowBiggerThanPeriod=1" flag into the collider :

Example :
InsertionSortCollider(
			[Bo1_Box_Aabb(),
			Bo1_Sphere_Aabb()],
			allowBiggerThanPeriod=True,
			verletDist=0.)

So it's a weird concept : walls will be non-periodic but spheres will.
That's the reason why you will need to oversize walls, if your spheres
travel through N perdiod in the periodic direction, the walls size must
be at least N*period_size.

François

-- 
You received this question notification because you are a member of
yade-users, which is an answer contact for Yade.