← Back to team overview

yade-users team mailing list archive

Re: [Question #695192]: wall.append(O.bodies.append(pfacet) in triaxial test


Question #695192 on Yade changed:

    Status: Answered => Open

Hanying Zhang is still having a problem:
> I would suggest to use boxes for the walls since they are completely flat. It would minimize the number of problems for a start.
I have tried boxes for the walls and it worked well except boxes can't interact with cylinder, and that's why I use Pfacet for walls.

> First problem: triax.wallIds is undefined, so triax engine doesn't know where to find walls .
So I should define a wallIDs list then append Pfacet-wall to it?

> Second problem: you can't impose the motion of a PFacet, it always follow the motion of the nodes. One exception is if you clump facet+nodes into a clump for each of the walls. Then the clumps' ids should be used to define wallIds.
One question, there are 8 nodes and I need to generate 12 Pfacets, so every node would be reused if I clump pfacet+nodes into a clump for each of the walls. Is that workable?

Besides, I find only Ig2_GridConnection_PFacet_ScGeom can generate contact between cylinder and other kind of wall-like body, and I have been working on fiber-reinforced sand so cylinder is the kind of body that I must use. Don't really know is there any other body that I can use to simulate wall?
Or, can I just impose the motion of a Pfacet without triax engine?Is there any attribute like O.area.force(I make up it)? 
My idea is that:
(1)Generate sample by Pfacets pushing spheres and cylinders till they achieve some certain positions(in a smaller box) 
(2)triaxial test on sample.

Thanks for your help!

You received this question notification because your team yade-users is
an answer contact for Yade.