Yade

Using PFacet/facetBox

Asked by Nishant

In my simulations, sphere falls under gravity on a plate. The plate is modeled using facetbox. The sphere falls at the center of bottom. When the sphere touches the plate, I expect one interaction.

Here is a MWE:
#--------------------------------------------------

from yade import pack, qt, export
import numpy as np

#-----
##Defining Box
O.periodic=False

# No side boundaries, just a plane!
O.bodies.append(geom.facetBox(center=(0.5,0.5,0.5),extents = (0.5,0.5,0.5), wallMask= 16))

#-----
O.bodies.append([sphere((0.5,0.5,0.2001),radius=0.2)]) # add a sphere

#-----
##Engine Definition

O.engines=[
 ForceResetter(),
 InsertionSortCollider([Bo1_Sphere_Aabb(),Bo1_Facet_Aabb()],allowBiggerThanPeriod=True),
 InteractionLoop(
  [Ig2_Sphere_Sphere_ScGeom(),Ig2_Facet_Sphere_ScGeom()],
  [Ip2_FrictMat_FrictMat_FrictPhys()],
  [Law2_ScGeom_FrictPhys_CundallStrack()]
   ),
 PyRunner(command = 'output()', iterPeriod = 1, label = "checker"),
 NewtonIntegrator(gravity=(0,0,-9.81),damping=0.0)
]

#-----

def output():
 if (O.iter < 20):
  count_intrs = 0.0
  for i in O.interactions:
   if i.isReal:
    count_intrs += 1.0
  print "time:", O.time, "Sphere position", (O.bodies[-1].state.pos[2]), "contacts", count_intrs, "force", O.forces.f(O.bodies[-1].id)[2]
 else:
  print "done"
  O.pause()

#-----

O.dt = 5e-1*PWaveTimeStep()
O.run()

#--------------------------------------------------
#--------------------------------------------------

Using facetbox, I see two interactions and a stiffer contact. How to resolve this? Maybe Pfacet, as suggested to me here [1]? Anyone solved a similar problem?

Thanks

[1]: https://answers.launchpad.net/yade/+question/665675

Question information

Language:
English Edit question
Status:
Answered
For:
Yade Edit question
Assignee:
No assignee Edit question
Last query:
Last reply:
Revision history for this message
Bruno Chareyre (bruno-chareyre) said : 		#1

Hi,
This is a well-known downside of modeling one single surface as a collection of independent facets.
Easy solutions: use just one Facet for the plane, or a Box, or a Wall.
PFacets offer a more general solution which also applies to curved surfaces.
Bruno

Revision history for this message
Nishant (nishantk) said : 		#2

Thank you Bruno for your reply.

Actually, I had over-simplified the question in the post.

I am working on cyclic compression of railway ballast bed by concrete Sleeper using yade. The Sleeper has a height and covers ONLY a part of the bed. Therefore, I decided to use the facetbox to model Sleeper - suggested here as well[1].

There are possibilities of contacts at the edges and corners of the Sleeper. These contacts are stiff with Facetbox due to multiple contact points.

This is the problem I am hoping to resolve. Any suggestion in this matter on Sleeper modeling would be very helpful.

Regards,

[1]: https://answers.launchpad.net/yade/+question/230299

Revision history for this message
Chareyre (bruno-chareyre-9) said : 		#3

In that case PFacets seems to be the best option.
Cheers

Le sam. 5 mai 2018 10:22, Nishant <email address hidden> a
écrit :

> Question #668496 on Yade changed:
> https://answers.launchpad.net/yade/+question/668496
>
> Status: Answered => Open
>
> Nishant is still having a problem:
> Thank you Bruno for your reply.
>
> Actually, I had over-simplified the question in the post.
>
> I am working on cyclic compression of railway ballast bed by concrete
> Sleeper using yade. The Sleeper has a height and covers ONLY a part of
> the bed. Therefore, I decided to use the facetbox to model Sleeper -
> suggested here as well[1].
>
> There are possibilities of contacts at the edges and corners of the
> Sleeper. These contacts are stiff with Facetbox due to multiple contact
> points.
>
> This is the problem I am hoping to resolve. Any suggestion in this
> matter on Sleeper modeling would be very helpful.
>
> Regards,
>
>
> [1]: https://answers.launchpad.net/yade/+question/230299
>
> --
> You received this question notification because you are subscribed to
> the question.
>
>

Can you help with this problem?

Provide an answer of your own, or ask Nishant for more information if necessary.

To post a message you must log in.