Using Elmer to model silly putty deforming

General discussion about Elmer
Post Reply
student
Posts: 1
Joined: 05 Dec 2018, 01:02
Antispam: Yes

Using Elmer to model silly putty deforming

Post by student »

Hi all,

For a school project my partner and I are trying to model the viscous deformation of silly putty. We would like to model how, over time, a conical silly putty object would deform to a flatter, pancake-shaped object. We were thinking we would use the Navier-Stokes solver capability of Elmer, but are having trouble finding examples to build our project off of. We did run the examples listed on the courses material page on the Elmer ice website. One of my primary questions is wether it is possible to actually accomplish this project with Elmer? Also, can changing shapes in the x and y directions be modeled with Elmer (the cone of silly putty expands horizontally as it flattens into a pancake shape)? Any help is appreciated! Thanks!
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: Using Elmer to model silly putty deforming

Post by raback »

Hi student

The problem seems simple but to model it may be rather difficult.

You can either use a Eulerian or Lagrangian method to model the deformation. Eulerian method has a fixed grid and can deal with all kinds of geometries. However, the shape is define by a signed distance function so it is never too sharp and you need a lot of element to have decent accuracy.

Lagrangian methods allow the mesh to deform. The downside is that the quality of the mesh quickly deteriorates. For example, if you really have cone you introduce a corner point that probably will be difficult to track. However, having a ~half-sphere could be numerically more stable. Then you could probably maintain some mesh quality when the height decays by a factor of few.

Ideal method would combine the Lagrangian approach with adaptive remeshing.

Also there is a question whether the fluid is really viscous or perhaps you need a visco-elastic model. There is one recent one in Elmer but I fear it is not up to large displacement cases.

All this said, you might try to start from test cases "freesurf_sloshing_2d". Replace the mesh with half-sphere and use some slip-coefficient for the ground. If you use no-slip velocity your mesh will have problems much earlier.

The Elmer/ICE community could have even more suitable starting points. Google for the "Bueler Elmer ice" for some info. The Bueler profile is an analytical shape for a steady-state toy glacier. Maybe somebody in the community would have a pointer to some test case.

-Peter
Post Reply