UMAT, shellsolver, elasticsolve

General discussion about Elmer
Post Reply
Jose97
Posts: 2
Joined: 24 Apr 2023, 22:24
Antispam: Yes
Location: Spain

UMAT, shellsolver, elasticsolve

Post by Jose97 »

Hello to the Elmer community,
I would like to carry out a simulation of a 2D mesh (shell), applying normal pressure to the surface, if UMAT cannot be implemented with shellsolver, is it posible to do this kind of simulation with elasticsolver?
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: UMAT, shellsolver, elasticsolve

Post by raback »

Hi Jose

What kind of material law would the UMAT have? ElasticSolver does indeed have UMAT interface which to my knowledge is lacking in the ShellSolver.

-Peter
Jose97
Posts: 2
Joined: 24 Apr 2023, 22:24
Antispam: Yes
Location: Spain

Re: UMAT, shellsolver, elasticsolve

Post by Jose97 »

Hi Peter,

My material is the Holzapfel model, a hyperelastic anisotropic material for biological tissues application.
It depends on 6 input material parameters (PROPS) that are supposed to vary the stress-strain curve shape, depending on their values.

I noticed that ShellSolver uses St Venant hyperelastic material, so I can only work with ElasticSolver. I've already finished implementing my UMAT model, and I'm running test simulations of a cubic mesh with hexahedral elements that elmer offers as an example, constraining displacement on 5 faces and applying tensile force on the sixth (attached are some results visualization).

Unfortunately, the stress-strain response I get is practically linear and my UMAT material is highly nonlinear, can elmer's ElasticSolver calculate this type of highly nonlinear behavior materials? if so, then I am doing something wrong in my UMAT.

Jose
Attachments
Simulation_cubic_mesh.png
(815.75 KiB) Not downloaded yet
kevinarden
Posts: 2237
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: UMAT, shellsolver, elasticsolve

Post by kevinarden »

There are two types of non-linear behaviors in solid mechanics, geometric large displacements which elasticsolver does. And there is material non-linearities which is up to the material model to do. Elmer does not have built in material non-linear models it is up to the UMAT to handle it.
You are not go to see the non-linear effects is a steady state analyses, the output is only at the end. If you want to see the non-linear behavior it needs to be transient so that every step is output. See tutorial 8.
https://www.nic.funet.fi/pub/sci/physic ... orials.pdf

However, UMAT is not supported in transient, so simulation type scanning is used to get a timestep type output.
Usually the load is ramped up with a function in a material non-linear situation for example
Force 1 = Variable Time; Real MATC "0.006*tx"
Post Reply