9 Replies Latest reply on Feb 7, 2009 3:13 PM by Bill McEachern

Nonlinear Indention Simulation in CW2008

Eric Hill Feb 5, 2009 4:13 PM

I am trying to use COSMOS 2008 Adv Pro to model a flat indentor pushing into a semi-infinite plane of a relatively soft material. The load on the indentor is high enough to yield the soft material in the plane, which leaves behind a permanent impression in the soft material when the indentor is withdrawn. As an output of the analysis, I would like to produce a load-displacement curve for the indenting operation.

The indentor material is a linear elastic material with a high elastic modulus. The soft plane material is a non-linear with a defined stress-strain curve that looks like a classic plastically deforming metal. The yield strength of the plane material is significantly below the indentor's yield strength.

My non-linear static model behaves well thoughout the elastic range, but when the plastic deformation becomes significant, the simulation diverges and the solver chokes.

Is this typical behavior for COSMOS Non-linear when dealing with these sorts of problems? It's my interpretation that the simulation runs smoothly until the deformation of mesh in the soft material is big enough that the deformation of the mesh elements reaches some critical value. I've seen some papers describing using FEA to simulate this type of phenomenon, but non using any of the COSMOS products.

Please let me know if you have any experience with running this type of simuation. Your input will be much appreciated.

Thanks,
Eric

Nonlinear Indention Simulation in CW2008

Bill McEachern Feb 5, 2009 4:34 PM (in response to Eric Hill)

Hi Eric,

Cosmos can only handle a max strain in the order of 10%. The other thing you should check is that you need a "true stress strain" relationship. It is easy to convert from and nominal or engineering stress strain relationship. Just put trues stress strain into google and it comes up pretty quick. The other thing I try to do is go with a bilinear model first, which in reality is a pretty decent approximation of a yielding metal for a lot of applications. You never want to have a true negative stiffness. If that works then I woudl say you have a problematic stress strain curve. If it doesn't well things get tougher as large element distrotions can cause problems. Cosmos does not support in solution remeshing so on that one I have no help that I know works. You could try mesh tuning - trial and error - and see if anything gets the solution to advance.
Like Show 0 Likes(0)

Actions
Nonlinear Indention Simulation in CW2008

Pete Yodis Feb 5, 2009 5:18 PM (in response to Eric Hill)

Can't answer for Cosmos (Simulation), but I think Abaqus would handle this. Abaqus is a DS Systemmes product now and I think can read in SolidWorks files.
Like Show 0 Likes(0)

Actions
- Nonlinear Indention Simulation in CW2008
  
  Pete Yodis Feb 5, 2009 5:40 PM (in response to Pete Yodis)
  
  http://www.simulia.com/products/abaqus_explicit.html
  
  Like Show 0 Likes(0)
  
  Actions
Nonlinear Indention Simulation in CW2008

Eric Hill Feb 5, 2009 5:21 PM (in response to Eric Hill)

Yes, Abacus would definitely do it. I've seen technical papers that describe simulations in Abacus that are very similar to what I am trying to accomplish in COSMOS (Simulation).
Like Show 0 Likes(0)

Actions
- Nonlinear Indention Simulation in CW2008
  
  Pete Yodis Feb 5, 2009 5:50 PM (in response to Eric Hill)
  
  papers like this?...
  
  http://www.simulia.com/academics/research_elastic.html
  
  Like Show 0 Likes(0)
  
  Actions
Nonlinear Indention Simulation in CW2008

Eric Hill Feb 5, 2009 6:00 PM (in response to Eric Hill)

Yes, indeed. Though my material models are a bit simpler.
Like Show 0 Likes(0)

Actions
- Nonlinear Indention Simulation in CW2008
  
  Pete Yodis Feb 6, 2009 12:45 PM (in response to Eric Hill)
  
  Good Luck
  
  Like Show 0 Likes(0)
  
  Actions
Nonlinear Indention Simulation in CW2008

Anthony Botting Feb 6, 2009 1:03 PM (in response to Eric Hill)

It might be worth examining the "hyperelastic" material model in Simulation, as I believe the elements can strain quite a bit more than the other models. However I do know that some elements get so distorted that the math breaks-down and the solver routines don't like that. I know ANSYS package has a re-meshing algorithm that stops at a point where the mesh is too distorted, re-meshes the domain, applies all the known strain fields and continues. This package is (last I check) about $30K.
Like Show 0 Likes(0)

Actions
- Nonlinear Indention Simulation in CW2008
  
  Bill McEachern Feb 7, 2009 3:13 PM (in response to Anthony Botting)
  
  ANSYS would not be on my list of codes for this class of problem - NL is not their strong suit though they do have some capabilities......ABAQUS, MARC, AMPS or LS Dyna would all be better choices for comparable money with ABAQUS being the most robust of the lot in my experience.
  
  Like Show 0 Likes(0)
  
  Actions

Go to original post