[I]Sorry, really long post. Trying to be as detailed as possible.[/I]
Can someone please look over my hyperelastic material model setup? I am having a hard time getting my model to converge when I refine my mesh to have a higher mesh density.
I am trying to simulate the indentation of a hyperelastic Neo-Hookean flat plate (made from ECOFLEX 00-30) with a rigid cylindrical indenter (Aluminum 6061) that has a flat bottom.
To validate the material model, I empirically determine the displacement of the indenter into my silicone flat plate by applying a force using a Mecmesin force tester.
When using Abaqus/Standard and hourglass control, my model converged with a mesh of 1800 C3D8RH elements. When comparing the Abaqus results against empirical, I had a 20% error. To check if my material model is correct, I refined my models mesh and generated a new mesh with 7935 elements of C3D8. But now, my model does not converge and error messages show that there is excessive element distortion in my model. Ive tried applying a small displacement (-1E-12) to ensure contact with the indenter and the silicone.
When I converted the model to Abaqus/Explicit with distortion control (enhanced), my analysis failed to run. I tried running with two different interaction methods (Kinematic and Penalty). With kinematic contact method, my analysis finished, but I saw unrealistic material behavior. This was just to try out explicit and understand the different types of method. With penalty contact method, my analysis failed to finish because there is only one element with excessive distortion. What gives? I cant apply ALE adaptive meshing because of hyperelastic material.
I triple checked all my units (English units lbs, in, psi) and found no issue. Am I missing something with my setup?
I can upload the .CAE or .inp files for more information.
Heres the setup.
[B]Model Setup Quarter Symmetry[/B]
The indenter is modeled as analytically rigid. (Assumption for analytically rigid is based on the Youngs modulus of the indenter being magnitudes larger than the hyperelastic materials modulus.)
The plate is modeled as a quarter section with Neo-hookean material properties for ECOFLEX 00-30. (Material has a Poissons ratio of 0.495 or 0.5 fully incompressible.)
[B]Boundary Conditions[/B]
Indenters reference point is fixed in all directions and rotation at initial step.
[INDENT]o During the analysis, I remove the BC of the direction where the applied load is located. (U3)[/INDENT]
Base of the plate is encastre.
Symmetry around U1 and U2 are placed to fulfill the symmetry model.
A surface to surface contact is applied with the following interaction properties:
[INDENT]o Frictionless[/INDENT]
[INDENT]o Hard Contact[/INDENT]
[INDENT]o Default enforcement[/INDENT]
Contact Initialization:
[INDENT]o Initial overclosures with resolve with strainfree adjustments.[/INDENT]
[B]Loading[/B]
Apply concentrated force at RP on indenter in the U3 direction in analysis step (Step-1).
[INDENT]o Ramp set up to be smooth step.[/INDENT]
[B]Mesh[/B]
C3D8 (Hexa elements) for mesh.
First model 1800 C3D8RH elements using Abaqus/Standard
Second model 7935 C3D8 elements using Abaqus/Explicit
Thanks in advance for your help.
-David