Negative Jacobians - spherical indenter

[jesicandrea]

Hello everyone,

I am trying to run a model of 3mm of spherical indentation on an isotropic elastic soft tissue. I have remeshed the contact zone and modified: two pass, auto penalty and lagrangian in the contact; the type of contact: facet on facet and node on facet for sliding contact; I am trying tension compression interface too; the time step also was changed to 1000 with a step size of 0.001. But it always dies for the depth of the indentation.

If someone knows what I am doing wrong... please tell me.

monocapa1.zip

Thank you!

[maas]

Hi,

I've attached a slightly modified version that runs for all the way to normal termination. I've tweaked a few parameters:

1. Turned off the must-point curve for dtmax since this causing too large time steps. (I could have modified this curve too, but felt it was easier to modify the time stepping parameters directly.)
2. I've turned on the auto-penalty. This usually is a good place to start. Then you can tweak the penalty parameter to get better convergence.
3. I turned on augmented Lagrangian. After tweaking penalty parameters that get me close, I turn this on to reduce contact gap further.

Like I said, this file works for me now, but please give it a try on your end and let us know if you have any further questions.

Cheers,

Steve

[jesicandrea]

Hello Steve,

You are right, the model run and the file was very useful. Thank you very much.

I have one more question: the deformable box is neo-Hookean, I thought it could be linear, do you think there are large deformations for being isotropic linear?. I just want to know a logic reason, but it will not change the model because I have experimental data so I will run the optimization file.

Best regards,

Jesica

[maas]

Hi Jesica,

The "isotropic elastic" material is only valid for small strains and large deformations. However, in your model you are getting fairly large strains (~60%). Under these conditions the isotropic elastic material may not produce realistic answers, although I did not do an actual comparison. It might be an interesting exercise to see if the predicted stresses and strains would be very different. In any case, I prefer the neo-Hookean material since it remains valid even for large strains.

Cheers,

Steve

[jesicandrea]

Hello Steve,

I have a new problem, I did another indentation model, simpler. I used the conditions that you proposed me before. When I have a regular mesh with the same size of elements, the model works and the indentation depth is 4mm, but when I activated the x-ratio and y-ratio on the mesh parameters, concentrating the elements in the center of te specimen, the zone where there are more deformations... the model does not finish.

I tried changing the aspect ratio of the center elements, I don't know why remeshed does not work.

https://drive.google.com/open?id=0B1GPvU5qJTIpMlhLVm9TZE9KUGM

Thanks in advance