3 point bending test mice femur and tibia - ABAQUS

[EliSera]

I am modeling a 3 point bending test in Abaqus on a mouse bone (both tibia and femur) and in particular I want to assess the crack insurgence at the diaphysis level.
I was able to do it in a 2D femur geometry by calculating the STATUS XFEM

When I moved in a 3D tibia geometry I face some issues:
1) I had difficulty in assembling the bone with two supports as constraints, and with a load beam, therefore I just applied a load as y-displacement at the diaphysis and I constrained the bone at the epiphyses with encastres.

Even if I obtained realistic values of Von Mises Stress

, I would like to move in a more faithful set-up as the one proposed by @magrimmelprez in the past 16 Dec 2016 in this forum. The researcher conducted a similar study with a 3PB test on a mice femur and human clavicle. Is there a way to interact with this researcher?

2) I tried to enrich the model with the special feature of a crack, but when I run the job, the software was not able to allow the crack generation and propagation. What can I do?

Thank you

 

[FEA way]

Can you share the link to that forum thread ? It would be good to know what you mean by "more faithful set-up" proposed there.

How was your crack definition different from the 2D model ? How was the 2D model constrained and loaded ?

 

[EliSera]

1)This is the link of the forum thread:

https://www.eng-tips.com/viewthread.cfm?qid=418549

2) The crack definition and domain was not different between 2D and 3D, meaning that I select all the bone geometry as crack growth and propagation, but the initial geometry imported in Abaqus is of course different because the 3D is a polysurface file obtained by segmentation of uCT images, instead the 2D is a mask obtained by microscope images.

3) In the 2D geometry, the load was applied in the origin of the circle load as a 0.5mm of displacement in the vertical direction, and the bone was constrained with encastre at the epiphyses.

 

[FEA way]

The approach described in that forum thread (supports and punch modeled as rigid cylindrical surfaces) is indeed the best way and I think that you should try it as well. It's a typical method when simulating 3-point bending.

Have you considered modeling the initial crack and specifying it for XFEM ? This is also commonly done when analyzing the crack propagation.