Please help - how to partition un-symmetrical parts 1

[cjn9999]

Hello - please I am looking for some help on how to partition bone, the spine in particular, and want to make for instance a cortical layer, a cancellous core, and also be able to make the posterior elements a certain material property. After this I plan on using meshes of varying density in the model. My goal is to test an interbody device in this setting.

But the main thing is, if you have curves, how do you partition them? I can take a block or circle for instance, and make one part of it bone, cause you can select an edge. When I try to select a curved line on the vertebrae, it says something about an error, and cancels me from going farther.

The vertebral bodies have a crust or outer layer of about 12000 MPa and about 0.5 mm thick. I belief the posterior elements, the stuff in back, is all about 3500 MPa (don't quote me on that). The inside of the main thick parts of the bodies, the top and bottom, superior and inferior parts, have cortical core of about 100 MPa or so, and makes up the bulk of bone.

Any help on this would be appreciated.

Chris

 

[rstupplebeen]

This is a common problem that is faced in biomechanics. I suggest searching the biomechanics forum and if you have access to a CAD package it's respective forum. Here is a link to a SolidWorks thread.

Geometrically this can be a very difficult problem since when you offset the surface to get the inner boundary all curvatures of the part need to be larger than the offset distance.

In Abaqus you could skin the part with the shell elements to represent the outer layer.

If you posted your CAD geometry I could give it a try in SolidWorks.

I hope this helps.

Rob Stupplebeen

https://sites.google.com/site/robertkstupplebeen/

 

[cjn9999]

Much appreciated for the response. I will attach the IGS file for the L2-L3 motion segment.

If you have ABAQUS this would be very very helpful if you could do it in there, but SolidWorks would be great as well.

All of this is tricky, but more tricky than the skin is assigning material to the posterior elements.

Please anything you do, list details of what was done to the model, the elements, material property, etc.

Best and thanks,

Chris

 

[cjn9999]

The previous links do not work. If anyone wants to help with these vertebral models, and add material and element properties, including posterior elements etc., that would be much appreciated.

Here is the updated link and download, the file is pretty large:

http://www.mediafire.com/?6os0wcp8x4upeu4

 

[rstupplebeen]

Chris,
I looked at the IGS in SolidWorks and checked for minimum curvature which is in the image below. Basically the curves are too tight to have a 0.5mm thick offset. Imagine that I said a 1mm diameter hollow ball had a 2mm wall thickness. Basically it is impossible geometry. You may be able to get around this with a significant amount of geometric manipulation and possibly using a variable cortical bone thickness.

Rob Stupplebeen

https://sites.google.com/site/robertkstupplebeen/

 

[rstupplebeen]

In abaqus there are some imprecise surfaces that can be repaired in abaqus or a CAD package. You also have many faces which combining with virtual topology could reduce poor element geometry and possibly count. It can be found in the mesh module>Tools>Virtual topology.

1 surface is twisted and causing the mesher to fail. Please see the picture below. The face highlighted in red is the bad one. I used virtual topology to combine this surface with the surrounding ones by right clicking on the face and choosing 'adjacent faces'.

After you complete these steps you will need to skin your part with shell elements. I hope this helps.

Rob Stupplebeen

https://sites.google.com/site/robertkstupplebeen/

 

[cjn9999]

Response from Robert via email (posts not on eng-tips because of file size issue, thus communication via email also):

I have only heard of relating bone density to material properties not element size. The shelling is traditionally for laminate structures such as carbon fiber over a foam core. Basically you can mesh the solid with the softer material. Then you can apply a shell element on the surface with differing properties for each face if you would like. The surface elements share the same nodes as the solid elements.

My thoughts:

Yes I think element type is changed in different areas of vertebrae, and also number of elements. I actually don't think element size is changed as I had thought, but in the literature they typically have charts mentioning used vertebral body finite element properties, and I think the mentioned items from me are on base.

 

[cjn9999]

To give an update everybody, here is the link to the human vertebral body scan:

http://www.mediafire.com/?6os0wcp8x4upeu4

It is about 116 MB. I was wondering if anyone can offer help as to how to apply material properties like 12000 MPa to the cortical bone, 3500 MPa to the posterior elements, and maybe 130 MPa to the cancellous bone, as well as a detailed explanation of the geometric boundaries for these bones and procedure in ABAQUS. I will attempt myself to do the shell elements as the cortical surface, but beyond that I might be stuck.

Any help would be appreciated, and many thanks to Robert.

 

[cjn9999]

Any current comments on this?