Fracture of teeth 4

[FENovice]

Hello Group,

I am trying to predict fracture of teeth.

The approach has been to 3d scan a tooth and recreate the geometry in ceramic material by rapid prototyping techniques.
The sample teeth are then tested on a load test machine.

Also material samples were tensile tested and i got a nonlinear stress strain curve used in the FE analysis.

Anyway the basic problem is that when i recreate the actual failure load in the FEmodel the stress is way higher than the UTS of the tensile sample.

I have looked at vaious failue theories - max pinicipal, mises, tresca and looked at strains too...

There is a very sharp corner where the crack initiates when the test samples fail and this is where i get my peak stress but it is way higher (around 600Mpa) than the UTS from the tensile test (the UTS is only around 60Mpa)...

I have a good mesh and loads of elements around the stress concentration and no abrupt changes in stress across elements... Everyting seems correct but stress is just too high...?

Does anyone have any advice?

Thanks,

 

[cloa]

BTW these are the biomechanical teeth you are talking about, not teeth of a mechine?

Tensile testing of ceramic material(maybe its composite? is unusual as it is affected by the initial flaw size. Wouldn`t a bending and or compression?test more appropriately model the material.

 

[Ron]

Yes...I think you have a stress-strain curve problem. First, you have a non-linear stress-strain curve on a rigid, brittle material...doesn't sound right.

Secondly, your compression modulus and tensile modulus can be significantly different.

Third, do you have any idea what Poisson's ratio for the material is? You can troubleshoot your data, looking to see if all those ratios are in sinc or not.

Also, when you are molding your specimens for material properties testing, is there a chance that they are not representative of the molding process for the actual teeth, such as density, uniformity, or inclusions.

 

[FENovice]

Hi,

Sorry its not really a true ceramic though it is supposed to behave in a similar manner - the material is bluestone - data sheet here

http://www.3dsystems.com/products/datafiles/datasheets/SLA/DS_Accura_Bluestone_US.pdf

It is more a kind of plastic but it is brittle with no real plastic deformation.

The stress strain curve is also pretty straight - I just included the small nonlinearity in case it made a difference. It diddn't...

Thanks you your replies,

 

[dik]

I did a bunch of work on teeth about 40 years ago... had to do with steel pins and amalgams failing. High tensile stresses due to the difference in materials. At the root of the pins, you have high tensile stresses created. The amalgam is like concrete; it has a high compressive strength and a low tensile strength, hence the fatigue tensile failure.

The compressive forces generated by your jaw muscles can be significant and can cause similar problems if there is not an 'exact' match for the biting surface.

Dik

 

[FENovice]

To Rons queries,

I don't have a measured value to possions ratio - so i will try some variance of the value here and see if i can get better correlation between the test data.

The specimiens should be fairly free from inclusions etc as the SLA process is fairly consistent however the parts are built in layers so this may cause some issues...

I would have been less surprised if the physical test samples failed earlier than predicted in FE but it is the reverse - the FE says they should fail at perhaps 1/10 of the actual load they can withstand...

 

[DCBII]

It sounds like the problem could be in how you've modeled the tooth. Finite element models produce infinite stresses at sharp corners. Sharp corners create a mathematical singularity. Try meshing your model into smaller pieces and watch the stresses jump even higher. The finer the mesh the higher the stresses. This is not what would happen in reality.

You can avoid this by putting a small fillet on the corner. I ran into the same problem designing a rectangular hole in a steel plate.

 

[FENovice]

To advice from DCBII,

Thanks for your thoughts..

The geometry i am using is completely smooth and tanget - there is a very small radius as opposed to a sharp corner and it IS radiused and has a ton of elemenets across it - It may be that the prototyped samples are not quite the same (although they are produced directly from the 3d file) and i will look into this.... but i doen't believe this is the smoking gun as to why it doesn't tie up...

Thanks...,

 

[msquared48]

You need to look at the wedging effect in the tooth dynamics here. Where the food gets trapped and compression is applied, particularly hard foods, a wedging action happens, imparting a splitting and spreading force to the tooth matrix, ie, trension stresses. If the tooth is properly shaped, this should not happen, but due to wear, can with time.

Mike McCann
MMC Engineering
Motto: KISS
Motivation: Don't ask