FEM analysis of a strip tensile test

[richard]

Hi everybody

Happy new year 2008! I pretty new in this Forum and I do not know whether I could get some nice hints there about my current modelling issues. I just give it a try!

I am now facing some difficulties trying to explain the differences between FEM and experimental results on a strip tensile test. My model is considerably underestimating the initial stiffness of the test specimen. One explanation lies in the fact that my model does not properly account for transverse stiffness of the chords (I am modelling them using embedded beams in ABAQUS). But this is for sure not enough. The other explanation would probably come from the rubber modelling itself and this is the reason why I posted my question on that forum!:)

For the considered rubber I have extensive experimental data (uniaxial, biaxial and planar test). I am using a YEOH model which seems to give pretty good approximation of my test data. The available uniaxial test starts with a compression phase (let s say from A (origin) to B) and then the specimen is loaded in tension (from B to C). This is this portion B-C of the exp data that I have considered to calibrate my model. The point is that the orginal stiffness at A is much larger than when BC crosses the Y-axis again when loaded back in tension. My question is: Am I not underestimating the initial rubber stiffness in tension knowing that in my strip tensile test the rubber is directly loaded in tension from a stress and strain free state? In case I would be underestimating it would you have guys some hints (litterature, papers) on how to input relevant uniaxial test data? Any suggestions is very welcome

Many thanks for your input in advance

Best regards

Richard

[Jorgen]

It sounds like your material might be undergoing Mullins softening. In the original compression phase (A to B) the stiffness is high since the material is "damaged". In the next phase (B to C) the material is softer due to the initial deformation.

What FE software do you use? What rubber is it and what is the concentration of filler particles?

- Jorgen

[richard]

Hi Jorgen

Many thanks for your reply. The FE software I am using is ABAQUS 6.7-4.
Concerning the type of rubber I cannot tell you as our customer does not want to tell us waht rubber it is for confidentiality issues. We just know it as rubber 754 which of course is not of big help! I can just tell you the rubber is used in for making tire.

The filler particule concentration is 11.1% (Volume of chords over volume of rubber). Is that what you mean?

So apparently we are considering experimental data for a already damage rubber (which occurs during first compression stage A to B), whereas my strip test considered an undamaged piece of rubber which is going to be primarily loaded in tension.

Is there a way to approximately define this first tension curve for the undamaged rubber based on the uniaxial cyclic test data we have? Shall we use a different modelling strategy?

Many thanks for your time

Regards

Richard

[Jorgen]

You might want to try the MULLINS model in ABAQUS. That model can be combined with the YEOH model.

- Jorgen