Riveting analysis with sealant

[amar]

Hello,
This is my first post on the website so I apologise if I make any mistakes.I am doing a riveting analysis in ABAQUS/standard. It has two aluminum plates joined together by an standard aluminum rivet. It involves large deformations and plasticity.
Now, I want to introduce a sealant in between the two plates. I want to model the sealant as an elastomeric material using hyperelastic material. Since the rivet will compress the plates I want compressive stress strain data. I dont have that but I had gone through the abaqus examples problems and they have constants (Mooney rivlin etc.) fitted for rubber gaskets etc. I was thinking of using those.
Also, this website also has lot of data for uniaxial compression for different rubber materials. The data is given for different strain rates. If I use that data do I input a sepcific strain rate data in abaqus? or do I have to input data at all three strain rates? Is the data suitable for my analysis?
Since purpose of sealant is to seal the plates do I use regular contact in abaqus or do I use Tie constraints for adhseive and plates?
Any suggestions, help will be greatly appreciated. Thanking you ina dvance.
Amar Atre
Student
Gatech

[Jorgen]

Hi Amar,

Without having all the information about the problem that you are trying to solve, I assume that the sealant is going to have a rather small impact on the mechanical response of your system. Hence, the constitutive model representation of the sealant does not have to be perfect. I would therefore start with a simple representation of the sealant - using a hyperelastic model. I would not use Mooney Rivlin, but instead Neohookean, Arruda-Boyce eight-chain, or the Yeoh model. If you need a more accurate model use the *Hysteresis model within ABAQUS.

You can use contact, tie constaints, or simple having the sealant and the plates share nodes. Which of these options is best will depend on how the adhesive is applied.

Jorgen

[amaratre]

Hello Dr.Bergstrom,
Thanks for your reply. Heres a picture of my model that might be useful to picture it

http://my.netomat.net/gtg506p/rivet1/

The sealant will be in between the plates. I will start off with ahyperelastic material model. Abaqus manual has some example problems where they have hyperelastic material constants. But most of their data is for Mooney rivlin or polynomial models.

I am doing a static analysis. One thing I am not clear/familliar with is that on the website I see that modst of the data is with hysteresis loops. Does that matter in a static analysis? Also which strain rate data do I use? How does that make a difference?

amar

[Jorgen]

Hello Amar,

You can certainly use the Mooney-Rivlin model, if that is most convenient. Just be aware that it is not the most accurate or robust model.

The experimental data on the website is data from "real" experiments. In these experiments you typically see hysteresis when you apply a complete loading-unloading cycle. It turns out that the amount of hysteresis is typically decreasing for slower experimental tests. In fact, there are indications that for certain elastomers the amount of hysteresis goes to zero as the deformation rate goes to zero.

In your particular application you can either do a static analysis as you indicated. To do this you simply need a hyperelastic material model for the sealant. In this case you should probably use the slow deformation rate data. Note a hyperelastic model cannot capture the influence of strain rate.

The other approach that you can take is to use use a more advanced material model for the elastomer. Using, for example the *Hysteresis model, you can simulate and predict the relaxation in stress in the adhesive with time, temperature, or whatever else you might be interested in.

Jorgen