Need help : Hyperelastic+Hyperfoam+Viscoelatic material

[x6v87qe]

Hi, Forum

First, I am so glad to find this nice place. Currently, I have encountered some problems with my project that may need your help

The material to be modeled is consisting of three parts: inner core (rubber/cork mixture), casing cork layer wounded with yarn and the leather cover. As this composite structure will be subject to impact simulation, therefore local large strain will be expected.

My supervisor ask me to model those components respectively, so what I’m gonna do is to model those components separately and assembly them as a whole.

Some raw thoughts: For the inner core my choice is Hyperelastic+ viscoelastic or Hyperelastic + hysteretic as energy loses must be account.
The cork layer: Hyperfoam+ viscoelastic
The leather cover: elastic+ viscoelastic ?
I’ve noticed that some investigators were using SHPB to capture the dynamic response of those soft materials. However this seems impossible for me.
My questions are:

1. For the leather part I am not sure the material model? And I didn’t find any FEM modeling against leather.

2. What exactly model could you advise me for Hyperelastic, e.g. MR, Neo, BB, etc.

3. If dynamic experiment is necessary what else experiment can I choose except SHPB? Any other suggestions?

4. For the visco property (used to characterize the energy loose). What experiments should I do? Should I do creep or relaxation test at a very high strain rate? How could this happen?

5. For viscoelastic, according to ABAQUS manual, it says both Long-term and short–term could be applied for, then what should I choose?

6. As strain rate is an essential consideration for this case. Is it possible for me to do some quasi-static experiments as a start point then manually tune it to fit the physical test curve (e.g. dropping force-deflection graph)? If this could be applied then the approach would be:

Please also note: The software adopted here is ABAQUS and I am only can do some Uniaxial tests. I’ve done some dropping test (as a whole ) already and gained the force-deflection curve including hysteretic loop.
Thanks for any help in advance !

[Jorgen]

It sounds like you have an challenging problem ahead of you. First, here are two questions:
:arrow: What strain rates and strain magnitudes are you interested in?
:arrow: Are you interested in both loading and unloading?

And here are a few comments:

:arrow: With SHPB I am assuming that you are referring to the split hopkinson pressure bar test.

:arrow: There are probably not many well developed material models for leather. What I would do is to first perform experiments on the leather and then use that information to decide what model is most appropriate.

:arrow: In some cases you can determine the dynamic response of a material for a given range of strain rates, and then extrapolate that information to faster rates using insight into the deformation mechanisms of the material.

:arrow: For the rubber (and perhaps all three materials) I would start with the Neo-Hookean model, and then perhaps switch over to the Arruda-Boyce 8-chain model, or the Bergstrom-Boyce model.

:arrow: Since you are focusing on impact loading, short-term stress relaxation tests are more relevant than long-term tests.

- Jorgen

[x6v87qe]

Dear Dr. Jorgen

Thanks very much for you kind reply. Very helpful comments!!! Answer to you two questions

:arrow: I’m not quite sure about the strain rate. All I know is the impact speed will be between 30~40 m/s (against rigid objective) So, approximately a few hundred per second??

:arrow: Yes. I’m interested in both loading and unloading to account for the energy losses. That’s why I include viscoelastic property.

Another few questions upon your comments:

:arrow: Actually I have got two proposed options to determine a dynamic response
1. A compromised way. I can conduct a dropping test which uses a flat anvil drop from certain height then calculate the stress/strain from measured force/displacement curve. The dropping can give a max 10m/s speed, better then static, but still behind the real condition (30~40m/s)

2. As you mentioned above to extrapolate that information to faster rates from static test. Indeed, that’s most investigators did. On extrapolation, as I see, most of the authors will just manually justify the parameters until the numerical simulation accord the force/displacement curve. Some even using semi-robust way like numerical parameters inversion, etc .Probably, this is the way I’m gonna do, Artificial neural network + Generic algorithm optimization. Is that sounds OK?
Can you please also infor me some examples on how to get “insight into the deformation mechanisms of the material “?


:arrow: THIS MAY BE THE BIGGEST PROBLEM. How to characterize the visco property under a dynamic condition? Should I do a relaxation test with a high strain rate? As I know a relaxation or creep test will normally take minutes with low strain rate rather then million seconds with high strain rate. So, how can I extract the instantaneous shear modulus by a common relaxation test setting? Is there any examples?

:arrow: On your paper, you mentioned that there’s a built-in BB model in the ABAQUS. Unfortunate, I can not find it in the ABAQUS 6.6. For rubber-like material, there’s another way to account for energy looses (except +visco) -“Hysteresis”. However the manual does not mention how to define those parameters, so it seems impractical. Is this the BB model you saying? Then how to conduct it ?


Hope, this large number of questions wouldn’t freak you out. Many thanks for your precious time, you really help me out

[Jorgen]

A few more comments:

:arrow: OK, you are mostly interested in very high strain-rates, and both loading and unloading.

:arrow: The drop test sound very useful. Can you explain how you would measure the force/displacement data from a drop test?

:arrow: There are some models (such as the Bergstrom-Boyce (BB) model) which can be calibrated for a relatively narrow range of strain rates and then be used to predict the behavior at much higher strain rates. This approach has been shown to work well for certain elastomers, and might also work for your materials.

:arrow: What I would do to address your problem is to run uniaxial compression (or tension) tests at a wide range of strain rates. I would run the tests using loading followed by unloading. I would then fit the BB-model to your experimental data. And then I would simulate the drop test to verify that the model is accurate also at very high strain rates.

There are other ways to attack this problem using stress relaxation tests, etc.

:arrow: A version of the BB model is a built-in feature of ABAQUS/Standard. To use this feature you need to use the *Hysteresis keyword. It is not always trivial to find the material parameters since the model is non-linear. I have written my own "material parameter fitting" software (that is commercially (http://www.polymerfem.com/modules.php?name=User_Subroutines) available), you can write your own using e.g. Matlab. Also, FYI, I have developed a version of the BB-model that is more advanced than the ABAQUS built-in version. My version (http://www.polymerfem.com/modules.php?name=User_Subroutines) works for both implicit and explicit simulations, and has support for Mullins effect type damage and temperature.

[x6v87qe]

Thanks Dr Bergstrom
Now I have got a much clearer picture. It seems that BB model would be a great choice.

:arrow: To calculate the force /displacement data ,I need a force plate to capture the impact force and the displacement was measured by integrating twice the acceleration.

:arrow: On conduct BB model, what type of the Hyperelastic mode should I choose to conjunct with the Hysteresis behavior? In ABAQUS 6.6 I presume it to be the Arrdua-Boyce is it ?

If this is the case, so, what I’m gonna do is to conduct static test first and fit it with the Arrdua-Boyce model and then define the parameters for Hysteresis by verify the physical testing? For parameter identification, what algorithm you use for fitting ?

:arrow: When you mentioned “other ways to attack this problem using stress relaxation tests”, can you give some more infor about it? An example would be great.