Polymer FEM newstarter

[courghan]

Hi, i am just starting using fem software for thermal related analysis, in particulare i need to find a reliable-enough way to simulate creep behavior of wood and hdpe composite material. I have access to a very well equipped lab, but as this field is quite new for all of us here, i would be really grateful if any of you could give me any advice on which software could be the best and even on books or guides i could use in my research.

thanks a lot.

Marco

[Jorgen]

Hello Marco,

Polymers, such as HDPE, are tricky in that their creep behavior is strongly dependent on temperature, applied load, and time. At "small enough" loads you can use linear viscoleasticity (which is built-in into virtually all FE software). Linear viscoelasticity is nice because it is easy to use and it is easy to find the needed material parameters from experimental data. But Linear viscoelasticity is not always a good option since it only gives accurate predictions for very small loads. What "small loads" mean will depend on the temperature and the polymer material.

There are more advanced and accurate models that can be used to predict the non-linear viscoelastic response. These more advanced model are in my mind better general purpose models since they do not have the same limitation of small loads and displacements. Which model approach is most appropriate in your case will depend on the details of what you are looking for and how accurately you want to know the creep behavior.

I cannot claim to be an expert on creep behavior of wood, but since wood can be considered a composite polymeric material, I assume that you will qualitatively see the same response as for traditional engineering polymers.

You should be able to experimentally characterize the viscoelastic properties of your material using standard creep or stress relaxation tests. I would also perform a few uniaxial tension or compression tests to get a more complete picture of the material response.

About software, there are many different FE software packages available. All of the major codes are good. For the testing program, take a look at the ASTM standards.

Best of luck,
Jorgen

[courghan]

Hi Jorgen,

thanks for answering my questions! The range of temperature i am interested in is about 40-70 Celsius and the loads will be static but not really small. I thought at the beginning about performing dmta tests on my material to obtain the E-modulus variation among temperature and then perform a linear static analysis with the modulus corresponding to my temperature. I know this will give me just a very rough idea on the material behavior but do you think this will be just a loss of time? Would you advice me to do even some tests in weatherometer?

Thanks

Marco

[Jorgen]

Hello Marco,

Your temperature range is relatively narrow, which makes it easier :)

DMTA tests are a good starting point, since they are quick and provide a lot of information. The approach that you descibed (where you take E to be a function of temperature) is a reasonable starting point. The major limitation of this approach is that you will not be able to predict any creep since you are using a temperature-dependent linear elastic model.

A more advanced approach would be to use the DMTA data to calibrate a linear viscoelastic model. That way you can predict some aspects of the creep behavior. The major limitation of this approach is that it is only valid for small loads (where the response is still linear viscoelastic).

Now, if you want to go even more advanced you can run larger strain stress relaxation or creep experiments, fit an advanced material model to the data, and then run general FE simulations.

Jorgen

[courghan]

Hi Jorgen,

thanks for your advice, but i still have a question for you.
How do i put together a material model?? I have never used ansys with advanced material models (i only used to type in the modulus ed poisson ratio...) and i have noone here that can help me as anyone in the company know less than me about that.
What do you think of Ansys Workbench 9.0?? Do you know if it is reliable for analysis??

thanks

[Jorgen]

Hello Courghan,

As you know, you'll need one material model for each material that you want to simulate. These material models can be be considered to consist of 2 parts:

Material Model = Constitutive Model + Material Parameters

What you need to do is to select one constitutive model (i.e. set of equations describing the material response) for each material, and then determine the needed material parameters from experimental data. If you would like more hands-on assistant, send me a private message (http://polymerfem.com/modules.php?name=Private_Messages).

Since I am not too familiar with Ansys Workbench, I recommend Ansys Mechanical.