PDA

View Full Version : Material Model For PTFE/RTFE



anishthampi
2009-01-07, 00:19
Hello:)
Im very new to this forum
Im currently working on a metal-plastic contact analysis between steel and RTFE ((15% glass fibre reinforced Teflon).
I want to find the contact pressure and stresses devoloped in the RTFE seal(washer) to test the leakage problem for a valve
Im using Altair Hyperworks for my analysis.
I had done some linear static analysis with linear elastic material for RTFE.
But im not so confident with my results as the material is nonlinear.
So now im trying to do the analysis with a nonlinear material. But im still struggling to choose material model for RTFE.
I have gone through lot of threads related to my problem in polymerfem.
Then i could find that the best material model for RTFE is DNF model.
But it is custom made and not avalible with hyperworks. DNF posses both viscoelastic and viscoplastic behavior. But for my software i can choose only one type at a time either viscoelastic or viscoplastic.
Im intrested only in contact pressure and stresses, can any body suggest me a good and easy material model

I got following options in hyperworks
1. Hyperelastic - Ogden Model
2. Elasto-plastic - Jhonson cook model with vonmises hardening
3. Elasto-plastic - Gurson model (von mises with viscoplastic flow)
4. Viscoelastic - Generalized Kelvin-Voigt
5. Viscoelastic - Generalized Maxwell-Kelvin
6. Viscoelastic - Boltzman
7. Viscoplastic - Closed cell, elasto-plastic foam

Please suggest one from above or any other material model.
Also help me in finding the parameters for these models(methods, what type of testing required)

Thanks & Regards

Anish Thampi

Jorgen
2009-01-08, 04:43
If you have to stick with one of those models, then I recommend the Johnson Cook plasticity model. You should be able to calibrate this material model using, for example, uniaxial stress-strain data.

To find the material parameters you will need to do one the following:
- use trial-and-error
- write your own parameter extraction software using Matlab, Mathematica, etc
- have someone do it for you (consulting)
- purchase some parameter extraction software

Best of luck,
Jorgen

anishthampi
2009-01-08, 05:42
Thanks Jorgen For your helpful reply
As per your recomendation I will use jhonson cook material model for my RTFE.

For this material card i need to provide the following parameters

1. Initial density
2. Young’s modulus
3. Poisson’s ratio
4. Plasticity yield stress
5. Plasticity hardening parameter
6. Plasticity hardening exponent
7. Failure plastic strain
8. Plasticity maximum stress
9. Strain rate coefficient
10.Reference strain rate

Is it necessary to get all these values as im interested only in stress and contact pressure results.

Right now im having only Youngs modulus, Poissons ratio and density. Can i able to calculate other parameters if i got a stress-strain curve.
Can i able to get these parameters from material testing? if yes, please suggest what type of test should i go for.

anishthampi
2009-01-18, 11:36
Hello Jorgen
Im waiting for your answer
Please spare some time for me

Thanks in advance

Regards
Anish Thampi

Jorgen
2009-01-25, 05:41
You will need experimental data at two or more deformation rates to determine the appropriate material parametes. At this time, Abaqus does not provide a tool to determine the material parameters.

I have developed my own tool that is based on Matlab and that is commercially available. As an alternative you can develop your own tool to determine the parameters, our you can approximate them through trial-and-error.

anishthampi
2009-05-27, 05:15
As per your suggesion i have tried jhonson cook material model in Hyperworks. But due to some other practical difficulties we didnt get right results in Hyperworks. So i will be trying the same analysis in Ansys. Hence for the same material (RTFE), which material model should I go for in Ansys.
Also what inputs will i need for that material model?
Are there any specific tests which I should perform for those input parameters?
Thanks in advance
Regards,
Anish

Jorgen
2009-05-27, 20:58
How about trying the Johnson-Cook model also here...

-Jorgen