Notifications
Clear all

Which material model to use if only one direction test data is availble

3 Posts
2 Users
0 Likes
340 Views
Seiden
Posts: 4
Topic starter
(@Seiden)
New Member
Joined: 7 years ago

Dear Sir,

I have only one direction (tension) stress -strain data (Engineering) is availble for a particular elastomer rubber. I want to model it into FEA. My doubt is as a start which materrial model I should use. How to approximately assume other direction stress strain behaviour with one direction data .

Topic Tags
2 Replies
bw_composite
Posts: 124
(@bw_composite)
Estimable Member
Joined: 15 years ago

Dear Vyasmanish,

my opinion is that there is no material model better than others in reproducing an unknown behaviour. In other words, in your situations I would struggle to suggest any material model capable of giving you a better fit over arbitrary deformation modes, etc. by simply knowing an engineering tensile stress-strain curve.

In order to maximise my chances I would try to choose a model with some physical foundations. It also depends on the type of rubber: for example, if I remember correctly using 2 coefficients-Mooney Rivlin systematically overestimates the stiffness in biaxial modes for SBR, heavily flled, rubbers.

If your analyses does not involve excessively large strains, if the prevailing deformation mode is tensile, if there is not very pronounced volumetric/deviatoric coupling, maybe you could get away with it. I would take example from the great physicists of the past, who, whenever anything is unknown, postulated it to be constant or linear. I would then choose a Neo-Hookean and see where I get to.

Let us know!

Topic Tags
2 Replies
Seiden
Posts: 4
Topic starter
(@Seiden)
New Member
Joined: 7 years ago

[QUOTE=Muzialis,13251]Dear Vyasmanish,

my opinion is that there is no material model better than others in reproducing an unknown behaviour. In other words, in your situations I would struggle to suggest any material model capable of giving you a better fit over arbitrary deformation modes, etc. by simply knowing an engineering tensile stress-strain curve.

In order to maximise my chances I would try to choose a model with some physical foundations. It also depends on the type of rubber: for example, if I remember correctly using 2 coefficients-Mooney Rivlin systematically overestimates the stiffness in biaxial modes for SBR, heavily flled, rubbers.

If your analyses does not involve excessively large strains, if the prevailing deformation mode is tensile, if there is not very pronounced volumetric/deviatoric coupling, maybe you could get away with it. I would take example from the great physicists of the past, who, whenever anything is unknown, postulated it to be constant or linear. I would then choose a Neo-Hookean and see where I get to.

Let us know![/QUOTE]

Thanks a lot Muzialis

Manish Vyas

Reply
Share: