Search
Close this search box.
Notifications
Clear all

Calibrating Material Test data for Natural Rubber using ABAQUS 6.9-EF1

2 Posts
2 Users
0 Likes
531 Views
Posts: 4
Topic starter
(@camicastano)
New Member
Joined: 14 years ago

Hello everyone,

I am new to polymerFEM and also quite new to Abaqus, therefore if some of my questions seem strange and some of my description is hard to understand I appologise. I am more than happy to clarify anything I wrote.

I have a question as to how to calibrate the material properties in Abaqus 6.9-EF1 to get the same results as the material tests that were made for the Uniaxial, Biaxial and Planar loading.

This is what I have done so far: I received the test results of the specimen. Each specimen was loaded several times for four different stresses and due to the mullins effect four loops were seen due to the unloading path being different to the loading path. for each test I created a primary curve that connects the peaks of all the four different sress tests (i.e. connecting the peaks of the loops). The primary curves were used as input data for the various Abaqus Uniaxial, Biaxial and Planar data requirements. Then a mullins effect was included in the material properties in Abaqus. This I did to include the damage that occures to the rubber. The mullins effect was calibrated using the conastants one can input. I also used the unloading path of the test data of the specimen as input for the mullins effect to see if there is any difference between the two.

Next I tested to see if my material properties were any good. I used a single cubic element and performed Uniaxial, Biaxial, and Planar simulations. I am able to somewhat get similar loading and unloading paths compared to the real life test but the problem is that I cant find constants that will give me good correlation with all the physical test results. Either the Uniaxial is quite good or the Planar or the Biaxial but not all at once. Furthermore if say i only take the Uniaxial and try to find the right constants for that I am not able to find constants for the mullins effect that will give me the same path for the four different stress tests (one is low stress and several cycles, one is medium stress several cycles, one is high stress several cycles and one is very high stress several cycles). Either the first path (loop) is ok or the second or the third or the fourth. Additionally the shape of the path is not the same. I get rather almond shaped paths when the real life tests are sort of S-shaped paths.

Does anyone know how to get good correlation to all test resluts (Uni-,Biaxial & Planat) and also how to get the same shape as the test results?

Any help is much appreciated and would go a long way in completing my master thesis. 🙂

Kind Regards

Cliff

Topic Tags
1 Reply
Posts: 7
(@formsjon)
Active Member
Joined: 12 years ago

[QUOTE=Cliff,6589]Hello everyone,

I am new to polymerFEM and also quite new to Abaqus, therefore if some of my questions seem strange and some of my description is hard to understand I appologise. I am more than happy to clarify anything I wrote.

I have a question as to how to calibrate the material properties in Abaqus 6.9-EF1 to get the same results as the material tests that were made for the Uniaxial, Biaxial and Planar loading.

This is what I have done so far: I received the test results of the specimen. Each specimen was loaded several times for four different stresses and due to the mullins effect four loops were seen due to the unloading path being different to the loading path. for each test I created a primary curve that connects the peaks of all the four different sress tests (i.e. connecting the peaks of the loops). The primary curves were used as input data for the various Abaqus Uniaxial, Biaxial and Planar data requirements. Then a mullins effect was included in the material properties in Abaqus. This I did to include the damage that occures to the rubber. The mullins effect was calibrated using the conastants one can input. I also used the unloading path of the test data of the specimen as input for the mullins effect to see if there is any difference between the two.

Next I tested to see if my material properties were any good. I used a single cubic element and performed Uniaxial, Biaxial, and Planar simulations. I am able to somewhat get similar loading and unloading paths compared to the real life test but the problem is that I cant find constants that will give me good correlation with all the physical test results. Either the Uniaxial is quite good or the Planar or the Biaxial but not all at once. Furthermore if say i only take the Uniaxial and try to find the right constants for that I am not able to find constants for the mullins effect that will give me the same path for the four different stress tests (one is low stress and several cycles, one is medium stress several cycles, one is high stress several cycles and one is very high stress several cycles). Either the first path (loop) is ok or the second or the third or the fourth. Additionally the shape of the path is not the same. I get rather almond shaped paths when the real life tests are sort of S-shaped paths.

Does anyone know how to get good correlation to all test resluts (Uni-,Biaxial & Planat) and also how to get the same shape as the test results?

Any help is much appreciated and would go a long way in completing my master thesis. 🙂

Kind Regards

Cliff

Hi,I think you should do the calibration separably for each test. I mean you will get some constants for uniaxial and some other constants for the bi-axial test, depending on your test and simulation.

Kind regards

Ali

Topic Tags
1 Reply
Share: