Displacement-Controlled Indentation of Rubber Membrane--Reaction Forces too Low

[EMJohnsen]

Hello,

I am trying to model the deflection of a rubber "membrane". The "membrane" is 250 mm in diameter and 0.5 mm in thickness; it is meshed with C3D6H elements and is modeled as an incompressible, neo-Hookean material with C10 = 0.17 MPa. The material is deformed 127 mm via displacement-controlled loading by a rigid body indentor with a radius of 50 mm. Standard "Hard Contact" with a friction coefficient of 0 describe the contact between the two surfaces. The model converges with the desired displacement, but the reaction forces are significantly lower than expected.

If there is any insight into what I may be doing wrong or leaving out, it would be greatly appreciated!

Thank you!

 

[Me09]

check the material property. Mooney–Rivlin method is an alternative to neo-Hookean method.

 

[yw5aj]

+1 to Me09. Neo-Hookean is non-linear per se, but IMO it is not "non-linear enough". Try Mooney-Rivlin, and if that wouldn't work I'd suggest try Ogden. If Ogden still won't work, then probably you could share your inp file with us?

 

[EMJohnsen]

Thank you for your suggestions; I tried them with little success. I've also tried changing parameters one-by-one to see where the issue may be coming from, but I just can't resolve things. My reaction forces are still an order of magnitude or so lower than the desired results.

I have included my input file below. As a reference, at 50 mm displacement, the reaction force should be about 25 N; 100 mm displacement, 130 N reaction force; 127 mm, 200 N reaction force.

Thank you again!!

 

[rstupplebeen]

My colleague and myself have faced similar issues transferring from test data into hyperelastic material properties. We utilized Isight to automate the optimization of material properties to correlate to a battery of physical tests. It is imperative to have a mixture of loading modes to correctly characterize a material such as shear, tension, compression, and equal-biaxial. For our application it was not feasible to have standard testing so we had various mixed modes. From here we created FEMs of each test. Then we let Isight automate the coefficients of the properties to minimize the RMS error of the test and simulation data. Carrying these simulations out on a variety of material models yielded the optimal model and coefficients. Once this is done validate the model on a tests that were not included in the simulation to see if it is predictive. I hope that this helps.


Rob Stupplebeen

http://www.OptimalDevice.com

 

[IceBreakerSours]

Hi Rob:

Do you know of any other codes that let you carry out material parameter identification by driving Abaqus?

Are you new to this forum? If so, please read these FAQ:

http://www.eng-tips.com/faqs.cfm?fid=376

http://www.eng-tips.com/faqs.cfm?fid=1083

 

[EMJohnsen]

Thank you Rob; very informative! However, your suggestion does not entirely apply to my current situation. I am trying to re-create an Abaqus model from literature as well as a CGMD model created in my lab--both of which yield the same results. With my simulation, however, my forces are several orders of magnitudes lower than the desired results. I've checked my units countless times, and do not think the error is there. Are there any other suggestions of where I should check?

Thank you again!

 

[rstupplebeen]

Try checking the size of your parts. I know some neutral CAD formats export in certain units. I hope this helps.

Rob Stupplebeen

http://www.OptimalDevice.com

 

[EMJohnsen]

Parts all check out. This is such a puzzle!

My thought is that the discrepancy is related to either the interaction or mesh; however, I've checked each setting one-by-one and can't seem to resolve the issue. Are there specific settings that can alter the accuracy of the model?

 

[rstupplebeen]

IceBreakSours, Sorry I missed your post. You should be able to carry it out with Python pre and post scripting but that will require a bunch more programing than using Isight.

EMJohnsen, I do not currently have Abaqus but will soon. If you post your problem or a scrubbed version someone usually steps up to help. It usually turns out to be a minor slip-up like referencing the wrong material in a property or something else unpredictable.

I hope this helps.

Rob Stupplebeen

http://www.OptimalDevice.com

 

[EMJohnsen]

Here is my problem statement and input file:

I am modeling an indentation of an incompressible, hyperelastic material. The following are the specifications used:

Membrane diameter = 5 um
Indentor diameter = 0.5 um
Membrane thickness = 0.005 um
C10 = 2.7e-9 N/(um^2)

The material is to be indented 2500 nm.

The following plot are the results obtained via CGMD:

http://files.engineering.com/getfil...c7a-4911-b236-0589e0ca7253&file=CGMD_Plot.jpg

As you can see, my results do not match, and I just can't seem to find the issue. If you could please help it would be GREATLY appreciated!