Polymer material definition using tensile test data (linear part Hyperelastic material model)
Hello
I am trying to define a Polymer material based on tensile test data. my main issue is the polymer enters the plastic zone after 70% strain.
I am using this polymer in an application that works only in the elastic zone of the material (entering the plastic zone for my application means failure).
I used Yeoh 3rd order (hyperelastic model) to define only the elastic part of the curve ( less than 70% strain) and my simulation ( using static analysis in ansys work bench) works well, is this right? if not I would appreciate it if you suggest the right method for defining the material.
I checked the forum and I understood that Hyperelastic materials should not be used with plastic materials.
I uploaded the tensile test figure for the polymer.
Best Answers

bsista Member Posts: 92
Short answer: You can use hyperelastic models to represent the elastic deformation of polymer plastics before they yield.
Long answer: Hyperelastic materials are meant for modeling nonlinear elastic deformation when there's very little to no plastic deformation. A key characteristic is that the stressstrain response is monotonically increasing. So, it does not account for loss of stiffness due to damage or plasticity. Modeling polymers is always tricky, people either use hyperelastic models when the application does not require plastic deformation, or use metal plasticity models when plastic deformation is of interest. Ideally one must develop a new constitutive model and implement it using subroutines which is an advanced and involved procedure so for practical purposes one can either use hyperelasticity or metal plasticity based models depending on their application. However, one must use the correct data to calibrate either of these material models. For instance, if you wish to use hyperelastic maodel, then make sure that the experimental data (stressstrain curve) is monotonically increasing. If the curve begins to dip down, then clip off that portion and use only the portion to the left of it.
A word of caution: In your case, from the figure, the curve starts to dip down way before 70% strain (probably around 30% strain). Make sure that you're using only the monotonically increasing part of the curve while curvefitting the hyperelastic model.

A_zaghloul Member Posts: 5
Hi bsista,
Thank you for your answer.
I would appreciate it if you suggest specific reference that I can use it to reference this part if available.
The only thing that makes me confused I calculated the modulus of elasticity form the curve for 10 % strain, the value will be around 25 MPa ( calculated from the slope of the stressstrain curve), and I will assume the poison Poisson ratio=0.49 My question is when I used these values instead of the hyperelastic model the solution did not converge.
and I got this error
"internal solution magnitude limit was exceeded ansys workbench"
shall I stick to the hyperelastic model?

A_zaghloul Member Posts: 5
Hi bsista,
I think I fixed the problem by increasing the number of sub steps in my simulation.
I would appreciate it if you suggest specific reference that I can use it to reference (the hyper elastic part) if available.
Answers
Short answer: You can use hyperelastic models to represent the elastic deformation of polymer plastics before they yield.
Long answer: Hyperelastic materials are meant for modeling nonlinear elastic deformation when there's very little to no plastic deformation. A key characteristic is that the stressstrain response is monotonically increasing. So, it does not account for loss of stiffness due to damage or plasticity. Modeling polymers is always tricky, people either use hyperelastic models when the application does not require plastic deformation, or use metal plasticity models when plastic deformation is of interest. Ideally one must develop a new constitutive model and implement it using subroutines which is an advanced and involved procedure so for practical purposes one can either use hyperelasticity or metal plasticity based models depending on their application. However, one must use the correct data to calibrate either of these material models. For instance, if you wish to use hyperelastic maodel, then make sure that the experimental data (stressstrain curve) is monotonically increasing. If the curve begins to dip down, then clip off that portion and use only the portion to the left of it.
A word of caution: In your case, from the figure, the curve starts to dip down way before 70% strain (probably around 30% strain). Make sure that you're using only the monotonically increasing part of the curve while curvefitting the hyperelastic model.
Hi bsista,
Thank you for your answer.
I would appreciate it if you suggest specific reference that I can use it to reference this part if available.
The only thing that makes me confused I calculated the modulus of elasticity form the curve for 10 % strain, the value will be around 25 MPa ( calculated from the slope of the stressstrain curve), and I will assume the poison Poisson ratio=0.49 My question is when I used these values instead of the hyperelastic model the solution did not converge.
and I got this error
"internal solution magnitude limit was exceeded ansys workbench"
shall I stick to the hyperelastic model?
Hi bsista,
I think I fixed the problem by increasing the number of sub steps in my simulation.
I would appreciate it if you suggest specific reference that I can use it to reference (the hyper elastic part) if available.
hi, is it okay to use ONLY uniaxial test data for curve fitting? I could get good fit with this curve, but getting different biaxial and shear data on using different models. How safe is it to use ONLY uniaxial test data? The curve I am interested is of 23 deg C, and shown part of it is all elastic. I wanted to Load till 66 Mpa. Kindly help..