I have some stress-strain curve with hysteresis of my elastomer ! The test was the uniaxial compression.
* How to determine the parameters of a viscoelastic model under abaqus ?? or how to model a viscoelastic behavior under abaqus with my unique test ?
If it's impossible, have you a solution for me ?
ps: what exactly the enter parameters for your umat bbmodel
Thank's , Eddy.
If you want to model the hysteresis of an elastomer using ABAQUS there are only a few material models that you can use, for example:
- Linear viscoelasticity.
- A user material model such as the BB model.
You can find the material parameters for both of these models using a similar approach:
1. Find an initial guess/estimation of the material parameters using a graphical technique and an understanding of how the material model. For example, the initial Young's modulus of the material is a material parameter in both of these material models.
2. Optimize the material parameters using a numerical optimization technique. This can be done in different ways using different algorithms. I have had good luck with the Nelder-Mead simplex algorithm. I have also written a specific material parameter determination program that can estimate the material parameters semi-automatically from experimental data.
Best of luck,
We have experimental data from uniaxial compression tests of a rubber (stress and stretch in the direction of compression).
When the parameters of Helmh. function are estimated by some optimization technique only from the stress in the direction of compression, we can ex post calculate some small stress in the transversal direction. If the condition of zero stresses in transversal direction is added to optimization process, we get unrealistic results. What to do?
In uniform uniaxial compression there is not stress in the transverse direction, hence I recommend that you impose that constraint!
There can be different reasons why you get strange results. Perhaps your optimization process is not free from bugs, perhaps you are using a hyperelastic strain energy function that is not unconditionally stable (Drucker's stable).
I recommend that you try the optimization again, but this time using an unconditionally stable model such as the neo-Hookean or 8-chain models.
So far as viscoelastic parameters are concerned in order to apply linear viscoelasticity modelling I have positive experience with an experimental data evaluation software called ViscoData (see the INTERNET). It has considerable advantages over appropriate built-in features of some FE-codes like ABAQUS, MARC and ANSYS. It provides you with a set of PRONY parameters in less than 5 minutes when raw data are available.