Hi,

Which type of tests are you trying to match? Monotonic? Cyclic? fast or slow?

There is a lot of misconception about the nonlinear behavior viscoelastic materials, and most literature is not clear about it. Hyperelasticity alone is not enough for capturing all the involved nonlinear phenomena, and depending on that, you may need a more complex model.

Depending on what are you trying to represent, the calibration procedure can become complex and time consuming too, so it is good to aim for a good model, but not more than necessary. Otherwise you will have trouble in calibrating it.

You can check these works that aim to calibrate a model for detailed FE simulations:

1) Reyes SI, Vassiliou MF, Konstantinidis D. Experimental characterization and constitutive modeling of thermoplastic polyurethane under complex uniaxial loading. J Mech Phys Solids 2024;186:105582. https://doi.org/10.1016/j.jmps.2024.105582.

2)Reyes SI, Lopez EL, Vassiliou MF. Validation of a three-dimensional finite element constitutive modeling approach for a thermoplastic polyurethane calibrated with uniaxial tests. Eng Struct 2025;322:119044. https://doi.org/10.1016/J.ENGSTRUCT.2024.119044.

Feel free to reach back if you have further questions.

Best,

Hi,

Thank you very much for your response and the paper recommendations. I really appreciate the insights.

In this study, I performed both monotonic and cyclic dynamic tests under a range of frequencies, displacement amplitude and temperatures, one key insight I am exploring further a modified fractional standard linear solid (FSLS) model that incorporates both the Payne effect and a temperature-frequency equivalence principle. Interestingly, I also used bilayer cohesive constitutive modeling in Abaqus to simulate interface damage a significant issue I have observed in my own experiments. In terms of calibration, I employed MATLAB for fitting the Prony series parameters of the viscoelastic layer, using least squares based on experimental data, which I found to be a very systematic approach.

You are absolutely right. relying on hyper-elasticity alone has not been sufficient in my case. I have tried coupling a hyperelastic model with time-domain viscoelasticity in Abaqus, but the results still diverge from the experimental data, especially at higher frequencies or under fast loading. I am starting to suspect that other nonlinear effects, perhaps even temperature-related softening or Mullin’s effect, could be playing a role.

Your point about not overcomplicating the model is well taken. The Fractional standard linear solid model strikes a good balance between physical interpretability and computational manageability, so I am leaning toward implementing that it seems to provide a higher-fidelity representation without becoming unreasonably complex.

In light of your advice, I am considering a more refined calibration approach. I have been collecting DMA data to build a Prony series and am exploring TTS for master curve construction. I am also reviewing my boundary conditions and experimental setup replication in the FEA. small misalignments or fixture stiffness may be affecting the outcome more than I initially thought.

Thanks again for highlighting the complexity of calibrating these models. I will definitely go through the references you suggested. the focus on thermoplastic polyurethane under uniaxial loading seems relevant to what I am studying.

Looking forward to diving deeper into this, and I may follow up with questions as I refine the model and compare results again.

Best regards,

What material model you are using?