The identification of constitutive parameters in engineering and biological materials, especially those with complex geometries and mechanical behaviors, has long been a challenge. The emergence of Physics-Informed Neural Networks (PINNs) has provided promising solutions, although current frameworks have limitations when it comes to advanced constitutive laws and incorporating experimental data. This paper introduces a new PINN-based framework specifically designed to identify material parameters for soft materials, particularly those with complex constitutive behaviors, under large deformation in plane stress conditions. Notably, our model focuses on training PINNs using multi-modal time-dependent experimental datasets that include full-field deformation and loading history, ensuring the algorithm’s robustness even in the presence of noisy data. Our results demonstrate that our framework can accurately determine constitutive parameters of the incompressible Arruda-Boyce model for samples with intricate geometries, maintaining an error rate below 5%, even with a 5% experimental noise level. We believe that our framework paves the way for a transformative approach to modulus identification in complex solids, particularly those with intricate geometries and constitutive behaviors.
Physics-Informed Neural Networks for Determining Constitutive Parameters in Complex Hyperelastic Solids (arXiv:2308.15640v1 [cond-mat.mtrl-sci])
