Efficient Identification of Stiffness Parameters in Hyperelastic Models for Coated Woven Fabrics Based on a Single Experiment - Comparative Study Using Real Experiment

Abstract

Background

Determining the optimal number of classical biaxial tests and the selection of appropriate stress ratios for characterizing the complex behavior of coated woven fabrics is a challenge. Thus, the reliability of models calibrated using these homogeneous tests in predicting engineering structural performance may be limited.

Objective

A recently proposed experimental-numerical approach has been shown in a feasibility study to offer a new perspective for the characterization of nonlinear anisotropic thin materials (Makhool and Balzani Experimental Mech 64:353–375 2025). The goal here is not only to validate the robustness and efficiency of this much less expensive identification framework using real experimental data from a single experiment on coated woven fabrics, but also to demonstrate that the parameters obtained based thereon may even improve the accuracy of structural analysis.

Methods

On the basis of the Equilibrium Gap Method, provided that the material parameters are linear in the constitutive equations, a quadratic objective function is formulated, enabling a unique identification. Key of this framework are inhomogeneous full-field kinematics obtained from a single experimental setup designed to excite all essential deformation modes in the material model. Prior to incorporating the data into the discretized equilibrium equations, a pre-processing step is conducted, involving interpolation and extrapolation of the displacements on the scattered speckle pattern to map the values onto the computational grid.

Results

The analysis validates the efficiency of the framework for the identification of material parameters using a single inhomogeneous displacement field captured by Digital Image Correlation. Further, through examining a boundary value problem replicating a modified experimental setup, highly predictive numerical results that closely match experimental data are obtained incorporating the identified parameters. In contrast, the acquired response using parameters derived from a classical fitting method shows notable deviations. The difference becomes particularly significant for a complex roof structure investigated as an example engineering problem.

Conclusion

Based on real experiments, the experimental-numerical framework turns out to indeed allow for an efficient and unique identification of material parameters in hyperelastic models for coated woven fabrics. Since a variety of different stress-ratios are integrated at once in the experiment by considering inhomogeneous kinematics, also an improved accuracy in structural simulations is found.