A genetic-algorithm-optimised viscoelastic fitting tool for polymers and polymer nanocomposites

An automated tool for analysing the viscoelastic properties of viscous materials in the frequency domain using dynamic mechanical analysis (DMA) data has been developed and made publicly available (https://doi.org/10.5281/zenodo.15075601). The least-squares method and a genetic algorithm were employed for master curve generation and Prony series fitting, respectively. The MATLAB-based tool overcomes the issue of negative modulus values commonly encountered in traditional linear fitting methods, while offering a user-friendly interface and high computational efficiency. The robustness of the tool was also evaluated through two distinct case studies. The first case focused on a recyclable epoxy resin, where the coefficient of determination (R²) between the fitted data and the experimental DMA data reached 0.997 with 11 Prony series terms. The derived viscoelastic parameters were further validated by comparing the stress relaxation behaviour predicted by a finite element model with experimental test results of the recyclable resin. In the second case, the viscoelastic behaviour of an epoxy nanocomposite modified with carbon nanotubes was investigated using the automated tool, demonstrating its versatility and effectiveness in characterising different material systems.