This is an unedited manuscript accepted for publication and provided as an Article in Press for early access at the author’s request. The article will undergo copyediting, typesetting, and galley proof review before final publication. Please be aware that errors may be identified during production that could affect the content. All legal disclaimers of the journal apply.
Gopu Vijayalaxmi,
Kumaresan M,
C Indirani,
T. Vaishnavi,
S.Gopalakrishnan,
M.Krishnamurthy,
- Assistant Professor, Department of Computer Science and Engineering (Artificial Intelligence & MachineLearning), Vaagdevi College of Engineering , Bollikunta Warangal, Telangana, India
- Assistant Professor, Department of Information Technology,Dr. N.G.P. Institute of Technology, Coimbatore, Tamil Nadu, India
- Assistant Professor III, Department of Mathematics, Bannari Amman Institute of Technology, Sathyamangalam, Erode, Tamil Nadu, India
- Assistant Professor, Department of Electronics and Communication Engineering, Vel Tech Rangarajan Dr. Sagunthala R&D Institute of Science and Technology, Chennai, Tamil Nadu, India
- Professor, Department of Electronics and Communication and Engineering, Saveetha School of Engineering, Saveetha Institute of Medical and Technical Sciences, Thandalam, Chennai, Tamil Nadu, India
- Assistant Professor, Department of Electronics and Communication Engineering PSNA College of Engineering and Technology, Dindigul, Tamil Nadu, India
Abstract
Knowing the polymers elasticity and mechanical strength is essential in the characterization of functional polymer materials, for rapid characterization and to avoid extensive experiments. Traditional polymer characterization methods are often time consuming and resource intensive and the mechanical response of polymers can be nonlinear, thus complicating the prediction of the properties. These limitations can be overcome by introducing a Variational Temporal Fusion Transformer (VaTFT) for the deep learning prediction of polymer elasticity and mechanical strength, as suggested in this paper. The idea is to model compact representations from polymer material data as probabilistic representations using a Variational Autoencoder (VAE) and to model complex feature interactions using a Temporal Fusion Transformer (TFT). VAE extracts informative latent features and discards redundant information, while the TFT applies attention based feature selection and contextual fusion modeling of the nonlinear relationship well. The resulting representation is then used to make mechanical property predictions, and is utilized for performance evaluation. Proposed VaTFT model was then compared to other models including MLP, CNN, LSTM, BiLSTM, GRU, Transformer and VAE-DNN. VaTFT achieved 98.72% accuracy, 98.41% precision, 98.58% recall, and 98.49% F1-score, outperforming VAE-DNN with 94.16% accuracy. Moreover, VaTFT achieved the lowest RMSE value of 0.041, 63.39% lower than 0.112 of VAE-DNN, which shows 63.39% improvement in prediction accuracy. The training time is 146 s, demonstrating the efficiency of training for the model. The results support the efficacy of VaTFT in accurate and efficient predicting of polymer mechanical properties.
Keywords: Polymers elasticity, Deep Learning, Polymer Material Characterization, Mechanical strength, Transformer Model and Industrial Application.
References
1. I. P. Malashin, D. Martysyuk, V. Nelyub, A. Borodulin, A. Gantimurov, and V. Tynchenko, “Deep learning for property prediction of natural fiber polymer composites,” Scientific Reports, vol. 15, p. 27837, 2025.
2. W. Ge, R. De Silva, Y. Fan, S. A. Sisson, and M. H. Stenzel, “Machine learning in polymer research,” Advanced Materials, vol. 37, p. 2413695, 2025.
3. Y. Liang, X. Wei, Y. Peng, X. Wang, and X. Niu, “A review on recent applications of machine learning in mechanical properties of composites,” Polymer Composites, vol. 46, pp. 1939-1960, 2025.
4. H. Sharma, G. Arora, M. K. Singh, V. Ayyappan, P. Bhowmik, S. M. Rangappa, et al., “Review of machine learning approaches for predicting mechanical behavior of composite materials,” Discover Applied Sciences, vol. 7, p. 1238, 2025.
5. N. Mathiazhagan, S. Palaniyappan, M. Tamilselvan, and N. K. Sivakumar, “Recycling potential of 3D printed support waste as eco-friendly polymeric composites,” Progress in Rubber, Plastics and Recycling Technology, vol. 42, pp. 137-157, 2026.
6. S. Kingsley, P. J. Jayarin, K. Sreenu, S. M. Ali, K. Solayan, and S. Ganesan, “Reinforcement Learning for Real-Time Traffic Management: A Robust DQN-Based Signal Control Approach with OpenTAPAS Integration,” VFAST Transactions on Software Engineering, vol. 14, pp. 143-162, 2026.
7. M. H. Niaki, M. Pashaian, and M. G. Ahangari, “Experimental assessment of the efficiency of deep learning method in predicting the mechanical properties of polymer concretes and composites,” Journal of Building Engineering, vol. 76, p. 107107, 2023.
8. P. Pattanayak, J. Giri, H. A. La-Lohedan, N. K. Akkasali, A. K. Meher, S. K. Bohidar, et al., “Numerical prediction and experimental verification of dynamic deflection responses of damaged hybrid (luffa-and carbon-nanotube-reinforced) polymeric composite panel structure,” International Journal of Applied Mechanics, vol. 17, p. 2450123, 2025.
9. S. Pitta, S. Gopalakrishnan, and S. R. Chand, “A hybrid approach for intrusion detection in cloud-integrated WSNs: AqRO for feature selection and CapTAD for temporal dynamics,” International Journal of System Assurance Engineering and Management, vol. 17, pp. 1523-1545, 2026.
10. S. Olhan, B. Antil, and P. Maimí, “Machine learning in next-generation polymer composites: recent advances and perspectives,” Applied Composite Materials, vol. 33, p. 54, 2026.
11. D. Suresh, M. S. Sheela, P. Sunitha, and S. Gopalakrishnan, “Physics-Guided Deep Learning with Bayesian Optimization for Enhanced River Streamflow Prediction,” Water Resources Management, vol. 40, p. 127, 2026.
12. Koltsakidis, E. K. Tzimtzimis, and D. Tzetzis, “Machine Learning for Predicting Mechanical Properties of 3D-Printed Polymers from Process Parameters: A Review,” Polymers, vol. 18, p. 499, 2026.
13. E. Champa-Bujaico, P. García-Díaz, and A. M. Díez-Pascual, “Machine learning for property prediction and optimization of polymeric nanocomposites: a state-of-the-art,” International Journal of Molecular Sciences, vol. 23, p. 10712, 2022.
14. L. Gao, J. Lin, L. Wang, and L. Du, “Machine learning-assisted design of advanced polymeric materials,” Accounts of Materials Research, vol. 5, pp. 571-584, 2024.
15. A. Sharma, T. Mukhopadhyay, S. M. Rangappa, S. Siengchin, and V. Kushvaha, “Advances in computational intelligence of polymer composite materials: machine learning assisted modeling, analysis and design,” Archives of Computational Methods in Engineering, vol. 29, pp. 3341-3385, 2022.
16. C. Xie, H. Qiu, L. Liu, Y. You, H. Li, Y. Li, et al., “Machine learning approaches in polymer science: progress and fundamental for a new paradigm,” SmartMat, vol. 6, p. e1320, 2025.
17. . Yue, J. He, L. Tao, and Y. Li, “High-throughput screening and prediction of high modulus of resilience polymers using explainable machine learning,” Journal of Chemical Theory and Computation, vol. 19, pp. 4641-4653, 2023.
18. Tian, Y., Wang, Y., Zhang, Z., Chen, B., & Du, S. (2024, May). Joma: Demystifying multilayer transformers via joint dynamics of mlp and attention. In International Conference on Learning Representations (Vol. 2024, pp. 36365-36390).
19. Zhang, F., Guo, H., Wang, W., Du, J., Liu, Y., & Wang, H. (2025). Fast prediction of curing residual stress in polymer‐matrix composites by a deep learning method. Polymer Composites, 46(9), 8424-8441.
20. Zhan, S., Li, Z., Zhao, H., Liu, Z., Li, Q., Ji, S., … & Liu, J. (2026). Transfer learning for polymer mechanics: a fusion approach to bridge molecular dynamics simulations and experiments in SSBR. Macromolecular Rapid Communications, 47(12), e00386.
21. Jian, Y., Hu, P., Zhou, Q., Zhang, N., Cai, D. A., Zhou, G., & Wang, X. (2025). A novel bidirectional LSTM network model for very high cycle random fatigue performance of CFRP composite thin plates. International Journal of Fatigue, 190, 108627.
22. Parvez, M. A., & Mehedi, I. M. (2025). High-accuracy polymer property detection via Pareto-optimized SMILES-based deep learning. Polymers, 17(13), 1801.
23. Nigam, V., Chandrasekhar, A., & Barati Farimani, A. (2026). Polymer-agent: Large language model agent for polymer design. Journal of Chemical Information and Modeling, 66(9), 5145-5154.
24. Yang, H., Gao, J., Heng, F., Cheng, Q., & Liu, Y. (2024). A deep neural network method for LCF life prediction of metal materials with small sample experimental data. Metals and Materials International, 30(6), 1667-1686.
Journal of Polymer & Composites
| Volume | 14 | |
| 05 | ||
| Received | 18/08/2026 | |
| Accepted | 06/10/2026 | |
| Published | 15/10/2026 | |
| Publication Time | 58 Days |