STM Journals Peer-reviewed research from Consortium e-Learning Network

Journal of Polymer & Composites Original Research

Data-Driven Digital Twin Model for Real-Time Strength Estimation in Polymeric Materials

  1. Makrand V. Kulkarni School of Sciences, Sanjivani University, Kopargaon
  2. Kalpana G. Joshi School of Engineering and Technology, Sanjivani University, Kopargaon
  3. Shanthi Kumaraguru Department of Information Technology, D Y Patil College of Engineering, Akurdi, Pune
  4. Vinit Kotak Department of Electrical Engineering, Shah & Anchor Kutchhi Engineering College, Mumbai
  5. Prashant V. Thokal Department of Electrical Engineering, Sanjivani College of Engineering, Kopargaon
  6. Prashant M. Yawalkar Department of Computer Engineering, MET’s Institute of Engineering, BKC, Nashik
  7. Vijay More Department of Computer Engineering, MET’s Institute of Engineering, BKC, Nashik

Abstract

The real-time prediction of mechanical properties in polymeric materials is essential for ensuring quality, consistency, and operational efficiency in modern manufacturing systems. As industrial processes become increasingly complex, traditional trial-and-error approaches to material characterization are no longer sufficient to meet the demands of high-throughput production environments. This study introduces a digital twin-integrated machine learning approach for the real-time estimation of tensile strength in polymeric materials by combining simulation-driven insights with advanced data-centric modeling techniques. A comprehensive digital twin framework is developed to accurately replicate the physical and thermomechanical behavior of polymer processing, establishing a virtual counterpart that mirrors real-world manufacturing conditions with high fidelity.Ensemble machine learning models—namely Random Forest and Gradient Boosting—are employed for predictive analysis and benchmarked against one another to determine the most effective modeling strategy. The models are trained on a synthetic dataset comprising 500 instances across six widely used polymer categories (PE, PP, PVC, PET, PA, and PS), incorporating eight significant process variables that capture the multidimensional nature of polymer behavior. Performance evaluation reveals that the Gradient Boosting model outperforms competing approaches, achieving an R² of 0.9101, a Mean Absolute Error of 6.61 MPa, and a Root Mean Square Error of 8.20 MPa, collectively indicating strong predictive capability and generalization. Feature importance analysis further highlights molecular weight, crystallinity, and processing temperature as the primary factors governing tensile strength outcomes. The proposed system enables continuous monitoring and real-time prediction within dynamic manufacturing environments, thereby supporting proactive process adjustments and improved product quality control. Overall, this work advances the evolution of smart manufacturing under the Industry 4.0 paradigm by seamlessly integrating physical and digital systems to enable data-driven optimization and proactive control of material properties.

Keywords

References (28)

  1. Yang H, Karimi S. Machine learning-driven digital twin for strength prediction of dissimilar adhesive joints under environmental aging. Journal of Composite Materials. 2025;59(24):2831-2851. doi:10.1177/00219983251342147
  2. Luo S, Karimi S. Adaptive digital twin framework for monitoring and predicting the performance of composite adhesive joints. The Journal of Adhesion. 2025;102(4):321-350. doi:10.1080/00218464.2025.2535687
  3. Hürkamp A, Gellrich S, Ossowski T, Beuscher J, Thiede S, Herrmann C, et al. Combining Simulation and Machine Learning as Digital Twin for the Manufacturing of Overmolded Thermoplastic Composites. Journal of Manufacturing and Materials Processing. 2020;4(3):92. doi:10.3390/jmmp4030092
  4. Butt J, Mohaghegh V. Combining Digital Twin and Machine Learning for the Fused Filament Fabrication Process. Metals. 2022;13(1):24. doi:10.3390/met13010024
  5. Khan I, Al Rashid A, Koç M. Integration of machine learning and digital twin in additive manufacturing of polymeric-based materials and products. Progress in Additive Manufacturing. 2025;10(12):10685-10737. doi:10.1007/s40964-025-01257-4
  6. Suess et al., "Polyester resin–quartz composites in the age of artificial intelligence and digital twins," 2025.
  7. Aysa NH, Karimi S. Integrating machine learning and digital twin for strength prediction of CFRP/aluminum adhesive joints under hygrothermal conditions. Polymer Composites. 2025;46(14):13236-13255. doi:10.1002/pc.29928
  8. Mishra A, Jatti VS, Messele Sefene E, Jatti AV, Sisay AD, Khedkar NK, et al. Machine learning-assisted pattern recognition algorithms for estimating ultimate tensile strength in fused deposition modelled polylactic acid specimens. Materials Technology. 2023;39(1). doi:10.1080/10667857.2023.2295089
  9. Ji X, Wang J, Gong Y, Bao Y, Yang Y. Experimental study on the mechanical property of carbon fiber reinforced polymer with the combined application of digital twin and reduced order algorithm. Polymer Composites. 2024;45(12):10761-10774. doi:10.1002/pc.28505
  10. Zhan S, Li Z, Zhao H, Liu Z, Li Q, Ji S, et al. Transfer Learning for Polymer Mechanics: A Fusion Approach to Bridge Molecular Dynamics Simulations and Experiments in SSBR. Macromolecular Rapid Communications. 2025;47(12). doi:10.1002/marc.202500386
  11. B S S, Hamaekers J. Feature-based prediction of properties of cross-linked epoxy polymers by molecular dynamics and machine learning techniques. Modelling and Simulation in Materials Science and Engineering. 2025;33(6):065010. doi:10.1088/1361-651x/adf56c
  12. Menon A, Thompson-Colón JA, Washburn NR. Hierarchical Machine Learning Model for Mechanical Property Predictions of Polyurethane Elastomers From Small Datasets. Frontiers in Materials. 2019;6. doi:10.3389/fmats.2019.00087
  13. Wang B, Li Z, Xu Z, Sun Z, Tian K. Digital twin modeling for structural strength monitoring via transfer learning-based multi-source data fusion. Mechanical Systems and Signal Processing. 2023;200:110625. doi:10.1016/j.ymssp.2023.110625
  14. Khayyami, "Predicting mechanical properties of polymer films after extrusion coating using supervised machine learning algorithms," 2019.
  15. Stieber S, Hoffmann A, Schiendorfer A, Reif W, Beyrle M, Faber J, et al. Towards Real-time Process Monitoring and Machine Learning for Manufacturing Composite Structures. 2020 25th IEEE International Conference on Emerging Technologies and Factory Automation (ETFA). 2020:1455-1458. doi:10.1109/etfa46521.2020.9212097
  16. Huang Y, Chen Q, Zhang Z, Gao K, Hu A, Dong Y, et al. A Machine Learning Framework to Predict the Tensile Stress of Natural Rubber: Based on Molecular Dynamics Simulation Data. Polymers. 2022;14(9):1897. doi:10.3390/polym14091897
  17. Utku A, Kısmet Y, Can Ü. Prediction of Mechanical Properties of Synthetic Waste Reinforced Polyolefins with GA-LSTM Hybrid Model. Muş Alparslan Üniversitesi Fen Bilimleri Dergisi. 2024;12(2):114-125. doi:10.18586/msufbd.1535577
  18. Cravero F, Ponzoni I, Diaz MF. Improving predictive modeling of polymeric materials using a hybrid approach of machine learning and expert intervention. Proceedings of the 21th LACCEI International Multi-Conference for Engineering, Education and Technology (LACCEI 2023): “Leadership in Education and Innovation in Engineering in the Framework of Global Transformations: Integration and Alliances for Integral Development”. 2023. doi:10.18687/laccei2023.1.1.1572
  19. Zhou S, Zhao W, Wan Z, Qiu H, Huang X, Sun ZY. Multimodal Machine Learning with 3D‐Weighted‐Matrix Encoding for High‐Throughput Design of High‐Performance Polyurethanes. Macromolecular Rapid Communications. 2025;47(12). doi:10.1002/marc.202500471
  20. Xu X, Wang G, Xuan S, Shan Y, Yang H, Yao X. Digital twin and cross-scale mechanical interaction for fabric rubber composites considering model uncertainties. Composites Science and Technology. 2024;248:110431. doi:10.1016/j.compscitech.2024.110431
  21. Yang et al., "Physics augmented machine learning discovery of composition-dependent constitutive laws for 3D printed digital materials," arXiv preprint, 2025, doi:10.48550/arxiv.2507.02991.
  22. Anvari A, Karimi S, Latifi A. A digital twin and machine learning approach for real-time fatigue life prediction of CFRP-to-aluminum adhesive joints subjected to environmental aging. The International Journal of Advanced Manufacturing Technology. 2025;139(5-6):2783-2801. doi:10.1007/s00170-025-16042-4
  23. Gyabaah et al., "Machine learning-assisted polymer and polymer composite design for additive manufacturing," 2025.
  24. Palanisamy S, Kalimuthu M, Azeez A, Palaniappan M, Dharmalingam S, Nagarajan R, et al. Wear Properties and Post-Moisture Absorption Mechanical Behavior of Kenaf/Banana-Fiber-Reinforced Epoxy Composites. Fibers. 2022;10(4):32. doi:10.3390/fib10040032
  25. Aruchamy K, Karuppusamy M, Krishnakumar S, Palanisamy S, Jayamani M, Sureshkumar K, et al. Enhancement of mechanical properties of hybrid polymer composites using palmyra palm and coconut sheath fibers: The role of tamarind shell powder. BioResources. 2024;20(1):698-724. doi:10.15376/biores.20.1.698-724
  26. Ayrilmis N, Kanat G, Yildiz Avsar E, Palanisamy S, Ashori A. Utilizing waste manhole covers and fibreboard as reinforcing fillers for thermoplastic composites. Journal of Reinforced Plastics and Composites. 2024;44(17-18):1108-1118. doi:10.1177/07316844241238507
  27. Ramasubbu R, Kayambu A, Palanisamy S, Ayrilmis N. Mechanical properties of epoxy composites reinforced with Areca catechu fibers containing silicon carbide. BioResources. 2024;19(2):2353-2370. doi:10.15376/biores.19.2.2353-2370
  28. Palanisamy S, Murugesan TM, Palaniappan M, Santulli C, Ayrilmis N. Fostering sustainability: The environmental advantages of natural fiber composite materials – a mini review. Environmental Research and Technology. 2024;7(2):256-269. doi:10.35208/ert.1397380
Support