An informed regression-based knowledge distillation framework for simultaneously predicting physical and mechanical properties of thermoset epoxy polymers

Machine Learning


  • Jin, F.-L., Lee, X., Park, S.-J. Synthesis and applications of epoxy resins: a review. J. Indus. Engineering chemistry. 291–11 (2015).

    Article CAS Google Scholar

  • Akira Shundo, Shin Yamamoto, Kazuya Tanaka, Network formation and physical properties of epoxy resins with a view to future practical application. jacks o 21522–1542 (2022).

    Article CAS PubMed PubMed Central Google Scholar

  • Rahman, MM & Akhtarul Islam, M. Application of epoxy resins in building materials: Progress and prospects. Polymer bull. 791949-1975 (2022).

    Article CAS Google Scholar

  • Amamoto, Y. A data-driven approach to structure-property relationships in polymer science for prediction and understanding. polymer j. 54957–967 (2022).

    Article CAS Google Scholar

  • Zeng, M. Graph convolutional neural networks for polymer property prediction. arXiv preprint arXiv:1811.06231 (2018).

  • Kuenneth, C. et al. Polymer informatics with multitask learning. pattern 2100238 (2021).

    Article CAS PubMed PubMed Central Google Scholar

  • Tao, L., Varshney, V. & Li, Y. Benchmarking machine learning models for polymer informatics: The glass transition temperature example. J. Chem. information model. 615395–5413 (2021).

    Article CAS PubMed Google Scholar

  • Jin, K., Luo, H., Wang, Z., Wang, H. & Tao, J. Composition optimization of high-performance epoxy resins based on molecular dynamics and machine learning. meter. Death. 194108932 (2020).

    Article CAS Google Scholar

  • Luo, H., Jin, K., Tao, J., Wang, H. Property prediction and design of self-healing epoxy resins by combining molecular dynamics simulation and backpropagation neural network. meter. Resolution experience value 8045308 (2021).

    Article CAS Google Scholar

  • Liu, B. et al. Optimizing the performance of shape memory epoxy polymers based on machine learning. Polymer advance technology. 331222–1232 (2022).

    Article CAS Google Scholar

  • Higuchi, C., Horvath, D., Marcou, G., Yoshizawa, K. & Varnek, A. Prediction of glass transition temperatures of linear homo/heteropolymers and crosslinked epoxy resins. ACS Applications Polymer Materials. 11430–1442 (2019).

    Article CAS Google Scholar

  • Meier, S., Albuquerque, RQ, Demleitner, M. & Ruckdäschel, H. Modeling the glass transition temperature of epoxy systems: a machine learning study. J. Mater. Science. 5713991–14002 (2022).

    Article ADS CAS Google Scholar

  • Hu, Y., Zhao, W., Wang, L., Lin, J. & Du, L. Machine learning-assisted design of highly tough thermoset polymers. ACS Application Meter. interface 1455004–55016 (2022).

    Article CAS PubMed Google Scholar

  • Sindu, B. & Hamekers, J. Feature-based prediction of properties of cross-linked epoxy polymers using molecular dynamics and machine learning techniques. model. Simul. meter. Science. engineering 33065010 (2025).

    Article ADS Google Scholar

  • Yan, C., Feng, X., Wick, C., Peters, A. & Li, G. Machine learning aided the discovery of a new thermoset shape memory polymer based on a small training dataset. polymer 214123351 (2021).

    Article CAS Google Scholar

  • Yan, C., Feng, X. & Li, G. From drug molecules to thermosetting shape memory polymers: a machine learning approach. ACS Application Meter. interface 1360508–60521 (2021).

    Article CAS PubMed Google Scholar

  • Liu, B., Wang, Y., Rabczuk, T., Olofsson, T. & Lu, W. Multiscale modeling of thermal conductivity of polyurethane incorporating phase change materials using physics-based neural networks. Update. energy 220119565 (2024).

    Article CAS Google Scholar

  • Liu, B. et al. Explainable machine learning for multiscale thermal conductivity modeling of polymer nanocomposites using uncertainty quantification. component. structure. 370119292 (2025).

    Article CAS Google Scholar

  • Mahmud, KR, Wang, L., Chen, J. & Hassan, S. A graph neural network-based deep learning framework for predicting thermomechanical behavior of thermoset shape memory polymers. polymer 128771 (2025).

  • von Luden, L. et al. Informed Machine Learning – Classification and exploration to integrate prior knowledge into learning systems. IEEE Trans. Please know. data engineering 35614–633 (2021).

    Google Scholar

  • Garcia, FG, Soares, BG, Pita, VJ, Sánchez, R. & Rieumont, J. Mechanical properties of epoxy networks based on DGEBA and aliphatic amines. J. Appl. polymer science. 1062047-2055 (2007).

    Article CAS Google Scholar

  • González Garcia, F., Leyva, ME, Oliveira, MG, De Queiroz, AA A. & SimoEs, AZ Influence of curing agent chemical structure on mechanical and adhesive properties of epoxy polymers. J. Appl. polymer science. 1172213–2219 (2010).

  • Knorr, DB Jr. et al. Glass transition dependence of ultrafast strain rate response in amine-cured epoxy resins. polymer 535917–5923 (2012).

    Article CAS Google Scholar

  • Bellenger, V., Dhaoui, W., Morel, E. & Verdu, J. Packing density of amine-crosslinked stoichiometric epoxy networks. J. Appl. polymer science. 35563–571 (1988).

    Article CAS Google Scholar

  • Pineda, Á. FE, Garcia, FG, Soares, BG, Simões, AZ & da Silva, EL Comparative study of glycerol diglycidyl ether/aliphatic amine networks. Ageist. 648–65 (2017).

    Google Scholar

  • Prozonic, TM Effect of epoxy network structure on toughness. PhD thesis (Lehigh University School, 2012).

  • Lee, J. & Yee, A. The role of inherent matrix toughness on the failure of glass bead-filled epoxies. polymer 418375–8385 (2000).

    Article CAS Google Scholar

  • Puruksawan, S., Lambard, G., Samitsu, S., Sodeyama, K., Naito, M. Prediction and optimization of epoxy bond strength from small datasets using active learning. Science. technology. Advanced meter. 201010–1021 (2019).

    Article CAS PubMed PubMed Central Google Scholar

  • Prolongo, S.G., del Rosario, G. & Ureña, A. Comparative study on the adhesive properties of different epoxy resins. internal. J. glue. Glue. 26125–132 (2006).

    Article CAS Google Scholar

  • Jang, J. Effect of particle size on mechanical and thermal properties. \(SiO_2\) Particulate polymer composite. Doctoral dissertation (University of Delaware School, 2012).

  • Selby, K. & Miller, L. Fracture toughness and mechanical behavior of epoxy resins. J. Mater. Science. 1012–24 (1975).

    Article ADS CAS Google Scholar

  • Kinloch, A. Mechanics and failure mechanisms of thermosetting epoxy polymers. in Epoxy resins and composites I. 45–67 (Springer, 2005).

  • Aziz, Maine Study the mechanical properties of epoxy resins cured using various curing agents and at constant curing times and temperatures. engineering technology. J. 29 (2011).

  • Liu, Z., Erhan, S. & Calvert, P. Solid free-form fabrication of epoxidized soybean oil/epoxy composites using bis- or polyalkyleneamine curing agents. component. Part A Applied Example Science. Manufacturer 3887–93 (2007).

    Article CAS Google Scholar

  • Vanlandingham, M., Eduljee, R. & Gillespie, Jr. Relationship between stoichiometry, microstructure, and properties of amine-cured epoxies. J. Appl. Polym. Science. 71699–712 (1999).

    Article CAS Google Scholar

  • Elmadi, A. et al. Effect of strain rate and silica filler content on the compressive behavior of rtm6 epoxy-based nanocomposites. polymer 133735 (2021).

    Article CAS PubMed PubMed Central Google Scholar

  • Behzadi, S. & Jones, FR Yielding behavior of model epoxy matrices of fiber-reinforced composites: Effects of strain rate and temperature. J. Macromol. Science. Part B Physics. 44993–1005 (2005).

    Article ADS CAS Google Scholar

  • Sun, L. et al. Mechanical properties of surface-functionalized SWCNT/epoxy composites. carbon 46320–328 (2008).

    Article CAS Google Scholar

  • Littell, JD et al. Measuring tensile, compressive, and shear stress-strain curves of epoxy resins over a wide range of strain rates using small specimens. J. Aerosp. engineering twenty one162–173 (2008).

    Article Google Scholar

  • Jordan, JL, Foley, JR, Siviour, CR Mechanical properties of Epon 826/DEA epoxy. Mecha. Depends on the time. meter. 12249–272 (2008).

    Article ADS CAS Google Scholar

  • Jakubinek, MB et al. Nano-reinforced epoxy and adhesive bonding incorporating boron nitride nanotubes. internal. J. glue. Glue. 84194–201 (2018).

    Article CAS Google Scholar

  • Fard, MY, Liu, Y. & Chattopadhyay, A. Characterization of epoxy resins including strain rate effects using digital image correlation system. J. Aerosp. engineering twenty five308–319 (2012).

    Article Google Scholar

  • Chen, W., Lu, F. & Cheng, M. Tensile and compression testing of two polymers under quasi-static and dynamic loading. Polymer test. twenty one113–121 (2002).

    Article CAS Google Scholar

  • Rostamiyan, Y., Fereidoon, A. & Mashhadzadeh, AH Experimental study on mechanical properties of epoxy-based nanocomposites using polymer alloying and various nanoreinforcements (nanofibers, nanolayers, and nanoparticulate materials). Science. Engineering component. meter. twenty two591–598 (2015).

    Article CAS Google Scholar

  • Kochergin, YS, Grigorenko, T., Popova, O. & Samoilova, E. Properties of epoxy polymers cured with polyoxypropylene diamine. Polym. Science. Sir. D 3231–234 (2010).

    Article Google Scholar

  • Morrill, JA, Jensen, RE, Madison, PH & Chabalowski, CF Prediction of formulation dependence of glass transition temperature of amine-epoxy copolymers using QSPR based on AM1 method. J. Chem. Calculate information. Science. 44912–920 (2004).

    Article CAS PubMed Google Scholar

  • Paszke, A. Pytorch: An imperative-style high-performance deep learning library. Advanced neural information processes. system. 32 (2019).

  • Weininger, D. Smiles, Chemical Language and Information Systems. 1. Introduction to methodology and encoding rules. J. Chem. Calculate information. Science. 2831–36 (1988).

  • Rdkit: Open source chemoinformatics. https://www.rdkit.org.



  • Source link