Development and prediction of properties of epoxy compositions using machine learning methods

Mishurov K.S., Monakhov A.D., Sarychev I.A. S
UDC
004.85
DOI
10.18577/2307-6046-2026-0-1-162-173
Article PDF (Russian) (832.44 KB)
How to cite
Mishurov K.S., Monakhov A.D., Sarychev I.A. S Development and prediction of properties of epoxy compositions using machine learning methods // Proceedings of VIAM. 2026. No. 1. DOI: 10.18577/2307-6046-2026-0-1-162-173. URL: https://test.viam.ru/en/journal/2026/1/12
Keywords
epoxy binder, adhesive, machine learning, active learning, Bayesian optimization, prediction of properties, design
Abstract

The article analyzes the application of machine learning methods for the development of epoxy compositions with specified properties. Approaches to the formation of training data sets based on limited number of experiments are considered. Special attention is paid to the methods of active learning and Bayesian optimization methods, which allow efficient planning of experiments. Practical examples of using machine learning models for predicting the properties of developed compositions are given. The advantages of combining different machine learning methods for solving complex problems in materials science are shown.

Reference list
  1. Kablov E.N., Chursova L.V., Babin A.N., Mukhametov R.R., Panina N.N. Developments of FSUE VIAM in the field of melt binders for polymer composite materials. Polimernye materialy i tekhnologii, 2016, vol. 2, no. 2, pp. 37–42.
  2. Kablov E.N. New generation materials and digital technologies for their processing. Vestnik Rossiyskoy akademii nauk, 2020, vol. 90, no. 4, pp. 331–334. DOI: 10.31857/S0869587320040052.
  3. Malysheva G.V., Marakhovskiy P.S., Barinov D.Ya., Nikolaev E.V. Optimization of the curing modes of fiber-glass based on epoxy binder. Aviation materials and technologies, 2023, no. 2 (71), pp. 94–103. Available at: http://www.journal.viam.ru (accessed: April 17, 2025). DOI: 10.18577/2713-0193-2023-0-2-94-103.
  4. Tkachuk A.I., Donetsky K.I., Terekhov I.V., Karavaev R.Yu. The use of thermosetting matrices for the manufacture of polymer composite materials by the non-autoclave molding methods. Aviation materials and technology, 2021, no. 1 (62), pp. 22–33. Available at: https://www.journal.viam.ru (accessed: April 17, 2025). DOI: 10.18577/2713-0193-2021-0-1-22-33.
  5. Kablov E.N., Chursova L.V., Lukina N.F., Kutsevich K.E., Rubtsova E.V., Petrova A.P. Study of epoxy-polysulfone polymer systems as the basis for high-strength adhesives for aviation purposes. Klei. Germetiki. Tekhnologii, 2017, no. 3, pp. 7–12.
  6. Moshinsky L. Epoxy resins and hardeners. Tel Aviv: Arcadia press Ltd., 1995, 370 p.
  7. Adler Yu.P., Markova E.V., Granovsky Yu.V. Experimental design in the search for optimal conditions. Moscow: Nauka, 1976, 140 p.
  8. Montgomery D.K. Experimental design and data analysis. Trans. from Engl. Leningrad: Shipbuilding, 1980, 384 p.
  9. Oreshko E.I., Erasov V.S., Sibayev I.G., Lutsenko A.N., Shershak P.V. Machine learning algorithms (review). Part 1. Classification and regression tasks. Linear algorithms in machine learning. Application of machine learning algorithms for calculating the strength characteristics of materials. Aviation materials and technologies, 2022, no. 3 (68), pp. 130‒146. Available at: http://www.journal.viam.ru (accessed: April 20, 2025). DOI: 10.18577/2713-0193-2022-0-3-130-146.
  10. Oreshko E.I., Erasov V.S., Sibayev I.G., Lutsenko A.N., Shershak P.V. Machine learning algorithms (review). Part 2. Machine learning metrics. Decision trees and ensembles of decision trees. An algorithm for neural network to forecast properties of ferrite-martensite steel. Aviation materials and technologies, 2022, no. 4 (69), pp. 132–146. Available at: http://www.journal.viam.ru (accessed: April 20, 2025). DOI: 10.18577/2713-0193-2022-0-4-132-146.
  11. Shonichev I.D., Tynchenko V.S., Borodulin A.S., Muzyka S.S. Application of machine learning technologies in the design and production of structures from composite materials. Vse materialy. Entsiklopedicheskiy spravochnik, 2024, no. 6, pp. 18–27. DOI: 10.31044/1994-6260-2024-0-6-18-27.
  12. Monakhov A.D., Yakovlev N.O. Application of the deep learning method in studying crack resistance characteristics. Trudy VIAM, 2024, no. 6 (136), pp. 80‒91. Available at: http://www.viam-works.ru (accessed: April 20, 2025). DOI: 10.18577/2307-6046-2024-0-6-80-91.
  13. Xu P., Ji X., Li M. et al. Small data machine learning in materials science. npj Computational Materials, 2023, vol. 9, art. 42. DOI: 10.1038/s41524-023-01000-z.
  14. Serrano L. Grokking Machine Learning. St. Petersburg: Piter, 2022, 320 p.
  15. Scholle F. Deep Learning in Python. St. Petersburg: Piter, 2021, 400 p.
  16. Choosing the right estimator. Scikit-learn. Machine learning in Python. Available at: http://scikit- learn.org/stable/tutorial/machine_learning_map (accessed: April 16, 2025).
  17. Theobald O. Machine Learning for Absolute Beginners. Moscow: Bombora, 2024, 208 p.
  18. Sutubalov A.I., Podzhivotov N.Yu., Shershak P.V., Yakovlev N.O. Assessment of the adhesive strength of the fiber–polymer matrix system using the nonparametric bootstrap method. Trudy VIAM, 2024, no. 11 (141), pp. 100‒112. Available at: http://www.viam-works.ru (accessed: April 17, 2025). DOI: 10.18577/2307-6046-2024-0-11-100-112.
  19. Aggarwal C.C. Data Mining: The Textbook. N.-Y.: Springer, 2015, 734 p.
  20. Monarch R. Machine Learning with Human Interaction. Moscow: DMK Press, 2022, 498 p.
  21. Nguyen K. Bayesian Optimization in Practice. Astana: ALIST, 2024, 416 p.
  22. Taniguchi S., Uemura K., Tamaki S. et al. Multi-objective Optimization of the Epoxy Matrix System Using Machine Learning. Results in Materials, 2023, vol. 17 (1), art. 100376. DOI: 10.1016/j.rinma.2023.100376.
  23. Akiba T.T., Sano S., Yanase T. Optuna: a next-generation hyperparameter optimization framework. Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (KDD’19). N.-Y., 2019, pp. 2623–2631. DOI: 10.1145/3292500.3330701.
  24. Rishal H. Grokking Artificial Intelligence Algorithms. St. Petersburg: Piter, 2023, 368 p.
  25. Pruksawan S., Lambard G., Samitsu S. et al. Prediction and optimization of epoxy adhesive strength from a small dataset through active learning. Science and Technology of Advanced Materials, 2019, vol. 20 (1), pp. 1010–1021. DOI: 10.1080/14686996.2019.1673670.
  26. Albuquerque R.Q., Rothenhäusler F., Ruckdäschel H. Designing formulations of bio-based, multicomponent epoxy resin systems via machine learning. MRS Bulletin, 2024, no. 49, pp. 59–70. DOI: 10.1557/s43577-023-00504-9.