Use of An Artificial Neural Networks for Optimizing Technological Parameters of High-Alloy Surfacing Processes
DOI:
https://doi.org/10.17072/1993-0550-2026-3-87-98Keywords:
artificial neural network, genetic algorithm, optimization, arc surfacing, powder strip, transfer coefficient, machine learning, weldingAbstract
The development of high-alloy compositions for arc surfacing using powder strips is a complex multi-parameter problem that involves ensuring the maximum transfer of alloying elements from the electrode to the deposited metal, while accounting for losses due to spattering and evaporation. The relevance of the study stems from the need to improve the accuracy of predicting the chemical composition of the deposited layer and to reduce the amount of experimental work. The study aims to develop a method for optimizing technological parameters of the surfacing process based on a combination of an artificial neural network and a genetic algorithm. Artificial neural networks were used to model the dependence of transfer coefficients on technological parameters, while a genetic algorithm was applied to determine their optimal values. The following parameters were found to be optimal: current – 400 A, electrode extension – 30 mm, and ferrotitanium fraction – 0.4 mm. The proposed approach made it possible to increase the transfer coefficients by an average of 11.2 %, including titanium by 30.7 %, manganese by 7.9 %, and carbon by 12.5 %. The results confirm the effectiveness of the proposed approach for optimizing surfacing processes.
References
Chigarev V. V., Zarechenskiy D. A., Belik A. G. Optimisation of the composition and melting parameters of powder strips with the exothermic mixture in the filler // Welding International. 2016. Vol. 30, № 7. P. 557–559.
Шалимов М. П., Вотинова Е. Б. Расчетно-экспериментальная методика оценки параметров перехода элементов в металл шва при дуговой сварке и совершенствование состава шихты порошковой проволоки // Вестник ПНИПУ. Машиностроение, материаловедение. 2019. Т. 21, № 1. С. 28–35.
Козырев Н. А., Крюков Р. Е., Усольцев А. А., Кибко Н. В., Бащенко Л. П. Разработка новой порошковой проволоки на основе пыли газоочистки силикомарганца // Известия вузов. Черная металлургия. 2021. Т. 64, № 12. С. 863–869.
Зареченский Д. А., Ашла А. А., Шпак А. А. Особенности легирования наплавленного металла при наплавке порошковыми лентами // Наука – общество – технологии – 2025: материалы Всерос. науч.-практ. конф. М.: Московский политехнический университет, 2025. С. 167–171.
Чигарев В. В., Белик А. Г. Порошковые ленты для наплавки // Сварочное производство. 2011. № 8. С. 38–44.
Nagesh D. S., Datta G. L. Prediction of weld bead geometry and penetration in shielded metal arc welding using artificial neural networks // Journal of Materials Processing Technology. 2002. Vol. 123, № 2. P. 303–312.
Ates H. Prediction of gas metal arc welding parameters based on artificial neural networks // Materials & Design. 2007. Vol. 28, № 7. P. 2015–2023.
Zhang Y., Chen J., Zhang Y. Application of artificial intelligence in welding process optimization: A review // Journal of Manufacturing Processes. 2020. Vol. 58. P. 1024–1043.
Katherasan D. et al. Simulation and parameter optimization of flux cored arc welding using artificial neural network and particle swarm optimization algorithm // Journal of Intelligent Manufacturing. 2014. Vol. 25. P. 67–76.
Deb K. Multi-Objective Optimization Using Evolutionary Algorithms. New York: John Wiley & Sons, 2001. 518 p.
Rao R. V., Kalyankar V. D. Parameter optimization of welding processes using a new optimization algorithm // Materials and Manufacturing Processes. 2013. Vol. 28, № 8. P. 918–923.
Goldberg D. E. Genetic Algorithms in Search, Optimization and Machine Learning. Reading, MA: Addison-Wesley, 1989. 412 p.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Роман Владиславович Ковальчик, Наталья Юрьевна Ротанева, Денис Александрович Зареченский
This work is licensed under a Creative Commons Attribution 4.0 International License.
Articles are published under license Creative Commons Attribution 4.0 International (CC BY 4.0).
