Use of auto machine learning, artificial intelligence, for predictive modeling of metallurgical properties of hot-rolled steel products
Alisson Paulo de Oliveira; Leonardo Sene de Lourenço
Abstract
This study aimed at predictive modeling, with Artificial Intelligence (AI), of the mechanical properties of sections produced from High Strength and Low Alloy (HSLA) steel, hot-rolled. The models were based on historical data of mechanical properties, along with the chemical composition of heats and the parameters of the rolling process. An AutoMachine Learning platform was used. This tool can test dozens of algorithms to achieve the lowest error. Simplified models were built based on statistical analyses of the database, while expanded models were developed using all available data. Despite mathematical precision, the models were developed to be metallurgically coherent with scientific trends. The results aligned well with expected trends in most cases. It was possible to evaluate the isolated effect of variables. The expanded models were able to generate predictions with lower statistical errors. Data variability is an important factor for the success of predictive models. Such models allow alloy design to be defined with greater precision, leading to reduced production costs and a better understanding of the effects of input variables. Data-driven decision-making is enhanced with AI.
Keywords
Referências
1 Pan G, Wang F, Shang C, Wu H, Wu G, Gao J, et al. Advances in machine learning- and artificial intelligence-assisted material design of steels. International Journal of Minerals Metallurgy and Materials. 2023;30(6):1003-1024. http://doi.org/10.1007/s12613-022-2595-0.
2 Duan H, He H, Yue S, Cao M, Zhao Y, Zhang Z, et al. Analysis of high-cycle fatigue life prediction of 304 stainless steel based on deep learning. JOM. 2023;75(11):4586-4595. http://doi.org/10.1007/s11837-023-06042-8.
3 Srivastava A, Patra P, Jha R. AHSS applications in Industry 4.0: determination of optimum processing parameters during coiling process through unsupervised machine learning approach. Materials Today. Communications. 2022;31:103625. http://doi.org/10.1016/j.mtcomm.2022.103625.
4 Hamdi A, Merghache SM. Application of artificial neural networks (ANN) and gray relational analysis (GRA) to modeling and optimization of the material ratio curve parameters when turning hard steel. International Journal of Advanced Manufacturing Technology. 2023;124(10):3657-3670. http://doi.org/10.1007/s00170-023-10833-3.
5 Czinege I, Harangozó D. Application of artificial neural networks for characterisation of formability properties of sheet metals. Int J Lightweight Mater Manuf. 2024;7(1):37-44. http://doi.org/10.1016/j.ijlmm.2023.08.003.
6 Monroe R. Data-driven design properties for cast carbon steels. Inter Metalcast. 2024;18: 2756-2777. https://doi. org/10.1007/s40962-023-01195-3.
7 Shang C, Wang C, Wu H, Liu WY, Chen YM, Pan GF, et al. Improved data-driven performance of Charpy impact toughness via literature-assisted production data in pipeline steel. Science China. Technological Sciences. 2023;66(7):2069-2079. http://doi.org/10.1007/s11431-023-2372-x.
8 Durodola JF. Machine learning for design, phase transformation and mechanical properties of alloys. Progress in Materials Science. 2022;123:100797. http://doi.org/10.1016/j.pmatsci.2021.100797.
9 Chen M, Zhang T, Gong Z, Zuo W, Wang Z, Zong L, et al. Mechanical properties and microstructure characteristics of wire arc additively manufactured high-strength steels. Engineering Structures. 2024;300:117092. http://doi. org/10.1016/j.engstruct.2023.117092. 10 Shi Z, Du L, He X, Gao X, Wu H, Liu Y, et al. Prediction model of yield strength of V–N steel hot-rolled plate based on machine learning algorithm. JOM. 2023;75(5):1750-1762. http://doi.org/10.1007/s11837-023-05773-y.
11 Altuğ M, Söyler H. Optimization with artificial intelligence of the machinability of Hardox steel, which is exposed to different processes. Scientific Reports. 2023;13(1):14100. http://doi.org/10.1038/s41598-023-40710-8.
12 Kong BO, Kim MS, Kim BH, Lee JH. Prediction of creep life using an explainable artificial intelligence technique and alloy design based on the genetic algorithm in creep-strength-enhanced ferritic 9% Cr steel. Metals and Materials International. 2023;29(5):1334-1345. http://doi.org/10.1007/s12540-022-01312-7.
13 Chakraborty S, Chattopadhyay P, Ghosh S, Datta S. Incorporation of prior knowledge in neural network model for continuous cooling of steel using genetic algorithm. Applied Soft Computing. 2017;58:297-306. http://doi. org/10.1016/j.asoc.2017.05.001.
14 Jozaghi T, Wang C, Arroyave R, Karaman I. Design of alumina-forming austenitic stainless steel using genetic algorithms. Materials & Design. 2020;186:108198. http://doi.org/10.1016/j.matdes.2019.108198.
15 Tsutsui K, Terasaki H, Maemura T, Hayashi K, Moriguchi K, Morito S. Morito. Microstructural diagram for steel based on crystallography with machine learning. Computational Materials Science. 2019;159:403-411. http://doi. org/10.1016/j.commatsci.2018.12.003.
16 Wang C, Shen C, Huo X, Zhang C, Xu W. Design of comprehensive mechanical properties by machine learning and high-throughput optimization algorithm in RAFM steels. Nuclear Engineering and Technology. 2020;52(5):1008- 1012. http://doi.org/10.1016/j.net.2019.10.014.
17 Herbst C. O que é AutoML e quais são as suas vantagens? [Internet]. 2022 [cited 2024 Jun 10]. Available at: https:// www.eldorado.org.br/en/blog/o-que-e-automl-e-quais-sao-as-suas-vantagens/
18 Chadha B, Juwe S. Agile machine learning with DataRobot: automate each step of the machine learning life cycle, from understanding problems to delivering value. Birmingham: Packt Publishing, Limited; 2021.
19 ASTM A572/A572M-21e1 (2021). Standard Specification for High-Strength Low-Alloy Columbium-Vanadium Structural Steel. ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959. United States, 4 p. Available from: https://www.astm.org/a0572_a0572m-21e01.html, doi: 10.1520/A0572_ A0572M-21E01
20 Graux A, Cazottes S, De Castro D, San Martín D, Capdevila C, Cabrera JM, et al. Precipitation and grain growth modelling in Ti-Nb microalloyed steels. Materialia. 2019;5:100233. http://doi.org/10.1016/j.mtla.2019.100233.
21 Wang F, Zheng X, Long J, Zheng K, Zheng Z. Effects of zirconium on the structure and mechanical properties of high-strength low-alloy steels under quenched or tempered conditions. Steel Research International. 2022;93(11):2200352. http://doi.org/10.1002/srin.202200352.
22 Moon J, Jang M-H, Kang J-Y, Lee T-H. The negative effect of Zr addition on the high temperature strength in alumina-forming austenitic stainless steels. Materials Characterization. 2014;87:12-18. http://doi.org/10.1016/j. matchar.2013.10.029.
23 Oliveira AP. Prediction model of mechanical properties of hot-rolled structural beams: an approach in artificial neural networks [thesis]. Belo Horizonte: Universidade Federal de Minas Gerais; 2008.
Submetido em:
17/10/2024
Aceito em:
18/11/2024
Facebook
Google+
Twitter
LinkedIn
Mendeley
StumbleUpon
CiteULike
Reddit
Email