NGBoost algorithm-based prediction of mechanical properties of a hot-rolled strip and its interpretability research with ANOVA values

Hongyi Wu; Jinwen Jin; Zhiwei Li; Hongyi Wu; Jinwen Jin; Zhiwei Li

doi:10.3934/math.20241578

AIMS Mathematics

2024, Volume 9, Issue 11: 33000-33022. doi: 10.3934/math.20241578

Previous Article Next Article

Research article Special Issues

NGBoost algorithm-based prediction of mechanical properties of a hot-rolled strip and its interpretability research with ANOVA values

Institute of New Iron and Steel Materials, Ningbo Iron & Steel Co., Ltd, Zhejiang, 315800, China

Received: 23 August 2024 Revised: 11 November 2024 Accepted: 12 November 2024 Published: 21 November 2024
MSC : 62F07, 62J05

Hot-rolled strip steel is an essential material extensively used in various industrial fields, with its mechanical properties being critical to product quality and engineering design. This article presents a method for predicting the mechanical properties of hot-rolled strip steel using the NGBoost (natural gradient boosting) algorithm. The study focused on predicting tensile strength, yield strength, and elongation of hot-rolled strip steel and compared the predictive results with those obtained from the gradient boosting algorithm, Lasso regression, and decision tree algorithms. The results indicated that the NGBoost algorithm performs well on average coverage error (ACE) and prediction interval absolute width (PIAW) values at different confidence levels, demonstrating strong predictive performance. Furthermore, the analysis of variance (ANOVA) method was employed to identify factors that significantly impact mechanical performance, providing theoretical support for optimizing design schemes and enhancing structural safety and reliability.
- hot-rolled strip steel,
- mechanical properties,
- NGBoost algorithm,
- ANOVA value
Citation: Hongyi Wu, Jinwen Jin, Zhiwei Li. NGBoost algorithm-based prediction of mechanical properties of a hot-rolled strip and its interpretability research with ANOVA values[J]. AIMS Mathematics, 2024, 9(11): 33000-33022. doi: 10.3934/math.20241578

Related Papers:

Abstract

Hot-rolled strip steel is an essential material extensively used in various industrial fields, with its mechanical properties being critical to product quality and engineering design. This article presents a method for predicting the mechanical properties of hot-rolled strip steel using the NGBoost (natural gradient boosting) algorithm. The study focused on predicting tensile strength, yield strength, and elongation of hot-rolled strip steel and compared the predictive results with those obtained from the gradient boosting algorithm, Lasso regression, and decision tree algorithms. The results indicated that the NGBoost algorithm performs well on average coverage error (ACE) and prediction interval absolute width (PIAW) values at different confidence levels, demonstrating strong predictive performance. Furthermore, the analysis of variance (ANOVA) method was employed to identify factors that significantly impact mechanical performance, providing theoretical support for optimizing design schemes and enhancing structural safety and reliability.

References

[1]	X. L. Wang, J. D. Chi, Intelligent manufacturing promotes the transformation and upgrading of the iron and steel industry, Metall. Ind. Autom., 42 (2018), 1–5.
[2]	H. Yada, Prediction of microstructural changes and mechanical properties in hot strip rolling, In: Proceedings of the Metallurgical Society of the Canadian Institute of Mining & Metallurgy, 1988,105–119. https://doi.org/10.1016/B978-0-08-035770-6.50012-X
[3]	M. Suehiro, K. Sato, Y. Tsukano, H. Yada, T. Senuma, Y. Matsumura, Computer modeling of microstructural change and strength of low carbon steel in hot strip rolling, Transact. ISIJ, 29 (1987), 439–445. https://doi.org/10.2355/isijinternational1966.27.439 doi: 10.2355/isijinternational1966.27.439
[4]	T. Senuma, M. Suehiro, H. Yada, Mathematical models for predicting microstructural evolution and mechanical properties of hot strips, ISIJ Int., 32 (1992), 423–432. https://doi.org/10.2355/isijinternational.32.423 doi: 10.2355/isijinternational.32.423
[5]	P. D. Hodgson, R. K. Gibb, A mathematical model to predict the mechanical properties of hot rolled C-Mn and microalloyed steels, ISIJ Int., 32 (1992), 1329–1338. https://doi.org/10.2355/isijinternational.32.1329 doi: 10.2355/isijinternational.32.1329
[6]	Y. B. Xu, X. H. Liu, G. D. Wang, Prediction-control model for microstructure and property of hot-rolled steel strip and their application, J. Iron Steel Res., 14 (2002), 65–68. https://doi.org/10.1080/09503150208411543 doi: 10.1080/09503150208411543
[7]	Y. Gan, The R & D of process modeling in thin slab hot strip rolling, Iron Steel, 38 (2003), 10–15.
[8]	J. Guo, R. Jia, R. Su, Y. Zhao, Identification of FIR systems with binary-valued observations against data tampering attacks, IEEE T. Syst. Man Cy.-S., 53 (2023), 5861–5873. https://doi.org/10.1109/TSMC.2023.3276352 doi: 10.1109/TSMC.2023.3276352
[9]	J. Guo, X. Wang, W. Xue, Y. Zhao, System identification with binary-valued observations under data tampering attacks, IEEE T. Automat. Contr., 66 (2021), 3825–3832. https://doi.org/10.1109/TAC.2020.3029325 doi: 10.1109/TAC.2020.3029325
[10]	S. Y. Huang, Prediction of tensile strength of hot rolled strips based on deep belief network, Wuhan Univ. Sci. Technol., 2022.
[11]	Y. Song, B. Li, C. Liu, F. F. Li, Prediction model of mechanical properties of hot rolled strip based on improved stacked self-encoder, Metall. Ind. Autom., 44 (2020).
[12]	F. Zhao, Research on prediction of mechanical properties for hot rolled strip based on big data and XGBoost, Metall. Autom. Design Res. Inst., 2020.
[13]	X. Zhang, Prediction of mechanical properties of hot rolled strips based on support vector quantile regression, Wuhan Univ. Sci. Technol., 2020.
[14]	W. Yang, W. Li, Y. Zhao, B. K. Yan, W. B. Wang, Mechanical property prediction of steel and influence factors selection based on random forests, Iron Steel, 53 (2018), 44–49.
[15]	E. Díaz, G. Spagnoli, Natural gradient boosting for probabilistic prediction of soaked CBR values using an explainable artificial intelligence approach, Buildings, 14 (2024), 352. https://doi.org/10.3390/buildings14020352 doi: 10.3390/buildings14020352
[16]	S. Z. Chen, D. C. Feng, W. J. Wang, E. Taciroglu, Probabilistic machine-learning methods for performance prediction of structure and infrastructures through natural gradient boosting, J. Struct. Eng., 148 (2022), 04022096.
[17]	X. R. Ma, X. L. Wang, S. Z. Chen, Trustworthy machine learning-enhanced 3D concrete printing: Predicting bond strength and designing reinforcement embedment length, Automat. Constr., 168 (2024), 105754. https://doi.org/10.1016/j.autcon.2024.105754 doi: 10.1016/j.autcon.2024.105754
[18]	O. Mamun, M. F. N. Taufique, M. Wenzlick, J. Hawk, R. Devanathan, Uncertainty quantification for Bayesian active learning in rupture life prediction of ferritic steels, Sci. Rep., 2083 (2022), 12. https://doi.org/10.1038/s41598-022-06051-8 doi: 10.1038/s41598-022-06051-8
[19]	J. Y. Ding, D. C. Feng, E. Brunesi, F. Parisi, G. Wu, Efficient seismic fragility analysis method utilizing ground motion clustering and probabilistic machine learning, Eng. Struct., 294 (2023), 116739. https://doi.org/10.1016/j.engstruct.2023.116739 doi: 10.1016/j.engstruct.2023.116739
[20]	T. Duan, A. Anand, D. Y. Ding, K. K. Thai, S. Basu, A. Ng, et al., NGBoost: Natural gradient boosting for probabilistic prediction, In: Proceedings of 37th International Conference on Machine Learning(ICML), 2020, 2690–2700.
[21]	B. S. Li, C. J. Pang, D. C. Cheng, Interpretable wind power probabilistic prediction based on NGBoost, Zhejiang Elec. Power, 42 (2023), 28–36.
[22]	Statistics Group, Institute of Mathematics, Chinese Academy of Sciences, Variance Analysis, Science Press, 1977.
[23]	H. C. Shen, Y. Zhao, M. Lang, Analysis on the application of variance analysis in the detection of moisture in iron ore, China Sci. Technol. Period. Database Ind. A, 2023, 20–22.
[24]	J. Kang, S. X. Zhang, Q. P. Zhang, B. Gao, Z. H. Yan, Fault diagnosis method of transformer based on ANOVA and BO-SVM, High Volt. Eng., 49 (2023), 1882–1891.
[25]	F. Li, Y. Song, C. Liu, R. Jia, B. Li, Research on error distribution modeling of mechanical performance prediction model for hot rolled strip, Metall. Ind. Autom., 43 (2019), 28–33.
[26]	C. Wan, Y. H. Song, Theories, methodologies and applications of probabilistic forecasting for power systems with renewable energy sources, Autom. Electr. Power Syst., 45 (2021), 2–16.

Reader Comments

Your name:*

Email:*
© 2024 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)