运用主成分分析和深度神经网络预测钢包炉合金元素收得率的研究

信自成; 张江山; 金玉; 郑瑾; 刘青

doi:10.1007/s12613-021-2409-9

Volume 30 Issue 2

Feb. 2023

Turn off MathJax

图(8) / 表(4)

数据统计

计量

文章访问数: 899
HTML全文浏览量: 323
PDF下载量: 88
被引次数: 0

文章导航 > 矿物冶金与材料学报（英文） > 2023 > 30(2): 335-344

Zicheng Xin, Jiangshan Zhang, Yu Jin, Jin Zheng, and Qing Liu, Predicting the alloying element yield in a ladle furnace using principal component analysis and deep neural network, Int. J. Miner. Metall. Mater., 30(2023), No. 2, pp. 335-344. https://doi.org/10.1007/s12613-021-2409-9

Cite this article as:

Zicheng Xin, Jiangshan Zhang, Yu Jin, Jin Zheng, and Qing Liu, Predicting the alloying element yield in a ladle furnace using principal component analysis and deep neural network, Int. J. Miner. Metall. Mater., 30(2023), No. 2, pp. 335-344. https://doi.org/10.1007/s12613-021-2409-9

Citation:

PDF( 1100 KB)

引用本文 PDF XML SpringerLink

研究论文

运用主成分分析和深度神经网络预测钢包炉合金元素收得率的研究

1)
北京科技大学钢铁冶金新技术国家重点实验室, 北京100083, 中国
2)
河北钢铁集团有限公司唐山分公司, 唐山063000, 中国

通讯作者:
张江山 E-mail: zjsustb@163.com

刘青 E-mail: qliu@ustb.edu.cn

文章亮点

(1) 运用主成分分析法对影响LF精炼过程合金元素收得率预测的数据进行处理，有效简化了实际生产数据结构。

(2) 融合SGDM算法和L2正则化方法，构建了适用于LF精炼过程合金元素收得率预测的深度神经网络。

(3) 采用主成分分析和深度神经网络相结合的方法，实现了LF精炼过程合金元素收得率的精准预测。

中文摘要

钢包炉（ladle furnace，LF）精炼的核心任务之一是进行钢水合金化过程，以保证钢包到达连铸平台后满足钢水成分要求。钢水合金化也是较为复杂的物理化学反应过程，其不仅对钢水温度有影响，而且对钢水质量、合金料消耗量均有影响。因此，实现LF精炼过程中合金元素收得率的精准预测具有重要意义。本研究首先采用主成分分析法（principal component analysis，PCA）对模型输入变量进行处理，用于简化实际生产数据结构；然后，融合SGDM算法和L2正则化方法，构建了深度神经网络（deep neural network，DNN）；最后，采用主成分分析和深度神经网络相结合的方法，建立了合金元素收得率预测模型。结果表明，Si元素收得率命中率在±1%, ±3%和 ±5%误差范围内分别为54.0%、93.8%和98.8%；Mn元素收得率命中率在±1%, ±2%和 ±3%误差范围内分别为77.0%、96.3%和99.5%。此外，PCA–DNN模型的预测精度优于参考炉次法、多元线性回归模型、改进BP神经网络模型和DNN模型，有助于实现钢水成分的“窄窗口”控制。本研究中合金元素收得率预测模型的建立，也可为LF智能精炼合金化控制模型的开发提供支撑。

关键词:

Research Article

Predicting the alloying element yield in a ladle furnace using principal component analysis and deep neural network

+ Author Affiliations

Abstract

Abstract

The composition control of molten steel is one of the main functions in the ladle furnace (LF) refining process. In this study, a feasible model was established to predict the alloying element yield using principal component analysis (PCA) and deep neural network (DNN). The PCA was used to eliminate collinearity and reduce the dimension of the input variables, and then the data processed by PCA were used to establish the DNN model. The prediction hit ratios for the Si element yield in the error ranges of ±1%, ±3%, and ±5% are 54.0%, 93.8%, and 98.8%, respectively, whereas those of the Mn element yield in the error ranges of ±1%, ±2%, and ±3% are 77.0%, 96.3%, and 99.5%, respectively, in the PCA–DNN model. The results demonstrate that the PCA–DNN model performs better than the known models, such as the reference heat method, multiple linear regression, modified backpropagation, and DNN model. Meanwhile, the accurate prediction of the alloying element yield can greatly contribute to realizing a “narrow window” control of composition in molten steel. The construction of the prediction model for the element yield can also provide a reference for the development of an alloying control model in LF intelligent refining in the modern iron and steel industry.
- ladle furnace;
- element yield;
- principal component analysis;
- deep neural network;
- statistical evaluation

FullText(HTML)

Supplements(0)

References(38)

References

[1]	J. Liu, Artificial intelligence drives changes in metallurgical industry, Iron Steel, 55(2020), No. 6, p. 1. doi: 10.13228/j.boyuan.issn0449-749x.20200191
[2]	J. Li, LF Refining Technology, Metallurgical Industry Press, Beijing, 2012.
[3]	P. Yu, D. P. Zhan, Z. H. Jiang, D. L. Li, X. D. Yin, and Z. G. Ma, Development of a terminal composition prediction model for steel refining with ladle furnace, J. Mater. Metall., 5(2006), No. 1, p. 20.
[4]	G.B. Li, C.L. Zhao, S.H. Zhao, L.J. Wang, and W.W. Zhang, Development of LF refining composition prediction model, Angang Technol., 2009(4), p. 26.
[5]	N.K. Nath, K. Mandal, A.K. Singh, B. Basu, C. Bhanu, S. Kumar, and A. Ghosh, Ladle furnace on-line reckoner for prediction and control of steel temperature and composition, Ironmaking Steelmaking, 33(2006), No. 2, p. 140. doi: 10.1179/174328106X80082
[6]	W.S. Cheng, S.G. Tang, Q.Z. Liu, and M.R. Fei, R&D of the ladle furnace mathematic model, [in] Proceedings of International Conference on Machine Learning and Cybernetics, Beijing, p. 566.
[7]	M. Seike, R. Sakao, H. Dei, H. Yamaguchi, T. Muroi, and S. Tsuda, Development of LFV guide control system using the expert system, CAMP-ISIJ, 7(1994), No. 5, p. 1260.
[8]	X.W. Gao, A.A. Zhang, and Q.L. Wei, Neural network based prediction of endpoint in ladle refining process, J. Northeast. Univ. Nat. Sci., 26(2005), No. 8, p. 726.
[9]	Z. Xu and Z.Z. Mao, Analysis and prediction of influencing factor on element recovery in ladle furnace, Iron Steel, 47(2012), No. 3, p. 34.
[10]	G.B. Huang, Z. Bai, L.L.C. Kasun, and C.M. Vong, Local receptive fields based extreme learning machine, IEEE Comput. Intell. Mag., 10(2015), No. 2, p. 18. doi: 10.1109/MCI.2015.2405316
[11]	V. N. Vapnik, The Nature of Statistical Learning Theory, Springer-Verlag, Berlin, 1995.
[12]	V. N. Vapnik, Statistical Learning Theory, John Wiley and Sons, New York, 1998.
[13]	L. Lin and J.Q. Zeng, Consideration of green intelligent steel processes and narrow window stability control technology on steel quality, Int. J. Miner. Metall. Mater., 28(2021), No. 8, p. 1264. doi: 10.1007/s12613-020-2246-2
[14]	S.H. Kwon, D.G. Hong, and C.H. Yim, Prediction of hot ductility of steels from elemental composition and thermal history by deep neural networks, Ironmaking Steelmaking, 47(2020), No. 10, p. 1176. doi: 10.1080/03019233.2019.1699358
[15]	J.P. Yang, J.S. Zhang, W.D. Guo, S. Gao, and Q. Liu, End-point temperature preset of molten steel in the final refining unit based on an integration of deep neural network and multi-process operation simulation, ISIJ Int., 61(2021), No. 7, p. 2100. doi: 10.2355/isijinternational.ISIJINT-2020-540
[16]	I. Mohanty, R. Banerjee, A. Santara, S. Kundu, and P. Mitra, Prediction of properties over the length of the coil during thermo-mechanical processing using DNN, Ironmaking Steelmaking, 48(2021), No. 8, p. 953. doi: 10.1080/03019233.2020.1848303
[17]	S.W. Wu, J. Yang, and G.M. Cao, Prediction of the Charpy V-Notch impact energy of low carbon steel using a shallow neural network and deep learning, Int. J. Miner. Metall. Mater., 28(2021), No. 8, p. 1309. doi: 10.1007/s12613-020-2168-z
[18]	Z.C. Xin, J.S. Zhang, J.G. Zhang, Y. Jin, J. Zheng, and Q. Liu, Mathematical modelling and plant trial on slagging regime in a ladle furnace for high-efficiency desulphurization, Ironmaking Steelmaking, 48(2021), No. 9, p. 1123. doi: 10.1080/03019233.2021.1935143
[19]	K. Pearson, Mathematical contributions to the theory of evolution. III. regression, heredity, and panmixia, Philos. Trans. R. Soc. London, Ser. A, 187, p. 253.
[20]	Z. Zhang, L.L. Cao, W.H. Lin, J.K. Sun, X.M. Feng, and Q. Liu, Improved prediction model for BOF end-point manganese content based on IPSO-RELM method, Chin. J. Eng., 41(2019), No. 8, p. 1052. doi: 10.13374/j.issn2095-9389.2019.08.011
[21]	K.X. Zhou, W.H. Lin, J.K. Sun, X.M. Feng, W. Fang, and Q. Liu, A prediction model to calculate Mn yield during BOF alloying process using improved extreme learning machine, J. Cent. South Univ. (Sci. Technol.), 52(2021), No. 5, p. 1399. doi: 10.11817/j.issn.1672-7207.2021.05.001
[22]	S. Valle, W.H. Li, and S.J. Qin, Selection of the number of principal components: The variance of the reconstruction error criterion with a comparison to other methods, Ind. Eng. Chem. Res., 38(1999), No. 11, p. 4389. doi: 10.1021/ie990110i
[23]	K. Wu, X.Z. Liu, X.X. Zhang, and Y. Miao, Feature extraction of hot strip rolling data based on PCA-DBN, Metall. Ind. Autom., 44(2020), No. 3, p. 21.
[24]	Y.L. Huang, Y.F. Liu, H. Huang, and B.L. Zheng, Prediction model of TPC reception iron amount based on PCA-GA-BP, Control Eng. China, 16(2009), No. 4, p. 446.
[25]	C. Chen, N. Wang, and M. Chen, Prediction model of end-point phosphorus content in consteel electric furnace based on PCA-extra tree model, ISIJ Int., 61(2021), No. 6, p. 1908. doi: 10.2355/isijinternational.ISIJINT-2020-615
[26]	Subagyo and G.A. Brooks, Online monitoring of dynamic slag behavior in ladle metallurgy, ISIJ Int., 43(2003), No. 8, p. 1286. doi: 10.2355/isijinternational.43.1286
[27]	G.E. Hinton and R.R. Salakhutdinov, Reducing the dimensionality of data with neural networks, Science, 313(2006), No. 5786, p. 504. doi: 10.1126/science.1127647
[28]	Z. H. Zhou, Machine Learning, Tsinghua University Press, Beijing, 2016.
[29]	Y. LeCun, Y. Bengio, and G. Hinton, Deep learning, Nature, 521(2015), No. 7553, p. 436. doi: 10.1038/nature14539
[30]	G.W. Song, B.A. Tama, J. Park, J.Y. Hwang, J. Bang, S.J. Park, and S. Lee, Temperature control optimization in a steel-making continuous casting process using a multimodal deep learning approach, Steel Res. Int., 90(2019), No. 12, art. No. 1900321. doi: 10.1002/srin.201900321
[31]	C.A. Myers and T. Nakagaki, Prediction of nucleation lag time from elemental composition and temperature for iron and steelmaking slags using deep neural networks, ISIJ Int., 59(2019), No. 4, p. 687. doi: 10.2355/isijinternational.ISIJINT-2018-338
[32]	S. Feng, H.Y. Zhou, and H.B. Dong, Using deep neural network with small dataset to predict material defects, Mater. Des., 162(2019), p. 300. doi: 10.1016/j.matdes.2018.11.060
[33]	M. Ranzato, F.J. Huang, Y.L. Boureau, and Y. LeCun, Unsupervised learning of invariant feature hierarchies with applications to object recognition, [in] 2007 IEEE Conference on Computer Vision and Pattern Recognition, Minneapolis, p. 1.
[34]	S.H. Wang, P. Phillips, Y.X. Sui, B. Liu, M. Yang, and H. Cheng, Classification of Alzheimer’s disease based on eight-layer convolutional neural network with leaky rectified linear unit and max pooling, J. Med. Syst., 42(2018), No. 85, art. No. 85(2018)
[35]	N. Qian, On the momentum term in gradient descent learning algorithms, Neural Netw., 12(1999), No. 1, p. 145. doi: 10.1016/S0893-6080(98)00116-6
[36]	I. Loshchilov and F. Hutter, Decoupled weight decay regularization, [in] 7th International Conference on Learning Representations (ICLR), New Orleans, 2019, p. 1.
[37]	M.H. Zhao, S.S. Zhong, X.Y. Fu, B.P. Tang, and M. Pecht, Deep residual shrinkage networks for fault diagnosis, IEEE Trans. Ind. Inf., 16(2020), No. 7, p. 4681. doi: 10.1109/TII.2019.2943898
[38]	S. Samarasinghe, Neural Networks for Applied Sciences and Engineering, Auerbach Publications, New York, 2006.