LS-SVM Blast Furnace Molten Iron Temperature Prediction based on Bayesian Optimization

Siyuan Wu; Nielei Yao; Jingyi Jia

doi:10.6919/ICJE.202505_11(5).0032

Authors

Siyuan Wu
Nielei Yao
Jingyi Jia

DOI:

https://doi.org/10.6919/ICJE.202505_11(5).0032

Keywords:

Iron Temperature; CatBoost; Least Squares Support Vector Machine; Bayesian Optimization.

Abstract

Blast furnace molten iron temperature prediction is a complex nonlinear problem, which is characterized by data uncertainty, nonlinearity and high dimensionality. In this paper, taking the molten iron temperature as the research object, the original data obtained are firstly normalized to exclude the influence of abnormal data on the prediction of blast furnace molten iron temperature, and then all the feature parameters affecting the molten iron temperature are feature-selected by using CatBoost algorithm and SHAP analysis to obtain the correlation feature parameters that significantly affect the molten iron temperature. On this basis, the LS-SVM iron temperature prediction model based on Bayesian optimization is established to predict the blast furnace iron temperature. The results show that the model can realize effective prediction of molten iron temperature, and the prediction accuracy is as high as 90.16%, which has better prediction effect of blast furnace molten iron temperature.

Downloads

Download data is not yet available.

References

[1] Adjustment and upgrading plan for the iron and steel industry (2016-2020)[J]. China Steel Industry, 2016, (12): 5-13.

[2] FENG Ting, LIU Xiangguan. Variable coefficient regression model for blast furnace temperature prediction and control[J]. Journal of Zhejiang University: Engineering Edition, 2007,41(10):1743.

[3] Zeng Jiusun, Liu Xiangguan. Application of principal component regression and partial least squares in blast furnace smelting[J]. Journal of Zhejiang University: Science Edition,2009,36(1):33.

[4] LI Qihui, LIU Xiangguan. Fuzzy prediction function control of silicon content in blast furnace molten iron[J]. Journal of Zhejiang University: Science Edition, 2007, 39(3): 104.

[5] LIN Jia-Taki,ZHOU Yang-Zhong,CHEN Dong-Yuan,et al. Inverse system control of radial magnetic levitation of a bearingless flux-switching motor rotor based on least-squares support vector machine decoupling[J/OL]. Journal of Electrotechnology,1-13[2024-12-27].

[6] Mao Yiming,Cheng Yanyan. Sparrow search algorithm to improve LSSVM for network intrusion detection[J]. Journal of Taiyuan College (Natural Science Edition),2024,42(04):64-69.

[7] ZHU Xinyu,XU Siyuan,XIAO Hongliang,et al. Adaptive cyclic engine performance optimization control based on Bayesian optimization[J/OL]. Journal of Aerospace Dynamics,1-13[2024-12-27].

[8] TAN Yong,CHEN Ji,YANG Zhongmin,et al. Slope stability prediction method based on CatBoost integrated learning[J]. Science, Technology and Engineering,2024,24(30):13153-13160.