Explainable MRF-BBAPM with self-learning for predicting compressive strength of oxidized pellet

Abstract

The compressive strength of oxidized pellets is a key indicator for evaluating pellet quality and stability. Accurate prediction of its variation trend is essential for improving production efficiency and optimizing process parameters. However, due to the high dimensionality and strong nonlinearity of compressive strength prediction, existing models still face limitations in terms of reliability, applicability, and generalization. This study proposes the metallurgical-random forest-based Bayesian optimized bidirectional gated recurrent unit (BiGRU) attention prediction model (MRF-BBAPM) model, which employs feature selection guided by metallurgical mechanisms and random forest to enhance model efficiency and relevance. The BiGRU network parameters are optimized using Bayesian optimization, and an attention mechanism is incorporated to focus on critical features, further improving model performance. The SHapley Additive exPlanations (SHAP) method is introduced to quantify the contribution of each feature to the prediction results, revealing the model’s decision-making process and enhancing its interpretability and reliability. The model also incorporates a self-learning mechanism that automatically updates and optimizes itself based on weekly prediction errors. Experimental results show that the proposed model achieves a mean absolute error of 80.58 N (2.77% of the mean) and a root mean square error of 95.75 N (3.29% of the mean) in predicting pellet compressive strength, demonstrating strong stability and reliability in real-world applications. This method provides effective data support for accurate prediction of pellet compressive strength and informed decision-making in production.

Keywords

compressive strength of oxidized pellet / metallurgical mechanisms / random forests / Bayesian optimization / bidirectional gated recurrent unit / SHapley Additive exPlanations

References

[1] X.H. Zhang, N. Wang, Z.Y. Zhou, Z.Y. Ning, and M. Chen, Effect of combined ferrous burden composition and ore–coke interaction on blast furnace burden distribution, Powder Technol., 449(2025), art. No. 120416.

[2] Zhao ZC, Tang J, Chu MS, et al.. Direct reduction swelling behavior of pellets in hydrogen-based shaft furnaces under typical atmospheres. Int. J. Miner. Metall. Mater., 2022, 29(10): 1891.

[3] Y. Shi, D.Q. Zhu, J. Pan, Z.Q. Guo, S.H. Lu, and M.J. Xu, Improving hydrogen-rich gas-based shaft furnace direct reduction of fired hematite pellets by modifying basicity, Powder Technol., 408(2022), art. No. 117782.

[4] Wang C, Xu CY, Liu ZJ, Wang YZ, Wang RR, Ma LM. Effect of organic binders on the activation and properties of indurated magnetite pellets. Int. J. Miner. Metall. Mater., 2021, 28(7): 1145.

[5] Y.Z. Wang, Y. Xu, X.R. Song, Q.K. Sun, J.L. Zhang, and Z.J. Liu, Novel method for temperature prediction in rotary kiln process through machine learning and CFD, Powder Technol., 439(2024), art. No. 119649.

[6] Zhao ZJ, Saxén H, Wang SH, et al.. Simulation study on the effect of increasing pellet proportion on burden distribution in the blast furnace. Ironmaking Steelmaking, 2024, 51(1): 23.

[7] Jiang TY, Xue T, Li ZZ, Yang AM, Li J, Zhang ZQ. Application of intelligent algorithms in pellet manufacturing process. China Metall., 2022, 32(5): 25

[8] Shi Q, Tang J, Chu MS. Key issues and progress of industrial big data-based intelligent blast furnace ironmaking technology. Int. J. Miner. Metall. Mater., 2023, 30(9): 1651.

[9] Shi Q, Tang J, Chu MS. Process metallurgy and data-driven prediction and feedback of blast furnace heat indicators. Int. J. Miner. Metall. Mater., 2024, 31(6): 1228.

[10] Q. Shi, J. Tang, and M.S. Chu, Evaluation, prediction, and feedback of blast furnace hearth activity based on data-driven analysis and process metallurgy, Steel Res. Int., 95(2024), No. 2, art. No. 2300385.

[11] Xin ZC, Zhang JS, Peng KX, et al.. Explainable machine learning model for predicting molten steel temperature in the LF refining process. Int. J. Miner. Metall. Mater., 2024, 31(12): 2657.

[12] Liu WX, Bai YJ, Zhang C, Wang ZJ, Yang AM, Wu MY. PSO-DFNN: A particle swarm optimization enabled deep fuzzy neural network for predicting the pellet strength. Alex. Eng. J., 2024, 106: 505.

[13] Z.H. Xu, Z.J. Wang, X.W. Qi, B. Bai, and J.M. Zhi, Prediction of green properties of flux pellets based on improved generalized regression neural network, Metals, 12(2022), No. 11, art. No. 1840.

[14] Dwarapudi S, Gupta PK, Rao SM. Prediction of iron ore pellet strength using artificial neural network model. ISIJ Int., 2007, 47(1): 67.

[15] H.L. Yan, X.L. Zhou, L. Gao, et al., Prediction of compressive strength of biomass-humic acid limonite pellets using artificial neural network model, Materials, 16(2023), No. 14, art. No. 5184.

[16] Yang AM, Zhuansun YX, Shi Y, Liu HX, Chen YJ, Li RS. IoT system for pellet proportioning based on BAS intelligent recommendation model. IEEE Trans. Ind. Inform., 2021, 17(2): 934.

[17] Liu WX, Li J, Yang AM, Zhang YZ, Xin ZC. Influence mechanism of basicity on strength of magnesium fluxed pellets. Iron Steel, 2021, 56(1): 28

[18] Shen FM, Gao QJ, Jiang X, Wei G, Zheng HY. Effect of magnesia on the compressive strength of pellets. Int. J. Miner. Metall. Mater., 2014, 21(5): 431.

[19] H. Guo, X. Jiang, F.M. Shen, H.Y. Zheng, Q.J. Gao, and X. Zhang, Influence of SiO₂ on the compressive strength and reduction-melting of pellets, Metals, 9(2019), No. 8, art. No. 852.

[20] Z.Z. Li, Y.F. Li, Y.S. Duan, A.M. Yang, Z.H. Xu, and J.M. Zhi, Study on the basic characteristics of iron ore powder with different particle sizes, Minerals, 12(2022), No. 8, art. No. 973.

[21] Zhang JL, Wang ZY, Xing XD, Liu ZJ. Effect of aluminum oxide on the compressive strength of pellets. Int. J. Miner. Metall. Mater., 2014, 21(4): 339.

[22] Wang XL, Jin YC, Schmitt S, Olhofer M. Recent advances in Bayesian optimization. ACM Comput. Surv., 2023, 55(13s): 1

[23] Y.Q. Gui, D.W. Zhan, and T.R. Li, Taking another step: A simple approach to high-dimensional Bayesian optimization, Inf. Sci., 679(2024), art. No. 121056.

[24] Zhang HT, Fu HD, Shen YH, Xie JX. Rapid design of secondary deformation-aging parameters for ultra-low Co content Cu–Ni–Co–Si–X alloy via Bayesian optimization machine learning. Int. J. Miner. Metall. Mater., 2022, 29(6): 1197.

[25] Z.Y. Jiang, Q.M. Tan, N. Li, J.X. Che, and X.K. Tan, A novel BiGRU multi-step wind power forecasting approach based on multi-label integration random forest feature selection and neural network clustering, Energy Convers. Manage., 319(2024), art. No. 118904.

[26] C.L. Huang, T. Zhou, W.D. Li, H.J. Yu, R.X. Li, and J.J. Fang, A coupled model integrating dual attention mechanism into BiGRU-RED for multi-step-ahead streamflow forecasting, J. Hydrol., 645(2024), art. No. 132137.

[27] S. Mangalathu, S.H. Hwang, and J.S. Jeon, Failure mode and effects analysis of RC members based on machine-learning-based SHapley Additive exPlanations (SHAP) approach, Eng. Struct., 219(2020), art. No. 110927.

[28] J.Y. Zhang, X.L. Ma, J.L. Zhang, et al., Insights into geospatial heterogeneity of landslide susceptibility based on the SHAPXGBoost model, J. Environ. Manage., 332(2023), art. No. 117357.

[29] X.H. Hu, S.Z. Chen, J.H. Zhao, R. Wang, and W. Liu, Risk identification and prediction model for continuous-lane-change vehicles considering driving style, Expert Syst. Appl., 259(2025), art. No. 125292.

[30] Z.H. Wang, K.W. Luo, H.S. Yu, K. Feng, and H. Ding, NO_x Emission prediction of heavy-duty diesel vehicles based on Bayesian optimization-gated recurrent unit algorithm, Energy, 292(2024), art. No. 130559.

RIGHTS & PERMISSIONS

University of Science and Technology Beijing