Research on Sintering Burn-Through Point Prediction Based on Multi Optimization Algorithms and Deep Learning_Vol. 3 No. 4 (JIEM 2025)_Journal of Industry and Engineering Management (ISSN: 2959-0612)

Home > Journal of Industry and Engineering Management (ISSN: 2959-0612) > Vol. 3 No. 4 (JIEM 2025) >

Research on Sintering Burn-Through Point Prediction Based on Multi Optimization Algorithms and Deep Learning

DOI: https://doi.org/10.62517/jiem.202503408

Author(s)

Zhen Dong1, Menghui Bai1, Haifeng Song2

Affiliation(s)

1College of Artificial Intelligence, North China University of Science and Technology, Tangshan, Hebei, China 2College of Science, North China University of Science and Technology, Tangshan, Hebei, China

Abstract

With the collaborative evolution of computing power and algorithms, machine learning has become the core supporting technology for solving complex problems in metallurgical engineering. Burn-Through Point (BTP) soft sensing faces typical big data challenges such as strong coupling of multiple variables, non-stationary across time scales, and high-frequency noise. This paper proposes a prediction model that integrates Osprey Optimization Algorithm (OOA), Time varying Filter Empirical Mode Decomposition (TVF-EMD), Dwarf Mongoose Optimiza-tion (DMO), and Gated Recurrent Unit (GRU). Firstly, perform missing value completion and Min Max normalization on industrial measured data sampled at 10 months and 30 minutes; Subsequently, with the minimization of sample entropy as the fitness, OOA is used to globally optimize the bandwidth threshold and B-spline order of TVFEMD, achieving efficient de-noising and multi-scale decomposition. All IMFs are reorganized into high, medium, and low frequency collaborative mode functions (Co-IMF) through K-Means clustering based on sample entropy; Further use VMD deep narrowband reconstruction to enhance key intermediate frequency features for Co-IMF0, which has the highest energy proportion. In the prediction stage, DMO adaptively searches for hyperparameters such as batch_size and epoch of GRU to form a DMO-GRU predictor, and outputs weighted superposition of the three sub signals. Compared with seven baseline models including TVFEMD-GRU and VMD-PSO-LSTM, this method reduces the RMSE from 2.86 to 1.89 and improves the R2 to 0.990 on the test set, demonstrating robustness to high-frequency noise, in-sensitivity to local optima of hyperparameters, and ac-curate capture of key modal information. the research provides a replicable new paradigm for green, low-carbon, and intelligent manufacturing in the sinter-ing process of the steel industry.

Keywords

Burn-Through Point; Multiscale Decom-Posetion; OOA-TVFEMD; DMO-GRU

References

[1] Jianlin Lin, DeJin Kong. Study on “Industrial Internet” and “Industrie 4.0”[J]. Strategic Study of CAE, 2015, 17(07):141-144. [2] Kesha Guo. the Trend of Industrial Structure Adjustment and Upgrading in China and the Thoughts of Policy Adjustment during the “14th Five-Year Plan”[J]. China Industrial Economics, 2019, (07):24-41. [3] KoŜtial I, Dorĉák E, Terpak J. Optimal control of the sintering process [J]. IFAC Proceedings Volumes, 2005, 38(1):280-285. [4] Fumin LI, Jucai Hou, Xiaojie Liu, et al. Development status and prospect of smart manufacturing of sintering system [J]. Sintering and Pelletizing, 2024, 49(01):27-34. [5] Xuefeng Zhang, Jingjing Tang, Liusong Huang, et al. Research progress of intelligent sintering prediction technology based on big data [J]. Metallurgical Industry Automation, 2024, 48(04):64-76. [6] Shanwen Xu. End-point Control Method and System Research of Sintering Process [J]. Metallurgical Power, 2023, (03):99-102+107. [7] Bing Xu, Shilai Huang, Yanyan Yang, et al. Development and Application of a Balanced Material Flow Control Model for Sintering [J]. Hebei Metallurgy, 2021, (10):54-56+84. [8] Xu Jiang. Application of intelligent analysis by big data in blast furnace ironmaking [J]. Modern Transportation and Metallurgical Materials, 2023, 3(01):83-88. [9] Lingkun Chen. Analysis of intelligent control system used for high efficiency smelting in blast furnace [J]. Metallurgical Industry Automation, 2021, 45(03):2-10. [10] Patisson F, Bellot J P, Ablitzer D. Study of moisture transfer during the strand sintering process [J]. Metallurgical Transactions B, 1990, 21(1):37-47. [11] Ren Y, Huang C, Jiang Y, et al. Neural network prediction model for sinter mixture water content based on KPCA-GA optimization [J]. Metals, 2022, 12(8):1287. [12] Rahman R, Otridge J, Pal R. IntegratedMRF: random forest-based framework for integrating prediction from different data types [J]. Bioinformatics, 2017, 33(9):1407-1410. [13] Jie Hu, Sheng Du, Min Wu, et al. Advanced Control and Intelligent Optimization of Raw-materials Preparation Process in Ironmaking [J]. Information and Control, 2018, 47(04):411-420. [14] Fubin Wang, Rui Wang, Chen Wu. Short-Term Prediction of Sintering State Based on Improved Random Forest [J]. Laser & Optoelectronics Progress, 2022, 59(18):382-388. [15] Wang S H, Li H F, Zhang Y J, et al. A hybrid ensemble model based on ELM and improved AdaBoost. RT algorithm for predicting the iron ore sintering characters [J]. Computational intelligence and neuroscience, 2019, 2019(1):4164296. [16] Wei Chen, Pengfei Wu, Baoxiang Wang, et al. Intelligent optimization system of burden structure in the whole process of sintering-blast furnace iron-making [J]. Sintering and Pelletizing, 2020, 45(05):8-13+24. [17] Tao Liu. Improving the quality of sintered ore under variable raw material conditions [J]. Hebei Metallurgy, 2022(007):000. DOI:10.13630/j. cnki. 13-1172.2022.0706. [18] Zengchao Tian, Wenlong Zhan, Jiangya Kang, et al. Development of DNN-based sintering endpoint prediction and intelligent adjustment system [J]. Sintering and Pelletizing, 2023, 48(06):101-108. [19] Liu S, Liu X, Lyu Q, et al. Comprehensive system based on a DNN and LSTM for predicting sinter composition [J]. Applied Soft Computing, 2020, 95:106574. [20] Jinyang Wang, Zhaoxia Wu, Zhongzheng Li, et al. Prediction Model of Burning Through Point Based on JITL-XGBoost [J]. Journal of Northeastern University(Natural Science), 2025, 46(02):28-34+41. [21] Yueyang Luo, Bocun He, Xinmin Zhang, et al. Research on multi-step prediction, method for sintering burn-through point prediction based on spatio-temporal encoding task decoding [J]. Metallurgical Industry Automation, 2025, 49(02):43-52. [22] Zhen Zhang, Xiaojie Liu, Song Liu, et al. Research on combined forecasting model of return fine amount based on deep learning [J]. Sintering and Pelletizing, 2023, 48(01):57-66. DOI:10.13403/j. sjqt. 2023.01.009. [23] Yuan Ding, Bin Wang, Jinchong Yan, et al. Prediction of burning through point based on two-dimensional interval autoregressive model [J]. Sintering and Pelletizing, 2017, 42(03):1-6+15. [24] Cao W, Zhang Y, She J, et al. A dynamic subspace model for predicting burn-through point in iron sintering process [J]. Information Sciences, 2018, 466:1-12. [25] Yan F, Zhang X, Yang C, et al. Data‐driven modelling methods in sintering process: Current research status and perspectives [J]. the Canadian Journal of Chemical Engineering, 2023, 101(8):4506-4522. [26] Xinpeng Qu, Chunjie Yang, Jinsong Meng, et al. Digital twin driven sintering endpoint prediction method [J]. Metallurgical Industry Automation, 2024, 48(S1):12-17. [27] Chen X, Lan T, Shi X, et al. A semi-supervised linear–nonlinear least-square learning network for prediction of carbon efficiency in iron ore sintering process [J]. Control Engineering Practice, 2020, 100:104454. [28] Dehghani M, Trojovský P. Osprey optimization algorithm: A new bio-inspired metaheuristic algorithm for solving engineering optimization problems [J]. Frontiers in Mechanical Engineering, 2023, 8:1126450. [29] Liu, B., Cai, G., Qian, J., Song, T., & Zhang, Q. (2023). Machine Learning Model Training and Practice: A Study on Constructing a Novel Drug Detection System. International Journal of Computer Science and Information Technology, 1(1), 139-146.