Prediction of Sintering Flue Gas Based on TCN Convolutional Neural Network

DOI: https://doi.org/10.62517/jiem.202603109

Author(s)

Weimin Yin1, Jiayi Zhang1, Jiaming Xu2, Yanbo Hu1, Tonghui Zhao1, Junhu Wang1, Jie Li3,*

Affiliation(s)

1CHINA MCC22 GROUP CORPORATION LTD, Tangshan, China 2College of Science, North China University of Science and Technology, Tangshan, China 3College of Metallurgy and Energy, North China University of Science and Technology, Tangshan, China *Corresponding Author

Abstract

The CO concentration in the sintering flue gas during the sintering process is an important indicator reflecting the combustion state and permeability. Accurate short-term prediction of CO concentration helps in process stability and anomaly detection. However, in industrial environments, numerous sensor variables, significant data noise, and operational drift complicate high-dimensional modeling, making rapid deployment challenging. This paper proposes a lightweight short-term prediction framework based on causal dilated convolutions (TCN) for sintering flue gas data with a two-minute sampling interval to predict the target variable, cleaned_co. First, based on process mechanisms, relevant variables related to flue gas and exhaust are aggregated and selected, forming an input set with only 11 features. Next, multivariate sequences are constructed into supervised samples using a sliding window approach, employing causal convolutions to prevent future information leakage, and dilated convolutions to expand the temporal receptive field, thereby capturing the temporal dependencies between combustion and airflow. Experiments on real-world datasets validate the effectiveness of the proposed method. the results demonstrate that the model tracks the overall trend of CO concentration well within a 90-minute historical window, achieving stable prediction performance in the normal operating range. This method provides a concise and feasible technical approach for online modeling and rapid deployment of the sintering process, with fewer features.

Keywords

CO Concentration; Short-Term Forecasting; Industrial Sensor Variables; Causal Dilated Convolution (TCN)

References

[1] He, K., Li, H., Kang, J., et al. (2025). "Pilot-scale practice of catalytic oxidation for CO in sintering flue gas. " Sintering and Pelletizing, 50(05), 102-108. DOI: 10.13403/j. sjqt. 2025.05.084. [2] Sun, Q., Hao, X., Li, J., et al. (2025). "Mechanism of CO and NO synergistic control in the sintering process. " China Metallurgy, 35(10), 118-129. DOI: 10.13228/j. boyuan. issn1006-9356.20250285. [3] Wang, J., et al. CO reduction in sintering flue gas by CFD-ML for process parameters optimization. Journal of Cleaner Production, 2025. [4] Long, H., Ding, L., Zhao, H., et al. (2023). "Progress in CO emission reduction in typical steel production flue gases. " Steel, 58(08), 1-12+24. DOI: 10.13228/j. boyuan. issn0449-749x. 20230132. [5] Pei, Y., Liao, J., Zhang, J., et al. (2019). "Discussion on CO reduction in iron ore sintering process. " Sintering and Pelletizing, 44(01), 69-73. DOI: 10.13403/j. sjqt. 2019.01.016. [6] Li, Q., Li, G., Yin, G., et al. (2019). "Study and practice of CO emission reduction on a 550 m² sintering machine at Zhongtian Steel. " Sintering and Pelletizing, 44(04), 70-73. DOI: 10.13403/j. sjqt. 2019.04.065. [7] Li, J., Xu, R., Zhang, W., et al. (2025). "Current status and prospects of CO emission reduction in sintering flue gas. " China Metallurgy, 35(07), 109-120. DOI: 10.13228/j. boyuan. issn1006-9356.20250135. [8] Yang, C., et al. Prediction and optimization of flue pressure in sintering process based on machine learning algorithms., 2025. [9] Zhu, T. Y., et al. Numerical simulation of CO emission in a sintering pot under sintering flue gas recirculation (SFGR). Chemical Engineering Journal, 2023. [10] Guo, X., et al. Optimizing sintering air volume for enhanced lean gas (CO, H₂, CH₄) generation in sintering flue gas. Scientific Reports, 2025. [11] Sun, J., et al. Intensive carbon combustion in a sintering packed bed via surface steam spraying. Fuel, 2023. [12] Liu, S., et al. A Prediction System of Burn-through Point Based on GBDT Algorithm and Decision Rules. ISIJ International, 2019. [13] Chen, X., et al. Burn-through point prediction and control based on multi-period dynamic spatio-temporal extraction., 2025. [14] Li, X., et al. Prediction model of burn-through point with data correction based on feature matching and TCN-GRU., 2024. [15] Li, Y., et al. Sintering Quality Prediction Model Based on Semi-Supervised Dynamic Feature Extraction Framework. Sensors, 2022. [16] Li, Y., et al. A Soft Sensor Model of Sintering Process Quality Index Based on Multi-Source Data Fusion., 2023. [17] Xu, Z., et al. Burn-through point prediction based on deep learning feature selection and ensemble learning., 2023. [18] Yan, F., et al. Data-driven modelling methods in sintering process: current status and prospects. Canadian Journal of Chemical Engineering, 2023. [19] Hu, M., et al. Review of Intelligent Modeling for Sintering Process Under Variable Operating Conditions. Processes, 2025.