Sintered CO Emission Prediction Based on CSSM-GBRT

DOI: https://doi.org/10.62517/jike.202604138

Author(s)

Hu Yanbo1, Yin Weiming1, Wang Yuyang2, Zhang Jiayi1, Zhao Tonghui1, Yang Aimin2,3*

Affiliation(s)

1China 22nd Metallurgical Group Co., Ltd., Tangshan, China 2College of Science, North China University of Science and Technology, Tangshan, China 3Hebei Province Engineering Research Center for Iron Ore Selection and Intelligent Pre-Iron Processes, Tangshan, China *Corresponding Author

Abstract

As the core process for CO emissions in the steel industry, accurately predicting the CO generation trend during the sintering process is the key prerequisite for understanding its formation mechanism and grasping emission patterns. It is also an essential foundation for optimizing sintering process parameters and promoting related technological innovations and upgrades. To address the challenge of precise prediction caused by the nonlinearity, lag, and volatility of CO emissions in steel sintering, a CSSM-GBRT hybrid model is proposed. This model uses LSTM to deeply extract long-term dependent features from multidimensional sequences such as wind box temperature, oxygen content, and quicklime flow, then applies GBRT for secondary correction of the LSTM residuals, realizing a dual mechanism- and data-driven approach. Experimental results demonstrate that CSSM-GBRT achieves a coefficient of determination R² of 0.9514, reduces RMSE to 98.52 ppm, and reaches a MAPE of only 0.84%. SHAP analysis indicates that cumulative compressed air volume, O₂ concentration in the main flue, and the temperature on the northern side of the wind box are the key factors influencing CO emissions. By integrating key sintering process parameters with emission data, this model can deeply elucidate the intrinsic relationships between solid fuel combustion efficiency, high-temperature reduction reaction intensity, and CO generation, enabling precise prediction of emission trends at different stages, such as low-temperature, low-oxygen high-CO and high-temperature, high-oxygen low-CO conditions, thereby providing technical support for source control and efficient management of CO emissions in the steel industry.

Keywords

Steel Sintering; CO Emission Prediction; Temporal Dynamic Changes; Nonlinear Correlation; Process Parameter Optimization

References

[1] WEN X, HAN D. Global Warming Accelerates: 2024 Projectedto Become the Hottest Year on Record Again [J]. Ecological Economy, 2025, 41(02):1-4. [2] ZHANG L N. Application and development research of low carbon technology in iron and steel industry [J]. Energy for Metallurgical Industry, 2023, 42(2):3. [3] LIU C, YIN P, CHEN R J, et al. Ambient carbon monoxide and cardiovascular mortality: A nationwide time-series analy-sis in 272 cities in China [J]. Lancet Planetary Health, 2018, 2(1):e12. [4] LIAO J Y, TAN X L, ZHANG Z, et al. Research on Pre-cooling Technology and Equipment of Sinter [J]. Journal of Wuhan University of Science and Technology, 2024, 47(03):174-181. [5] SHAO Y J, XU L, LIU X P, et al. Discussion on solution of Carbon neutrality in China's steelp roduction [J]. China Metallurgy, 2022, 32(4):1. [6] ZHANG F M. Development prospects and meth-ods on low-carbon technology in ironmaking system [J]. Iron and Steel, 2022, 57(9):11. [7] FAN X, YU Z, GAN M, et al. Appropriate technology pa-rameters of iron ore sintering process with flue gas recircula-tion [J]. ISIJ International, 2014, 54(11):2541. [8] Xie, Ty., Zhang, Jg., Li, Lj. et al. Attention-Based Convolutional Aggregation: An Efficient Model for Off-Gas Profile Forecasting and Dynamic Pre-Control of BOF Steelmaking. Int J Comput Intell Syst 17, 312(2024). [9] Zheng R, Bao Y, Zhao L, Xing L. Prediction of steelmaking process variables using K-medoids and a time-aware LSTM network. Heliyon. 2024 Jun 14; 10(12):e32901. [10] Shengwei L, Zeping T, Haroon M. Estimation of transport CO2 emissions using machine learning algorithm [J]. Transportation Research Part D: Transport and Environment, 2024, 133:104276. [11] ZHONG Wei. Lithium-Ion Battery Capacity Prediction Based on LSTM Models [J]. Automobile and New Powertrain, 2025, 8(01):6-10. [12] LIU C S, QU J S, GE Y J, et al. LSTM model-based prediction of carbon emissions from China's transportation sector [J]. China Environmental Science, 2023, 43(05):2574-2582. [13] LIU W Q, CHENG Y, LI J, etal. Effect of particle size of coke powder on CO emission in sintering flue gas [J]. Journal of Ironand Steel Research, 2020, 32(7):633. [14] LONG H M, DING L, TAO J J, et al. Analysis on resource utilization of spent V2O5-WO3/TiO2 catalyst produced in sinteringflue gas [J]. IronandSteel, 2022, 57(7):162. [15] LONG H M, DING L, ZHAO H X, et al. Research progress of CO removal in flue gas of typical steel production process [J]. Iron and Steel, 2023, 58(8):1. [16] ZHOU H, MA P N, LAI Z Y, et al. Harmless treatment of waste selective catalytic reduction catalysts during ironore sintering process [J]. Journal of Cleaner Production, 2020, 275:122954. [17] CAOG Y D, CHEN S, HU C. Numerical simulation of biomass gas blending in 350 MW coal-fired boiler under oxygen enrich-ment [J]. Modeling and Simulation, 2023, 12(5):4294. [18] CHEN Y, LIU C, WANG Z C, et al. CO process control optimization technology and technique based on gas circulation sintering [J]. Sintering and Pelletizing, 2024, 49(6):129. [19] WU H L, LUO Y F, ZHOU J H, et al. Influence of oxygen enrichment and flue gas circulation on quality index of sinter and CO emissions [J]. Journal of Central South University (Science and Technology), 2022, 53(4):1179. [20] HU J L, ZHOU H Y, LIU Q, et al. Mechanism of CO generation in iron ore sintering process and key technologies for emission reduction [J]. China Metallurgy, 2025, 35(2):94. [21] WANG Y G, YANG X F, LI J. Theoretical studies of CO oxidation with lattice oxygen on Co₃O₄ surfaces [J]. Chinese Journal of Catalysis, 2016, 37(1):193. [22] WANG Y F, YANG T, WANG H Y, et al. Application of steam injection in iron ore sintering: Fuel combustion efficiency and CO emissions [J]. Journal of Iron and Steel Research International, 2023, 30(1):31. [23] YU Y B, ZHAO J J, HAN X, et al. Influence of calcination and pretreatment conditions on the activity of Co₃O₄ for CO oxidation [J]. Chinese Journal of Catalysis, 2013, 34(2):283. [24] JONES J, XIONG H F, DELARIVA A T, et al. Thermally stable single-atom platinum-on-ceria catalysts via atom trapping [J]. Science, 2016, 353(6295):150. [25] THORMÄHLEN P, SKOGLUNDH M, FRIDELL E, et al. Low-temperature CO oxidation over platinum and cobalt oxide catalysts [J]. Journal of Catalysis, 1999, 188(2):300. [26] LOO C E. Role of coke size in sintering of a hematite ore blend [J]. Ironmaking and Steelmaking, 1991, 18(1). [27] MURAKAMI K, SUGAWARA K, KAWAGUCHI T. Analysis of combustion rate of various carbon materials for iron ore sintering process [J]. ISIJ International, 2013, 53(9):1580. [28] Hrithik P. M., Rehman Z M, Dar A A, et al. Forecasting CO₂ Emissions in India: A Time Series Analysis Using ARIMA [J]. Processes, 2024, 12(12):2699-2699. [29] Dar A A, Jain A, Malhotra M, et al. Time Series Analysis with ARIMA for Historical Stock Data and Future Projections [J]. Soft Computing, 2024, 28(21):1-12. [30] ZHU S M, XIA H, LYU X Z, et al. Condition Prediction of Reactor Coolant Pump in Nuclear Power Plants Based on the Combination of ARIMA and LSTM [J]. Nuclear Power Engineering, 2022, 43(02):246-253. [31] CHEN H F, WANG H, LI Y, et al. Short-Term Wind Speed Prediction by Combining Two-Step Decomposition and ARIMA-LSTM [J]. Acta Energiae Solaris Sinica, 2024, 45(02):164-171. [32] WANG Y W, MA S C. Time Series Forecasting Based on ARIMA-DLSTM Hybrid Model [J]. Computer Applications and Software, 2021, 38(02):291-298. [33] HU J H, XIONG L, DAN B B, et al. Prediction Model of Hot Metal KR Desulfurization Based on Improved CNN-LSTM and RF [J]. Journal of Wuhan University of Science and Technology, 2024, 47(04):254-263.