Blast furnace (BF) permeability index is a crucial parameter that can quickly, intuitively, and comprehensively reflect the furnace condition. Accurate prediction of this index is crucial for optimising production efficiency and ensuring the stable operation of the BF. In this study, a hybrid permeability index prediction model is constructed by combining a least squares support vector machine and an artificial neural network, using the mean shift clustering algorithm (MSCA) to classify the BF conditions is applied. The results show that the MSCA algorithm shows remarkable precision in classifying the stable and unstable operating states of BF, achieving an impressive accuracy rate of 93.98%. The hybrid prediction model could accurately predict the permeability index and has a mean absolute error of 0.6877, a mean square error of 0.4721 and an R2 of 0.9215, highlighting its robust predictive performance. These findings underscore the practical significance of our model in enhancing BF operational efficiency and reliability.
AgrawalA. Blast furnace performance under varying pellet proportion. Trans Indian Inst Met2019; 72: 777–787.
2.
ShiauJS. Effect of CRI and CSR on the softening-melting behavior of burden in the high temperature zone. Metall Res Technol2020; 116: 625.
3.
JiangDWWangZYLiKJ,et al.Analysis of blast furnace permeability regulation strategy based on machine learning. Steel Res Int2023; 95: 2300590.
4.
LiHYLiXLiuXJ, et al.Prediction of blast furnace parameters using feature engineering and stacking algorithm. Ironmak Steelmak2023; 40: 65–73.
5.
LiuXJZhangYJLiX, et al.Prediction for permeability index of blast furnace based on VMD-PSO-BP model. J Iron Steel Res Int2024; 31: 573–583.
6.
SunWQWangZHWangQ. Hybrid event-, mechanism- and data-driven prediction of blast furnace gas generation. Energy2020; 199: 117497.
7.
ZhangSJiangHHYinYX, et al.The prediction of the gas utilization ratio based on TS fuzzy neural network and particle swarm optimization. Sensors2018; 18: 625.
8.
LiYJZhangSYinYX, et al.A novel online sequential extreme learning machine for gas utilization ratio prediction in blast furnaces. Sensors2018; 17: 1847.
9.
YilmazOBasEEgriogluE. The training of pi-sigma artificial neural networks with differential evolution algorithm for forecasting. Comput Econ2022; 59: 1699–1711.
10.
XuHWQinWSunYN, et al.Attention mechanism-based deep learning for heat load prediction in blast furnace ironmaking process. J Intell Manuf2024; 35: 1207–1220.
11.
HanIJYuanTFLeeJY, et al.Learned prediction of compressive strength of GGBFS concrete using hybrid artificial neural network models. Materials2019; 12: 3708.
12.
ChenBJJiangTWenJ, et al.Review of pellets and blast furnace slag research progress: the effects of MgO on metallurgical properties. Ironmak Steelmak2023; 50: 1022–1036.
13.
LiZNChuMSLiuZG, et al.Furnace heat prediction and control model and its application to large blast furnace. High Temp Mater Process2019; 38: 884–891.
14.
TanKKLiZZHanY, et al.Research and application of coupled mechanism and data-driven prediction of blast furnace permeability index. Appl Sci (Basel)2023; 13: 9556.
15.
KouMYZhouHHongZB, et al.Numerical analysis of effects of different blast parameters on the gas and burden distribution characteristics inside blast furnace. ISIJ Int2020; 60: 856–864.
16.
ZhangLMHuaCCTangYG,et al.Ill-posed Echo State Network based on L-curve method for prediction of blast furnace gas flow. Neural Process Lett2016; 43: 98–113.
17.
JiangLLZhaoYSMengYM, et al.Numerical simulation of co-injection of pulverized coal and blast furnace gas separated by a membrane. Ironmak Steelmak2021; 48: 324–333.
18.
LingYBalachandarS. Simulation and scaling analysis of a spherical particle-laden blast wave. Shock Waves2018; 28: 545–558.
19.
ChuLMCuiGM. Predicting blast furnace permeability index: a deep learning approach with limited time-series data. Metall Res Technol2024; 121: 215.
20.
TunckayaY. Performance assessment of permeability index prediction in an ironmaking process via soft computing techniques. Proc Inst Mech Eng, Part E2017; 231: 1101–1113.
21.
SuXLZhangSYinYX,et al.Prediction model of permeability index for blast furnace based on the improved multi-layer extreme learning machine and wavelet transform. J Franklin Inst-Eng Appl Math2018; 355: 1663–1691.
22.
LuYAZhouHYuanF, et al.Study on feasibility and superiority of preparing TiO2-rich burdens by composite agglomeration process. JOM2023; 75: 3424–3434.
23.
GolovchenkoADychkovskyiRPazynichY, et al.Some aspects of the control for the radial distribution of burden material and gas flow in the blast furnace. Energies2020; 13: 923.
24.
CuiGMChenRMaX,et al.Decision-making optimization of coal injection volume based on evaluation of blast furnace condition. J Control Decis2020; 35: 2803–2809.
25.
LiCXueQGZuoHB, et al.Pressure drop and gas flow in an oxygen blast furnace analyzed by a combination of experimentation and a porous model. Metals2023; 13: 455.
26.
RogozhnikovSPRogozhnikovIS. Mathematical model for estimating the optimal distribution of ore charge along the blast furnace radius based on zonal balances. Metallurgist2020; 63: 1024–1032.
27.
LiuSHSunWQLiWD,et al.Prediction of blast furnace gas generation based on data quality improvement strategy. J Iron Steel Res Int2023; 30: 864–874.
28.
SalmunIShapiroIOpherI,et al.PLDA-based mean shift speakers’ short segments clustering. Comput Speech Lang2017; 45: 411–436.
29.
HyrienOBaranA. Fast nonparametric density-based clustering of large datasets using a stochastic approximation mean-shift algorithm. J Comput Graph Stat2016; 25: 899–916.
30.
LiuYSZhouSBLiuWX, et al.Least-squares support vector machine and successive projection algorithm for quantitative analysis of cotton-polyester textile by near infrared spectroscopy. J Near Infrared Spec2018; 26: 34–43.
31.
ZhangWPYangQKumarM,et al.Application of improved least squares support vector machine in the forecast of daily water consumption. Wirel Pers Commun2018; 102: 3589–3602.
32.
WuYCFengJW. Development and application of artificial neural network. Wirel Pers Commun2018; 102: 1645–1656.
33.
MousapourAHajipourARashidiMM,et al.Performance evaluation of an irreversible Miller cycle comparing FTT (finite-time thermodynamics) analysis and ANN (artificial neural network) prediction. Energies2016; 94: 100–109.
