The permeability index is an important variable to reflect the operating state of a blast furnace quickly, intuitively and comprehensively. Given the multi-scale, nonlinear and massive characteristics of the blast furnace smelting process data, a prediction model of permeability index based on the WD-NL-transformer (wavelet de-noising-nonlinear-transformer) method was proposed. The key variables affecting permeability index were selected from multiple perspectives by Pearson correlation coefficient and copula entropy. The extreme outliers were optimised using box plots and Lagrange linear interpolation. A data set was constructed after wavelet de-noising to build the prediction model with nonlinear processing capability, time-series prediction capability, and early stopping method. The results show that the MSE of WD-NL-transformer is 0.0093, with the RMSE of 0.0086. Meanwhile, the R2 is 0.9859, and the running time is 19.59 s, which means the model has great prediction accuracy and inference speed. The proposed model could provide theoretical support for advanced control of permeability index, which is of practical significance to ensure the stability of blast furnace smelting and accelerate the metallurgical process of intelligent automation.
ZhouDDXuKXuG,et al.The production of large blast furnaces of China in 2017 [J]. Ironmaking Steelmaking2020; 47: 316–321.
2.
LiHYLiXLiuXJ, et al.Prediction of blast furnace parameters using feature engineering and stacking algorithm [J]. Ironmaking Steelmaking2022; 49: 283–296.
3.
RenYZXingXDWangBR, et al.Prediction model for silicon content of hot metal based on PSO-TCN [J]. Metall Mater Trans B-Process Metall Mater Process Sci2024; 55: 2837–2846.
4.
XiJHXingXDZhengZY, et al.Effect of feature selection on the prediction model of FeO content in sinter [J]. JOM2023; 75: 5930–5939.
5.
ShimomuraYNishikawaKArinoS, et al.On the internal state of the lumpy zone of blast furnace [J]. Trans Iron Steel Inst Jpn1977; 17: 381–390.
6.
PanYZZuoHBWangJS, et al.Review on improving gas permeability of blast furnace [J]. J Iron Steel Res Int2020; 27: 121–131.
7.
LouSWYangCJWuP, et al.Fault diagnosis of blast furnace iron-making process with a novel deep stationary kernel learning support vector machine approach [J]. IEEE Trans Instrum Meas2022; 71: 3521913.
8.
GhoshSViswanathanNNBallalNB. Flow phenomena in the dripping zone of blast furnace - A review [J]. Steel Res Int2017; 88: 1600440.
9.
GolovchenkoADychkovskyiRPazynichY, et al.Some aspects of the control for the radial distribution of burden material and gas flow in the blast furnace [J]. Energies2020; 13: 923.
10.
LinAQiuGBZhangXL, et al.Development and application of comprehensive diagnostic control model for blast furnace production [J]. Iron Steel2022; 57: 41–56.
11.
PangYSAdemaAVanDS, et al.Modelling of phenomena affecting blast furnace burden permeability using the discrete element method (DEM) - A review [J]. Powder Technol2023; 415: 118161.
12.
ChenLGFengHJXieZH. Generalized thermodynamic optimization for iron and steel production processes: theoretical exploration and application cases [J]. Entropy2016; 18: 353.
13.
KoekemoerALuckosA. Effect of material type and particle size distribution on pressure drop in packed beds of large particles: extending the Ergun equation [J]. Fuel2015; 158: 232–238.
14.
WangHTChuMSGuoTL, et al.Mathematical simulation on blast furnace operation of coke oven gas injection in combination with top gas recycling [J]. Steel Res Int2016; 87: 539–549.
15.
LiQTangRWangS,et al.A coupled LES-LBM-IMB-DEM modeling for evaluating pressure drop of a heterogeneous alternating-layer packed bed [J]. Chem Eng J2022; 433: 133529.
16.
LiuYRShenYS. Three-dimensional modelling of charcoal combustion in an industrial scale blast furnace [J]. Fuel2019; 258: 116088.
17.
ShenYSYuABAustinPR,et al.CFD Study of in-furnace phenomena of pulverised coal injection in blast furnace: effects of operating conditions [J]. Powder Technol2012; 223: 27–38.
18.
SongJLXingXDPangZG,et al.Prediction of silicon content in the hot metal of a blast furnace based on FPA-BP model [J]. Metals2023; 13: 918.
19.
WangSLiQ. Prediction model of blast furnace permeability by combining big data with neural network [J]. Iron Steel2023; 58: 46–53.
20.
LiuXJZhangYJLiX, et al.Prediction for permeability index of blast furnace based on VMD-PSO-BP model [J]. J Iron Steel Res Int2024; 31: 573–583.
21.
ChenHHWuGDLiJX, et al.Research advances on deep learning recommendation based on attention mechanism [J]. Comput Eng Sci2021; 43: 370–380.
22.
KeFCWangWQTanWC, et al.A hierarchical transformer model for session-aware knowledge tracing [J]. Knowl-Based Syst2024; 284: 111300.
23.
RaffelCShazeerNRobertsA, et al.Exploring the limits of transfer learning with a unified text-to-text transformer [J]. J Mach Learn Res2020; 21: 140.
24.
ZhangEYCheokADPanZ, et al.From turing to transformers: a comprehensive review and tutorial on the evolution and applications of generative transformer models[J]. Sci2023; 5: 46.
25.
XuCSLiJYFengB,et al.A financial time-series prediction model based on multiplex attention and linear transformer structure [J]. Appl Sci-Basel2023; 13: 5175.
26.
ZhuTYKuangLPiaoCZ, et al.Population-specific glucose prediction in diabetes care with transformer-based deep learning on the edge [J]. IEEE Trans Biomed Circuits Syst2024; 18: 236–246.
27.
MaJSunZQ. Mutual information is copula entropy [J]. Tsinghua Sci Technol2011; 16: 51–54.
28.
HanMRenWJLiuXX. Joint mutual information-based input variable selection for multivariate time series modeling [J]. Eng Appl Artif Intell2015; 37: 250–257.
KraskovAStogbauerHGrassbergerP. Estimating mutual information [J]. Phys Rev E Stat Nonlin Soft Matter Phys2004; 69: 232–238.
31.
YehKCKwanKC. A comparison of numerical integrating algorithms by trapezoidal, Lagrange, and spline approximation [J]. J Pharmacokinet Biopharm1978; 6: 79–98.
32.
BinZGuoGLHuangGF,et al.A sampled values interface method for substation led based on the linear Lagrange interpolation algorithm [J]. Automation Electric Power Syst2007; 31: 86–90.
33.
PalSMitraM. Empirical mode decomposition based ECG enhancement and QRS detection [J]. Comput Biol Med2012; 42: 83–92.
34.
GongQAnYPLuoSF. Image denoising combined with wavelet transform based on region segmentation [J]. Comput Eng Appl2010; 46: 191–194.
35.
ZhuHLGuoTYQuLS. Signal de-noising by improving soft thresholding on the dyadic wavelet transform [J]. Acta Autom Sin2004; 30: 199–206.
36.
LangdonWB. Genetic programming and evolvable machines at 20 [J]. Genet Program Evolvable Mach2020; 21: 205–217.
37.
LiuSLiuXJLyuQ,et al.Comprehensive system based on a DNN and LSTM for predicting sinter composition [J]. Appl Soft Comput2020; 95: 106574.
38.
MahjoubSChrifi-AlaouiLMarhicB,et al.Predicting energy consumption using LSTM, multi-layer GRU and drop-GRU neural networks [J]. Sensors2022; 22: 4062.
39.
XiongSLiuZXMinCD, et al.Compressive strength prediction of cemented backfill containing phosphate tailings using extreme gradient boosting optimized by whale optimization algorithm [J]. Materials2023; 16: 08.
