基于最优工况迁移的高炉铁水硅含量预测方法

蒋朝辉; 许川; 桂卫华; 蒋珂

doi:10.16383/j.aas.c200980

基于最优工况迁移的高炉铁水硅含量预测方法

doi: 10.16383/j.aas.c200980

蒋朝辉^{1, 2,},
许川^1,,
桂卫华^{1, 2,},
蒋珂^1,

1.
中南大学自动化学院长沙 410000
2.
鹏城实验室深圳 518000

基金项目: 国家自然科学基金(61773406, 61988101), 中南大学中央高校基本科研任务业务费专项资金(2020zzts572)资助

详细信息

作者简介:
蒋朝辉：中南大学自动化学院教授, 鹏城实验室研究员. 2011年获得中南大学博士学位. 主要研究方向为光电信息感知, 图像处理, 人工智能, 工业VR和智能优化控制. E-mail: jzh0903@csu.edu.cn

许川：中南大学自动化学院博士研究生. 主要研究方向为复杂工业过程建模, 数据分析和机器学习. 本文通信作者. E-mail: csuxuchuan@csu.edu.cn

桂卫华：中国工程院院士, 中南大学自动化学院教授, 鹏城实验室研究员. 1981年获得中南矿冶学院硕士学位. 主要研究方向为复杂工业过程建模与最优控制, 分布式鲁棒控制和故障诊断. E-mail: gwh@csu.edu.cn

蒋珂：中南大学自动化学院博士研究生. 2019年获得中南大学硕士学位. 主要研究方向为数据驱动的工业过程建模与控制, 过程数据分析和机器学习. E-mail: jiangke@csu.edu.cn

计量
- 文章访问数: 1413
- HTML全文浏览量: 302
- PDF下载量: 286
- 被引次数: 0
出版历程
- 收稿日期: 2020-11-25
- 网络出版日期: 2021-05-15
- 刊出日期: 2022-01-25

Prediction Method of Hot Metal Silicon Content in Blast Furnace Based on Optimal Smelting Condition Migration

JIANG Zhao-Hui^{1, 2
,},
XU Chuan^1
,,
GUI Wei-Hua^{1, 2
,},
JIANG Ke^1
,

1.
School of Automation, Central South University, Changsha 410000
2.
Peng Cheng Laboratory, Shenzhen 518000

Funds: Supported by National Natural Science Foundation of China (61773406, 61988101), and Central South University Central University Basic Scientific Research Task Business Expenses Special Funds (2020zzts572)

More Information

Author Bio:
JIANG Zhao-Hui　Professor at the School of Automation, Central South University. Professor at the Peng Cheng Laboratory. He received his Ph. D. degree from Central South University in 2011. His research interest covers photoelectric information perception, image processing, artificial intelligence, industrial VR, and intelligent optimization control

XU Chuan　Ph. D. candidate at the School of Automation, Central South University. His research interest covers complex industrial process modeling, data analysis, and machine learning. Corresponding author of this paper

GUI Wei-Hua　Academician of Chinese Academy of Engineering, and professor at the School of Automation, Central South University. Professor at the Peng Cheng Laboratory. He received his master degree from Central South Institute of Mining and Metallurgy in 1981. His research interest covers modeling and optimal control of complex industrial process, distributed robust control, and fault diagnoses

JINAG Ke　Ph. D. candidate at the School of Automation, Central South University. She received her master degree from Central South University in 2019. Her research interest covers data-based modeling and control of industrial process, process data analysis, and machine learning

摘要

摘要: 高炉铁水硅含量是铁水品质与炉况的重要表征, 冶炼过程关键参数频繁波动及大时滞特性给高炉铁水硅含量预测带来了巨大挑战. 提出一种基于最优工况迁移的高炉铁水硅含量预测方法. 首先, 针对过程变量频繁波动问题, 提出基于邦费罗尼指数的自适应密度峰值聚类算法, 实现对高炉冶炼过程变量的工况划分, 并建立不同工况硅含量预测子模型. 其次, 针对冶炼过程的大时滞特性, 定义相邻时间节点间的硅含量工况迁移代价函数, 并提出多源路径寻优算法, 实现冶炼过程中硅含量最优工况迁移路径及当前时刻硅含量最优预测值的求解. 最后, 基于工业现场数据验证了所提方法的有效性与准确性.
- 高炉炼铁 /
- 铁水硅含量 /
- 预测 /
- 工况迁移 /
- 密度峰值聚类
Abstract: The hot metal silicon content in blast furnace can characterize the hot metal quality and the condition of blast furnace. It poses a great challenge to the online prediction of silicon content because of the frequent fluctuation of smelting parameters and the existence of large time delay during the ironmaking process. This paper proposes an algorithm for predicting the hot metal silicon content in blast furnace based on optimal smelting condition migration. Firstly, arming at the frequent fluctuation of smelting process variables, an adaptive density peak clustering algorithm based on the Bonferroni index to dynamically cluster the process variables of blast furnace ironmaking process is proposed, which can obtain clusters of different smelting conditions, and establish sub-models for different smelting conditions. Secondly, to mitigate the large time delay of blast furnace ironmaking process, this paper defines the silicon content migration cost function between adjacent time nodes, and proposes a multi-source path optimization algorithm to solve the optimal migration path of silicon content during the smelting process and the optimal prediction value of silicon content at the current time. Finally, the effectiveness and accuracy of the proposed method are verified based on the industrial field data.
- Blast furnace ironmaking /
- hot metal silicon content /
- prediction /
- smelting condition migration /
- density peak clustering

HTML全文

图 1 高炉炼铁工艺

Fig. 1 Blast furnace ironmaking process

下载: 全尺寸图片幻灯片

图 2 基于最优工况迁移的建模策略

Fig. 2 Modeling strategy based on optimal smelting condition migration

下载: 全尺寸图片幻灯片

图 3 Elman神经网络结构

Fig. 3 Structure of Elman neural network

下载: 全尺寸图片幻灯片

图 4 滑动窗口采样

Fig. 4 Sliding window sampling

下载: 全尺寸图片幻灯片

图 5 过程变量工况隶属度与模型预测

Fig. 5 Process variable membership degree matching and model prediction

下载: 全尺寸图片幻灯片

图 6 铁水硅含量工况迁移图

Fig. 6 Smelting condition migration diagram of hot metal silicon content

下载: 全尺寸图片幻灯片

图 7 相邻时间节点工况迁移代价函数

Fig. 7 Smelting condition migration cost function of adjacent node

下载: 全尺寸图片幻灯片

图 8 相邻时间节点连接图

Fig. 8 Connection graph of adjacent node

下载: 全尺寸图片幻灯片

图 9 邦费罗尼指数曲线

Fig. 9 Bonferroni index curve

下载: 全尺寸图片幻灯片

图 10 聚类中心决策图

Fig. 10 Decision diagram of cluster center

下载: 全尺寸图片幻灯片

图 11 聚类中心截断系数

Fig. 11 Truncation coefficient of cluster center

下载: 全尺寸图片幻灯片

图 12 4种工况聚类簇

Fig. 12 Clusters of 4 smelting conditions

下载: 全尺寸图片幻灯片

图 13 最优工况迁移模型硅含量预测结果

Fig. 13 Prediction of silicon content in hot metal based on optimal smelting condition migration model

下载: 全尺寸图片幻灯片

图 14 Elman网络预测结果

Fig. 14 Prediction of Elman network

下载: 全尺寸图片幻灯片

图 15 Elman-Adaboost网络预测结果

Fig. 15 Prediction of Elman-Adaboost network

下载: 全尺寸图片幻灯片

图 16 FEEMD-Adaboost-Elman网络预测结果

Fig. 16 Prediction of FEEMD-Adaboost-Elman network

下载: 全尺寸图片幻灯片

图 17 模型预测误差

Fig. 17 Model prediction error curve

下载: 全尺寸图片幻灯片

图 18 硅含量预测值与实际值散点图

Fig. 18 The scatter plot of observed and predicted

下载: 全尺寸图片幻灯片

表 1 过程变量MIC相关性系数

Table 1 MIC correlation coefficient of process variables

过程变量	MIC 系数	过程变量	MIC 系数
富氧率	0.291	总压差	0.204
透气性指数	0.270	炉腹煤气指数	0.278
标准风速	0.275	热风压力	0.268
富氧流量	0.218	实际风速	0.173
冷风流量	0.264	冷风温度	0.209
鼓风动能	0.204	热风温度	0.213
设定喷煤量	0.241	顶温下降管	0.209
理论燃烧温度	0.248	铁水红外温度	0.291
顶压	0.195	顶温	0.292
富氧压力	0.229	鼓风湿度	0.179
冷风压力	0.197	阻力系数	0.204

下载: 导出CSV

表 2 聚类中心截断标志

Table 2 Cluster center truncation flag

序号	1	2	3	4	5	6
截断系数	3.00	4.02	42.30	52.50	28.02	24.34

下载: 导出CSV

表 3 寻优算法耗时对比

Table 3 Comparison of the time consumption of optimization algorithms

寻优算法	节点数
寻优算法	40	80	120	160	200
Floyd 算法耗时 (ms)	3.20 × 10⁴	2.72 × 10⁵	8.96 × 10⁵	2.09 × 10⁶	4.05 × 10⁶
本文算法耗时(ms)	3	8	11	13	18

下载: 导出CSV

表 4 模型性能对比

Table 4 Model performance comparison

模型类别	性能指标
模型类别	数值预测命中率 (%)	趋势预测准确率 (%)	预测均方误差
工况迁移预测模型	88	82	0.0043
Elman 网络	79	69	0.0069
Elman-Adaboost	85	71	0.0054
FEEMD-Adaboost-Elman	86	74	0.0049

下载: 导出CSV

参考文献(33)

[1]	Biswas A K. Principles of Blast Furnace Ironmaking —— Theory and Practice. Brisbane: Cootha Publishing House, 1981. 1−12
[2]	Zhou Ping, Zhang Shuai, Dai Peng. Recursive Learning Based Bilinear Subspace Identification for Online Modeling and Predictive Control of a Complicated Industrial Process. IEEE Access, 2020, 8: 6253-62541.
[3]	Saxén H, Gao Chuan-Hou, Gao Zhi-Wei. Data-driven time discrete models for dynamic prediction of the hot metal silicon content in the blast furnace—A review. IEEE Transactions on Industrial Informatics, 2012, 9(4): 2213-2225.
[4]	Onorin O P, Spirin N A. Features of blast furnace transient processes. Metallurgist, 2017, 61(1): 121-126.
[5]	Spirin N A, Onorin O P, Istomin A S, Lavrov V V, Gurin I A. Study of transition processes of blast-furnace smelting by the mathematical model method. In: Proceedings of the 2018 IOP Conference Series, Materials Science and Engineering. Novokuznetsk, Russia: Institute of Physics Publishing, 2018. 012−073
[6]	Spirin N, Onorin O, Alexander I. Prediction of blast furnace thermal state in real-time operation. Solid State Phenomena, 2020, 299: 518-523. doi: 10.4028/www.scientific.net/SSP.299.518
[7]	Spirin N. A, Polinov A A, Gurin I A, Pishnograev SN. Information System for Real-Time Prediction of the Silicon Content of Iron in a Blast Furnace. Metallurgist, 2020, 63(9): 898-905.
[8]	Östermark R, Saxen H. VARMAX-modelling of blast furnace process variables. European Journal of Operational Research, 1996, 90(1): 85-101. doi: 10.1016/0377-2217(94)00304-1
[9]	Saxen H, Östermark R. State realization with exogenous variables-A test on blast furnace data. European journal of operational research, 1996, 89(1): 34-52. doi: 10.1016/S0377-2217(96)90050-8
[10]	Bhattacharya T. Prediction of silicon content in blast furnace hot metal using partial least squares. ISIJ international, 2005, 45(12): 1943-1945. doi: 10.2355/isijinternational.45.1943
[11]	Li Jun-Peng, Hua Chang-Chun, Yang Yan-Na, Guan Xin-Ping. Bayesian block structure sparse based T–S fuzzy modeling for dynamic prediction of hot metal silicon content in the blast furnace. IEEE Transactions on Industrial Electronics, 2017, 65(6): 4933-4942.
[12]	Xu Xia, Hua Chang-Chun, Tang Ying-Gan, Guan Xing-Ping. Modeling of the hot metal silicon content in blast furnace using support vector machine optimized by an improved particle swarm optimizer. Neural Computing and Applications, 2016, 27(6): 1451-1461. doi: 10.1007/s00521-015-1951-7
[13]	Han Y, Li J, Yang X L, Liu W X, Zhang Y Z. Dynamic prediction research of silicon content in hot metal driven by big data in blast furnace smelting process under hadoop cloud platform. Complexity, DOI: 10.1155/2018/8079697
[14]	Xu X, Hua C C, Tang Y G, Guan X P. Wiener model identification of blast furnace ironmaking process based on laguerre filter and linear programming support vector regression. In: Proceedings of the 2014 International Joint Conference on Neural Networks. Beijing, China: IEEE Press, 2014. 2198−2204
[15]	Zhou Ping, Li Wen-Peng, Wang Hong. Robust Online Sequential RVFLNs for Data Modeling of Dynamic Time-Varying Systems With Application of an Ironmaking Blast Furnace. IEEE Transactions on Cybernetics, 2020, 50(11): 4783-4795. doi: 10.1109/TCYB.2019.2920483
[16]	郜传厚, 渐令, 陈积明, 孙优贤. 复杂高炉炼铁过程的数据驱动建模及预测算法. 自动化学报, 2009, 35(06): 725-730. doi: 10.3724/SP.J.1004.2009.00725 Gao Chuan-Hou, Jian Ling, Chen Ji-Ming, Sun You-Xian. Data-driven modeling and prediction algorithm for complex blast furnace ironmaking process. Acta Automatica Sinica, 2009, 35(06): 725-730. doi: 10.3724/SP.J.1004.2009.00725
[17]	David S F, David F F, Machado M L P. Artificial neural network model for predict of silicon content in hot metal blast furnace. Materials Science Forum, 2016, 869: 572-577. doi: 10.4028/www.scientific.net/MSF.869.572
[18]	宋菁华, 杨春节, 周哲. 改进型 EMD-Elman 神经网络在铁水硅含量预测中的应用. 化工学报, 2016, 67(3): 729-735. Song Jing-Hua, Yang Chun-Jie, Zhou Zhe. Application of improved EMD-Elman neural network in prediction of silicon content in molten iron. CIESC Journal, 2016, 67(3): 729-735.
[19]	Jiang Ke, Jiang Zhao-Hui, Xie Yon-Fang, Chen Zhi-Peng. Classification of silicon content variation trend based on fusion of multilevel features in blast furnace ironmaking. Information Sciences, 2020, 521: 32-45. doi: 10.1016/j.ins.2020.02.039
[20]	周平, 张丽, 李温鹏, 戴鹏, 柴天佑. 集成自编码与PCA的高炉多元铁水质量随机权神经网络建模. 自动化学报, 2018, 44(10): 1799-1811. Zhou Ping, Zhang Li, Li Wen-Peng, Dai Peng, Chai Tian-You. Modeling of blast furnace multi-element molten iron quality with random weight neural network based on self-encoding and PCA. Acta Automatica Sinica, 2018, 44(10): 1799-1811.
[21]	Jian Ling, Song Yun-Quan, Shen Shu-Qian, Wang Yan, Yin Hai-Qing. Adaptive least squares support vector machine predictor for blast furnace ironmaking process. ISIJ International, 2015, 55(4): 845-850. doi: 10.2355/isijinternational.55.845
[22]	Zeng Jiu-Sun, Liu Xiang-Guan, Gao Chuan-Hou, Luo Shi-Hua, Jian Ling. Wiener model identification of blast furnace ironmaking process. ISIJ International, 2008, 48(12): 1734-1738. doi: 10.2355/isijinternational.48.1734
[23]	蒋朝辉, 董梦林, 桂卫华, 阳春华, 谢永芳. 基于Bootstrap的高炉铁水硅含量二维预报. 自动化学报, 2016, 42(05): 715-723. Jiang Zhao-Hui, Dong Meng-Lin, Gui Wei-Hua, Yang Chun-Hua, Xie Yon-Fang. Two-dimensional prediction for silicon content of hot metal of blast furnace based on bootstrap. Acta Automatica Sinica, 2016, 42 (5): 715-723.
[24]	李温鹏, 周平. 高炉铁水质量鲁棒正则化随机权神经网络建模. 自动化学报, 2020, 46(04): 721-733. Li Wen-Peng, Zhou Ping. Wen Liang. Blast furnace hot metal quality robust regularization random weight neural network modeling. Acta Automatica Sinica, 2020, 46(04): 721-733.
[25]	温亮, 周平. 基于多参数灵敏度分析与遗传优化的铁水质量无模型自适应控制. 自动化学报. Wen Liang, Zhou Ping. Model-free adaptive control of molten iron quality based on multi-parameter sensitivity analysis and genetic optimization. Acta Automatica Sinica, to be published.
[26]	Rodriguez, A, Laio A. Clustering by fast search and find of density peaks. Science, 2014, 344(6191): 1492-1496. doi: 10.1126/science.1242072
[27]	Martin B, Elena, Silber J. The Bonferroni index and the measurement of distributional change. Metron, 2017, 75(1): 1-16. doi: 10.1007/s40300-016-0105-8
[28]	孙甜, 凌卫新. 基于模拟退火的 Levenberg-Marquardt 算法在神经网络中的应用. 科学技术与工程, 2008(18): 5189−5192 Sun Tian, Ling Wei-Xin. Application of Levenberg-Marquardt algorithm based on simulated annealing in neural network. Science Technology and Engineering, 2008(18): 5189−5192
[29]	Reshef, D N, Reshef Y A. Detecting novel associations in large data sets. Science, 2011, 334(6062): 1518-1524. doi: 10.1126/science.1205438
[30]	Pan Dong, Jiang Zhao-Hui, Chen Zhi-Peng. Temperature measurement and compensation method of blast furnace molten iron based on infrared computer vision. IEEE Transactions on Instrumentation and Measurement, 2018, 68(10): 3576-3588.
[31]	庄田, 杨春节. 基于Elman-Adaboost强预测器的铁水硅含量预测方法. 冶金自动化, 2017, 41(04): 1-6+17. Zhuang Tian, Yang Chun-Jie. Prediction method of silicon content in molten iron based on Elman-Adaboost strong predictor. Metallurgical Automation, 2017, 41(04): 1-6+17.
[32]	王凯, 毕贵红, 高晗, 蒲娴怡, 陈仕龙. 基于改进快速集合经验模态分解和Elman-Adaboost的短期风速预测方法. 电力科学与工程, 2020, 36(05): 32-39. doi: 10.3969/j.ISSN.1672-0792.2020.05.005 Wang Kai, Bi Gui-Gong, Gao Han, Pu Xian-Yi, Chen Shi-Long. Short-term Wind Speed Prediction Method Based on Improved Fast Ensemble Empirical Mode Decomposition and Elman-Adaboost. Electric Power Science and Engineering, 2020, 36(05): 32-39. doi: 10.3969/j.ISSN.1672-0792.2020.05.005
[33]	蒋珂, 蒋朝辉, 谢永芳, 潘冬, 桂卫华. 大型高炉铁水硅含量变化趋势的智能预报. 控制工程, 2020, 27(03): 540-546. Jiang Ke, Jiang Zhao-Hui, Xie Yon-Fang, Pan Dong, Gui Wei-Hua. Intelligent prediction of silicon content change trend in molten iron of large blast furnace. Control Engineering, 2020, 27 (03): 540-546.