作者：古志远¹, 吕东澔¹, 李向丽², 张勇¹, 代学冬¹

作者单位：1. 内蒙古科技大学，内蒙古 包头 014010;
2. 常熟理工学院电气与自动化工程学院，江苏 苏州 215500

关键词：高炉炼铁;经验知识;神经网络;高炉炉温

摘要：

针对复杂高炉冶炼过程具有大滞后等特点，为提高高炉炉温预测精度，提出一种经验知识与数据相结合的炉温融合预测方法。首先，根据高炉经验知识，分析各变量在高炉内的滞后关系，以及在滞后时间内停留在高炉内部形成的累积关系，累积量对当前炉温造成的影响。通过累积量进行相关性分析，合理地确定输入变量。然后，将铁水温度与铁水硅含量融合来更好地表征炉温。最后，通过神经网络利用累积量作为输入建立经验知识与数据相结合的高炉炉温融合预测模型。实验中使用某钢厂高炉生产数据进行仿真，结果表明累积量模型具有良好的性能，可为高炉炉温预测提供新思路。

Fusion prediction of blast furnace temperature based on combination of knowledge and data
GU Zhiyuan¹, Lü Donghao¹, LI Xiangli², ZHANG Yong¹, DAI Xuedong¹
1. Inner Mongolia University of Science and Technology, Baotou 014010, China;
2. School of Electrical and Automation Engineering, Changshu Institute of Technology, Suzhou 215500, China
Abstract: In view of the large lag in the complex blast furnace smelting process, in order to improve the accuracy of blast furnace temperature prediction, a furnace temperature fusion prediction method combining empirical knowledge and data was proposed. First, this paper analyzed experience of the blast furnace, obtained the hysteresis relationship of each variable in the blast furnace and each variable had a cumulative relationship within the lag time in the blast furnace, and the accumulations affected the current blast furnace temperature. Selecting the input variable reasonably by correlation analysis of the accumulations of each variable. Then a method of fusing the temperature of the molten iron with the silicon content of the molten iron was proposed to better characterize the temperature of the blast furnace. Finally, based on combining knowledge of experience and data, a neural network was used to establish the fusion model for prediction of the blast furnace temperature by accumulations as inputs. In the experiment, the data came from blast furnace production in a steel plant and this data is used for simulation, the model has good performance and provides new ideas for prediction of the blast furnace temperature.
Keywords: blast furnace ironmaking;knowledge of experience;neural network;blast furnace temperature
2024, 50(3):19-28  收稿日期: 2021-12-29;收到修改稿日期: 2022-02-27
基金项目: 国家自然科学基金资助项目(61763038，61763039，61803049)；内蒙古自然科学基金项目(2019BS06004，2020LH06006)
作者简介: 古志远(1995-),男,河南焦作市人,硕士研究生,专业方向为高炉炉温的建模与控制。
参考文献
[1] 周平, 李瑞峰, 郭东伟, 等. 高炉炼铁过程多元铁水质量指标多输出支持向量回归建模[J]. 控制理论与应用, 2016, 33(6): 727-734
ZHOU P, LI R F, GUO D W, et al. Multi-output support vector regression modeling for multivariate molten iron quality indices in blast furnace ironmaking process[J]. Control Theory & Applications, 2016, 33(6): 727-734
[2] 崔桂梅, 陈荣, 于凯, 等. 基于多尺度分解的ELM炉温预测研究[J]. 控制工程, 2020, 27(11): 1901-1906
CUI G M, CHEN R, YU K, et al. The study on temperature prediction of ELM furnace based on multiscale decomposition[J]. Control Engineering of China, 2020, 27(11): 1901-1906
[3] 蒋朝辉, 董梦林, 桂卫华, 等. 基于Bootstrap的高炉铁水硅含量二维预报[J]. 自动化学报, 2016, 42(5): 715-723
JIANG Z H, DONG M L, GUI W H, et al. Two-dimensional prediction for silicon content of hot metal of blast furnace based on Bootstrap[J]. Acta Automatica Sinica, 2016, 42(5): 715-723
[4] 方一鸣, 赵晓东, 张攀, 等. 基于改进灰狼算法和多核极限学习机的铁水硅含量预测建模[J]. 控制理论与应用, 2020, 37(7): 1644-1654
FANG Y M, ZHAO X D, ZHANG P, et al. Prediction modeling of silicon content in liquid iron based on multiple kernel extreme learning machine and improved grey wolf optimizer[J]. Control Theory & Applications, 2020, 37(7): 1644-1654
[5] 尹林子, 李乐, 蒋朝辉. 基于粗糙集理论与神经网络的铁水硅含量预测[J]. 钢铁研究学报, 2019, 31(8): 689-695
YIN L Z, LI L, JIANG Z H. Prediction of silicon content in hot metal using neural network and rough set theory[J]. Journal of Iron and Steel Research, 2019, 31(8): 689-695
[6] 崔桂梅, 李静, 张勇, 等. 基于T-S模糊神经网络模型的高炉铁水温度预测建模[J]. 钢铁, 2013, 48(11): 11-15
CUI G M, LI J, ZHANG Y, et al. Prediction modeling study for blast furnace hot metal temperature based on T-S fuzzy neural network model[J]. Iron and Steel, 2013, 48(11): 11-15
[7] ZHANG H G, YIN Y X, ZHANG S. An improved ELM algorithm for the measurement of hot metal temperature in blast furnace[J]. Neurocomputing, 2016, 174: 232-237
[8] 刘琦. 高炉强化后的基本操作制度选择[J]. 钢铁, 2004, 39(3): 4-10
LIU Q. Basic regime for BF of intensified operation[J]. Iron & Steel, 2004, 39(3): 4-10
[9] 王明海. 炼铁原理与工艺[M]. 北京: 冶金工业出版社, 2012.
[10] 陈春俊, 杨露, 何智颖, 等. ARIMA-BP神经网络高速列车隧道压力波预测模型研究[J]. 中国测试, 2021, 47(10): 80-86
CHEN C J, YANG L, HE Z Y, et al. Research on tunnel pressure wave prediction model of high-speed train based on ARIMA-BP neural network[J]. China Measurement & Test, 2021, 47(10): 80-86
[11] 崔桂梅, 李静, 张勇, 等. 高炉铁水温度的多元时间序列建模和预测[J]. 钢铁研究学报, 2014, 26(4): 33-37
CUI G M, LI J, ZHANG Y, et al. Multivariate time series modeling research for blast furnace hot iron temperature[J]. Journal of Iron and Steel Research, 2014, 26(4): 33-37
[12] 罗世华, 陈坤. 基于偏态深度分类的高炉硅含量及波动预测[J]. 控制与决策, 2021, 36(2): 491-497
LUO S H, CHEN K. Prediction of blast furnace silicon content and fluctuation based on skewness depth classification[J]. Control and Decision, 2021, 36(2): 491-497
[13] 范广权. 高炉炼铁操作[M]. 北京: 冶金工业出版社, 2010.
[14] 邬肖敏, 李世平, 程双江. 基于小波神经网络和PSO的动态误差溯源方法研究[J]. 中国测试, 2014, 40(6): 27-30
WU X M, LI S P, CHENG S J. Research of dynamic error tracing method based on wavelet neural network and PSO[J]. China Measurement & Test, 2014, 40(6): 27-30
[15] 吴新忠, 耿柯, 陈昌. 基于IGOA-RBF的矿用风压传感器温度补偿研究[J]. 中国测试, 2021, 47(6): 137-143
WU X Z, GENG K, CHEN C. Research on temperature compensation of mine wind pressure sensor based on IGOA-RBF neural network[J]. China Measurement & Test, 2021, 47(6): 137-143