Prediction model of end-point phosphorus content of converter based on FA-ELM
GAO Fang1, BAO Yan-ping1, WANG Min1, LIU Yu2, HUANG Yong-sheng2, SUN Guang-tao2
1. State Key Lab of Advanced Metallurgy, University of Science and Technology Beijing, Beijing 100083, China; 2. The Third Steelmaking Plant, Zenith Steel Group Co., Ltd., Changzhou 213011, Jiangsu, China
Abstract：Accurate prediction of end-point phosphorus content in converter is of great significance to realize automatic tapping and improve converter smelting efficiency. In order to realize the narrow window control of end-point phosphorus content in converter and provide guidance for operation process, it is necessary to establish a more accurate prediction model of end-point phosphorus content. Based on the analysis of the smelting process and dephosphorization thermodynamics, 12 observable indexes affecting the dephosphorization process are selected to build the end-point phosphorus content prediction index system. Then, by using factor analysis (FA) to reduce the dimension of the data, six derived variables are obtained, which are used as the model input and the end-point phosphorus content is used as the model output, and the end-point phosphorus content prediction model based on the overrun learning machine (ELM) is established. By comparing the prediction results of the ELM model with the BP neural network, it is found that the ELM model has a higher fitting degree, R2 = 0.778 7, MAPE=0.106 0, and the furnace number with prediction error within ±0.003 0% accounts for 86.67%. Therefore, compared with the BP neural network model, the ELM model has higher accuracy and better generalization ability.
高放, 包燕平, 王敏, 刘宇, 黄永生, 孙光涛. 基于FA-ELM的转炉终点磷含量预测模型[J]. 钢铁, 2020, 55(12): 24-30.
GAO Fang, BAO Yan-ping, WANG Min, LIU Yu, HUANG Yong-sheng, SUN Guang-tao. Prediction model of end-point phosphorus content of converter based on FA-ELM[J]. Iron and Steel, 2020, 55(12): 24-30.
[1] 李光辉, 刘青. 转炉炼钢过程工艺控制的发展与展望[J]. 钢铁研究学报,2013,25(1):1.(LI Guang-hui, LIU Qing. Present status and prospect of BOF steelmaking process control[J]. Journal of Iron and Steel Research,2013, 25(1):1.) [2] 汪淼,李胜利,高闯,等.80 t转炉终点预报TSVR模型精度[J].钢铁,2020,55(7):53.(WANG Miao, LI Sheng-li, GAO Chuang, et al. End-point prediction TSVR model accuracy of 80t BOF steelmaking[J]. Iron and Steel, 2020,55(7):53.) [3] 田晓涛,冯春松,胡正彪,等.灰色模型在转炉终点温度预测中的应用[J].中国冶金,2020,30(3):51.(TIAN Xiao-tao, FENG Chun-song, HU Zheng-biao, et al. Application of gray model in converter steel-making endpoint temperature prediction[J]. China Metallurgy, 2020,30(3):51.) [4] Klanke S, Löpke M, Uebber N, et al. Advanced data-driven prediction modelsfor BOF endpoint detection[C]//Association for Iron and Steel Technology 2017 Proceedings. Nashville:[s.n.], 2017:1307. [5] DONG X L, DONG S. The converter steelmaking end point prediction model based on RBF neural network[J]. Applied Mechanics and Materials, 2014, 577 (1): 98. [6] 李长荣, 赵浩文, 谢祥, 等. 基于L-M算法BP神经网络的转炉炼钢终点磷含量预报[J]. 钢铁, 2011, 46(4): 13.(LI Chang-rong, ZHAO Hao-wen, XIE Xiang, et al. Prediction of end-point phosphorus content for BOF based on LM BP neural network[J], Iron and Steel, 2011,46 (4):13. ) [7] LIU H, WANG B, XIONG X. Basic oxygen furnace steelmaking end-point prediction based on computer vision and general regression neural network[J]. Optik-International Journal for Light and Electron Optics, 2014, 125(18):5241. [8] 张壮, 曹玲玲, 林文辉, 等. 基于IPSO-RELM转炉冶炼终点锰含量预测模型[J]. 工程科学学报, 2019, 41(8):1052.(ZHANG Zhuang, CAO Ling-ling, LIN Wen-hui, et al. Improved prediction model for BOF end-point manganese content based on IPSO-RELM method[J]. Chinese Journal of Engineering, 2019, 41(8):1052.) [9] HE F, ZHANG L. Prediction model of end-point phosphorus content in BOF steelmaking process based on PCA and BP neural network[J]. Journal of Process Control, 2018, 66:51. [10] 张长鲁. 基于因子分析的煤质指标降维及其可视化应用[J]. 煤炭技术,2017, 36(1):310.(ZHANG Chang-lu. Dimension reduction for coal quality index and its visual application based on factor analysis[J]. Coal Technology,2017, 36(1):310.) [11] 巩知乐, 张少龙, 廖海斌. 因子分析判别准则的线性降维方法研究[J]. 计算机工程与应用,2016, 52(3):145.(GONG Zhi-le, ZHANG Shao-long, LIAO Hai-bin. Linear dimension reduction by factors analysis discriminant criterion[J]. Computer Engineering and Applications,2016, 52(3):145.) [12] DING S, JIA W, SU C, et al. Research of neural network algorithm based on factor analysis and cluster analysis[J]. Neural Computing and Applications, 2011, 20(2): 297. [13] 陈恒志, 杨建平, 卢新春, 等. 基于极限学习机(ELM)的连铸坯质量预测[J]. 工程科学学报, 2018(7):815.(CHEN Heng-zhi, YANG Jian-ping, LU Xin-chun, et al. Quality prediction of the continuous casting bloom based on the extreme learning Machine[J]. Chinese Journal of Engineering,2018(7):815.) [14] OU T Y, CHENG C Y, CHEN P J, et al. Dynamic cost forecasting model based on extreme learning machine-A case study in steel plant[J]. Computers and Industrial Engineering, 2016, 101:544. [15] 智建国,吴伟,高琦,等.大型转炉优化造渣工艺提高脱磷的效果[J].钢铁,2020,55(7):72.(ZHI Jian-guo, WU Wei, GAO Qi, et al. Effect of improving dephosphorization of molten steel by optimizing slagging in large converters[J]. Iron and Steel, 2020,55(7):72.) [16] Zebende G F. DCCA cross-correlation coefficient: Quantifying level of cross-correlation[J]. Physica a Statistical Mechanics and its Applications, 2011, 390(4):614. [17] 李孜军, 王晓磊, 石东平. 硫化矿石低温氧化性指标的相关性分析[J]. 科技导报, 2011, 29(36):28.(LI Zi-jun, WANG Xiao-lei, SHI Dong-ping. Analysis of the correlation among indexes of sulfide ores in oxidation process at ambient temperature[J]. Science and Technology Review, 2011, 29(36):28.)
2020, Vol. 55 
