Prediction model of converter gas production based on deep-learning
BAO Xiangjun1, CHEN Kai1, LI Xiuping2, YANG Xiaojing1, LIU Xiao3, CHEN Guang1
1. School of Energy and Environment, Anhui University of Technology, Maanshan 243000, Anhui, China; 2. Steel Industry Green and Intelligent Manufacturing Technology Center, China Iron and Steel Research Institute Group Co., Ltd., Beijing 100081, China; 3. Iron and Steel Research Institute Co., Ltd., Beijing 100081, China
Abstract:The prediction of the occurrence of converter gas provides important support for the micro differential pressure control at the converter inlet, the improvement of converter gas recovery quality, and the overall gas balance scheduling of the plant. Based on the actual occurrence data of converter gas during the blowing process of a certain steel plant, a deep learning method was used to establish three prediction models for converter gas occurrence: BP neural network, LSTM long short memory neural network, and RBFNN radial basis function neural network. The effects of three parameters, namely prediction steps, input sample size, and hidden unit number, on the accuracy and computational efficiency of the prediction model were compared and analyzed. The research results indicate that the prediction accuracy of the model decreases with the increase of prediction steps, and choosing 30 step prediction is more in line with the actual needs of steel mills. As the sample input increases, there is no significant change in the accuracy of LSTM, while the accuracy of BP shows a decreasing trend. The accuracy of RBF first increases significantly and then slowly decreases. The prediction efficiency of LSTM showed no significant change, BP significantly decreased, and RBF remained unchanged. When the three models are under the optimal sample input and 30 step prediction conditions, the accuracy of LSTM remains basically unchanged as the number of hidden units increases. BP first slightly increases and then slowly decreases, while RBF first increases significantly and then remains stable, and then decreases significantly. The prediction efficiency of LSTM has slightly decreased, BP has significantly decreased, and RBF remains unchanged. Finally, under the condition of 30 step prediction, the optimal parameter conditions for LSTM, BP, and RBF models are as follows: LSTM sample input quantity is 125, hidden unit number is 135, ERMS minimum is 13.38, and training duration is 4.7 min; The input amount of BP samples is 50, the number of hidden units is 60, and the minimum ERMS is 31.46, with a training duration of 16.8 min; The input amount of RBF samples is 210, and the number of hidden units is 210. At this time, the minimum ERMS is 2.07, and the training duration is 1.2 min. Compared with actual data, RBF has the best prediction effect. By using the prediction results of converter gas generation to regulate the speed of the fan, the micro differential pressure at the furnace mouth can be maintained in a more stable state, reducing the suction air volume, and improving the heat value of the recovered gas.
包向军, 陈凯, 郦秀萍, 杨筱静, 刘骁, 陈光. 基于深度学习的转炉煤气发生量预测模型[J]. 钢铁, 2024, 59(1): 67-74.
BAO Xiangjun, CHEN Kai, LI Xiuping, YANG Xiaojing, LIU Xiao, CHEN Guang. Prediction model of converter gas production based on deep-learning[J]. Iron and Steel, 2024, 59(1): 67-74.
[1] 于鹏飞,曾加庆,林腾昌,等.国内转炉煤气回收概况与研究展望[J].铸造技术,2018,39(1):240.(YU P F, ZENG J Q, LIN T C, et al. Overview and research prospects of converter gas recovery in China[J]. Foundry Technology, 2018,39(1):240.) [2] 上官方钦,干磊,周继程,等.钢铁工业副产煤气资源化利用分析及案例[J].钢铁,2019,54(7):114.(SHANGGUAN F Q, GAN L, ZHOU J C, et al. Analysis and case study on the resource utilization of byproduct gas in the iron and steel industry[J]. Iron and Steel, 2019,54 (7): 114.) [3] 张琦,蔡九菊,王建军,等.钢铁厂煤气资源的回收与利用[J].钢铁,2009,44(12):95.(ZHANG Q, CAI J J, WANG J J, et al. Recovery and utilization of gas resources in steel mills [J]. Iron and Steel, 2009,44 (12):95.) [4] 张福明,张德国,张凌义,等.大型转炉煤气干法除尘技术研究与应用[J].钢铁,2013,48(2):1.(ZHANG F M, ZHANG D G, ZHANG L Y, et al. Research and application of dry dust removal technology for large converter gas[J]. Iron and Steel, 2013,48 (2):1.) [5] 张立峰.炼钢技术的发展历程和未来展望(Ⅱ):炼钢的未来展望[J].钢铁,2023,58(1):1.(ZHANG L F. Development history and future prospects of steelmaking technology (II): Future prospects of steelmaking[J]. Iron and Steel, 2023,58 (1):1) [6] 韩仁德,龙次考,黄磊.湛钢350 t转炉煤气回收潜力分析[J].冶金能源,2023,42(2):49.(HAN R D,LONG C K,HUANG L. Analysis of converter gas recovery potential of 350 t converter in Zhansteel[J]. Energy for Metallurgical Industry,2023,42(2):49.) [7] 张宁,吴浩.唐钢新区提高转炉煤气回收措施[J].冶金能源,2022,41(3):45.(ZHANG N,WU H. Measures for improving converter gas recovery in Tanggang new area[J]. Energy for Metallurgical Industry, 2022,41(3):45.) [8] 文旭林,彭澜欣,梁超松,等.柳钢控产期间煤气平衡措施[J].冶金能源,2023,42(4):13.(WEN X L,PENG L X,LIANG C S,et al.Gas balance countermeasure during production control in Liuzhou Steel[J]. Energy for Metallurgical Industry, 2023,42(4):13.) [9] 肖冬峰,杨春节,宋执环.基于改进BP网络的高炉煤气发生量预测模型[J].浙江大学学报(工学版),2012,36(11):2103.(XIAO D F, YANG C J, SONG Z H. A prediction model for blast furnace gas production based on improved BP network [J]. Journal of Zhejiang University (Engineering), 2012,36 (11):2103.) [10] ZHAO J,LIU, Q,PEDRYCZ W,et al. Effective noise estimation-based online prediction for byproduct gas system in steel industry[J]. IEEE Transactions on Industrial Informatics,2012,8(4):953. [11] 熊文真,赵娜.基于ARIMA-MC模型钢铁企业高炉煤气发生量预测[J].工业安全与环保,2016,42(10):100.(XIONG W Z, ZHAO N. Prediction of blast furnace gas production in steel enterprises based on ARIMA-MC model[J]. Industrial Safety and Environmental Protection, 2016,42 (10):100.) [12] 李志刚,纪月,任雄朝.基于LSTM与ARIMA组合模型的高炉煤气产生量预测[J].铸造技术,2018,39(11):2456.(LI Z G, JI Y, REN X C. Prediction of blast furnace gas production based on the combination model of LSTM and ARIMA [J]. Foundry Technology, 2018,39 (11):2456.) [13] 包向军,翁思浩,陈光,等.基于时序模型的高炉煤气发生量多步预测对比[J].钢铁,2022,57(9):166.(BAO X J, WENG S H, CHEN G, et al. Comparison of multi-step prediction of blast furnace gas generation based on time series models[J]. Iron and Steel, 2022,57 (9):166.) [14] 张琦,李鸿亮,赵晓宇,等.高炉煤气产生量与消耗量动态预测模型及应用[J].哈尔滨工业大学学报,2016,48(1):101.(ZHANG Q, LI H L, ZHAO X Y, et al. Dynamic prediction model and application of blast furnace gas production and consumption[J]. Journal of Harbin Institute of Technology, 2016,48 (1):101.) [15] 贺东风,官竹林,胡正彪.基于分类和ARIMA-WT-LSTM模型的高炉煤气产生量预测[J].冶金自动化,2022,46(2):103.(HE D F, GUAN Z L, HU Z B. Prediction of blast furnace gas production based on classification and ARIMA-WT-LSTM model[J]. Metallurgical Automation, 2022,46 (2):103.) [16] 李红娟,王建军,王华,等.建立PNN-HP-ENN-LSSVM模型预测钢铁企业高炉煤气发生量[J].过程工程学报,2013,13(3):451.(LI H J, WANG J J, WANG H, et al. Establishing PNN-HP-ENN-LSSVM model to predict blast furnace gas production in iron and steel enterprises[J]. Journal of Process Engineering, 2013,13 (3):451.) [17] 汤晓燕. 基于炼钢节奏估计的转炉煤气发生量长期预测[D]. 大连:大连理工大学,2014.(TANG X Y. Long Term Prediction of Converter Gas Generation Based on Steelmaking Rhythm Estimation[D]. Dalian:Dalian University of Technology, 2014) [18] 韩旭. 基于深度学习的转炉煤气产消流量预测与应用[D]. 大连:大连理工大学,2022.(HAN X. Prediction and Application of Converter Gas Production and Consumption Flow Based on Deep Learning[D]. Dalian:Dalian University of Technology, 2022.) [19] 费佳杰,吴定会,范俊岩,等.基于生产间歇改进Elman的转炉煤气发生量预测[J/OL].系统仿真学报: 1[2023-08-24].(FEI J J, WU D H, FAN J Y, et al. Prediction of converter gas generation based on production batch improvement Elman [J/OL]. Journal of System Simulation: 1[2023-08-24].) [20] 李红娟,熊文真.钢铁企业副产煤气预测及优化调度[J].钢铁,2016,51(8):90.(LI H J, XIONG W Z. Prediction and optimized scheduling of by-product gas in steel enterprises [J]. Iron and Steel, 2016,51(8):90.) [21] 施琦,赵贤聪,白皓,等.钢铁企业副产煤气短周期优化调度模型[J].钢铁,2016,51(8):81.(SHI Q, ZHAO X C, BAI H, et al. Short cycle optimization scheduling model for by-product gas in steel enterprises[J]. Iron and Steel, 2016,51(8):81.) [22] 周双.PLC控制系统在转炉烟气干法除尘及回收系统中的应用[J].新型工业化,2019,9(11):59.(ZHOU S. Application of PLC control system in dry dust removal and recovery system of converter flue gas[J]. New Industrialization, 2019, 9(11):59.) [23] 李雄.干法除尘风机控制系统在涟钢210 t转炉的应用[J].新型工业化,2020,10(6):16.(LI X. Application of dry dust removal fan control system in Liangang 210 t converter[J]. New Industrialization, 2020,10(6):16.) [24] 张新建.基于吹炼特性的转炉LT法炉口微差压控制研究方法[J].冶金自动化,2020,44(2):43.(ZHANG X J. Research method for micro differential pressure control of converter LT method based on blowing characteristics[J]. Metallurgical Industry Automation, 2020,44(2):43.) [25] 李春香.提高不锈钢转炉煤气回收热值和回收量的实践[J].冶金动力,2020(6):20.(LI C X. Practice of improving the heat value and recovery capacity of stainless steel converter gas recovery[J]. Metallurgical Power, 2020(6):20.) [26] 李末卓,张军国,戴雨翔,等.唐钢新区200 t转炉高效冶炼生产实践[J].中国冶金,2023,33(3):117.(LI M Z, ZHANG J G, DAI Y X, et al. High efficiency smelting production practice of 200 t converter in Tanggang New Area[J]. China Metallurgy, 2023,33(3):117.) [27] 王杰,曾加庆,杨利彬,等.转炉炼钢过程的精细化控制及协同优化[J].钢铁,2022,57(5):55.(WANG J, ZENG J Q, YANG L B, et al. Fine control and collaborative optimization of converter steelmaking process[J]. Iron and Steel, 2022,57(5):55.) [28] 陈均.干法除尘工艺在200 t半钢炼钢转炉上的应用[J].钢铁,2016,51(7):89.(CHEN J. Application of dry dust removal process on a 200 t half steel steelmaking converter[J]. Iron and Steel, 2016,51(7):89.) [29] WANG Y J,ZHU Z Y,SHA A X, et al. Low cycle fatigue life prediction of titanium alloy using genetic algorithm-optimized BP artificial neural network[J]. International Journal of Fatigue,2023,172:1. [30] 赵路朋,吴铿,朱利,等.基于BP神经网络的烧结矿性能预报模型[J].钢铁,2017,52(9):11.(ZHAO L P, WU K, ZHU L, et al. A prediction model for sintering ore performance based on BP neural network[J]. Iron and Steel, 2017,52 (9):11.) [31] HAN D Y,LIU P,XIE K, et al. An attention-based LSTM model for long-term runoff forecasting and factor recognition[J]. Environmental Research Letters,2023,18(2). [32] YANG X,LI Y M,YU X X, et al. Regional/Single station zenith tropospheric delay combination prediction model based on radial basis function neural network and improved long short-term memory[J]. Atmosphere,2023,14(2). [33] 边海涛,杨荃,钟恬,等.基于RBF神经网络的轧辊偏心补偿控制[J].钢铁,2007,42(11):42.(BIAN H T, YANG Q, ZHONG T, et al. Roll eccentricity compensation control based on RBF neural network[J]. Iron and Steel, 2007,42(11):42.)