作者：周翼男, 崔桂梅, 皮理想, 刘伟, 王东旭
作者单位：内蒙古科技大学信息工程学院流程工业综合自动化重点实验室，内蒙古 包头 014010
关键词：转炉炼钢;贝叶斯优化算法;梯度提升决策树;终点预测
摘要：
为提高转炉炼钢终点碳含量和温度预报精度，提出基于贝叶斯优化梯度提升决策树（boa_gbdt）的转炉炼钢终点碳含量和温度预测模型，将其与基础模型径向基函数（rbf）、支持向量机（svm）、梯度提升决策树（gbdt）以及贝叶斯优化的径向基函数（boa_rbf）、支持向量机（boa_svm）终点碳温预测模型对比分析。实验结果表明：boa_gbdt各项误差指标最小，命中率最高，终点时刻碳含量在±0.01%误差区间内命中率为96.2%；终点温度在±10 ℃误差区间内命中率为92.1%。贝叶斯优化算法能够显著提升模型性能，更准确地判断转炉炼钢终点碳含量和温度，为吹炼出符合要求的钢水提供较为可靠的依据。
prediction of converter steelmaking end point based on bayesian optimization gbdt
zhou yi'nan, cui guimei, pi lixiang, liu wei, wang dongxu
key laboratory of comprehensive automation of process industry, school of information engineering, inner mongolia university of science and technology, baotou 014010, china
abstract: in order to improve the prediction accuracy of end point carbon content and temperature in converter steelmaking, a prediction model of end point carbon content and temperature in converter steelmaking based on bayesian optimized gradient lifting decision tree (boa_gbdt) was proposed. it is compared and analyzed with the end-point carbon temperature prediction models of the base model radial basis function(rbf), support vector machine (svm), gradient boosting decision tree (gbdt) and bayesian algorithm optimized radial basis function(boa_rbf), support vector machine (boa_svm) end point carbon temperature prediction model. the experimental results show that boa_gbdt has the smallest error index and the highest hit rate. the hit rate of carbon content at the end point is 96.2% within the error interval of ±0.01%; the hit rate of the end temperature is 92.1% within the error interval of ±10 ℃. bayesian optimization algorithm can significantly improve the performance of the model, more accurately judge the end-point carbon content and temperature of converter steelmaking, and provide a more reliable basis for converting qualified molten steel.
keywords: converter steelmaking;bayesian optimization algorithm;gradient boosting decision tree;endpoint prediction
2024, 50(7):33-39 收稿日期: 2022-07-24;收到修改稿日期: 2022-09-29
基金项目: 国家自然科学基金项目（61763039）
作者简介: 周翼男（1997-），女，吉林长春市人，硕士研究生，专业方向为转炉终点预测建模与优化控制。
参考文献
[1] 曾鹏飞, 刘辉. 基于二次相似性度量的即时学习转炉炼钢终点碳温软测量方法[j]. 计算机集成制造系统, 2021, 27(5): 1429-1439.
zeng p f, liu h. soft-sensing method for end-point carbon temperature of converter steelmaking based on quadratic similarity measurement[j]. computer integrated manufacturing systems, 2021, 27(5): 1429-1439.
[2] 李超, 刘辉. 改进mtbcd火焰图像特征提取的转炉炼钢终点碳含量预测[j/ol]. 计算机集成制造系统: 1-22 [2022-05-30]. http://kns.cnki.net/kcms/detail/11.5946.tp.20210428.1806.020.html.
li c, liu h. carbon content prediction of converter steelmaking end-point based on improved mtbcd flame image feature extraction[j/ol]. computer integrated manufacturing systems: 1-22 [2022-05-30]. http://kns.cnki.net/kcms/detail/11.5946.tp.20210428.1806.020.html.
[3] 熊倩, 刘辉, 刘旭琛. 基于lnn-dpc加权集成学习的转炉炼钢终点碳温软测量方法[j/ol]. 计算机集成制造系统: 1-18 [2022-05-30]. http://kns.cnki.net/kcms/detail/11.5946.tp.20201202.1719.007.html.
xiong q, liu h, liu x c. soft measurement method of endpoint carbon content and temperature of converter steelmaking based on lnn-dpc weighted ensemble learning[j/ol]. computer integrated manufacturing systems: 1-18 [2022-05-30]. http://kns.cnki.net/kcms/detail/11.5946.tp.20201202.1719.007.html.
[4] 徐钢, 黎敏, 徐金梧, 等. 基于函数型数字孪生模型的转炉炼钢终点碳控制技术[j]. 工程科学学报, 2019, 41(4): 521-527.
xu g, li m, xu j w, et al. control technology of end-point carbon in converter steelmaking based on functional digital twin model[j]. chinese journal of engineering, 2019, 41(4): 521-527.
[5] 张彩军, 韩阳, 何世宇, 等. 炉口火焰光谱驱动的炼钢终点控制[j]. 仪器仪表学报, 2018, 39(1): 24-33.
zhang c j, han y, hu s y, et al. furnace mouse flame spectrum driven steelmaking end control[j]. chinese journal of scientific instrument, 2018, 39(1): 24-33.
[6] 谢书明, 柴天佑, 陶钧. 一种转炉炼钢动态终点预报的新方法[j]. 自动化学报, 2001(1): 136-139.
xie s m, chai t y, tao j. a kind of newmethodfor ld dynamicendpoint prediction[j]. acta automatica sinica, 2001(1): 136-139.
[7] cox i j , lewis r w , ransing r s , et al. application of neural computing in basic oxygen steelmaking[j]. journal of materials processing tech, 2002, 120(1-3): 310-315.
[8] cemalettin k, harun t, recep a, et al. bofy-fuzzy logic control for the basic oxygen furnace (bof)[j]. robotics and autonomous systems, 2004, 49(3-4): 193-205.
[9] han m, li y, cao z j. hybrid intelligent control of bof oxygen volume and coolant addition[j]. neurocomputing, 2014, 123(123): 415-423.
[10] 周志华. 机器学习[m]. 北京: 清华大学出版社, 2016: 170-173.
zhou z h. machine learning[m]. beijing: tsinghua university press, 2016: 170-173.
[11] friedman j h. greedy function approximation: a gradient boosting machine[j]. annals of statistics, 2001, 29(5): 1189-1232.
[12] 崔佳旭, 杨博. 贝叶斯优化方法和应用综述[j]. 软件学报, 2018, 29(10): 3068-3090.
cui j x, yang b. survey on bayesian optimization methodology and applications[j]. journal of software, 2018, 29(10): 3068-3090.
[13] shahriari b , swersky k , wang z , et al. taking the human out of the loop: a review of bayesian optimization[j]. proceedings of the ieee, 2015, 104(1): 148-175.
[14] wu j , chen x y , zhang h , et al. hyperparameter optimization for machine learning models based on bayesian optimization[j]. 电子科技学刊: 英文版, 2019, 17(1): 15.
[15] rasmussen c e, williams c k i. gaussian processes for machine learning[j].intl. journal of neural systems, 2006, 103(14): 429.
[16] 张豹, 刘琼, 吴细宝, 等. 基于集成学习的涡扇发动机剩余寿命预测[j]. 中国测试, 2022, 48(7): 47-52.
zhang b, liu q, wu x b, et al. remaining useful lifetime prediction of turbofan engine based on ensemble learning[j]. china measurement ＆ test, 2022, 48(7): 47-52.
[17] 崔桂梅, 刘伟, 张帅, 等. 基于差分进化支持向量机的轧制力预测[j]. 中国测试, 2021, 47(8): 83-88.
cui g m, liu w, zhang s, et al. rolling force prediction based on differential evolution support vectormachine[j]. china measurement ＆ test, 2021, 47(8): 83-88.
