基于集成学习的脱硫剂加入量预测方法  

Method for predicting desulfurizer dosage based on ensemble learning

作　　者：方一飞 但斌斌[1] 吴经纬[2] 容芷君[1] 都李平 罗钟邱 FANG Yifei;DAN Binbin;WU Jingwei;RONG Zhijun;DU Liping;LUO Zhongqiu(Key Laboratory of Metallurgical Equipment and Control of Ministry of Education,Wuhan University of Science and Technology,Wuhan 430081,China;Intelligent Manufacturing Division,WISDRI Engineering and Research Incorporation Limited,Wuhan 430223,China;Baoxin Software(Wuhan)Co.,Ltd,Wuhan 430080,China)

机构地区：[1]武汉科技大学冶金装备及其控制教育部重点实验室,湖北武汉430081 [2]中冶南方工程技术有限公司智能制造事业部,湖北武汉430223 [3]宝信软件(武汉)有限公司,湖北武汉430080

出　　处：《武汉科技大学学报》2024年第5期361-367,共7页Journal of Wuhan University of Science and Technology

基　　金：国家自然科学基金项目(51475340);湖北省重点研发计划项目(YFXM2022000556).

摘　　要：为解决铁水预脱硫过程中脱硫剂加入量控制问题,提出一种基于集成学习的脱硫剂加入量预测方法。首先,对原始数据进行预处理,将空值、重复值、0值以及不符合工艺规范的数据行删除,并使用LOF算法结合专家经验剔除异常值;其次,基于GBDT算法计算每个输入特征的重要性占比,进行特征筛选;最后,采用Optuna超参数自动寻优框架对预测模型调优,寻找最佳超参数组合,预测脱硫剂加入量。利用某钢厂铁水预处理过程中的实际生产数据,分别采用XGBoost、RF、GBDT以及LightGBM等方法构建预测模型并进行对比试验。其中XGBoost模型的拟合精度(R^(2))、均方根误差(RMSE)、平均绝对误差(MAE)以及平均绝对百分比误差(MAPE)分别为0.8962、198.245、119.726以及7.897%,相较于其它模型均是最优。To control the amount of desulfurizer added in the pre-desulfurization process of hot metal,a method for desulfurizer dosage prediction based on ensemble learning was proposed.Firstly,the original data was preprocessed,null values,duplicate values,0 values and data lines that did not meet the process specifications were deleted,and the outliers were eliminated using local outliers factor(LOF)algorithm combined with expert experience.Secondly,the importance ratio of each input feature was calculated based on gradient boosting decision tree(GBDT)algorithm,and the feature was screened.Finally,Optuna was used to optimize the prediction model,find the best combination of hyperparameters,and predict the amount of desulfurizer added.Based on the actual production data of a steel mill,the prediction models were constructed by using extreme gradient boosting(XGBoost),random forest(RF),GBDT and light gradient boosting machine(LightGBM),etc.The coefficient of determination(R^(2)),root mean square error(RMSE),mean absolute error(MAE)and mean absolute percentage error(MAPE)of XGBoost model are 0.8962,198.245,119.726 and 7.897%,respectively,which are the best compared with other models.

关 键 词：脱硫剂加入量 铁水预脱硫 局部异常因子 Optuna算法 极限梯度提升树 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]