摘要
终点碳含量和温度的窄窗口控制是衡量转炉冶炼水平和实现智能制造的主要目标.本文基于某钢厂Q235钢种转炉实际生产数据,采用卡方分箱法并结合所划定的理想终点目标区间对转炉操作工艺参数进行了分箱,借助WOE编码判别确定了影响转炉理想终点目标的关键工艺参数,采用逻辑回归算法对编码后数据进行了评分.所构建的转炉操作工艺评价模型能够对操作工艺参数进行合理评分,所判别不符合转炉理想终点目标区间的数据召回率为84%、精准率为88%.此外,评价结果表明铁水碳含量及硅含量是影响该炼钢厂转炉终点目标的主要因素,因此,优化铁水条件有利于实现对转炉终点目标窄窗口控制之目的.
Narrow window control of endpoint carbon content and temperature of converter are the main target for measuring the refining level and achieving intelligent manufacturing.According to the practical data of Q235 steel obtained in converter production,the operation process parameters were clustered based on the chi-square boxing method combined with ideal target interval of endpoint carbon content and temperature.At the same time,the key process parameters affecting the ideal endpoint of the converter were determined by means of WOE code discrimination,and the coded data was scored by logistic regression algorithm.The evaluation model of converter operation process established in the present work can reasonably evaluate the operating process parameters,and the recall ratio of the converter production data that does not meet the ideal endpoint target interval is 84%,and the accuracy ratio of evaluation model is 88%.In addition,the evaluation result indicates that the carbon content and silicon content of hot metal are the main factors affecting the convert endpoint of this steel plant,and thus,optimizing the hot metal condition is proper to attain the purpose of narrow window control on the converter endpoint.
作者
陈超
王楠
于海洋
陈敏
Chen Chao;Wang Nan;Yu Haiyang;Chen Min(School of Metallurgy,Northeastern University,Shenyang 110819,China)
出处
《材料与冶金学报》
CAS
北大核心
2019年第2期87-91,共5页
Journal of Materials and Metallurgy
基金
十三五国家重点研发计划(2017YFB0304201
2017YFB0304203
2016YFB0300602)
关键词
转炉工艺参数
窄窗口控制
工艺模型
卡方分箱
逻辑回归
converter processing parameters
narrow window control
processing
chi-square boxing method
logistic regression
