摘要
分析了原燃料条件较差、炉况稳定程度较低和检测仪表不多的湘潭钢铁公司２号高炉铁水含硅量数据的统计学特征。按照现代控制论的观点，将高炉视作多输入－单输出系统。当系统处于相对稳定的状态时，可以用若干个操作变量和模型预报偏差的多项式预报下一铁次的含硅量；根据赤池法则确定模型中包括的自变量及其阶数。应用该模型对９１５炉铁次的含硅量作离线预报研究，结果证明了模型的适用性。
s:The statistical characteristics of hot metal silicon data were concluded for Xianggang BF2 which was fed with raw materials of poor qualities and equipped with insufficient measuring sensors.The blast furnace was considered as a multi inputsingle output system in accordance with the contemporary control theory,and the silicon content of the coming cast might be forecast with a time series model under conditions when the furnace was running relatively smoothly.The Akaike criterion was applied to the model identification.This model has been used off line in hot metal silicon prediction for 915 casts,and as a result,its usefulness has been justified.
关键词
高炉
铁水
预报
数学模型
硅含量
blast furnace
hot metal silicon
prediction
mathematical model
