摘要
在高炉炼铁生产过程中,铁水硅含量反映高炉炉温,预测和控制炉温对高炉生产的节能、降耗、顺行至关重要.基于包钢6号高炉生产数据,建立了RBF神经网络铁水硅含量预测模型.研究表明:考虑时滞因素的RBF神经网络模型,当误差范围<±0.10时,预报准确率达到了85%,其准确度高于不考虑时滞因素的RBF神经网络模型,对在线预测高炉铁水硅含量具有实用价值.
In blast furnace iron making process,the silicon content in hot metal is usually used to measure the furnace temperature.The control and prediction of furnace temperature is vitally important for energy conservation and cost reduction.Based on data collected from NO.6 blast furnace of Baotou Steel,a RBF(Radial Basis Function)neural network was developed and employed in the prediction of silicon content in hot metal of blast furnaces.The results show that the RBF model acquires better results if time delay is taken into consideration,and prediction accuracy is higher than that of without time delay.Prediction accuracy of RBF model with time delay reached 85% within the permissible error ±0.10.The RBF neural network prediction model with time delay is useful in online monitoring the silicon content in hot metal.
出处
《内蒙古大学学报（自然科学版）》
CAS
CSCD
北大核心
2012年第2期188-191,共4页
Journal of Inner Mongolia University：Natural Science Edition
基金
国家自然科学基金资助项目(51064019)
教育部"春晖计划"合作项目(Z2009-1-01053)
内蒙古自然科学基金资助项目(2010MS0911)
关键词
铁水硅含量
炉温预测
RBF神经网络
时滞
silicon content in hot metal
temperature prediction
RBF neural network
time delay
