基于神经网络的高炉铁水硅和硫含量预报模型  

PREDICTION MODEL BASED ON NERRAL NET FOR SILICON AND SULFUR CONTENTS IN HOT METAL

作　　者：于卓颖[1] 郑涛[1] 

机构地区：[1]唐山国丰钢铁有限公司,河北唐山063300

出　　处：《河北冶金》2015年第3期38-41,共4页Hebei Metallurgy

摘　　要：采用基于自学习的参考炉次法,建立了反应高炉炉温和铁水质量的预报模型,对炼铁过程铁水硅含量和硫含量进行预报,建立了基于BP神经网络的高炉铁水硅含量和硫含量预报模型。用国内某高炉的生产数据进行模型训练,经预报结果数据验证,想要通过现有直接获取的高炉参数很难准确同时预报铁水硅含量和硫含量,但基本能准确预报铁水硫含量的变化趋势。Adopting referring heats method based on self-study,a prediction model is built for temperature of blast furnace and hot metal quality. With it the silicon and sulfur contents in hot metal in iron-making can be predicted,and with those a prediction model based on BP neural net is built for silicon and sulfur contents of blast furnace hot metal. The model is trained with the production data of some blast furnace at home.It is showed from the data check that it is difficult to accurately predict silicon and sulfur contents in hot metal at the same time with the directly-got blast furnace parameters,but the change trend of sulfur content in hot metal can be basically accurately predicted.

关 键 词：高炉过程 BP神经网络 铁水硅含量 铁水硫含量 预报模型 

分 类 号：TF513[冶金工程—钢铁冶金]