基于BP神经网络的连铸板坯裂纹预报  

Crack Forecast of Continuous Casting Slab based on BP Neural Network

作　　者：孙洁[1] 王俊霞[1] 王沛利 

机构地区：[1]河北理工大学计算机与自动控制学院,河北唐山063009 [2]河北唐山建工中专,河北唐山063004

出　　处：《铸造技术》2009年第4期471-473,共3页Foundry Technology

摘　　要：分析了连铸板坯裂纹的成因,并根据连铸生产的特点和工艺条件,结合国内外一些先进钢铁厂板坯裂纹的防止对策,引入BP神经元网络,采用神经网络与专家系统相结合的方法,运用热电偶温度数据,结合结晶器液面、保护渣、钢号以及钢液中C、S、P含量等数据,实现板坯表面裂纹预报。The causes of continuous casting slab crack are analyzed. Based on the production characteristics and technical conditions of continuous casting along with the prevention countermeasure of national and international iron and steel plants, BP Neural network is adopted in this investigation. Through integration of neural networks and expert systems, the slab surface crack can be forecasted by temperature, liquid level in the crystallizer, preventing slag, steel grade and C, S, P content.

关 键 词：连铸板坯 裂纹 专家系统 神经网络 

分 类 号：TF777[冶金工程—钢铁冶金]