基于BP神经网络的连铸板坯裂纹预报  

Crack Forecast with Continuous Casting Based on BP Neural Network

作　　者：孙洁[1] 王俊霞[1] 

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

出　　处：《河北理工大学学报（自然科学版）》2009年第2期67-71,共5页Journal of Hebei Polytechnic University:Social Science Edition

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、钢液C,S,P含量等数据实现板坯表面裂纹预报。The causes of continue cast slab crack are analyzed.According to the feature and crafts condition of continuous casting,combined with the prevental countermeasures of board in some national and foreign advanced iron and steel plants,adopt with the BP Neural network,Using neural networks and expert systems integration approach,thermocouple temperature data,combined with liquid crystal,powder,steel,steel of C,S,P,and other data content to achieve slab surface cracks forecast.

关 键 词：连铸 裂纹 专家系统 神经网络 

分 类 号：TP338.6[自动化与计算机技术—计算机系统结构]