基于BP神经网络的薄板坯连铸连轧预报模型的建立  

Modeling of Prediction of Thin Slab Casting and Rolling Using BP Artificial Neural Network

作　　者：靳冰[1] 梁智学[1] 杨宏群[1] 

机构地区：[1]南阳理工学院,南阳473004

出　　处：《计算机与数字工程》2013年第6期1027-1029,共3页Computer & Digital Engineering

摘　　要：对某钢厂SPHC钢卷的碳、硅、锰、磷、硫的质量分数进行整理,获取力学性能参考数据(屈服强度,抗拉强度,延伸率),利用BP神经网络建立起其间的关系网络模型进行组织性能预报。研究表明:BP神经网络薄板坯连铸连轧性能预测模型具有很高的预测精度,可很好地描述多个元素含量变化下的薄板坯连铸连轧的复杂的非线性力学性能,解决了回归分析只能针对单个力学性能进行建模的问题。Reference data of（the yield strength, tensile strength, elongation ratio） mechanical properties were obtained through data sorting of C.Mn,Si,P,S of SPHC steel coil; their relationship network model was built by BP artificial neural network to do structure and property prediction. The results show that the built model of thin slab casting and rolling has certain accuracy, it can be described more ele- ments of the thin slab continuous casting variation in content of rolling of complex non-linear mechanics performance and solve the problem of single mechanical performance modeling of regression analysis.

关 键 词：BP神经网络 薄板坯连铸连轧 组织性能预报 

分 类 号：TG146.23[一般工业技术—材料科学与工程]