摘要
通过研究转炉冶炼终点磷、硫含量的影响因素.确定了影响冶炼终点的控制变量,根据人工神经网络技术,对常用BP算法进行改进,建立了基于神经网络的转炉冶炼终点双节点输出模型,实现了对终点钢水磷、硫含量同时进行预报,选取现场实际生产数据为样本,对模型进行训练,使模型对磷、硫含量的预报误差在±0.003%的命中率均达到了85%以上,预报精度达到了炼钢工艺的要求。
According to the research of the factors of End-point Contents phosphor and sulfur in Converter, the dominative variable of the Model of End-point for Converter smelting is fixed. According to the improvement of artificial neural network technic, this article establishes the Model for P and S of End-point for Converter smelting Based on Neural Network. Produced data are chosen as the sample and the model is trained to make it approach to the prediction of dynamic control.
出处
《河北理工大学学报（自然科学版）》
CAS
2008年第3期26-29,40,共5页
Journal of Hebei Polytechnic University:Social Science Edition
关键词
转炉冶炼
神经网络
终点成分
预报模型
converter smelting
neural network
end point predcting content
forecast model
