BP神经网络模型在铁水脱磷中的应用  

The Application of Artificial Neural Network in Hot Metal Dephosphorization

作　　者：郭湛[1] 王志明[1] 周庆成[1] 郭上型[2] 

机构地区：[1]马鞍山钢铁股份有限公司 [2]安徽工业大学

出　　处：《安徽冶金》2005年第4期41-43,共3页Anhui Metallurgy

摘　　要：使用转炉渣配制的脱磷剂对铁水进行预脱磷,并建立BP神经网络模型,通过对该网络模型的检验,发现:模型的预报值与实验值吻合较好;转炉渣剂合适加入量约为10%;Fe2O3代替转炉渣的合适替代量约为5%;BaCO3代替转炉渣的合适替代量约为10%;脱磷剂中(P2O5)含量的增加会导致脱磷率的降低。Dephosphorizing agent made of converter slag is used to perform a hot metal pre-dephosphorization,and a BP artificial neural network is set up.It is found through testing the network model that the forecasted value meets the experimental value well,the suitable addition of converter slag agent is about 10%,the suitable addition is about 5% if Fe_2O_3 is used instead of converter slag,the suitable addition is about 10% if BaCO_3 is used and dephosphorization rate decreases with the increase of P_2O_5 in dephosphorizing agent.

关 键 词：BP神经网络 转炉渣 铁水 脱磷 

分 类 号：TF704.4[冶金工程—钢铁冶金]