摘要
针对鞍钢铁水罐喷吹CaO+Mg粉剂复合脱硫过程,建立了基于BP神经网络的铁水预处理终点硫含量预报模型。用鞍钢的1 000炉数据进行模型训练,经60炉数据验证表明,有5%的炉次预报值与实际值完全一致,有76.67%的炉次误差≤0.003%,平均误差为0.002 5%。
Based on the productive practice of CaO + Mg powder co- injection in Ansteel, the prediction model of final sulfur content for hot metal pretreatment was established. The data of 1 000 heats were used to train the model and among others 60 heats were randomly selected as the test samples respectively. The results shown that, the predicted values of the 5% heats of the total test heats were the same as the actual values, the absolute error of the 76.67% heats was less than 0.003%, and the average error was 0.002 5%.
出处
《金属材料与冶金工程》
CAS
2009年第3期58-61,共4页
Metal Materials and Metallurgy Engineering
关键词
铁水
神经网络
硫含量
预报
模型
hot metal
neural network
sulfur content
prediction
model