摘要
针对高炉初渣、中间渣组分多变特别是FeO含量高等问题,系统研究了CaO-SiO2-Al2O3-FeO-MgO五元渣系的黏度及组分对黏度的影响规律,并建立了基于WEB的神经网络-遗传算法(ANN-GA)系统的高炉渣黏度预报模型。结果表明,该模型对高FeO渣系的黏度预报值与试验结果吻合较好,误差在20%以内。通过模型预报获得的各因素对渣黏度影响的规律与文献及试验结果一致。
The primary slag and the intermediate slag of the blast furnace had complex compositions and were especially rich in FeO.Aiming at these problems,the viscosity of the CaO-SiO2-Al2O3-FeO-MgO melt was investigated as well as the influencing laws of each component to viscosity.Then a viscosity prediction model was established on the basis of WEB-based neural network-genetic algorithm（ANN-GA） system.Verification result of this model indicates that the error of this model is basically within 20% which is acceptable.The univariate analyses attained by the ANN-GA system are well consistent with literature results.
出处
《钢铁》
CAS
CSCD
北大核心
2012年第7期20-25,共6页
Iron and Steel
关键词
高炉
初渣
中间渣
黏度预报
神经网络
blast furnace
primary slag
intermediate slag
viscosity prediction
ANN-GA
