对烧结矿FeO含量预测的数学模型研究  

A study on mathematical model of prediction of FeO content in the sintering ore

作　　者：张帆[1] 魏国[1] 庞巍[1,2] 高强健[1] 陈伟亮[1] 温秋林[1] 杜钢[1] 

机构地区：[1]东北大学材料与冶金学院,沈阳110819 [2]鞍钢集团矿业公司东烧厂,辽宁鞍山114000

出　　处：《材料与冶金学报》2013年第3期159-162,168,共5页Journal of Materials and Metallurgy

基　　金：国家自然科学基金资助项目(51074040);(51074206)

摘　　要：分析烧结生产中影响烧结矿FeO含量的众多因素,选择碱度、配煤量、一次温度、制粒效果、加水量、料层厚度、点火温度、煤气流量等8个工艺参数以及4种矿粉配比作为FeO含量预报模型的输入变量.分别采用BP神经网络、RBF神经网络、SVM 3种进行建模预测.预测结果表明,SVM预测性能优于BP神经网络,RBF神经网络优于SVM.Many factors affecting FeO content in sinter were analyzed and the eight technological parameters such as basicity,blending coal ratio,first temperature,pelleting effect,adding water volume,bed depth,ignition temperature and coal gas flow and proportion of the four kinds of fine iron ore were chosen as the input variables to form a model to predict FeO content in the sintering ore.Three algorithms: BP neural network,RBF neural network,and SVM were selected.The results showed that predicting performance of SVM is better than that of BP neural network and predicting performance of RBF neural network is better than that of SVM.

关 键 词：FeO含量预测 BP SVM RBF 

分 类 号：TF046.4[冶金工程—冶金物理化学]