基于神经网络的单种铁矿粉烧结基础特性预报模型  

A Single Kind of Iron Ore Sintering Basic Characteristics Prediction Model Based on Neural Network

作　　者：李小成[1] 李福民[1] 陈丽芳[2] 吕庆[1] 

机构地区：[1]河北联合大学冶金与能源学院,河北唐山063009 [2]河北联合大学理学院,河北唐山063009

出　　处：《河北联合大学学报（自然科学版）》2014年第3期16-18,共3页Journal of Hebei Polytechnic University:Social Science Edition

摘　　要：利用神经网络可以以任意精度逼近任意复杂非线性函数的特点，建立了铁矿粉烧结基础特性的预报模型，并对常用铁矿粉中的10种矿的同化性、液相流动性和粘结相强度进行了预报，模型的准确度达到了75%~90%。该预报模型的构建对于烧结配矿决策、提高烧结矿质量具有指导意义。Using the characteristics that neural networks can approximate any arbitrary precision with any complex nonlinear function,a model was established to predict the iron ore sintering basic characteristics,and then forecast the assimilation,the flow of liquid and the strength of the binding phase,the accuracy achieved 75% ~ 90% . This prediction model has guiding significance for sinter ore decision and improving the quality of sinter.

关 键 词：铁矿粉 化学成分 烧结基础特性 神经网络 

分 类 号：TF046.4[冶金工程—冶金物理化学]