转炉冶炼工艺中合金加入量的确定及NN建模  

Determine and NN Modeling of Alloy Addition Amount in Converter Smelting

作　　者：肖纯[1] 张毅[2] 徐卫平[3] 郑家良[2] 

机构地区：[1]贵州师范大学材料与建筑工程学院,贵州贵阳550059 [2]水城钢铁(集团)有限责任公司炼钢厂,贵州水城553028 [3]贵州师范大学机电工程学院,贵州贵阳550059

出　　处：《冶金设备》2009年第3期13-17,共5页Metallurgical Equipment

基　　金：贵州省科技基金(黔科合J字(2009)2054号);水城钢铁(集团)有限责任公司(技字200810号)资助项目

摘　　要：简述炼钢冶炼工艺的特点及其复杂性,分析了冶炼工艺的决策过程,建立了转炉冶炼工艺中硅铁、硅锰合金加入量的神经网络(NN)模型。针对冶炼工艺数据的特点,研究了神经网络输入输出数据的预处理方法,引入基于对数变换的数据预处理算法,测试表明效果较优。所建立的NN模型能真实反映转炉冶炼的工艺特点,模型值和实测值相差较小,能实现在给定成品成分要求下合金加入量的自动选择。The features and the complexity of the steel smelting process are briefly introduced in this paper. Analysed the decision-making process, the neural networks （NN） model of Si-Fe, Si-Mn alloy addition in converter smelting process is established. According to the characteristics of process data, input and output data pre-pro- cessing methods are discussed, and logarithmic transform algorithms is introduced in the pre-processing which showed good effect. The little difference between calculation and measured values showed that NN model can truly reflect the characteristics of the smelting process, and automaticly choose the request of alloy addition for a given product ingredients.

关 键 词：建模 神经网络 转炉冶炼 合金加入量 数据预处理 

分 类 号：TF711[冶金工程—钢铁冶金]