基于RBF神经网络的转炉冶炼中低碳铬铁终点磷含量预报模型的研究  

Research on Prediction Model of End-point Phosphorus Converter for Content Smelting Medium-Low Carbon Ferrochrome Based on RBF Neural Network

作　　者：邱东[1] 戴文娟[1] 

机构地区：[1]长春工业大学电气与电子工程学院,长春130012

出　　处：《计算机测量与控制》2014年第9期3020-3023,共4页Computer Measurement &Control

基　　金：吉林省科技发展计划项目(20120420)

摘　　要：中低碳铬铁冶炼工艺复杂,杂质磷含量的高低是影响铬铁产品质量的重要因素;通过研究转炉冶炼中低碳铬铁铁水脱磷预处理的反应特性及热力学条件,分析了影响中低碳铬铁合金终点磷含量的重要因素;基于中钢吉铁辽阳公司转炉冶炼中低碳铬铁的生产工艺及样本数据,建立了基于RBF人工神经网络的转炉冶炼中低碳铬铁终点磷含量预报模型,实现了对冶炼过程终点中低碳铬铁磷含量的在线预报与分析;仿真结果表明,该模型预报精度在±0.003%范围内命中率达到85.7%,为改进冶炼工艺、提高产品质量提供了重要的理论依据。Medium-Low Carbon ferrochrome smelting is a complex process. The discretion of the impurity content of phosphorus is an important factor affecting the quality of Ferrochrome products. The important influence factors of converter smelting ferrochrome end-point phosphorus content was concluded based on analyzing response characteristics of hot metal dephosphorization pretreatment and thermodynamic condition, the control variables of the end phosphorus content was fixed; According to Sinosteel Jilin Ferroalloys co. , Ltd converter smelting medium and low carbon ferrochrome technology and production data, a prediction model for converter smelting ferrochrome end-point phosphorus content has been established based on RBF artificial neural network in accordance with the ferrochrome smelting process for online prediction of end phosphorus content. Results show that the hit rate of the prediction model is 85.7% with the error ±0. 003%, provides important theoretical basis for the improvement of smelting process and product quality.

关 键 词：RBF神经网络 转炉 终点磷含量 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]