基于大数据的高炉炉温监测预警系统  

Blast furnace temperature monitoring and early warning system based on big data

作　　者：刘小杰 张玉洁 李欣 刘然 张智峰 陈树军 LIU Xiao-jie;ZHANG Yu-jie;LI Xin;LIU Ran;ZHANG Zhi-feng;CHEN Shu-jun(School of Metallurgy and Energy,North China University of Science and Technology,Tangshan 063210,Hebei,China;Chengde Branch,HBIS Group Co.,Ltd.,Chengde 067000,Hebei,China)

机构地区：[1]华北理工大学冶金与能源学院,河北唐山063210 [2]河钢集团有限公司承德分公司,河北承德067000

出　　处：《中国冶金》2023年第2期98-105,140,共9页China Metallurgy

基　　金：国家自然科学基金青年基金项目(项目编号:52004096)。

摘　　要：随着新一轮产业变革和技术革命的兴起,钢铁制造业正在由高碳向低碳、由低端向高端转型升级。为实现高效率、低能耗、长寿命、低污染的综合目标,现代炼铁工艺逐渐趋于绿色化、智能化。高炉为一个非线性、大时滞的黑箱化系统,高温高压的环境使得高炉炉温的测量和控制都不易实现。利用铁水硅含量、铁水温度与高炉炉温的正相关性,建立基于大数据分析的铁水硅含量以及铁水温度预测模型,间接实现对炉温的预测。首先,利用经过异常值、缺失值以及归一化处理后的高炉标准数据数据集,通过多角度的相关性分析方法完成对模型输入变量的选取;然后,通过模型综合对比,建立基于Adaboost模型的铁水硅含量、铁水温度预测模型;最后,结合计算机技术建立高炉炉温监测预警系统。该系统的应用不仅有效解决了传统冶炼工艺带来的弊端,而且能起到延长设备生命周期、提前预测炉况走向等作用,有效推动了高炉智能化转型。With the rise of a new round for industrial and technical revolution,iron and steel manufacturing is transforming and upgrading from high-carbon to low-carbon and from lower end to higher end.In order to achieve the goal of high-efficiency,low-energy,long-life and low pollution,modern ironmaking process gradually leads to arenization and intelligentially.Blast furnace is nonlinear and long time-delay black-boxed system,its high temperature and pressure environment makes that measurement and control of furnace temperature is not easy to achieve.Based on the positive correlation between hot metal silicon content,hot metal temperature and blast furnace temperature,prediction models of hot metal silicon content and hot metal temperature based on big data analysis were established,which indirectly realized the prediction of furnace temperature.Firstly,the input variables of the models were selected using the blast furnace standard data set after the processing of outliers,missing values and normalization,and through multi angle correlation analysis.And then from contrasting different models comprehensively,prediction models of the silicon content and temperature of hot metal were set up based on Adaboost model.Finally,blast furnace temperature monitoring and early warning system based on big data system was established with computer technology.This system not only solves the malpractice of traditional ironmaking processing,but also plays the role of prolonging the device life cycle and predicting the trend of furnace condition in advance,which promots the intelligent transformation effectively.

关 键 词：铁水硅含量 铁水温度 大数据 Adaboost模型 线性回归模型 

分 类 号：TF53[冶金工程—钢铁冶金] TP311.13[自动化与计算机技术—计算机软件与理论]