钢铁冶炼的风炉拱顶温度建模仿真研究  

Modeling of Vault Temperature of Hot Blast Stove Based on Wavelet Filter

作　　者：崔桂梅[1] 薄宇[1] 

机构地区：[1]内蒙古科技大学信息工程学院,内蒙古包头014010

出　　处：《计算机仿真》2016年第8期292-296,共5页Computer Simulation

基　　金：国家自然科学基金资助项目(61164018);内蒙古科技厅项目(41402060423)

摘　　要：针对钢铁冶炼对高炉生产节能降耗要求的不断提高,需提高对热风炉的燃烧效率。由于热风炉拱顶温度是热风炉燃烧控制系统是一个大惯性、纯滞后、多元参数分布非线性的系统,要实现高效的热风炉燃烧控制,就需要提前预测拱顶温度值,常规的数学建模较难实现。通过采集某钢厂热风炉的现场燃烧数据,在传统的数据预处理基础上,引入小波滤波对数据进行滤波分析,建立热风炉拱顶温度的BP神经网络模型。仿真结果表明模型准确性好。可以看出与传统的数据预处理所建立的神经网络模型相比,小波分析进行滤波使得滤波后残留噪声较小,同时实现降噪滤波与异常值剔除,所建立的神经网络模型预测更加准确。With the increasing requirements of blast furnace energy saving,the efficiency requirements of hot blast stove are also increasing. Arch temperature is an important control parameter in combustion process of hot blast stove. Combustion system is a big inertia,pure lag,multivariate parameter distribution nonlinear system,it is necessary to predict arch temperature to achieve efficient stove combustion control and conventional mathematical modeling is difficult to achieve. By collecting the data from combustion process of hot blast stove,based on the traditional data preprocessing,the wavelet filter is introduced to analyze the data to complete data filtering,the model of BP neural network is established for the arch temperature of hot blast stove. The simulation results show that the model is accurate,and compared with the neural network model established by traditional data preprocessing data,the proposed wavelet filter makes small noise after filtering,and the data filtering and elimination are completed at the same time.

关 键 词：热风炉 拱顶温度 小波滤波 神经网络 数学模型 

分 类 号：N945.12[自然科学总论—系统科学] TP391.9[自动化与计算机技术—计算机应用技术]