轉(zhuǎn)爐煙氣與爐口火焰光譜特征的相關(guān)性分析及應(yīng)用

摘" 要： 針對轉(zhuǎn)爐煙氣存在干擾性和滯后性的問題，利用與其同源產(chǎn)生的光譜信息，修正煙道安裝環(huán)境對系統(tǒng)數(shù)據(jù)采集的擾動，研究轉(zhuǎn)爐煙氣的光譜分析新技術(shù)。為探究兩者之間的相關(guān)性，首先利用多重分形理論對光譜數(shù)據(jù)的能量值、連續(xù)譜峰值、離散譜峰值比和爐氣數(shù)據(jù)的變化率特征進(jìn)行運算，求取四者之間的關(guān)聯(lián)性；然后使用粒子群算法優(yōu)化支持向量機(jī)預(yù)測模型（PSO?SVM），以光譜數(shù)據(jù)的能量值、連續(xù)譜峰值和離散譜峰值比作為輸入，以爐氣數(shù)據(jù)的變化率特征作為輸出，進(jìn)行預(yù)測實驗。結(jié)果表明：能量值、連續(xù)譜峰值等4個指標(biāo)的多重分形譜最大值均接近于1.011 492，譜寬均接近于0.000 014 476，

且實驗曲線對稱性相同，表明多重分形譜具有極強(qiáng)的自相似性和復(fù)雜性；同時，預(yù)測模型實驗結(jié)果差異性的均方誤差為0.012 5。

關(guān)鍵詞： 轉(zhuǎn)爐煙氣； 火焰光譜； 光譜特征分析； 相關(guān)性分析； 多重分形； PSO?SVM

中圖分類號： TN247?34" " " " " " " " " " " " " " 文獻(xiàn)標(biāo)識碼： A" " " " " " " " " " " "文章編號： 1004?373X（2025）02?0162?07

Correlation analysis and application of converter flue gas and flame spectrum characteristics of furnace inlet

ZHANG Ruicheng1， LIU Yonglong1， HAN Yang2

（1. School of Electrical Engineering， North China University of Science and Technology， Tangshan 063210， China;

2. College of Metallurgy and Energy， North China University of Science and Technology， Tangshan 063210， China）

Abstract： In allusion to the problems of interference and hysteresis of converter flue gas， a new spectral analysis technology of converter flue gas is researched by using the spectral information generated from the same origin to correct the disturbance of flue installation environment to system data acquisition. In order to explore the correlation between the two， the multifractal theory is used to calculate the characteristics of the energy value， the peak ratio of the continuum spectrum and the peak ratio of the discrete spectrum of the spectral data and the change rate of the furnace gas data， so as to find the correlation between the four. The particle swarm optimization support vector machine （PSO?SVM） prediction model is used to conduct the Prediction experiment by taking the energy value， continuum peak and discrete peak ratio of the spectral data as the input， and the change rate characteristics of the furnace gas data as the output. The results show that the maximum values of the multifractal spectra of the four indexes such as energy value and continuous spectrum peak， are close to 1.011 492， the spectral widths are close to 0.000 014 476， and the symmetry of the experimental curves are the same， indicating that the multifractal spectrum has strong self?similarity and complexity. The mean square error of the difference of experimental results of the PSO?SVM prediction model is 0.012 5.

Keywords： converter flue gas; flame spectrum; spectral characteristic analysis; correlation analysis; multifractal; PSO?SVM

0" 引" 言

當(dāng)前，實現(xiàn)轉(zhuǎn)爐冶煉終點動態(tài)控制常用的技術(shù)為煙氣分析技術(shù)[1]。煙氣分析技術(shù)是通過在煙道處安裝煙氣分析質(zhì)譜儀，對冶煉中產(chǎn)生的煙氣成分和流量進(jìn)行持續(xù)檢測分析，結(jié)合供氧信息對熔池中脫碳速率進(jìn)行實時計算，從而實現(xiàn)轉(zhuǎn)爐爐內(nèi)碳溫的在線預(yù)報和全程動態(tài)矯正[2]。然而對鋼水中碳含量和溫度值的動態(tài)預(yù)報精度取決于煙道安裝環(huán)境和數(shù)據(jù)信息同步狀況，其魯棒性不高，且存在滯后性，導(dǎo)致控制精度并不理想[3]。爐口火焰中各組分構(gòu)成和濃度始終在發(fā)生變化，氧化還原反應(yīng)的過程和強(qiáng)度始終在發(fā)生變化，以及煙氣中不同組分的輻射光譜動態(tài)疊加，最終導(dǎo)致輻射光譜構(gòu)成極為復(fù)雜，并且會隨吹煉進(jìn)程差異產(chǎn)生相同時刻下不同波長的光譜輻射差異。……