摘要
由于劣质煤的存在,燃煤时锅炉内极易出现结渣问题,不仅降低了煤的利用率、电厂的经济效益,且使火电厂存在安全隐患,如管壁超温爆管等,不符合国家节能减排的发展战略要求。因此炉内结渣是影响火电机组和气化工艺可靠运行的关键因素之一,准确预测灰熔点可以提前调整炉膛出口温度以避免结渣。采用激光诱导击穿光谱仪(LIBS)对33个火电厂常用煤燃烧后得到的煤灰样品采集其中所有元素的光谱,分别建立煤灰中的所有元素的谱线强度与煤灰熔点的随机森林模型、支持向量机回归模型和线性回归模型,直接预测煤灰熔点温度。采用基于马氏距离(MD)的异常数据剔除算法和基于稀疏矩阵的基线估计与降噪算法(BEADS),对粉煤灰样的全光谱数据进行了预处理。将全部数据划分为70%训练集和30%测试集,并以平均相对误差(MRE)作为判断模型拟合程度的标准,其中随机森林模型对粉煤灰熔点的预测平均相对误差(MRE)为54.74%,支持向量机回归模型的预测平均相对误差为60.08%,而线性回归模型的预测平均相对误差达到了9.78%。研究结果表明,线性回归模型对煤灰熔点的预测结果更准确。将便携式LIBS光谱仪与机器学习算法相结合,用于煤灰熔点的快速检测,可有效避免锅炉结渣问题,实现火电厂安全性和经济效益的同步提高,有广阔的发展前景。
Because of the existence of low quality coal,the slagging problem is easy to occur in the boiler when coal is burned,which not only reduces the utilization ratio of coal and the economic benefit of the power plant,but also causes the safety hazard of the the power plant,such as pipe wall overtemperature explosion,etc.,which does not meet the requirements of the national development strategy of energy conservation and emission reduction.Therefore,slagging in furnace is one of the key factors affecting the reliable operation of thermal power units and gasification process.In this paper,LIBS was used to collect the spectra of all the elements in the coal ash samples from 33 thermal power Laser-induced breakdown spectroscopy,a random forest model,a support vector machine regression model and a linear regression model were developed to predict the ash fusion temperature.The full spectrum data of fly ash samples were preprocessed by using the algorithm of removing abnormal data based on mahalanobis distance(MD)and the algorithm of baseline estimation and denoising based on sparse matrix(BEADS).All the data were divided into 70%training set and 30%test set,and the average relative error(MRE)was used as the criterion to judge the fitting degree of the model,the mean relative error(MRE)of the random forest model and the support vector machine regression model were 54.74%and 60.08%,respectively,and the average relative error of linear regression model was 9.78%.The results showed that the linear regression model was more accurate in predicting the ash melting point.In this study,a portable LIBS spectrometer combined with machine learning algorithm was used to detect the melting point of coal ash,which could effectively avoid the problem of slagging in boiler,and simultaneously improve the safety and economic benefit of thermal power plant,which has a wide development prospection.
作者
鄢嘉懿
王艺陶
李燕
YAN Jiayi;WANG Yitao;LI Yan(School of Chemistry and Chemical Engineering,Nanjing University of Science and Technology,Nanjing,Jiangsu 210094,China)
出处
《中国无机分析化学》
CAS
北大核心
2024年第2期191-196,共6页
Chinese Journal of Inorganic Analytical Chemistry
基金
国家自然科学基金资助项目(51676100,21207066)。
关键词
激光诱导击穿光谱
粉煤灰
灰熔点
随机森林模型
支持向量机
线性回归
laser-induced breakdown spectroscopy
fly ash
ash melting point
random forest
support vector machine
linear regression
