Incorporating empirical knowledge into data-driven variable selection for quantitative analysis of coal ash content by laser-induced breakdown spectroscopy

作     者：吕一涵 宋惟然 侯宗余 王哲 Yihan LYU;Weiran SONG;Zongyu HOU;Zhe WANG

作者机构：State Key Laboratory of Power System Operation and ControlInternational Joint Laboratory on Low Carbon Clean Energy InnovationDepartment of Energy and Power EngineeringTsinghua UniversityBeijing 100084People’s Republic of China School of Energy and Electrical EngineeringQinghai UniversityXining 810016People’s Republic of China Tsinghua UniversityShanxi Research Institute Clean EnergyTaiyuan 030032People’s Republic of China 

出 版 物：《Plasma Science and Technology》 (等离子体科学和技术（英文版）)

年 卷 期：2024年第26卷第7期

页      面：148-156页

核心收录：

学科分类：080901[工学-物理电子学] 0809[工学-电子科学与技术（可授工学、理学学位）] 08[工学] 080401[工学-精密仪器及机械] 0804[工学-仪器科学与技术] 0803[工学-光学工程] 

基　　金：financial supports from National Natural Science Foundation of China(No.62205172) Huaneng Group Science and Technology Research Project(No.HNKJ22-H105) Tsinghua University Initiative Scientific Research Program and the International Joint Mission on Climate Change and Carbon Neutrality 

主　　题：laser-induced breakdown spectroscopy(LIBS) coal ash content quantitative analysis variable selection empirical knowledge partial least squares regression(PLSR) 

摘      要：Laser-induced breakdown spectroscopy(LIBS)has become a widely used atomic spectroscopic technique for rapid coal ***,the vast amount of spectral information in LIBS contains signal uncertainty,which can affect its quantification *** this work,we propose a hybrid variable selection method to improve the performance of LIBS *** variables are first identified using Pearson s correlation coefficient,mutual information,least absolute shrinkage and selection operator(LASSO)and random forest,and then filtered and combined with empirical variables related to fingerprint elements of coal ash ***,these variables are fed into a partial least squares regression(PLSR).Additionally,in some models,certain variables unrelated to ash content are removed manually to study the impact of variable deselection on model *** proposed hybrid strategy was tested on three LIBS datasets for quantitative analysis of coal ash content and compared with the corresponding data-driven baseline *** is significantly better than the variable selection only method based on empirical knowledge and in most cases outperforms the baseline *** results showed that on all three datasets the hybrid strategy for variable selection combining empirical knowledge and data-driven algorithms achieved the lowest root mean square error of prediction(RMSEP)values of 1.605,3.478 and 1.647,respectively,which were significantly lower than those obtained from multiple linear regression using only 12 empirical variables,which are 1.959,3.718 and 2.181,*** LASSO-PLSR model with empirical support and 20 selected variables exhibited a significantly improved performance after variable deselection,with RMSEP values dropping from 1.635,3.962 and 1.647 to 1.483,3.086 and 1.567,*** results demonstrate that using empirical knowledge as a support for datadriven variable selection can be a viable approach to improve the acc