Quantitative analysis of multiple elements in steel by laser-induced breakdown spectroscopy based on Kalman filter and machine learning
REN Xiangxu1,2, KONG Linghua*1,2,3, ZHENG Jishi4, DING Zhigang1 LIAN Guofu1,2, YANG Jiacheng1,2
1. School of Mechanical and Automotive Engineering,Fujian University of Technology,Fuzhou 350118,China; 2. Digital Fujian Industrial Manufacturing IOT Lab,Fujian University of Technology,Fuzhou 350118,China; 3. Fujian Key Laboratory of Intelligent Machining Technology and Equipment,Fujian University of Technology, Fuzhou 350118,China; 4. School of Transportation,Fujian University of Technology,Fuzhou 350118,China
Abstract：In order to solve the problem of quantitative analysis of trace elements in the process of iron and steel smelting, the elements including Mn, P, S and C in steel samples were quantitatively analyzed offline, and the preliminary research work was carried out offline for the online detection of molten steel composition in molten state. The spectral characteristics of steel were studied by laser-induced breakdown spectroscopy (LIBS). Kalman filter (KF) was proposed to denoise the spectrum and combined with partial least squares regression (PLSR), support vector machine regression (SVR) and random forest (RF) machine learning models to establish the quantitative analysis model of each element. The structural parameters of the model were optimized by K-fold cross validation and grid search method. The quantitative analysis of elements in steel was realized based on LIBS technology. The results showed that the KF-SVR prediction model had the best prediction performance for Mn and C. The values of determination coefficient (R2) of test set were all up to 0.999 9, and the root mean square error (RMSE) was 0.020 8% and 0.021 3%, respectively. The KF-PLSR prediction model had the best prediction performance for P and S. The R2 values of their test sets were 0.999 8 and 0.999 9 respectively, and RMSE were 0.006 0% and 0.002 8%,respectively. The R2 values of three machine learning models after combined with KF were all higher than 0.996. The results demonstrated that the use of KF in the preprocessing of spectral data of LIBS could effectively improve the signal-to-noise ratio and spectral quality. After combined with machine learning model, it could achieve high precision in quantitative analysis of micro elements in steel. It had a good application value to realize the rapid and high precision quantitative analysis of micro elements in the field of metallurgy, so as to control the properties and quality of steel in real time. This study also provided a basis for further online detection of molten steel composition in molten state.
任祥旭, 孔令华, 郑积仕, 丁志刚, 练国富, 杨嘉诚. 基于卡尔曼滤波与机器学习的激光诱导击穿光谱钢铁多元素定量分析探究[J]. 冶金分析, 2023, 43(6): 19-29.
REN Xiangxu, KONG Linghua, ZHENG Jishi, DING Zhigang LIAN Guofu, YANG Jiacheng. Quantitative analysis of multiple elements in steel by laser-induced breakdown spectroscopy based on Kalman filter and machine learning. , 2023, 43(6): 19-29.
[1] Moncayo S,Manzoor S,Rosales J D,et al.Qualitative and quantitative analysis of milk for the detection of adulteration by laser induced breakdown spectroscopy (LIBS)[J].Food Chemistry,2017,232:322-328. [2] Zhao X,Zhao C,Du X,et al.Detecting and mapping harmful chemicals in fruit and vegetables using nanoparticle-enhanced laser-induced breakdown spectroscopy[J].Scientific Reports,2019,9(1):1-10. [3] Senesi G S,Dell′Aglio M,Gaudiuso R,et al.Heavy metal concentrations in soils as determined by laser-induced breakdown spectroscopy (LIBS),with special emphasis on chromium[J].Environmental Research,2009,109(4):413-420. [4] 王亚文,张勇,陈雄飞,等.激光诱导击穿光谱在金属材料在线分析方面的应用进展[J].冶金分析,2020,40(12):7-13. WANG Yawen,ZHANG Yong,CHEN Xiongfei,et al.Application progress of laser-induced breakdown spectroscopy in online analysis of metal material[J].Metallurgical Analysis,2020,40(12):7-13. [5] Sturm V,Fleige R,de Kanter M,et al.Laser-induced breakdown spectroscopy for 24/7 automatic liquid slag analysis at a steel works[J].Analytical Chemistry,2014,86(19):9687-9692. [6] Lednev V N,Dormidonov A E,Sdvizhenskii P A,et al.Compact diode-pumped Nd:YAG laser for remote analysis of low-alloy steels by laser-induced breakdown spectroscopy[J].Journal of Analytical Atomic Spectrometry,2018,33(2):294-303. [7] Mohamed N,Rifai K,Selmani S,et al.Chemical and mineralogical mapping of platinum-group element ore samples using laser-induced breakdown spectroscopy and micro-X-ray fluorescence[J].Geostandards and Geoanalytical Research,2021,45(3):539-550. [8] Senesi G S.Laser-induced breakdown spectroscopy (LIBS) applied to terrestrial and extraterrestrial analogue geomaterials with emphasis to minerals and rocks[J].Earth-Science Reviews,2014,139:231-267. [9] Idrees B S,Wang Q,Khan M N,et al.In-vitro study on the identification of gastrointestinal stromal tumor tissues using laser-induced breakdown spectroscopy with chemometric methods[J].Biomedical Optics Express,2022,13(1):26-38. [10] Guirado S,Fortes F J,Laserna J J.Elemental analysis of materials in an underwater archeological shipwreck using a novel remote laser-induced breakdown spectroscopy system[J].Talanta,2015,137:182-188. [11] Kashiwakura S,Wagatsuma K.Selection of atomic emission lines on the mutual identification of austenitic stainless steels with a combination of laser-induced breakdown spectroscopy (LIBS) and partial-least-square regression (PLSR)[J].ISIJ International,2020,60(6):1245-1253. [12] Owolabi T O,Gondal M A.Quantitative analysis of LIBS spectra using hybrid chemometric models through fusion of extreme learning machines and support vector regression[J].Journal of Intelligent & Fuzzy Systems,2018,35(6):6277-6286. [13] Chang F,Lu H,Sun H,et al.Assessment of the performance of quantitative feature-based transfer learning LIBS analysis of chromium in high temperature alloy steel samples[J].Journal of Analytical Atomic Spectrometry,2020,35(11):2639-2648. [14] Dai Y,Song C,Gao X,et al.Quantitative determination of Al-Cu-Mg-Fe-Ni aluminum alloy using laser-induced breakdown spectroscopy combined with LASSO-LSSVM regression[J].Journal of Analytical Atomic Spectrometry,2021,36(8):1634-1642. [15] 谢远明,孙兰香,袁德成,等.基于互信息特征筛选偏最小二乘的激光诱导击穿光谱铁矿浆定量分析[J].冶金分析,2022,42(1):18-24. XIE Yuanming,SUN Lanxiang,YUAN Decheng,et al.Quantitative analysis of iron slurry based on laser induced breakdown spectroscopy combined with mutual information feature selection partial least squares method [J].Metallurgical Analysis,2022,42(1):18-24. [16] Park S,Gil M S,Im H,et al.Measurement noise recommendation for efficient Kalman filtering over a large amount of sensor data[J].Sensors,2019,19(5):1168. [17] 陈鹏,齐超,刘人玮,等.基于支持向量机回归的LIBS飞灰含碳量定量分析[J].光学学报,2022,42(9):278-285. CHEN Peng,QI Chao,LIU Renwei,et al.Quantitative analysis of carbon content in fly ash using LIBS based on support vector machine regression[J].Acta Optica Sinica,2022,42(9):278-285. [18] Chen T,Men J,Zhao M,et al.The spectral fusion of laser-induced breakdown spectroscopy (LIBS) and mid-infrared spectroscopy (MIR) coupled with random forest (RF) for the quantitative analysis of soil pH[J].Journal of Analytical Atomic Spectrometry,2021,36(5):1084-1092. [19] Zhang T,Liang L,Wang K,et al.A novel approach for the quantitative analysis of multiple elements in steel based on laser-induced breakdown spectroscopy (LIBS) and random forest regression (RFR)[J].Journal of Analytical Atomic Spectrometry,2014,29(12):2323-2329. [20] Wold S,Esbensen K,Geladi P.Principal components analysis[J].Chemometrics and Intelligent Laboratory Systems,1987,2(1/2/3):37-52. [21] Mac＇kiewicz A,Ratajczak W.Principal components analysis (PCA)[J].Computers & Geosciences,1993,19(3):303-342.
