Quantitative analysis of chromium by fiber-optic laser-inducedbreakdown spectroscopy of steels for nuclear power plant based on support vector regression and random forest regression
SHI Mingxin1,2, WANG Aosong1, HUANG Dapeng1YU Han2, ZHANG Zhi2, WU Jian2
1. State Key Laboratory of Nuclear Safety Monitoring Technology and Equipment,China National Nuclear Engineering Co.,Ltd.,Shenzhen 518172,China; 2. School of Electrical Engineering,Xi′an Jiaotong University,Xi′an 710049,China
Abstract：The state detection technology of key equipment in nuclear power plants was an important guarantee for the safety and economy of nuclear power.Laser-induced breakdown spectroscopy (LIBS) was a new method for on-line measurement of elemental composition and structural states of materials.The fiber-optic laser-induced breakdown spectroscopy (fiber-optic LIBS,FO-LIBS) system which was based on the optical fiber transmission of the laser and plasma radiation was developed and the typical spectra of 20 standard steel samples and 3 steel samples to be tested were obtained.Based on support vector regression(SVR) and random forest regression(RFR),effective calibration models for quantitative measurement of Cr element were established.The optimal training effects of calibration models were obtained by screening the input spectral lines and determining the number of groups.The cross-validation results of support vector regression and the calibration results of random forest regression proved that the two algorithms both had excellent generalization ability.The determination coefficient (R2) of Cr element calibrated by the two methods reached more than 0.98 and the relative standard deviation could be limited to 10%.Based on the calibration model,the content of Cr element in three steel samples was obtained.SVR predicted 19.38%(standard value 19.33%),0.190%(standard value 0.110%) and 0.986%(standard value 0.990%),and RFR predicted 18.57%,0.16% and 0.951%,respectively.The research laid a theoretical foundation for the development of optical fiber LIBS equipment.
时铭鑫, 王傲松, 黄大鹏, 余涵, 张智, 吴坚. 基于支持向量机和随机森林算法结合光纤式激光诱导击穿光谱定量检测核电用钢中铬[J]. 冶金分析, 2021, 41(1): 30-40.
SHI Mingxin, WANG Aosong, HUANG DapengYU Han, ZHANG Zhi, WU Jian. Quantitative analysis of chromium by fiber-optic laser-inducedbreakdown spectroscopy of steels for nuclear power plant based on support vector regression and random forest regression. , 2021, 41(1): 30-40.
董超芳,肖葵,刘智勇,等.核电环境下流体加速腐蚀行为及其研究进展[J].科技导报,2010,28(10):96-100.DONG Chaofang,XIAO Kui,LIU Zhiyong,et al.A review of flow-accelerated corrosion in nuclear power industry[J].Science and Technology Review,2010,28(10):96-100.
[2]
JIANG S,CHAI F,SU H,et al.Influence of chromium on the flow-accelerated corrosion behavior of low alloy steels in 3.5% NaCl solution[J].Corrosion Science,2017,123(15):217-227.
[3]
CHAI F,JIANG S,YANG C F.Effect of Cr on characteristic of rust layer formed on low alloy steel in flow-accelerated corrosion environment[J].Journal of Iron and Steel Research International,2016,23(6):602-607.
[4]
FORTES F J,MOROS J,LUCENA P,et al.Laser-induced breakdown spectroscopy[J].Analytical Chemistry,2013,85(2):640-669.
[5]
HAHN D W,OMENETTO N.Laser-induced breakdown spectroscopy (LIBS),part I:review of basic diagnostics and plasma-article interactions:still-challenging issues within the analytical plasma community[J].Applied Spectroscopy,2012,64(12) 335A-366A.
[6]
张天龙,吴珊,汤宏胜,等.化学计量学在激光诱导击穿光谱分析中的研究进展[J].分析化学,2015,43(6):939-948.ZHANG Tianlong,WU Shan,TANG Hongsheng,et al.Progress of chemometrics in laser-induced breakdown spectroscopic analysis[J].Chinese Journal of Analytical Chemistry,2015,43(6):939-948.
[7]
ERLER A,RIEBE D,BEITZ T,et al.Oil nutrient detection for precision agriculture using handheld laser-induced breakdown spectroscopy (LIBS) and multivariate regression methods (PLSR,Lasso and GPR)[J].Sensors,2020,20(2):418.
[8]
FENG J,WANG Z,WEST L,et al.A PLS model based on dominant factor for coal analysis using laser-induced breakdown spectroscopy[J].Analytical and Bioanalytical Chemistry,2011,400(10):3261-3271.
[9]
杨友良,王禄,马翠红.改进粒子群算法优化SVR的LIBS钢液元素定量分析[J].激光与光电子学进展,2020,57(5):053002-1-053002-8.YANG Youliang,WANG Lu,MA Cuihong.Quantitative analysis of liquid steel element in LIBS using SVR improved by particle swarm optimization[J].Laser &Optoelectronics Progress,2020,57(5):053002-1-053002-8.
[10]
ZHANG T,WU S,DONG J,et al.Quantitative and classification analysis of slag samples by laser induced breakdown spectroscopy (LIBS) coupled with support vector machine (SVM) and partial least square (PLS) methods[J].Journal of Analytical Atomic Spectrometry,2015,30(2):368-374.
[11]
ZHANG T,LIANG L,WANG K,et al.A novel approach for quantitative analysis of multi-elements in steels based on laser-induced breakdown spectroscopy(LIBS) and random forest regression(RFR)[J].Journal of Analytical Atomic Spectrometry,2014,29(12):2323-2329.
[12]
LU S,DONG M,HUANG J,et al.Estimation of the aging grade of T91 steel by laser-induced breakdown spectroscopy coupled with support vector machines[J].Spectrochimica Acta Part B:Atomic Spectroscopy,2017,140:35-43.
