Effective Selection and Application of Ethanol Characteristic Spectrum in Gasoline
-
摘要: 近红外光谱快速分析技术是检测汽油中乙醇含量的主要方法之一,光谱谱段的选择是影响快检模型预测准确性的重要因素。本研究建立了一种基于有效特征谱段的近红外光谱快速分析方法,提高了汽油中乙醇含量检测的准确度。通过对比不同浓度乙醇含量的汽油近红外光谱图,确定了汽油中乙醇分子的有效特征谱段是4524.183~5044.869 cm−1和5985.961~7108.329 cm−1。选择最优的近红外光谱预处理方法,分别使用近红外光谱全谱段和有效特征谱段进行建模分析。使用特征谱段建立的数据模型相关参数如下:交叉验证均方根误差(RMSECV)是0.5849,内部交叉验证相关系数(
${{R}}_{\rm{CV}}^{{2}}$ )是0.9991,预测均方根误差(RMSEP)是0.6090,预测集外部验证相关系数(${{R}}_{\rm{P}}^{{2}}$ )是0.9989。相较于全波长建模分析,使用特征谱段建立模型的RMSECV降低了30.27%, RMSEP降低了18.58%。综上,使用特征谱段建立的模型准确度较高,能够满足汽油中乙醇含量快速分析的需求。Abstract: Near-infrared spectroscopy rapid analysis is one of the main methods to detect the ethanol content in gasoline, and the selection of the spectrum is an important factor affecting the prediction accuracy of the rapid detection model. In this study, a rapid analysis method of near-infrared spectroscopy based on an effective characteristic spectrum was established to improve the accuracy of detection of ethanol content in gasoline. By comparing the near-infrared spectra of gasoline with different concentrations of ethanol content, the effective characteristic spectrum of ethanol molecules in gasoline was determined to be 4524.183~5044.869 cm−1 and 5985.961~7108.329 cm−1. The optimal pre-processing method for near-infrared spectroscopy was chosen, and modeling analysis was carried out using the full-spectrum and effective characteristic spectrum. The relevant parameters of the model established using the effective characteristic spectrum are as follows: the root-mean-square error of cross-validation (RMSECV) is 0.5849, the internal cross-validation correlation coefficient (${{R}}_{\text{CV}}^{\text{2}}$ ) is 0.9991, the root-mean-square error of prediction (RMSEP) is 0.6090, and the external verification correlation coefficient (${{R}}_{\text{P}}^{\text{2}}$ ) of the prediction set is 0.9989. Compared with the full-spectrum modeling analysis, the RMSECV was reduced by 30.27 %, and the RMSEP was reduced by 18.58%. In conclusion, the quantitative analysis model established by using the characteristic spectrum has higher accuracy and can meet the need for rapid analysis of ethanol content in gasoline. -
表 1 不同预处理方法的全波长建模模型性能比较
Table 1. Performance comparison of full-spectrum modeling models with different pre-processing methods
预处理方法 校正集 预测集 RMSECV $ {{R}}_{\text{CV}}^{\text{2}} $ RMSEP $ {{R}}_{\text{P}}^{\text{2}} $ 原始光谱 1.8817 0.9980 1.4163 0.9942 一阶导数 0.8388 0.9984 0.7480 0.9982 标准正态变换 1.1763 0.9909 1.6889 0.9792 矢量归一化 1.4365 0.9944 1.6948 0.9918 多元散射校正 1.3668 0.9952 2.1895 0.9862 Savitzky-Golay卷积平滑 0.8606 0.9980 1.3472 0.9948 表 2 不同预处理方法的特征谱段建模模型性能比较
Table 2. Performance comparison of characteristic spectrum modeling models with different pre-processing methods
预处理方法 校正集 预测集 RMSECV $ {{R}}_{\text{CV}}^{\text{2}} $ RMSEP $ {{R}}_{\text{P}}^{\text{2}} $ 原始光谱 0.9681 0.9984 0.9848 0.9986 一阶导数 0.5849 0.9991 0.6090 0.9989 标准正态变换 0.9524 0.9932 1.6504 0.9922 矢量归一化 0.9765 0.9985 1.4466 0.9934 多元散射校正 0.9583 0.9950 1.2998 0.9882 Savitzky-Golay卷积平滑 0.7011 0.9989 0.8940 0.9977 -
[1] FERNANDES H L, RAIMUNDO I M, PASQUINI C, et al. Simultaneous determination of methanol and ethanol in gasoline using NIR spectroscopy: Effect of gasoline composition[J]. Talanta, 2008, 75(3): 804-810. doi: 10.1016/j.talanta.2007.12.025 [2] ARDEBILI S M S, SOLMAZ H, IPCI D, et al. A review on higher alcohol of fusel oil as a renewable fuel for internal combustion engines: Applications, challenges, and global potential[J]. Fuel, 2020, 279: 11851. [3] 谭宁. 核燃料棒长度自动测量系统有效性验证研究[J]. 计量科学与技术, 2021, 65(3): 58-62. [4] AWAD O I, MAMAT R, ALI O M, et al. Alcohol and ether as alternative fuels in spark ignition engine: A review[J]. Renewable and Sustainable Energy Reviews, 2018, 82: 2586-2605. doi: 10.1016/j.rser.2017.09.074 [5] ZHAO Q H, WU X P, CHEN D Z. Emissions effect of gasoline and ethanol gasoline on vehicle exhaust CO/HC[J]. Applied Mechanics and Materials, 2014, 442: 197-200. [6] IKHSANIA A, ILMINNAFIK N, FACHRI B A. Heating ethanol-gasoline fuel mixtures to improve performance and reduce exhaust emissions at gasoline engine - A Review[C]. International Conference on Climate Change and Sustainability in ASEAN, AIP Conference Proceedings. New York: AMER INST PHYSICS, 2020: 0200241-0200247. [7] HU J, LIU Y, HAO Y, et al. Qualitative discrimination and quantitative determination model research of methanol gasoline and ethanol gasoline[J]. Spectroscopy and Spectral Analysis, 2020, 40(5): 1640-1644. [8] ARAPATSAKOS C I, KARKANIS A N, SPARE P D. Gas emissions and engine behavior when gasoline‐alcohol mixtures are used[J]. Environmental Technology, 2003, 24: 1069-1077. doi: 10.1080/09593330309385647 [9] 国家质量监督检验检疫总局. 车用乙醇汽油(E10): GB/T 18351-2017 [S]. 北京: 国家标准出版社, 2017. [10] 国家能源局. 汽油中醚类和醇类的测定 气相色谱: NB/SH/T 0663-2014 [S]. 北京: 中国石化出版社, 2014. [11] HEILMANN J, HEUMANN K. Sulfur trace determination in petroleum products by isotope dilution ICP-MS using direct injection by thermal vaporization (TV-ICP-IDMS)[J]. Analytical and Bioanalytical Chemistry, 2009, 393: 393-397. doi: 10.1007/s00216-008-2387-z [12] TEIXEIRA L S G, OLIVEIRA F S, DOS SANTOS H C, et al. Multivariate calibration in Fourier transform infrared spectrometry as a tool to detect adulterations in Brazilian gasoline[J]. Fuel, 2008, 87(3): 346-352. [13] 欧阳爱国, 刘军, 王亚平. 乙醇汽油含量的近红外光谱检测研究[J]. 激光与红外, 2012, 42(8): 901-904. doi: 10.3969/j.issn.1001-5078.2012.08.014 [14] MOREIRA M, FELICIO A, DE FRANCA J. Evaluation of the influence of sample variability on the calibration of a NIR photometer for quantification of ethanol in gasoline[J]. Measurement Science and Technology, 2019, 30(7): 075501. doi: 10.1088/1361-6501/ab16af [15] LUTZ O M D, BONN G K, RODE B M, et al. Reproducible quantification of ethanol in gasoline via a customized mobile near-infrared spectrometer[J]. Analytica Chimica Acta, 2014, 826: 61-68. doi: 10.1016/j.aca.2014.04.002 [16] OUYANG A G, LIU J. Classification and determination of alcohol in gasoline using NIR spectroscopy and the successive projections algorithm for variable selection[J]. Measurement Science and Technology, 2013, 24(2): 025502. doi: 10.1088/0957-0233/24/2/025502 [17] MABOOD F, BOQUE R, HAMAED A, et al. Near-Infrared spectroscopy coupled with multivariate methods for the characterization of ethanol adulteration in premium 91 gasoline[J]. Energy Fuels, 2017, 31(7): 7591-7597. doi: 10.1021/acs.energyfuels.7b00870 [18] CARNEIRO H S P, MEDEIROS A R B, OLIVEIRA F C C, et al. Determination of ethanol fuel adulteration by methanol using partial least-squares models based on fourier transform techniques[J]. Energy Fuels, 2008, 22(4): 2767-2770. doi: 10.1021/ef8000218 [19] PRIETO N, PAWLUCZYK O, DUGAN M E R, et al. A Review of the Principles and applications of near-infrared spectroscopy to characterize meat, Fat, and Meat Products[J]. Applied J Spectroscopy, 2017, 71(7): 1403-1426. doi: 10.1177/0003702817709299 [20] 刘喆, 宋小卫, 吴晓凤, 等. 正己烷中石油类标准物质的探讨及分析[J]. 计量科学与技术, 2021, 65(6): 39-44. doi: 10.12338/j.issn.2096-9015.2020.9001 [21] YUN Y H, LI H D, DENG B C, et al. An overview of variable selection methods in multivariate analysis of near-infrared spectra[J]. Trends in Analytical Chemistry, 2019, 103: 102-115. [22] HOSSEINI E, GHASEMI J B, DARAEI B, et al. Near-infrared spectroscopy and machine learning-based classification and calibration methods in detection and measurement of anionic surfactant in milk[J]. Journal of Food Composition and Analysis, 2021, 104: 104170. doi: 10.1016/j.jfca.2021.104170 [23] 褚小立, 王艳斌, 陆婉珍. 近红外光谱定量校正模型的建立及应用[J]. 理化检验·分析手册, 2008, 44(8): 796-800. [24] 魏传喆, 潘江, 宋夏红, 等. 一种直接驱动式振动管密度测量装置的研制[J]. 计量学报, 2021, 42(10): 1294-1298. doi: 10.3969/j.issn.1000-1158.2021.10.06 [25] SCHLEGEL L B, SCHUBERT-ZSILAVECZ M, ABDEL-TAWAB M. Quantification of active ingredients in semi-solid pharmaceutical formulations by near infrared spectroscopy[J]. Journal of Pharmaceutical and Biomedical Analysis, 2017, 142: 178-189. doi: 10.1016/j.jpba.2017.04.048