基于自编码器的汽油分子组成预测

doi:10.12422/j.issn.1672-6952.2024.02.001

摘要/Abstract

摘要：

汽油分子在线调和技术需要快速获取各种类型组分油的详细分子组成信息。开发了基于自编码器的汽油分子组成快速解析方法，该方法可由近红外光谱直接预测汽油详细单体烃的组成。构建的汽油分子组成自编码器模型可挖掘汽油组成的潜在特征，并利用潜在特征解码恢复原始分子组成。利用神经网络算法关联近红外光谱特征信息与汽油组成的潜在特征，并采用加氢汽油验证了模型的准确性。结果表明，平均绝对误差为0.033。开发的模型将自编码器算法应用在传统的石油化工过程中，对汽油分子在线调和与实时优化具有重要的指导意义。

关键词: 在线调和, 自编码器, 分子组成, 快速解析, 近红外光谱

Abstract:

Gasoline molecular blending technology on?line requires rapid access to detailed molecular composition information of various types of component oils. In this paper, an autoencoder?based method for the rapid resolution of gasoline molecular composition is developed, which can directly predict the detailed monomeric hydrocarbon composition of gasoline from near?infrared spectra. The constructed autoencoder model of gasoline molecular composition can explore the potential features and recover the original molecular composition by decoding the potential features. The artificial neural network algorithm is used to correlate the NIR spectral information with the potential features of gasoline composition. The accuracy of the model is verified by using hydrogenated gasoline with the average absolute error is 0.033. The model developed in this work applies the current popular autoencoder algorithm to the traditional petrochemical process, which is an important guideline for blending online and real?time optimization of gasoline molecules.

Key words: Blending on?line, Autoencoder, Molecular composition, Rapid resolution, Near?infrared spectra

中图分类号:

TE626

蔡广庆, 胡益炯, 李春澎, 纪晔, 王弘历. 基于自编码器的汽油分子组成预测[J]. 辽宁石油化工大学学报, 2024, 44(2): 1-6.

Guangqing CAI, Yijiong HU, Chunpeng LI, Ye JI, Hongli WANG. Gasoline Molecular Composition Prediction Based on Autoencoder Algorithm[J]. Journal of Liaoning Petrochemical University, 2024, 44(2): 1-6.

图/表 8

图1 近红外光谱快速预测汽油分子组成模型构建的示意图

Fig.1 Schematic diagram of model construction for rapid prediction of gasoline molecular composition by near infrared spectroscopy

图2 107组汽油样品的近红外光谱图

Fig.2 Near?infrared spectra of 107 gasoline samples

图3 汽油的典型气相色谱图

Fig.3 The typical gas chromatogram of gasoline

图4 近红外光谱和气相色谱数据在二维坐标上的分类可视化图

Fig.4 Classification visualization of near?infrared spectra and gas chromatogram data on two?dimensional coordinates

图5 汽油分子组成自编码器模型示意图

Fig.5 Schematic diagram of gasoline molecular composition autoencoder model

图6 汽油分子组成自编码器的训练结果

Fig.6 Training results of gasoline molecular composition autoencoder

图7 测试集加氢汽油的预测值与实验值的对比结果

Fig. 7 Comparison of predicted and experimental values of hydrogenated gasoline in the test set

图8 加氢汽油的预测值和实验值对比结果

Fig.8 Parity plot of predicted and experimental values of hydrogenated gasoline

参考文献 21

1	STEWART W E. Predict octanes for gasoline blending［J］. Petroleum Refinery， 1956， 38（12）： 135⁃139.
2	AUCKLAND M H T， CHARNOCK D J. The development of linear blending indices for petroleum properties［J］. Journal of the Institute of Petroleum， 1969， 55（545）： 322⁃329.
3	RUSIN M H， CHUNG H S， MARSHALL J F. A "transformation" method for calculating the research and motor octane numbers of gasoline blends［J］. Industrial & Engineering Chemistry Fundamentals， 1981， 20（3）： 195⁃204.
4	MULLER A. New method produces accurate octane blending values［J］. Oil & Gas Journal， 1992， 90（12）： 80⁃90.
5	HEALY W C， JR C W M， PETERSON R T. A new approach to blending octane［J］. Proc Am Inst， 1959， 39（3）： 132⁃136.
6	TWU C H， COON J E. Predict octane numbers using a generalized interaction method： Clean fuels technology［J］. Hydrocarbon Processing， 1996， 75（2）： 51⁃56.
7	BROWN J M， SUNDARAM A， SAEGER R B， et al. Estimating detailed compositional information from limited analytical data： US⁃8682597⁃B2［P］. 2014⁃03⁃25.
8	NEUROCK M， NIGAM A， LIBANATI C， et al. Monte carlo simulation of complex reaction systems： Molecular structure and reactivity in modelling heavy oils［J］. Chemical Engineering Science， 1990， 45（8）： 2083⁃2088.
9	NEUROCK M， NIGAM A， TRAUTH D， et al. Molecular representation of complex hydrocarbon feedstocks through efficient characterization and stochastic algorithms［J］. Chemical Engineering Science， 1994， 49（24）： 4153⁃4177.
10	CAMPBELL D M， KLEIN M T. Construction of a molecular representation of a complex feedstock by Monte Carlo and quadrature methods［J］. Applied Catalysis A： General， 1997， 160（1）： 41⁃54.
11	TRAUTH D M， STARK S M， PETTI T F， et al. Representation of the molecular structure of petroleum resid through characterization and Monte Carlo modeling［J］. Energy & Fuels， 1994， 8（3）： 576⁃580.
12	PAN Y B， YANG B L， ZHOU X W. Feedstock molecular reconstruction for secondary reactions of fluid catalytic cracking gasoline by maximum information entropy method［J］. Chemical Engineering Journal， 2015， 281： 945⁃952.
13	HUDEBINE D， VERSTRAETE J J. Reconstruction of petroleum feedstocks by entropy maximization. Application to FCC gasolines［J］. Oil & Gas Science and Technology， 2011， 66（3）： 437⁃460.
14	ZHANG Y. A molecular approach for characterisation and property predictions of petroleum mixtures with applications to refinery modelling［D］. England： University of Manchester， 1999.
15	CUI C， ZHANG L Z， MA Y J， et al. Computer⁃aided gasoline compositional model development based on GC⁃FID analysis［J］. Energy & Fuels， 2018， 32（8）： 8366⁃8373.
16	CUI C， BILLA T， ZHANG L Z， et al. Molecular representation of the petroleum gasoline fraction［J］. Energy & Fuels， 2018， 32（2）： 1525⁃1533.
17	LITANI⁃BARZILAI I， SELA I， BULATOV V， et al. On⁃line remote prediction of gasoline properties by combined optical methods［J］. Analytica Chimica Acta， 1997， 339（1/2）： 193⁃199.
18	史月华，陆勇，徐光明，等. 主成分回归残差神经网络校正算法用于近红外光谱快速测定汽油辛烷值［J］. 分析化学， 2001， 29（1）： 87⁃91.
	SHI Y H， LU Y， XU G M， et al. Principal component regression residual artificial neural network calibration algorithm applied in near infrared fast measurement of gasoline octane number［J］. Chinese Journal of Analytical Chemistry， 2001， 29（1）： 87⁃91.
19	KELLY J J， BARLOW C H， JINGUJI T M， et al. Prediction of gasoline octane numbers from near⁃infrared spectral features in the range 660⁃1215 nm［J］. Analytical Chemistry， 1989， 61（4）： 313⁃320.
20	史永刚，刘绍璞，宋世远，等. 基于支持向量机的汽油族组成近红外光谱分析方法研究［J］. 分析测试学报， 2007， 26（3）： 343⁃346.
	SHI Y G， LIU S P， SONG S Y， et al. Near infrared analysis of major classes of hydrocarbon constituents in gasoline with least square⁃support vector machine［J］. Journal of Instrumental Analysis， 2007， 26（3）： 343⁃346.
21	刘秋芳，褚小立，陈瀑，等. 基于近红外光谱快速预测石脑油单体烃分子组成［J］. 石油炼制与化工， 2022， 53（1）： 86⁃92.
	LIU Q F， CHU X L， CHEN B， et al. Rapid prediction of hydrocarbon molecular composition of naphtha based on near infrared spectroscopy［J］. Petroleum Processing and Petrochemicals， 2022， 53（1）： 86⁃92.

[1]	王金明,于金媛,李忠虎. 近红外CO光谱采集及建模方法研究[J]. 辽宁石油化工大学学报, 2017, 37(3): 54-57.
[2]	张宇佳. MIR和NIR光谱技术比较及在烃源岩分析中的应用[J]. 辽宁石油化工大学学报, 2013, 33(3): 19-23.