skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Prediction of gasoline octane numbers from near-infrared spectral features in the range 660-1215 nm

Abstract

The feasibility of predicting the quality parameters of gasoline from its absorption spectrum in the wavelength range 660-1215 nm was investigated. In this spectral region, vibrational overtones and combination bands of CH groups of methyl, methylene, aromatic, and olefinic functions were observed. With the aid of multivariate statistics, the spectral features could be correlated to various quality parameters of gasoline such as octane number. As an example, multivariate analysis of the spectra of 43 unleaded gasoline samples yielded a three-wavelength prediction equation for pump octane that gave excellent correlations (R/sup 2/ = 0.95; standard error of estimate, 0.3-0.4 octane number; standard error of prediction, 0.4-0.5 octane number) with the ASTM motor determined octane numbers. Independent multivariate analysis using partial least-squares (PLS) regression yielded similar results. An additional set of nine samples from the Pacific Coast Exchange Group of the ASTM were examined for ten different quality parameters (research and motor octane numbers, Reid vapor pressure, API gravity, bromine number, lead, sulfur, aromatic, olefinic, and saturate contents). Regression analysis of the spectra results in correlation of nine of the ten properties with R/sup 2/ values ranging from 0.94 to 0.99 and standard errors near the independent reference test values.

Authors:
; ; ;
Publication Date:
Research Org.:
Univ. of Washington, Seattle (USA)
OSTI Identifier:
6271943
Resource Type:
Journal Article
Journal Name:
Anal. Chem.; (United States)
Additional Journal Information:
Journal Volume: 61:4
Country of Publication:
United States
Language:
English
Subject:
02 PETROLEUM; 37 INORGANIC, ORGANIC, PHYSICAL AND ANALYTICAL CHEMISTRY; ALKANES; ABSORPTION SPECTROSCOPY; ALKENES; ANTIKNOCK RATINGS; PREDICTION EQUATIONS; AROMATICS; GASOLINE; LEAD; SULFUR; API GRAVITY; BROMINE NUMBER; EXPERIMENTAL DATA; INFRARED SPECTRA; MULTIVARIATE ANALYSIS; QUANTITATIVE CHEMICAL ANALYSIS; REGRESSION ANALYSIS; VAPOR PRESSURE; CHEMICAL ANALYSIS; DATA; DENSITY; ELEMENTS; EQUATIONS; FUELS; HYDROCARBONS; INFORMATION; LIQUID FUELS; MATHEMATICS; METALS; NONMETALS; NUMERICAL DATA; ORGANIC COMPOUNDS; PETROLEUM PRODUCTS; PHYSICAL PROPERTIES; SPECTRA; SPECTROSCOPY; STATISTICS; THERMODYNAMIC PROPERTIES; 023000* - Petroleum- Properties & Composition; 400102 - Chemical & Spectral Procedures

Citation Formats

Kelly, J J, Barlow, C H, Jinguji, T M, and Callis, J B. Prediction of gasoline octane numbers from near-infrared spectral features in the range 660-1215 nm. United States: N. p., 1989. Web. doi:10.1021/ac00179a007.
Kelly, J J, Barlow, C H, Jinguji, T M, & Callis, J B. Prediction of gasoline octane numbers from near-infrared spectral features in the range 660-1215 nm. United States. https://doi.org/10.1021/ac00179a007
Kelly, J J, Barlow, C H, Jinguji, T M, and Callis, J B. 1989. "Prediction of gasoline octane numbers from near-infrared spectral features in the range 660-1215 nm". United States. https://doi.org/10.1021/ac00179a007.
@article{osti_6271943,
title = {Prediction of gasoline octane numbers from near-infrared spectral features in the range 660-1215 nm},
author = {Kelly, J J and Barlow, C H and Jinguji, T M and Callis, J B},
abstractNote = {The feasibility of predicting the quality parameters of gasoline from its absorption spectrum in the wavelength range 660-1215 nm was investigated. In this spectral region, vibrational overtones and combination bands of CH groups of methyl, methylene, aromatic, and olefinic functions were observed. With the aid of multivariate statistics, the spectral features could be correlated to various quality parameters of gasoline such as octane number. As an example, multivariate analysis of the spectra of 43 unleaded gasoline samples yielded a three-wavelength prediction equation for pump octane that gave excellent correlations (R/sup 2/ = 0.95; standard error of estimate, 0.3-0.4 octane number; standard error of prediction, 0.4-0.5 octane number) with the ASTM motor determined octane numbers. Independent multivariate analysis using partial least-squares (PLS) regression yielded similar results. An additional set of nine samples from the Pacific Coast Exchange Group of the ASTM were examined for ten different quality parameters (research and motor octane numbers, Reid vapor pressure, API gravity, bromine number, lead, sulfur, aromatic, olefinic, and saturate contents). Regression analysis of the spectra results in correlation of nine of the ten properties with R/sup 2/ values ranging from 0.94 to 0.99 and standard errors near the independent reference test values.},
doi = {10.1021/ac00179a007},
url = {https://www.osti.gov/biblio/6271943}, journal = {Anal. Chem.; (United States)},
number = ,
volume = 61:4,
place = {United States},
year = {Wed Feb 15 00:00:00 EST 1989},
month = {Wed Feb 15 00:00:00 EST 1989}
}