Modeling of the Properties of Gasoline and Petroleum Fractions Using a Robust Scheme

Abstract

Gasoline is one of the most recognized products of the petroleum industry due to its use as a liquid fuel worldwide. As a result, it is of great importance to accurately determine the properties of gasoline, so as to evaluate its quality. In this article, an effective mathematical and predictive strategy, namely least squares support vector machines (LSSVM) is applied to predict some gasoline properties, viz. specific gravity (SG), motor octane number (MON), research octane number (RON), and Reid vapor pressure (RVP). A comprehensive error analysis is also undertaken to compare the values predicted from the model with actual data which enables one to evaluate the performance of the model developed in this study. The results indicate that the model developed has reasonable accuracy and prediction capability. The correlation indices, R 2 , are 0.990, 0.933, 0.955, and 0.920 for SG, MON, RON, and RVP, respectively.

Predicting Octane Number Using Nuclear Magnetic Resonance Spectroscopy and Artificial Neural Networks

“…Several correlations and methods have been reported in the literature to predict the ON of pure hydrocarbons, − PRFs, , toluene primary reference fuels (TPRFs), − gasoline compounds, naphtha, , gasolines, − gasoline with ethanol, ,− and petroleum fractions . The inputs for these models have been generated by utilizing different analytical techniques such as Fourier transform infrared (FT-IR) spectroscopy, ,,, flame emission spectroscopy, nuclear magnetic resonance (NMR) spectroscopy, ,, dispersive fiber-optic Raman spectroscopy, dielectric spectroscopy, gas chromatography, distillation curves, thermal wave interferometry and ignition delay time (IDT) measured in an ignition quality tester (IQT) .…”

“…Several correlations and methods have been reported in the literature to predict the ON of pure hydrocarbons, 3−6 PRFs, 7,8 toluene primary reference fuels (TPRFs), 8−11 gasoline compounds, 12 naphtha, 13,14 gasolines, 15−22 gasoline with ethanol, 7,23−27 and petroleum fractions. 16 The inputs for these models have been generated by utilizing different analytical techniques such as Fourier transform infrared (FT-IR) spectroscopy, 5,13,28,29 flame emission spectroscopy, 30 nuclear magnetic resonance (NMR) spectroscopy, 23,31,32 dispersive fiber-optic Raman spectroscopy, 19 dielectric spectroscopy, 18 gas chromatography, 17 distillation curves, 15 thermal wave interferometry 21 and ignition delay time (IDT) measured in an ignition quality tester (IQT). 22 The data from these techniques have been analyzed by a number of statistical and theoretical methods such as multiple linear regression (MLR), 23 partial least-squares (PLS), 30 quantitative structure property relationship (QSPR), 3,4 response surface methodology, 10,33 and artificial neural networks (ANNs) 6,33−35 to process the data and yield the prediction models.…”

Reliable method for the determination of surfactant retention in porous media during chemical flooding oil recovery

“…They reported that the method proposed can estimate the nitrogen-oil minimum miscibility pressure with an average absolute relative error of 10.02%. Kamari et al [27] applied least square support vector machine for prediction of gasoline properties, viz. specific gravity, motor and research octane number (RON), and Reid vapor pressure.…”

Parametric and Nonparametric Approaches of Reid Vapor Pressure Prediction for Gasoline Containing Oxygenates: A Comparative Analysis Using Partial Least Squares, Nonlinear, and LOWESS Regression Modelling Strategies with Physical Properties

“…Several linear and nonlinear multivariate regression techniques, including partial least squares regression (PLS), artificial neural network (ANN), support vector machine (SVM), and principal components regression (PCR) [12][13][14][15], have been used to successfully predict the RVP of gasoline based on data from spectral analysis or physical properties. Although these studies predict the RVP of gasoline by the regression calibration methods, using either spec-troscopic analysis or physical properties has reached good standard error values, but still, it is required to explore other regression methods to overcome the difficulties that might arise due to various processes, adulteration, and blending causing tremendous variability of gasoline types.…”