Measurement of Diesel Cetane Number Using Near Infrared Spectra and Multivariate Calibration

Abstract

Abstract.Near infrared spectroscopy (NIRS) was investigated for measurement of diesel cetane. Three types of pretreatments including standard normal variate (SNV), multiplicative scattering correction (MSC) and Savitzky-Golay smoothing combining first derivative were adopted to eliminate the system noises and external disturbances. Then, partial least squares (PLS) and least squares-support vector machine (LS-SVM) methods were implemented for calibration models. Simultaneously, the performance of least squares-support vector machine (LS-SVM) method was compared with three kinds of dimension reduction input, including principal components (PCs), latent variables (LVs), and effective wavelengths (EWs). The best predictions were obtained with LS-SVM-LVs model for diesel cetane number (R 2 Pre =0.557, RMSEP = 2.169 and residual prediction deviation (RPD) =1.61), which was deemed as good model predictions. It is recommended to adopt LS-SVM-LVs model technique for higher accuracy measurement of the selected diesel cetane number with NIR, in comparison with LS-SVM-PCs and LS-SVM-EWs model techniques.

Diesel cetane number estimation from NIR spectra of hydrocracking total effluent

“…The developed model uses 9 LVs to explain by about 99% the variance of the studied property. Considering the most recent studies regarding the estimation of diesel cetane number from NIR spectroscopy, the model developed in this study presents an RMSEP (2.0) comparable to the one obtained by Zhan et al [20] (2.0) but presenting a better r²P (0.96 vs. 0.55). Additionally, compared to the regression method employed by them (LS-SVM), by using the PLS method in this study, the obtained model was interpretable, helping to understand the chemical information of the total effluent having an impact on the diesel cetane number.…”

“…Regarding the diesel cetane number estimation using NIR spectroscopy, the most recent studies are reported by Zhan et al [20] and Barra et al [21]. In the first study, a least squares-support vector machine (LS-SVM) regression model was developed with errors of calibration (1.8) and prediction (2.0) lower than the reproducibility of the ASTM D613-01 standard method (~3.3).…”