Rapid measurement of biodiesel conversion rate by near-infrared spectroscopy

Yucong Liu, Jiangang Lu, Long

Abstract

This paper introduces a method based on the near-infrared spectroscopy for rapid measurement of the biodiesel conversion rate during the transesterification of triolein with methanol. To build the regression model between the conversion rate and the spectral date, the real conversion rate of glycerides to methyl esters was analyzed by Proton Nuclear Magnetic Resonance (1HNMR) spectroscopy, and several steps were taken to the obtained spectra. Firstly, the raw spectra of all samples obtained were preprocessed by Savitzky-Golay smoothing method. Moreover, eighteen effective wavelength points were selected by successive projections algorithm (SPA). Finally, the partial least squares (PLS) model was constructed with the selected 18 wavelength points. The eventual regression model was validated and the Root Mean Square Error of Prediction (RMSEP) was 0.0115, coefficient of determination (R2) was 0.9877. The total average absolute deviation of the measurements is 1.09%.

Relevant Publications in Journal of Chemical and Pharmaceutical Research