Industrial Hydrogenation Process Monitoring Using Ultra-compact Near-Infrared Spectrometer and Chemometrics

Abstract

The hydrogenation process of soybean oil is monitored through time-consuming methodologies that demand sample preparation and produce chemical residues. Thus, it is necessary to develop faster low-cost waste-free instrumentation methodologies. The aim of this work was to evaluate an ultra-compact near-infrared spectrometer in tandem with the partial least squares regression (PLSR) or support vector regression (SVR) in the control of the hydrogenation process. Models were used to predict the saturated fatty acids (SFA), unsaturated fatty acids (UFA), monounsaturated fatty acids (MUFA), trans fatty acids (TFA), polyunsaturated fatty acids (PUFA), and the iodine value (IV). The values predicted by the PLSR and SVR models were compared to the experimental values obtained by gas chromatography. A methodology for feature selection was also assessed, which was able to reduce by up to 85% the variables used in the models without loss of performance. The values obtained for root mean square error of cross validation, root mean square error of calibration, root mean square error of prediction, and r 2 remained very close for both PLSR and SVR. Regarding RSD, all values were above 5% for the PLSR models, whereas for the SVR, the RSD presented values lower than 5% for IV and UFA. It is worth noting that the spectrometer used has low cost, effortless assembly, and easy handling, which allows its use in any environment. Through the results obtained, it was demonstrated that the ultra-compact NIRS spectrometer in tandem with PLSR or SVR represent an alternative to monitor the industrial hydrogenation process of soybean oil.