Predicting feedstock and percent composition for blends of biodiesel with conventional diesel using chemometrics and gas chromatography-mass spectrometry.

The two main goals of the analytical method described herein were to (1) use principal component analysis (PCA), hierarchical clustering (HCA) and K-nearest neighbors (KNN) to determine the feedstock source of blends of biodiesel and conventional diesel (feedstocks were two sources of soy, two strains of jatropha, and a local feedstock) and (2) use a partial least squares (PLS) model built specifically for each feedstock to determine the percent composition of the blend. The chemometric models were built using training sets composed of total ion current chromatograms from gas chromatography-quadrupole mass spectrometry (GC-qMS) using a polar column. The models were used to semi-automatically determine feedstock and blend percent composition of independent test set samples. The PLS predictions for jatropha blends had RMSEC=0.6, RMSECV=1.2, and RMSEP=1.4. The PLS predictions for soy blends had RMSEC=0.5, RMSECV=0.8, and RMSEP=1.2. The average relative error in predicted test set sample compositions was 5% for jatropha blends and 4% for soy blends.

Predicting percent composition of blends of biodiesel and conventional diesel using gas chromatography-mass spectrometry, comprehensive two-dimensional gas chromatography-mass spectrometry, and partial least squares analysis.

The percent composition of blends of biodiesel and conventional diesel from a variety of retail sources were modeled and predicted using partial least squares (PLS) analysis applied to gas chromatography-total-ion-current mass spectrometry (GC-TIC), gas chromatography-mass spectrometry (GC-MS), comprehensive two-dimensional gas chromatography-total-ion-current mass spectrometry (GCxGC-TIC) and comprehensive two-dimensional gas chromatography-mass spectrometry (GCxGC-MS) separations of the blends. In all four cases, the PLS predictions for a test set of chromatograms were plotted versus the actual blend percent composition. The GC-TIC plot produced a best-fit line with slope=0.773 and y-intercept=2.89, and the average percent error of prediction was 12.0%. The GC-MS plot produced a best-fit line with slope=0.864 and y-intercept=1.72, and the average percent error of prediction was improved to 6.89%. The GCxGC-TIC plot produced a best-fit line with slope=0.983 and y-intercept=0.680, and the average percent error was slightly improved to 6.16%. The GCxGC-MS plot produced a best-fit line with slope=0.980 and y-intercept=0.620, and the average percent error was 6.12%. The GCxGC models performed best presumably due to the multidimensional advantage of higher dimensional instrumentation providing more chemical selectivity. All the PLS models used 3 latent variables. The chemical components that differentiate the blend percent compositions are reported.