Partial Least Squares (PLS) and Multiple Linear Correlations between Heithaus Stability Parameters (Po) and the Colloidal Instability Indices (CII) with the 1H Nuclear Magnetic Resonance (NMR) Spectra of Colombian Crude Oils

Abstract

Various methods were developed to predict the stability of Colombian crude oils, in which the integral areas of the resonance signals from 12 regions of 1H nuclear magnetic resonance (NMR) spectra of 29 widely different crude oils were correlated with the stability parameter of Heithaus (Po) and the colloidal instability index (CII). Correlations between the NMR spectra and properties were obtained using partial least-squares (PLS) regression and multiple linear regression (MLR). The prediction models for Po and CII by PLS had coefficients of determination (R2) of >98 and >99%, respectively, while the cross-validation values (CV, q2) ranged from 89 to 90%, respectively. The models obtained from MLR showed a high adjusted R2 (R2ad) for Po and a lower R2ad for CII. The R2 values of the prediction models for Po ranged from 97 to 98%. The use of these predictive methods is faster, more environmentally friendly, and less expensive than conventional methods. Of the set of crude oils used in this study, it was o...