Oil Quality Prediction

Oil quality is a central issue in the economic assessment of liquid hydrocarbon accumulations. Poor oil quality typically diminishes resource value, and can make some prospects uneconomic. The primary sub-surface processes governing oil quality include source rock type and thermal evolution of the source rocks (pre-migration), as well as biodegradation, water-washing and phase separation (post-migration). The combined effects of these processes are often reflected in the chemical composition of crude oils.

GSI's has built a geochemical database of over 800 oils representing the major basins and stratigraphic horizons. Oils have been characterized for physical properties (API gravity, weight percent sulfur, heavy metal content, viscosity and pour point, gross composition) in addition to the use of state-of-the-art oil fingerprinting techniques such as gas chromatography, stable carbon isotopes and analysis of biomarker distributions. The oils are classfied into compositionally-distinct oil families using multivariate statistical methods, in addition, to their thermal maturity and degree of alteration.

These data are used as a training set for an oil quality prediction model. Principal Component Regression (PCR) and Partial Least Squares (PLS) are used to arrive at a predictive algorithm between a quantitative sample property, (for example, oil quality measures such as API gravity, pour point, viscosity and sulfur content), and several independent variables (measured geochemical properties such as C15+ bulk composition. alkane and isoprenoid distributions, triterpane and sterane biomarkers, and fluorescence characteristics).