Adapting Desorption Mass Spectrometry and Pattern Recognition Techniques to Petroleum Fluid Correlation Studies
James C. Hickey, S. L. Durfee
Petroleum explorationists are often faced with determining the relationship between the products of wells completed in lithologies that may have some spatial or communicative relationship. Conventional methods of sampling and analysis are often time consuming and expensive. A new method for the sampling, analysis, and computerized data interpretation of the C2-C16 fraction of crude oil and natural gas is reported here.
Controlled temperature headspace sampling of crude oils and direct pressure equilibrated natural gas exposure of carbon adsorption wires has been successfully applied to the sampling of the volatile fractions of petroleum fluids. Thermal vacuum desorption followed by mass spectrometric analysis of these volatile organic compounds is a rapid and sensitive method for obtaining detailed information of the distribution (fingerprint) of the components in a given sample; however, the resulting information is too complex for direct human interpretation. Techniques of computerized chemical pattern recognition such as principal components analysis (PCA) with graphical rotation, discriminant analysis, and similarity analysis (SIMCA) have proven useful in establishing the relationships between p tentially correlated samples via the fingerprints of their volatile fractions.
Studies have been conducted on multiple samples from numerous continental basins. The results of several of these studies will be presented to demonstrate the applicability of this new, rapid, cost-efficient approach to correlation studies.
AAPG Search and Discovery Article #91038©1987 AAPG Annual Convention, Los Angeles, California, June 7-10, 1987.