Ata Sagnak, Texaco E&P, & Harold Gurrola & George Asquith, Texas Tech University
The determination of the correlation between petrophysical data, and core data is essential in the determination of reservoir rock properties. The goal of this study was to design a computer algorithm, which will use the geophysical inverse theory to deduce different reservoir facies, from well log responses, by utilizing statistical relationships between the well logs and core derived lithofacies information. In most oil and gas fields only a minority of the wells are cored. As a result, determination of reservoir rock properties is mainly dependent on the interpretation of geophysical log data. An objective approach to analyze well log data to determine the reservoir rock properties, would speed up the interpretive process and would also enable the researcher to correlate information between wells and incorporate a-prior knowledge into their interpretation. Detailed core and petrographical analysis was conducted as the first step in establishing statistical relationships of lithological data from four cored wells. Petrographically sixteen major rock types (lithofacies) were identified. Secondly, a valid forward model, which is a requirement in any successful inversion process, was constructed by the usage of mathematical formulation of well logs, and the petrographical data obtained from cores. Well logs used for this study include Neutron, Spectral Density including Bulk Density and Photoelectric Absorption Index (Pe), and Borehole Compensated Sonic (Interval Transit Time) logs. A unity constraint was also used as a supplementary log data. Thirdly, an inversion method, which can incorporate the results of core analysis and petrographical information as apriori geologic information, as potential constrains in the inversion itself was determined. The facies observed between wells in a given oil and gas field are related to each another. Therefore, using the a-priori information from the cored wells, with reservoir facies control, to determine if results from selected well(s) can provide lithofacies information in the remaining wells would improve the reservoir management efforts. An inversion algorithm, using a-priori geologic information was tested on four cored wells. Inversion method was tested for different a-priori geologic information cases where the examined facies were grouped in sixteen, fourteen, eleven, and five lithofacies classes. The robustness of the inversion results was found to depend on the a-priori information provided. Lithofacies inversion, using five lithofacies classes, showed reliable estimations for the purpose of depicting possible reservoir versus non-reservoir zones. The method used was also able to reliable identify most of the eleven facies classes. Inversion, using sixteen and fourteen different facies, showed to be a valuable tool in determining the gross lithology.