Reservoir Properties Estimation Using Flow Zone Indicator and Artificial Neural Network Integration: A Case Study

Abstract In today's financially constrained business landscape, companies often grapple with challenges related to allocating capital expenses, resulting in a scarcity of reservoir characterization data. This shortage necessitates the optimization of existing data and the estimation of unavailable reservoir properties. While classical correlations in core analysis traditionally used porosity to predict permeability, the intricate interplay of lithology and pore geometry renders this approach unreliable for exclusive permeability estimation from porosity. This study aims to advance the understanding of the Tortonian reservoir in the Gamma oil field by exploring the combined application of Flow Zone Indicator (FZI), Artificial Neural Network (ANN), and Convergent Interpolation (CI) methodologies. Utilizing data from an exploratory well and four appraisal wells, the study seeks to model the intricate non-linear relationships between Tortonian reservoir properties, determine effective porosity, estimate permeability for uncured wells, and create a comprehensive permeability map for the Tortonian oil reservoir. The results reveal the presence of three distinct rock types within the Tortonian reservoirs and successfully establish estimates for effective porosity and permeability logs. Notably, the generated permeability map demonstrates a direct correlation with the porosity map, validating the proposed methodology. Through the integrated use of FZI, ANN, and CI techniques, the reliability of the porosity-permeability relationship is significantly enhanced, achieving an impressive accuracy of 90%. This study effectively models the nuanced non-linear porosity-permeability relationship within the Tortonian reservoir, offering an economically viable means to enhance reservoir characterization within the constraints of a limited capital budget and accessible data sources.