A Quadruple-Porosity Model for Consistent Petrophysical Evaluation of Naturally Fractured Vuggy Reservoirs

Abstract

Summary Several dual- and triple-porosity models have been proposed for quantifying the porosity exponent (m) in multiporosity reservoirs. Total porosity (ϕ) is usually portioned into the matrix (ϕb) and vuggy porosity, which includes separate vugs (SVGs) and connected vugs (CVGs). As a result, in their majority, the existing petrophysical models were developed and applied mostly without any distinction between the various types of CVGs despite their specific pore geometries, which critically determine the properties of the rock/fluid systems. For instance, unlike otherwise CVGs, natural fractures (NFs) and microcracks that have low pore-aspect-ratio values are highly compressible; this can cause their closure and lead to increasing m values. In this paper, we proposed a quadruple-porosity model that accounts for NFs (ϕ2 or ϕf) and CVGs (ϕc), in addition to ϕb and SVGs (ϕnc) separately, as distinct input variables to ensure accurate determination of m in composite reservoirs. The approach was based on the volume-model method and rules of electric-resistance networks in porous media. Computed water-saturation values used to validate the model show significant improvement and close agreement with the laboratory measurements, demonstrating the applicability of the proposed model for accurate prediction of m in naturally fractured vuggy reservoirs. New correlations that consider the pore-type diversity were generated using a plot of ϕ vs. m, obtained with the proposed quadruple-porosity model. The procedure involved sorting the ϕ/m scattering points using pore-type mixing and relative abundance of specific porosity. It allowed defining consistent ϕ/m relationships, with determination coefficients of 0.7 to 0.9. This suggests that m varies with the pore-structure types; this was further demonstrated with a rock-frame flexibility factor (γ) used as a proxy to cluster the scattering points. The established correlations can alternatively be applied to reasonably predict m using detailed prior knowledge of pore-type description.