Classifying rock lithofacies using petrophysical data

Abstract

This study automates a type-curve technique for estimating the rock pore-geometric factor (λ) from capillary pressure measurements. The pore-geometric factor is determined by matching the actual rock capillary pressure versus wetting-phase saturation (Pc–Sw) profile with that obtained from the Brooks and Corey model (1966 J. Irrigation Drainage Proc. Am. Soc. Civ. Eng. 61–88). The pore-geometric factor values are validated by comparing the actual measured rock permeability to the permeability values estimated using the Wyllie and Gardner model (1958 World Oil (April issue) 210–28). Petrophysical data for both carbonate and sandstone rocks, along with the pore-geometric factor derived from the type-curve matching, are used in a discriminant analysis for the purpose of developing a model for rock typing. The petrophysical parameters include rock porosity (), irreducible water saturation (Swi), permeability (k), the threshold capillary-entry-pressure (Pd), a pore-shape factor (β), and a flow-impedance parameter (n) which is a property that reflects the flow impedance caused by the irreducible wetting-phase saturation. The results of the discriminant analysis indicate that five of the parameters (, k, Pd, λ and n) are sufficient for classifying rocks according to two broad lithology classes: sandstones and carbonates. The analysis reveals the existence of a significant discriminant function that is mostly sensitive to the pore-geometric factor values (λ). A discriminant-analysis classification model that honours both static and dynamic petrophysical rock properties is, therefore, introduced. When tested on two distinct data sets, the discriminant-analysis model was able to predict the correct lithofacies for approximately 95% of the tested samples. A comprehensive database of the experimentally collected petrophysical properties of 215 carbonate and sandstone rocks is provided with this study.