Residual oil saturation in Pre-salt Brazilian carbonates: A multi-approach core-to-log case study

Abstract

Accurate residual oil saturation (Sor) measurement is paramount for successful reservoir management and has significant economic implications. Considerable uncertainty surrounds its estimation due to inaccurate direct measurements and non-representative lab results. This case study combines well logging, petrography, RCAL, and SCAL to estimate and understand Sor in Pre-salt reservoirs in Santos Basin, Brazil. The results indicate that the distribution of Sor is strongly influenced by lithology. In-situ facies have higher Sor values than reworked facies due to their greater textural and porous complexity. It has been observed that laboratory experiments and well logging results show a correlation between irreducible water saturation, porosity, and permeability with Sor, thus validating the upscaling of laboratory data. The dielectric tool provides the most accurate results in Sor logging evaluation. This is due to its shallow depth of investigation, high sampling rate, and the elimination of uncertainty associated with electrical parameters and salinity used in the Archie equation. Nuclear magnetic resonance (NMR) logging can qualitatively identify regions with higher remaining oil, which can be correlated with areas of high Sor. Still, it cannot quantitatively directly assess residual saturation by the conventional cutoff approach. Conversely, machine learning results indicate that NMR T2 times can be used to predict Sor values obtained by dielectric logging.