Laboratory determination of residual oil saturation (ROS) in carbonate cores is sometimes uncertain due to wide pore size distribution, core scale heterogeneity, and complex wettability. The values obtained in laboratory tests may vary depending on flow rates, the type of samples (plugs or whole cores), and sample preparation techniques.
The purpose of our study is to integrate modern flow visualization technology with conventional laboratory tools to provide a comprehensive picture of waterflood recovery behavior in four carbonate cores. Our data set comprise thin sections; mercury injection; 3–D porosity distribution; oilfloods and waterfloods on cleaned samples; and 3–D flow imaging of miscible floods, oilfloods, and waterfloods on restored state samples.
The 3–D Computed Tomography (CT) images allowed us to understand the reasons for decrease in oil saturation observed with increased pressure drop in the corefloods; whether this is due to capillary-end effects, core scale heterogeneity, or actual reduction in ROS. We find that ROS values under field-rate flooding conditions (~ 1 psi/foot, Nc* ~ 10-8, lateral flood) are in the 30%-60% PV range. These ROS values reduce significantly as the pressure gradient applied during the floods is raised from field values to the much higher-pressure gradients sometimes used in laboratory testing (~ 100 psi/ft, Nc ~ 10-6). The carbonate samples with large pore-throat aspect ratios have the largest ROS values and the biggest variation with the pressure drop used in the waterfloods.
Introduction
ROS from Laboratory Floods.
Most of the early laboratory waterfloods reported in the literature were conducted at high flow rates (Rapoport and Leas, 1953) to eliminate the capillary end effect at the outlet of the core sample. This method is still appropriate for very strongly water wet or very strongly oil-wet rocks. In the former case, waterflooding is a strong imbibition process, and we have a well-defined residual oil saturation (ROS) that is constant over a very wide range of flow rates. The ROS decreases only when flow rates become so large (Nc >10%5) that the trapped blobs of oil are mobilized. In the case of strongly oil-wet media, waterflooding is a drainage process. The waterflooded oil is in the form of continuos films, and residual oil saturation is theoretically close to zero. The issue here is one of determining the oil relative permeability at low oil saturation.
Many reservoir rocks are now thought to be have intermediate or mixed wettability (Cuiec, 1991). Waterfloods in such systems are potentially a combination of drainage and imbibition. Laboratory measurements become a problem as we may experience a drainage capillary end effect, encouraging use of high rates; and there may also be a rate dependence in the imbibition trapping mechanism, suggesting the use of reservoir rates (Heaviside, 1991). Any decrease in the oil remaining in the core with increase in waterflood rate can be due to reduction in capillary end effects and/or due to change in the microscopic trapping mechanisms. The first factor is a laboratory artifact and does not reflect behavior in the field. The change in microscopic trapping is the true change in ROS and can be exploited by changing field conditions.