Evaluation and Comparison of Residual Oil Saturation Determination Techniques

Abstract

SummaryThe amount and distribution of residual oil saturation (ROS) are critical parameters for determining whether to apply an EOR process to a reservoir. A brief review of available ROS techniques is presented, indicating advantages, limitations, problems, and possible improvements of each technique. Advantages and disadvantages of each ROS-determination technique are summarized. Screening criteria for determining the best ROS technique under certain wellbore or reservoir conditions are presented.This paper also presents results from comparisons of ROS measurements obtained from the literature as calculated from resistivity logs, pulsed neutron capture (PNC) logs, pressure coring, single-well tracer tests, nuclear magnetism logs (NML), carbon/oxygen (C/O) logs, and electromagnetic propagation tool (EPT) measurements. In this study, the ROS measured by each method is compared with that determined by other methods conducted in the same well. The comparison shows that average values of ROS determined by C/O log, PNC-LIL (log-inject-log), and single-well tracer test do not differ statistically when compared with other methods. The resistivity log tends to give higher than average [2 saturation units (s.u.)] ROS measurements, while pressure coring tends to give lower than average (4 s.u.) ROS values. EPT and NML show deviations of about 8 s.u. of ROS values from other methods, which indicates a statistically significant difference.ROS vertical profiles obtained by two different methods from the same well were compared to eliminate the ROS variation resulting from formation depth. The vertical profiles based on ROS zoning and foot-by-foot measurements were studied to provide more "resolution" for comparisons. The results show that discrepancies in measurement methods are more pronounced when vertical profiles are divided into different zones. This could mean that the discrepancies are much greater for some zones than for others. This approach offers the possibility of studying ROS-method discrepancies as a function of different ROS values.