As in many mature carbonate reservoirs, determination of remaining oil and oil-water contact in the Grayburg-San Andres dolomite formation at Vacuum Field (Permian Basin) is challenged by unknown salinity (due to decades of water flood) and variable electrical rock properties (due to the presence of anhydrite nodules, vugs, and possibly variable wettability). To address these shortcomings, a new protocol combining Log-Inject-Log (LIL) NMR and Relaxation-Diffusion (T2-D) NMR was tested to improve accuracy in remaining oil estimation. In classic LIL-NMR, manganese mud doping is used to decrease the T2 relaxation time of water between two successive NMR T2 log passes. The difference between the two NMR T2 logs is attributed to water, while the NMR signal remaining after doping is attributed to oil. However, it becomes very difficult to correctly differentiate oil from water if doping fails. Because oil signal develops along the D dimension in T2-D logging, while water does not, T2-D NMR makes it possible to identify whether mud doping was successful. In turn, unflushed oil saturation and wettability can be interpreted with more reliability, which was verified against sponge core measurements. Using only one T2-D NMR pass after manganese mud doping also alleviates the need for a pre-doping log pass. The new protocol therefore improves the quality control and simplifies the data acquisition of traditional LIL NMR method.
The Vacuum Field in the US Permian Basin is a typical example of mature oilfield that has been subject to decades of secondary and tertiary floods. Little data was acquired below first produced water during initial field development. Over the years, water floods (started in 1977) and CO2 floods (started in 1997) were implemented to increase recovery of the main pay intervals but hardly tapped oil volumes present under the original first produced water depth.