Water-alternating-gas (WAG) injection has been used widely for improving oil recovery by combining the high sweep efficiency of Waterflooding (WF) and microscopic displacement efficiency of gas injection (GI). The process also improves the vertical sweep efficiency by reducing gravity segregation. Although, the majority of oil reservoirs are mixed-wet and most of the successful WAG injection schemes involve low gas/oil interfacial tension (IFT) due to the injection gas being either high-pressure hydrocarbon gas or CO2, the existing predictive approaches are based on water-wet conditions and high gas/oil IFT systems. Reliable laboratory data on WAG injections under realistic reservoir conditions (i.e., mixed-wet and low gas/oil IFT) are invaluable for better understanding of the complex multi-phase and multi-physics processes involved in WAG injection. Such measured data would also be vital for a proper assessment of the formulations available in commercial reservoir simulators for accounting for cycle hysteresis taking place in WAG injection. The objective of the present work is to provide unique set of experimental data with the associated theoretical studies under mixed-wet and low gas/oil IFT conditions.
We report the results of a comprehensive series of coreflood experiments carried out under different pressures corresponding to three different levels of gas/oil IFT namely, ultra-low, intermediate, and high gas/oil IFT values of 0.04, 0.15, and 2.70 mN.m-1 in mixed-wet rocks. Coreflood experiments included waterflooding (WF), gas injection (GI) and two WAG injection scenarios at each IFT value. In the first series of WAG experiments, fluid injection started with water injection (I) followed by gas injection (D), and this cyclic injection of water and gas was repeated in four cycles (WAG-IDIDIDID). In the second series of WAG experiments, the test started with gas injection (D) followed by water injection (I), and this cyclic injection of water and gas was repeated four times (WAG-DIDIDIDI). In addition to these experiments, for the high and ultra-low gas/oil IFT systems, SWAG injection experiments have also been performed with SWAG ratio of unity (Qg/Qw = 1).
The results showed that the performance of GI was higher in the case of lower IFT condition compared to high-IFT system. The effect of gas/oil IFT was more pronounced in high permeable mixed-wet rock than it was in low permeable mixed-wet system. That is the improvement in recovery obtained by reducing gas/oil IFT, was more significant for high permeability core than it was for the low permeability one. Interestingly, for all IFT values tested, WF performance was better than GI under mixed-wet condition. This is in contrast with our previous results obtained under water-wet conditions. The results also showed that under mixed-wet conditions, for the three gas/oil IFT levels tested, WAG injections outperformed WF and GI. For the ultra-low IFT condition, oil recovery by the WAG-IDIDIDID experiment was higher than that of the WAG-DIDIDIDI experiment. However, at the other two IFT values, WAG-DIDIDIDI outperformed WAG-IDIDIDID injection scenario. For WAG-IDIDIDID, the lower the gas/oil IFT the higher the ultimate oil recovery; conversely, for the WAG-DIDIDIDI injection scenario, oil recovery performance was better for the high IFT condition rather than the ultra-low IFT case. Our results show considerably higher injectivities during WF periods of the ultra-low IFT WAG injections compared to high-IFT WAG injections. In general, injectivity was lower for the WAG-DI injection scenarios compared to the WAG-ID. The effect of gas/oil IFT on oil recovery was more significant under three-phase flow (WAG injections) compared to the two-phase flow (primary GI).Trapped gas saturations Sgt (for the same Sgi) were found to be higher under higher IFT conditions, and the trend of Sgt vs. Sgi curve was significantly affected by the sequence of fluid injection during WAG injection (DIDIDIDI or IDIDIDID). This is especially true for intermediate and high IFT conditions.
The results show that trapping models such as Land, Carlson, and Jerauld models cannot capture the observed trend of trapped gas saturations accurately, under the conditions of our experiments. This is especially true for the WAG-DIDIDIDI injection scenarios in which, contrary to the WAG-IDIDIDID injections, Sgtw values are not necessarily higher for the case with higher initial gas saturation. This shows the importance of developing new trapping models for non-water-wet systems. In addition, the results show that the reduction coefficient in Sor adjustment formula of the WAG-Hysteresis model is a function of both gas/oil IFT and fluid injection sequence and it also depends on the rock permeability. These further highlight the importance of performing laboratory experiments under representative reservoir and operational conditions.
The current study presents details of a new and rarely available set of experimental data measured under more realistic reservoir conditions of mixed-wet and low gas/oil IFT. In addition, the paper offers novel insights into the mechanisms involved in three-phase flow taking place in WAG injection for different gas/oil IFT values and different WAG injection scenarios. The presented experimental data can be used to assess the validity of three-phase relative permeability and hysteresis models. The results can also be used to obtain representative three-phase relative permeabilities for each phase (gas, oil and water) and their corresponding hysteretic behaviour for different cycles of WAG injection.