Investigation of enhanced oil recovery by Water Alternating Gas (WAG) injection in sandstone and carbonate rocks

Abstract

Water-Alternating-Gas (WAG) injection is regarded as a very efficient and successful EOR technique for lowering the residual oil in a porous medium, especially from water-flooded/gas-flooded hydrocarbon reservoirs. However, as a special case of three-phase flow, WAG injection involves complex physics and mechanisms in the process of oil recovery, for which there is still an incomplete understanding. In the literature, laboratory data on WAG processes are very limited, especially for certain conditions, such as, the ultra-low oil/gas interfacial tension (IFT), weakly/non-water-wet, non-uniform wettability and different rock types. For numerical simulation, reliable three-phase relative permeability (kr) data with their hysteresis effects are crucial for optimizing the predictions of WAG injection performance in oil reservoirs. Although the current models of three-phase relative permeability (e.g. Stone I, Stone II or Baker) and its hysteresis (e.g. WAG hysteresis) are widely applied for reservoir rocks which are non-water-wet, for their good description of pore geometry and wettability, their application is limited for a water-wet system. Since these correlations, which are available in the most widely used reservoir simulators, were developed on the basis of idealizing the rock and simplifying the assumptions, none of these correlations are able to account for the behaviours of the complex mechanisms, multi-phase flow, multi-physics processes and hysteresis phenomena involved in oil recovery by WAG flooding, especially for certain reservoir fluid conditions (e.g. ultra-low gas-oil IFT) and for oil reservoirs that are characterized by weakly-water-wet and mixed-wet rocks. Thus, reliable experimental data at realistic reservoir conditions are needed to improve our understanding of the actual mechanisms, complex physics, multi-phase flow and hysteresis behaviours underlying the oil recovery by WAG injection and to develop improved models and methodologies for reliable predictions of the performance of oil reservoirs under WAG injection. In this thesis, an extensive series of WAG coreflood experiments is reported, in which several important fluid/flow/rock characteristic properties (e.g. gas/oil IFT scaling, gas viscosity, oil-gas/oil-water viscosity ratio, steady and unsteady displacements, cyclic hysteresis, rock wettability, rock type, and remaining oil and gas saturations) and operation parameters (slug size, injection order and injection strategy) have been systematically investigated. The results of these experimental investigations are discussed in detail in terms of oil recovery, injectivity (or differential pressure) and average saturation profile (or saturation trajectories). Because the coreflood experiments were carried out on Clashach sandstone and Indiana limestone core samples the content of this thesis can be divided into two parts as follows: The first part presents the results of the investigation of the above-mentioned WAG parameters in sandstone rock with two different wettability systems (mixed-wet and weakly-water-wet). Investigating the effect of design/operation parameters on the performance of WAG flood under immiscible displacement (gas-oil IFT = 2.7 mN.m-1 ) revealed a better efficiency and higher recovery performance for short water and gas slugs compared to the large cycle injections. Comparison with the SWAG flood shows that SWAG is the upper limit of oil recovery by small slug water and gas injections. WAG injection efficiency was investigated at intermediate-miscible displacement (gas/oil IFT = 0.15 mN.m-1 ) and under mixed-wet conditions. The results are compared to those published at near-miscible and immiscible displacements to investigate the impact of IFT on the performance of WAG injection. The results showed that WAG injection performance increases as gas/oil IFT decreases and becomes optimum when approaching the critical pressure. The effect of actual rock wettability on the performance of WAG injection at near-miscible conditions (gas/oil IFT = 0.04 mN.m-1 ) is investigated for a weakly-water system, and then compared to mixed-wet and water-wet regimes. The comparison reveals that, regardless of the type of wettability, the performance of oil recovery by near-miscible WAG is considerably superior. Even though oil recovery efficiency by water flood increases as the direction of wettability changes from water-wet towards mixed-wet, passing through a weakly-water-wet stage, its performance by WAG injection, post waterflooding, decreased for the same direction of wettability changes. Two WAG experiments with two different binary-hydrocarbon fluid systems (C1-nC4 and C1-nC10) were performed under near-miscible and weakly water-wet conditions to investigate the impact of gas viscosity on gas and WAG injections. It was found that the cyclic oil recovery efficiency by water slugs was higher in C1-nC4 WAG than that in C1- nC10. In contrast, for gas cycles, it was higher in C1-nC10 that in C1-nC4. Furthermore, the gap between oil recoveries increased as the number of gas cycles increased. Two-phase and three-phase SS-kr experiments were performed under near-miscible and weakly water-wet conditions. Comparing these results with those for the unsteady state indicated significant differences in the recovery mechanisms, due to the difference in the nature of the displacement experiments. This highlights that the differences in the kr values between SS and USS in the two-phase and three-phase regions are not only pertinent to the non-uniqueness problem that is known to be associated with USS- kr , but also to the nature of the differences in the displacements, flow and mechanisms involved in oil recovery by SS and USS experiments. In the second part, a systematically acquired set of experimental data in which the effect of rock type on two-phase and three-phase flow and displacements have been investigated is presented. All experiments were conducted on a 45 mD water-wet limestone core and their results were compared to those obtained from a 65 mD water-wet sandstone core with similar physical properties to those of the carbonate. These laboratory results revealed that there is an intrinsic heterogeneity (vugs) in the internal pore structure of the carbonate rock. Comparison of WAG injection results revealed that WAG injection in sandstone, with 88.5 % (IOIP %) ultimate oil recovery outperformed that in carbonate, with 71 % (IOIP %). Finally, investigation of the effects of all the above pertinent WAG parameters on the remaining oil and gas saturations revealed that the order of fluid injection, gas/oil IFT and rock type are the most effective parameters on the slope of So,rem vs. Sg,rem trend line, which is represented by α in the WAG hysteresis model.