Multiphase flow and hysteresis phenomena in oil recovery by water alternating gas (WAG) injection
Fatemi, S. Mobeen
MetadataShow full item record
Water flooding and gas injection are two widely used improved oil recovery techniques that can be applied individually or combined as water alternating gas (WAG) or simultaneous gas and water (SWAG) injection. Laboratory data on WAG and SWAG injections for non-water-wet systems are very limited especially for near-miscible (very low IFT) gas-oil systems. Near-miscible gas injection represents a number of processes of great importance to reservoir engineers including high pressure hydrocarbon gas injection and CO2 flooding. Simulation of these processes (WAG and SWAG injections) requires three-phase relative permeability (kr) data. Most of the existing three-phase relative permeability correlations (such as Stone-I, Stone-II or Baker) have been developed for water-wet conditions and are unable to adequately account for all the complex multi-phase and multi-physics processes involved in these oil recovery techniques. Another major problem in the prediction of the performance of Water Alternating Gas (WAG) process is the uncertainty associated with the changes in three-phase relative permeability (kr) values of oil, gas and water in different cycles, which is known as cyclic hysteresis. The current approach in the industry (except hysteresis model proposed by Larsen and Skauge) is to use two-phase bounding imbibition and drainage relative permeabilities along with a two-phase hysteresis model (such as Land, Carlson or Killough to generate two-phase scanning curves) and input the result into a three-phase correlation (Stone-I, Stone-II, Baker etc) to simulate hysteresis in WAG injection. The other approach in the industry to account for hysteresis in WAG injection is the WAG-hysteresis model (proposed by Larsen and Skauge) coupled with Stone-I correlation. None of these models and approaches is developed and assessed based for low oil/gas IFT and/or nonwater- wet system. Nevertheless, the majority of oil reservoirs are believed to be mixedwet and hence, prediction of the performance of WAG injection in these reservoirs is associated with significant uncertainties. Accurate determination of relative permeability values and their hysteresis behaviour is crucial for obtaining a reliable prediction of the performance of water-alternating-gas (WAG) injection in oil reservoirs. Performing reliable laboratory experiments is the key to evaluating the performance of these oil recovery techniques under different reservoir and operational conditions. The experimental data can be also used for assessment of different relative permeability and hysteresis models, and developing new methodologies for reliable simulation of WAG and SWAG injections (if required). The content of the thesis can be divided into two sections: a) two-phase flow and b) three-phase flow. I present the results of comprehensive series of two-phase and three-phase (WAG injections) coreflood experiments for a gas/oil system at near-miscible (IFT= 0.04 mN.m-1) conditions. Two different cores; a high-permeability (1000 mD) and a lower permeability (65 mD) core were used in the experiments and both water-wet and mixedwet conditions were examined. Experimental data have been used to obtain reliable relative permeabilities and investigate their cyclic hysteresis behavior. In the first section of the thesis (two-phase flow), effects of different parameters such as permeability, wettability (water-wet and mixed-wet), immobile water and saturation history on two-phase flow of oil and gas at near-miscible condition have been investigated. Contrary to the open literature reports which are based on high IFT oil/gas, the results (for very-low oil/gas IFT) showed the importance of the wettability and immobile water saturation on the recovery profiles and estimated relative permeabilities. In addition contrary to the near-miscible liquid-liquid systems, it was observed significant hysteresis effect in the gas-liquid system. I have also investigated different two-phase systems (gas-oil, gas-water and oil-water) in mixed-wet systems. This is crucial, considering the importance of the two-phase systems as a backbone to better understand three-phase flow as well as their importance as an input to two-phase hysteresis models (for simulation of WAG including hysteresis). The investigation in this study shows that currently available two-phase hysteresis models in simulators (Carlson and Killough) are not able to capture the observed cyclic hysteresis behavior in these systems. The results suggest that for mixed-wet systems, it is necessary to consider irreversible hysteresis loops for both the wetting and non-wetting phases. Such capability currently does not exist in reservoir simulators due to lack of appropriate predictive tools. Results highlight the differences between cyclic hysteresis behaviors of the relative permeabilities in these three systems. In the second section of the thesis, I first evaluated the performance of different injection scenarios in the mixed-wet system. These processes include primary waterflooding (WF), primary gasflooding (GF), WAG injection (either starting with water injection or gas injection), and SWAG injection (with different gas/water ratios). For some of these processes (WF, GF and WAG injection started with primary WF) the effect of wettability was also investigated. The results show that in both the water-wet and mixed-wet cores, the performance of WAG injection is better than water injection and gas injection alone. The results show that in mixed-wet core, oil recovery by the WAG test which had started with water injection was higher than the WAG test started with gas injection. WAG injections had superior performance over SWAG injections. SWAG performed better compared to primary gas injection. However, surprisingly, SWAG resulted in lower oil recovery compared to primary waterflood in the mixed-wet system. Compared to the other injection strategies, a very high pressure drop across the core was observed during SWAG injection indicating injectivity problems with the application of the process in mixed-wet rocks. Using results of the WAG injection experiments, I also investigated the cyclic hysteresis effect on three-phase relative permeabilities of each phase (gas, oil and water). The results show the importance of properly accounting for irreversible kr hysteresis loops in the processes involving cyclic injection under three-phase flow conditions. Gas relative permeability (krg) dropped in successive cycles under both water-wet and mixed-wet conditions. krg hysteresis was larger in the water-wet system compared to the mixed-wet case. The results also reveal saturation history dependency for oil relative permeability (kro), which tends to increase in successive gas injection periods. The improvement in kro was larger in the water-wet system. In both water-wet and mixed-wet systems, the largest krw hysteresis happens for the transition from two-phase (oil/water system) to three-phase system (from 1st water injection into 1st gas injection) and the subsequent WAG cycles does not show much hysteresis for krw in the experiments. I addressed some serious shortcomings of the existing reservoir simulators for reliable simulation of oil recovery processes involving three-phase flow and flow reversal.