In this work, the reservoir dimensions used are 8000*6000*1000 ft for X, Y, and Z directions. The reservoir model consists of 50 grid blocks in the horizontal direction and 40 grid blocks in the vertical direction. The number of blocks was selected by running multiple scenarios of different block numbers, and the used grid number was found to be representative and computationally accepted. It should be noted that the same number of grids was used for all scenarios in order to minimize the uncertainty and improve the work reliability. Also, the reservoir formation was divided into five layers to account for the reservoir heterogeneity. The first three layers represent the oil zones, while the water aquifer is represented by the two layers at the bottom. The porosity distribution is 18, 17, 10, 18, and 15% for the five layers, while the layers’ permeability was varied between 75 and 100 mD. A thin layer of 5 ft thickness and 0.01 mD permeability was placed between the water aquifer and the reservoir formations, to help in the slow release of CO₂. Moreover, other reservoir properties such as the reservoir pressure of 4,000 psi, minimum miscibility pressure (MMP) of 1600 psi, and gas oil contact of 9,000 ft were used, which are the average values for typical reservoirs in the region. Figure 1 shows the reservoir model used in this study.

In addition, the well properties are defined as follows: the maximum injection pressure for injectors is 5,500 psi, the operating bottom hole pressure (BHP) for producers is 1800 psi, the maximum allowable gas–oil ratio (GOR) is 8000 SCF/bbl, and the maximum allowable water cut is 0.99. Also, a well radius of 0.625 ft and a skin factor of 1.5 were used, as suggested by many studies (Jacob, 1947; Bresciani et al., 2020). Figures 2, 3 show the relative permeability data used in this work for oil–water and gas–oil systems, respectively. Also, the relative permeability to oil was determined as a function of gas, water, and oil saturation, as shown in Figure 4. The relative permeability profiles were obtained using Correy’s model available in CMG software (Correy et al., 1956).

The simulation work was carried out using CMG (Computer Modeling Group) software, and CMG Launcher 2015.10 (Builder 2015.1) was used in this work. Various scenarios of gas injection were simulated. All the simulations were started with the residual oil saturation (S_or) condition, representing depleted oil reservoir conditions. Three main scenarios were examined in this work: CO₂ injection into the reservoir, CO₂ injection into the aquifer, and CO₂ injection into the aquifer followed by water injection. CO₂ is highly soluble in water (Xing et al., 2012); the CO₂ solubility in the formation brine is a strong function of pressure, temperature, and salinity (Duan and Sun 2003; Duan et al., 2006). Increasing the pressure showed a rapid increase in the CO₂ solubility, while the CO₂ solubility decreased with the increasing temperature or salinity. The models developed by Duan and Sun (2003) and Duan et al. (2006) were used in this work to estimate the CO₂ solubility in water.

In addition, in all scenarios, three well schemes were utilized: horizontal wells (injectors and producers), vertical wells, and vertical injectors and horizontal producers. Therefore, a total of nine scenarios are discussed in this work. The profiles of oil recovery, reservoir pressure, and CO₂ saturation are analyzed. The reservoir performance was examined for a period of 20 years. Around 7,928 tons of carbon dioxide can be sequestered within the reservoir/aquifer formations, on average. It should be noted that the wells’ number and locations can affect the simulation results; hence, a consistent reservoir/well model was used in this work in order to minimize the uncertainty and improve the work reliability.

In this case, carbon dioxide is injected directly into the reservoir to enhance oil recovery through multiple mechanisms: oil swelling, viscosity reduction, and oil displacement. Three well configurations were used in this case. The first configuration is that both the producer and the injector were drilled horizontally, where one producer and one injector were used. The second configuration is that both producers and injectors were drilled vertically, where three producers and three injectors were used. The third configuration is that the injector wells were drilled vertically, while the producer wells were drilled horizontally, where three vertical gas injectors and one horizontal producer were used. Figure 5 shows the average oil saturation for the studied well configurations. In this work, the simulation duration is 20 years, and the maximum bottom hole pressure of 5,500 psi was used for the injectors, while a maximum oil production rate of 30000 bbl/day was defined for the producers. For all cases, the average oil saturation decreases with time until reaching an oil saturation less than 10%, indicating that CO₂ injection is effectively enhancing the oil production. Using three injectors and three producers showed the lowest oil saturation, followed by using horizontal wells for injection and production. However, the case of using three vertical injectors and one horizontal producer showed relatively high residual saturation (around 9.5%) revealing relatively low oil recovery compared to other examined scenarios. Also, an injection period of 10 years could be selected as the optimum injection duration, as no further reduction was observed in the oil saturation after 10 years.

Figure 6 shows the oil recovery factor for the studied well configurations; all cases show a recovery factor more than 65% of the original oil-in-place (OOIP). The highest oil recovery of 72% of the OOIP was achieved using three vertical injectors and three vertical producers, while using horizontal wells showed relatively comparable results: an oil recovery of 70% of the OOIP. However, using three vertical injectors and one horizontal producer showed the lowest oil recovery (around 66%) among the studied cases. The main reason for the difference in oil recovery for different scenarios could be the contact area between the wellbore and the reservoir. Usually, increasing this contact area can lead to an increase in oil recovery. Overall, all studied scenarios showed oil recovery more than 65% of the OOIP, indicating effective CO₂ treatment.

In this case, three well schemes were examined by changing the well type from horizontal to vertical for injectors and producers. The first case is that all wells were drilled horizontally, where one producer and one injector were used. The second case is that all wells were drilled vertically, where three producers and three injectors were used. Finally, horizontal producers and vertical injectors were studied, where three vertical gas injectors and one horizontal producer were used. Similar to the previous case of CO₂ injection into the reservoir, a simulation time of 20 years was used to examine the recovery profiles. Also, the maximum pressure of 5,500 psi was used for the injectors, and the maximum production of 30000 bbl/day was defined.

Figure 7 shows the profiles of the residual oil saturation for the three studied scenarios. In all cases, CO₂ was injected only into the aquifer and no CO₂ was injected into the reservoir. It should be notated that a thin layer of low permeability is placed between the aquifer and the reservoir to enable the slow release of CO₂ from the aquifer to the reservoir. For all cases, the minimum oil saturation was achieved after 10 years of injection, and no reduction was observed in oil saturation by continuing CO₂ injection after 10 years. Therefore, the optimum CO₂ injection duration could be selected as 10 years. Among all studied cases, three vertical injectors and three vertical producers showed the lowest oil saturation followed by the case of using one horizontal injector and one horizontal producer, where the residual oil saturation of around 9% was achieved. However, using three vertical injectors and one horizontal producer showed the poorest performance, where the residual oil saturation of 13% was obtained. Moreover, the oil recovery profiles for all studied cases are presented in Figure 8, where CO₂ was injected only into the aquifer. Using three vertical injectors and three vertical producers showed the highest oil recovery of 69% of the OOIP, followed by using a horizontal injector and producer, where an oil recovery of 68% of the original oil-in-place was observed. However, using three vertical injectors and one horizontal producer gave an oil recovery of 58% of the OOIP. Overall, using three vertical injectors and three vertical producers or using one horizontal injector and one horizontal producer showed very efficient oil recovery as indicated by the residual oil saturation and oil recovery profiles.

The performance of CO₂ into the aquifer was assisted by injecting water using a sequential approach; CO₂ was injected, and then waterflooding was implemented. Different well types were used similar to the previously discussed cases; all wells are horizontal, all wells are vertical, and producers are horizontal and injectors are vertical. Also, in all scenarios, a maximum injection pressure of 5,500 psi was used, and a maximum production rate of 30000 bbl/day was applied. Figure 9 shows the average oil saturation during the injection of CO₂ followed by waterflooding treatment. Using horizontal producers and injectors outperforms all other scenarios, where the residual oil saturation of 21% was achieved as opposed to 23.5% observed for other cases. In contrast to all previous scenarios, the residual oil saturation decreases with time even after 10 years of injection. Also, applying waterflooding showed lower performance, which could be attributed to increasing the volume of trapped oil within the injected water. Furthermore, the oil recovery factor for all studied cases is shown in Figure 10. More oil is recovered with time, and no recovery plateau is observed. However, for all cases, the maximum oil recovery is less than 30% of the OOIP indicating poor recovery performance. Injecting water could have resulted in the trapping of more oil. Also, CO₂ is in contact with less volume of oil. In this case, the minimum oil recovery is 18%, and the maximum recovery is around 28%, which was achieved using one horizontal injector and one horizontal producer. Overall, all cases of vertical and horizontal wells showed lower recovery compared to the cases where waterflooding was not implemented.

The oil recovery for all studied scenarios (nine cases) is listed in Table 1 . It should be noted that the same number of wells and configurations was used in the three main scenarios (CO₂ injection into the reservoir, CO₂ injection into the aquifer, and CO₂ injection into the aquifer assisted by waterflooding) in order to minimize the uncertainty of the simulation results. The highest oil recovery (73% of the OOIP) was obtained using CO₂ injection into the depleted reservoir and utilizing three vertical injectors and three vertical producers, while the minimum oil recovery (18% of the OOIP) was achieved for injecting CO₂ into the aquifer followed by waterflooding, where three vertical injectors and one horizontal producer were used. Also, injecting CO₂ into the aquifer without applying waterflooding showed a comparable oil recovery compared to the case of direct injection of CO₂ into the reservoir. The simulation results showed that the injected CO₂ will be stored within the reservoir and aquifer systems, as a small amount of CO₂ (around 12% of the injected volume) was produced. Around 80–90% of the injected CO₂ will be sequestered in the reservoir formations, and the produced gases can be re-injected into the reservoir. Utilizing a constant injection approach, the estimated volume of CO₂ that can be sequestered in the underground formations is around 7,928 tons, on average.

Moreover, the performance of CO₂ injection is compared to that of the conventional waterflooding treatment. Figure 11 shows the oil recovery factor vs. time for CO₂ injection into the aquifer and waterflooding cases, where all wells are drilled horizontally. CO₂ injection showed considerably higher recovery compared to water flooding: 68 vs. 12% of the OOIP. The CO₂ approach outperforms water injection although one horizontal well was used for injecting CO₂, while two horizontal wells were used for water injection. Moreover, the injection of CO₂ into the aquifer provides dual benefits: CO₂ sequestration and attractive oil production. Injecting carbon dioxide into the aquifer results in slow release of CO₂, which will interact with the oil and reduce the oil viscosity.

In this work, the geochemical reactions between the injected carbon dioxide and the reservoir/aquifer systems were not examined. Injecting CO₂ may alter the rock properties such as porosity and permeability, due to the geochemical reactions (Cui et al., 2021). Hence, it is recommended to conduct a geochemical study to assess the changes in the reservoir and aquifer properties. Moreover, the distance between the injection and production wells can significantly affect the oil recovery profiles. However, in this work, we did not change the well location, as the main objective of the current study is to compare the performance of three main scenarios (CO₂ injection into the reservoir, CO₂ injection into the aquifer, and CO₂ injection into the aquifer followed by water injection). Further studies can be carried out to assess the impact of well locations on oil recovery during different CO₂ injection approaches.