Economic Optimization of Enhanced Oil Recovery and Carbon Storage Using Mixed Dimethyl Ether-Impure CO₂ Solvent in a Heterogeneous Reservoir

Abstract

CO₂ is the main solvent used in enhanced oil recovery (EOR). However, its low density and viscosity compared to oil cause a decrease in sweep efficiency. Recently, dimethyl ether (DME), which is more efficient than CO₂, has been introduced into the process. DME improves oil recovery by reducing minimum miscible pressure (MMP), interfacial tension (IFT), and oil viscosity. Since DME is an expensive solvent, price reduction and appropriate injection scenarios are needed for economic feasibility. In this study, a compositional model was developed to inject DME with impure CO₂ streams, where the CO₂ was derived from one of these three purification methods: dehydration, double flash, and distillation. It was assumed that such a mixed solvent was injected into a heterogeneous reservoir where gravity override was maximized. As a result, lower oil recovery is achieved for the higher impurity content of the CO₂ stream, lower DME content, and more heterogeneous reservoir. When a high-purity CO₂ stream is used, the change in oil recovery according to DME content and heterogeneity of the reservoir is increased. When the lowest-purity CO₂ stream is used, the net present value (NPV) is the highest. For a homogeneous reservoir, the NPV is highest for all impure CO₂ streams. This optimization indicates a greater impact on revenue from reduced CO₂ purchase cost than on profit loss due to reduced oil recovery by impurities. Additional benefits can be expected when considering solvent reuse and carbon capture and storage (CCS) credits.

1. Introduction

Enhanced oil recovery (EOR) is a method of increasing the amount of oil recovery by injecting a fluid such as steam, polymer, or surfactant into the oil reservoir, ultimately increasing economic returns [1,2]. CO₂ has often been used as a solvent in EOR due to its environmental friendliness as well as being mixed with oil under temperature and pressure conditions in the reservoir [3]. However, the density of CO₂ is slightly lower than that of oil, and the viscosity is much lower [4]. Therefore, the viscous fingering effect, reduction in sweep efficiency, early breakthrough, and unfavorable mobility ratios occur [5,6]. To solve this problem, dimethyl ether (DME), used in fuel additives, aerosol propellants, and heating fuels, was recently introduced as a solvent for EOR [7,8,9]. DME is converted from methane-rich syngas produced by techniques such as methane reforming [10]. Therefore, gas can be liquefied into DME and procured from stranded gas fields that are not utilized due to lack of productivity near oil fields, such as the factories of Royal Dutch Shell that produce liquid synthetic fuel from natural gas in Malaysia [11]. DME is a substance with a critical temperature of 399.4 K, a critical pressure of 5.264 MPa, and a molecular weight of 40.07 [12]. DME is non-toxic and is soluble in both oleic and aqueous phases due to its slight polarity. When DME is injected into the reservoir, it reduces the minimum miscible pressure (MMP) and interfacial tension (IFT), thereby lowering the oil viscosity. It also improves oil mobility by lowering the oil density through oil swelling and can be extracted from chase water and reused [11,12,13,14,15]. Recent EOR studies related to DME have mainly focused on DME-enhanced waterflooding (DEW) [16,17]. In addition, studies on the phase behavior of the oil system at the core scale focused on DME and heavy oil have been conducted [18,19,20,21,22]. DME is an efficient but expensive solvent, so a thorough economic analysis is required [23]. However, most existing studies have focused on oil recovery rather than economic evaluation.

This study aims to suggest a cost-effective measure in which DME and CO₂ are co-injected into heterogeneous reservoirs. It was intended to derive an appropriate optimal value between improving oil recovery and reducing the CO₂ purchase cost using relatively inexpensive impure CO₂. The change in oil properties according to solvent injection was analyzed by applying a compositional model with a water-alternating-gas (WAG) method, known to be effective in CO₂ injection. In this study, the changes in sweep efficiency, displacement efficiency, and oil recovery according to DME concentration and impurity of CO₂ are analyzed considering factors such as oil saturation, oil viscosity, interfacial tension (IFT), and breakthrough. In addition, the optimal injection scenario for maximizing net present value (NPV) is derived considering oil price, impure CO₂ stream price, solvent reuse, and carbon capture and storage (CCS) credits.

2. Methodology

2.1. Fluid Modeling

For fluid modeling, the Winprop 2024.30 software from Computer Modeling Group (CMG) was used [24]. Based on the W3 fluid model from the Weyburn field in Saskatchewan, Canada [19], the phase behavior of the injection fluid including reservoir oil and DME was calculated using the Peng–Robinson (PR) equation of state (EOS) [25]. The components and properties of each component of the fluid for EOS calculation in fluid modeling are shown in

Table 1. Oil components and properties of each component of fluid for EOS calculation [26,27].

Component	Mole Fraction	Critical Pressure (MPa)	Critical Temperature (K)	Molecular Weight
N2	0.0207	3.394	126.2	28.01
CO2	0.0074	7.376	304.2	44.01
H2S	0.0012	8.936	373.2	34.08
C1	0.0749	4.600	190.6	16.04
C2	0.0422	4.883	305.4	30.07
C3	0.0785	4.245	369.8	44.09
C4	0.0655	3.722	416.5	58.12
C5	0.0459	3.379	464.9	72.15
C6–9	0.2155	3.019	556.4	102.5
C10–17	0.2202	2.017	692.3	184
C18–27	0.1027	1.327	808.4	306.2
C28+	0.1252	1.155	915.5	585.61
DME	0.0001	5.442	400.3	46.07

[26,27].

Table 2. Binary interaction coefficients between each component of oil and DME [29].

	N2	CO2	H2S	C1	C2	C3	DME	C4	C5	C6–9	C10–17	C18–27	C28
N2	0.00	–	–	–	–	–	–	–	–	–	–	–	–
CO2	0.00	0.00	–	–	–	–	–	–	–	–	–	–	–
H2S	0.13	0.14	0.00	–	–	–	–	–	–	–	–	–	–
C1	0.25	0.11	0.07	0.00	–	–	–	–	–	–	–	–	–
C2	0.01	0.13	0.09	0.00	0.00	–	–	–	–	–	–	–	–
C3	0.09	0.13	0.08	0.01	0.00	0.00	–	–	–	–	–	–	–
DME	0.10	0.00	0.00	0.29	0.25	0.25	0.00	–	–	–	–	–	–
C4	0.10	0.12	0.08	0.02	0.01	0.00	0.25	0.00	–	–	–	–	–
C5	0.10	0.12	0.07	0.02	0.01	0.00	0.25	0.00	0.00	–	–	–	–
C6–9	0.11	0.12	0.05	0.03	0.02	0.01	0.02	0.00	0.00	0.00	–	–
C10–17	0.11	0.12	0.05	0.06	0.04	0.02	0.08	0.02	0.01	0.01	0.00	–	–
C18–27	0.11	0.12	0.05	0.09	0.06	0.05	0.08	0.03	0.03	0.02	0.00	0.00	–
C28	0.11	0.12	0.05	0.01	0.09	0.07	0.08	0.05	0.05	0.03	0.01	0.00	0.00

shows the values of binary interaction coefficients between each component of the oil and DME [28].

Table 3. Comparison of the fluid model and experimental data from W3 fluid [18,28].

Parameters	Fluid Model	Weyburn W3
Saturation Pressure (MPa)	4.78	4.92
Oil Density at Saturation Pressure (kg/m3)	805.7	806.4
Oil Viscosity at Saturation Pressure (mPa·s)	1.75	1.76
Formation Volume Factor	1.11	1.12
Oil Gravity (°API)	35	31

compares the values used in the fluid model, which were generated through regression analysis, with experimental data from W3 fluid [18,28]. The phase envelopes generated from the fluid modeling with CO₂ injection and DME-CO₂ injection are compared in Figure 1.

2.2. Reservoir Modeling

CMG’s Builder software 20224.30 was used to simulate the hypothetical reservoir. A two-dimensional cross-sectional heterogeneous model was developed in this study to examine the efficiency of the CO₂-DME WAG process in a heterogeneous reservoir. The reservoir model consisted of 52 × 1 × 20 cells, and the size of each cell was 10 ft × 10 ft × 5 ft. Figure 2 shows the schematic of the reservoir model. As shown in

Table 4. Reservoir cases categorized by standard deviation.

Case	Standard Deviation $(\mathsf{\sigma })$
1	0.00
2	5.34
3	10.68
4	21.35

, reservoir modeling was executed considering four heterogeneous reservoirs with the same mean and different standard deviations, where each layer of the reservoir model had a different permeability. A gravity override phenomenon in which less-dense fluid flows preferentially on the top of a reservoir unit and denser fluid flows at the bottom results in early breakthrough and decreased oil productivity. To analyze the effect in the unfavorable case where the gravity override effect is maximized, the model considered a heterogeneous reservoir with a high-permeability layer toward the top. The layer permeability in each case is shown in

Table 5. Permeability distributions for different cases of standard deviation.

	Standard Deviation	0.00	5.34	10.68	21.35
Layer
1		50	59	68	86
2			58	66	82
3			57	64	78
4			56	62	74
5			55	60	70
6			54	58	66
7			53	56	62
8			52	54	58
9			51	52	54
10			50	50	50
11			50	50	50
12			49	48	46
13			48	46	42
14			47	44	38
15			46	42	34
16			45	40	30
17			44	38	26
18			43	36	22
19			42	34	18
20			41	32	14

. In addition, since DME is less effective in the oil-wet reservoir and can change wettability from more oil-wet to more water-wet [30,31,32,33,34], an unfavorable case is assumed using the relative permeability of the oil-wet reservoir as shown in Figure 3 [33].

2.3. Injection Design

In this study, a CO₂ stream containing impurities was used to reduce costs. Impurities such as N₂ and SO₂ increase MMP, which negatively affects oil recovery [34]. The composition and price of the CO₂ stream obtained from the flue gas vary depending on the compression and purification method [35]. Three CO₂ stream compositions are captured through different compression and purification processes from the same flue gas.

Table 6. Composition and price of CO₂ streams with different purification methods.

	Compression and Dehydration (Case 1)	Double Flash CPU (Case 2)	Distillation CPU (Case 3)
CO2 (wt.%)	77.4	96.5	100
Ar (wt.%)	3.19	0.545	-
N2 (wt.%)	11.4	1.55	-
O2 (wt.%)	7.96	1.36	-
SO2 (ppmv)	200	200	-
Price (USD/ton)	16.1	24.1	28.1

shows that CO₂ purity and price increase proportionally in three CO₂ streams with different compositions [28]. The proportions of CO₂ and DME in the injection gas for the four heterogeneous reservoirs are shown in

Table 7. Cases categorized by DME content.

Case	Injection Gas Composition
1	100% CO2 Stream + 0% DME
2	95% CO2 Stream + 5% DME
3	90% CO2 Stream + 10% DME
4	80% CO2 Stream + 20% DME

.

For the injection, the WAG method was used. In the four reservoir models, the impure CO₂ stream was injected with different DME proportions. For the second recovery, water flooding is performed for the first 3 years and the WAG cycle is performed for the next 6 years (1:1 WAG ratio). Finally, in order to recover DME dissolved in water, water flooding was performed for 3 years with chase water. As shown in Figure 4, the design consists of 12 years of injection [28]. The initial reservoir properties from the literature are shown in

Table 8. Initial conditions of the reservoir.

Properties	Values
Initial Pressure (MPa)	13.8
Grid Top (ft)	4000
Temperature (°F)	145
Initial Oil Saturation	0.88
Average Permeability (md)	50
Porosity	0.3
Total Injection (PV)	1.5

[36].

2.4. Economic Analysis

Since DME is an expensive solvent, an analysis of economic feasibility is essential [37]. The greater is the amount of impurities in the CO₂ stream, the lower is the purchase cost. In this study, NPV was calculated when the oil price was 30, 60, and 90 USD/bbl by changing the injection gas composition [38]. In addition to the benefits of oil recovery, solvent reuse and CCS credits are considered to maximize economic benefits. NPV is calculated as follows [39]:

$$\mathrm{N}\mathrm{P}\mathrm{V}={\sum }_{n=1}^N{\displaystyle \frac{{\mathrm{O}\mathrm{i}\mathrm{l}}_{{\mathrm{r}\mathrm{e}\mathrm{v}\mathrm{e}\mathrm{n}\mathrm{u}\mathrm{e}}_n}+{\mathrm{C}\mathrm{a}\mathrm{r}\mathrm{b}\mathrm{o}\mathrm{n}}_{{\mathrm{p}\mathrm{r}\mathrm{i}\mathrm{c}\mathrm{e}}_n}-{\mathrm{R}\mathrm{e}\mathrm{c}\mathrm{u}\mathrm{r}\mathrm{r}\mathrm{e}\mathrm{n}\mathrm{t}}_{{\mathrm{c}\mathrm{o}\mathrm{s}\mathrm{t}}_n}}{{\left(1+r\right)}^n}}$$

(1)

where ${\mathrm{O}\mathrm{i}\mathrm{l}}_{{\mathrm{r}\mathrm{e}\mathrm{v}\mathrm{e}\mathrm{n}\mathrm{u}\mathrm{e}}_n}$ represents revenue from oil production at the nth year, ${\mathrm{C}\mathrm{a}\mathrm{r}\mathrm{b}\mathrm{o}\mathrm{n}}_{{\mathrm{p}\mathrm{r}\mathrm{i}\mathrm{c}\mathrm{e}}_n}$ denotes the price of carbon as incentive for carbon storage at the nth year, and ${\mathrm{R}\mathrm{e}\mathrm{c}\mathrm{u}\mathrm{r}\mathrm{r}\mathrm{e}\mathrm{n}\mathrm{t}}_{{\mathrm{c}\mathrm{o}\mathrm{s}\mathrm{t}}_n}$ is recurrent operation cost at the nth year.

$${\mathrm{O}\mathrm{i}\mathrm{l}}_{{\mathrm{r}\mathrm{e}\mathrm{v}\mathrm{e}\mathrm{n}\mathrm{u}\mathrm{e}}_n}=\left[Q_{op\left(n\right)}-Q_{op\left(n-1\right)}\right]\left({\mathrm{O}\mathrm{i}\mathrm{l}}_{\mathrm{p}\mathrm{r}\mathrm{i}\mathrm{c}\mathrm{e}}\right)$$

(2)

$${\mathrm{C}\mathrm{a}\mathrm{r}\mathrm{b}\mathrm{o}\mathrm{n}}_{{\mathrm{p}\mathrm{r}\mathrm{i}\mathrm{c}\mathrm{e}}_n}=\left\{\left[Q_{gi\left(n\right)}-Q_{gi\left(n-1\right)}\right]-\left[Q_{gp\left(n\right)}-Q_{gp\left(n-1\right)}\right]\right\}(\mathrm{S}\mathrm{t}\mathrm{o}\mathrm{r}\mathrm{a}\mathrm{g}\mathrm{e} \mathrm{t}\mathrm{a}\mathrm{x})$$

(3)

$${\mathrm{R}\mathrm{e}\mathrm{c}\mathrm{u}\mathrm{r}\mathrm{r}\mathrm{e}\mathrm{n}\mathrm{t}}_{{\mathrm{c}\mathrm{o}\mathrm{s}\mathrm{t}}_n}={\mathrm{G}\mathrm{a}\mathrm{s}}_{{\mathrm{p}\mathrm{u}\mathrm{r}}_n}+{\mathrm{G}\mathrm{a}\mathrm{s}}_{{\mathrm{r}\mathrm{e}\mathrm{c}\mathrm{y}\mathrm{c}\mathrm{l}\mathrm{i}\mathrm{n}\mathrm{g}}_n}+{\mathrm{W}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{r}}_{{\mathrm{c}\mathrm{o}\mathrm{s}\mathrm{t}}_n}$$

(4)

where $Q_{op\left(n\right)}$ and $Q_{op\left(n-1\right)}$ denote cumulative oil production till the nth and (n − 1)th year, ${\mathrm{O}\mathrm{i}\mathrm{l}}_{\mathrm{p}\mathrm{r}\mathrm{i}\mathrm{c}\mathrm{e}}$ is the price of oil, $Q_{gi\left(n\right)}$ and $Q_{gi\left(n-1\right)}$ represent cumulative gas injection till the nth and (n − 1)th year, $Q_{gp\left(n\right)}$ and $Q_{gp\left(n-1\right)}$ represent cumulative gas production till the nth and (n − 1)th year, $\mathrm{S}\mathrm{t}\mathrm{o}\mathrm{r}\mathrm{a}\mathrm{g}\mathrm{e} \mathrm{t}\mathrm{a}\mathrm{x}$ is the tax credit for carbon storage, ${\mathrm{G}\mathrm{a}\mathrm{s}}_{{\mathrm{p}\mathrm{u}\mathrm{r}}_n}$ is the CO₂ purchase cost at the nth year, ${\mathrm{G}\mathrm{a}\mathrm{s}}_{{\mathrm{r}\mathrm{e}\mathrm{c}\mathrm{y}\mathrm{c}\mathrm{l}\mathrm{i}\mathrm{n}\mathrm{g}}_n}$ represents the CO₂ recycling cost at the nth year, and ${\mathrm{W}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{r}}_{{\mathrm{c}\mathrm{o}\mathrm{s}\mathrm{t}}_n}$ is the produced water management cost at the nth year.

Table 9. Variables and values used for economic analysis.

Variables	Values
DME Price (USD/ton)	538
CO2 Price (USD/ton)	16.1, 24.1, 28.1
Water Injection Cost (USD/bbl)	1
Oil Price (USD/bbl)	30, 60, 90
Yearly Discount Rate (%)	10
DME Recycling Cost (USD/ton)	254
CO2 Recycling Cost (USD/ton)	12
CO2 Credit Price (USD/ton)	27.61

shows the variables used in the NPV calculation [40,41,42,43].

3. Results and Discussion

3.1. Vertical Sweep Efficiency

The vertical sweep efficiencies for different reservoir heterogeneities, DME contents, and CO₂ stream purities were compared. Figure 5 shows the change in oil viscosity when 0.5 PV solvent is injected into different reservoirs at a DME content of 0% using the dehydration CO₂ stream with lowest purity. Figure 6 and Figure 7 show the same case when using a double flash and distillation CO₂ stream. Figure 8, Figure 9 and Figure 10 show the oil viscosity when 0.5 PV of solvent is injected when the DME content is 20% using dehydration, double flash, and distillation CO₂ streams. When the DME content is high, the oil viscosity decreases more significantly, and the area swept by the injection fluid also increases. If the DME is 0%, there is no deep blue area with oil viscosity less than 0.5, but if the DME is 20%, double flash and distillation CO₂ streams increase the deep blue area significantly. The oil viscosity near the injection well is greater than initial oil viscosity, because the light and intermediate components are displaced and heavy components accounted for a greater percentage of the residual oil [44]. This shows that DME significantly reduces the oil viscosity and confirms that impurities contained in the CO₂ stream adversely affect the oil viscosity improvement. The higher the purity of the CO₂ stream, the better it mixes with the oil, and the more the oil viscosity decreases, which decreases to at least 0.54 cp for dehydration CO₂ streams, 0.46 cp for double flash CO₂ streams, and 0.45 cp for distillation CO₂ streams with 20% of DME. The more heterogeneous the reservoir, the lower the slope in front of the swept area. When the DME content is 20%, the slope decreased from 1.74 to 0.94, 0.54, and 0.45 with the dehydration CO₂ stream when the standard deviation changed from 0.00 to 5.34, 10.68, and 21.35, respectively. When the double flash CO₂ stream was used, the slope decreased from 5.02 to 3.84, 2.19, and 0.86 for the cases with standard deviation of 0.00, 5.34, 10.68, and 21.35. The higher the purity of the CO₂ stream, the steeper the slope, resulting in greater contact with oil in the lower region of the reservoir. In all cases, there is a region after the injection slug where its oil viscosity is the same as or greater than the initial oil viscosity. This is because the light components of the oil are displaced by the solvent, and only the heavy components remain, and the oil viscosity increases again for all cases.

Figure 11 shows the change in IFT at the representative point of the reservoir (13, 1, 5) when the DME content was 20% in dehydration, double flash, and distillation CO₂ streams. The grid (13, 1, 5) could show the change in IFT among the points that were swept in all cases of the purity of the CO₂ stream, reservoir heterogeneity, and DME content. Therefore, it is set as a representative point for comparison. Figure 12 shows results for a standard deviation of 21.35. When the reservoir is more heterogeneous, the time to reach the peak is shorter; the higher the DME content, the lower the overall interfacial tension after the peak. A lower IFT means that the oil and gas are miscible. Currently, oil swelling occurs when the oil is mixed with the light component and the oil viscosity decreases. Therefore, the mobility of the oil is finally increased, and the oil recovery is improved. In a double flash CO₂ stream, the IFT at the end of production in the most heterogeneous reservoir decreased from 1.55 to 1.06, 0.55, and 0.19 dyne/cm as the DME content increased from 0 to 5, 10, and 20%, respectively. In addition, at a DME content of 20%, the time to reach the peak decreased from 1044 to 1009, 983, and 949 days as the reservoir heterogeneity (standard deviation) increased from 0.00 to 5.34, 10.68, and 21.35, respectively. When a distillation CO₂ stream is used, IFT increases again after the end of the WAG cycle, passing through the back of the swept area by the DME-CO₂ mixed fluid due to chase water.

Table 10. Breakthrough time (days): (a) dehydration; (b) double flash; (c) distillation CO₂ stream.

	Standard Deviation	0.00	5.34	10.68	21.35
DME Content (%)
0		1046	1005	973	930
5		1050	1008	976	932
10		1054	1012	980	936
20		1077	1039	1010	967
(a)
	Standard Deviation	0.00	5.34	10.68	21.35
DME Content (%)
0		1266	1179	1119	1054
5		1286	1198	1134	1064
10		1325	1224	1151	1073
20		1448	1221	1150	1101
(b)
	Standard Deviation	0.00	5.34	10.68	21.35
DME Content (%)
0		1301	1206	1137	1065
5		1322	1221	1149	1071
10		1358	1246	1167	1079
20		1472	1233	1160	1107
(c)

shows the time to reach the breakthrough. In the most heterogeneous reservoir, dehydration, double flash, and distillation CO₂ streams are used when the DME content is 20%, and the times to reach the breakthrough are 967, 1101, and 1107 days, respectively. In the event of an early breakthrough, the injection fluid is continuously produced there, negatively affecting oil production. The more heterogeneous the reservoir and the lower the DME content, the earlier the breakthrough occurs. In addition, breakthrough is delayed when a high-purity CO₂ stream is used.

3.2. Displacement Efficiency

The displacement efficiencies were analyzed by comparing breakthrough time and oil saturation. Figure 13 shows the change in oil saturation when 1.0 PV solvent is injected into different reservoirs when the DME content is 0% in a dehydration CO₂ stream with the lowest CO₂ purity, Figure 14 shows results for a double flash CO₂ stream, and Figure 15 shows results for a distillation CO₂ stream. Figure 16, Figure 17 and Figure 18 show the same case when the DME is 20%. When the DME content is 0%, the area with an oil saturation of 0.3 or more accounts for more than 90% of the total in all cases. When the DME content is 20%, the red area with an oil saturation of 0.5 or more is similar. The green area with an oil saturation of 0.3 or less is significantly increased. This demonstrates that DME slightly improves gravity override and can be mixed with more oil at the same volume, significantly improving the displacement efficiency. Also, if the DME content is the same, as the reservoir heterogeneity increases, the area of the region with an oil saturation of 0.5 or more in the lower part is increased.

3.3. Recovery

Oil recovery is expressed as the product of displacement efficiency and sweep efficiency, and the sweep efficiency is calculated by multiplying vertical sweep efficiency and areal sweep efficiency [45]. Since this reservoir model is 2D cross-sectional, oil recovery is a function of vertical sweep efficiency and displacement efficiency. Figure 19 and Figure 20 show the oil recovery using dehydration, double flash, and distillation CO₂ streams depending on the reservoir heterogeneity and the DME content. When using the dehydration CO₂ stream at a DME content of 20%, the oil recovery decreased by 8.9%, from 59.0 to 56.1, 53.8, and 50.1 with increasing reservoir heterogeneity. In the case of double flash CO₂ streams, the decrease is about 14.6% from 70.0 to 64.5, 60.7, and 55.4, which is a greater decrease than that of dehydration. Finally, when distillation CO₂ streams were used, oil recovery was reduced by 15% from 70.5 to 65.1, 61.2, and 55.9. In addition, as the DME content increased from 0 to 5, 10, and 20% in the most heterogeneous reservoir, the oil recovery increased by about 4% from 46.1 to 46.7, 47.6, and 50.1 in a dehydration CO₂ stream, by about 7% from 48.5 to 50.1, 51.8, and 55.4 in a double flash CO₂ stream, and by about 7% from 48.5 to 50.1, 51.8, and 55.4 in a distillation CO₂ stream, respectively. The difference in oil recovery according to the DME content and the heterogeneity of the reservoir increases as the purity of the CO₂ stream increases. This shows that oil recovery is more affected by impurities contained in the CO₂ stream than the DME content.

3.4. Economic Evaluation

DME has a positive effect on sweep efficiency and displacement efficiency, increasing oil recovery. In addition, it is obvious that the oil recovery decreases as the impurity content of the CO₂ stream increases. However, the optimal injection scenario based on the economic analysis is different from that based on maximum oil recovery criteria. Since DME is an expensive solvent, prices must be considered to maximize the economic benefit. Previously, it was confirmed that oil recovery decreased as the reservoir became heterogeneous. Therefore, the optimum injection scenario for maximum NPV was designed for different CO₂ streams and oil prices of 30, 60, and 90 USD/bbl in the most heterogeneous reservoir (

Table 11. Economic evaluation without consideration of CO₂ credits when the standard deviation is 21.35 in a dehydration CO₂ stream.

Oil Prices		30 USD/bbl		60 USD/bbl		90 USD/bbl
	Objective Function	NPV (USD)	Oil Recovery (%)	NPV (USD)	Oil Recovery (%)	NPV (USD)	Oil Recovery (%)
Case
Without Solvent Reuse	CO2 100% + DME 0% (NPV Max at 30 USD/bbl)	225,424	46.1	-	-	-	-
	CO2 89% + DME 11% (NPV Max at 60 USD/bbl)	-	-	489,640	47.7	-	-
	CO2 80% + DME 20% (NPV Max at 90 USD/bbl)	-	-	-	-	703,712	50.1
Solvent Reuse	CO2 100% + DME 0% (NPV Max at 30 USD/bbl)	250,150	46.1	-	-	-	-
	CO2 80% + DME 20% (NPV Max at 60 USD/bbl)	-	-	511,307	50.1	-	-
	CO2 62% + DME 38% (NPV Max at 90 USD/bbl)	-	-	-	-	770,286	56.6

,

Table 12. Economic evaluation with consideration of CO₂ credits when the standard deviation is 21.35 in a dehydration CO₂ stream.

Oil Prices		30 USD/bbl		60 USD/bbl		90 USD/bbl
	Objective Function	NPV (USD)	Oil Recovery (%)	NPV (USD)	Oil Recovery (%)	NPV (USD)	Oil Recovery (%)
Case
Without Solvent Reuse + Credit	CO2 97% + DME 3% (NPV Max at 30 USD/bbl)	254,615	46.4	-	-	-	-
	CO2 77% + DME 23% (NPV Max at 60 USD/bbl)	-	-	529,805	50.9	-	-
	CO2 51% + DME 49% (NPV Max at 90 USD/bbl)	-	-	-	-	772,999	61.0
Solvent Reuse + Credit	CO2 95% + DME 5% (NPV Max at 30 USD/bbl)	289,711	46.7	-	-	-	-
	CO2 68% + DME 32% (NPV Max at 60 USD/bbl)	-	-	574,132	54.5	-	-
	CO2 34% + DME 66% (NPV Max at 90 USD/bbl)	-	-	-	-	853,472	68.2

,

Table 13. Economic evaluation without consideration of CO₂ credits when the standard deviation is 21.35 in a double flash CO₂ stream.

Oil Prices		30 USD/bbl		60 USD/bbl		90 USD/bbl
	Objective Function	NPV (USD)	Oil Recovery (%)	NPV (USD)	Oil Recovery (%)	NPV (USD)	Oil Recovery (%)
Case
Without Solvent Reuse	CO2 100% + DME 0% (NPV Max at 30 USD/bbl)	198,718	48.5	-	-	-	-
	CO2 93% + DME 7% (NPV Max at 60 USD/bbl)	-	-	451,669	50.3	-	-
	CO2 86% + DME 14% (NPV Max at 90 USD/bbl)	-	-	-	-	690,238	52.5
Solvent Reuse	CO2 100% + DME 0% (NPV Max at 30 USD/bbl)	221,803	48.5	-	-	-	-
	CO2 85% + DME 15% (NPV Max at 60 USD/bbl)	-	-	500,972	53.0	-	-
	CO2 72% + DME 28% (NPV Max at 90 USD/bbl)	-	-	-	-	735,943	58.8

,

Table 14. Economic evaluation with consideration of CO₂ credits when the standard deviation is 21.35 in a double flash CO₂ stream.

Oil Prices		30 USD/bbl		60 USD/bbl		90 USD/bbl
	Objective Function	NPV (USD)	Oil Recovery (%)	NPV (USD)	Oil Recovery (%)	NPV (USD)	Oil Recovery (%)
Case
Without Solvent Reuse + Credit	CO2 100% + DME 0% (NPV Max at 30 USD/bbl)	242,988	48.5	-	-	-	-
	CO2 82% + DME 18% (NPV Max at 60 USD/bbl)	-	-	517,757	54.7	-	-
	CO2 63% + DME 37% (NPV Max at 90 USD/bbl)	-	-	-	-	736,005	62.2
Solvent Reuse + Credit	CO2 100% + DME 0% (NPV Max at 30 USD/bbl)	275,330	48.5	-	-	-	-
	CO2 74% + DME 26% (NPV Max at 60 USD/bbl)	-	-	546,201	58.0	-	-
	CO2 50% + DME 50% (NPV Max at 90 USD/bbl)	-	-	-	-	826,911	69.1

,

Table 15. Economic evaluation without consideration of CO₂ credits when the standard deviation is 21.35 in a distillation CO₂ stream.

Oil Prices		30 USD/bbl		60 USD/bbl		90 USD/bbl
	Objective Function	NPV (USD)	Oil Recovery (%)	NPV (USD)	Oil Recovery (%)	NPV (USD)	Oil Recovery (%)
Case
Without Solvent Reuse	CO2 100% + DME 0% (NPV Max at 30 USD/bbl)	179,063	48.9	-	-	-	-
	CO2 96% + DME 4% (NPV Max at 60 USD/bbl)	-	-	415,007	50.2	-	-
	CO2 90% + DME 10% (NPV Max at 90 USD/bbl)	-	-	-	-	686,245	52.1
Solvent Reuse	CO2 100% + DME 0% (NPV Max at 30 USD/bbl)	201,896	48.9	-	-	-	-
	CO2 90% + DME 10% (NPV Max at 60 USD/bbl)	-	-	496,372	52.4	-	-
	CO2 81% + DME 19% (NPV Max at 90 USD/bbl)	-	-	-	-	711,332	55.0

and

Table 16. Economic evaluation with consideration of CO₂ credit when the standard deviation is 21.35 in a distillation CO₂ stream.

Oil Prices		30 USD/bbl		60 USD/bbl		90 USD/bbl
	Objective Function	NPV (USD)	Oil Recovery (%)	NPV (USD)	Oil Recovery (%)	NPV (USD)	Oil Recovery (%)
Case
Without Solvent Reuse + Credit	CO2 100% + DME 0% (NPV Max at 30 USD/bbl)	198,334	48.9	-	-	-	-
	CO2 83% + DME 17% (NPV Max at 60 USD/bbl)	-	-	499,021	54.3	-	-
	CO2 67% + DME 33% (NPV Max at 90 USD/bbl)	-	-	-	-	700,065	61.6
Solvent Reuse + Credit	CO2 100% + DME 0% (NPV Max at 30 USD/bbl)	223,851	48.9	-	-	-	-
	CO2 77% + DME 23% (NPV Max at 60 USD/bbl)	-	-	527,010	57.3	-	-
	CO2 58% + DME 42% (NPV Max at 90 USD/bbl)	-	-	-	-	773,804	65.0

).

In a dehydration stream, the maximum NPV is obtained when only CO₂ is used at an oil price of 30 USD/bbl. When the oil price rises, NPV is maximized by injecting CO₂ and DME together. Considering CO₂ credits, even if the oil price is 30 USD/bbl, injection with a small amount of DME is preferable to using only CO₂. In addition, the amount of DME injected to achieve the maximum NPV increases. When only CO₂ was used in the double flash stream, the recovery increased due to higher CO₂ purity, and NPV decreased due to the increased CO₂ price. When oil prices are high, the DME content in the maximum NPV scenario is lower than that of dehydration. Even considering CO₂ credits, it is optimal not to use DME for low oil prices. Distillation, in which oil recovery is highest, is most disadvantageous in terms of economic feasibility due to the high CO₂ price.

4. Limitations and Recommendations

A compositional model was developed in which impure CO₂ and DME were mixed and injected into a hypothetical reservoir model. Since this study focused on the effects of impurities in the gas stream with a 2D-layer reservoir model, there are limitations in identifying the complex behavior of real fields with various types of heterogeneities. In addition, when CO₂ is dissolved in water, it can react with minerals in carbonate rock reservoirs and change the physical properties of the reservoir.

To obtain more accurate results and expand the scope of applications, the following research topics are suggested.

• A field-scale reservoir model must be built using real reservoir data. In order to reflect real reservoir heterogeneity, which has a significant impact on oil recovery, both layered heterogeneity and random areal heterogeneity must be included. In this case, the collection of field-scale rock and fluid data is also important.
• A more sophisticated equation of state must be introduced for fluid modeling.
• By incorporating a geochemical reaction model, it is possible to reflect changes in porosity and permeability due to reactions between minerals and brine in carbonate rocks.

5. Conclusions

Based on compositional modeling of a water-alternating impure CO₂ process, oil recovery was calculated based on vertical sweep efficiency and displacement efficiency, and NPV was used as a function for optimization.

• The higher the heterogeneity of the reservoir, the lower the DME content; the lower the purity of the CO₂ stream, the lower the slope in front of the swept area. The higher the CO₂ purity, the greater the area with an oil viscosity of 0.5 cp or less. For the IFT, the higher the CO₂ stream purity, the smaller the peak value; the higher the DME content, the greater the difference after the peak; and the higher the reservoir heterogeneity, the shorter the time to reach the peak.
• The more heterogeneous the reservoir, the higher the remaining oil saturation (0.5 or more) in the lower part of the reservoir. In addition, the higher the DME content and the higher the CO₂ purity, the greater the area with oil saturation of 0.3 or less. As CO₂ purity increases, the breakthrough is delayed.
• When the dehydration and double flash CO₂ streams were used, oil recovery decreased by 8.9%, from 59.0 to 50.1%, and by 14.6%, from 70.0 to 55.4%, respectively, due to increased reservoir heterogeneity when the DME content was 20%. As the DME content increased when the reservoir was heterogeneous, the oil recovery for dehydration and double flash increased by 4%, from 46.1 to 50.1%, and by 6.9%, from 48.5 to 55.4%, respectively. The oil recovery decreases as the reservoir becomes heterogeneous and the DME content decreases. In addition, the oil recovery increases as the purity of the CO₂ stream increases.
• Since cost-cutting of the CO₂ stream has a greater impact than reduction in oil recovery due to CO₂ impurities, lower-purity CO₂ leads to a higher NPV. For a given CO₂ stream, the higher the DME content injected in the optimal scenario, the higher the oil price. Solvent reuse and CO₂ credits are economically beneficial. With the less expensive dehydration CO₂ stream, injection with some DME maximizes economic feasibility considering CO₂ credits even when oil prices are low. For the dehydration CO₂ stream, solvent reuse and CO₂ credits with oil prices of USD 90, the DME content of 66% has a maximum NPV of USD 853,472, which is 121% of the maximum value of USD 703,712 from the DME content of 20% without solvent reuse and CO₂ credits. For double flash CO₂ stream is used, it is a 3.2% improvement over the NPV maximum value of USD 826,911 with 50% DME.