Analysis of the Energy Efficiency of Gas Extraction from Lake Kivu

2.1. Thermodynamic Model

A fluid model for accurately predicting the phase behavior of extracted water and dissolved gases at low pressures (1–15 bar) is crucial for estimating energy production and consumption, as it provides the amounts of CH₄ and CO₂ in the gas phase at each degassing pressure. While single-component gas solubility in water or brine has been extensively studied,³⁷ there is limited research on CO₂–CH₄–H₂O ternary mixtures. High concentrations of CO₂ in produced gas have a negative effect on the efficiency of the power plant and increase the total energy requirement for gas processing purposes. The water content of the produced gas also could cause corrosion problems. Therefore, a model that can precisely predict the phase behavior, vapor fraction, and mutual solubilities of the water–CH₄–CO₂ ternary system is required.

Two modeling approaches exist for gas–water equilibrium: using an equation of state (EoS) to calculate fugacity in both phases or combining an EoS for fugacity calculations in the vapor phase and an activity model for the aqueous phase. However, many models are limited to specific conditions and focus on binary systems. Recent works by Zirrahi et al.,³⁷ Zuo et al.,³⁸ Shabani and Vilcáez,³⁹ Li et al.,⁴⁰ and Ziabakhsh-Ganji and Kooi⁴¹ have modeled CH₄–CO₂–water/brine systems, with good agreement with experiment. Zirrahi et al. used a noniterative approach with the Peng–Robinson equation of state (PR-EoS) and Henry’s constant, requiring fewer parameters and offering faster convergence.

In this work, we used the thermodynamic model that was developed by Zirrahi et al.³⁷ In this approach, the thermodynamic model is based on a fugacity calculation of the vapor phase using PR-EoS and Henry’s law approaches for the liquid phase. The reference pressure is considered to be 1 bar. The mole fraction of the dissolved gases in the aqueous phase is given by

1

where y_i, ϕ_i, γ_i, and ν^∞_i are the mole fraction in the vapor phase, fugacity coefficient, activity coefficient in the mole fraction scale, and partial molar volume, respectively. H_i(T,1) represents Henry’s constant at a reference pressure. Zirrahi et al. used a nonrandom mixing rule. However, in our case, since H₂S concentration is very low and CH₄ and CO₂ mole fractions compose the majority of the gas phase, the conventional mixing rule is used. Moreover, we are not dealing with high temperatures and pressures. Thus, conventional random mixing rules given below are adequate for tuning the interaction parameters of the components in a gas mixture.

2

3

Parameters needed for implementing eq 1, Henry’s constant and partial molar volume, for a ternary system of CH₄–CO₂–water were obtained from literature.^42,43 For CO₂, ν^∞_CO2(g) = 32.6 cm³/mol is used, and CO₂ Henry’s constant is given by⁴²

4

where T is in °C, Henry’s constant is in bar molality scale, and coefficients c₁, c₂, and c₃ are given in Table 3. Partial molar volume and Henry’s constant of CH₄ are given by eqs 5 and 6, respectively.⁴³

5

6

where T is in °C in eq 5, whereas in eq 6 τ = 0.01T, where T and H(T,1) are in Kelvin and Pa/mole fraction, respectively. The water content in the gas phase is given by

7

where

8

9

Subscript i represents dissolved gases, CO₂ and CH₄, ν^∞_H2O = 18.18 cm³/mol, and K⁰_H2O is the equilibrium constant of water at 1 bar as given by⁴²

10

where coefficients are given in Table 3, T is in °C, and the water equilibrium constant is on the molality scale in bar. Since the mixture parameters (eq 1) are calculated based on the mixing rule, the water content of the gas phase is required, resulting in an iteration. The same assumption proposed by Spycher et al.⁴² is applied, and the water mole fraction in the gas phase is assumed to be zero for mixing rule calculations.

In electrolyte solutions, salinity can significantly impact the gas solubility. However, Lake Kivu’s low salinity (1000–6000 ppm or 0.0170–0.1027 molal NaCl) results in a correction factor near unity, based on Zirrahi et al., indicating salinity’s effect on gas phase water content is negligible. This simplifies the calculations by avoiding the need to parametrize dissolved gases and ions interaction.

The fluid model was calibrated using experimental data from the literature⁴⁴ and validated by comparing predicted gas solubility with published measurements,^45,46 confirming its accuracy under various conditions. The comparison with the literature data is reported in the Results section.

The density of degassed water is the most important thermophysical property for the precise estimation of lake stability. In Lake Kivu, four components contribute to density: temperature, salinity, CH₄, and CO₂.⁴⁷ The precise estimation of water density as a function of temperature can be determined using an equation of state developed by Chen and Millero⁴⁸ as given by

11

where ρ is in g/cm³ and T is in °C. The effect of other contributors also should be implemented in determining the water density,⁴⁷ which can be included as follows:

12

where the coefficient of haline contraction, β_s, and contraction coefficient for CO₂ and CH₄ are shown in Table 4.

2.2. Energy Efficiency

The thermodynamic model described earlier was utilized to predict the water degassing efficiency at various separator pressures. Following the approach proposed by Bolson et al.³⁶ (hereinafter called the theoretical approach), we determined the optimal separator pressure in terms of energy efficiency per m³ of the extracted water. The theoretical approach considers the energy contained in CH₄ and the predefined energy required to separate the two components, CO₂. It also assumes pure methane with a lower heating value (LHV) of 50 MJ/kg and total CO₂ removal to calculate the net produced power. However, in our simulation-based approach, we calculated the equivalent generated electricity based on energy and material balance from process simulation results, gas turbine manufacturer data sheet, and reported heat rate. Then the equipment’s power requirements for CO₂ removal and heat duty were calculated using energy balance from process simulation results. Finally, net power per cubic meter of extracted water was calculated.

In amine-based carbon capture, besides power consumption for rotary equipment, reboiler heat consumption is the main energy consumer. A range of 4–10 MJ/kg CO₂ captured is reported as a reboiler heat consumption based on monoethanolamide (MEA) and solvent flow rate.^51,52 It was also found that the process parameters and amine type affect the regeneration energy consumption.⁵² Although the required power for extraction and reinjection of degassed water and amine plant power consumption could be neglected compared to reboiler heat duty, the wide range of 2–8 MJ/kg CO₂ captured is considered in energy efficiency calculation in the theoretical approach. This wide range provides insight into the feasibility of the project in the case of very high energy consumption and the trend between energy penalty and optimum pressure. This range could also provide valuable information for optimistic and pessimistic cases of the amine-scrubbing technique.

2.3. Simulation-Based Scenarios

2.3.1. Amine Scrubbing Scenarios

The fluid model was integrated with the process simulation to assess the energy requirements of amine scrubbing. Since achieving complete carbon dioxide removal is neither economically nor practically feasible, and a small portion of the extracted methane is inevitably lost as methane slips during the process,³⁴ the simulation process helps to quantify these limitations. The output from the fluid model serves as input for the amine scrubbing simulation, enabling the evaluation of the CO₂ capture efficiency, methane production, and overall power consumption of the plant. Moreover, the calculated optimal pressure based on the fluid model was reevaluated in the case of a cogeneration scenario, where the energy requirement for CO₂ removal can be reduced, leading to new optimal pressure calculations. To determine the energy efficiency of the amine-based carbon separation approach, we integrated the water degassing results from the fluid model with the Aspen-HYSYS V14.1. In this approach, the primary objective is to determine the maximum net power that can be delivered per cubic meter of extracted water (eq 13). Therefore, the power required to operate the rotary equipment, cooler, and the heat duty of the reboiler are calculated based on process simulation results. Additionally, the power generated by the gas turbine generator is used to determine the system’s energy efficiency at different separator pressures.

13

where Power_net is the net output power of the extraction process in kWhe/m³ water, Power_generator is the electricity generated by the gas turbine generator, Power_consumption is the power required to run the process equipment and provided heat duty of the reboiler, and Water_extracted is the water extraction rate in m³/h.

Figure 4 shows the workflow that is followed in this work. For this purpose, the following assumptions were made for the analysis: the isentropic efficiency of the pumps and compressor is 85%, and the pressure profile through the regenerator and absorber columns is linear. It is also assumed that the lean-amine cooler, raw gas cooler, and condenser are air-cooled exchangers with no heat loss in the heat exchangers to the outside environment. The overall power consumption of the fans is calculated based on a motor efficiency of 95%. Additionally, the power consumption includes the contribution of compressor duty needed to increase the gas pressure from the separator pressure to the required absorber column pressure for the amine/wash scrubber tower. However, the compressor power required to overcome pressure loss along the pipeline for delivering gas to the onshore plant is excluded, as it depends on the distance between the offshore barge and the onshore facility.

Figure 4.

Open in a new tab

The following equation is used to calculate the air-cooled exchangers’ fan power consumption.⁵³

14

where Q_air is the actual air flow rate in ft³/min, ΔP_b is the pressure drop across the bundle in inches of water (0.6 in. of water is assumed), and η_f is the fan efficiency (0.7 is considered).

For process simulation purposes, power consumption, amount of CH₄ production, and CO₂ captured are calculated based on the initial design and parameters outlined in Table 5. Estimated power requirements can help identify potential areas for energy savings and improve the overall efficiency of the operation. For instance, compressor duty may be reduced by optimizing the absorber pressure to a lower feasible level while maintaining the CO₂ removal efficiency and increasing the final net power output. Although the power consumption for the amine-based CO₂ separation process is expected to be higher than that of water scrubbing, its environmental benefits, such as preventing the reinjection of CO₂ into the lake and consequently mitigating acidification in the biozone, could justify its adoption. If the additional power requirement remains within an acceptable range, stakeholders may prioritize the long-term ecological advantages of the amine-based method over water scrubbing.

Table 5. Amine Scrubbing Parameters for the Initial Simulation Design.

parameter	initial value
Water extraction rate	31,100 m3/h
Absorber, Regenerator column: No. of Tray	20
Absorber column pressure	35 bar
Regenerator column pressure	2 bar
Raw gas feed Temperature	25 °C
Lean amine temperature	31 °C

Open in a new tab

Furthermore, the simulation results offer insights into potential improvements in the design and operation of the extraction system. For example, if the power consumption for CO₂ capture is disproportionately high, alternative methods or optimizations can be explored to reduce the energy use. Similarly, if methane production is lower, then expected adjustments to the separator pressure or flow rates can be made to enhance yield.

2.3.2. Water Scrubbing Scenario

Water scrubbing is one of the dominant techniques in biogas upgrading.³⁴ Water scrubbing is currently used for the removal of CO₂ at both KivuWatt and KP56. However, this approach necessitates reinjecting sour water to the bottom of the biozone, raising concerns regarding its environmental impact on the ecology and living species. Although amine scrubbing could serve as an alternative approach for CO₂ removal and mitigate the environmental risks associated with water scrubbing, its energy requirements might pose a challenge. While the amine approach may use less power than water scrubbing, its heat duty is significant.³⁴ Therefore, a comparative study of water scrubbing and amine scrubbing was conducted in terms of energy efficiency and optimum pressure to provide a comprehensive evaluation.

This study compares the energy efficiency and environmental sustainability of the CO₂ removal methods in methane extraction. By analyzing water and amine scrubbing techniques, we aim to identify optimal pressure conditions that maximize power efficiency and minimize environmental impact, guiding methane extraction project designs.

Aspen-HYSYS V14.1 process simulation was used to conduct energy and material balance calculations. The wash tower operating pressure and temperature were set to 7 bar and 25 °C. The total power requirements of the water scrubbing process are consumed by the compressor, extraction and reinjection pumps, and gas air cooler. The pump power requirement was calculated using

15

where ρ (kg/m³) is the water density, g (m²/s) is the gravitational acceleration, Q (m³/s) is the flow rate, H (m) is the total head, and η (fraction) is the pump efficiency. The total head is estimated by eq 16.⁵³

16

where Δp_f (Pa) is the pressure drop due to friction and ΔP (Pa) and Δz (m) are the pressure and elevation difference between the extraction point and scrubbing tower, respectively. The water rate was set to match the objective of amine scrubbing, achieving the same CO₂ removal for biogas upgrading purposes. This determination ensures that both scrubbing methods aim for maximal CO₂ capture efficiency, providing a consistent basis for comparison. By setting the objective, the study can accurately evaluate the effectiveness and energy consumption of each method. Table 6 lists the assumed values based on the current projects in Lake Kivu^27,54 to estimate the total head.

Table 6. Assumptions for Calculating the Total Head of the Pump in the Water Scrubbing Process.

parameter	value
Pipe diameter	24 in.
Water extraction depth	20 m
Discharge sour water depth	80 m
Pipe length	600 ma
Wash tower height	15 m
Pipe material	HDPE
Pump efficiency	75%

Open in a new tab

^a

Pipe length assumption value is based on the onshore treatment plant configuration.

2.4. Early and Late Reservoir Life Energy Efficiency

The decline in gas concentrations due to extraction from the resource zone and subsequent dilution is inevitable. Currently, with two projects having initiated gas extraction, the impact may be minimal. However, if other concessionaires begin gas production, the water extraction rate would increase significantly, leading to more substantial effects. A 1-D simulation study by Wüest et al.²³ showed that, with a water extraction rate of 70 m³/s under Plan B (scenario RZ5), methane concentration could decrease by half after 30 years of production. Although their model includes some uncertainties⁵⁵ and assumes constant gas removal efficiency over a 100-year simulation, the results provide valuable insights for predicting energy efficiency in the later stages of reservoir life. In this study, we used their simulation results after 50 years of production as inputs for the fluid model to estimate the energy efficiency for both amine and water scrubbing methods at obtained optimum pressure after a decline in gas concentration.

3.1. Fluid Model Validation

The thermodynamic model was calibrated using data provided by Qin et al.,⁴⁴ and the results are presented in Figure 5. The model exhibited an average absolute relative deviation (AARD) of 5.98% for methane and 5% for carbon dioxide, demonstrating an acceptable level of accuracy in capturing the solubility behavior within the ternary system. To assess the model’s accuracy in predicting solubility, the calculated solubilities were compared with additional experimental data from Al Ghafri et al.⁴⁵ and Kastanidis et al.⁴⁶Figure 6 illustrates the predicted solubility of the CH₄–CO₂ mixture in water, highlighting the model’s performance.

3.2. Removal Efficiency of Degassed Water

As discussed, MP recommends the percentage of removal of CO₂ from the extracted water for three different production scenarios. Understanding the removal percentage of each dissolved gas at various separator pressures is crucial for determining the optimum pressure. Based on the fluid model, the removal efficiency of degassed water, the in situ percentage of each dissolved gas in the vapor phase, at different pressures, is shown in Figure 7. The water sample depth is selected from the lower resource zone. The figure indicates the percentage of each gas that transitioned from the water phase to the vapor phase at varying separator degassing pressures. Based on the in situ dissolved gas concentrations and the gas concentrations at different pressures, the removal efficiency of each gas has been determined (Table S1). As can be seen, the removal efficiency of CH₄ has a linear relationship with pressure, while CO₂ demonstrates a nonlinear trend and is significantly released when pressure falls below 7 bar. At 1 bar, CH₄ removal is about 100% and CO₂ removal is around 70% of the in situ values. However, the overall amount of CO₂ released is approximately four times the amount of CH_4. This is primarily because the in situ dissolved CO₂ is around five times greater than CH₄ in place. It should be noted that these results depend on the dissolved gas concentration at the extraction depth. Currently, the resource zone can be harvested economically. Therefore, only a sample from the resource zone is considered to address the amount of dissolved gas that can be evolved per m³ of the extracted water. As the separator pressure drops below 20 bar, both dissolved gases begin to evolve into the vapor phase, with CH₄ exhibiting a significantly higher percentage removal than that of CO₂ due to its lower solubility in water.

3.3. Energy-Efficient Optimum Pressure

The net electricity generated per m³ of extracted water is contingent upon gross energy production, which can be estimated from the amount of harvested methane and the energy utilized for CO₂ removal. A lower separator pressure tends to increase the quantity of methane and, consequently, the gross energy output. However, this also results in a higher concentration of CO₂ in the produced raw gas, thereby necessitating greater energy for the CO₂ removal. Optimizing the net electricity generated per m³ extracted water involves determining the optimal separator pressure that balances these factors.

Figure 8 and Table 7 show the net energy efficiency of the extracted water sampled at lower resource zone and optimum pressure using Bolson et al.'s approach, respectively.³⁶ In an optimistic case (2 MJ/kg of CO₂ captured), the maximum net power that can be generated is about 1.15 kWhe/m³ of the extracted water at a separator pressure of around 4.5 bar. In other words, to deliver 25 MW to the electricity grid, at least 6.04 m³/s of water should be extracted in the current status of the lake. In a pessimistic case (8 MJ/kg of CO₂ captured), 9.14 m³/s would be the minimum water extraction rate. All these rates are based on the reported efficiency of the gas turbine generator, which is around 38.6%. In case of lower efficiency, the extraction rate would be higher. For example, considering 25% efficiency of the electricity generator, the water extraction rate to deliver 25 MW would be 9.3 and 14.1 m³/s for optimistic and pessimistic CO₂ capture energy consumption scenarios, respectively.

Based on Figure 8, it is also noted that at separator pressure around 2 bar, which is the current operating pressure of the KivuWatt and Kivu56 projects,^27,54 the energy consumption exceeding 4 MJ/kg CO₂ captured is unfavorable regarding energy efficiency. This is because the net electricity generated becomes very low or even zero in the case of higher energy penalties due to the huge amount of CO₂ released to the gas phase at lower pressure. The utilized theoretical approach based on the fluid model also confirmed that the amine-scrubbing method would be more energy efficient at separator operating pressures higher than 5 bar.³⁶

In order to estimate the energy efficiency in a realistic scenario compared to the theoretical approach, the fluid model’s flash calculation output has been coupled with process simulation (Figure 9). For a more robust comparative analysis, the primary objective of the process simulation was to capture nearly all of the CO₂ and deliver high-purity sweet gas. Since different types of amines vary in terms of energy utilization, we conducted a sensitivity analysis on various amine types (Table 8). Based on this analysis, we selected a mixture of the primary amine monoethanolamine (MEA) at 10 wt % and the tertiary amine methyldiethanolamine (MDEA) at 40 wt %. We considered the parameters in Table 5 in all simulation cases. Figure 10 reveals the net power efficiency based on the energy and material balance outputs of the process simulation. Three cases of energy consumption are considered. First, just the reboiler duty is considered as the energy consumption term in the net power calculation of amine scrubbing. The average value of reboiler heat consumption for CO₂ removal was approximately 4.11 MJ/kg CO₂ captured, which is close to the reported literature value.⁵⁶ In the second case, the power requirement of other process equipment was added to the net power estimation (reboiler and power consumption curve). It highlights the impact of power consumption on overall energy efficiency. Finally, in the third case, the energy efficiency of the cogeneration scenario has been evaluated. In this scenario, only the power consumption by other equipment is considered to be the only energy penalty, and reboiler duty is assumed to be provided by heat recovery from the power plant.

For the first two cases, where reboiler duty was included in the energy consumption, the power efficiency was found to be 0.88 and 0.83 kWhe/m³ of extracted water, respectively. Although the power consumption of other processes is negligible compared to the reboiler duty, including these additional energy penalties decreases the net power efficiency by 7%. In both cases, the trend and optimum pressure are the same (Table 9). Thus, we can conclude that, regardless of reboiler duty, power consumption has no impact on optimum pressure, and the reboiler duty is the only parameter that indicates the optimum pressure from an energy efficiency perspective.

The blue curve in Figure 10 represents the net power efficiency based on the theoretical approach by Bolson et al., assuming an energy penalty of 4 MJ/kg of CO₂ captured. This approach demonstrates higher energy efficiency than the simulation-based results, while the trend and optimum pressure remain approximately the same. The difference arises from the material balance. The theoretical approach assumes that all CH₄ and CO₂ in the gas phase are separated and the equivalent gross energy and energy penalties are factored into the net energy calculation. In contrast, in the process simulation, some content of both gases may slip through the process and remain in the amine solution. In other words, the theoretical approach provides the upper bound of energy efficiency for the specified scenario.

Further, the power requirements for water extraction and degassed water reinjection are not considered in any scenario’s calculation, as they are negligible compared to other power consumption and would remain the same regardless of the water or amine scrubbing technique. However, the offshore load still needs to be included. In both the offshore configuration of the KivuWatt project and the semioffshore configuration of the Kivu56 project, part of the produced gas or power is consumed by compressors and coolers installed on offshore barges to transport the gas to onshore facilities. Therefore, the final net energy generated per m³ of extracted water would be lower than shown in this work, necessitating a higher water extraction rate to deliver the same amount of power to the grid. If the extraction configuration changes to an onshore installation (no extraction barge), other parameters should be included, which will be discussed later.

Figure 10 also reveals that apart from the cogeneration scenario, all other cases exhibit the same trend at the same pressure range. Beyond 7 bar, net power decreases linearly with pressure. The removal percentages of both gases in Figure 7 indicate that methane and carbon dioxide evolve into the gas phase linearly between 20 and 7 bar. However, methane is released at a higher percentage due to its lower solubility. As expected, in this pressure range, net power increases as pressure decreases since methane removal reaches up to 70%, while the level of CO₂ removal remains around 10%. Between approximately 3 and 7 bar, all curves reach their peak values. This is primarily due to the removal of CO₂ beginning its exponential behavior, as shown in Figure 7. Below 3 bar, the release of CO₂ into the gas phase increases from 25% to approximately 80%, significantly impacting net power, leading to a sharp decline in the net power curve, as seen in Figure 10.

The overall trend in the cogeneration scenario is similar to those of the other cases, except that waste heat recovery enables a shift toward a lower pressure range while also increasing net power. In this scenario, the net power reaches its peak value between 2 and 4 bar.

The cogeneration scenario, wherein the reboiler heat duty is supplied through heat recovery from power generation, demonstrated a significant impact on the power efficiency. This configuration increases the net power efficiency to 1.2 kWhe/m³ of water at the optimum pressure of 3 bar, reflecting a 36% improvement. In this scenario, only the power consumption by rotary and other process equipment is considered as the energy penalty. Under this scenario, even complete degassing or a separator pressure of approximately 2 bar appears to be acceptable from an energy efficiency perspective, unlike the first two cases, where a pressure around 2 bar is not recommended. The substantial increase in net power efficiency can be attributed to the effective utilization of waste heat from power generation, which reduces the overall energy demand for the reboiler. This highlights the potential benefits of integrating heat recovery systems within methane extraction and processing operations. By harnessing waste heat, the energy requirements for CO₂ removal are significantly reduced, leading to higher overall efficiency.

Moreover, the findings suggest that lower separator pressures, which were previously deemed inefficient due to high CO₂ removal costs, become viable in the cogeneration scenario. This is because heat recovery mitigates the energy penalties associated with CO₂ separation, making lower pressures more feasible. Consequently, the operational flexibility is enhanced, allowing for a broader range of pressures to be utilized without compromising the energy efficiency. These results were based on the initial design given in Table 5. Therefore, any adjustments in operational conditions, such as pressure, temperature, amine flow rate, or concentration, that can reduce the process equipment’s power consumption will contribute to higher energy efficiency.

In the cogeneration scenario, our analysis shows that approximately 0.5 MJ/kg of captured CO₂ is consumed (Table 10) for CO₂ removal, achieving a carbon capture efficiency of about 99%. If the theoretical approach is applied for the case of 0.5 MJ/kg of CO₂ as energy penalty, the optimum pressure and the corresponding net captured energy would be 2 bar and 1.38 kWhe/m³, respectively. In contrast, our comprehensive process simulation approach shows 13% less optimum net power than the theoretical one at 3 bar (Figure 11). Both approaches also follow the same trend, which is consistent with previous results (Figure 10). The trend of net energy indicates that a decrease in energy consumption in CO₂ removal not only lowers the optimum pressure but also provides the opportunity to meet the mandatory requirements of management prescription.¹⁵

The results showed that amine scrubbing could be an ideal technique for CO₂ removal in the Lake Kivu gas extraction project compared to water scrubbing, since this approach completely mitigates the environmental impacts on the biozone, and there is no concern regarding the ecology of the biozone. However, in terms of energy efficiency, amine scrubbing might be less efficient than water scrubbing. Thus, another comparative study has been conducted to highlight the amine scrubbing and water scrubbing techniques in terms of energy efficiency. Figure 12 shows the process flow diagram for water scrubbing.

Figure 13 compares the outcomes of amine scrubbing, the cogeneration scenario, and water scrubbing. The data (Table 11) indicate that the water scrubbing technique results in a 44% increase in net power per m³ of extracted water at the optimum pressure. Water scrubbing yields 1.19 kWhe/m³ of water at 3 bar, whereas amine absorption produces 0.83 kWhe/m³ of water at 6 bar. However, the cogeneration scenario significantly improves energy efficiency, achieving results comparable to those of the water scrubbing method. Consequently, it can be concluded that amine absorption is a viable CO₂ removal method for the Lake Kivu gas extraction project if integrated with the cogeneration scenario. This approach offers the same net energy as water scrubbing while preserving the ecological balance of the biozone. Otherwise, water scrubbing provides a net energy considerably higher than that of amine scrubbing. Although amine scrubbing can effectively address environmental concerns related to discharge of washwater into the biozone, the substantial reduction in net energy generation could significantly influence the final decision of stakeholders.

According to the available reports,^26,54 the Kivu56 project can expect a net power efficiency of 0.85 and 0.95 kWhe/m³ of water when accounting for total power consumption and only onshore power consumption, respectively. KivuWatt’s net power efficiency is estimated at around 0.79 kWhe/m³ of water. Due to the lack of detailed information regarding power consumption for specific equipment and the distribution between onshore and offshore loads, a direct comparison with the current reported values is impossible. Nevertheless, since the results of this study are based on the assumption of onshore loads only, they provide a clearer understanding of the expected magnitudes of these values and demonstrate that our approach could align closely with current real-world cases.

Based on the literature, power consumption for biogas upgrading is reported as ranging from 0.16 to 0.3 kWh per normal m³ of input biogas for water scrubbing and 0.12 to 0.14 kWh per normal m³ for amine scrubbing.^34,57 In this study, the calculated average power consumption for both methods is 0.13 kWh per normal m³ of input biogas. The lower power consumption observed in our study compared with the literature for the water scrubbing method could indeed be attributed to the fact that the water is not recycled and circulated back to the wash tower. Without the need for water regeneration systems through air stripper columns, the process design becomes simpler. This simplification can lead to lower operational energy costs, as fewer components mean fewer opportunities for energy losses. Regarding heat consumption, the calculated average energy requirement for amine scrubbing is 0.53 kWh per normal m³ of input biogas, which closely approximates the reported value of 0.55 kWh per normal m³ in the literature.³⁴ The reported values and our findings indicate no substantial difference in power consumption between amine scrubbing and water scrubbing. While amine scrubbing generally requires less electrical power than water scrubbing in biogas upgrading processes, its high heat energy demand significantly affects the overall energy efficiency of the project. However, the close alignment of the calculated and literature values for both power, heat consumption, and the energy efficiency demonstrates that the methodologies, assumptions, and simulation model used in this study are robust, underscoring the accuracy of the study’s modeling approach.

KivuWatt and Kivu56 project designs were based on the power production of 26 and 56 MW from a water extraction rate of 9.2 and 18.4 m³/s, respectively.^27,54 We assume that the net power delivered to the grid and the operating conditions are the same as the reported values. In this case, the net power efficiencies of the projects are 0.79 and 0.85 kWhe per m³ of extracted water, respectively. It should be noted that these values are also based on the inclusion of offshore power consumption. In contrast, our study is based on onshore gas treatment facilities, and the associated compression and power consumptions for transporting the extracted gas to onshore facilities are not included. In the case of including other power requirements, the net energy will be lower. However, the study’s main objective is to build a robust workflow that highlights the project’s energy efficiency trend and provides a good approximation of the power efficiency of real-world scenarios. It can be concluded that the optimum net energy efficiency of gas extraction from Lake Kivu could be between 0.8 and 1 kWhe/m³ extracted water, considering the current dissolved gas concentration at the resource zone and energy conversion efficiency. These estimated values are based on the results of this work and the current operating projects in Lake Kivu. Notably, the offshore barge’s power requirements depend on the distance of the offshore barge from the power plant and the location of the biogas upgrading plant since the biogas upgrading facilities could be installed on the offshore barge or onshore.

According to MP’s conservative scenario for power generation from the RZ, approximately 400 MW is anticipated over a 50-year operational period.¹⁵ If the net and gross power efficiencies are considered to be 0.85 and 1 kWhe/m³ of extracted water, respectively, the estimated water extraction rates would be 111 and 130 m³/s for each case. As the methane concentration will gradually decrease following the commencement of extraction, a higher extraction rate will be necessary to sustain power generation. Simulation model studies conducted by Wüest et al.²³ indicated that with RZ extraction scenarios and an extraction rate of 70 m³/s, the average expected energy generation over a 50-year period would be around 200 MW, as the CH₄ concentration decreased significantly by a factor of 2 after 25 to 30 years of operation. Therefore, achieving 400 MW of electricity over 50 years of operation, as illustrated by the MP expert group with the current production strategy and power conversion method, appears to be doubtful. The main challenge is how long the proposed power generation can be sustained. Thus, monitoring dissolved gas resources and gas and power production data is required to validate the plateau period. Currently, Plan B (seeTable 2) is being implemented, utilizing a single-cycle power generation approach in the power plant since the project’s inception, in which a comprehensive study of its energy efficiency has been addressed in this work.

Another important aspect of gas exploitation is the plateau period, which refers to the duration over which gas production and power generation can be maintained at a constant level. Once this period ends, gas production begins to decline. This decline typically coincides with a decreased dissolved gas concentration due to continuous extraction and resource zone dilution.

We conducted a study to estimate the energy efficiency after 50 years of production. The forecasted concentration of dissolved gases by Wüest et al.²³ was used as input to predict the removal efficiency for CH₄ and CO₂. We calculated the efficiency of amine scrubbing, considering a cogeneration scenario, and water scrubbing at an optimal pressure of 3 bar, assuming no change in the dissolved gas removal efficiency. The results indicated that energy efficiency in both technologies decreased from 1.2 kWhe/m³ of extracted water to approximately 0.6 kWhe/m³ (Table 12). In other words, cutting the gas concentration in half reduces energy efficiency by at least a factor of 2. Consequently, the number of gas extraction facilities would need to be doubled to maintain the power generation levels.

However, the primary uncertainty lies in determining when the decline in the gas concentration will begin and the decline rate. This uncertainty is highly dependent on design parameters such as the extraction plan, extraction and reinjection depths, CH₄ and CO₂ removal efficiency, and especially the extraction rate. It is worth noting that the simulation study by Wüest et al. was based on an extraction rate of 70 m³/s. In contrast, current concessionaire plans (Table 1) suggest that the total extraction rate will be much higher than 70 m³/s, potentially leading to dilution of the resource zone much sooner than anticipated.

3.4. Density Control

In this study, we examined the energy efficiency of gas extraction from Lake Kivu, focusing on different technologies for the removal of CO₂. The sensitivity analysis of energy efficiency of surface facilities, particularly separator pressure, was the main parameter of interest. However, any gas extraction scenario must prioritize maintaining the stratification of the lake (ensuring safety), maximizing CH₄ harvesting (economic efficiency), and considering the lake’s ecological impact (environmental concerns).¹⁶ The separator pressure is a crucial factor that influences the project’s energy efficiency and the first two primary objectives, the safety of the lake and economic efficiency. It determines the amount of CO₂ reinjected into the RZ, which affects the density of the degassed water plume and the rate of dilution of the RZ.

From a safety perspective, preserving the main chemocline must be the primary objective of any gas extraction operation. Thus, CO₂ removal from degassed water becomes the sole method of density control, governed by separator pressure to avoid interfering with the density structure in the lake and to make sure the degassed water discharge plume is displaced downwardly and mixed with the stratified ambient layer.¹⁵ Therefore, maintaining the integrity of the main chemocline (250–260 m) and the integrity of the resource zone density gradient (300–320 m) depend on the behavior of the degassed water discharge plume. As discussed, the separator pressure also influences the optimum pressure for energy efficiency. In cases in which there is a conflict between the optimum pressures for density control (safety) and energy efficiency, priority must be given to density control.

Figure 14 illustrates the calculated density of degassed water using eq 12 and the predicted removal efficiency of the fluid model at various degassing pressures. It is assumed that the degassing process at the separator has no impact on the temperature and salinity of the degassed water, and the dissolved gas concentration is the only variable parameter at different pressures. According to MP’s recommendation for CO₂ removal (Table 2), Figure 14 indicates that the degassing pressure should be less than 3 bar to meet the percentage of CO₂ removal. In terms of energy efficiency, the water scrubbing technique offers more flexibility concerning the separator pressure. However, utilizing amine scrubbing presents challenges, as the analysis shows that the pressure required for density control is significantly below the optimum pressure from an energy efficiency perspective. Our analysis showed that in the case of amine scrubbing, the optimum pressure would be above 5 bar, while from a density control perspective, it is recommended to be less than 3 bar. When amine scrubbing is employed, it must be coupled with cogeneration to achieve the aforementioned goals. Implementing such a scenario ensures that the project’s energy efficiency remains comparable to water scrubbing while also addressing the concern of density control of the degassed plume through appropriate degassing pressure without significantly compromising energy efficiency. Additionally, this approach mitigates the negative environmental impact associated with wash water.

4.1. Uncertainty Analysis

Uncertainty is an inherent aspect of any data analysis or comparative study. In this context, the primary uncertainty in the reported values for net energy generated per cubic meter of water lies in the efficiency and performance of power conversion equipment such as the gas turbine. Typically, the reported heat rate is based on the standard conditions. However, air temperature and humidity variations can significantly impact the generator’s final efficiency, leading to differences in net electricity generation. Despite these variations, the net energy versus pressure trend is likely to remain consistent. The efficiency of rotary equipment, such as compressors and pumps, also introduces uncertainty into the results, as lower efficiency results in higher power consumption. Additionally, the measurement of dissolved gas concentrations contributes to uncertainty, as these measurements are crucial for determining the feed composition of the fluid model. Future monitoring programs can provide valuable data for verifying forecasted simulation data and hypotheses regarding methane generation rates.^8,16 If new measurements are obtained, the provided workflow (discussed later) should be reimplemented to account for propagated errors due to uncertainties.

The rate at which the concentration declines is another significant uncertainty affecting the power production plan. The results of the present study are based on the latest measurements from conducted studies in 2018.⁸ Consequently, any changes in the concentration will lead to variations in the output. One-dimensional simulation models can provide forecasts for future CH₄ and CO₂ concentrations; however, these models are also subject to uncertainties, including the rate of CH₄ formation, the location and rate of subaquatic inflows, and the variable efficiency of CO₂ removal.

The decline in concentration also highlights another uncertainty, which is the economic limit of the CH₄ concentration. This limit is critical for determining the plateau period or how long the power generation project can continue to operate. As the concentration decreases, a decline in power generation becomes inevitable, given that the surface facilities and risers were designed based on a specific water extraction rate. Therefore, future technological advancements and production data will be crucial in determining the concentration economic limit at which the project remains economically viable.

4.2. Offshore and Onshore Facility Configuration

In addition to addressing environmental concerns associated with the biozone, amine scrubbing provides significant advantages, including efficient removal of CO₂ with less methane slip compared to water scrubbing.³⁴ Furthermore, this process allows for CO₂ utilization by achieving 98% purity of captured carbon, in contrast to water scrubbing, where CO₂ is reinjected into the biozone. However, water scrubbing systems can be designed and installed in both offshore barges (seeFigure 2) and onshore areas, whereas amine scrubbing can only be implemented onshore due to the complexity of its flow process and the requirement for a larger footprint. Moreover, as previously discussed, amine scrubbing can achieve the same energy efficiency as water scrubbing if its heat energy requirements are met through a cogeneration scenario. Consequently, the amine scrubbing with the cogeneration scenario plant should be constructed near the power plant in an onshore area. Currently, KivuWatt and Kivu56 projects are operated based on offshore and semioffshore layouts, respectively.

In this study, we only considered the power required for compressing and cooling the gas outlet from the separator under separator pressure and temperature conditions for CO₂ removal. We excluded the energy requirement for transporting the gas to the onshore power plant as this depends on factors such as pipe size, length, and gas volume. While the results of this study can be applied to both offshore and semioffshore configurations, an onshore configuration, where surface facilities are installed onshore without the use of barges, presents additional challenges in terms of process complexity, energy requirements, and fluid flow. In this scenario, the intake and discharge risers for extracted and degassed water could span kilometers. Power requirements for water extraction and the reinjection of degassed water must be considered, as they are not negligible and the siphoning effect alone may be insufficient. Additionally, managing potential two-phase fluid flow through inclined pipes of such lengths could pose significant challenges. Therefore, comprehensive nodal analysis studies should be conducted to address these issues effectively.

4.3. Limitation and Challenges

As discussed earlier, certain assumptions have been made for this simulation-based study. One of the key assumptions in this work is the use of air-cooled heat exchangers, which may present certain limitations and challenges. In cases of higher ambient temperatures, where the temperature approach does not meet expectations, a larger heat transfer area would be required. Consequently, this would necessitate additional space for plant construction, posing a significant challenge in areas with limited land availability near the lake.

Alternatively, using a cooling tower instead of air-cooled exchangers could mitigate space constraints due to water’s higher heat transfer coefficient.⁵³ However, this approach has drawbacks, including increased power consumption, which could reduce the net power output compared with the results obtained in this study. Additionally, water treatment requirements could present further challenges when using water as a coolant in heat exchangers. Therefore, the feasibility assessment in terms of energy efficiency remains constrained by the assumptions outlined in this study.

In the cogeneration scenario, the proposed pathway is viable only if the power plant is constructed in the proximity of the amine scrubbing facility. If the project’s final objective is to deliver gas to the grid rather than generate electricity on-site, this scenario would not be applicable. Thus, beyond the facility configuration, the availability of a nearby power plant introduces a critical limitation for the amine scrubbing method. In contrast, the water scrubbing technique offers a key advantage by being less dependent on the power plant or offshore/onshore facility location.

The lack of real-case data presents a significant challenge for simulation-based studies. This study relies on the limited data available from the currently operating projects in Lake Kivu. Access to more detailed real-case data would enhance confidence in the reliability and robustness of the simulation model, enabling the further optimization of operational conditions.

4.4. Block to Block Gas Extraction

In the work of Bolson et al., the concept of a block harvesting approach for Lake Kivu was discussed.³⁶ They proposed dividing the lake’s surface into 5 × 5 km² blocks with no or little cross-flow, suggesting that once a block is depleted of gas, the extraction barge should be relocated to another block. However, this strategy faces significant challenges due to the cross-flow between the blocks, where mixing occurs on a time scale of months or a few years.⁵⁸ Consequently, the gas resources cannot be effectively harvested using a block strategy or managed independently by two countries. Any changes in gas concentration or gas resource depletion due to local extraction can impact the basin-wide concentration, and the assumption of little or no cross-flow block exploitation is not valid in Lake Kivu.

This analysis highlights the complexity of resource management in Lake Kivu. The interconnected nature of the lake’s gas reserves requires a more holistic approach to extraction that considers the entire lake system rather than isolated blocks. A more integrated production strategy that accounts for the natural mixing processes and ensures equitable resource distribution is essential for sustainable gas extraction. Additionally, cross-border cooperation and comprehensive monitoring of gas concentrations are crucial to adapting to changes in the lake’s dynamics and maintaining the viability of the extraction projects. This underscores the need for collaborative management strategies that transcend political boundaries and focus on the long-term sustainability of Lake Kivu’s resources.

4.5. Integrated Production System

This study primarily focused on the influence of separator pressure on the energy efficiency of gas extraction. However, energy efficiency cannot be considered as a sole parameter to set the separator pressure, and further studies should be conducted to determine the final pressure that meets all of the essential criteria. Therefore, developing an integrated production system for Lake Kivu necessitates a multifaceted approach that balances energy efficiency, lake stability, and economic viability. A key component of this system is the optimization of separator pressure, which directly influences the net energy efficiency, methane recovery, and stability of the lake.

From the net energy efficiency standpoint, adjusting the separator pressure can minimize the energy required for CO₂ removal and processing, thereby maximizing the net energy output. This involves balancing the energy consumption with the energy generated from methane combustion to achieve optimal efficiency. Second, regarding lake stability, maintaining Lake Kivu’s stratification and preserving the main chemocline is crucial to preventing gas outbursts. The pressure must be managed to control the density of the degassed water plume, minimizing the risk of disrupting the lake’s natural stratification and preventing short-circuiting of gas flows, which requires optimal CO₂ removal that is interdependent with energy efficiency. Finally, economic efficiency considerations require maximizing methane recovery, as this directly correlates with the profitability of the extraction project. Optimal pressure enhances the methane yield, ensuring that the economic returns justify the extraction efforts.

The first two studies can be conducted independently by each concessioner since they primarily depend on facility design and capacity. However, the extraction rate is the primary parameter for lake-scale studies, as any extraction project, regardless of its location, can impact basin-wide parameters within a few months or years. Therefore, to conduct such studies for optimization or forecast analysis, the extraction rate, depth of extraction and reinjection in the resource zone, and CH₄ and CO₂ removal efficiency of other concessioners must be known and incorporated into lake model studies (full field model).

Nodal analysis should also be included in the production system. Two-phase flow in pipes is important since, in the future, the concentration of dissolved gases will decrease, the siphoning effect might not be sufficient to bring the water upward to the surface, and extra energy might be required, negatively affecting the project’s net energy yield. Additionally, inlet separator pressure will impact the efficiency of siphoning or the required power. In the case of onshore configurations, where the pipe length is on the order of kilometers, nodal analysis is necessary to ensure the vertical lift performance of inclined pipelines.

Based on these considerations, an integrated production system, such as those described in Figure 15, where all contributing factors are integrated, is used to determine a final separator pressure. Once this pressure is established, computational fluid dynamic (CFD) studies are essential to ensure that the injection of degassed water does not result in short-circuiting. This involves simulating the flow and mixing dynamics within the lake to verify that the injected water does not return too quickly to the intake risers, which could lead to inefficiencies. These integrated strategies and modeling efforts highlight the importance of a holistic approach to gas extraction in Lake Kivu. By considering the interdependencies between energy efficiency, lake stability, and economic viability, the extraction project can be managed sustainably, benefiting all stakeholders while preserving the lake’s environmental and unique characteristics.

In the context of the current methane extraction projects in Lake Kivu, this comparative analysis will guide the selection of CO₂ removal technologies that balance energy efficiency with environmental protection. The findings will also contribute to the broader field of biogas upgrading, providing valuable insights into the trade-offs and synergies between different scrubbing methods. Future research should further explore the long-term impacts of these technologies on both the material and energy balances and the ecological health of methane extraction sites, ensuring a holistic approach to sustainable energy production.

This study provides new insights into gas production; however, investment, profitability, and environmental aspects are not discussed or evaluated in this paper, and the optimization was based solely on the project’s energy efficiency. For example, studies show that the optimum pressure for the water scrubbing method is around 3 bar, while for amine scrubbing, it is around 6 bar. While lower pressure could also meet the criteria for density control, lower operating pressure requires larger separators and compressors, which present additional challenges in design, installation, and, consequently, investment cost. Therefore, in addition to technical analysis, an economic analysis is critical for decision-making regarding the CO₂ removal technique and separation conditions.