- 1 - 

 

Selection of a technique to separate carbon dioxide from 

methane for recovery of natural gas at Lake Kivu 

MSc (50/50) RESEARCH REPORT 

Prepared by 

Hermann Ekini Ntini  

0512451Y 

Submitted to 

School of Chemical and Metallurgical Engineering, Faculty of Engineering and the Built 

Environment, University of the Witwatersrand, Johannesburg, South Africa 

 

Supervisor(s):  

Professor Diakanua Nkazi 

 Dr. Elie Mukaya  
 

November 2023 



 - 2 - 

Declaration 
 

I hereby declare that this Research Report entitled “Selection of a technique to separate carbon 

dioxide from methane for recovery of natural gas at Lake Kivu” was written by me.  

 

The review of the published documentations, the simulation and its discussions were put forth 

based on my understanding and analysis. This research report was not submitted for any degree 

before, either in this or in any other Universities. All used ideas were acknowledged at the 

respective place in the text.  

 

 

 

__ ______________________________________ 

Hermann Ekini Ntini  

(Student Number: 0512451Y) 

 

 

 

 

 



 - 3 - 

Acknowledgement 

I would like to express my deep feelings of gratitude to the late Professor Diakanua Nkazi and to 

Doctor Elie Mukaya for the guidance and support received during this research. 

I would like to recognize my family Delphin Kadim Ntini,  Modestine Yalankay Ngolo,            

Carine Sesep, Ayel-e-Nziam Manuella-Michelle Ntini, Nathanael Ekini Sesep Ntini and     

Malaika-Sabine Yalankay Ntini for their faith in me and sacrifices made towards the completion 

of this study.  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



 - 4 - 

Abstract  
 

Lake Kivu is situated between the Democratic Republic of Congo (DRC) and Rwanda. It is known 

to contain large amount of dissolved carbon dioxide and methane. It is termed a killer Lake due to 

the toxic nature of these gases, which could emerge on the surface during a catastrophic eruption 

and cause massive devastation in this region. Extracting these toxic gases proves to be crucial to 

avoid natural disasters and to afford economic benefits in the form of electricity generation or 

energy export. 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



 - 5 - 

Table of Contents  

DECLARATION .................................................................................................................................... - 2 - 

ACKNOWLEDGEMENT ..................................................................................................................... - 3 - 

ABSTRACT ............................................................................................................................................ - 4 - 

TABLE OF CONTENTS ....................................................................................................................... - 5 - 

LIST OF FIGURES ............................................................................................................................... - 7 - 

LIST OF TABLES ................................................................................................................................. - 8 - 

LIST OF ABBREVIATIONS ................................................................................................................ - 9 - 

1 INTRODUCTION ....................................................................................................................... - 10 - 

 PROBLEM STATEMENT ..................................................................................................... - 11 - 

 RESEARCH OBJECTIVES .................................................................................................. - 12 - 

2 LITERATURE REVIEW ........................................................................................................... - 13 - 

 OVERVIEW OF GAS EXTRACTION AT LAKE KIVU .................................................................. - 13 - 

 METHODS FOR SEPARATION OF BINARY MIXTURE OF CARBON DIOXIDE AND METHANE ... - 14 - 

 ABSORPTION OF CARBON DIOXIDE IN ALKALINE SALT SOLUTIONS ..................................... - 14 - 

 CRYOGENIC SEPARATION OF CARBON DIOXIDE IN NATURAL GAS RECOVERY .................... - 15 - 

 ADSORPTION OF CARBON DIOXIDE ON POROUS SOLIDS AND MEMBRANES .......................... - 17 - 

 ECONOMIC EVALUATION CRITERIA FOR PRELIMINARY TECHNOLOGY SELECTION ........... - 19 - 

3 METHODOLOGY ...................................................................................................................... - 21 - 

 DESIGN BASIS AND ASSUMPTIONS ........................................................................................... - 21 - 

5.1.1. Survey of data from Lake Kivu ..................................................................................... - 21 - 

5.1.2. Production rate ............................................................................................................... - 21 - 

5.1.3. Annual operating time and shutdown duration ........................................................... - 22 - 

5.1.4. Material of construction ................................................................................................. - 22 - 

 MODELLING IN ASPEN PLUS V12 AND SIZING OF PROCESS EQUIPMENT ............................. - 22 - 

 Thermodynamic method ................................................................................................ - 22 - 

 Selection of major process equipment .......................................................................... - 23 - 

 PRELIMINARY COST ESTIMATE ............................................................................................... - 30 - 

Criteria for selection of utilities .................................................................................................. - 31 - 

4 DISCUSSION ............................................................................................................................... - 33 - 

 DEGASSING SECTION ............................................................................................................... - 33 - 

 ABSORPTION ............................................................................................................................. - 36 - 

 ADSORPTION ............................................................................................................................ - 44 - 

 COLD SEPARATION................................................................................................................... - 48 - 

 BENEFITS OF THE THREE SEPARATION TECHNIQUES ............................................................ - 52 - 

 COMPARISON OF YIELD AND NATURAL GAS PRODUCT PURITY ............................................ - 53 - 

 COMPARISON OF GROSS INCOME FROM SALES OF ELECTRICITY ......................................... - 54 - 

 COMPARISON BASED ON TOTAL CAPITAL INVESTMENT ........................................................ - 54 - 

 COMPARISON BASED ON NET ANNUAL INCOME ..................................................................... - 55 - 

 COMPARISON BASED ON INTERNAL RATE OF RETURN ...................................................... - 56 - 

5 CONCLUSION AND RECOMMENDATIONS ....................................................................... - 58 - 

REFERENCES ..................................................................................................................................... - 61 - 



 - 6 - 

APPENDIX A – ASPEN PLUS V12 SIMULATION RESULTS ..................................................... - 64 - 

APPENDIX B –EQUIPMENT SIZING, ESTIMATION OF CAPITAL COSTS AND OPERATING 

COSTS ................................................................................................................................................... - 68 - 

APPENDIX C – CALCULATIONS OF INTERNAL RATE OF RETURN .................................. - 86 - 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



 - 7 - 

List of Figures 

Figure 1Process flow diagram................................................................................................ - 33 - 

Figure 2 Process flow diagram............................................................................................... - 36 - 

Figure 3 Process flow diagram............................................................................................... - 44 - 

Figure 4 Process flow diagram............................................................................................... - 48 - 

 

  



 - 8 - 

List of Tables 

Table 1 Properties of Raschig rings packed columns ....................................................................... - 25 - 

Table 2 Properties of packed bed of activated carbon ...................................................................... - 27 - 

Table 3 Overall heat transfer coefficient ............................................................................................ - 30 - 

Table 4 Equipment costing equations[15] .......................................................................................... - 30 - 

Table 5 Sizing of major pieces of process equipment (Appendices A and B) ................................. - 34 - 

Table 6 Purchased cost of major pieces of process equipment (Appendices A and B)[15] ........... - 35 - 

Table 7 Sizing of major pieces of process equipment (Appendices A and B)[15] ........................... - 38 - 

Table 8 Sizing of major pieces of process equipment (Appendices A and B)[15] ........................... - 39 - 

Table 9 Purchased cost of major pieces of process equipment (Appendices A and B)[15] ........... - 41 - 

Table 10 Purchased cost of major pieces of process equipment (Appendices A and B)[15] ......... - 42 - 

Table 11 Major annual cost of utilities (Appendices A and B)[15] .................................................. - 43 - 

Table 12 Sizing of major pieces of process equipment (Appendices A and B)[15] ......................... - 46 - 

Table 13 Purchased cost of major pieces of process equipment (Appendices A and B)[15] ......... - 47 - 

Table 14 Major annual cost of utilities (Appendices A and B)[15] .................................................. - 47 - 

Table 15 Sizing of major pieces of process equipment (Appendices A and B)[15] ......................... - 49 - 

Table 16 Purchased cost of major pieces of process equipment (Appendices A and B)[15] ......... - 50 - 

Table 17 Major annual cost of utilities (Appendices A and B)[15] .................................................. - 51 - 

Table 18 Economic parameters for profitability analysis ................................................................ - 53 - 

Table 19 Yield and product purity ..................................................................................................... - 53 - 

Table 20 Gross annual income ............................................................................................................ - 54 - 

Table 21 Total Capital investment (Appendix B) .............................................................................. - 54 - 

Table 22  Annual operating cost (Appendix B) ................................................................................. - 55 - 

Table 23 Net annual income (Appendix B) ........................................................................................ - 55 - 

Table 24 Internal rate of return (refer to Appendix C) .................................................................... - 56 - 

 

 

 

 

 

 

 

 

 

 

 

 



 - 9 - 

 

                                                                 

 

List of abbreviations  
Aspen Plus V12 – Aspen Plus Version 12 

CEPCI – Chemical Plant Cost Index 

CH4 – Methane 

CO2 – Carbon dioxide 

DRC – Democratic Republic of Congo 

ENRTL-SR – Electrolyte non-random two liquid thermodynamic model with Soave-Redlich  

                       with Soave-Redlich equation of state        

H2O – Water                                                                                                             

KPI – Kibuye Power Limited Stage 1  

PSA – Pressure swing adsorption 

SRK – Soave Redlich Kwong  

TSA – Temperature swings adsorption  

VPSA – Vacuum pressure swing adsorption 

 

  



 - 10 - 

1 INTRODUCTION 
 

Lake Kivu situated between the Democratic Republic of Congo (DRC) and Rwanda, along with 

Lake Monoun and Lake Nyos in Cameroun, are three continental unique water systems due to their 

high content of dissolved carbon dioxide in deep water [1],[2],[30]. In Lake Kivu, the estimated 

volume of dissolved gases is about 300 km3 of carbon dioxide and about 60 km3 of methane 

[1],[30]. Catastrophic limnic eruptions, where large clouds of carbon dioxide were released from 

deep waters have occurred at Lakes Monoun and Nyos in Cameroun in a recent past leading to the 

killing of thousands of human lives and decimation of ecosystems around both Lakes [1],[30] .  

 

Lake Kivu particularly contains much larger amount of dissolved methane along with carbon 

dioxide in deep water. Degassing of the Lake by industrial processes is a lifesaving precaution but 

may still pose a concern on global warming due to the potential for venting of greenhouse gases 

into the atmosphere.  

 

It is therefore necessary to degas the Lake but capture the carbon dioxide for storage at safer 

location or for other uses. The high purity methane recovered may be used for electricity generation 

or commercialized as natural gas for exportation. 

 

Three separation techniques have been shortlisted as strong candidates for recover methane from 

carbon dioxide at Lake Kivu [3]:  

• absorption of carbon dioxide in an alkaline solution of mixed salt,  

• adsorption of carbon dioxide on activated carbon, 

• cold separation of carbon dioxide in liquid form. 

The aim of this study was therefore to investigate and select a gas separation technique amongst 

three alternatives methods of gas separation. Initial flowsheets were developed for the three 

techniques. The total capital investment and the annual operating costs were estimated for each 

flowsheet.  

 



 - 11 - 

Separation by absorption in an alkaline solvent has proved to allow a high yield or recovery of 

methane in the product stream, however at an excessivecost of energy as discussed in literature 

[31].  

 

Cold separation or a distillation utilizing the difference in dew point between carbon dioxide and 

methane (cryogenic process) offers the benefit of a product stream with a high purity of methane. 

However, results obtained in this study points out that the cold separation technique is more 

expensive than the absorption technique with regards to the required cost of capital investment 

[32].  

 

After calculations and comparison of the internal rate of return for the three flowsheets, adsorption 

on activated carbon has been retained as the most economically profitable route to consider for 

recovering methane from Lake Kivu with the intent of supplying international energy markets. 

Results of simulations in Aspen Plus prove that adsorption is the most cost-competitive amongst 

the three techniques when recovering methane from reservoir formations of high content of carbon 

dioxide due to its lower energy consumption and smaller plant equipment footprint [8].  

 

The main disadvantage with adsorption in activated carbon is its lowest yield since a significant 

amount of methane is also adsorbed along with carbon dioxide in activated carbon [26], [Appendix 

B, section e]. Despite this fact, adsorption remains the most cost-effective technique recommended 

in this research in order to extract methane with a more favorable economic profitability from Lake 

Kivu. 

 

 PROBLEM STATEMENT 

An initiative by Kivuwatt, is currently undertaken by a joint consortium between the government 

of Rwanda and private investors to degas Lake Kivu [1]. The initiative generates about 25 MW of 

electricity with gas recovered from the Lake at the Kibuye power station in Rwanda. The degassing 

rate is not currently sufficient to prevent future accumulation of dissolved gases in deep waters, 

therefore still needs to be substantially increased to mitigate the potential risk of a catastrophic 

limnic eruption, especially by maintaining safe concentration of dissolved gases below saturation 

levels[2].  



 - 12 - 

 

Methane can be separated from carbon dioxide after the gas mixture has been extracted from the 

Lake.  Methane may be further used for electricity generation or exported as natural gas. Carbon 

dioxide may be moved to a safer location for storage, sequestration or other uses after being 

separated from the extracted gas. Currently the two greenhouse gases end up in the atmosphere 

after degassing operations, which may affect health and life.  

 

Improved separation processes are therefore necessary to be investigated and developed after 

degassing has taken place at Lake Kivu in support of global efforts for greenhouse gas reduction. 

Such separation steps should be designed as cost-effective to recover high purity methane for 

electricity generation or for exportation and high purity carbon dioxide for storage, sequestration 

or other uses. The separation technology thus must be selected, such that the overall process design 

leads to a reasonable investment cost and to profitable unit operations with low energy 

consumption.  

 

 RESEARCH OBJECTIVES 

The aim of this study was to investigate and select a gas separation technique that can assist in 

methane gas production, and reducing the carbon dioxide concentration, which may emerge from 

solution and cause devastation to the eco-systems around the Lake. The objectives of this study 

were therefore: 

• to conduct steady-state simulation of each separation technique in Aspen Plus V12 process 

simulator in order to achieve a specified purity and production rate of methane as required 

for natural gas export, and calculate 

o the yield of methane of each technique, 

o the cost of capture of carbon dioxide by each technique;  

• to compare and select the optimal alternative amongst the three selected techniques 

(Absorption, adsorption and cold separation) based on criteria of yield and cost of gas 

capture thus obtained.  

 



 - 13 - 

2 LITERATURE REVIEW 
 

 Overview of gas extraction at Lake Kivu 

Lake Kivu is known to contain large amounts of carbon dioxide (CO2) and methane (CH4) in the 

deepest layers of water near the bottom surface[1]. Reserves are estimated at 60 billion Nm3 of 

methane and 300 billion Nm3 of carbon dioxide. The Lake reaches maximum depth of 480 m and 

gases are found dissolved from a depth of 260 m. Layers of water are stratified from more dense 

layers containing the dissolved gases below 260 m to less dense and gas-free layers from 260 m to 

the surface. Carbon dioxide originates from magma processes found in volcanic grounds under the 

Lake. Methane results from the reduction of carbon dioxide, which enters the bottom of the Lake, 

by deep water micro-organisms and from degradation of biological sediments settling from upper 

layers to the bottom of the Lake [1].  

An independent expert committee issued a report in 2006 for a feasibility study for the Kibuye 

Stage 1 Power Limited (KP1) project [1]. The KP1project was a partnership between the 

Government of Rwanda and Dane Associates Limited for a pilot plant, then for a 35 MW power 

plant to be located at Kibuye, Rwanda. The purpose of the committee was primarily to recommend 

safe ways of extracting gas for the power plant while ensuring that the stability of various water 

layers above the bottom is still maintained. The committee indicated that substantial amount of 

methane continuously accumulates every year at the bottom of the Lake. The elevated hydrostatic 

pressure allows gases to dissolve and remain stable in solution. The concentration of dissolved 

gases was still below the saturation level in 2006. The natural rate of accumulation of methane has 

been estimated to 250 million Nm3 per year and if no action is taken to degas the Lake, a 

catastrophic limnic eruption would result within the next 100 years due to saturation of the gases 

in deep water. The committee recommended therefore to extract gases from depth of 320 m or 

deeper. A minimum of three 35MW-power stations producing in total 105 MW was deemed 

enough to achieve the degassing rate of 250 million Nm3 of methane per year.   

A 25 MW power plant fueled by gas extracted from Lake Kivu was commissioned in 2015 at 

Kibuye by KivuWatt, a subsidiary of ContourGlobal [2]. The next phase of the process is to add a 



 - 14 - 

generation capacity of 75 MW, which will total to about 110 MW, therefore exceeding the 

minimum rate of degassing recommended in 2006.  

 Methods for separation of binary mixture of carbon dioxide and 

methane 

Both methane and carbon dioxide extracted from Lake Kivu are greenhouse gases. Their extraction 

process must therefore minimize venting to atmosphere as this undermines efforts to combat 

climate change. An extraction process at Lake Kivu should therefore recover high purity methane 

for power generation or for sale as natural gas, since methane is the main valuable product. High 

purity carbon dioxide will also be recovered and may be considered as a by-product gas from the 

Lake, which is then available for storage, sequestration or exportation for other uses. 

The three commonly used methods to separate carbon dioxide are absorption in an alkaline solvent, 

cryogenic separation and adsorption on a membrane in industrial applications aimed at reducing 

greenhouse gas emissions [3]. 

  Absorption of carbon dioxide in alkaline salt solutions 

Zhang, Ye and Lu [4] have studied the vapor-liquid equilibrium behavior of aqueous solutions of 

potassium carbonate, bicarbonate and carbon dioxide. The effect of the concentration of potassium 

carbonate, the temperature of the solution and the initial load of carbon dioxide in solution on the 

total pressure and ratio of partial pressure of water to carbon dioxide was specifically investigated. 

It was found that when varying the concentration of potassium carbonate in the initial solution at 

high total pressure and high temperature, the ratio of the partial pressure of carbon dioxide to water 

at equilibrium was higher for solutions with a high initial concentration of potassium carbonate. 

Further increasing the temperature of the solution did not increase the partial pressure of carbon 

dioxide in the vapor phase. This result showed that when carbon dioxide is absorbed in a slurry 

with a high concentration of mixed potassium carbonate and bicarbonate, it may be recovered from 

that concentrated solution in a stripper operating at high pressure and high temperature and a much 

lower amount of evaporation of water from the solution would result. The vapor phase which 

leaves the stripper at high pressure contains a higher amount of carbon dioxide and little water, 

which can be removed by a condenser with a low duty. The carbon dioxide may then be 



 - 15 - 

compressed for storage also with a lower duty of the compressor since already leaving the stripper 

at high pressure. The high concentration of the potassium carbonate in the solvent allows the 

stripper to operate with a lower duty of the reboiler to achieve high partial pressure of carbon 

dioxide in the overhead of the stripper. This results in total in less energy to desorb carbon dioxide 

from the solution and to compress it for storage compared to conventional absorption in 

monoethanolamine solutions. The overall absorption and stripping process were therefore 

considered as an alternative method which is less energy-intensive for capture of carbon dioxide 

from flue gas of power stations. 

Jayaweera, Jayaweera, Elmore, Bao and Bhamidi [5] have discussed the use of mixed salt 

solutions to absorb CO2 from post-combustion gas streams. Mixed solutions of ammonia and 

potassium carbonate at a bench scale and pilot scale experimental setup have been tested and were 

proven successful to remove CO2 from flue gas at high loading in a flue gas from a coal-fired 

power plant. The process consists of a two-stage absorption step, first in an ammonia-rich solvent 

then in a potassium carbonate –rich solvent. The absorbed CO2 was then stripped from the CO2 

rich solution in a second solvent regeneration. At the regeneration step, the ammonia-rich solution 

is recovered at the bottom of the column while the potassium carbonate-rich solution at an 

intermediate level on the column. The mixed salt absorption technology was found to provide the 

following benefits which outweighed the conventional neat ammonia, neat potassium carbonate or 

monoethanolamine processes: reduced emissions of ammonia and hazardous wastes, reduced 

cooling requirement since absorption occurs at low temperature with less heat of reaction 

produced, higher absorption rate, stable condition of salt in dissolved phase with no precipitation 

in solid form, lower duty of the reboiler at the solvent regeneration step and lower requirement of 

the energy to compress the pure CO2 which is released from solution at high pressure in the 

regeneration column. The technology has been deemed economically feasible for post-combustion 

and pre-combustion CO2 capture due to its reduced cost of 40$ per ton of CO2 captured. 

 Cryogenic separation of carbon dioxide in natural gas recovery 

In applications which involve separating water from methane during extraction of natural gas on 

offshore platforms, the priority in process design is to reduce the size and the weight of the 

separation equipment to install on the platform [6]. This is commonly achieved via cryogenic gas 



 - 16 - 

separation as supersonic gas flows through a turbine. The natural gas at high pressure is accelerated 

through a turbine such that the temperature of the exhaust gas from the turbine drops below the 

freezing point of water at the lower outlet pressure of the turbine. Under such conditions, water is 

removed as dry ice from the natural gas. Hammera, Wahla, Anantharamana, Berstada and Lervåg 

[6] have conducted a simulation of a case to capture carbon dioxide in the form of dry ice for flue 

gases leaving a coal-fired power station by using a Laval nozzle (convergent-divergent nozzle) as 

the separation equipment. The Laval nozzle was selected since it is a small-size cold separation 

equipment with a lower footprint in comparison to supersonic turbines used on natural gas 

platforms to remove water from methane. The Peng-Robinson equation of state was used as the 

thermodynamic method of simulation in Hysys. The flowsheet was simulated to reduce the power 

consumption of the flue gas compressor upstream of the Laval nozzle. The duty of the feed 

compressor could be reduced with the Laval nozzle in comparison to the typical supersonic gas 

turbine used offshore in the Hysys simulation. The study concluded that the Laval nozzle could be 

used to separate carbon dioxide in dry ice form from the flue gas leaving the coal-fired power 

station by a small-sized process equipment with a similar energy-intensity than the conventional 

absorption in monoethanolamine. The benefit of this process is that it would require a lower capital 

investment than conventional absorption in monoethanolamine. 

The aim of the study conducted by Mohamad [7] was to develop a mathematical model of the 

cryogenic process for separation of carbon dioxide from methane by desublimation or deposition 

of carbon dioxide in solid form on a packed bed at atmospheric pressure. Cryogenic separation of 

carbon dioxide from natural gas has been carried out at industrial scale by Exxon [7] during 

extraction of methane in gas deposits containing high concentration of carbon dioxide.  The 

process of separation occurs in a cycle consisting of three phases: cooling of the bed, deposition 

of carbon dioxide onto the bed and regeneration of the bed. In the cooling phase, liquid nitrogen 

is passed through the bed and cools it to a cryogenic temperature below the deposition point of 

carbon dioxide (-78.5 °C) but still above the dew point of methane (-160°C) at atmospheric 

pressure. The next phase is the capture phase whereby carbon dioxide deposits in solid form on 

spherical particles which makes up the bed as its mixture with methane is passed through the 

cooled bed.  The last phase is the bed regeneration phase whereby the deposited carbon dioxide 



 - 17 - 

sublimates into a hotter gaseous stream of pure carbon dioxide passing through the bed. Water 

which deposited as ice during the capture phase also evaporates during the regeneration phase. 

An experimental setup consisting of an insulated horizontal pipe filled with spherical marbles as 

particles forming the packed bed was designed to obtain characteristics of the bed such as the void 

fraction and the mean particle diameter, necessary to calculate heat and mass transfer coefficients 

in order to derive the profile of the deposition rate of carbon dioxide along the axial length of the 

packed bed. Simulations to close the material and energy balance across the packed bed for various 

molar concentrations of carbon dioxide in the gas mixture introduced in the packed bed were 

performed. These simulations demonstrated that a minimum space-time of 2 minutes was required 

to achieve a zero concentration of carbon dioxide in the gas at the outlet of the packed bed. Below 

2 minutes, the contact time was not enough to completely deposit all the carbon dioxide contained 

in the gas at inlet of the bed. The optimal temperature range at atmospheric pressure was found to 

be between -80°C and -110°C for complete removal of carbon dioxide as solid deposited on the 

particles. The rate of deposition of carbon dioxide also increased with its molar concentration in 

the feed to the bed. Molar concentrations of up to 90% carbon dioxide in the feed led to high rate 

of separation from the mixture with methane. This revealed that cryogenic separation of carbon 

dioxide from methane was adequate for natural gas mixtures with a high concentration of carbon 

dioxide. A limited number of experiments were then conducted to verify the concentration profile 

obtained for a mixture of 70% CO2 and 30% CH4 fed from gas cylinders into the packed bed. The 

bed was previously cooled with liquid nitrogen at cryogenic conditions. The experiments 

demonstrated that carbon dioxide was completely removed at the outlet of the pipe but the actual 

concentration profile could not be verified due to constraints on installing many gas sample points 

along the axial length of the pipe used during the experiments as a plug flow reactor.   

  Adsorption of carbon dioxide on porous solids and membranes 

Adsorption is also being considered as an alternative technique to capture carbon dioxide from gas 

mixtures due to its low energy requirements, its low capital cost and to the ease with which the 

adsorbent may regenerated by modulation of the pressure or temperature of the system to desorb 

the carbon dioxide after the adsorption step [8]. Hence three known types of adsorption are 

pressure wing adsorption (PSA), temperature swing adsorption (TSA) and vacuum pressure swing 



 - 18 - 

adsorption (VPSA) depending on whether pressure or temperature is the variable modulated to 

regenerate the adsorbent. Though not yet available for post-combustion capture of CO2 from flue 

gas due to the low capacity and selectivity of commercially available adsorbents, adsorption is 

considered for use in the separation of CO2 from CH4 in natural gas wells which are often not 

considered viable for exploitation during exploration of natural gas due the high content of carbon 

dioxide.    

Abdullah, Idris, Shamsudin, Kim and Othman [8] conducted an experiment with the aim of 

selecting the best adsorbent for separation of CO2 from a binary mixture with CH4 at an initial 

concentration of 70% CO2 by volume. Four adsorbents were considered in the study: Zeolite, 

which is crystalline aluminosilicates, Zirconium-benzene dicarboxylate, activated carbon made of 

material from a tropical hibiscus tree and activated carbon made from the shell of palm oil fruits. 

The four adsorbents were first analysed for three key characteristics which are mean pore size, 

particle size distribution and nitrogen adsorption capacity in cm3 per gram of adsorbent under 

standard conditions of pressure and temperature. These key characteristics determine adsorption 

capacity. Activated carbon made from shell of palm oil fruit was found to be the best adsorbent 

with a pore size of about 20 Å, its nitrogen adsorption capacity was as high as 220 cm3 per gram 

and its mean particle diameter 112μm. Tests were carried out with the four adsorbents at a pressure 

of 3 bar and at ambient temperature. Activated carbon made from the shell of palm oil fruit yielded 

superior affinity for carbon dioxide than methane: methane was recovered with a yield of 89% 

during adsorption at 3 bar and the purity of the carbon dioxide in the gas recovered during 

desorption at atmospheric pressure reached 94%. 

He and Hägg [9] have evaluated the feasibility of different techniques to capture carbon dioxide 

based on energy consumption and cost of capture of carbon dioxide as criteria of comparison for 

fossil-fuel power stations. Energy consumption was quantified by thermal efficiency, i.e., the ratio 

of the net power exported by the power station including the carbon dioxide capture section to the 

heat generated by burning coal. Cost was defined as the investment cost of the process per unit 

amount of carbon dioxide captured.  A baseline was defined for the net power output from the 

plant with the carbon capture section included. The most competitive technique would be the one 

which would achieve the baseline net power output of an existing plant without the carbon-capture 

facility at a higher thermal efficiency and lower cost of carbon capture. Two techniques were 



 - 19 - 

investigated via simulations in HYSYS: adsorption on a fixed-site carrier membrane and 

conventional absorption in monoethanolamine. The results of the simulation achieved in terms of 

power consumption of auxiliary equipment such as compressors, pumps, expanders needed, were 

used to size process equipment relevant to each technique, therefore, to estimate utilities and 

capital cost for each case. Adsorption was found to have a higher thermal efficiency and a lower 

cost of capture than conventional absorption in monoethanolamine.  

 Economic evaluation criteria for preliminary technology selection 
 

Sinnott [11], Peters and Timmerhaus[14] have discussed in detail various measures to assess the 

economic viability of an investment in a process or chemical plant by use of some reliable 

profitability measures. At an early stage during the design of a manufacturing process, alternative 

process flowsheets are normally obtained showing the sequence of major process equipment 

required to manufacture the finished chemical product or commodity. The calculations of the 

internal rate of return, also known as the discounted cash-flow rate of return is one of the simple 

methods to evaluate the economic performance of different options of flowsheet at an early stage 

of the project. At this point, only preliminary estimates of the cost of major equipment are available 

to the designer within an accuracy of ± 30%. This level of accuracy is enough to highlight design 

options that are worth further exploring in later stages of process design.   

 

The internal rate of return takes into consideration the time-value of money over the entire life 

cycle of the project, thus offers a comprehensive to evaluate the profitability over the duration of 

the project which is the same for different alternative design routes being investigated. A simple 

definition of the internal rate of return is that it is the present-day interest rate which is just enough 

to ensure that the total capital invested in the present is paid back exactly at the end of the life 

cycle of the project. The internal rate of return is the value of r that satisfies the equation below 

over a time t, which is the life cycle of the process equipment before major renewal of the 

equipment is undertaken. For the purpose of this study, the depreciation of equipment over years 

of the project is excluded and the net annual profit is the difference between the gross annual 

income from sales of the commodity and the major annual cost of utilities required to achieve the 

desired separation between methane and carbon dioxide.  



 - 20 - 

𝑇𝑜𝑡𝑎𝑙 𝐶𝑎𝑝𝑖𝑡𝑎𝑙 𝐼𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡 − ∑
𝑁𝑒𝑡 𝑎𝑛𝑛𝑢𝑎𝑙 𝑝𝑟𝑜𝑓𝑖𝑡

(1 + 𝑟)𝑡

𝑡

𝑘=0

= 0 

Recommended values of the internal rate of return in order to deem a preliminary flowsheet 

worth of further exploration span between 20 and 30%.    



 - 21 - 

3 METHODOLOGY 

The purpose of the research project is to compare three different potential techniques to separate 

carbon dioxide from methane during recovery of natural gas from Lake Kivu. This requires a 

definition of a rate of flow, common to the three techniques, at which the methane is to be produced 

during recovery via each given separation technique. This rate of flow serves as the basis for 

calculations to determine the size of equipment and the amount of utilities to be available for each 

technique. These latter quantities can then be used to obtain order-of-magnitude estimate of the 

capital and operating cost of each technique, hence enables comparison and selection of the cost-

efficient amongst the three.  

The study has been conducted following the four stages listed below. 

  Design basis and assumptions 
 

5.1.1. Survey of data from Lake Kivu  

Considering elements discussed in the preceding section, an estimated annual production rate of 

250 million Nm3 of methane is the degassing rate needed to maintain concentrations of dissolved 

gases at safe level in Lake Kivu. This comparative study will therefore be based on extracting high 

purity methane at a rate of 250 million Nm3 per year at a depth of 320 m. The pressure at that depth 

is about 32 bar and the temperature 23°C. The dissolved gas concentrations at 320 m are 16 mmol/l 

of methane and 80 mmol/l of carbon dioxide [1].  

5.1.2. Production rate  

The annual volumetric flow rate required to degas Lake Kivu in order to maintain the concentration 

of dissolved toxic gases below safe level in its deep layers has been estimated as 250 million 

Nm3/year of methane at a purity of at least 98% (mol % or vol%). 98% methane is enough as a 

criterion to deem the recovered natural gas fit for commercialization in the world energy market 

[22]. 

 



 - 22 - 

5.1.3. Annual operating time and shutdown duration  

The online time for the separation process is 330 days per year, hence the shutdown period for 

maintenance is 35 days per year. A storage of methane produced should be available to ensure that 

product is still delivered to the market during the 35 days per year when operation is posed for 

annual maintenance. To achieve an extraction rate of 250 million Nm3/year of methane during 330 

days of operation, the equivalent molar flow rate is about 391mol/s of 98% (mol %). 

5.1.4. Material of construction 

The presence of carbon dioxide dissolved in water implies that the presence of weakly acidic 

conditions in the streams of water extracted from the Lake. To ensure that the equipment to be 

installed for the methane recovery operation withstand the corrosive environment due to presence 

of dissolved carbon dioxide, stainless steel has been selected as material of construction for all 

equipment in this investigation. 

  Modelling in Aspen Plus V12 and sizing of process equipment 

The objective of each separation technique to be investigated will be to recover 250 million Nm3 

per year of methane at 98% purity. 98% is considered as acceptable for export as natural gas. The 

volume (molar) composition of the gas mixture from Lake Kivu after extraction from the bottom 

consists mainly of 63% carbon dioxide, 36% methane and less than 1% of water vapor. At this 

stage, water has been separated from the gas. The initial temperature of the gas mixture fed to the 

separation equipment will be 20 °C and the starting pressure will vary according to each separation 

technique. This is the condition of the gas after having been extracted and separated from water at 

the surface obtained from an initial simulation in Aspen Plus V12.  

The simulation in Aspen generates the preliminary size of major process equipment and the 

amount of utilities required for each of the three techniques being investigated. 

 Thermodynamic method 

For each technique, an appropriate thermodynamic method to represent the behavior of carbon 

dioxide and methane mixture has been selected as recommended in Aspen Plus. Each technique 

has been simulated with the objective of achieving a final production rate of 250 million Nm3 per 



 - 23 - 

year of methane at 98% purity. The Soave-Redlich-Kwong (SRK)   and the Peng Robinson cubic 

equation of state with the Boston-Mathias (PR-BM) alpha function cubic equations of state are 

used as property method in Aspen Plus for the flowsheets of cold separation  and adsorption 

respectively. They are particularly recommended for gas processing, refinery and petrochemical 

applications. 

For the case of the absorption flowsheet where carbonate and bicarbonate ions are present as ions 

in solution. The ENRTL-SR property method was selected in Aspen Plus V12. It is based on 

the Symmetric Electrolyte NRTL property model with the following considerations: 

• the Redlich-Kwong equation of state for vapor phase properties 

• the symmetric reference state (pure fused salts) for ionic species. 

• Henry's law for solubility of supercritical gases. 

• symmetric Electrolyte NRTL method of handling zwitterions 

 Selection of major process equipment 

a. Piping for gas extraction 

Carbon steel piping with a schedule 40 is selected for the piping to be submerged into the deep 

layers of the Lake in order to extract dissolved gases and route them to the surface. The maximum 

available diameter in the equipment cost estimator used is 0.508 m [15]. The length of the piping 

is taken as 320 m since this is the depth from which dissolved gas are extracted from the Lake.   

The goal of the simulation in Aspen Plus V12 is to calculate the flow rate per piping with diameter 

of 0.508 m and length of 320 m required to achieve  a gas phase emerging at the surface of the 

Lake with a molar content of 63 mol% carbon dioxide and 36 mol % methane. The starting liquid 

at a depth of 320 m has a composition of 8 mmol % carbon dioxide and 1.6 mmol% methane, the 

rest being Lake water. The piping is to ensure that the two-phase flow regime consists of gas 

bubbles dispersed in the liquid phase as the mixture is drawn upwards to the surface. 

 

 

file:///C:/ProgramData/AspenTech/Aspen%20Plus%20V12.0/HtmlHelp/Subsystems/ref2/Content/html/peng_robinson_boston_mathias.htm
file:///C:/ProgramData/AspenTech/Aspen%20Plus%20V12.0/HtmlHelp/Subsystems/ref2/Content/html/peng_robinson_boston_mathias.htm
file:///C:/ProgramData/AspenTech/Aspen%20Plus%20V12.0/HtmlHelp/Subsystems/ref2/Content/html/Symmetric_Electrolyte_NRTL_Activity_Coefficient_Model.htm
file:///C:/ProgramData/AspenTech/Aspen%20Plus%20V12.0/HtmlHelp/Subsystems/ref2/Content/html/redlich_kwong2.htm
file:///C:/ProgramData/AspenTech/Aspen%20Plus%20V12.0/HtmlHelp/Subsystems/ref2/Content/html/henry_sconstant.htm


 - 24 - 

b. Gas-liquid separators or flash drums   

Gas-liquid separators are required in the different sections of the flowsheets developed in this study 

to allow recovery of methane as the key profit component in the gas phase in which it emerges 

along with carbon dioxide after extraction from the Lake. Horizontal gas-liquid separators are 

sized by the Souders-Brown method [24] which is recommended as enough for preliminary sizing 

of gas-liquid separation vessels. Separation essentially occurs as liquids droplets flow vertically 

downwards, being removed from the bulk gas flow that occur upwards.  

The liquids are considered to settle downwards under the influence of gravity as spherical droplets. 

A maximum allowable gas velocity is defined which must be less than or equal to the terminal 

settling velocity of the droplets from the gas phase. Should this gas velocity exceed the maximum, 

liquid droplets will be entrained upwards in the gas stream and separation will not take place. 

Therefore, the dimensions of the gas-liquid separator vessels are chosen to ensure that the gas 

velocity remains below the maximum allowable. The sizing equations used in this approach are 

therefore listed below.  

The maximum allowable velocity of the gas in the separator is calculated by 

𝑣𝑔𝑎𝑠 𝑚𝑎𝑥 = 𝐾𝑠 ∗ √
(𝜌𝑙 − 𝜌𝑔)

𝜌𝑔)
              (1) 

Where Ks is a sizing parameter given by 𝐾𝑠 = 0.152 ∗ (𝐿/3.05)0.56;  

𝜌𝑙 is the liquid density in kg/m3 and 𝜌𝑔  is the gas density in kg/m3. 

The minimum vessel diameter to ensure that the maximum gas velocity is not exceeded is then 

calculated as 

𝐷𝑚𝑖𝑛 = 𝐾𝑠 ∗ √
4
𝜋 ∗ 𝑞𝑎

𝑣𝑔𝑎𝑠 𝑛𝑎𝑥
                    (2) 

where qa is the gas flow rate at the actual flowing condition in m3/s. 

The available diameter of the vessel chosen is 4 m and the length-to-diameter is selected as 4. This 

implies that all gas-liquid separators are 4 m wide in diameter and 16 m long in length. The above 



 - 25 - 

equation is then applied with material balance data from the simulation performed for each 

particular vessel in Aspen in order to calculate the minimum diameter, ensuring that it is less than 

4 m, which is the standard size chosen for all vessels in this study. 

 

c. Compressors 

The duty of compressors is calculated in Aspen Plus V12 by specifying the required discharge 

pressure. Referring to Smith et al [12], it is assumed that isentropic compression takes place with 

both an isentropic and mechanical efficiency set at 70%.  

d. Expansion valve 

An expansion valve is required in the flow sheet of cold separation to enable the pressure drop 

required to generate the Joule-Thomson effect for cooling the gas below the boiling point of the 

mixture. At such low temperature, conditions are sufficiently cryogenic to achieve the removal of 

most of the CO2 as liquid phase at the bottom of the gas-liquid separator drum.    It is enough in 

Aspen Plus V12 to specify the pressure at the outlet of the pressure reduction valve.  

e. Separation columns 

Absorption and stripping columns 

Packed beds are preferred in this study due to large liquid flow rates required for absorption into 

the liquid phase and desorption from the liquid phase of carbon dioxide [13]. The second reason 

for the choice of packed columns is the corrosive nature of solutions of carbonate or bi-carbonate 

ions present in the solution (alkaline media) where absorption or desorption of carbon dioxide 

takes place. 

 Table 1 Properties of Raschig rings packed columns 

  Properties of Raschig rings packed columns 

Aspen Plus V12 

Porosity 0.64 

Particle diameter 13 mm 

Bed spec surface, Sb 367.0/m 

Available maximum column diameter 3 m 

Column height 6 m 



 - 26 - 

The choice of ceramic packing consisting of Raschig rings is therefore motivated by the above two 

considerations. Standard available columns with a maximum diameter of 3 m and a height of 6 m 

are selected [15]. The choice of the large diameter is motivated by the large flow rates through the 

packing, which allows moderate velocities of the gas and liquid phases contacting each other in 

the columns, thus low pressure drop across the packing. Low pressure drop is particularly 

necessary to ensure that the pressure in the columns remain high enough to minimize losses of 

water through evaporation, hence ensuring that concentrations of dissolved salts in the liquid 

phases are maintained below saturations levels. High evaporation of water from solutions could 

result into the precipitation of salts from solution and blockages in the process. 

Adsorption and desorption columns 

Some considerations have been noted while selecting adsorption as possible technique to separate 

methane from carbon dioxide in the gas extracted from Lake Kivu. For adsorption to be 

competitive to remove carbon dioxide and generate methane at high purity, activated carbon is to 

be used as the adsorbent. The choice of activated carbon is motivated by the fact that this 

commercial adsorbent has a high selectivity for carbon dioxide over methane. This is explained by 

an affinity for activated carbon to interact better with carbon dioxide than with methane as carbon 

dioxide is a more polar compound compared to the non-polar methane.  

Abdullah et al [26]have conducted experiments to assess the performance of adsorption of carbon 

dioxide on activated carbon. Adsorption of carbon dioxide on activated carbon follows a type 1 

isotherm, which is characteristic of monolayer absorption. According to this model, at constant 

temperature, the saturation concentration of CO2 in the adsorbed phase increases with the partial 

pressure of CO2 in the gas phase until a plateau is reached at higher pressures.  The Freunlich 

equation has been found to provide a good model for the adsorption isotherm of CO2 on activated 

carbon. The selectivity of carbon dioxide over methane on activated carbon has been recorded 

against pressure by Heuchel et al [25] at pressures of up to 15 bar, the selectivity value ranges 

between values of 1 and 10 at 5 bar. 

Kinetics data on the adsorption of CO2 at various pressures and temperatures on activated carbon 

have been reported by Singh and Kumar [27]. Results obtained were found to fit a pseudo second 

order kinetic model where the rate of adsorption of CO2 is proportional to the square of free 



 - 27 - 

adsorption sites available in the pores of the adsorbent. The free adsorption site available in the 

pores is expressed as the difference between the saturation concentration and the instantaneous 

concentration in the adsorbed phase.  

Properties of beds of activated carbon have been sourced and derived from results obtained from 

the various sources. Those properties are reported in the table 2. 

Table 2 Properties of packed bed of activated carbon 

Properties of packed bed of activated carbon 

Hauchhum and Mahanta (2014) 

Porosity 0.5  
Bulk density 350 kg/m3 

Particle diameter 0.92 Mm 

Pore volume 0.068 cm3/g 

Bed spec surface, Sb 543.0 /m 

Total spec surface, S 1087.0 /m 

Freunlich equation 

 𝐶𝑠𝑎𝑡 = 𝐾 ∗ 𝑃𝐶𝑂2

1

𝑛  where K = 0.505; n =1.701 at 5 bar, 298K 

 𝐶𝑠𝑎𝑡 , saturation concentration (mg CO2/g activated carbon)  and 𝑃𝐶𝑂2, partial 

pressure (bar) 

𝐶𝑠𝑎𝑡 = 227.0 mg/g at 5 bar and 298 K 

Heuchel et al (1999) 

Selectivity range 1-10 

Singh and Kumar (2015) 

Rate of adsorption,  𝑟𝐶𝑂2 = 𝑘 ∗ (𝑀𝑠𝑎𝑡 − 𝑀)2                                                

(mol/kg/s) 

𝑘, CO2 adsorption rate constant (5 bar, 298K)             0.03 g AC/mg CO2/min 

                                                                                               0.022 kg AC/mol 

CO2/s 

 𝑀𝑠𝑎𝑡 = 𝐶𝑠𝑎𝑡/44 =227.0/44 = 5.15 mol/kg and  

 𝑀 instantaneous concentration of CO2 in adsorbed phase in mol/kg 

 0.03 g/(mg.min) 

The contact time between the gas and the adsorbent should also be long enough to enable enough 

adsorption of the carbon dioxide as the gas travels through the bed of adsorbent. This implies that 

the bed should be sized such that the superficial velocity of carbon dioxide is low enough to allow 

a longer contact time with the adsorbent.  The adsorption step has slower kinetics and is the rate- 

limiting step as the gas passes through the adsorbent. To enable a low superficial velocity, the 



 - 28 - 

diameter of the packed column containing the bed of activated carbon should be selected to be 

sufficiently large.  

The low superficial velocity also enables to minimize pressure losses as the gas flows across the 

bed. The advantage of minimal pressure drop is that the operating pressure is maintained constant 

throughout the length of the bed and sufficiently high to enable adsorption at high rates throughout 

the bed. Adsorption occurs at faster rates as the operating pressure is increased.  

As the gas extraction rate from the Lake will be large, in order to maintain the superficial velocity 

low enough through the adsorption bed, the gas flow will be split across many columns of identical 

diameter containing the same mass of adsorbent and arranged in parallel. In this manner, the flow 

rate through a single column is small enough to achieve a long contact time and an insignificant 

pressured drop.  

As more carbon dioxide is adsorbed in each column, its bed of activated carbon becomes saturated 

with this compound and needs to be regenerated for the next cycle of adsorption. Each column will 

therefore undergo a cycle of adsorption and desorption. Desorption will be achieved by lowering 

the pressure in the column to vacuum conditions to allow carbon dioxide to return to the gas phase 

and free the bed of activated carbon for the next adsorption cycle, i.e., pressure swing adsorption. 

Each column undergoing an adsorption cycle will have a counterpart column undergoing the 

desorption cycle such that the production rate can be maintained by swinging each time between 

the saturated and the regenerated column. 

f. Pumps 

The power requirement for centrifugal pumps is calculated in Aspen Plus V12 by specifying the 

required discharge pressure. The overall efficiency of centrifugal pump is assumed as 75% (Smith, 

Van Ness and Abbott) [12],[23].   

g. Heat exchangers 

Shell-and-tube heat exchangers are chosen as the preferred standard type used in the process 

industries. The design with the floating head is recommended for the ease of maintenance of the 



 - 29 - 

tube bundles that it provides [11]. Counter-current flow is the preferred arrangement used in all 

heat exchangers in the study. 

The standard dimensions used for the tubes on all heat exchangers are outside diameter 0.019m; 

square pitch 0.0254m; tube length 4.88m and design pressure 690 kPa. Stainless steel is the 

material of construction of choice [15]. 

The sizing exercise consists of obtaining the duty of the heat transfer from the flowsheet model 

developed in Aspen. The log-mean temperature difference is then calculated for the counter-

current flow arrangement. Given the typical ranges for values of the overall heat transfer 

coefficient reported by Sinnott [11] (Coulson and Richardson, volume 6), the area required for 

heat transfer is then calculated through the set of formula listed below. 

The log-mean temperature difference is first calculated for the case of counter-current flow of 

material by 

∆T𝐿𝑀 =
(𝑇ℎ𝑜𝑡 𝑖𝑛−𝑡𝑐𝑜𝑙𝑑 𝑜𝑢𝑡)−(𝑇ℎ𝑜𝑡 𝑜𝑢𝑡−𝑡𝑐𝑜𝑙𝑑 𝑖𝑛)

𝑙𝑛(
𝑇ℎ𝑜𝑡 𝑖𝑛−𝑡𝑐𝑜𝑙𝑑 𝑜𝑢𝑡
𝑇ℎ𝑜𝑡 𝑜𝑢𝑡−𝑡𝑐𝑜𝑙𝑑 𝑖𝑛

)
          (3)       

then the heat transfer area as  

A =
𝑄

𝑈 ∗ ∆𝑇𝐿𝑀
                                                             (4) 

quantities are ∆T𝐿𝑀 log-mean temperature difference (℃); 𝑇ℎ𝑜𝑡 𝑖𝑛the temperature of the hot stream 

entering the heat exchanger (℃);  𝑇ℎ𝑜𝑡 𝑜𝑢the temperature of the hot stream leaving the heat 

exchanger (℃); 𝑡𝑐𝑜𝑙𝑑 𝑖𝑛the temperature of the cold stream entering the heat exchanger (℃);  

𝑡𝑐𝑜𝑙𝑑 𝑜𝑢𝑡the temperature of the cold stream leaving the heat exchanger (℃); U the overall heat 

transfer coefficient (kW/m2/℃); and A the required area in m2. 

Overall heat transfer coefficient relevant to this study are listed in table 3 [11]. 

 

 

 



 - 30 - 

Table 3 Overall heat transfer coefficient 

Type of heat transfer Hot stream Cold stream U (kW/m2/℃) 

Heater Medium pressure steam 

at 150℃ 

Aqueous solution 1 

Cooler Gas Cooling water at 19℃ 0.3 

Cooler Aqueous solution Cooling water at 19℃ 0.8 

Condenser Aqueous 

vapor/condensing gas 

Cooling water at 19℃ 1 

Vaporizer or reboiler Medium pressure steam 

at 150℃ 

Aqueous solution 1.5 

Exchanger 1 aqueous solution aqueous solution 0.8 

Exchanger 2 Gas gas 0.01 

 

 Preliminary cost estimate 

Table 4 Equipment costing equations [15] 

 Equipment Size (S) 

Min - Max 

Purchased Cost 

(Cp)  

Constants in 

equation 

Centrifugal compressor, 

motor-driven 

75 – 6000 kW (power) 𝐶𝑝 = 𝑎 ∗ 𝑆𝑏 𝑎 = 2193.2; 
𝑏 = 0.9435 

Centrifugal pump [23] 6 – 174 kW (power) 𝐶𝑝 = 𝑎 ∗ exp (𝑏 ∗ 𝑆) 𝑎 = 8870.3; 
𝑏 = 0.0124 

Turbo-blower 1 – 4.72 m3/s (flow) 𝐶𝑝 = 𝑎 ∗ 𝑆𝑏 𝑎 = 45868; 
𝑏 = 0.616 

Expander, stainless steel 

gate valve, 1035 kPa rating 

0.0508 – 0.305 m 

(diameter) 

𝐶𝑝 = 𝑎 ∗ 𝑆2 + 𝑏 ∗ 𝑆 + 𝑐 𝑎 = 139255; 
𝑏 = 8071.3; 

𝑐 = 1018.9; 

Shell-and-tube heat 

exchanger 

9.3 – 1000 m2 (area) 𝐶𝑝 = 𝑎 ∗ 𝑆2 + 𝑏 ∗ 𝑆 + 𝑐 𝑎
= −0.04 𝑡𝑜 − 0.06 

𝑏 = 234.41; 

𝑐 = 9000 𝑡𝑜 20 000; 

Absorption/stripping 

column, diameter 3m 

6 – 50 m (height) 𝐶𝑝 = 𝑎 ∗ 𝑆+b 𝑎 = 18291 

𝑏 = 86158 

Ceramic packing for 

absorption/stripping 

columns diameter 3m 

5.3 – 56 m (height) 𝐶𝑝 = 𝑎 ∗ 𝑆+b 𝑎 = 5474; 
𝑏 = 3454 

Adsorption/desorption 

column, diameter 3m 

1 – 20 m (height) 𝐶𝑝 = 𝑎 ∗ 𝑆+b 𝑎 = 71317 

𝑏 = −31.625 

Piping for degassing 

section 

1 – 1 000 000 m 

(length) 

𝐶𝑝 = 𝑎 ∗ 𝑆 𝑎 = 128.5375 

Gas-liquid drum 

horizontal, diameter 4m, 

stainless steel, design 

pressure 1035 kPa [23] 

7.9- 53.7 m (height) 𝐶𝑝 = 𝑎 ∗ S + b 𝑎 = 12291; 
𝑏 = 50577 



 - 31 - 

A preliminary flowsheet has been obtained for each technique after simulation in Aspen Plus 

V12. The flowsheet reports the sequence of major pieces of process equipment, inlet and outlet 

process conditions as well as preliminary sizes of equipment. Bulk equipment sizes have been 

used to make preliminary estimates of the installed capital cost and of the cost of utilities for 

each technique of separation under this investigation. 

The purchase cost of process equipment is estimated using the technique as recommended as 

described by Sinnott [11]; and Peters and Timmerhaus [14], [15]. The technique consists of first 

on inferring the unknown cost by scaling up the calculated size from the Aspen Plus models in 

relation to the size and cost of the same type of process equipment as known at a previous point in 

time. 

Equations that approximate the online cost estimator of Peters and Timmerhaus (2002) are listed 

in table 4[15]. An equation is given for each type of equipment for a selected range of size or 

capacity. These are the equations used to estimate the purchased cost of each piece of equipment 

with the year 2002 as basis of reference for cost inflation. 

The scaled-up cost of all equipment is then summed up to yield the total cost at the previous point 

in time. This previous total cost is then updated to current time by multiplying by a factor which 

is the ratio of the equipment cost index at present time to the same cost index at the previous point 

in time. The chemical plant cost index (CEPCI) published by the Chemical Engineering journal 

was used as the index for cost inflation over time in this research [15]. The total purchased cost is 

then multiplied by a cost installation factor in order to obtain the total installed cost of the 

equipment for each technique investigated. 

Criteria for selection of utilities 

a. Electricity  

Electricity is available at a unit price of 0.095$/kWh in the local regions of the Democratic republic 

of Congo and Rwanda within the vicinity of Lake Kivu [18]. 

b. Steam   



 - 32 - 

Medium pressure steam is supplied at 500 kPa and at a temperature of 150 ℃ [20].  

c. Cooling water  

Cooling water is essentially water taken from the Lake supplied to the process at a temperature of 

19℃ [21]. 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



 - 33 - 

4 DISCUSSION  
 

  Degassing section 

S1
DM1

S2

20°C,
 3200 kPa

20°C,
127 kPa

20°C,
127 kPa

LN1

S0

S3

20°C, 
127 kPa

0.0016 mol% CH4
0.008 mol% CO2

99.9904 mol% H2O 

0.0000 mol% CH4
0.0052 mol% CO2

99.9948 mol% H2O 

36.0 mol% CH4
63.0 mol% CO2
1.0 mol% H2O 

 
Figure 1Process flow diagram 

Gas from the deep layers of the Lake is recovered in the degassing section. This is the first section 

of the plant, which is common to all three techniques of separation.  

In the degassing section, water containing the dissolved gases rises from the bottom to the surface 

of the Lake through the piping system denoted LN1. This upwards flow is driven by the difference 

between the high pressure at the bottom of the Lake (3200 kPa) and the nearly atmospheric 

pressure at the surface (127 kPa). The flow happens at constant temperature of 20℃, which is the 

temperature in the deep layers of the Lake. As the pressure rises, gases emerged from the dissolved 

phase into the gas phase, hence the steam flowing up in the piping system LN1 consists of a gas-

liquid flow. The piping system LN1 is submerged in the Lake. It emerges on the surface where it 

connects to the gas-liquid separator drums set DM1. During the degassing process, most of the 



 - 34 - 

methane is recovered from the liquid into the gas phase as well as up to more than half of carbon 

dioxide. Most of the water containing the remaining amount of carbon dioxide is returned to the 

Lake at a higher elevation than the deeper layers of origin. 

The calculated flow rate from the bottom of the Lake is 92 000 000 kmol/hr. This is the molar flow 

rate required to produce 250 million Nm3 per year of methane, which is necessary to keep the 

concentration levels of dissolved gases below the safe level of 16 mmol/l of methane and 80 

mmol/l of carbon dioxide. 

The gas phase extracted in the degassing section contains about 36 mol% methane and 63 mol % 

carbon dioxide. This is the starting concentration for the next separation section where the purity 

of carbon dioxide is expected to be increased to at least 98 mol %. The molar flow rate of gas to 

be extracted in the degassing section depends on the separation efficiency of each of the three 

techniques to be explored in the next separation section.  

Results of simulation in Aspen Plus V12 for the degassing section are presented in Appendix A – 

Aspen Plus V12 simulation results.  

The vertical length of the submerged pipe from the bottom of the Lake to the surface is equal to 

320 m, which is the average depth at which dissolved gases are present in water at the bottom of 

the Lake. A standard diameter for carbon steel piping available is 0.508 m15. For a single pipe 

with a diameter of 0.508 m and a vertical length of 320 m, the calculated flow rate to achieve 

process conditions was found to be 100 000 kmol/hr. This flow rate ensures that the flow regime 

in the pipe remains in a stable form where the emerging gas is dispersed as fine bubbles in 

Table 5 Sizing of major pieces of process equipment (Appendices A and B) 

Equipment Type, size parameters per unit equipment Quantity 

Degassing piping, LN1 Carbon steel welded pipe, Sch 40 

Available diameter: 0.508 m  

Flow per pipeline: 100 000 kmol/hr 

Total flow: 92 000 000 kmol/hr 

920 

Degassing drums, DM1 

 

Type: horizontal tank 

Diameter: 4 m 

Length: 16 m 

Minimum diameter required: 3.1 m 

1 



 - 35 - 

the liquid phase. To achieve the total flow rate of 92 000 000 kmol/hr, the number of single pipes 

with diameter of 0.508 m is 920.  

The purpose of the degassing section is only limited to siphon dissolved gases out of the liquid 

phases. This merely requires carbon steel piping which is submerged into the deep layers of the 

Lake from where water containing dissolved gases is extracted. Gases then emerge out of solution 

as the mixture flows upwards to the surface of the Lake.  

Gas-liquid separators at the surface allow the gas phase to be removed as overhead while the liquid 

phase is routed back into the Lake. As such degassing only occurs via two-phase flow process and 

does not require utilities in the form of electricity, steam or cooling water.  

Table 6 Purchased cost of major pieces of process equipment (Appendices A and B) [15] 

Equipment Type Cost per 

unit 

Quantity Subtotal 

Degassing 

piping, LN1 

Carbon steel welded 

pipe, Sch 40 

Available diameter: 

0.508 m 

$41 132 920 $37 841 440  

Degassing 

drums, DM1 

Type: horizontal 

tank 

Diameter: 4 m 

Length: 16 m 

$247 233 1 $247 233 

Subtotal (2002), CEPCI 390.4 $38 088 673 

Subtotal (2022), CEPCI 877.5 $85 611 707 

The total purchased cost of the degassing section is therefore limited only to the approximate 

capital cost of the pipeline and the gas liquid separators. This is the present-day capital cost 

reported as $85 611 707 in table 6. The cost to purchase piping represents the major portion of the 

capital investment for this section. 

 

 

 



 - 36 - 

  Absorption 
 

S1

137°C, 
400 kPa

DM2

ES1

VL1

VL2

ES2

DM3

DM4

PC1

S3

S4 S5

S7

S9

S10

S11

S15

S17

S19

S16

S2

HX1

S13

ES3

S8

ES4
S14

44°C, 
400 kPa

20°C, 
400 kPa

20°C, 
100 kPa

20°C, 
100 kPa

97.0 mol% CH4
1.0 mol% CO2
2.0 mol% H2O
1462 kmol/hr

56°C, 
400 kPa

50°C, 
400 kPa

54°C, 
100 kPa

92°C, 
400 kPa

97°C, 
100 kPa

50°C, 
100 kPa

97°C, 
100 kPa

140°C, 
500 kPa

140°C, 
400 kPa

50°C, 
100 kPa

50°C, 
100 kPa

50°C, 
100 kPa

50°C, 
100 kPa

50°C, 
100 kPa

50°C, 
400 kPa

87.5.0 mol% CO2
12.2 mol% H2O
0.3 mol% CH4
1599 kmol/hr

0 kmol/hr

2 mol% K2CO3
98 mol% H2O

204 450 kmol/hr

36 mol%    CH4
63 mol%    CO2
1 mol%   H2O

   4076 kmol/hr

ES5

S12

ES6

KC1

S0

20°C, 
127 kPa

S6

 
Figure 2 Process flow diagram 

Gas from the degassing section received at 20℃ is compressed by KC1 from 127 to 400 kPa. It 

then enters the absorption columns system VL1 at 137℃ and 400 kPa. The absorption columns 

are packed columns where gas containing about 63 mol% carbon dioxide and 36 mol% methane 

is brought in contact with an alkaline solution of mixed potassium carbonate and potassium 

bicarbonate.  Absorption of carbon dioxide takes place from the gas phase into the alkaline liquid 

phase through the equilibrium chemical reaction below where carbon dioxide is converted into the 

bicarbonate ion: 

CO2(g) + K2CO3 (aq) + H2O (l) ↔ 2KHCO3 (aq)  

Absorption takes place at a range of temperatures between 44 to 56℃ at a pressure of 400 kPa. 

unreacted Methane passes through the columns and most of it exits in the overhead gas from the 

columns. This overhead gas is cooled in the water cooler ES1. Water that has evaporated from the 

column is condensed at outlet of ES1 and collected as liquid phase with some dissolved carbon 



 - 37 - 

dioxide in the gas-liquid separator system DM2. The overhead gas from DM2 is the desired 

product methane recovered at 97 mol %. 

The bicarbonate-rich solution leaving the absorption step flows through the heat integration heat 

exchangers HX1. Further heating is achieved through the steam heater ES3 at the outlet of which 

the bi-carbonate rich solution enters stripper columns system VL2. VL2 is fitted with the reboiler 

ES5. CO2 is stripped from the liquid phase at a pressure of 100 kPa and a range of temperature 

between 97 and 140 ℃, exits at the overhead of VL2 with mostly evaporated water. The overhead 

gas emerging from the stripping columns is cooled in ES2. Evaporated water from VL2 condenses 

at the outlet of ES2 and is recovered in the drums DM3, then in the mixing tank DM4. Emerging 

as gas phase from DM3 is a side stream containing at least 87 mol % CO2 with the balance being 

evaporated water. 

The solution leaving at the outlet of the reboiler ES5 is mostly a lean solution of potassium 

bicarbonate. This lean solution is cooled in HX1, further in ES4 and collected in the gas-liquid 

separators DM4. In DM4, fitted with the heater ES6; condensed liquid from DM2 and DM3 

systems are also collected in DM4. The solution in DM4 is bicarbonate-lean and is pumped back 

to the absorption section by PC1. 

Separation by absorption produces natural gas containing 97% methane by mole (volume). The 

yield of this method for methane is close to 97%, meaning almost all the methane extracted from 

the Lake is recovered in the natural gas produced by this technique. The amount of methane lost 

with carbon dioxide is very little.  

Results of simulation in Aspen Plus V12 for separation by absorption section are presented in 

Appendix A – Aspen Plus V12 simulation results. Sizing parameters for the absorption flowsheet 

are reported in table 7 below. 

 

 

 



 - 38 - 

Table 7 Sizing of major pieces of process equipment (Appendices A and B) [15] 

Equipment Type, size parameters per unit equipment Quantity 

Feed compressors, 

KC1 

Type:                                                centrifugal motor 

driven 

Total load:                                       7 336 kW 

Available load per unit:                  6 000 kW 

2 

Lean solvent to 

absorption 

columns pumps, 

PC1 

Type:                                                centrifugal motor 

driven 

Total load:                                        523 kW 

Available load per unit:                  174 kW 

3 

Absorption 

columns, VL1 

Type:                                     packed column, stainless 

steel 

Height:                                    6 m 

Diameter:                               3 m 

Packing type: Ceramic (Raschid) ring, 13 mm diameter,  

                                                 6m high 

1 

Stripping columns, 

VL2 

Type:                                     packed column, stainless 

steel 

Height:                                    6 m 

Diameter:                               3 m 

Packing type: Ceramic (Raschid) ring, 13 mm diameter, 

                                                 6m high 

1 

Heat exchangers, 

rich solvents- lean 

solvent, HX1 

Type: shell-and-tube, floating head, square pitch design 

Total duty:                                         130 000 kW 

Heat transfer coefficient:               0.8 kW/m2/℃ (Liquid-

liquid) 

Mean-temperature difference:     5 ℃ 

Total area:                                        34 916 m2 

Available area per unit:                   1 000 m2 

35 

Absorption 

columns overhead 

gas condensers, 

ES1 

Type: condenser, floating head, square pitch design 

Total duty:                                        655 kW 

Heat transfer coefficient:               0.3 kW/m2/℃ 

(gas/water) 

Mean-temperature difference:     8 ℃ 

Total area:                                         293 m2 

Available area per unit:                   293 m2 

1 

Stripping columns 

overhead gas 

condensers, ES2 

Type: condenser, floating head, square pitch design 

Total duty:                                        656 647 kW 

Heat transfer coefficient:               1 kW/m2/℃ (aqueous 

vapor/water) 

Mean-temperature difference:     51 ℃ 

Total area:                                         12 892 m2 

Available area per unit:                  1 000 m2 

13 

 
 



 - 39 - 

Table 8 Sizing of major pieces of process equipment (Appendices A and B) [15] 

Equipment Type, size parameters Quantity 

Stripping 

columns feed 

heaters, ES3 

Type: heater, floating head, square pitch design 

Total duty:                                         249 148 kW 

Heat transfer coefficient:               1 kW/m2/℃ (steam-

aqueous liquid) 

Mean-temperature difference:     29 ℃ 

Total area:                                         8 742 m2 

Available area per unit:                   1 000 m2 

9 

Lean solvent from 

stripping columns 

coolers, ES4 

Type: cooler, floating head, square pitch design 

Total duty:                                         25 908 kW 

Heat transfer coefficient:               0.8 kW/m2/℃ (water-

water) 

Mean-temperature difference:     35 ℃ 

Total area:                                         929 m2 

Available area per unit:                   929 m2 

1 

Stripping 

columns reboilers 

in VL2 

Type: vaporizer, floating head, square pitch design 

Total duty:                                         1 kW 

Heat transfer coefficient:               1.5 kW/m2/℃ (steam-

aqueous solution) 

Mean-temperature difference:     53 ℃ 

Total area:                                         0.013 m2 

Available area per unit:                   9.3 m2 

1 

Mixing tank 

heaters in DM4 

Type: heater, floating head, square pitch design 

Total duty:                                         50 kW 

Heat transfer coefficient:               1 kW/m2/℃ (steam-

aqueous liquid) 

Mean-temperature difference:     100 ℃ 

Total area:                                         0.82 m2 

Available area per unit:                   9.3 m2 

1 

Gas-liquid 

separators, DM2 

Type: horizontal tank 

Minimum diameter required: 1.0 m 

Diameter: 4 m 

Length: 16 m 

1 

Gas-liquid 

separators, DM3 

Type: horizontal tank 

Minimum diameter required: 1.2 m 

Diameter: 4 m 

Length: 16 m 

1 

Mixing tank, 

DM4 

Type: vertical storage tank 

Minimum diameter required: 7.5 m 

Diameter: 8 m 

Length: 16 m 

Volume: 804 m3 

1 

 



 - 40 - 

The duty of the feed compressor KC1 has been calculated using efficiencies as discussed in Smith 

et al [12], it is assumed that isentropic compression takes place with both the isentropic and 

mechanical efficiency set at 70%. A pump is only required in the case of absorption used as a 

technique of separation. The overall efficiency of the centrifugal pump PC1 is assumed as 75% 

(Smith, Van Ness and Abbott) [12].   

The RADFRAC packed absorption model was used in Aspen Plus V12 to estimate the required 

number of packed beds (number of equilibrium stages) for absorption of carbon dioxide into the 

alkaline solution in the first column VL1 and the  stripping of the carbon dioxide from the alkaline 

solution in the second column VL2. 

The duty of heat exchangers HX1, ES1, ES2, ES3, ES4, ES5 and ES6 are also reported in table 7. 

The log-mean temperature difference is then calculated for the counter-current flow arrangement. 

Given the typical ranges for values of the overall heat transfer coefficient reported by Sinnott [11] 

(Coulson and Richardson, volume 6), the area required for heat transfer is then calculated for each 

heat exchanger. 

Horizontal gas-liquid separators DM1, DM2 and DM3 are sized by the Souders-Brown 

method[24]. Values for recommended length to diameter ratios and available diameters are from 

the online cost estimator for horizontal tanks [15],[24]. 

 

 

 

 

 

 

 



 - 41 - 

Table 9 Purchased cost of major pieces of process equipment (Appendices A and B) [15] 

Equipment Type Cost per unit Quantity Subtotal 

Feed 

compressors, 

KC1 

centrifugal motor 

driven, 6 000 kW 

$8 049 592   2 $16 099 184 

Lean solvent to 

absorption 

columns 

pumps, PC1 

centrifugal motor 

driven, 174 kW 

 

$76 968 3 $230 904 

Absorption 

columns, VL1 

Packed column,  

D=3m; H = 6m 

$195 904 1 $195 904 

Ceramic packing 

(porcelain, Raschid 

ring); D=3m; H = 

6m 

$36 298 1 $36 298 

Stripping 

columns, VL2 

Packed column,  

D=3m; H = 6m  

$195 904 1 $195 904 

Ceramic packing 

(porcelain, Raschid 

ring); D=3m; H = 

6m 

$36 298 1 $36 298 

Heat 

exchangers, 

rich solvents- 

lean solvent, 

HX1 

Type: shell-and-

tube, floating head, 

square pitch design, 

1 000 m2 

 

$201 130 35 $7 039 550 

Absorption 

columns 

overhead gas 

condensers, 

ES1 

Type: condenser, 

floating head, square 

pitch design, 293 m2 

 

 $80 470 1 $80 470 

Stripping 

columns 

overhead gas 

condensers, 

ES2 

Type: condenser, 

floating head, square 

pitch design, 1000 

m2 

 

$201 130 13 $2 614 690 

Stripping 

columns feed 

heaters, ES3 

Type: heater, 

floating head, square 

pitch design, 1000 

m2 

$201 130 9 $1 810 170 

Lean solvent 

from stripping 

columns 

coolers, ES4 

Type: cooler, 

floating head, square 

pitch design, 929 m2 

 

$187 254 1 $187 254 



 - 42 - 

 

Table 10 Purchased cost of major pieces of process equipment (Appendices A and B) [15] 

Equipment Type Cost per 

unit 

Quantity Subtotal 

Stripping columns 

reboilers, ES5 

Type: vaporizer, 

floating head, square 

pitch design, 9.3 m2 

 

$11 978 1 $11 978 

Mixing tank 

heaters, ES6 

Type: heater, floating 

head, square pitch 

design, 1000 m2 

 

$11 978 1 $11 978 

Gas-liquid 

separator drums, 

DM2 

Type: horizontal tank 

Diameter: 4 m 

Length: 16 m 

$247 233 1 $247 233 

Gas-liquid 

separator drums, 

DM3 

Type: horizontal tank 

Diameter: 4 m 

Length: 16 m 

$247 233 1 $247 233 

Mixing tank, 

DM4 

Type: vertical tank 

Diameter: 8 m 

Height: 16 m 

Volume: 804 m3 

$194 378 1 $194 378 

Subtotal (2002), CEPCI 390.4 $29 085 490 

Subtotal for Absorption section (2022), CEPCI 877.5 $65 375 301 

Subtotal for Degassing section (2022), CEPCI 877.5 $85 611 707 

Total purchase cost of major equipment $150 987 008 

Total capital investment (Lang Factor = 5.7 for Fluid processing 

plant) 

$860 625 945 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



 - 43 - 

 
 

Table 11 Major annual cost of utilities (Appendices A and B) [15] 

Utilities Equipment & 

consumption 

Annual 

consumption 

Cost per unit Subtotal 

Electricity 

generated from 

local gas 

power plant 

Compressors KC1 

7 336 kW 

58 101 120 kWh 0.095 $/kWh $5 913 112 

Pumps PC1 

523 kW 

4 142 160 kWh 

Medium 

pressure steam,  

4 bar and    

150 ℃, at 

current cost of 

energy from 

methane 

source [20] 

Heaters ES3 

249.148 MW 

7 103 707 776 MJ 0.007 $/MJ 

[20] 

$49 736 133 

Reboilers ES5 

0.001 MW 

28 512 MJ 

Heaters ES6 

0.050 MW 

1 425 600 MJ 

Cooling water, 

mostly river 

water, 

 19℃ [21] 

Condensers ES1, 

0.655 MW 

18 675 360 MJ 0.000264 $/MJ 

[21] 

$5 140 472 

Condensers ES2, 

656.647 MW 

18 722 319 264 MJ 

Coolers ES4, 

25.908 MW 

738 688 896 MJ 

Total $60 789 717 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



 - 44 - 

 Adsorption 
 

KC1

S0

S1

20°C,
100 kPa

187°C, 
500 kPa

ES1

VL1 VL2

S5

S4

S8

S6

S7
S8

S9

S11

S2

S10

V1

V2

V3

FA1

25°C, 
500 kPa

25°C, 
500 kPa

25°C, 
500 kPa

25°C, 
500 kPa

25°C, 
500 kPa

25°C, 
500 kPa

25°C, 
50 kPa

25°C, 
50 kPa

25°C, 
50 kPa

25°C, 
50 kPa

36.0 mol% CH4
63.0 mol% CO2 

1 mol% H2O
5948 kmol/hr

15.0 mol% CH4
85.0 mol% CO2 
4389 kmol/hr

97.0 mol% CH4
2.5 mol% CO2
0.5 mol% H2O 
1504 kmol/hr

DM2

S3

S12

25°C, 
500 kPa

25°C, 
500 kPa

<0.001 mol% CH4
<0.001 mol% CO2

>99.999 mol% H2O 
55 kmol/hr

 
Figure 3 Process flow diagram  

Gas from the degassing drums system DM1 is compressed to 500 kPa by the compression system 

KC1 then cooled in the water cooler ES1 to 25℃. Higher pressures and lower temperatures favor 

faster kinetics of adsorption and higher saturation concentrations of the adsorbed carbon dioxide 

in activated carbon. Condensate at the outlet of ES1 is removed in the set DM2 of gas-liquid 

separation drums. The overhead gas entering the adsorption columns consists of mostly 36 mol% 

CH4 and 64 mol% CO2. 

The overhead gas is thus introduced at 25℃ and 500 kPa in the set VL1 of packed adsorption 

columns that are online. The other set VL2 of packed adsorption columns are offline for 

regeneration of the adsorbent. The control valves V1, V2 and V3 are positioned such that gas the 

paths from DM2 to VL1 and at the outlet of VL1 to V2 are open to allow adsorption of CO2; the 

paths from DM2 to VL2 and at the outlet of VL2 to V2 are closed; the path at the outlet of VL1 to 



 - 45 - 

the blowers systems FA1 is closed while the path at the outlet of VL2 to the set of vacuum pumps 

FA1 is open to allow desorption of CO2 from VL2.  

Adsorption of CO2 happens isothermally and nearly at constant pressure given the size of the 

packed columns that are selected so that pressure drop across the packing is minimized.  At the 

same time some CH4 is adsorbed although only in very small rates since the chosen operating 

conditions allow higher selectivity adsorption of CO2 over CH4 in activated carbon. At the outlet 

of the set of adsorption columns VL1, control valve V2 permits the flow of product containing 98 

mol% CH4 to exit the system. At 500 kPa and 25℃, the saturation concentration of CO2 in 

activated carbon is reported as about 227 mg CO2/g of adsorbent by Lalhmingsanga et al (2014). 

The breakthrough time is the time at which this concentration is reached. At the breakthrough time, 

the purity of methane will begin to drop below 95 mol % at the outlet of VL1. The set VL1 of 

adsorption columns will be switched offline to desorption mode while the set VL2 will be switched 

online to adsorption mode by swinging the positions of control valves V1, V2 and V3.  

Desorption is done via pressure swing where the pressure in the saturated adsorption columns is 

reduced to 50 kPa or below via the suction of the blowers FA1. This low suction pressure created 

by FA1 allows most of the adsorbed CO2 to return to the gas phase and be delivered at 85.0 mol% 

or higher.  

Natural gas recovered via the adsorption technique also contains 97% methane by molar or volume 

fraction. However, adsorption has a low yield since only about 68% of methane in the gas extracted 

from the Lake is recovered in the natural gas produced from the Lake while a significant 32% of 

methane initially is lost in the by-product gas containing mostly carbon dioxide. 

 

 

 

 

 

 

 

 

 

 

 



 - 46 - 

Table 12 Sizing of major pieces of process equipment (Appendices A and B) [15] 

Equipment Type, size parameters Quantity 

Feed compressors, 

KC1 

Type: centrifugal, motor-driven 

Load: 15 774 kW 

Available load per unit: 6 000 kW 

3 

Feed coolers, ES1 Type: cooler, floating head, square pitch design 

Total duty:                                         11 646 kW 

Heat transfer coefficient:               0.3 kW/m2/℃ (gas- 

water) 

Mean-temperature difference:     49 ℃ 

Total area:                                         799 m2 

Available area per unit:                   799 m2 

1 

Gas-liquid separators, 

DM2 

Type: horizontal tank 

Minimum diameter required: 1.5 m 

Diameter: 4 m 

Length: 16 m 

1 

Adsorption/desorption 

columns, VL1 and 

VL2 

Type:                                     packed column 

Height:                                    0.225 m 

Diameter:                               3 m 

Mass of activated carbon:   5.6 ton 

10 

Blowers, FA1 Type: heavy-duty, centrifugal blower 

Volume flow rate: 217 066 m3/hr 

Available volume flow rate: 16 992 m3/hr 

13 

 

For the feed compressors KC1 and the blowers FA1, it is assumed again that an isentropic 

compression takes place with both the isentropic and mechanical efficiency set at 70% as discussed 

in Smith et al [12].  

Properties of beds of activated carbon have been sourced from the work of Hauchhum et al; 

Heuchel et al [25]; Singh et al [27]. The simulation has produced the length of the packed bed of 

activated carbon reactor required to achieve adsorption of CO2 so that the desired flow rate and 

purity of CH4 is achieved at the outlet of the packed bed. This length is then used to estimate the 

mass and the cost of the adsorption beds in VL1 and VL2. 

The log-mean temperature difference for the counter-current flow arrangement and the typical 

overall heat transfer coefficient reported by Sinnott [11] (Coulson and Richardson, volume 6) were 

used to estimate the area required for heat transfer for the feed coolers ES1. 



 - 47 - 

Horizontal gas-liquid separators DM2 were also sized by the Souders-Brown method using the 

online cost estimator [15],[24]. 

Table 13 Purchased cost of major pieces of process equipment (Appendices A and B) [15] 

Equipment Type Cost per 

unit 

Quantity Subtotal 

Feed compressors, 

KC1 

centrifugal motor driven, 6 

000 kW 

$8 049 592   3 $24 148 776 

Feed coolers, ES1 Type: cooler, floating head, 

square pitch design, 799 m2 

$164 452 1 $164 452 

Feed gas liquid-

separator drums, DM2 

Type: horizontal tank 

Diameter: 4 m 

Length: 16 m 

$247 233 1 $247 233 

Adsorption/desorption 

columns, VL1 and 

VL2 

packed column 

 

$16 014.7 10 $160 147 

Blowers, FA1 Type: heavy-duty, 

centrifugal blower 

Available volume flow rate: 

16 992 m3/hr 

$119 305 13 $1 550 965 

Subtotal, CEPCI 390.4 $26 271 573 

Subtotal, CEPCI 877.5 $59 050 475 

Adsorbent (2022)[16], [17] Activated carbon [16], [17] $3 000/ton 5.6 ton $16 800 

Subtotal for Adsorption section (2022), CEPCI 877.5 $59 050 491 

Subtotal for Degassing section (2022), CEPCI 877.5 $85 611 707 

Total purchase cost of major equipment $144 662 198 

Total capital investment (Lang Factor = 5.7 for Fluid processing plant) $824 574 529 
 

Table 14 Major annual cost of utilities (Appendices A and B) [15] 

Utilities Equipment & 

consumption 

Annual 

consumption 

Cost per unit Subtotal 

Electricity 

generated from 

local gas 

power plant 

Compressors 

KC1, 15 774 kW 

 

124 930 080 kWh 0.095 $/kWh $20 630 808 

Vacuum pumps 

FA1 

11 646 kW 

92 236 320 kWh 

Cooling water, 

mostly river 

water, 

 19℃ 

Coolers ES1,  

11.646 MW 

332 050 752 MJ 0.000264 $/MJ  $87 661 

Total $20 718 469 



 - 48 - 

  Cold separation 

 

 

 

KC1

S1

S2

20°C,
127 kPa 200°C, 

750 kPa

DM2

S9

S5

S4 S7

S11

S12

S13

S6

S8

HX1

HX2

180°C,
 750 kPa

180°C,
 100 kPa

-120°C,
750 kPa

-152°C,
20 kPa

-85°C,
750 kPa

-134°C,
 750 kPa

-152°C,
 20 kPa

178°C,
20 kPa

-152°C,
20 kPa

176°C,
20 kPa

180°C,
 750 kPa

ES1
S3

36 mol% CH4
64 mol% CO2
1 mol% H2O 

4510 kmol/hr

98.1 mol% CH4
1.9 mol% CO2
0.0 mol% H2O
1449 kmol/hr

4.3 mol% CH4
95.7 mol% CO2
0.0 mol% H2O
131 kmol/hr

S10

EXP

 
Figure 4 Process flow diagram 

The driving factor for cold separation is the difference in boiling point or in volatility between CO2 

and CH4 at low operating temperature. When a binary mixture of CO2 and CH4 is cooled, CO2 

being heavier than CH4 is expected to condense first while most of the CH4 to remain in the gas 

phase.  

Gas from the overhead drums DM1 is compressed to 400 kPa. The compressed gas leaving the 

drums DM1 travels through a series of heat exchangers HX1, HX2, HX3 and HX4. These are a 

series of heat integration heat exchangers where hot gas leaving the compressor system KC1 at 

130℃ is cooled to -102℃. The gas-liquid mixture leaving the cooling stage at -102℃ is collected 

in the system of separators DM2. Overhead gas from DM2 is further expanded from 400 kPa to a 

100 kPa in the gas turbine system TB1. At the outlet of the turbine, a small portion of CO2 and a 

trace of H2O still in the gas phase is condensed and recovered as liquid in the separators DM3, 

heated up through the heat integration heat exchanger HX4, then the steam heater ES1 before being 



 - 49 - 

recovered at nearly ambient conditions. The gas from DM3 is also heated up in the effluent-feed 

exchanger systems HX2, this gas leaving DM3 is the product CH4 containing more than 98 mol%.   

The condensed liquid from DM2 contains most of the CO2 in the liquid phase, it is vaporized back 

in the effluent-feed exchanger HX3, expanded in the gas turbine systems TB2 and further passed 

in the heat exchanger HX1 to begin cooling the gas entering the compressors system KC1 from 

the cooling train. The product from HX1 is mostly carbon dioxide removed at more than 92 mol%. 

Table 15 Sizing of major pieces of process equipment (Appendices A and B) [15] 

Equipment Type, size parameters per unit equipment Quantity 

Feed compressors, 

KC1 

Type:                                                centrifugal motor 

driven 

Total load:                                        11 788 kW 

Available load per unit:                  6 000 kW 

2 

Heat exchanger, 

HX1 

Type: shell-and-tube, floating head, square pitch design 

Total duty:                                         4 720 kW 

Heat transfer coefficient:               0.01 kW/m2/℃ (gas-gas) 

Mean-temperature difference:     22 ℃ 

Total area:                                         21 455 m2 

Available area per unit:                   1 000 m2 

22 

Heat exchanger, 

HX2 

Type: shell-and-tube, floating head, square pitch design 

Total duty:                                         24 182 kW 

Heat transfer coefficient:               0.01 kW/m2/℃ (gas-gas) 

Mean-temperature difference:     25 ℃ 

Total area:                                         96 728 m2 

Available area per unit:                   1 000 m2 

97 

Cooler, ES1 Type: condenser, floating head, square pitch design 

Total duty:                                         1013 kW 

Heat transfer coefficient:              1 kW/m2/℃ (condensing 

gas - water) 

Mean-temperature difference:     171 ℃ 

Total area:                                          6 m2 

Available area per unit:                   9.3 m2 

1 

Expansion valve, 

EXP 

Type: stainless steel gate valve 

Available diameter: 0.305 m 

1 

Degassing drums, 

DM2 

Type: horizontal tank 

Diameter: 4 m 

Length: 16 m 

Minimum diameter required: 1.7 m 

1 



 - 50 - 

Utilizing cold separation, 88% of methane from the degassing section is recovered in the natural 

gas produced at a purity of 98% by mole or volume. About 12% is lost in the carbon dioxide by-

product stream from the process. 

Table 16 Purchased cost of major pieces of process equipment (Appendices A and B) [15] 

Equipment Type Cost per 

unit 

Quantity Subtotal 

Feed compressors, 

KC1 

centrifugal motor driven, 

6000 kW 

$8 049 592   2 $16 099 184 

Heat exchanger, 

HX1 

exchanger, floating head, 

square pitch design, 1000 

m2 

 

$201 130 22 $4 424 860 

Heat exchanger, 

HX2 

exchanger, floating head, 

square pitch design, 1000 

m2 

 

$201 130 97 $19 509 610 

Cooler, ES1 Type: vaporizer, floating 

head, square pitch design, 

9.3 m2 

 

$11 978 1 $11 978 

Expansion valve, 

EXP 

Type: stainless steel gate 

valve 

Available diameter: 0.305 m 

$16 435        1 $16 435        

Degassing drum, 

DM2 

Type: horizontal tank 

Diameter: 4 m 

Length: 16 m 

$247 233 1 $247 233 

Subtotal (2002), CEPCI 390.4 $40 371 352 

Subtotal for Cold separation (2022), CEPCI 877.5 $90 742 473 

Subtotal for Degassing section (2022), CEPCI 877.5 $85 611 707 

Total purchase cost of major equipment $176,354,180 

Total capital investment (Lang Factor = 5.7 for Fluid processing plant) $1 005 218 826 
 

  



 - 51 - 

 

Table 17 Major annual cost of utilities (Appendices A and B) [15] 

Utilities Equipment & 

consumption 

Annual 

consumption 

Cost per unit Subtotal 

Electricity 

generated from 

local gas 

power plant 

Compressors KC1 

11 788 kW 

93 360 960 kWh 0.095 $/kWh $8 869 291 

Cooling water, 

mostly river 

water, 

 19℃ 

Coolers ES1, 

1.013 MW 

28 882 656 MJ 0.000264 $/MJ $7 625 

Total $8 876 916 
 

 

  



 - 52 - 

  Benefits of the three separation techniques  
 

Two important benefits have been defined as part of initiatives for the degassing of Lake Kivu. 

The first goal to achieve is to extract toxic dissolved gases that accumulate in the deep layers of 

the Lake, hence enable their concentration to remain below levels safe enough to prevent the 

ecological devastation that could result on the surface should a natural volcanic eruption of high 

magnitude occurs in the region. The second goal is to separate methane from carbon dioxide as 

the methane accumulating in the deep layers of the Lake is a useful energy source for the 

generation of electricity than can be sold or exported to both local east African markets 

surrounding the Lake and much larger world markets than the east African great Lakes region. 

 

The first goal ensures environment safety for the region in case of a natural disaster impacting the 

stability of dissolved layers of gases in the Lake and does not necessarily pursue a direct economic 

benefit to result for the capital invested. The second goal builds on the opportunity of supplying 

energy to markets by beneficiation of the natural gas that accumulates through natural processes 

in the Lake. As electricity is generated and sold from this natural gas, financial income then results. 

An analysis is then possible to identify the economic worth of each of the three techniques explored 

in this study. By comparing the income generated by the local sale of electricity produced from 

the methane recovered via each of the three techniques to the respective cost of the capital, the 

three methods can be ranked according to a suitable measure of profitability by this preliminary 

study. This ranking then allows to select the technique with the most favorable measure of financial 

viability for further detailed engineering and appraisal. 

 

The measure of profitability to be used in this study estimate is the internal rate of return or 

discounted rate of return on the cash flow or net income from the generation of electricity from 

methane. The basis for this evaluation is reported in the table. The price of energy from methane 

is taken as 0.095$/ kWh18 as is currently in the local markets surrounding the region (DRC and 

Rwanda). The expected service life of the installation to recover methane is taken as 14 years as 

recommended by Peters and Timmerhaus[14] for the case of a natural gas production site. The 

annual operating time is 330 days. The minimum Wobbe Index of methane is reported as 11 452 

kCal/Nm3 and converted to 13.319 kWh/Nm3 of gas [28].  

 



 - 53 - 

Table 18 Economic parameters for profitability analysis 

Typical service life  

(Natural gas production plant) 

14 years 

Annual operating time 330 days 

Lower Wobbe index of methane 11 452 kCal/Nm3 or 13.319 kWh/ Nm3 

Price of electricity,  

Lake Kivu areas (2022) 

0.095$/ kWh 

Gross unit income  1.30 $/ Nm3 

 

The unit gross income from the sale of electricity is calculated as 1.30 $/Nm3 of methane produced. 

This unit income is common to all three techniques explored in this study.  

   

 Comparison of yield and natural gas product purity  
 

All three techniques of separation have the potential to generate natural gas at more than 75 mol 

% of methane, which is the minimum recommended purity for commercial natural gas.  The 

condition to generate at least 250 million Nm3/year of natural gas is also satisfied by all three 

techniques of separation.  

Table 19 Yield and product purity 

Method of 

separation 

Methane purity 

(mole or volume %) 

Yield (%) 

Absorption 97 97 

Adsorption 97 68 

Cold 

separation 

98 88 

 

The purity and the yield of methane achieved is not identical for all three cases. While cold 

separation allows to achieve the highest purity of 98 mol%, 97 mol % can be attained with both 

absorption and adsorption methods. Absorption in potassium bicarbonate has the highest yield of 

97% followed by cold separation with 88% and adsorption figures last with only 68%.  

 

 

 



 - 54 - 

 Comparison of gross income from sales of electricity  
 

 

 

 

 

 

 

The minimum production requirement was defined as 250 million Nm3/year. This would yield 325 

million $ worth of gross income from sales of electricity at the current price of about 0.095 $/ 

kWh[18]. This requirement is common to all three techniques. 

 

 Comparison based on total capital investment 
 

The techniques are evaluated by comparing their specific cost of capital to purchase the equipment 

and their specific cost of utilities. For each technique the total cost of the equipment includes the 

cost of the degassing section situated upstream of the actual section where separation between 

carbon dioxide and methane is realized. Results obtained are presented in the table below.  

 

Table 21 Total Capital investment (Appendix B) 

Method of separation Total capital investment  

Absorption $860 625 945 

Adsorption $824 574 529 

Cold separation $1 005 218 826 
 

Cold separation is the most expensive technique based on cost of capital needed to purchase the 

equipment.  The cost of heat exchangers required in the heat integration scheme that creates 

cryogenic conditions necessary to achieve the separation of gases is the primary drive of this 

capital cost followed by the cost of compression of the gas  received from the degassing section 

located upstream.  

 

Absorption is ranked second in terms of the cost of the required capital investment. Absorption 

also has a large footprint as it requires many pieces of equipment in order to achieve the desired 

Minimum annual 

methane production rate 

(million Nm3) 

 

Unit gross income 

($/Nm3) 

Annual gross income 

(million $) 

250 1.30 325 

Table 20 Gross annual income 



 - 55 - 

separation of gases. With this technique, compressors needed to increase the pressure of gas to 

higher values which enhance the rate of transfer of carbon dioxide in the alkaline solution of 

potassium carbonate represents the major contributor to the cost of the equipment to purchase.  

Adsorption in activated carbon requires the least total capital investment in order to recover natural 

gas from Lake Kivu. The cost of the compressors that are required to deliver the higher operating 

pressure at which the higher saturation concentration of carbon dioxide in activated carbon can be 

achieved represents the major portion in the capital investment when this technique is employed.  

 

 

 Comparison based on net annual income 

Based on results listed in the table above, absorption is the most expensive to operate, followed by 

adsorption and cold separation. 

Table 22  Annual operating cost (Appendix B) 

Method of separation Major annual cost of utilities  

Absorption $60 789 717 

Adsorption $20 718 469 

Cold separation $8 876 916 
 

The common cost contributor in adsorption and cold separation is that of compression that is 

required at the beginning of the separation steps when these two methods are to be employed. 

Absorption stands out as the most expensive technique due to the cost of medium pressure steam 

required when this method is selected. Medium pressure steam is required to raise the temperature 

of the solution rich in potassium carbonate leaving the absorption step as it enters the stripping 

section. This is the step where carbon dioxide is removed in the gas phase, leaving a solution which 

is lean in carbonate in order that can be recycled to the preceding absorption step.  

 

Table 23 Net annual income (Appendix B) 

Method of separation Net annual income  

Absorption $264 210 283 

Adso