Process Integration and Performance of Chilled Ammonia CO2 Capture Technology Post combustion technology in a novel aluminium manufacturing process Master of Science Thesis in the Master Degree Programme, Sustainable Energy Systems JESPER ARONSSON HENRIC BJÖRK Department of Energy and Environment Division of Energy Technology CHALMERS UNIVERSITY OF TECHNOLOGY Gothenburg, Sweden, 2011 Report No. T2011-355 REPORT NO. T2011-355 Process Integration and Performance of Chilled Ammonia CO2 Capture Technology Post combustion technology in a novel aluminium manufacturing process JESPER ARONSSON HENRIC BJÖRK Supervisor: M.Sc. Henrik Jilvero Examiner: Prof. Filip Johnsson Department of Energy and Environment CHALMERS UNIVERSITY OF TECHNOLOGY Gothenburg, Sweden 2011 Process Integration and Performance of Chilled Ammonia CO2 Capture Technology Post combustion technology in a novel aluminium manufacturing process JESPER ARONSSON HENRIC BJÖRK © JESPER ARONSSON, HENRIC BJÖRK, 2011 Technical report no T2011-355 Department of Energy and Environment Division of Energy Technology CHALMERS UNIVERSITY OF TECHNOLOGY SE-412 96 Gothenburg Sweden Telephone + 46 (0)31-772 1000 Cover: Schematic of the CAP, showing the CO2 capture process in the bottom cycle and the NH3 scrubbing process in the top cycle. Reproservice Gothenburg, Sweden 2011 i Acknowledgements This work has been performed with the much appreciated support and expertise from our supervisor M.Sc. Henrik Jilvero, who in an admirable way always seemed to have some extra time for our questions. Henrik has more than once saved us from falling helplessly from the winding road of despair called Aspen Plus. We would also like to thank our supplementary supervisors Associate Professor Klas Andersson and Ph.D. Fredrik Normann for providing additional guidance during project definition and start- up and in the end of our thesis work for commenting on our report and helping us decide which results to present and how to best present them, without using our information overload graphs. Many thanks also to Professor Filip Johnsson for providing us with this interesting project; we are grateful to have been given the opportunity of working with his team. When on the subject, we also thank the entire team at the Department of Energy and Environment for five inspiring months and we wish all of you the best in your respective work. Big thanks also to Hydro in general and Gunn-Iren Müller in particular for being very supportive throughout our work, for inviting us to tour their state of the art aluminium manufacturing plant and for being very open with process related information vital for the success of our work. Our collaborating partner in Norway, Tel-Tek, also deserves our gratitude for good cooperation, we consider ourselves fortunate to have worked with them and been a part of this project at such an early stage. Lastly, we wish all current and forthcoming participants good luck with the continuation of this project and we hope that at least we will live to see the day when CO2 capture is applied at the Sunndalsøra aluminium manufacturing plant. Sincerely, Jesper Aronsson & Henric Björk ii iii Abstract Global environmental concern with possible near-term political policies to increase cost for greenhouse gas emissions and an inclusion of aluminium production in the European emissions trading system have sparked an interest for CO2 capture within the aluminium industry. The aluminium manufacturing process produces large quantities of CO2 emissions through a non- combustion process, this limits the potential technologies for CO2 capture to post combustion only. A commercially available post combustion technology utilizes a solvent that is degraded in an oxidising environment, as is the case in aluminium manufacturing. The CO2 capture technology investigated in this study uses ammonia as solvent. Ammonia is not degraded by oxygen. The process integration of the chilled ammonia process CO2 capture technology in a novel aluminium manufacturing plant was investigated from a required reboiler duty and available heat from an integration point of view. The process gas leaving the aluminium manufacturing process is characterised by low CO2 concentration and high flow rates compared to applications such as coal-power plants and sour-gas removal from natural gas. The effects on performance and behaviour of the chilled ammonia process for these low CO2 concentrations have not been fully established and have therefore been investigated. The process integration was performed for a case with varying process gas flow rate and CO2 concentration. Gas properties were obtained from Hydro, the company operating the Sunndalsøra aluminium manufacturing plant studied in this report. To isolate and investigate the effects of varying CO2 concentration, a case with constant process gas flow rate was also studied. The CO2 concentration was varied between 4, 7, 10 and 15 vol%. To determine the amount of surplus heat available for process integration, a model of the process gas treatment train was created using the process simulation software Ebsilon Professional. The model consists of the existing process gas treatment train with the addition of two heat extractions for the top and bottom cycle process integration. The chilled ammonia process was described in two separate models in order to determine the required reboiler duties, one for CO2 capture in the bottom cycle and another for ammonia recovery in the top cycle. The simulation software Aspen Plus, handling heat and mass transfer, chemical reactions and complex solutions, was used for modelling. All components, except the top and bottom cycle absorbers, were modelled using an equilibrium calculation approach that assumes infinitely fast mass transfer and reaction kinetics in reactors. The absorbers were modelled using a rate-based approach in which also kinetics is considered, creating limiting factors for the speed and performance of the process. The CO2 capture rate was predetermined to 85% with a maximum ammonia slip of 10 ppmv. Through simulations it was observed that CO2 only has a moderate effect on the bottom cycle specific reboiler duty, in kJ/kg CO2 captured, but that the absolute reboiler duty increases as CO2 concentration and total CO2 content increases. The opposite effect was observed for the top cycle; with the absolute reboiler duty remaining relatively unchanged and the specific reboiler duty decreasing with increasing CO2 concentration. The reboiler temperatures were found to be 111°C for the top cycle and between 194-207°C for the bottom cycle. Through simulations of the process gas treatment train, it was established that surplus heat is available at temperatures sufficient for integration with the chilled ammonia process. It was found that the total heat demand for the chilled ammonia process could be covered to 50% by process integration for the best case scenario. The highest share of heat demand covered through process integration was achieved for the 4 vol% CO2 concentration case. The share then decreases with increasing CO2 concentration to 32% heat demand covered through integration for the 15 vol% case. The study concluded that increasing the CO2 concentration not necessarily will be economically favourable. The lower process gas flow rate associated with the higher CO2 concentration allows for the use of smaller process equipment, lowering investment costs. The lower share of heat demand covered through process integration will however increase operational costs meaning that the most economically feasible CO2 concentration is a trade-off between investment costs and operational costs. Keywords: Process integration, chilled ammonia process, CAP, post combustion, CO2 capture, CCS, carbon dioxide, aqueous ammonia, aluminium, Aspen Plus, process simulation, rate-based iv v Sammanfattning En växande global miljöhänsyn, förestående politiska styrmedel med målsättningen att höja kostnaderna för utsläpp av växthusgaser samt en inkludering av aluminiumproduktionen i det europeiska handelssystemet för utsläppsrätter har lett till ett ökat intresse för CO2-infångning från aluminiumindustrin. Aluminiumproduktion ger upphov till stora mängder CO2-utsläpp genom en process där ingen förbränning förekommer. Detta begränsar de CO2-infångningsmetoder som kan användas till de som avskiljer CO2 efter själva processen. En i dagsläget kommersiell teknologi av denna typ använder ett absorptionsmedel som degraderas i en syrerik miljö, likt den som återfinns i aluminiumproduktion. Den CO2-infångningsprocess som undersökts i denna avhandling använder ammoniak, som inte degraderas vid kontakt med syre, för att separera CO2 från processen. Processintegrering av CO2-infångning med kyld ammoniak som absorptionsmedel, i en innovativ aluminiumproduktionsprocess har undersökts utifrån ett perspektiv där värmebehov och tillgänglig spillvärme betraktats. Processgasen som bildas vid aluminiumproduktion karakteriseras av låg CO2-koncentration och stort gasflöde jämfört med processer som kolkraftanläggningar och naturgasrening. Hur CO2-infångningsprocessen med kyld ammoniak uppför sig vid låg CO2-koncentration är inte helt känt och har därför undersökts. Processintegreringen har utförts för ett fall där gasflöde och CO2-koncentration varierats. Processdata har erhållits från Hydro som driver den studerade aluminiumanläggningen i Sunndalsøra. För att kunna isolera och undersöka effekterna av enbart varierande CO2- koncentration undersöktes ytterligare ett fall där gas flödet behölls konstant. CO2- koncentrationen varierades mellan 4, 7, 10 och 15 vol%. För att bestämma mängden spillvärme som kan utnyttjas för processintegrering skapades en modell av gasreningsanläggningen i programmet Ebsilon Professional. Modellen består av den befintliga gasreningsanläggningen samt två värmeväxlare för processintegrering av CO2-infångningsprocessen. CO2-infångningsprocessen med kyld ammoniak beskrevs genom två separata modeller för att bestämma det totala värmebehovet. En modell där CO2-infångning i bottencykeln beskrevs och en annan modell för att beskriva infångningen av flyktig ammoniak i toppcykeln. Modellerna skapades i Aspen Plus, ett program som behandlar mass- och värmebalans, kemiska reaktioner och komplexa lösningar för att beskriva processen. Samtliga komponenter utom absorptionskolonnerna i topp- och bottencykeln modellerades utifrån ett jämviktsantagande där masstransport antas oändligt snabb och med reaktionsjämvikt i kolonnerna. De båda absorptionskolonnerna modellerades utifrån ett hastighetsbaserat antagande där också processens kinetik beaktas. Avskiljningsgraden av CO2 och de högsta tillåtna ammoniakutsläppen var förutbestämda till 85% respektive 10 ppmv. Genom simuleringar kunde det fastställas att CO2-koncentrationen endast har en måttlig påverkan på det specifika värmebehovet i bottencykeln, angett i kJ/kg infångad CO2, men att det absoluta värmebehovet ökar med ökande CO2-koncentration och total mängd CO2. I toppcykeln observerades den motsatta effekten, det absoluta värmebehovet förblir relativt oförändrad medan det specifika värmebehovet minskar med ökande CO2-koncentration. Temperaturnivåerna som behövs för regenerering fastställdes till 111°C för toppcykeln och mellan 194-207°C för bottencykeln. Genom simuleringar av gasreningsanläggningen konstaterades att spillvärme finns tillgänglig vid tillräckligt höga temperaturer för processintegrering av CO2-infångningsprocessen. Vidare fastställdes det att CO2-infångningsprocessens värmebehov kunde tillgodoses till 50% genom processintegrering av det fall med lägst totalt värmebehov. Den högsta andelen värmebehov som kunde tillgodoses genom processintegrering uppnåddes för en CO2- koncentration på 4 vol%. Andelen av tillgodosett värmebehov minskar därefter med ökande CO2-koncentration till 32% för en CO2-koncentration på 15 vol%. En av avhandlingens slutsatser var att en ökad CO2-koncentration inte nödvändigtvis är ekonomiskt fördelaktigt. Det lägre gasflödet förknippat med högre CO2-koncentration gör det möjligt att använda mindre CO2-infångningsutrustning vilket ger lägre investeringskostnader. Den lägre andelen av tillgodosett värmebehov genom processintegrering ger dock upphov till ökade driftskostnader. Detta betyder att den CO2-koncentration som är fördelaktig utifrån ett ekonomiskt perspektiv blir en avvägning mellan investeringskostnad och driftskostnad. vi vii Contents Acknowledgements ..................................................................................................................................... i Abstract ......................................................................................................................................................iii Sammanfattning ......................................................................................................................................... v Abbreviations and Symbols ..................................................................................................................... ix List of Figures ............................................................................................................................................ xi List of Tables ........................................................................................................................................... xiii 1. Introduction ............................................................................................................................................ 1 1.1. Aim and Scope ................................................................................................................................ 2 2. Hydro Sunndalsøra ................................................................................................................................ 3 2.1. Process Gas Treatment .................................................................................................................. 5 3. The Chilled Ammonia Process ............................................................................................................. 7 3.1. CO2 Capture Cycle ......................................................................................................................... 8 3.2. NH3 Scrubbing Cycle ................................................................................................................... 10 4. Method .................................................................................................................................................. 11 4.1. Process Data ................................................................................................................................. 11 4.2. Modelling Tools ............................................................................................................................ 12 4.3. Process Integration ....................................................................................................................... 13 5. Modelling .............................................................................................................................................. 15 5.1. Process Gas Treatment Model ................................................................................................... 15 5.2. CAP CO2 Capture Model ............................................................................................................ 18 5.2.1. Input Data ............................................................................................................................... 18 5.2.2. Absorber Modelling ............................................................................................................... 19 5.2.3. CO2 Capture Cycle ................................................................................................................. 20 5.2.4. NH3 Scrubbing Cycle ............................................................................................................. 23 5.2.5. Auxiliary Equipment ............................................................................................................. 25 6. Results ................................................................................................................................................... 27 6.1. CO2 Capture Cycle ....................................................................................................................... 27 6.2. Process Integration of the CAP .................................................................................................. 34 7. Conclusions ........................................................................................................................................... 37 8. References ............................................................................................................................................ 39 Appendix A – CO2 Concentration Properties Extrapolation ............................................................... I Appendix B – Dimensioning of Packed Columns ............................................................................... III viii ix Abbreviations and Symbols Abbreviations CCS Carbon Capture and Storage IEA The International Energy Agency EU The European Union EU ETS European Union Emissions Trading System MEA Monoethanolamine CAP The Chilled Ammonia Process PM Particulate Matter VLE Vapour-Liquid Equilibrium Symbols Greek Symbols Symbol Meaning Unit �� Gas phase film thickness m �� Liquid phase film thickness m Latin Symbols �� Film concentration of species A mol/m3 ��� Bulk concentration of species A mol/m3 ��� Bulk concentration of species B mol/m3 ∆ � Heat of reaction J/kg ∆ � ���� � Heat of solution J/kg �� Film partial pressure of species A Pa ��� Bulk partial pressure of species A Pa � Coordinate axis z - x xi List of Figures Figure 1. Schematic overview of a post combustion CO2 capture process integrated with an aluminium production process. The treated process gas enters the post combustion process where it is separated and sent for storage. .......................................................................................................... 1 Figure 2. Cross sectional view of the current generation of cells using the Hall-Héroult process for aluminium production. The product gas mixes fully with the cooling air resulting in a process gas with low CO2 concentration (0.9-1 vol%). Source: Hydro .................................................................... 3 Figure 3. Cross sectional view of a HAL Ultra cell with separate collector for product gas, lowering the volumetric flow of total process gas and increasing the CO2 concentration up to 4 vol%. Source: Hydro ................................................................................................................................. 4 Figure 4. Process gas treatment flow sheet at section 3 of the SU4. The process gas (yellow) is cleaned and scrubbed free from various contaminants throughout the gas treatment train. A heat exchanger supplies the local district heating network with hot water. ................................................. 5 Figure 5. Flowchart describing the CAP. CO2 is absorbed and stripped in the bottom cycle leaving pure CO2 and ammonia contaminated off-gas as the two outgoing streams. Off-gas is treated in the top cycle where ammonia is absorbed, stripped reintroduced to the CO2 absorption process. (15) ............................................................................................................................................................... 7 Figure 6. Bubble point pressures of a 28 wt% ammonia solvent with a temperature of 8°C as a function of the CO2 loading for equilibrium simulations. (16) ............................................................. 9 Figure 7. Schematic overview of the aluminium manufacturing process and proposed heat source for process integration of the CAP. ....................................................................................................... 13 Figure 8. Describes the model of the process gas treatment train. Grey boxes marks the different included sections. The green dashed line indicates the presently installed equipment at the existing SU4, section 3 ............................................................................................................................. 15 Figure 9. Possible heat extraction for process integration of the CAP is modelled by a heat consumer connected to a heat exchanger. Both the bottom and top cycle process integration heat exchangers are modelled accordingly. ................................................................................................... 16 Figure 10. The dry scrubber is modelled by a selective splitter and a heat extraction block to simulate the ash and alumina separation and temperature drop of the actual process. ................... 16 Figure 11. A liquid-gas heat exchanger and a heat consumer are used to model the district heating heat exchanger. ...................................................................................Fel! Bokmärket är inte definierat. Figure 12. The wet scrubber is modelled in four steps, saturation, heat extraction, SO2 removal and excess sea water drainage. ............................................................................................................... 17 Figure 13. Concentration profiles depending on chosen film theory. ................................................ 19 Figure 14. Illustrative process flowchart describing the CO2 capture cycle model. .......................... 20 Figure 15. The packed CO2 absorber column is modelled using Aspen RadFrac block. ................ 20 Figure 16. The packed CO2 regenerator column is modelled using Aspen RadFrac block. ........... 22 Figure 17. Illustrative process flow chart describing the NH3 scrubbing cycle model. ..................... 23 Figure 18. The packed NH3 absorber column is modelled using Aspen RadFrac block. ................ 23 Figure 19. The packed NH3 regenerator column is modelled using Aspen RadFrac block. ........... 24 xii Figure 20. Reboiler duty per unit mass of captured CO2 shown for different lean stream loadings and grouped by total solvent flow. The data shown is for the 7% CO2 Hydro case. ........................ 27 Figure 21. Reboiler duty averaged for different solvent flows, grouped by absorber CO2 loading, Hydro case. ............................................................................................................................................... 28 Figure 22. Reboiler duty averaged for different solvent flows, grouped by absorber CO2 loading, constant flow case. ................................................................................................................................... 28 Figure 23. Mole fraction ammonia in the stream leaving the bottom cycle regenerator, Hydro case. ........................................................................................................................................................... 29 Figure 24. Ammonia slip from the absorber, Hydro case. ................................................................... 29 Figure 25. Ammonia slip from the absorber, constant flow case. ....................................................... 30 Figure 26. Ammonia slip from absorber with respect to lean stream CO2 loading, Hydro case. .... 30 Figure 27. Top cycle reboiler duty and specific reboiler duty. Specific reboiler duty relates the reboiler duty to the amount of CO2 captured in the bottom cycle. Some values are missing due to non-convergent simulations. ................................................................................................................... 31 Figure 28. Heat of evaporation for water in the regenerator related to the reboiler duty. Some values are missing due to non-convergent simulations. ....................................................................... 32 Figure 29. Bottom cycle cooling need, Hydro case. ............................................................................. 33 Figure 30. Top cycle cooling need, Hydro case. Some values are missing due to non-convergent simulations. ............................................................................................................................................... 33 Figure 31. The required reboiler duty for different NH3 molality. A comparison between results from this study and studies made by Jilvero et al. (30) and Darde et al (16). ................................... 34 Figure 32. District heating, low and high temperature heat extraction available for process integration given different CO2 concentrations ranging from 4 to 15 vol%. ..................................... 35 Figure 33. Top and bottom cycle heat requirements compared to the available heat from process integration. ................................................................................................................................................ 35 Figure 34. Extrapolated values for temperatures and flow rates for 10 and 30% CO2 concentrations. ........................................................................................................................................... I Figure 35. Describing the general design and layout of a packed column. The diffusion occurs at the phase interface between liquid and vapour in the packed bed. Source: Grén et al. (28) ......... III Figure 36. Packed column flow rate correlation presented in a log-log graph. The ordinate (Ѳ) and abscissa (FLV) are flow rate parameters used to determine the vapour flow velocity and subsequently the column diameter. Source: Grén et al. (28) ............................................................. IV xiii List of Tables Table 1. Measured and estimated emission data before gas treatment for a HAL Ultra cell (4). .... 5 Table 2. Variables and values that make up the parametric study for the CO2 capture process .... 11 Table 3. The properties for process gas at various CO2 concentrations for the Hydro case. Temperature and pressure is 5°C and atmospheric respectively for each case when entering the CO2 capture process................................................................................................................................. 12 Table 4. The properties for process gas at various CO2 concentrations, adjusted to constant flow rate. Temperature and pressure is 5°C and atmospheric respectively for each case when entering the CO2 capture process. ......................................................................................................................... 12 Table 5. Properties and composition of the process gas leaving the HAL Ultra potline for the 4 vol% CO2 concentration case. ................................................................................................................ 15 Table 6. The properties of the process gas entering and exiting the CAP integration heat exchangers. The table presents the results for the 4 vol% CO2 concentration case. ........................ 16 Table 7. The properties of the process gas entering the process gas cooler and the properties and composition of the process as leaving the cooler. ................................................................................. 18 Table 8. The properties for process gas at various CO2 concentrations leaving the aluminium plant gas treatment train. Temperature and pressure is 5°C and atmospheric respectively for each case. .................................................................................................................................................................... 18 Table 9. Simulation technical design specifications and properties of the CO2 absorber and its connecting streams for a 4 vol% CO2 concentration case with 6.59 mol NH3/kg H2O. Numbers in brackets indicate parameters chosen for model stability and convergence and should not be used as actual design parameters. ................................................................................................................... 21 Table 10. Simulation technical design specifications and properties of the CO2 regenerator for a 4 vol% CO2 concentration case with 6.59 mol NH3/kg H2O. Numbers in brackets indicate parameters chosen for model stability and convergence and should not be used as actual design parameters. ............................................................................................................................................... 22 Table 11. Simulation technical design specifications and properties of the CO2 absorber and its connecting streams for a 4 vol% CO2 concentration case with 6.59 mol NH3/kg H2O. Numbers in brackets indicate parameters chosen for model stability and convergence and should not be used as actual design parameters. ................................................................................................................... 24 Table 12. Simulation technical design specifications and properties of the NH3 regenerator for a 4 vol% CO2 concentration case with 6.59 mol NH3/kg H2O. Numbers in brackets indicate parameters chosen for model stability and convergence and should not be used as actual design parameters. ............................................................................................................................................... 25 Table 13. Top and bottom cycle reboiler temperature for different CO2 concentrations. .............. 34 Table 14. Additional heat requirements and the share of required heat that is supplied by process integration. ................................................................................................................................................ 36 Table 15. Specified and calculated values for the bottom and top cycle absorbers using metal pall ring packing material. ............................................................................................................................... V xiv 1 1. Introduction Global environmental concern with possible near-term political policies to increase cost for greenhouse gas emissions have made industrial applications for carbon capture and storage (CCS) interesting. The aluminium manufacturing industry utilizes a highly energy intensive process with direct emissions of CO2. The international energy agency (IEA) estimates that for every tonne of aluminium produced, 1.52 tonnes of CO2 is released to the atmosphere (1). A directive (2) adopted by the European Union (EU) in April 2009 included aluminium production from alumina (primary) and aluminium production from recycled products (secondary) in the European Union Emissions Trading System (EU ETS) for the period 2013- 2020 (3). During this period, the aluminium industry will receive a high share of needed allowances without extra cost in order to keep European aluminium production competitive on the global market. For the period following 2020, the emission targets will most likely be more stringent, resulting in higher costs coupled to direct greenhouse gas emissions from aluminium production. The inclusion of direct emissions from the aluminium industry into the EU ETS will with present technology result in higher production costs, giving incentives for investigation of CCS integration into the production process. Present research about CCS is focused to a few technologies for CO2 capture. The most prominent technologies are pre combustion, oxy-fuel combustion and post combustion. Post combustion, which is the technology in focus for this project, refers to CO2 capture downstream of the main process (combustion, industrial manufacturing etc.), illustrated in Figure 1, making it suitable for retrofits at existing installations. In post combustion CO2 capture technology, CO2 is captured and separated from the flue gas stream by a solvent in an absorption reactor. The captured CO2 is then transported to a regeneration reactor where CO2 is released through heat addition, breaking the bonds formed in the absorber. The solvent is then recirculated back to the absorber for further CO2 capture. Figure 1. Schematic overview of a post combustion CO2 capture process integrated with an aluminium production process. The treated process gas enters the post combustion process where it is separated and sent for storage. There are several solvents suitable for CO2 capture available, all with various advantages and disadvantages depending on the application considered. Two technologies considered when integrating CO2 capture with aluminium production are the monoethanolamine process (MEA) and the chilled ammonia process (CAP) (4). The selective removal of CO2 at small and medium scale has been a commercial process for many years, for example for producing carbonated gas to be used in beverages. However, when considering CO2 capture from power plants or other industrial processes, like the aluminium manufacturing process considered here, there are only few large scale commercial examples. (5) Gas cleaning with MEA has been used in different industries since the 1930´s (6), for example removal of acid gases like CO2 from natural gas streams. MEA is an organic substance used because of its ability to form strong bonds with CO2. This allows for low partial pressures of CO2 in the flue gas. That is usually the case in power plants and, even more so, in the low CO2 2 concentration aluminium manufacturing process (7). The CAP and MEA systems are rather similar in function and structure; both make use of nitrogen based compounds for absorption; both use an absorber, a stripper and heating and cooling systems. There are however some significant differences. The use of amines for CO2 capture is a widely known technology. However, there are some inherent properties that make the use of MEA in certain processes unfavourable. Some of these properties are: high energy consumption for regeneration, degradation in an oxidising environment and corrosion of equipment due to salt formation (8). The flue gas entering the absorber in the MEA process must first be treated in order to remove most of the SO2 and NO2. They otherwise react and form heat stable salts with MEA. The heat stable salts are not dissolved in the regenerator, resulting in lower solvent absorption capacity and decreased performance for the process (8). Salts are formed also in the CAP. These salts however, dissolve when heated and further affects the process by increasing the CO2 absorption capacity of the solvent, decreasing the required solvent flow rate. The pressure in the regenerator, and consequently that of the exiting CO2 gas, is for the CAP about 30 bar. The high pressure results in lower energy requirement for compression compared to the MEA process in which the regenerator pressure is only slightly above atmospheric (1). Arguably, the most important difference between the two solvents, considering the studied application, is the solvent degradation resistance. MEA is mainly degraded by carbamate polymerisation; thermal degradation above 205°C and oxidative degradation, of which the degradation by an oxidizing environment is the most prominent one considering the studied application (9). In a combustion process oxygen reacts with fuel in an exothermic reaction, producing heat and combustion products. In a power production process a fairly complete combustion can be achieved, resulting in a low concentration of oxygen in the flue gas. In the aluminium production process, air is used for cooling purposes and thus, no combustion takes place. Hence the process gas leaving the aluminium process has, except for higher CO2 concentration and contaminants from the process, roughly the same composition as air and still contains roughly 21 vol% oxygen. Using MEA in such an oxidizing environment could result in severe degradation. Ammonia however, does not react with oxygen alone and is not degraded in an oxidizing environment like the one present in the aluminium production process. The CAP is therefore believed to have good potential when applied for CO2 capture from the aluminium manufacturing process. 1.1. Aim and Scope The overall aim of this work is to investigate how the performance of the CAP is affected by a low CO2 concentration source and to design and integrate such a process for use at a novel aluminium manufacturing plant. The aim is to investigate different operating conditions affiliated with the HAL Ultra aluminium production process, and the specific effects of varying process gas CO2 concentration, solvent flows and changes in solvent ammonia concentration from a heat duty requirement and ammonia emission restriction point of view. This work focuses on the CAP for capture of CO2 from the process gas produced in the aluminium manufacturing process. Many studies have been made regarding the CAP, including some with models and simulations similar in design to models in this study. However, most studies have been made with power production, combustion processes in mind. For the vast majority of coal, biomass and petroleum combustion processes, CO2 constitutes about 20 wt% of the total flue gas flow. The CO2 concentration expected from state-of-the-art aluminium cells is between 4- 10 vol%, lower than the 20 wt% of the described combustion processes. It is therefore of interest to investigate how the required heat duty for the CAP is affected by the lower CO2 concentration from the aluminium cells. Furthermore, post combustion CO2 capture applications for industrial processes are not presently conventional and studies of such possible applications for the primary aluminium process are scarce. This study investigates if, where and how surplus heat is available for integration with the aluminium production. 3 2. Hydro Sunndalsøra Primary aluminium is mainly produced in electrolytic cells through the Hall-Héroult process. In this process, alumina (Al2O3) is dissolved in a cryolitic solution (Na3AlF6). The base of the cell, in which the solution is contained, is insulated with refractory lining and floored with carbon blocks which act as process cathodes. To complete the electric circuit, carbon anodes are submerged in the solution. When an electric current is passed through the solution, via the carbon cathode and the carbon anode, alumina is reduced into aluminium. The carbon anode is consumed by the oxygen produced, which gives rise to emissions of CO2. Pure aluminium is deposited at the carbon cathode at the bottom of the cell where it is tapped off and cast. The Hall-Héroult process is continuous and alumina and pure aluminium are continuously added and removed at regular intervals. The melting point of the cryolitic solution, and therefore the temperature needed for the process, is close to 1 000°C. To keep the temperature from rising and cause structural damage to the cell, the process is cooled by an air flow above the cryolitic solution. Process gas is formed when cooling air mixes with CO2, SO2 and other compounds released from the cell. The process gas is transported from the cell for removal of SO2, particulate matter (PM), hydrogen fluorides (HF) and other pollutants before being released to the atmosphere (10). A development of aluminium cell technology is needed to facilitate the integration of post combustion CO2 capture with the aluminium production process. Aluminium smelting cells are generally using the Hall-Héroult process illustrated in Figure 2. In this process, CO2 formed at the carbon anode is mixed fully with the cooling air. Due to the large volumetric air flows needed for cooling, the process gas leaving the cell contains a low concentration of CO2, in the range of 0.9-1 vol%. Figure 2. Cross sectional view of the current generation of cells using the Hall-Héroult process for aluminium production. The product gas mixes fully with the cooling air resulting in a process gas with low CO2 concentration (0.9-1 vol%). Source: Hydro Low CO2 concentration and high volumetric gas flow rate requires large CO2 capture and gas scrubbing equipment to achieve desired CO2 capture rates. Large additional equipment means added investment cost and higher cost of operation. Additional costs that are not directly adding value to the final aluminium product results in an economically unfavourable investment with present costs associated with CO2 emissions. (11) To reduce cost and facilitate integration of CO2 capture, Hydro’s HAL Ultra cell, shown in Figure 3, is developed in such a way that most of the CO2 is separated from the rest of the cooling air stream at the point where the CO2 is released in the cell. This is achieved by introducing a new process gas collector in the cell and by reducing process gas flow rate, so that most gas released from the cryolitic solution can be collected and transported away for treatment. The novel design increases the CO2 concentration to about 4 vol%. The higher CO2 concentration and lower flow rate is believed to result in lower cost of both investment and operation, which enhances the possibility for introduction of CO2 capture in the primary aluminium manufacturing process (4). 4 Figure 3. Cross sectional view of a HAL Ultra cell with separate collector for product gas, lowering the volumetric flow of total process gas and increasing the CO2 concentration up to 4 vol%. Source: Hydro The studied primary aluminium production facility is located in western Norway, in Møre og Romsdal County. The facility’s location at the end of a fjord ensures reliable transportation routes by shipping and a large supply of cold sea water for process cooling. After the completion of a major modernization in 2004, the Sunndalsøra aluminium manufacturing facility is considered to be one of Europe’s most modern plants with an annual production capacity of 390 000 metric tonnes (3). The modernization included construction of a new line of aluminium cells called SU4, divided into three sections, in order to replace two lines of older cells. Section 1 of SU4 consists of 108 cells and the other two, section 2 and 3, each contain 116 cells. The 340 cells, which together make up SU4, all use the Hall-Héroult process. The yearly total primary aluminium production capacity of SU4 is approximately 276 000 metric tonnes, more than 70% of the total production capacity of the Sunndalsøra facility. For this project, section 3 of the SU4 is chosen as reference when considering heat extraction for process integration and process gas stream data. The pressure in the cells is kept at sub-atmospheric conditions in order to ensure that no process gas leaks from the cells into the atmosphere without gas treatment. The flow of cooling air and process gas from the cells during normal operation is 6 500 Nm3/h. In order to keep the pressure difference at 0.5 mm H2O even when cells are opened for anode exchange, additional suction is required. The anode exchange is limited to a few cells at a time with an increased air flow of 18 000 Nm3/h for those cells (4). The process gas stream leaves the cell and is merged with the gas streams from the other cells in section 3. A potline with HAL Ultra cells has reduced cooling air flow rate in order to increase the concentration of CO2 in the process gas. According to Hydro, achieving 4 vol% CO2 concentration requires a reduction of the cooling air flow to 1 600 Nm 3/h for normal operation but remain at 18 000 Nm3/h during anode exchange to ensure sufficient underpressure (4). The total process gas and cooling air flow rate using the HAL Ultra cell technology at section 3 is then: �112 ∙ 1 600 � 4 ∙ 18 000� 3 600⁄ " 69.78 &'() *⁄ + �1.1� The total process gas flow is reduced by a factor three compared to present operation. Emissions from a HAL Ultra cell with various cooling gas flow rates have been performed and measured or estimated by Hydro and the results are presented in Chapter 2.1. 5 Table 1. Measured and estimated emission data before gas treatment for a HAL Ultra cell (4). Flow rate [Nm3/h] Temperature [°C] Concentration of gas [mg/ Nm3] Concentration of gas [vol%] Concentration of total dust [mg/ Nm3] HF SO2 CO2 7 000 130 155 235 0.6 500 4 000 185 235 411 0.9 500 2 000 265 400 822 3.0 500 1 600 290 500 1 027 4.0 500 1 000 345 780 1 800 7.0 500 800 365 1 060 2 055 10.0 500 Table 1 shows that in order to reach the desired 4 vol% of CO2 in the product gas, the total flow rate need to be reduced to 1 600 Nm3/h. This will in turn affect the composition and properties of the process gas, resulting in higher temperatures and increased concentration of e.g. HF and SO2 in comparison to present operation. 2.1. Process Gas Treatment Due to contamination of mainly SO2, PM and HF, the process gas needs treatment before released to atmosphere. The process gas treatment train at SU4 is shown in Figure 4 and consists of several stages to ensure sufficient removal of contaminants. Figure 4. Process gas treatment flow sheet at section 3 of the SU4. The process gas (yellow) is cleaned and scrubbed free from various contaminants throughout the gas treatment train. A heat exchanger supplies the local district heating network with hot water. When leaving the cells, the process gas stream is mixed with the ingoing alumina (Al2O3), which is then carried with the gas to the dry cleaning plant. Whilst mixed, the fluoride in the process gas will react with the alumina, forming fluoride enriched alumina. The fluoride injection reduces the amount of HF and other fluoride compounds released to the atmosphere. Fluoride is one of the components that make up the cryolitic solution in the cells, meaning that the recirculation of fluoride through alumina reduces the need of cryolite makeup. The dry scrubber consists of a cyclone and baghouse filtration system and is used to remove PM and to separate the fluoride 6 enriched alumina from the process gas stream. The process gas continues through the scrubber whilst the heavier alumina particles are collected at the bottom and distributed to the cell. The section 3 process gas treatment train is connected to the Sunndalsøra district heating network by a heat exchanger. The heat exchanger is located between the dry and wet scrubbing plants, using heat that would otherwise be wasted. At present, the temperature of the process gas is about 100°C when entering and between 70-80°C when leaving the heat exchanger. The district heating water is heated from 60°C to 90°C in the heat exchanger and is then distributed to the Sunndalsøra district heating customers. (4) Cold sea water is introduced at the top of the wet scrubber, countercurrent to the process gas. The main task for the wet scrubber is to remove SO2 from the process gas. The wet scrubber also reduces the HF emissions but due to the high efficiency of the upstream dry scrubber, most HF is already removed (12). The sea water is discharged back into the sea after the wet scrubber at a slightly elevated temperature. The cleaned process gas is released to the atmosphere (4). 7 3. The Chilled Ammonia Process CO2 capture by the use of an aqueous ammonia solvent is a promising technology. Some expected advantages are: a comparatively inexpensive solvent; regeneration at pressure, minimizing solvent vaporization and pre-compressing the CO2 for transport; a low heat of reaction and low regeneration energy consumption (13) (14). However, ammonia vaporization may be so great that a separate NH3 capture cycle may be needed (15) and the need for cooling water is extensive. A schematic of the CAP is given in Figure 5. Treated and cooled process gas containing CO2 is passed through an absorption column where CO2 reacts with the chilled aqueous ammonia solvent in which the CO2 is absorbed. Typically 85-90% of the CO2 is absorbed before discharge to the atmosphere. The solvent, now containing CO2, is transported from the absorption column to a regenerator where the CO2 is released through heat addition, breaking the chemical bonds. The released CO2 stream, also containing water and residual ammonia, is first cleaned from water and ammonia then pressurized and sent for storage. The solvent is recycled back to the absorption column for further CO2 absorption. Figure 5. Flowchart describing the CAP. CO2 is absorbed and stripped in the bottom cycle leaving pure CO2 and ammonia contaminated off-gas as the two outgoing streams. Off-gas is treated in the top cycle where ammonia is absorbed, stripped reintroduced to the CO2 absorption process. (15) 8 3.1. CO2 Capture Cycle The absorber is of cross-flow type with the cooled CO2-lean solution, coming from the regenerator, entering from the top and the cooled flue gas entering from the bottom at atmospheric pressure (16). In their response to critique by IEA GHG (1), Alstom claim to use a packed column of proprietary design, capable of handling solids. They also state that heat exchangers are used to remove heat from the exothermic absorption process. Alstom claim that the CAP is tolerant enough towards SO2 to be fitted directly after conventional SO2-cleaning systems and that it can capture residual SO2, SO3, HCl and PM (1). According Gal (17), the optimal CO2/NH3 mole ratio (CO2 loading) in the absorber should be 0.33-0.67 but values as low as 0.25 or as high as 1 are possible. According to Darde et al. the rich CO2 loading should be in the range of 0.5-1 (16). Ratios in the higher end of the spectrum results in: a reduction of gaseous ammonia being released from the solution as a result of lower ammonia vapour pressure, an increase of undissolved solids and an increase in CO2 vapour pressure, reducing capture efficiency (13). Ratios in the lower end of the spectrum have the opposite effect. Increasing temperature has great effect on the system: it gives an increased solubility of solids, it increases ammonia vapour pressure which increases ammonia vaporization and it increases CO2 vapour pressure which reduces capture efficiency (13). For the CAP, IEA (1) state the following main chemical reactions and standard heats of formation in the liquid phase: �,-�.� � -,�/0� � ' )�/0� ↔ �,)2 � ' 34 ∆ � " 63.7 56/(89 �1.2� �,-�.� � 2' )�/0� ↔ ' -�,-2 �' 34 ∆ � " 74.1 56/(89 �1.3� �,-�.� � �,)-2�/0� � -,�/0� ↔ 2 �,)2 ∆ � " 26.4 56/(89 �1.4� ' ) � 4 ↔ ' 34 �1.5� �,)-2 � 4 ↔ �,)2 �1.6� And the following solid formation reactions and heat needed for dissolution: ' 34 � �,)2 ↔ ' 3 �,)�*� ∆ � ���� � " 28.4 56/(89 �1.7� ' 34 �' -�,-2 ↔ ' -�,-' 3�*� ∆ � ���� � " 15.9 56/(89 �1.8� 2' 34 � �,)-2 � -, ↔ �' 3�-�,) ∙ -,�*� �1.9� 4' 34 � �,)-2 � 2 �,)2 ↔ �' 3�-�,) ∙ 2' 3 �,)�*� �1.10� Thus, the favoured reaction would be the formation of bicarbonate ions, a reaction which is prevalent below 15°C (1). At higher CO2-loadings the solubility limit is reached and precipitation occurs. This implies that theCO2-rich stream out of the absorber could be in slurry form (16). The stream flows through a high pressure pump, compressing it to 3.45-172.4 bar in order to avoid unwanted CO2 vaporization (17). It should be noted that solids formation increases the total amount of heat needed for regeneration and that it demands use of equipment capable of handling solids (1). The solids formation however, allows for higher CO2 loadings, resulting in lower flow rates. The cold CO2-rich stream is fed to the regenerator system where it is heated up by an auxiliary heater or heat exchanger to 50-200°C, with a preferable range of 100-150°C, and then fed into the regenerator column (16) (17). Due to the strong temperature dependence of the solubility of the solids (13), they dissolve when the slurry is heated enough (above 80°C according to Kozak et al. (14)). The regenerator consists of a stripper column to which the rich, heated solution is fed at several different levels, according to Alstom in IEA GHG (1). The pressure in the column is elevated in order to demote water and ammonia vaporization while the solubility of CO2 in water is greatly decreased. This allows for a more pure CO2 stream to exit at the top of the column (16) (13). Alstom (1) state that they use a pressure of 30 bar at the bottom of the column. The reactions taking place are reversed compared to those presented as absorption reactions above. Because of water and ammonia contamination of the CO2 stream leaving the system, a cleaning system consisting of a wet scrubber is installed after the regenerator (16). 9 The stream of CO2-lean solvent is cooled and reintroduced into the absorber (17) at a lean CO2 loading of 0.25-0.67 (16). This range is suggested since CO2 loadings below approximately 0.54 increases the ammonia bubble point pressure drastically, seen in Figure 6, thus increasing the amount of ammonia in the gas phase, contaminating the stream leaving the absorber (16). Figure 6. Bubble point pressures of a 28 wt% ammonia solvent with a temperature of 8°C as a function of the CO2 loading for equilibrium simulations. (16) Because of formation of solids in the absorber there is need for a pump capable of handling slurry. The type of pump usually used for this sort of application is a centrifugal pump with open impellers and large clearances (18). When handling slurry, lower pump efficiency and lower developed head can be expected compared to pumping water alone (19). Transporting slurry gives rise to wear on the equipment and may also promote corrosion attacks associated with erosion in chemical applications (19). The heat exchanger transferring heat between the CO2-rich stream coming from the absorber and the CO2-lean stream coming from the regenerator needs to be able to handle slurry. Moretta (20) says that: “Spiral plate heat exchangers are ideal for cooling slurries and viscous fluids”. This is justified by that they are compact, have a decreased tendency of fouling, are easy to maintain and have increased heat transfer rates compared to other alternatives, opinions which are seconded by the heat exchanger manufacturing company Alfa Laval (21). Lee et al. (22) however use a tube- in-tube arrangement in their research, a setup which would allow for the use of a scraper if necessary to avoid fouling (18). Plate heat exchangers could be an alternative, as claimed in an advertisement in Process Heating (23). However, Couper et al. (18) claim that these are not suitable for use at pressures above 300 psig (20.68 bar). There is also a possibility for adding a heat exchanger in series with the reboiler in the CO2 regenerator. This would mean heat exchanging between the solution in the regenerator and with a gas stream coming directly from the potline. According to Hydro (4), this gas is contaminated, leading to fouling of the heat exchanger. Thus, the same kinds of heat exchangers could be used as the alternatives proposed for the rich/lean stream heat exchangers. The CAP has need for cooling of several streams. The process gas feed, the rich recirculation stream and the lean return stream have to be cooled and there is also be need for intercooling in the absorber itself. The process gas stream is washed from SO2 in a wet cleaning device which resembles a direct contact cooler, partially cooling it down. Because the preferred inlet temperature to the absorber is about 0-10°C, a supplementary cooler might be required to get sufficient cooling during the warmer parts of the. Examples of this sort of device are given by Couper et al. (18), as either absorption or compression refrigeration. Both of these types of 10 refrigeration can attain temperatures below 0°C. The two liquid streams in need of cooling can use the same kind of refrigeration cycles if needed, but should preferably be cooled by heat exchange with sea water. Hydro estimates that sea water will be available at a temperature of maximum 10°C all year round. In order to increase the solids content in the slurry, IEA (1) states that Alstom uses a hydrocyclone in combination with a slurry retention tank. The slurry is concentrated from a solids content of 10-15 wt% to 30-50 wt%. This, according to IEA, allows for a reduction in sensible heat needed for heating the slurry. 3.2. NH3 Scrubbing Cycle There is a possibility of having high ammonia slip from the CO2 capture cycle. Because of the ammonia slip there might be a need for a NH3 capture cycle, as mentioned by Mathias et al. (15). In this cycle, ammonia is absorbed into water to create an aqueous solution. The solution is fed to a regenerator where the ammonia solubility is lower because of the higher temperature. The regenerator pressure can be kept at atmospheric levels to reduce solubility (24). However, a slightly raised pressure is recommended in order to reduce vaporization of water. 11 4. Method To determine if the CAP can be integrated to the aluminium manufacturing process, two separate studies are carried out. The heat requirements for the CAP and how it is affected by low CO2 concentration is determined through modelling and simulations of the process. How much heat that can be extracted from the process gas and used for process integration is determined through modelling and simulations of the existing process gas treatment train and the added heat extractions. A model is created to investigate the effects of low CO2 concentration process gas on the required heat duty for CO2 capture. The outcome of the study is the reboiler duty and the specific reboiler duty1 of the CAP, required to achieve a predetermined CO2 capture rate and off-gas ammonia volumetric concentration. Knowledge of heat duty and temperature requirements is used to determine and propose available streams in the aluminium manufacturing process gas treatment train suitable for process integration of the CAP. Given the properties of the CAP, the model is built to handle solvent slip, reducing the solvent discharge to the environment to the regulated levels of 10 ppmv. To satisfy a discharge of exactly 10 ppmv requires a smaller reboiler step size than is used during simulations. Thus, the use of an ammonia discharge span instead of a fixed concentration is made as a simulation technical solution to increase convergence in the model. The ammonia concentration is restricted to within the span of 5-15 ppmv of the outgoing process gas. Heat duty requirement for sufficient solvent abatement is taken into account when the process integration is investigated. Table 2 presents the parametric study carried out for the CO2 capture cycle. In the study, CO2 concentration, solvent flow rate and the absorber CO2 loading is varied. The absorber CO2 loading is an important reference measure and it is defined as the number of moles of CO2 per moles of NH3 in the absorber at simulation start-up. The solvent flow rate is defined as the molar flow rate of the aqueous ammonia solution at start-up. Another important parameter, not varied specifically here, used for comparison with previous studies is the solvent composition. This is defined as the ammonia molality and describes the amount of ammonia in relation to water in the solvent, mol NH3/kg H2O. Table 2. Variables and values that make up the parametric study for the CO2 capture process Parameter Unit Low → High Concentration [vol% CO2] 4 7 10 15 Solvent flow rate [kmol/s] 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0 4.2 4.4 Absorber CO2 loading [mol CO2 / mol NH3] 0.25 0.3 0.35 0.4 To identify available surplus heat at sufficient temperature and to determine process gas properties and flow rates, a heat and mass balance simulation model of the aluminium manufacturing process gas treatment train is created. The model is used to determine process gas properties used as input to the CO2 capture model. Being the main heat source investigated for process integration, the process gas treatment train model is also used to determine where and how much heat that is available for heat integration of the CAP. 4.1. Process Data The CO2 capture cycle simulations are divided into two main cases. One case where the input values are obtained from Hydro, see Table 3, and another where the process gas flow rate is kept at a constant rate while the composition of CO2 and N2 is varied according to Table 4. They are referred to in the text as “Hydro case” and “constant flow case” respectively. The process data used for the Hydro case is based on measured and estimated results from the HAL Ultra process tests performed by Hydro. The provided data originates from different HAL Ultra operating conditions resulting in varying CO2 concentrations and flow rates for investigation of heat duty and process integration requirements and possibilities. For the process integration study, using 1 Defined as the energy needed to capture one unit mass of CO2 [kJ/kg CO2 captured] 12 input data provided by Hydro, varying the CO2 concentration also means varying the process gas flow rate. Table 3. The properties for process gas at various CO2 concentrations for the Hydro case. Temperature and pressure is 5°C and atmospheric respectively for each case when entering the CO2 capture process. Process gas out from cells Process gas entering the CO2 capture process Composition [wt%] CO2 concentration [vol%] Flow rate per cell [Nm3/h] Temperature [°C] Total mass flow [kg/s] N2 O2 CO2 H2O 4 1 600 290 82.43 71.2 22.4 5.9 0.528 7 1 000 345 60.40 67.2 22.0 10.2 0.520 10 800 365 53.07 63.4 21.7 14.4 0.512 15 487 379 41.59 57.4 21.1 21.0 0.498 Table 4. The properties for process gas at various CO2 concentrations, adjusted to constant flow rate. Temperature and pressure is 5°C and atmospheric respectively for each case when entering the CO2 capture process. Process gas out from cells Process gas entering the CO2 capture process Composition [wt%] CO2 concentration [vol%] Total mass flow [kg/s] N2 O2 CO2 H2O 4 60 72.6 21 5.9 0.5 7 60 68.3 21 10.2 0.5 10 60 64.1 21 14.4 0.5 15 60 57.4 21 21.0 0.5 4.2. Modelling Tools Two modelling software are used in this study, Ebsilon Professional 7.0 and Aspen Plus 7.2. The data provided by Hydro is limited to properties of untreated process gas, requiring initial simulations to obtain the CAP inlet gas stream. The Ebsilon model is therefore used to simulate the aluminium plant gas treatment train to obtain the process gas outlet conditions. The Ebsilon model is furthermore used to identify the amount of heat available in the process gas and is the model used when determining the location and specifications for process integration. The model used to simulate the CAP is built in Aspen which, except for mass and energy balance, also handles chemical reactions, complex solutions and the accompanying vapour-liquid equilibrium (VLE) calculations. The CAP model consists of two separate models. One model simulating the CO2 capture cycle, referred to as the bottom cycle, and another simulating scrubbing and recovery of ammonia slip, referred to as the top cycle. Both models are used to investigate the required heat duty to achieve desired CO2 capture and ammonia slip respectively. The absorbers and regenerators are modelled using the Aspen RadFrac component. The component is used for modelling of absorption and desorption. Two different calculation approaches can be used for RadFrac, equilibrium and rate-based calculation approach, more of which is described in 0. 13 4.3. Process Integration As a result of low working temperature for the bottom cycle absorber in relation to the high temperatures required in the regenerator, both heating and cooling demand is required. The CAP’ heating demand is the primary focus and it is determined for each of the different simulation cases. To investigate the possibility of process integration with the aluminium manufacturing process is therefore an important part of this report. With constant working temperatures of around 1 000°C, the aluminium cells are interesting from a heat integration point of view. Integration directly with cells would pose some advantages, the foremost being the high temperature of the cell compared to the process gas leaving the cell. However, bringing water in close proximity to the cells is prohibited for safety reasons. Leakage of water from such a heat exchanger system could result in steam explosion, posing a threat to both equipment and personnel. The heat integration of the CAP is therefore focused to the process gas treatment train, utilizing heat from the process gas as shown in Figure 7. Figure 7. Schematic overview of the aluminium manufacturing process and proposed heat source for process integration of the CAP. The heat consumers in the CAP are the bottom and top cycle reboilers and, while similar in principal and function, the working temperatures greatly differ between the two. The CO2 capture simulations provide separate information about the minimum required temperature for both the bottom and top cycle reboilers, hence determining the minimum temperature of the process gas stream at which process integration of the two cycles can be considered. For the process integration study, the simulation cases with lowest total reboiler duty for the combined bottom and top cycle is considered. This is done for the different CO2 concentrations and compared with the available heat from process integration with the aluminium process for the corresponding case. Limiting conditions in the existing process gas treatment train together with the required reboiler temperature levels determine the most suitable location for process integration. The maximum allowed temperature in the process gas dry scrubber is specified to 110°C, meaning that enough heat needs to be extracted to lower the temperature to that specified level. If more heat is extracted from the process gas stream, lowering the temperature below 110°C, downstream equipment like the district heating heat exchanger will experience declining temperature driving force and efficiency. Given the nature of the technological measures planned by Hydro made to increase outlet CO2 concentration, process gas flow rate is gradually decreased which in turn affects the potential for both process integration and district heating heat extraction. The district heating heat exchanger is not relocated or removed from its location in the existing process gas treatment train and output is optimized given the minimum required supply hot water temperature (90°C), return temperature (60°C) and the available heat in the process gas stream. The high heat development from exothermic reactions in the bottom cycle absorber, together with hot gas and solvent streams in need of cooling, results in a significant total cooling demand for the CAP. In this particular case, the location of the primary aluminium facility is beneficial when cooling requirements are considered. The close proximity to the sea and hence to cold sea water at constant temperature, greatly reduces the potential costs for cooling. The sea water, assumed to have a temperature of between 5-10°C when utilized in the process, is used both in the 14 process gas treatment train and for cooling in the CAP. The actual integration of cooling is assumed with certainty to be of minor significance compared to the heat integration, consisting mostly of electricity for pumping. The cooling integration will consequently not be treated in the results. 15 5. Modelling 5.1. Process Gas Treatment Model According to Hydro (4), the installed process gas treatment train would also be able to handle the decreased flow rate and higher concentration of contaminants expected from the future generation HAL Ultra cells considered in this report. The presented values and intermediate results in this model description are based on the HAL Ultra case with 4 vol% CO2 concentration. All heat exchangers in the model are assumed to have a minimum temperature difference of 10°C. The process modelling software Ebsilon Professional is used to acquire reliable properties and compositions for the process gas at the end of the process gas treatment train. The model can be divided into sections, shown in Figure 8, each describing a process in the gas treatment train. Figure 8. Describes the model of the process gas treatment train. Grey boxes mark the different included sections. The green dashed line indicates the presently installed equipment at the existing SU4, section 3 Knowledge about process gas properties and compositions at the cell outlet allows significant simplifications in the modelling. The process gas leaving the cells, presented earlier in Table 3 and Table 4 for all cases and in Table 5 for the 4 vol% CO2 concentration case, is generated and constitutes the main inlet condition of the model. Table 5. Properties and composition of the process gas leaving the HAL Ultra potline for the 4 vol% CO2 concentration case. HAL Ultra process gas Temperature [°C] Pressure [bar] Mass flow rate [kg/s] Process gas inlet 290 1.013 84.36 Composition [vol%] N2 O2 CO2 SO2 Ash Process gas inlet 75.26 20.70 4.0 0.0287 0.0000833 Process integration of the CAP reduces the need for supplementary heat addition. Heat exchangers, of which one is presented in Figure 9, are introduced at the beginning of the process gas treatment train. The heat extraction is modelled as two separate sections, one for heat integration of the bottom cycle reboiler, requiring high temperature heat, and another for the top cycle reboiler, working at lower temperature. Heat consumers connected to the heat exchangers simulate the reboilers in the CAP. The reboiler working temperature is assumed to be constant and the required heat exchanger temperatures are, with consideration to the minimum temperature difference, determined by the CO2 capture model to 121°C for the top cycle reboiler and to vary between 204-217°C for the bottom cycle reboiler. The top and bottom cycle heat extractions are presented in Table 6. Process gas outlet conditions after bottom and top cycle heat extractions are determined by the maximum temperature allowed at the dry scrubber inlet (4). 16 Figure 9. Possible heat extraction for process integration of the CAP is modelled by a heat consumer connected to a heat exchanger. Both the bottom and top cycle process integration heat exchangers are modelled accordingly. Table 6. The properties of the process gas entering and exiting the CAP integration heat exchangers. The table presents the results for the 4 vol% CO2 concentration case. High temperature heat extraction Temperature [°C] Pressure [bar] Mass flow rate [kg/s] Process gas inlet 290 1.013 84.36 Process gas outlet 206 1.013 84.36 Hot water/steam inlet 196 30 164 Hot water/steam outlet 206 30 164 High temperature heat available 7.37 [MW] Low temperature heat extraction Temperature [°C] Pressure [bar] Mass flow rate [kg/s] Process gas inlet 206 1.013 84.36 Process gas outlet 121 1.013 84.36 Hot water/steam inlet 111 3 172 Hot water/steam outlet 121 3 172 Low temperature heat available 7.31 [MW] As described in Chapter 2.1, alumina is mixed in with the process gas flow to react with and absorb fluoride from the process gas. The dry scrubber, shown in Figure 10, separates the fluoride enriched alumina and ash from the rest of the process gas stream. The dry scrubbing process is modelled by a selective splitter, separating all ash content from the process gas stream, and a heat extraction block, simulating the temperature drop in the scrubber. The heat extracted corresponds to the heat needed to attain actual dry scrubber outlet temperature. Figure 10. The dry scrubber is modelled by a selective splitter and a heat extraction block to simulate the ash and alumina separation and temperature drop of the actual process. 17 The heat exchanger connected to the Sunndalsøra district heating network, shown in Figure 11, is modelled by a heat exchanger and a heat consumer. The heat consumer simulates the hot water usage in the town of Sunndalsøra, extracting heat in order to reach the heat exchanger outlet temperature. Hot water is supplied to the district heating system at 90°C and returned at 60°C. Figure 11. A liquid-gas heat exchanger and a heat consumer are used to model the district heating heat exchanger. The wet scrubber, shown in Figure 12, is modelled in several steps in order to achieve a reliable simulation of the actual process. A wet scrubber resembles a direct contact cooler in that it introduces water directly into the process gas stream. The packed column wet scrubber is assumed to have an approach temperature of 1°C (13). Furthermore, sea water is assumed to be available at a maximum of 10°C throughout the year. (4) The introduction of sea water is simulated by saturating the process gas stream with water. Excess heat is extracted from the process gas stream in accordance to the maximum sea water temperature, lowering the process gas temperature to 11°C. To simulate SO2 removal a splitter removes all SO2 from the process gas. A drain is used to remove liquid water from the outlet process gas. Figure 12. The wet scrubber is modelled in four steps, saturation, heat extraction, SO2 removal and excess sea water drainage. A cooler is used in order to reach desired temperature of the exiting process gas. The temperature of the process gas is lowered from 11°C to 5°C. A drain is used to remove condensed water from the process gas stream. The properties and composition of the process gas exiting the cooler, presented in Table 7, are used as input to the CAP investigated in this report. 18 Table 7. The properties of the process gas entering the process gas cooler and the properties and composition of the process as leaving the cooler. Process gas cooler Temperature [°C] Pressure [bar] Mass flow rate [kg/s] Process gas inlet 10.85 1.013 82.66 Process gas outlet 5.0 1.013 82.43 Cooling duty in process gas cooler 577 [kW] Composition Process gas outlet N2 O2 CO2 H2O SO2 Ash vol% 74.64 20.53 3.97 0.86 - - wt% 71.17 22.36 5.94 0.53 - - 5.2. CAP CO2 Capture Model The simulations are carried out in two separate models, one describing CO2 capture in the bottom cycle and another describing NH3 abatement in the top cycle. The bottom cycle is illustrated in Figure 14 and the top cycle is illustrated in Figure 17. 5.2.1. Input Data The treated process gas is introduced to the CO2 capture process at 5°C and atmospheric pressure. The existing gas treatment is assumed to remove all contaminants like SO2, PM and HF. The inlet condition of the process gas is presented in Table 8 for the examined CO2 concentrations. Table 8. The properties for process gas at various CO2 concentrations leaving the aluminium plant gas treatment train. Temperature and pressure is 5°C and atmospheric respectively for each case. Process gas out from cells Process gas entering the CO2 capture process CO2 concentration [vol%] Total mass flow [kg/s] Composition [wt%] N2 O2 CO2 H2O M e a su re d / e st im a te d b 4 82.43 71.17 22.36 5.94 0.53 7 60.40 67.24 22.0 10.23 0.52 10 53.07 63.44 21.66 14.39 0.51 E x tr a p o la te d c 15 41.59 57.36 21.11 21.04 0.50 b) Measurements and estimations performed by Hydro (4). c) Extrapolated values from measurements and estimations performed by Hydro. Calculations further described in Appendix A – CO2 Concentration Properties Extrapolation. 19 5.2.2. Absorber Modelling The level of modelling complexity can be divided into five cases (25), the simplest being an equilibrium stage model with reaction equilibrium. Mass transfer is assumed to be infinitely fast within each stage and vapour and liquid phases are assumed to be at reaction equilibrium at the end of each stage (26). A more complex model is using a rate-based approach. The difference between equilibrium and rate-based calculations is that heat and mass transfer correlations are applied. A two-film model is used to describe heat and mass transfer in the gas-liquid interface, seen in Figure 13, as opposed to the two respective bulk phases. The two-film model assumes that there is no composition gradient in the bulk liquid due to high mixing and that all mass diffusion resistance is concentrated to the two film layers; where only steady state molecular diffusion occurs (27). In a reactive system such as CO2 absorption, this model is however not sufficient. Therefore, several types of reaction characteristics have to be taken into account, some of which are shown in Figure 13. The figure shows the partial pressure and concentration profiles in the gas and liquid films, depending on input conditions and reaction speed. The equilibrium stage model in Aspen Plus is equivalent to the simplest model with the addition of a bulk phase reaction kinetics assumption (8), increasing the accuracy of the model. According to Wang (8) equilibrium models are not sufficient when modelling amine absorption processes, why he suggests a rate-based approach being used instead. This is assumed to be the case for ammonia absorption processes as well and is why rate-based calculations with film and bulk phase reactions are applied. Onda et al. (1968) is used for mass transfer coefficient and interfacial area calculations and the Chilton and Colburn correlation is used for heat transfer coefficient calculations. Figure 13. Concentration profiles depending on chosen film theory. 20 5.2.3. CO2 Capture Cycle The main components in the bottom cycle CO2 capture model, shown in Figure 14, are the absorber and regenerator. Remaining components are described as auxiliary equipment at the end of this chapter. Figure 14. Illustrative process flowchart describing the CO2 capture cycle model. Simulations are performed through a set of cases with varying: process gas flow rate, containing individual CO2 concentrations in accordance to Table 8; solvent flow rate and solvent composition. A fixed design specification varies the regenerator reboiler duty in order to achieve a total CO2 capture rate of 85%. Reboiler duty, ammonia slip and reboiler temperature is generated as main results. The bottom cycle CO2 absorber, shown in Figure 15, is modelled with Aspen RadFrac packed column, using a rate-based approach. Rate-based modelling of the absorption column requires some additional specification of internal column design. The absorber diameter is calculated using a correlation by Grén et al. (28), this is further described in Appendix B – Dimensioning of Packed Column. The additional column design specifications; like height, number of stages, packing material and packing size, are all chosen for simulation technical reasons to achieve a stable model and should thus not be confused with actual absorber design specifications. The design parameters for one exemplified case with 4 vol% CO2 concentration are presented in Table 9. Figure 15. The packed CO2 absorber column is modelled using Aspen RadFrac block. 21 Table 9. Simulation technical design specifications and properties of the CO2 absorber and its connecting streams for a 4 vol% CO2 concentration case with 6.59 mol NH3/kg H2O. Numbers in brackets indicate parameters chosen for model stability and convergence and should not be seen as actual design parameters. Bottom cycle absorber design specifications CO2 absorber height (120) [m] CO2 absorber diameter 7.2 [m] Pressure 1.0132 [bar] Number of stages (40) [-] Packing type/material (Pall ring/Metal) [-] Packing size (16) [mm] CO2 absorber stream properties Stage Temperature [°C] Pressure [bar] Mass flow rate [kg/s] Process gas inlet 40 5 1.0132 82.43 CO2 absorber off-gas 1 6.85 1.0132 79.66 CO2-lean solution 1 / 5 5 1.0132 53.68 Aqueous ammonia make-up 30 5 1.0132 2.93 Total internal refrigeration duty in CO2 absorber 6. 7 [MW] Composition CO2 absorber off-gas N2 O2 CO2 H2O NH3 wt% 73.64 23.14 0.92 0.58 1.72 The process gas containing CO2 is fed into the absorber at the bottom stage. The returning CO2 lean solution, returning from the regenerator, is split into two equally sized streams and introduced to the absorber at stage 1 and 5. A make-up stream, introduced at stage 30, makes up for water and ammonia lost with the absorber and regenerator off-gas. Off-gas leaving the absorber at the top stage contains process gas, scrubbed free from most CO2 but contaminated by ammonia slip, constituting the reason for the ammonia scrubbing top cycle. Internal circulation and refrigeration is needed to counter the exothermic reactions and keep the working temperature in the absorber between the desired 0-20°C. The refrigeration is supplied by using utility exchangers, drawing solvent from one stage in the absorber, cooling it to 8°C and returning it at a higher stage. In the CO2 absorber model a total of 15 internal pumparounds, each with a flow rate of 1.5-2.5 kmol/s, are used to simulate the internal recirculation and to keep the temperature below 20°C at most times. The 15 internal pumparounds are chosen to achieve sufficient cooling and convergence in the model and does not necessarily reflect the reality in terms of amounts and placement. However, even in reality the CAP bottom cycle absorber is characterised by large internal recirculation (1). The CO2 regenerator is modelled as a packed column with the Aspen RadFrac block and the simulations are carried out at equilibrium, no internal column design is therefore specified. An equilibrium approach is a reasonable assumption since the temperature in the regenerator is very high, meaning high reaction speeds and lower impact from kinetics. The required reboiler duty and the temperature in the reboiler are determining factors when process integration is considered and hence of great interest for this study. 22 Figure 16. The packed CO2 regenerator column is modelled using Aspen RadFrac block. The regenerator has one inlet stream, the CO2-rich stream from the bottom stage of the absorber. The CO2-lean stream leaving the regenerator at the bottom is recycled back to the absorber for further absorption. The working pressure in the regenerator is 30 bar, allowing for desorption of CO2 whilst preventing vaporization of water and ammonia. The CO2 stream leaving the regenerator through the top contains a high percentage of CO2 mixed with water and ammonia. The water and ammonia in the stream can then be removed using a flash tank. The design parameters for one exemplified case with 4 vol% CO2 concentration are presented in Table 10. Table 10. Simulation technical design specifications and properties of the CO2 regenerator for a 4 vol% CO2 concentration case with 6.59 mol NH3/kg H2O. Numbers in brackets indicate parameters chosen for model stability and convergence and should not be seen as actual design parameters. Bottom cycle regenerator design specifications Pressure 30 [bar] Number of stages (15) [-] Regenerator stream properties Stage Temperature [°C] Pressure [bar] Mass flow rate [kg/s] Regenerator CO2 out 1 182.9 30 5.70 Reboiler temperature 212.7 [°C] Reboiler duty 13.9 [MW] Reboiler duty / kg of CO2 captured 3 337 [kJ/kg CO2 captured] Composition Regenerator CO2 out N2 O2 CO2 H2O NH3 wt% 1.410-2 9.1·10-3 73.0 20.4 6.5 23 5.2.4. NH3 Scrubbing Cycle Much like the bottom cycle, the top cycle is modelled with an absorber and regenerator as main components. The top cycle, illustrated in Figure 17, is simulated with a stand-alone model but with the inlet stream specified to have properties and flow rates equal to the absorber off-gas stream in the bottom cycle. Figure 17. Illustrative process flow chart describing the NH3 scrubbing cycle model. The function of the top cycle is to clean the bottom cycle off-gas from ammonia slip, leaving the final off-gas free from harmful concentrations of ammonia. The NH3 regenerator reboiler duty is varied to achieve a final ammonia slip of between 5-15 ppmv from the top cycle absorber, as suggested by Mathias, Paul M et al. (15). The top cycle NH3 absorber illustrated in Figure 18 is like its bottom cycle counterpart, modelled as a rate-based packed column. The internal column design parameters needed for rate-based calculations are presented for the same case as in the bottom cycle, with 4 vol% CO2 concentration, in Table 11. The off-gas from the bottom cycle is cooled, flashed and introduced at the last stage of the NH3 absorber. The NH3-rich and lean streams leaves and enters the top cycle absorber in the top and bottom stage respectively. Water lost from the top cycle through the absorber and regenerator off-gas is balanced with a water make-up stream introduced directly at the top of the absorber. Figure 18. The packed NH3 absorber column is modelled using Aspen RadFrac block. 24 Table 11. Simulation technical design specifications and properties of the CO2 absorber and its connecting streams for a 4 vol% CO2 concentration case with 6.59 mol NH3/kg H2O. Numbers in brackets indicate parameters chosen for model stability and convergence and should not be seen as actual design parameters. Top cycle absorber design specifications NH3 absorber height (60) [m] NH3 absorber diameter 6.5 [m] Pressure 1.0132 [bar] Number of stages (30) [-] Packing type/material (Pall ring/Metal) [-] Packing size (16) [mm] NH3 absorber stream properties Stage Temperature [°C] Pressure [bar] Mass flow rate [kg/s] Bottom cycle off-gas inlet 30 5 1.0132 78.11 NH3 absorber off-gas 1 5.01 1.0132 77.56 H2O make-up 1 5 1.0132 4.94 Composition NH3 absorber off-gas N2 O2 CO2 H2O NH3 wt% 75.63 23.76 5.8·10-2 0.55 5.9·10-4 The top cycle regenerator, illustrated in Figure 19, is an Aspen RadFrac equilibrium model of a packed column. Together with the bottom cycle reboiler, the required top cycle reboiler duty and temperature are of interest when considering process integration. The off-gas leaving the NH3 regenerator consists of mainly water and ammonia. No further treatment is considered in the model and it is assumed that the off-gas stream is flashed and the resulting ammonia stream is return back into the system, reducing need for and ammonia make-up. The properties of the top cycle regenerator are presented in Table 12. Figure 19. The packed NH3 regenerator column is modelled using Aspen RadFrac block. 25 Table 12. Simulation technical design specifications and properties of the NH3 regenerator for a 4 vol% CO2 concentration case with 6.59 mol NH3/kg H2O. Numbers in brackets indicate parameters chosen for model stability and convergence and should not be seen as actual design parameters. Top cycle regenerator design specifications Pressure 1.5 [bar] Number of stages (15) [-] NH3 regenerator reboiler properties Reboiler temperature 111.3 [°C] Reboiler duty 16.2 [MW] NH3 reboiler duty / kg of CO2 captured in bottom cycle 3 891 [kJ/kg CO2] 5.2.5. Auxiliary Equipment The auxiliary equipment includes the supporting equipment not directly coupled with the CO2 absorption/desorption and NH3 abatement. The components for the bottom and top cycles are in general modelled in the same way, sharing most of the settings. • Rich solvent pumps The pump is modelled as a pressure changer, increasing the pressure from the atmospheric pressure of the bottom and top cycle absorbers to the 30 and 1.5 bar respectively, the working pressures of the regenerators. No losses or efficiencies are considered when modelling the pump. • Rich/lean solvent heat exchangers The process internal heat exchangers are modelled with the Aspen HeatX heat exchanger block. The stream flow direction is set to countercurrent flow and the minimum hot outlet temperature approach is specified to 10°C as suggested by Kothandaraman, A (29) and Mathias, Paul M et al. (15). The model uses the shortcut calculation method and therefore no internal heat exchanger design specifications are needed. • Lean valves The lean valves are pressure changers, decreasing the pressure in the bottom and top cycles from 30 and 1.5 bar respectively back to atmospheric pressure. Adiabatic flash calculation is used for the specified outlet pressures. • Coolers Coolers are used to cool the CO2-lean and NH3-lean streams down to the desired absorber inlet temperature of 5°C, applicable for both the bottom and top absorber. Another cooler is used to cool the bottom cycle off-gas to 5°C. The bottom cycle off-gas cooler helps keeping the temperature in the top cycle absorber down and increases the effect of the bottom cycle off-gas flash tank by facilitating for more water vapour to be flashed to liquid. The coolers are modelled with the Aspen heater/cooler block, assuming no pressure loss and a specified outlet pressure of 5°C. • CO2-lean splitter The CO2-lean splitter is needed to enable the returning lean stream to be introduced to different stages in the bottom cycle absorber, helping with the convergence and improving the absorber temperature profile. The CO2-lean splitter splits the CO2-lean stream into two equally large flows and introduces them to the absorber at stages 1 and 5. • Bottom cycle off-gas flash tank The bottom cycle off-gas is flashed before it is introduced into the top cycle absorber. The concentration of ammonia in the off-gas stream is reduced by flashing water vapour into liquid and draining it, also removing the ammonia dissolved in the water. The stream is flashed at 5°C and atmospheric pressure. 26 27 6. Results 6.1. CO2 Capture Cycle It is rather complex to present how reboiler duty is affected by CO2 concentration only, since several other variables also come into play. Total solvent flow and CO2 loading in the absorber are also investigated in order to explain what parameters affect reboiler duty except from CO2 concentration. Ammonia slip is also of interest, as it affects the total system reboiler duty. The effects of absorber CO2 loading and solvent mole flow are shown in Figure 20. There is a tendency for reboiler duty to increase with increased absorber CO2 loading. Because there is less free ammonia available with higher absorber CO2 loading, the CO2 loading in the lean stream entering the absorber has to be lower. The driving force for the absorption reaction will otherwise decrease due to decreased difference in concentration between CO2 and NH3, causing less CO2 to be absorbed if other parameters are kept constant. The difference in reboiler duty is rather small, generally within 500 kW for a given solvent flow. It can also be noted that there for most cases is a slight increase in reboiler duty with increased solvent flow. The mole flow of ammonia is constant throughout each absorber CO2 loading case, why the concentration decreases with increased total solvent mole flow; leading back to the same reasoning about driving force. Because the total solvent flow is increased, there is also more liquid to heat in the reboiler. Figure 20. Reboiler duty per unit mass of captured CO2 shown for different lean stream loadings and grouped by total solvent flow. The data shown is for the 7% CO2 Hydro case. The simulations show that, for a given absorber CO2 loading, specific reboiler duty generally decreases with increased CO2 concentration. It is however possible that there might be a concentration where the reboiler duty is at a minimum as the simulation results, shown in Figure 21 and Figure 22, hint that reboiler duty might increase for the 15% CO2 case compared to the 10% CO2 case. The explanation to this could be that the residence time in the absorber is not sufficient in the 15% case. The absorption reaction is known to be slow, why this could be the limiting factor, forcing the CO2 loading in the lean stream to be lower and thereby increasing the amount of free ammonia in the 15% case compared to 10% case. Alas, this theory has to be rejected based on the simulation results as the lean stream CO2 loading on an average is higher in the 15% case than in the 10% case. Instead, the reason is believed to be that the concentration of ammonia in the CO2 out stream is unproportionally high in the case of 15% CO2 concentration compared to the lower ones, see Figure 23. Closer examination of the cases makes it possible to deduct where the required reboiler duty originates from. The major contribution is the sensible heat needed to heat the water in the reboiler, followed by the heat of reaction for the absorption reactions. When looking at the required heat per kg CO2 captured, it was found that the sensible heat decreases more than what the heat of reaction increases, with the exception of the 15% case. 0 500 1000 1500 2000 2500 3000 3500 4000 2 2.5 3 3.5 4 4.5 5 S p ec if ic r eb o il er d u ty [k J/ k g C O ₂captu re d ] Solvent mole flow [kmol/s] Absorber CO₂ loading 0.25 Absorber CO₂ loading 0.3 Absorber CO₂ loading 0.35 Absorber CO₂ loading 0.4 28 Figure 21. Reboiler duty averaged for different solvent flows, grouped by absorber CO2 loading, Hydro case. Figure 22. Reboiler duty averaged for different solvent flows, grouped by absorber CO2 loading, constant flow case. 0 1000 2000 3000 4000 5000 0 5 10 15 20 25 30 35 40 0.25 0.3 0.35 0.4 S p ec if ic r eb o il er d u ty [k J/ k g C O 2 ca p tu re d ] R eb o il er d u ry [ M W ] Absorber CO2 loading [mol CO2/mol NH3] Reboiler duty, 4% CO₂ case Reboiler duty, 7% CO₂ case Reboiler duty, 10% CO₂ case Reboiler duty, 15% CO₂ case Specific reboiler duty, 4% CO₂ case Specific reboiler duty, 7% CO₂ case Specific reboiler duty, 10% CO₂ case Specific reboiler duty, 15% CO₂ case 0 1000 2000 3000 4000 5000 0 5 10 15 20 25 30 35 40 0.3 0.35 0.4 S p ec if ic r eb o il er d u ty [k J/ k g C O 2 ca p tu re d ] R eb o il er d u ry [ M W ] Absorber CO2 loading [mol CO2/mol NH3] Reboiler duty, 4% CO₂ case Reboiler duty, 7% CO₂ case Reboiler duty, 10% CO₂ case Reboiler duty, 15% CO₂ case Specific reboiler duty, 4% CO₂ case Specific reboiler duty, 7% CO₂ case Specific reboiler duty, 10% CO₂ case Specific reboiler duty, 15% CO₂ case 29 Figure 23. Mole fraction ammonia in the stream leaving the bottom cycle regenerator, Hydro case. Slip from the absorber is of great significance due to legislation limiting the allowed ammonia concentration out from the system. The configuration also encompasses a separate top cycle designed to reduce ammonia slip to within legal limits. Figure 24 and Figure 25 show that absorber slip increases significantly with increasing NH3 molality, and thus with increasing CO2 concentration. This is believed to be because of the higher ammonia vapour pressure due to increased concentration. Figure 24. Ammonia slip from the absorber, Hydro case. 0 0.05 0.1 0.15 0.2 0.25 0.3 0.2 0.25 0.3 0.35 0.4 0.45 N H 3 m o le f ra ct io n i n r eg en er at o r o ff -g as Absorber CO2 loading [mol CO2/mol NH3] 4% CO₂ case 7% CO₂ case 10% CO₂ case 15% CO₂ case 0 2 4 6 8 10 12 14 16 18 20 0% 2% 4% 6% 8% 10% 12% 14% 16% 18% 0.25 0.3 0.35 0.4 N H 3 m o la li ty [ m o l N H 3/ k g H 2O ] A b so rb er N H 3 sl ip [ vo l% ] Absorber CO2 loading [mol CO2/mol NH3] Slip, 4% CO₂ case Slip, 7% CO₂ case Slip, 10% CO₂ case Slip, 15% CO₂ case Molality, 4% CO₂ case Molality, 7% CO₂ case Molality, 10% CO₂ case Molality, 15% CO₂ case 30 Figure 25. Ammonia slip from the absorber, constant flow case. An important parameter when looking at ammonia slip is the lean stream CO2 loading, shown in Figure 26. The results clearly show a decrease of slip with increased lean stream loading in the higher absorbent CO2 loading cases, as could be expected. When absorber CO2 loading decreases, the effect is more unclear. It could be expected due to the large increase of solvent NH3 molality as seen in Figure 24 and Figure 25, causing higher ammonia partial pressure and counteracting the effect. Figure 26. Ammonia slip from absorber with respect to lean stream CO2 loading, Hydro case. 0 5 10 15 20 25 0% 2% 4% 6% 8% 10% 12% 14% 16% 18% 0.3 0.35 0.4 N H 3 m o la li ty [ m o l N H 3/ k g H 2O ] A b so rb er N H 3 sl ip [ vo l% ] Absorber CO2 loading [mol CO2/mol NH3] Slip, 4% CO₂ case Slip, 7% CO₂ case Slip, 10% CO₂ case Slip, 15% CO₂ case Molality, 4% CO₂ case Molality, 7% CO₂ case Molality, 10% CO₂ case Molality, 15% CO₂ case 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35N H ₃ m o le f ra ct io n i n a b so rb er o ff -g as Lean stream CO₂ loading [mol CO₂/mol NH₃] 4 vol% CO₂ case, absorber CO₂ loading 0.25 4 vol% CO₂ case, absorber CO₂ loading 0.30 4 vol% CO₂ case, absorber CO₂ loading 0.35 4 vol% CO₂ case, absorber CO₂ loading 0.40 7 vol% CO₂ case, absorber CO₂ loading 0.25 7 vol% CO₂ case, absorber CO₂ loading 0.30 7 vol% CO₂ case, absorber CO₂ loading 0.35 7 vol% CO₂ case, absorber CO₂ loading 0.40 10 vol% CO₂ case, absorber CO₂ loading 0.25 10 vol% CO₂ case, absorber CO₂ loading 0.30 10 vol% CO₂ case, absorber CO₂ loading 0.35 10 vol% CO₂ case, absorber CO₂ loading 0.40 15 vol% CO₂ case, absorber CO₂ loading 0.25 15 vol% CO₂ case, absorber CO₂ loading 0.30 15 vol% CO₂ case, absorber CO₂ loading 0.35 15 vol% CO₂ case, absorber CO₂ loading 0.40 31 The temperature in the absorber is to be held below 20°C for good performance and preferably under 10°C for optimal performance. These figures are unfortunately not always achieved, and absorber temperatu