CO2 Capture With MEA: Integrating the Absorption Process and Steam Cycle of an Existing Coal-Fired Power Plant by Colin F. Alie A thesis presented to the University of Waterloo in ful?llment of the thesis requirement for the degree of Master of Applied Science in Chemical Engineering Waterloo, Ontario, Canada, 2004 ? Colin F. Alie 2004 I hereby declare that I am the sole author of this thesis. This is a true copy of the thesis, including any required ?nal revisions, as accepted by my examiners. I understand that my thesis may be made electronically available to the public. ii Abstract In Canada, coal-?red power plants are the largest anthropogenic point sources of atmospheric CO2. The most promising near-term strategy for mitigating CO2 emis- sions from these facilities is the post-combustion capture of CO2 using MEA (mo- noethanolamine) with subsequent geologic sequestration. While MEA absorption of CO2 from coal-derived ?ue gases on the scale proposed above is technologically feasible, MEA absorption is an energy intensive process and especially requires large quantities of low-pressure steam. It is the magnitude of the cost of providing this supplemental energy that is currently inhibiting the deployment of CO2 capture with MEA absorption as means of combatting global warming. The steam cycle of a power plant ejects large quantities of low-quality heat to the surroundings. Traditionally, this waste has had no economic value. However, at different times and in different places, it has been recognized that the diversion of lower quality streams could be bene?cial, for example, as an energy carrier for district heating systems. In a similar vein, using the waste heat from the power plant steam cycle to satisfy the heat requirements of a proposed CO2 capture plant would reduce the required outlay for supplemental utilities; the economic barrier to MEA absorption could be removed. In this thesis, state-of-the-art process simulation tools are used to model coal com- bustion, steam cycle, and MEA absorption processes. These disparate models are then combined to create a model of a coal-?red power plant with integrated CO2 capture. A sensitivity analysis on the integrated model is performed to ascertain the process vari- ables which most strongly in?uence the CO2 energy penalty. From the simulation results with this integrated model, it is clear that there is a sub- stantial thermodynamic advantage to diverting low-pressure steam from the steam cycle for use in the CO2 capture plant. During the course of the investigation, methodologies for using Aspen Plus? to predict column pressure pro?les and for converging the MEA absorption process ?owsheet were developed and are herein presented. iii Acknowledgements I would like extend my thanks and my appreciation to all those who have assisted me in preparing this thesis: ? Dr. Eric Croiset and Dr. Peter Douglas for the privilege of working with them and for their guidance and mentorship. ? Blair Seckington, personally, and Ontario Power Generation (OPG), as a whole, for their ?nancial and technical support. ? Dr. Thomas Duever, Graeme Lamb, Dr. William Anderson, Dennis Herman, and Wendy Irving who, when called upon, provided their insight and assistance. And, lastly, to my loving wife Amanda, who supported me through what will surely be the longest "two weeks" of our lives. . . iv Contents Acronyms and Abbreviations xiv Chemical Symbols and Formulae xvii Nomenclature xix 1 Introduction 1 1.1 Objective 1 1.2 Motivation 1 1.2.1 Fossil fuels, carbon dioxide, and climate change 1 1.2.2 Fossil fuels and electric power generation 2 1.2.3 Generating electricity while mitigating CO2 emissions 4 1.2.4 Capturing CO2 with MEA 8 1.3 Implementation 10 1.3.1 Selection of study basis 10 1.3.2 Selection of simulation software 10 1.3.3 Outline of thesis 11 2 Flue Gas Synthesis 13 2.1 Objective 13 2.2 Rationale 13 2.2.1 Model ?exibility 13 v 2.2.2 Model accuracy 14 2.3 Implementation 15 2.3.1 Specifying properties 15 2.3.2 Specifying streams 18 2.3.3 Specifying blocks 18 2.4 Model Validation 19 2.4.1 Coal heat of combustion 19 2.4.2 Flue gas ?ow rate 19 2.5 Conclusions and Recommendations 21 3 Simulation of Steam Cycle 22 3.1 Objective 22 3.2 Motivation 22 3.3 Points of emphasis 23 3.4 Implementation 24 3.4.1 Specifying properties 24 3.4.2 Specifying streams 26 3.4.3 Specifying blocks 26 3.5 Model validation 33 3.5.1 Property method 33 3.5.2 Steam temperature, pressure, and ?ow potential 34 3.5.3 Part-load power output and heat input 34 3.5.4 Turbine and unit heat rate 34 3.6 Conclusions and recommendations 39 4 Simulation of MEA Absorption Process 42 4.1 Objective 42 4.2 Motivation 42 4.2.1 Process ?owsheet evaluation 42 vi 4.2.2 Equipment design 45 4.2.3 Solvent selection 46 4.2.4 Optimizing process operating conditions 47 4.2.5 Process integration exploration 48 4.3 Points of emphasis 48 4.4 Implementation 49 4.4.1 Specifying properties 49 4.4.2 Specifying streams 52 4.4.3 Specifying blocks 53 4.5 Model Parameter elucidation 58 4.5.1 Property method selection 58 4.5.2 Absorber and Stripper internal con?guration 62 4.6 Conclusions and recommendations 71 5 Integration of Power Plant and MEA Absorption 73 5.1 Introduction 73 5.2 Implementation 76 5.2.1 Location of steam extraction and condensate re-injection . . . . 76 5.2.2 Maximum available steam for Stripper reboiler heating 82 5.2.3 Flue gas pre-conditioning 83 5.2.4 Stripper reboiler 84 5.2.5 Blower and CO2 Compressor 84 5.3 Process Simulation 84 5.3.1 Sensitivity of CO2 capture to recycle CO2 loading 84 5.3.2 Sensitivity of CO2 capture to Absorber height 87 5.3.3 Sensitivity of CO2 capture to Stripper height 89 5.4 Model validation 91 5.5 Conclusions and recommendations 94 vii 6 Conclusion and Future Work 95 6.1 Conclusion 95 6.2 Future work 97 A Conditions of steam at potential extraction locations 99 B Sieve Tray Column Hydrodynamic Design Recipe 101 B.1 Tower diameter 101 B.2 Downcomer ?ooding 103 B.3 Tray pressure drop 106 B.4 Downcomer seal 107 B.5 Weeping 107 C Steam Energy Calculations 108 D Comparison of Calculated CO2 Solubility With Experimental Values 111 E Aspen Plus Input ?le for Power Plant With Integrated MEA Absorption 116 Glossary 149 List of References 150 viii List of Tables 1.1 Canadian emissions of greenhouse gases, 2000 2 1.2 Electricity generation in Canada, 2000 3 1.3 Electricity generation from thermal power plants, 2000 3 1.4 Coal use across Canada, 2000 7 1.5 Age distribution of Canadian coal power plants: 2000 and 2010 8 2.1 Coal characteristics 16 2.2 Comparison of ? ?ch??? with observed GCV 19 2.3 Flue gas ?ow rate simulation input data 20 2.4 Comparison of calculated ?ue gas ?ow rate with observed values . . . . 20 2.5 Flue gas composition 21 3.1 Design inlet volumetric ?ow rates into turbine sections 27 3.2 Ratio of discharge pressure to inlet pressure for turbine groups 28 3.3 Fractional isentropic ef?ciency of turbine groups 29 3.4 Comparison of calculated internal power with design values 33 3.5 Comparison of calculated heat input with design values 34 4.1 Hydrodynamic performance neglect matrix 49 4.2 Property methods and model available for CO2-MEA-H2O system . . . 51 4.3 UOM's in MEA absorption process model 53 4.4 Design parameters for sizing and hydrodynamic evaluation of columns . 56 4.5 Survey of CO2 delivery pressures used in MEA absorption studies . . . 57 ix 4.6 Summary of results from Absorber study 66 4.7 MEA absorption process model initialization parameters 71 5.1 Scope of MEA absorption sensitivity analysis 84 5.2 MEA absorption process energy duties 92 5.3 Stripper reboiler speci?c heat duty 93 5.4 Summary of best cases from sensitivity studies 94 A.1 Base and part-load conditions in Nanticoke steam cycle 99 B.1 Required input for sizing and hydrodynamic evaluation of tray columns 102 C.1 Changes in steam internal energy in steam cycle 110 x List of Figures 1.1 Utilization of natural resources for electricity generation 4 1.2 Process ?ow diagram for CO2 removal via chemical absorption 9 2.1 Coal combustion simulation ?owsheet 15 3.1 Steam cycle simulation ?owsheet 25 3.2 High-pressure section pressure ratio at part-load 27 3.3 LP3 and LP4 stage groups' isentropic ef?ciency at part-load 30 3.4 Turbine 'bleed' steam ?ow rates at part-load 31 3.5 Boiler feed water temperature at part-load 32 3.6 Potential steam extraction locations in steam cycle 35 3.7 Steam temperature at part-load 35 3.8 Steam pressure at part-load 36 3.9 Steam ?ow rate at part-load 37 3.10 Turbine power output at part-load 37 3.11 Turbine heat duty at part-load 38 3.12 Main turbine Sankey diagram 38 3.13 Main turbine work and energy ?ows 40 3.14 Boiler feed water pump turbine mechanical power losses 40 3.15 Turbine and unit heat rate at part-load 41 4.1 Base MEA absorption process ?owsheet 43 4.2 Amine Guard FS? process ?owsheet 43 xi 4.3 Kerr-McGee/Lummus Crest Global MEA absorption process ?owsheet 44 4.4 'Split feed' MEA absorption process ?owsheet 45 4.5 MEA absorption simulation ?owsheet 50 4.6 Solubility of CO2 in 30 wt% MEA solution 59 4.7 Comparison of calculated VLE with experimental values at 40? C . . . . 60 4.8 Comparison of calculated VLE with experimental values at 120? C . . . 60 4.9 Residual analysis of VLE data — ?PCO2 vs αlean at 40? C 61 4.10 Residual analysis of VLE data — ?PCO2 vs αlean at 120? C 62 4.11 Sensitivity of Flean to Absorber height 64 4.12 Sensitivity of Absorber downcomer ?ooding to Absorber tray spacing . 67 4.13 Sensitivity of Qreb to Stripper height 69 4.14 Sensitivity of Stripper downcomer ?ooding to Stripper tray spacing . . 70 5.1 Enthalpy-entropy curve for power plant 74 5.2 Implication of steam extraction on steam cycle work and heat ?ows . . 75 5.3 Power plant with integrated MEA absorption simulation ?owsheet . . . 77 5.4 Base-load steam conditions in steam cycle 79 5.5 High-pressure section of Nanticoke turbine 80 5.6 Intermediate-pressure section of Nanticoke turbine 80 5.7 Low-pressure section of Nanticoke turbine 81 5.8 Lengthwise view of Nanticoke turbine 81 5.9 Sensitivity of power plant electricity output to steam extraction 82 5.10 Sensitivity of ? ?P? Absorber and Qreb to CO2 loading 85 5.11 Sensitivity of capture plant's electricity demand to CO2 loading 85 5.12 Sensitivity of power plant electricity output to CO2 loading 86 5.13 Sensitivity of ? ?P? Absorber and Qreb to Absorber height 87 5.14 Sensitivity of capture plant's electricity demand to Absorber height . . . 88 5.15 Sensitivity of power plant electricity output to Absorber height 88 5.16 Sensitivity of ? ?P? Absorber and Qreb to Stripper height 89 xii 5.17 Sensitivity of capture plant's electricity demand to Stripper height . . . 90 5.18 Sensitivity of power plant electricity output to Stripper height 90 6.1 In?uence of CO2 loading on plant thermal ef?ciency 96 6.2 In?uence of Absorber height on plant thermal ef?ciency 96 6.3 In?uence of Stripper height on plant thermal ef?ciency 97 D.1 Comparison of calculated VLE with experimental values at 0? C . . . . 111 D.2 Comparison of calculated VLE with experimental values at 25? C . . . . 112 D.3 Comparison of calculated VLE with experimental values at 40? C . . . . 112 D.4 Comparison of calculated VLE with experimental values at 60? C . . . . 113 D.5 Comparison of calculated VLE with experimental values at 80? C . . . . 113 D.6 Comparison of calculated VLE with experimental values at 100? C . . . 114 D.7 Comparison of calculated VLE with experimental values at 120? C . . . 114 D.8 Comparison of calculated VLE with experimental values at 150? C . . . 115 xiii Acronyms and Abbreviations ABB Asea Brown Boveri Ltd. AMP 2-amino-2-methyl-1-propanol ASME American Society of Mechanical Engineers CFC'S chloro?uorocarbons CORAL CO2-removal absorption liquid DEA diethanolamine DGA diglycolamine DIPA diisopropanolamine DTD drain temperture difference In a feed water preheater, the differ- ence in temperature between the condensate outlet and the feed water outlet. EOR enhanced oil recovery EOS equation of state FG ?ue gas GCV gross calori?c value GHG greenhouse gas HFC'S hydro?uorocarbons HP high-pressure IEA International Energy Agency xiv IGCC integrated gasi?cation combined cycle IP intermediate-pressure IPCC Intergovernmental Panel on Climate Change KEPCO Kansai Electric Power Company Inc. KP Kansai packing KS Kansai solvent LP.low-pressure MCR maximum continuous rating MDEA methyldiethanolamine MEA monoethanolamine MHI Mitsubishi Heavy Industries Ltd. N/A not available/not applicable NBS National Bureau of Standards NCV net calori?c value NGCC natural gas combined cycle NRC National Research Council OPG Ontario Power Generation PCC pulverized coal combustion PRB Powder River Basin SOFC solid oxide fuel cell TEA triethanolamine THR turbine heat rate TNO Short-form of of?cial dutch name Nederlandse Organisatie voor toegepast- natuurwetenschappelijk onderzoek TNO (Netherlands Organisation for Applied Scienti?c Research TNO, in english). xv TTD terminal temperature difference In a feed water preheater, the dif- ference between the saturation temperature of the steam and the tem- perature of the feed water or condensate outlet. UHR unit heat rate UOM unit operation model UOP Universal Oil Products LLC USLS U. S. low-sulphur VLE vapour-liquid equilibrium xvi Chemical Symbols and Formulae Ar argon C carbon CH4 methane Cl chlorine CO carbon monoxide CO2 carbon dioxide H hydrogen H2 molecular hydrogen H2O water HCl.hydrogen chloride HF.hydrogen ?uoride N nitrogen N2 molecular nitrogen N2O nitrous oxide NaOH sodium hydroxide NO nitrogen oxide NO2 nitrogen dioxide NOx nitrogen oxides xvii O2 molecular oxygen S sulphur SF6 sulphur hexa?uoride SO2 sulphur dioxide SOx sulphur oxides xviii Nomenclature Variables α CO2 loading Aa active area of tray Ah area of tray covered by holes a,b regression parameters EFA approach to entrainment ?ooding E electrical power output Cp speci?c heat capacity ? change in value dh diameter of holes in tray ?E electrical power loss ε roughness factor F liquid molar ?ow rate f length of weir speci?ed as a fraction of tray diameter G vapour mass ?ow rate ?ch speci?c heat of combustion ?f h speci?c heat of formation hc downcomer clearance; height of gap between tray and downcomer apron xix hw weir height H weight percent hydrogen in coal, wet basis h speci?c enthalpy k¤ 1 rate of reverse reaction k1 rate of forward reaction ky vapour phase mass transfer coef?cient k column tray spacing scale factor ? length L liquid mass ?ow rate ˙ m mass ?ow rate M molecular mass m mass η ef?ciency N molar ?ux N number of trays n number of moles ?P power loss ?P pressure drop P power Px partial pressure of component x P pressure Q heat ?ow rate q volumetric ?ow rate ρ density xx R gas constant σ surface tension tt tray thickness TS tray spacing T temperature U internal energy V volume w weight fraction x mole fraction y vapour phase mole fraction Subscripts ∞ conditions in bulk ?uid i conditions at interface abs pertaining to the Absorber bfpt pertaining to boiler feed water pump turbine shaft bleed pertaining to bleed stream boil pertaining to block BOIL in steam cycle b pertaining to power plant boiler col pertaining to column c conditions at 'cold-side' exc pertaining to exciter gen pertaining to generator terminal gross before applicable losses have been accounted for G pertaining to gas phase xxi in conditions at inlet lean pertaining to that part of the recycle loop of the MEA absorption process with a relatively low concentration of CO2 L pertaining to liquid phase mech mechanical net after applicable losses have been accounted for out conditions at outlet plant pertaining to the power plant reb pertaining to the Stripper reboiler reht pertaining to block REHT in steam cycle sta pertaining to station service str pertaining to the Stripper s isentropic conditions th thermal trans pertaining to main transformer w pertaining to the tray weir Superscripts ? property at standard state * denotes set point or optimal value d dry basis i in reference to i'th iteration min denotes minimum value m mineral matter-free basis sat property at saturated conditions d diameter xxii Chapter 1 Introduction 1.1 Objective Capturing substantial amounts of CO2 from the ?ue gas from a coal-?red power plant using amine absorption technology requires large amounts of energy, mostly in the form of heat. The objective of this thesis is to evaluate the feasibility of obtaining the heat required for amine absorption from the existing power plant. 1.2 Motivation 1.2.1 Fossil fuels, carbon dioxide, and climate change The greenhouse effect refers to the phenomenon whereby gases in the upper atmosphere absorb a portion of the heat radiated by the earth. It is estimated that the Earth's temper- ature is 33? C warmer than it would be if this energy were instead transmitted to space [18]. Increasingly, the by-products of human activity are enhancing this 'natural' green- house effect stimulating a change in climate with potentially devastating effects for the planet's inhabitants. The IPCC (Intergovernmental Panel on Climate Change) has identi?ed six anthro- pogenic gases with climate change potential: CO2, CH4, N2O, SF6, CFC'S (chloro?u- orocarbons), and HFC'S (hydro?uorocarbons). Table 1.1 shows Canadian emissions of these gases. The ?rst column of Table 1.1, Global Warming Potential, expresses each compound's ability to absorb heat radiation on a unit mass basis. While, of the six greenhouse gases, 1 Table 1.1: Canadian emissions of greenhouse gases, 2000 (Source: Environment Canada [47]) Global Warming 1990 2000 Potential [Mt] § Mt CO2 eq¨? Mt § Mt CO2 eq¨? CO2 1 472 472 571 571 CH4 21 3.5 73 4.4 91 N2O 310 0.17 53 0.17 54 HFC'S 40–1170 0.9 CFC'S 6500–9200 6 6 SF6 23900 2.9 2.3 CO2 has the lowest Global Warming Potential, it is has the largest global climate change impact because its total emissions are so much greater than the others. Thus, current efforts in preempting climate change focus on strategies for the reduction of CO2 emis- sions. 1.2.2 Fossil fuels and electric power generation Electricity is a means to an end and not an end in and of itself. We need energy that is chemical, thermal, mechanical, etc. and our societies have evolved or are evolving such that electrical energy is often an intermediate form. Energy cannot be created or destroyed; it may be changed from one form to another. "Electric power generation" is actually "energy conversion". The energy conversion process selected is often site speci?c — "you take what you can get". In Canada — a large country with varied geography, topology, and geology — there are many different types of power plants. Table 1.2 presents the installed generating capacity and the actual generation of electric energy categorized loosely by type of power plant. Most of Canada's electricity is hydroelectric with signi?cant contributions from 'con- ventional' steam, nuclear, and combustion turbine plants. The last four categories of power in Table 1.2 use non-renewable energy sources and the last three — 'conventional' steam, combustion turbine, and internal combustion — are the ones typically associated with CO2 emissions (hydrocarbon fueled). Currently, most of this thermal electricity, 93.1%, is produced by utilities. Table 1.3 shows the amount of electric energy these utilities generated from the various non-renewable fuels. 2 Table 1.2: Electricity generation in Canada, 2000 (Source: Statistics Canada [22]) Source Installed generating capacity Generation of electric energy [MW] [%] [MWh] [%] Hydro 67 407 60.6 354 548 767 60.5 Non-conventional 96 0.1 263 820 0.0 Nuclear 10 615 9.5 68 675 253 11.7 Conventional steam 27 721 24.9 143 262 501 24.5 Internal combustion 654 0.6 1 356 761 0.2 Combustion turbine 4 808 4.3 17 706 788 3.0 Total 111 301 100.0 585 813 890 100.0 Table 1.3: Electricity generation from thermal power plants, 2000 (Source: Statistics Canada [22]) Fuel Generation of electric energy [MWh] [%] Coal 106 429 553 49.5 Petroleum 10 600 250 4.9 Natural Gas 26 623 329 12.4 Wood 1 830 560 0.8 Uranium 68 675 251 31.9 Other 961 711 0.4 Total 215 120 654 100.0 3 1.2.3 Generating electricity while mitigating CO2 emissions A laudable goal is to reduce CO2 emissions suf?ciently to stabilize atmospheric CO2 concentrations at a 'comfortable' level. Of the total GHG (greenhouse gas) emissions shown in Table 1.1, 128 Mt in 2000, 17.6% of the total for that year, resulted from the combustion of fossil fuels for the production of heat and electricity. In contrast, 95 Mt of GHG emissions were produced for the same reasons as in 1990 representing 15.7% of that year's production. Apparently, doing nothing is not an option. So then, how can CO2 production be mitigated during electricity generation? Figure 1.1 identi?es ?ve useful demarcation points in the discussion: electricity coal petroleum natural gas wood other conventional steam internal combustion combustion turbine rivers, streams wind, tidal flows, solar uranium hydro non-conventional nuclear 5 4 3 2 carbon dioxide 1 Figure 1.1: Utilization of natural resources for electricity generation x Produce less electricity In Canada, it is inconceivable that a shortfall exist between electricity supply and demand. Therefore, it is not possible for utilities to produce less power than is demanded and have brownouts, for example. y Switch from CO2 emitting to non-CO2 emitting electricity sources In cases where there is a mix of CO2 emitting and non-CO2 emitting electricity sources, it is probably already true that non-CO2 emitting sources are used pref- erentially for economic reasons. For example, OPG (Ontario Power Generation), which owns 75% of the generating capacity in Ontario [21], uses its hydroelec- tric and nuclear capacity for base-load supply and its fossil fuel plants for peaking power [50]. There is potential for retiring CO2 emitting plants and building new non-CO2 emitting capacity.1 However, the non-CO2 emitting electricity options have other 1 For the record, even the 'non-CO2 emitting' power plants will have associated, albeit secondary, GHG 4 challenges which detract from their appeal. Nuclear power plants have relatively lengthy construction schedules (on the order of a decade) so a decision today to switch to nuclear power would not realize CO2 reductions in the short or near- medium term. Sources of electricity derived from the sun and/or wind are prob- lematic principally because of their intermittency. So, a large installed capacity of non-conventional power plants would need to accompanied by a large installed capacity of energy storage facilities or conventional power plants in order to keep the lights on when the sun isn't shining and/or the wind isn't blowing. There are also issues that are neither of a technical or economic nature that need to be dealt with in taking this course of action. For example, in the case of more nuclear power, there are serious public concerns regarding the safety of nuclear power plants and the disposal of nuclear waste. In the case of wind power, there is some resistance to turbines "littering" the landscape. While these concerns may seem irrational or frivolous to some, they exist and along with the technical and economic concerns, would have to be addressed. z Improve energy ef?ciency of energy conversion processes This occurs in three ways. One, upgrades are made to existing installations. For example, at OPG's Nanticoke Generating Station, new turbine blades installed in a couple of the units should improve the energy ef?ciency of these units by 1–2 percentage points [50]. Two, for existing process designs, technological advances allow new installations to operate more ef?ciently. For example, improvements in materials engineering has led to manufacture of steam boilers capable of working under higher pressures which has led to higher overall steam cycle ef?ciencies. Three, altogether new processes have been developed which allow conventional fu- els to be used more ef?ciently. For example, Canadian electric utility power plants using coal had an average thermal ef?ciency of 33.04% in 2000 [22]. In contrast, using coal in an IGCC (integrated gasi?cation combined cycle), ef?ciencies of up to 51% are proposed. { Use lower carbon intensity fuels This is commonly referred to as fuel-switching and almost always refers to sub- stituting natural gas for coal. A 'back of the envelope' calculation shows that 2.5 emissions. Examples of secondary emissions include releases of methane gas caused by the decomposition of organic material in regions ?ooded by hydroelectric dams, CO2 emissions associated with manufactur- ing cement used in construction and transportation of fuel and wastes to and from nuclear power plants. 5 times more CO2 is released if coal is used rather than natural gas to produce a given amount of heat.2 A major disadvantage to this speci?c substitution is the price of natural gas. Firstly, natural gas, on a unit energy basis, is more expensive than coal. Secondly, its price is subject to much more ?uctuation. Other disadvantages vis-` a-vis this particular fuel-switch are that natural gas is more dif?cult to transport and store than coal and its proven reserves are also much less (according to the National Energy Board [6, p 75], as of 1991, there were 91 years of domestic coal reserves versus just nine years of natural gas reserves). Bio-fuels could also represent a class of hydrocarbons with a lower carbon inten- sity than coal. The actual fuel combustion would be carbon-neutral; all associated carbon emissions would result from the ancillary collection, processing, and trans- portation activities. A full life cycle assessment would be necessary to determine if this type of fuel-switching is indeed bene?cial. | Capture and storage of CO2 The CO2 produced as part of the energy conversion process is captured prior to being released to the atmosphere and subsequently stored. The capture can be performed either pre-combustion or post-combustion and there are a number of potential storage destinations: aquifers, porous geologic formations, depleted oil and gas reservoirs, coal seams, deep ocean ?oor. CO2 capture and storage is a viable solution for CO2 wherever fossil fuels are used as an energy source and opportunities for storage exist [35, p 249]. In Canada, 23 coal ?red plants were used to create 106 TWh of the electricity generated in 2000. Table 1.4 shows the contribution that these coal plants made to the electrical generation capacity in each province. Several technologies are available for capturing CO2 from coal power plants: (a) Chemical absorption with amine solvents (b) O2/CO2 recycle combustion (oxy-fuel) (c) Cryogenics (d) Membrane separation either with or without absorption solvent 2 Energy content of bituminous coal and natural gas used in Ontario during 2000; natural gas assumed to be pure methane with speci?c gravity of 0.585; coal assumed to contain only carbon and hydrogen in ratio of 80:20 6 Table 1.4: Coal use across Canada, 2000 (Source: Statistics Canada [21]) Province Coal generating capacity Percent of in- stalled capacity [MW] [%] Nova Scotia 1 280 55.4 New Brunswick 570 13.6 Ontario 7 767 26.2 Manitoba 220 4.2 Saskatchewan 1 766 53.7 Alberta 5 900 60.1 Canada 17 503 15.7 Of the ?ve CO2-reduction ideas presented above, chemical absorption with amine solvents is the most promising near-term3 mitigation strategy for at least two reasons: 1. Table 1.5 shows the actual age distribution of Canadian coal-?red generating ca- pacity in 1998 and forecasts the 2010 distribution assuming that all of these plants remain in service. There is a substantial investment in coal-?red capacity in Canada and, with a coal-?red power plant having a nominal useful-life of 40 years, this capital stock will be available in the near- to medium term. Amine absorption capitalizes on this investment as it does not require modi?cation of the existing power plant.4 Converting these plants for oxy-fuel combustion or fuel-switching requires the plant boilers be replaced; switching to non-CO2 emitting power plants or IGCC would imply moth-balling the existing equipment. 2. Technology to remove acid gases from relatively dilute, low pressure vapour streams is commercially available. The process is used for natural gas sweetening and to provide a source of CO2 for various industrial processes: food processing, freez- ing, beverage carbonation, chilling, and enhanced oil recovery (EOR (enhanced 3In the more distant medium- and long-term, there is a particularly noteworthy technology which com- bines ideas z, {, and | presented above: SOFC (solid oxide fuel cell)'s. SOFC's, using synthesis gas generated from coal as a fuel source, would generate electricity more ef?ciently than either PCC (pul- verized coal combustion) or IGCC and produce a high-purity CO2 stream that is more or less ready for transportation and storage. Further CO2 mitigation and higher ef?ciency could be achieved by using natu- ral gas in lieu of coal. However, there are outstanding materials and systems issues that need to be resolved before this technology can be implemented on a utility scale. 4While it is true that amine absorption does not require modi?cations to the power plant, this thesis examines the bene?ts of extracting steam from the power plant for use in the CO2 capture plant. So, in this work, it cannot be said that the power plant is entirely left alone. 7 Table 1.5: Age distribution of Canadian coal power plants: 1998 and 2010 (Source: Statistics Canada [21])6 Age of units 1998 capacity 2010 capacity years [MW] [%] [MW] [%] 1–24 8 989 46 2 728 16 25–29 3 404 16 2 632 15 30–34 4 503 25 3 629 21 35–39 212 3 3 404 19 40 394 10 5 110 29 Total 17 503 100 17 503 100 oil recovery)).7 Oxy-fuel is in the demonstration stage only, IGCC technology is commercially available but not with CO2 capture (that part has yet to reach the demonstration stage), and membrane separation requires additional materials re- search and development before it becomes a possibility. 1.2.4 Capturing CO2 with MEA The general process ?ow diagram for amine absorption is shown in Figure 1.2. The fundamental underlying principle is the exothermic, reversible reaction between a weak acid (e.g., CO2) and a weak base (e.g., MEA) to form a soluble salt. The inlet gas is contacted counter-currently with 'lean' solvent in the Absorber. The acid gases are preferentially absorbed by the solution. The solution, 'enriched' with CO2, is pre-heated before entering the Stripper where, through the addition of heat, the reaction is reversed. From the bottom of the column, the 'lean' solvent exchanges heat with the 'rich' solvent entering the column and is recycled back to the Absorber. From the top, a high-purity (dry-basis) CO2 is produced. Large quantities of heat are required by the Stripper reboiler to regenerate the rich solvent; studies have shown that 0.37–1.90 kJ kg CO2 is needed8. For reference, a 500 mathrmMWe unit burning sub-bituminous coal emits about 500 000 kg/hr of CO2. Deciding where this heat is to come from is a fundamental part of the design of an MEA absorption plant. One approach is to include auxiliary heat and, maybe, power generating equipment as part of the design [55, 54, 14, 40]. The other alternative is to 7 That being said, it has never been used for the capture of CO2 on a scale that the wholesale scrubbing of power plant ?ue gas entails. 8see 5.4 for the source of this range and a detailed analysis of the energy requirements. 8 ABSORBER RICH_PUMP STRIPPER HEATX MIXER COOLER FLUE-ABS LEAN-ABS STACK RICH-PUM RICH-HX LEAN-HX RICH-STR LEAN-MIX MAKE-UP LEAN-COO CO2-COMP Figure 1.2: Process ?ow diagram for CO2 removal via chemical absorption extract the required heat from the existing power plant and it is this road "less travelled by" that is the focus of this thesis. Let me be the ?rst (and last?) to say that there is nothing wrong with the auxiliary approach. Doing so essentially obviates to need to modify the existing power plant and it provides ?exibility in determining the post-capture power output of the station (i.e., could increase electricity output to the grid, if so desired). In contrast, integrating the power plant with the capture plant, by extracting steam from the one for use in the other, will probably make the power plant more dif?cult (i.e., costly) to maintain and will de?nitely de-rate the facility. On the bright side, a design where the two units are linked should yield a higher overall thermal ef?ciency which implies a lower CO2 capture cost. The perceived contributions of this work are: 1. Evaluation of current "state-of-the-art" simulation tool for use in modelling MEA absorption processes. 2. Presentation of a methodology that allows one to successfully converge MEA ab- sorption process models that contain recycle streams with no manual intervention. 3. Demonstration of an approach for including pressure calculations in MEA absorp- tion model with a discussion of the bene?ts. 4. Improvement in accuracy of sensitivity analysis due to broadening of the scope of measured process variables to include the energy requirements of all unit opera- tions and not just the Stripper reboiler. 9 1.3 Implementation 1.3.1 Selection of study basis The basis for this work is a 500 MW unit from OPG's Nanticoke Generating Station. The Nanticoke Generating Station consists of eight Babcock and Wilcox units; each one is designed to generate about 3 3 106 lb hr of steam at 2400 psig and 1000? F with re-heat also to 1000? F. Reasons for this site selection are as follows: ? the Nanticoke Generation Station is the largest point source of CO2 emissions in the province of Ontario (1.7 Gt of CO2 emitted in 1999 which represented more than half (53%) of the CO2 emissions from power generation in that year [3]). ? OPG provided funding to support the work and access to operational data not avail- able in the open literature. ? The unit size and coastal location — it is situated on the northern shore of Lake Erie — correspond with the basis for CO2 capture studies chosen by the IEA (Interna- tional Energy Agency) GHG R&D Programme Test Network for CO2 Capture [32]. Therefore, it is expected that the results of this study will have immediate bene?t for that group. ? In contrast to perceived conventional wisdom, it has recently been demonstrated that, within the proximity of Nanticoke, there exists a potential CO2 sequestration reservoir capable of accepting many years worth of CO2 from a 500 MW unit [52]. As noted above, CO2 capture and sequestration as a CO2 mitigation strategy is only worth considering where opportunities for CO2 sequestration exist. Apparently, Nanticoke quali?es. 1.3.2 Selection of simulation software The choice was made to use Aspen Plus? for all process simulation work. At the time that the study began, two generic process simulation software suites were readily available: HYSYS?, marketed by Hyprotech, and Aspen Plus?, developed by Aspen Technology, Inc. The initial decision to use Aspen Plus? over HYSYS? was based upon reported lim- itations of HYSYS? in modelling Absorber and Stripper columns with large numbers of trays [55, p 46]. As the work progressed, other advantages and disadvantages associated with using Aspen Plus? for this work became apparent: 10 ? In May 2002, Aspen Technology, Inc. announced its acquisition of Hyprotech. Given that there was substantial overlap of the Aspen Plus? and HYSYS? product spaces, there was speculation, even among employees, that support for HYSYS? could be discontinued [19]. So, this seemed to reaf?rm the decision to use Aspen Plus? as being correct. ? There are a number of reports of Aspen Plus? being used for modelling amine absorption processes [1, 54, 55, 27, 25, 26, 20] and, also a report of Aspen Plus? being used for modelling a power plant steam cycle [48]. This prior record suggests that Aspen Plus? is suited to the current endeavour. ? Aspen Plus? is updated often. Three major software revisions have been used during this study. This change can be both good and bad: good, in that every new revision brings the promise of improvements and bad, in the sense that many changes occur beneath the threshold sensitivity of the user which can unknowingly cause discontinuities in the results. However, in this work, the software changes did not appear to affect the outcomes of the simulations. ? Aspen Plus? allows the incorporation of almost any arbitrary Fortran code which makes it ?exible and extensible. ? A major disadvantage, though, is that, being proprietary, there is no access to the underlying system design or source code which makes troubleshooting some be- haviour particularly onerous (i.e., requires building test cases to reverse engineer the software). 1.3.3 Outline of thesis Assessing the feasibility of using steam from the power plant to 'fuel' CO2 capture ne- cessitated a number of discrete activities. Each task is presented in its own chapter: ? Chapter 2 describes the development of a simple coal combustion model that esti- mates the resultant heat and ?ue gas production from burning a given quantity and quality of coal. ? Chapter 3 describes the development of a steam cycle model that accurately pre- dicts the power output and steam conditions of the power plant at part-load condi- tions. ? Chapter 4 discusses, in detail, the development of a model of CO2 capture using amine absorption. 11 ? Chapter 5 details the integration of the aforementioned three models to create a uni?ed model of a coal-?red power plant with amine absorption of CO2 where steam extraction from the power plant provides the heat required for capture. The last section, Chapter 6, evaluates the integrated scheme shown in Chapter 5 with scenarios where the additional energy of CO2 capture is provided by an auxiliary power plant. 12 Chapter 2 Flue Gas Synthesis 2.1 Objective The objective is to develop a model that is able to predict the ?ow rate and composition of ?ue gas and heat output for a particular power plant given knowledge about the fuel used, boiler operating conditions, and plant power output. 2.2 Rationale 2.2.1 Model ?exibility Including coal combustion as part of the overall model increases its ?exibility and, thereby, its usefulness. It allows the evaluation of the performance of MEA absorption for non-existent power plants or for fuels which are not currently in use. ? Nanticoke was originally designed to burn a high-sulphur, U.S. bituminous coal but now consumes a mixture of PRB (Powder River Basin) and USLS (U. S. low- sulphur) coals in order to mitigate SOxemissions [4]. In evaluating CO2 capture potential at Nanticoke, one scenario to consider is the return to high-sulphur, U.S. bituminous coal. Since this coal is not currently being used, the characteristics of its ?ue gas need to be estimated. ? The GHG R&D Programme Test Network for CO2 Capture has agreed upon a basis for conducting studies in CO2 capture [32]. The power plant is hypothetical so, indeed, a method for estimating the ?ue gas properties of this plant is required. 13 2.2.2 Model accuracy The accuracy of the combustion model is important as its outputs — ?ue gas composi- tion, ?ue gas ?ow rate, and speci?c heat output — affect the design, performance, and cost of MEA absorption. ? The mass ?ux of a component in the vapour phase can be expressed as the product of a driving force and the appropriate mass transfer coef?cient: N ky ? y∞ yi? The higher the concentration of CO2 in the ?ue gas, the faster it is absorbed by the solvent. Different fossil fuels generate ?ue gases with very different CO2 concen- trations. For example, ?ue gas with 14 mol% CO2 is typical for coal combustion; 8 mol% and 3 mol% CO2 is normal for ?ue gas resulting from the use of natural gas in a natural gas boiler and an NGCC (natural gas combined cycle), respectively. ? There are a several compounds, typically present in ?ue gas, to which MEA ab- sorption is particularly sensitive (e.g., O2, SOx, NOx). To a lesser or greater extent, the abundance of these molecules in the ?ue gas depends upon the composition of the fuel. The impacts on the design and operation of the capture process are many: – Additional pollution control equipment may be required to treat the ?ue gas upstream of MEA absorption. – The concentration of MEA may need to be restricted and/or additives may be required. – Additional make-up MEA may be required. ? The ?ue gas volumetric ?ow rate in?uences both the capital and operating costs of the MEA absorption process. – The volume of ?ue gas will determine the size of the ductwork and, more im- portantly, the size (and number) of Absorber required to capture the desired amount of CO2. – A Blower is required to push the ?ue gas through any and all pollution control equipment upstream of the MEA absorption process and to overcome the pressure drop in the Absorber. The volume of ?ue gas will determine the work duty of this equipment. 14 HTRANS SEPARATE COAL-IN AIR FLUE-GAS SOLIDS Q-DECOMP Q-FURN COAL-OUT IN-BURN EXHAUST DECOMP BURN Figure 2.1: Coal combustion simulation ?owsheet 2.3 Implementation The synthesis of the Aspen Plus? input ?le draws heavily from the example Modelling Coal Combustion included in the systems documentation [9, pp 3-1–3-23]. The simula- tion ?owsheet is given in Figure 2.1. Below are discussed the areas where the Aspen Plus? model development differs from the example problem. 2.3.1 Specifying properties Property Data The ultimate, proximate, and sulphur analyses is provided for the three coals of immedi- ate interest. Two of these coals are used at Nanticoke Power Generating Station [4] (i.e., PRB and USLS) and the last is speci?ed by the IEA for use in CO2 mitigation studies [32]. The characteristics of these coals are given in Table 2.1. In the case of PRB and USLS, with the absence of full analysis, component ash is speci?ed as a 'very poor coal' (i.e., coal with 100 wt% ash). The more rigorous approach is used for the IEA coal where the ash constituents are speci?ed and the ENTHGEN and DNSTYGEN property methods are used to calculate its enthalpy and density. Property Methods The property method is changed from IDEAL to PR-BM. PR-BM is recommended for coal combustion applications [11]. 15 Table 2.1: Coal characteristics Units PRB USLS IEA Proximate analysis (dry): Moisture % 28.1 7.5 9.5 Volatiles % 42.92 33.69 N/A Ash % 7.13 10.36 13.5 Fixed carbon % 49.95 55.95 N/A Ultimate analysis (dry): Carbon % 69.4 77.2 71.4 Hydrogen % 4.9 4.9 4.8 Nitrogen % 1.0 1.5 1.6 Sulphur % 0.4 1.0 1.0 Oxygen % 17.2 5.0 7.8 Ash % 7.1 10.4 13.5 High heating value: Dry kJ/kg 27637 31768 As ?red kJ/kg 19912 29385 Calori?c value: Gross MJ/kg 27.06 Net MJ/kg 25.87 16 Careful consideration is given to the manner in which enthalpy calculations for coal are handled in Aspen Plus?. Speci?c enthalpy of a coal is given by h ?f h T 298K Cp dT The Aspen Plus? coal enthalpy model is called HCOALGEN and its four option codes specify how enthalpy is calculated. 1. In HCOALGEN, heat of combustion is a GCV (gross calori?c value), is expressed in Btu/lb of coal on a dry, mineral-matter-free basis, and is controlled by the ?rst option code. There are ?ve correlations in Aspen Plus? for the calculation of ? ?ch? d m plus the ability for a user to specify ? ?ch? d directly. 2. The second option code selects one of two correlations for calculating the heat of formation, ?f h; the ?rst calculates heat of formation directly from the coal analyses and the other is based on the heat of combustion. The heat of combustion correlation assumes that combustion results in complete oxidation of all of the elements except for sulphatic sulphur and ash. The numer- ical coef?cients are combinations of stoichiometric coef?cients and the heats of formation of CO2, H2O, NO2, and HCl at 298.15 K. ?f h ? ?ch? d ? 1 418 106 wd H 3 278 105 wd C 9 264 104 wd S 2 418 106 wd N 1 426 104 wd Cl? 102 3. There are two correlations for calculating the heat capacity and these are selected via the third option code. ? The Kirov correlation identi?es ?ve coal constituents — moisture, ash, ?xed carbon, and primary and secondary volatile matter — and calculates the heat capacity as a weighted sum of cubic equations for each constituent. ? The second correlation is a cubic temperature equation with parameters re- gressed from data for three lignite and one bituminous coal. 4. The remaining option code in HCOALGEN allows the user to specify the enthalpy basis. Aspen Plus? can be instructed to use either: 17 ? elements in their standard states at 298.15 K and 1 atm or ? the component at 298.15 K. The Heat of Combustion approach is used to calculate ?f h and values of ? ?ch? d are entered directly, The Kirov correlation is used to calculate the heat capacity because it takes into account the coal analyses whereas the cubic equation correlation does not, and, ?nally, the enthalpy basis used is that of the component at 298.15 K (i.e., option code '6111'). 2.3.2 Specifying streams The ?owsheet has two inputs: AIR and COAL-IN. ? The composition of AIR is taken from literature [18, p 653] and is nominally 78% N2, 21% O2, and 1% Ar. AIR ?ow rate is calculated such that there is 21% excess O2 "in the ?ame". AIR temperature is set to the outlet temperature of the air from the secondary air heater and atmospheric pressure is used. ? COAL-IN composition is given by specifying the relative abundance of each type of coal. COAL-IN ?ow rate is set such that the target heat duty, Q!FURN , is achieved. As an example, Q!FURN can be calculated from the plant power output and overall ef?ciency: Q!FURN Etrans ηth plant COAL-IN temperature is set to the pulverizer outlet temperature and, again, at- mospheric pressure is assumed. 2.3.3 Specifying blocks HTRANS is modelled with the HEATER UOM (unit operation model) and is inserted between BURN and SEPARATE. This block removes from the combustion gases the useful heat transfered to the steam cycle. The temperature of the block is equivalent to the ?ue gas temperature at the economizer outlet. 18 2.4 Model Validation 2.4.1 Coal heat of combustion The standard heat of combustion is determined for three different coals whose properties are given in Table 2.1 using the simulation ?owsheet shown in Figure 2.1. A coal's standard heat of combustion, ? ?ch? ? , should be approximately equal to its NCV (net calori?c value). The NCV of the PRB and USLS coals is not available but can be calculated from the GCV by making an adjustment for pressure and the latent heat of vaporization of water [5]: NCV GCV 215 5 " J g#%$ wH The NCV of IEA coal is reported on a dry basis. This converted to an "as ?red" number via: NCV NCVd $ ? 1 wH2O? Table 2.2 compares the heat of combustion from the simulations with data obtained experimentally. Aspen Plus? calculates a heat of combustion which is slightly greater than the corresponding NCV. Table 2.2: Comparison of calculated standard heat of combustion with observed NCV Units USLS PRB IEA NCV kJ/kg 28149 18480 23412 ? ?ch?&? kJ/kg 28710 19535 24112 ? % 2.0 5.7 3.0 2.4.2 Flue gas ?ow rate The ?ue gas mass and volumetric ?ow rates at the economizer exit from a unit at Nan- ticoke Power Generating Station burning a 50/50 blend of PRB and USLS coals are estimated. The input data values and sources are shown in Table 2.3. Table 2.4 compares the ?ue gas ?ow rate from the simulation with observed values. The estimated mass and volumetric ?ow rates are moderately higher and lower, respec- tively, than what is observed at the plant. 19 Table 2.3: Flue gas ?ow rate simulation input data Units Value Source Overall plant: Egen kW 507611 [30] ηth plant % 36 [51] Q!FURN 106 Btu hr 4816 Streams: TAIR ? F 519 [31] TCOAL-IN ? F 160 [31] Blocks: THTRANS ? C 320 [4] Table 2.4: Comparison of calculated ?ue gas ?ow rate with observed values Mass Volumetric ' kg( hr) ' m3( hr) Actual 2424400 4182700 Simulated 2500291 4081180 ? 3.0 % -2.5 % 20 2.5 Conclusions and Recommendations ? The combustion model reasonably predicts the ?ue gas ?ow rate and heat output from a power plant boiler. ? For a 50/50 blend of PRB and USLS coals, Table 2.5 shows the ?ue gas compo- sition. Table 2.5: Flue gas composition Component mol % N2 72.86 CO2 13.58 H2O 8.18 O2 3.54 Ar 0.87 NO 0.50 CO 0.37 SO2 0.05 H2 0.04 21 Chapter 3 Simulation of Steam Cycle 3.1 Objective The objective is to develop a model that simulates the part-load performance of the steam cycle of a 500 MW unit at OPG's Nanticoke Generating Station. That is, to create a model that predicts the required heat input to the boiler, power output from the turbine, and conditions (i.e., temperature, pressure, ?ow rate) of steam and feed water throughout the steam cycle. 3.2 Motivation Including the steam cycle as part of the overall model increases its ?exibility and, thereby, its usefulness. It allows for the evaluation of the performance of MEA absorption when the power plant is operating at part-load and the exploration of different process integra- tion con?gurations. ? For a number of reasons (e.g., technical problems, desire to maintain reserve ca- pacity, lack of demand), plants operate at loads other than their MCR (maximum continuous rating). The effect of plant load on CO2 capture using MEA absorption can be studied. ? The heat and work duties of the MEA absorption process are considerable. It may be economically desirable for the large work and heat duties of the MEA absorption process to be reduced or paid for through process integration: 22 – use steam to provide heat for Stripper reboiler – use super-heat from steam destined for Stripper reboiler to pre-heat "rich" solvent – use steam to provide motive power for ?ue gas blower and CO2 compressors – use boiler feed water for cooling in-between CO2 compression stages ? Speaking strictly from the point of view of the steam cycle, process integration con?gurations differ from one another in terms of the location from which steam is extracted and, to a lesser extent, the position at which the condensate is re- injected. Each potential extraction location provides access to steam at a different temperature, pressure, and ?ow potential (i.e., limit to the quantity of ?uid that can be removed). Similarly, except for maybe the main and re-heat steam temperatures, changing plant load also changes the steam temperature and pressure throughout the process. The variations in steam quality affect the quantity of steam that needs to be diverted to the Stripper reboiler in order to satisfy a given heat duty which, in turn, increases or reduces the power output from the plant. 3.3 Points of emphasis The accuracy of this steam cycle model is important as its outputs — heat input to the boiler, power output from the turbine, and conditions of steam and feed water — affect the performance and cost of MEA absorption. ? As the plant load decreases, so too does ηth plant. More coal is required to produce each unit of power and, consequently, the quantity of ?ue gas emitted per unit of electricity produced increases. This will drive the speci?c cost of capture (i.e., cost per unit mass of CO2) upwards as the ?ue gas volumetric ?ow rate in?uences both the capital and operating costs of the MEA absorption process. ? Accurate plant power output estimation increases the con?dence with which the following two questions can be answered: 1. How much will MEA absorption de-rate the plant? 2. How does MEA absorption compare with other mitigation options? 23 3.4 Implementation OPG provided design heat balance of Nanticoke Generating Stations units 1–4 at 100%, 75%, and 50% load each of which displays the stream and equipment connectivity and provides the following information: ? for each stream, the mass ?ow rate and the temperature, pressure, and/or speci?c enthalpy.1 ? for each feed water pre-heater, the TTD (terminal temperature difference) and the DTD (drain temperture difference) ? the turbine and unit heat rate ? the main turbine Sankey diagram The simulation ?owsheet is shown in Figure 3.1. With the following notable excep- tions, it reproduces the ?ow diagram in the design heat balance: ? streams with ?ow rates less than 10000 lb/hr, except for ST-FPT1, are ignored ? pressure drop across piping and feed water pre-heaters is ignored ? packing and valve stem leakages are ignored The development of the Aspen Plus? input ?le is discussed below. 3.4.1 Specifying properties There are two property methods within Aspen Plus? indicated for use for steam cycle simulation: STEAM-TA and STEAMNBS. STEAM-TA is based upon 1967 ASME steam table correlations. STEAMNBS is based upon 1984 NBS (National Bureau of Standards)/NRC (National Research Council) steam table correlations and is reportedly the more accurate of the two. In spite of its purported inferiority, STEAM-TA is used as it more closely matches the 1936 Keenan and Keyes steam tables upon which the original design is based. 1In general, the speci?c enthalpy and only one of temperature and pressure are speci?ed for any given stream in the heat design balance. Using a software implementation of the ASME (American Society of Mechanical Engineers) 1967 steam tables [41], temperature, pressure, speci?c enthalpy, speci?c entropy, and speci?c volume are calculated for each stream at each plant load, where missing. The conditions of important streams are given in Appendix A. 24 HP_SEP1 VALVE1 HP1 HP_SEP2 REHT IP_SEP1 IP2 IP1 IP_SEP2 IP3 IP4 IP_SEP3 IP_SEP4 IP_SEP5 IP_COMB VALVE2 LP_SEP1 LP_SEP2 LP1 LP2 LP_SEP3 LP3 LP_SEP4 LP6 LP5 LP_SEP5 LP4 LP_COMB1 LP_COMB2 FWPUMP2 FWP_C FWPUMP1 CONDENSE CND_COMB FPT1 FPT2 FPT_COMB BOIL H2O-BOIL ST_MAIN ST-FPT1 ST-HP HP_1X ST-REHT ST-FWPA ST-IPX ST-HPX ST-IP IP_03 IP_02 IP_2X IP_12 ST-FWPC IP-1LP IP_3X1 IP_34 IP_3X2 ST-FPT2 ST-FWPB IP-4LP ST-FWPD ST-LP LP_056 LP_012 LP_02 LP_01 ST-FWPF LP_2X LP_23 ST-2FWPG LP_3CR LP_06 LP_05 ST-FWPE LP_5X LP_45 ST-5FWPG LP_4CR ST-CNDR ST-FWPG H2O-FWPA STFWP_AB STFWP_BC H2O-FWPB FPT_12 H2O-FWPC H2O-FWPD ST-FWPE STFWP_EF H2O-FWPE H2O-FWPF STFWP_GC H2O-FWPG H2O-MAIN H2O-CNDR H2-PUMP ST-FPT1 IP_4X STFPT-CN IN-PUMP FPT_1X STFWP_FG STFWP_DE Q_FWPA Q_FWPB Q_FWPD Q_FWPE Q_FWPF Q_FWPG FWP_A FWP_B FWP_D FWP_E FWP_F FWP_G Figure 3.1: Steam cycle simulation ?owsheet 25 3.4.2 Specifying streams Plant load is controlled by changing the ?ow rate of the boiler feed water, H2O-BOIL. This is the only stream speci?ed in the input ?le and it is initialized using values for the design heat balance at 100% load: Units Value T ? F 488 P psia 2700 ˙ m lb/hr 3358670 3.4.3 Specifying blocks The steam cycle model has four sections: 1. main and boiler feed water turbines 2. condenser 3. boiler feed water pre-heaters 4. economizer, boiler, super-heater, and re-heater The speci?cation of each of these sections is discussed in turn. Main and feed water pump turbines The main turbine drives the generator producing electrical power for plant consumption and output to the grid. The other, smaller turbine drives the boiler feed water pumps. As is done elsewhere [23, 17, 48], each turbine is modelled as a series of single turbine stages interspersed with ?ow mixers and splitters as indicated by the ?ow path in the heat design balance [30]. Table 3.1 shows the volumetric ?ow rate of steam entering the HP (high-pressure), IP (intermediate-pressure), and LP (low-pressure) sections of the turbine. The steam pressure is throttled at part-load to maintain a constant ?ow rate into the HP and IP sections. This behaviour is emulated using VALVE1 and VALVE2. It is expected that, at part-load, the pressure ratios of stages between the governing stages and the last stage will be approximately constant [23]. This fact is borne out by 26 Table 3.1: Design inlet volumetric ?ow rates into turbine sections (106 ft3 hr) Section Plant Load Mean Std Dev %RSD 100% 75% 50% HP 1 157 1 155 1 154 1 155 0 001 0 13 IP 4 522 4 530 4 541 4 531 0 009 0 21 LP 20 724 20 784 20 828 20 779 0 052 0 25 the data in Table 3.2 which shows the ratio of outlet pressure to inlet pressure for each of the compressor stage groups in Figure 3.1. For turbine groups where constant pressure ratio is not observed, other criteria is used for specifying the outlet pressure: ? The pressure ratio of HP is calculated using a function of the form: Pout Pin a ˙ min b Figure 3.2 gives "least-squares" estimates of the parameters a and b and shows that the proposed model does a good job at explaining the variation in pressure ratio at part-load. Pout Pin 4 82 10 ¤ 3 ˙ min 0 2944 HP section inlet mass ?ow rate / 106 lb hr HP pressure ratio 3 4 3 2 3 0 2 8 2 6 2 4 2 2 2 0 1 8 1 6 0 287 0 286 0 285 0 284 0 283 0 282 0 281 0 280 0 279 0 278 Figure 3.2: High-pressure section pressure ratio at part-load 27 Table 3.2: Ratio of discharge pressure to inlet pressure for turbine groups Main turbine Block Plant Load Mean Std Dev %RSD 100% 75% 50% HP 0 278 0 282 0 287 0 282 0 004 1 48 IP1 0 515 0 517 0 519 0 517 0 002 0 43 IP2 0 231 0 233 0 235 0 233 0 002 0 85 IP3 0 453 0 455 0 457 0 455 0 002 0 37 IP4 0 262 0 265 0 267 0 265 0 002 0 91 LP1 0 151 0 150 0 152 0 151 0 001 0 44 LP2 0 068 0 067 0 069 0 068 0 001 1 22 LP3 0 153 0 146 0 210 0 170 0 035 20 63 LP4 0 153 0 146 0 210 0 170 0 035 20 63 LP5 0 068 0 067 0 069 0 068 0 001 1 22 LP6 0 433 0 435 0 437 0 435 0 002 0 46 FP turbine Block Plant Load Mean Std Dev %RSD 100% 75% 50% FPT1 0 107 0 080 0 054 0 080 0 027 33 34 FPT2 0 003 0 003 0 004 0 003 0 001 22 79 28 ? The outlet pressure of FPT1 is set equal to that of the ST-FPT2. ? The outlet pressure of FPT2, LP3, and LP4 is set equal to that of the Condenser. Given the constant volumetric ?ow rates and stage pressure ratios, it is expected that the isentropic ef?ciencies between the governing stage and the last stage stay about the same at part load [59]. Table 3.3 shows that this is indeed the case for all turbine groups except LP3, LP4, and FPT1. These blocks require special consideration. Table 3.3: Fractional isentropic ef?ciency of turbine groups Main turbine Block Plant Load Mean Std Dev %RSD 100% 75% 50% HP 0 906 0 903 0 903 0 904 0 002 0 19 IP1 0 901 0 902 0 904 0 902 0 002 0 18 IP2 0 910 0 910 0 910 0 910 0 000 0 04 IP3 0 891 0 898 0 898 0 895 0 004 0 48 IP4 0 915 0 912 0 914 0 914 0 002 0 17 LP1 0 910 0 910 0 911 0 910 0 000 0 02 LP2 0 904 0 907 0 909 0 907 0 003 0 30 LP3 0 607 0 598 0 715 0 640 0 065 10 13 LP4 0 607 0 598 0 715 0 640 0 065 10 13 LP5 0 904 0 907 0 909 0 907 0 003 0 30 LP6 0 902 0 901 0 898 0 901 0 002 0 26 FP turbine Block Plant Load Mean Std Dev %RSD 100% 75% 50% FPT1 0 182 0 153 0 126 0 153 0 028 18 22 FPT2 0 801 0 786 0 798 0 795 0 008 1 02 ? The variation in the isentropic ef?ciency of FPT1 is ignored as its low magnitude, even at base load, coupled with the low mass ?ow rate of steam through this part of the turbine makes its contribution to the overall feed water pump turbine output negligible. Therefore, the mean value is used in all cases. 29 ? The ef?ciency of the last stage of a turbine is mostly dependent upon the annulus velocity [59]. Therefore, a model of the form ηs aqout b is proposed to describe the part-load behaviour of LP3 and LP4. Figure 3.3 gives "least-squares" estimates of the parameters a and b and compares the proposed model with the data from the design heat balance. The model does a fantastic job at explaining the variation in the isentropic ef?ciency of LP3 and LP4. ηs 0 4016qout 0 9867 LP section exit volumetric ?ow rate / 109 ft3 hr LP3 , LP4 isentropic ef?ciency 1 00 0 95 0 90 0 85 0 80 0 75 0 70 0 65 0 72 0 70 0 68 0 66 0 64 0 62 0 60 0 58 Figure 3.3: LP3 and LP4 stage groups' isentropic ef?ciency at part-load The ?nal aspect of turbine behaviour that needs the be addressed is the bleed steam mass ?ow rates at part-load. Steam is extracted from the main turbine to drive the boiler feed water pump turbine and to pre-heat the boiler feed water. It is proposed that the bleed steam mass ?ow rates vary as a function of the steam mass ?ow rate at the inlet of the turbine section: ˙ mbleed a ˙ min b 30 "Least-squares" estimates of the parameters a and b are obtained for bleed stream and the proposed model is compared with data from the design heat balance in Figure 3.4. The proposed model explains essentially all of the variation in the bleed steam ?ow rates at part-load. Steam mass ?ow rate at turbine section inlet / 106 lb0 hr Bleed steam mass ?ow rate / lb 1 hr 32 5 32 0 22 5 22 0 12 5 12 0 350000 300000 250000 200000 150000 100000 50000 0 a b [103 2] [103 lb4 hr] FWP A 12.30 -78.94 FWP B 5.39 -16.85 FWP C 5.10 -24.40 FWP D 5.24 -20.77 FWP E 6.31 -22.28 FWP F 4.16 -14.75 FWP G 6.17 -25.38 FPT1 7.00 FPT2 2.68 1.95 Figure 3.4: Turbine 'bleed' steam ?ow rates at part-load Feed water pre-heaters The feed water pre-heater section contains seven feed water pre-heaters, numbered A through G, and two pumps. Increasing the temperature of the boiler feed water increases the overall thermal ef?ciency of the power plant. Six of the feed water pre-heaters — A, B, D, E, F, G — are closed and the other, C, is open and also functions as a deaerator. The closed feed water pre-heaters are shell and tube heat exchangers. These units are usually modelled in Aspen Plus? using HEATX UOM's however, all attempts in this work to represent the feed water pre-heater section using HEATX UOM's met with fail- ure. It is believed that this dif?culty could have been overcome if detailed heat exchanger design information or design heat balances at additional plant loads were available. That not being the case, the example of Ong'iro et al. [48], where each closed feed water pre- heater is modelled as a pair of HEATER blocks, is instead followed. The open feed water heater is modelled using the UOM MIXER. For any regenerative cycle, the temperature to which the feed water is raised is a design variable that is ultimately ?xed by economic considerations. This is also true of the temperature rise that is to be accomplished by each pre-heater [57, p 290]. The 31 'cold-side' exit temperatures are found to vary with plant load according to the following model: Tc out a ln ˙ mc in b "Least-squares" estimates of the parameters a and b are calculated for each feed water pre-heater and the proposed model is compared with data from the design heat balance in Figure 3.5. All of the variation observed in the data is explained by the model. Feed water inlet mass ?ow rate / 106 lb0 hr Feed water outlet temperature / 5 F 32 4 32 2 32 0 22 8 22 6 22 4 22 2 22 0 12 8 12 6 12 4 12 2 500 450 400 350 300 250 200 150 100 a b [6 F] [6 F] FWP A 85.46 -796.3 FWP B 68.40 -627.2 FWP C 64.68 -621.2 FWP D 55.37 -527.4 FWP E 46.02 -440.5 FWP F 37.88 -375.2 FWP G 30.33 -299.6 Figure 3.5: Boiler feed water temperature at part-load The two feed water pumps are modelled using PUMP blocks. For FWPUMP1, in the absence of information in the design heat balance, the outlet pressure is selected such that it is marginally greater than that of the open feed water pre-heater, FWP C. For FWPUMP2, the internal shaft power of the Feed water pump turbine is used as the power input to the pump. In both cases, ef?ciency is calculated using ef?ciency curves for water in a centrifugal pump [10] . Condenser The condenser is modelled using a HEATER UOM. In the design heat balance, the condenser pressure is 1.4787 Hg at base load and 1.0797 Hg at both 75% and 50% load. It is not clear how condenser pressure changes with plant load. Therefore, the given value at base load is used at part-load. 32 Economizer, boiler, super-heater, and re-heater The economizer, boiler, and super-heater are represented using a single HEATER bock, BOIL, with outlet temperature and pressure of 1000? F and 2365 psia, respectively. The re-heater is represented by the HEATER block REHT with an outlet temperature of 1000? F and zero pressure drop. 3.5 Model validation 3.5.1 Property method The basis for the Nanticoke Generating Station design heat balance is the 1936 Keenan and Keyes steam tables; note that the property methods in Aspen Plus? are based upon either the 1967 ASME or the 1984 NBS/NRC steam tables. It would be expected that changes in the underlying property data will cause changes in some of the calculated performance values. Tables 3.4 and 3.5 show the results of simulations performed with the two property methods alongside the data from the design heat balance. In these simulations, the VALVE and COMPR outlet pressures, COMPR isentropic ef?ciencies, and FSPLIT outlet ?ow rates are set using data directly from the design heat balance. That is, none of the correlations or assumptions presented above are used. As such, any differences observed between the simulation results and the design data result from differences in steam properties. Several conclusions can be drawn. ? The similarity between the observed and calculated values suggests that there were no gross errors in the transcription of the data. ? STEAM-TA is slightly better at reproducing internal power and STEAMNBS is slightly better at reproducing heat input. That being said, either property method is suitable for steam cycle modelling. Table 3.4: Comparison of calculated internal power with design values (MW) Plant Load Design data STEAM-TA STEAMNBS theoretical calculated % diff calculated % diff 100% 516.97 517.73 0.15 518.37 0.27 75% 382.06 382.63 0.15 383.07 0.26 50% 255.04 256.31 0.50 256.57 0.60 33 Table 3.5: Comparison of calculated heat input with design values (MW) Plant Load Design data STEAM-TA STEAMNBS theoretical calculated % diff calculated % diff 100% 3919 3914 0.12 3917 0.03 75% 2938 2936 0.06 2939 0.02 50% 2016 2015 0.04 2016 0.02 3.5.2 Steam temperature, pressure, and ?ow potential There are several locations along the turbine where it is feasible to extract steam for process use: ? at the inlet of the HP, IP, and LP sections ? at the turbine outlet ? at locations where steam is already extracted for feed water pre-heating These locations are highlighted in the schematic of the turbine shown shown in Fig- ure 3.6. Predicted steam conditions and ?ow rate at part-load at these key locations are compared to the design heat balance data in Figures 3.7, 3.8, and 3.92. In all cases, the model successfully describes the changes in steam conditions and ?ow rate at part-load. 3.5.3 Part-load power output and heat input The internal power and heat input are estimated given boiler feed water ?ow rates from 1 6 106 to 3 4 106 lb/hr. This range covers plant performance from 50% to 100% of base-load. The simulation results are compared with the data from the design heat balance in Figures 3.10 and 3.11. Both in terms of power output and heat duty, the agreement between the model and the design data is very good. 3.5.4 Turbine and unit heat rate The relationship between the turbine internal power, the electrical output to the grid, and the associated losses that occur along the way are illustrated in Figure 3.12. 2In the case of locations A through G, the ?ow rate shown is that which is available at the particular location and not necessarily the amount that is extracted for feed water pre-heating. 34 IP LP A C B D F G CNDR E HP high pressure intermediate pressure low pressure low pressure Figure 3.6: Potential steam extraction locations in steam cycle CNDR G F E D, LP C B IP A HP Plant load Steam temperature / @ F 100% 90% 80% 70% 60% 50% 1100 0 1000 0 900 0 800 0 700 0 600 0 500 0 400 0 300 0 200 0 100 0 0 0 Figure 3.7: Steam temperature at part-load 35 A HP Plant load Steam pressure / psia 100% 90% 80% 70% 60% 50% 2400A 0 2200A 0 2000A 0 1800A 0 1600A 0 1400A 0 1200A 0 1000A 0 800A 0 600A 0 400A 0 200A 0 D, LP C B IP Plant load Steam pressure / psia 100% 90% 80% 70% 60% 50% 600A 0 500A 0 400A 0 300A 0 200A 0 100A 0 0A 0 CNDR G F E Plant load Steam pressure / psia 100% 90% 80% 70% 60% 50% 30A 0 25A 0 20A 0 15A 0 10A 0 5A 0 0A 0 Figure 3.8: Steam pressure at part-load 36 CNDR G F E LP D B, C IP HP, A Plant load Steam mass ?ow rate / 10 6 lb B hr 100% 90% 80% 70% 60% 50% 3 5 3 0 2 5 2 0 1 5 1 0 0 5 0 0 Figure 3.9: Steam ?ow rate at part-load Boiler feed water mass ?owrate / 106 lb hr Main and BFP turbine internal power / MW 3 4 3 2 3 0 2 8 2 6 2 4 2 2 2 0 1 8 1 6 550 500 450 400 350 300 250 Figure 3.10: Turbine power output at part-load 37 Boiler feed water mass ?owrate / 106 lb hr Heat input / MW 3 4 3 2 3 0 2 8 2 6 2 4 2 2 2 0 1 8 1 6 4000 3800 3600 3400 3200 3000 2800 2600 2400 2200 2000 1800 Figure 3.11: Turbine heat duty at part-load PSfrag replacements TURBINE GENERATOR EXCITER Pgross Eexciter Egen Etrans ? ?P ? mech ? ?P ? gen ? ?E? exciter ? ?E? station Figure 3.12: Main turbine Sankey diagram 38 Heat rate is an expression of the ef?ciency with which the internal power generated by the turbine is transformed into electrical energy. There are two "heat rates" given in the power plant design heat balances: THR (turbine heat rate) and UHR (unit heat rate). These can be calculated using the following expressions: THR ˙ mBOIL ? hBOIL out hBOIL in?C ˙ mREHT ? hREHT out hREHT in? Egen Pbfpt net UHR ˙ mBOIL ? hBOIL out hBOIL in?D ˙ mREHT ? hREHT out hREHT in? Etrans ηb th Values for each of the aforementioned power 'adjustments' are available in the in- cluded Sankey diagram. These are plotted versus plant load in Figures 3.13 and 3.14.3 Models proposed for each factor, parameters regressed from the data, and the output from the models shown as straight lines in Figures 3.13 and 3.14. The agreement is very good except, perhaps, in the case of the Boiler feed water pump turbine mechanical losses. However, given the small magnitude of these losses, the effect on THR and UHR calculation is negligible. Finally, using the above correlations, the turbine and unit heat rates, as a function of plant load, are calculated using the results from the Aspen Plus? steam cycle model. These results are compared to that offered in the design heat balances in Figure 3.15. As has come to be expected, the agreement between the design heat balance data and the results from the Aspen Plus? model is excellent. 3.6 Conclusions and recommendations ? The steam cycle model successfully predicts the part-load performance of the steam cycle of a 500 MW unit at OPG's Nanticoke Generating Station. ? The performance of the model at part-loads below 50% needs to be validated. 3 In Figures 3.13 through 3.15, the points represent data taken from the heat design balances and the lines represent simulation results. 39 ? ?E? station 11 10 exp ? Pgross 1000? 3 737 Eexciter 3 437 10 ¤ 3 Pgross 0 4078 ? ?P? gen 1 511 10 ¤ 2 Pgross 0 7343 ? ?P? mech 1 919 Main turbine internal power / MW Energy ?ow / MW 550 500 450 400 350 300 250 16 0 14 0 12 0 10 0 8 0 6 0 4 0 2 0 0 0 2 0 Figure 3.13: Main turbine work and energy ?ows ? ?P ? mech 4 534Pbfpt gross 42 44 design heat balance BFP turbine internal power / MW Mechanical power loss / kW 10 0 9 5 9 0 8 5 8 0 7 5 7 0 6 5 6 0 5 5 5 0 90 0 85 0 80 0 75 0 70 0 65 0 Figure 3.14: Boiler feed water pump turbine mechanical power losses 40 turbine HR unit HR Plant load Heat rate / BTU/kWh 100% 90% 80% 70% 60% 50% 9600 0 9400 0 9200 0 9000 0 8800 0 8600 0 8400 0 8200 0 8000 0 7800 0 7600 0 7400 0 Figure 3.15: Turbine and unit heat rate at part-load ? The model should be extended such that it is able to predict off-spec performance of the steam cycle (e.g., plant performance with one or more feed water pre-heaters off-line). This would allow the investigation of more complicated process integra- tion con?gurations. 41 Chapter 4 Simulation of MEA Absorption Process 4.1 Objective The objective of the work in this chapter is to develop an adaptable model that simulates the removal of CO2 from power plant ?ue gas using MEA absorption. In particular, the model should report the work and heat duties required to achieve a particular recovery of CO2 given a set of nominal equipment speci?cations and operating conditions. 4.2 Motivation Having a detailed, adaptable model of MEA absorption increases the ?exibility of the overall model and, thereby, its usefulness. It allows for the measurement of the sen- sitivity of the work and heat duties to changes in the process ?owsheet, the design of key equipment, the choice of solvent, and the nominal operating conditions. A detailed model also increases the number of process integration scenarios that can be examined. 4.2.1 Process ?owsheet evaluation ? For modelling CO2 capture from ?ue gas, the MEA absorption ?owsheet shown in Figure 4.1 is the one most frequently reported as being used [54, 55, 25, 26, 27, 20].1 1Singh et al. [54, 55] and Freguia et al. [25, 26, 27] do not close the recycle loop in their simulation ?ow sheets. 42 ABSORBER RICH_PUMP STRIPPER HEATX MIXER COOLER FLUE-ABS LEAN-ABS STACK RICH-PUM RICH-HX LEAN-HX RICH-STR LEAN-MIX MAKE-UP LEAN-COO CO2-COMP Figure 4.1: Base MEA absorption process ?owsheet ? When the Absorber operates at pressures greater than atmospheric, it makes sense to ?ash the rich solvent exiting the absorber. This is the case in UOP (Universal Oil Products LLC)'s Amine Guard FS? system (shown in Figure 4.2). ABSORBER RICH_PUMP STRIPPER HEATX MIXER COOLER FLUE-ABS LEAN-ABS STACK RICH-PUM RICH-HX LEAN-HX RICH-STR LEAN-MIX MAKE-UP LEAN-COO CO2-COMP FLASH LIQUID VAPOUR Figure 4.2: Amine Guard FS? process ?owsheet ? The ?owsheet of the Kerr-McGee/ABB (Asea Brown Boveri Ltd.) Lummus Global MEA absorption process is traditionally the same as that shown in Figure 4.1. However, Kerr-McGee/ABB Lummus Global now uses an "energy saving design" for new installations for their CO2 recovery system [13] and this modi?ed process ?owsheet is shown in Figure 4.3. In this new design, the rich solution is ?ashed af- ter leaving the cross-exchanger; the liquid from the Flash is the feed to the Stripper 43 and the vapours are mixed with the Stripper overhead vapours. ABSORBER RICH_PUMP STRIPPER HEATX MIXER COOLER FLUE-ABS LEAN-ABS STACK RICH-PUM RICH-HX LEAN-HX RICH-STR LEAN-MIX MAKE-UP LEAN-COO CO2-COMP FLASH LIQUID VAPOUR Figure 4.3: Kerr-McGee/Lummus Crest Global MEA absorption process ?owsheet In their examination of CO2 capture cost sensitivity to solvent type, concentration, and ?ow rate and to the number of trays in each of the Absorber and Stripper, the simulation ?owsheet description of Chakma et al. matches that shown in Figure 4.3 [14].2 ? Soave and Feliu have demonstrated that, in a distillation tower, reboiler heat duty can be signi?cantly lowered by only heating a fraction of the Stripper feed [58]. This implies that the ?owsheet shown in Figure 4.4 may be preferred vis-` a-vis those previously shown. ? The con?uence of very large ?ue gas ?ow rates, a desire for a high recovery of CO2, and the limits, in terms of diameter, with which separation columns can be constructed results in the necessity of multiple trains of Absorbers and/or Strippers. – Chapel et al., in a review of Fluor Daniel's Econamine FG (?ue gas)? pro- cess [15], state that CO2 capture is limited by absorber size (taken to be a maximum of 12.8 metres for circular cross-section). – In an overview of CO2 capture in Japan [65], Yokoyama mentions that the size of the Absorber dictates the required number of trains in the CO2 capture plant. This is not necessarily a bad thing as multiple trains provide ?exibility in the case of varying plant load. 2Curiously, in the process ?ow sheet shown and referenced by Chakma et al., the aforementioned FLASH unit is not visible. Also, there would be two streams leaving this FLASH unit: one liquid and one vapour. The liquid stream presumably ?ows to the amine-amine heat exchanger but the vapour stream destination is not obvious and is not stated in the article. 44 ABSORBER RICH_PUMP STRIPPER HEATX MIXER COOLER LEAN-ABS STACK RICH-SPL RICH-HX LEAN-HX RICH-HI (COLD) LEAN-MIX MAKE-UP LEAN-COO CO2-COMP FLUE-ABS RICH-LOW (HOT) RICH-PUM Figure 4.4: 'Split feed' MEA absorption process ?owsheet – Singh et al., simulated the recovery of 90% of the CO2 from the ?ue gas of a 400 MW coal-?red power plant using MEA absorption [54]. The ?ue gas is treated in four separate absorption/regeneration column trains with column diameters of approximately 10 metres; single-train would require an absorber with a diameter of 18 metres. – Desideri and Paolucci simulated the recovery of 90% of the CO2 from 350 MW power plants combusting natural gas and coal [20]. Three and four trains, respectively, are used to treat the ?ue gas from the plants. They state that 3000000 m3 hr is the maximum quantity of ?ue gas that can be handled economically in an Absorber. Presumably, the quantity of ?ue gas from the natural gas and coal cases is more than two or three times greater than this single-Absorber maximum, respectively. An adaptable model allows the effect of these changes, and others, to the process ?owsheet, to be easily studied. 4.2.2 Equipment design The column type (e.g., structured or random packing, valve or sieve trays) and the size of the mass transfer region (i.e., height of packing, number of trays) are important de- sign variables. The optimal design is not immediately apparent and involves a tradeoff between cost, availability, and performance. ? David Singh examined the sensitivity of the CO2 loading in the rich solvent stream to the number of stages in the Absorber [55]. 45 ? Freguia and Rochelle examine the relationship between Absorber and Stripper packing height and the reboiler heat duty [26, 27].3 ? One of the three thrusts taken by MHI (Mitsubishi Heavy Industries Ltd.) and KEPCO (Kansai Electric Power Company Inc.) in improving their CO2 recovery system is the development of packing materials with reduced pressure drop. This has led to the development of KP-1, a structured packing, which reduces the size of CO2 absorbers and the horsepower requirements of ?ue gas blowers [46, 43, 42]. ? Aroonwilas et al. examine the difference between selected random and structured packings on CO2 absorption [7]. A ?exible MEA absorption process model allows the performance of the different scenarios to be assessed. 4.2.3 Solvent selection There are a variety of amine-based solvents that are used, or potentially could be used, to capture CO2. ? Fluor Daniel's Econamine FG process uses an inhibited 30 wt% MEA solution. The inhibitor scavenges oxygen which has two bene?ts: allowing the use of carbon steel in construction and preventing oxygen from degrading MEA. The cost of inhibitor is 20% that of the make-up MEA [15]. ? UOP licenses the Amine Guard FS? process for acid gas removal. It makes use of Union Carbide's UCARSOL family of formulated amines. Corrosion inhibitors and quantitative removal of O2 and NOx allow amine concentrations in the range of 25–30 wt% to be used [15]. ? Kerr-McGee/ABB Lummus Global licenses technology for CO2 capture that uses an uninhibited MEA solution of either 15 or 20 wt% [15, 13]. ? MHI and KEPCO have jointly developed a sterically-hindered amine, dubbed KS-1, which has several stated advantages over MEA: lower regeneration temper- ature, lower regeneration energy, non-corrosive to carbon steel in the presence of oxygen up to 130? C, and less prone to degradation [15, 46, 43, 34]. 3The Absorber and Stripper diameters are kept constant. 46 As a follow-up, MHI and KEPCO efforts have yielded a second-generation sol- vent, dubbed KS-2, whose performance is marginally better than that of KS-1 [44]. As part of their research, some 80 different solvents were evaluated. ? Aroonwilas et al. compared the absorption performance of MEA, NaOH, and AMP [7]. ? Chakma et al. evaluated the CO2 absorption performance of aqueous solutions of MEA, DEA, DIPA, DGA, MDEA, and TEA [14]. ? Marion et al. presented an ABB-designed MEA absorption process where an op- timized mixture of MEA and MDEA is used to capture CO2 [38].4 ? Tontiwachwuthikul et al., in a study of the economic feasibility of CO2 capture for use in enhanced oil recovery, assessed the performance of both MEA and AMP [60]. ? Paul Feron, on behalf of TNO, discusses the development of CORAL (CO2- removal absorption liquid) which has the following stated advantages over MEA: stable operation with polyole?n membranes, better oxygen stability, less corro- sive, and has no loss of active component (i.e., does not degrade under operating conditions) [24]. The design of the MEA absorption model should not preclude the evaluation of dif- ferent solvents for use in capturing CO2.5 4.2.4 Optimizing process operating conditions With large heat and work duties at stake, sub-optimal operation of the process is strongly undesirable. ? Freguia and Rochelle examine the relationship between Absorber ? L G? and re- boiler heat duty [26, 27]. This is analogous to examining the relationship between lean solvent loading and reboiler heat duty. They also examine the relationship between Stripper pressure and reboiler heat duty. 4The MEA/MDEA mixture could not be made O2 tolerant. Therefore, oxygen is catalytically removed from the ?ue gas upstream of the Absorber. 5Note: There are two important solvent-related phenomena of particular interest to MEA absorption that a steady-state model cannot directly include: corrosion and solvent degradation. 47 ? Aroonwilas et al. examined the effect of ?ue gas ?ow rate, ?ue gas CO2 concen- tration, solvent ?ow rate, solvent concentration, and Absorber temperature on CO2 absorption [7]. Effects of the recycle stream CO2 loading, Absorber ?ue gas and lean solvent inlet temperature, Absorber vapour outlet pressure, Stripper reboiler pressure, amount of heat exchange between rich and lean solvent streams, Stripper condenser temperature, and CO2 compressor inter-cooling temperature on work and energy duties can be assessed. 4.2.5 Process integration exploration The principle contribution of this thesis is to begin to discern if the cost of CO2 capture can be reduced by integrating the MEA absorption process with the adjacent steam cy- cle (i.e., using power plant to provide MEA absorption process steam, power, and elec- tricity). A prerequisite is a process simulation model which includes all of the stream conditions and the process heat and work duties. 4.3 Points of emphasis ? The recycle loop in the simulation ?owsheet needs to be closed. ? The model should calculate the pressure pro?le of the Absorber and Stripper. Ad- ditionally, the model should assess the hydrodynamic performance of the columns. For whatever reason(s), column pressure pro?le and hydrodynamic performance has been overlooked in previous MEA absorption process simulation work. This neglect is manifest in three ways: x The Absorber and Stripper are speci?ed with constant pressures throughout the columns. y The pressure drop across a column is obviously dependent upon process op- erating conditions, column type, and column con?guration. However, the Absorber and Stripper pressures are never accordingly modi?ed when of these aspects is changed. z The reader is never informed that the column design is explicitly checked for stable and/or feasible operation (e.g. in the case of trayed columns: down- comer ?ooding, downcomer seal, weeping, etc.). 48 Table 4.1 lists references to MEA absorption simulation studies and indicates which of the above items apply. Table 4.1: Hydrodynamic performance neglect matrix6 x y z Chakma et al. [14] E E E Desideri and Paolucci [20] E E E Freguia et al. [25] F E E Freguia et al. [26, 27] N/A E E Singh [55] E E E Singh et al. [54] E F E This model needs to consider the hydrodynamic performance of the Absorber and Stripper. This feature enhances the model by allowing: – assessment of sensitivity of work required by the Blower, H2O Pump, Rich Pump, and CO2 Compressor to process design and operation. – more accurate representation of the Absorber and Stripper pressure pro?les. – provision of additional information regarding the feasibility of particular de- signs and process conditions. 4.4 Implementation There is nothing ingenious in the synthesis of the simulation ?owsheet. To the nomi- nal MEA absorption process ?owsheet from Figure 4.1 is prepended unit operations to precondition the ?ue gas prior to entering the Absorber and appended still more unit op- erations for the preparation of CO2 for transport via pipeline. The ?nal result is shown in Figure 4.5. The development of the Aspen Plus? input ?le is discussed below. 4.4.1 Specifying properties This section of the Aspen Plus? input ?le speci?es the solution chemistry and the prop- erty method or model that is to be used to calculate ?uid transport and thermodynamic properties. 6The checkmarks in the above table indicate that the particular group of authors is 'guilty' of the neglect referenced by the column heading. 49 BLOWER H2O_PUMP DCC CO2_COMP FLUE-BLO FLUE-DCC H2O-PUMP H2O-DCC H2O-OUT FLUE-ABS CO2 ST1 ST2 ST3 ABSORBER RICH_PUMP STRIPPER HEATX MIXER COOLER FLUE-ABS LEAN-ABS STACK RICH-PUM RICH-HX LEAN-HX RICH-STR LEAN-MIX MAKE-UP LEAN-COO CO2-COMP Figure 4.5: MEA absorption simulation ?owsheet The solution chemistry can be represented by equilibrium reactions 4.1 through 4.5. There is one class of property methods, one property model, and several property inserts that are indicated for use in modelling processes containing CO2, MEA, and H2O: the electrolyte NRTL methods, the AMINES property model, and the emea, kemea, mea, and kmea property inserts.7 These are listed and described in Table 4.2.8 2H2O G OH ¤ H3OH (4.1) CO2 2H2O G HCO ¤ 3 H3OH (4.2) HCO ¤ 3 H2O G CO2¤ 3 H3OH (4.3) RNHH3 H2O G RNH2 H3OH (4.4) RNHCOO ¤ H2O G RNH2 HCO ¤ 3 (4.5) 7The Pitzer-based property methods PITZER, PITZ-HG, and B-PITZER are also indicated for use for aqueous electrolyte solutions. Unfortunately, the Aspen Physical Property System does not contain interaction parameters involving MEA, CO2, or their derivatives. 8A complete description of these entities can be found in the software documentation [8]. 50 Table 4.2: Property methods and model available for CO2-MEA-H2O system Name Description ELECNRTL ENRTL-HG ENRTL-HF ? NRTL-RK method extended to accommodate interactions with ions in solution. Aspen Physical Property System contains binary and pair inter- action parameters and chemical equilibrium constants for systems con- taining CO2, H2S, MEA, and H2O with temperatures up to 120? C and amine concentrations up to 50 wt%.9 ? ENRTL-HF uses the "HF" EOS (equation of state) to calculate vapour phase fugacity whereas ELECNRTL uses Redlich-Kwong. "HF" EOS is able to account for the association (principally hexamerization) that occurs between HFmolecules at low pressure in the vapour phase. ? The "HG" variant differs from ELECNRTL in that it uses the Helgeson model to very accurately and ?exibly calculate standard enthalpy, entropy, Gibbs free energy, and volume for components in aqueous solutions. This adjustment improves the accuracy at high temperatures and pressures. AMINES ? This property model is valid for systems with temperatures of 32–138? C, a maximum CO2 loading of 0.5, and between 15–30 wt% MEA in so- lution. It uses the Kent-Eisenberg method for calculating K-values and enthalpy unless the amine concentration is outside of the recommended range in which case Chao-Seader correlation is used for K-value.10 9 Parameter values are taken from D.M. Austgen, G.T. Rochelle, X. Peng, and C.C. Chen, "A Model of Vapor-Liquid Equilibria in the Aqueous Acid Gas-Alkanolamine System Using the Electrolyte-NRTL Equation," Paper presented at the New Orleans AICHE Meeting, March 1988. 10Kent-Eisenberg and Chao-Seader correlations are only used to calculate fugacity of CO2 and H2S. 51 Property methods and model available for CO2-MEA-H2O system cont. . . Name Description emea kemea mea kmea ? emea uses the ELECNRTL property method and is indicated for systems containing CO2, H2S, MEA, and H2O with temperatures up to 120? C and amine concentrations up to 50 wt%. ? kemea is identical to emea except that reaction 4.4 is replaced with a pair of kinetic reactions: CO2 OH ¤ k1 PI HCO ¤ 3 HCO ¤ 3 kQ 1 RSI CO2 OH ¤ This substitution reportedly allows the system to be modelled more accu- rately when using RadFrac? or RateFrac? unit operation models. ? mea and kmea are analogous to emea and kemea except that they use the older SYSOP15M property method. 4.4.2 Specifying streams As a minimum, the conditions and ?ow rates of all input streams must be speci?ed. There are three such streams in Figure 4.5: FLUE-GAS, H2O-PUMP, and MAKE-UP. ? FLUE-GAS ?ow rate and composition is derived from the ?ue gas synthesis re- sults shown previously in Tables 2.4 and 2.5 with one modi?cation: the compo- nents O2, Ar, NO, CO, SO2, and H2 are not included in the MEA absorption problem de?nition. As it turns out, the time required for convergence of the Rate- Frac? UOM is strongly dependent upon the number of components present in the feed and the Stripper rarely converges with all nine components included. The implication of the decision not to include these components, most notably O2, NO, and SO2, on the accuracy of the simulation results is discussed at the end of this chapter. ? H2O-PUMP consists solely of water and its ?ow rate is adjusted such that the ?ue gas is cooled to the desired Absorber inlet temperature. It is assumed that the water is available at atmospheric pressure and a temperature of 12? C.11 11The value of 12T C is taken from [32] and represents the average summer inlet temperature for a sea- 52 ? MAKE-UP adds MEA and H2O to the process to exactly offset the small amounts that are lost from the top of the Absorber and as part of the Stripper distillate. The molar ?ow rates of MEA and H2O in this stream are calculated immediately prior to Mixer execution therefore any initial values suf?ce. It is assumed that this make- up solvent is available at atmospheric pressure and 25? C. Additionally, to ease (i.e., make possible?) ?owsheet convergence, an initial speci?- cation is given to each of the two tear streams: LEAN-ABS and LEAN-HX.12 4.4.3 Specifying blocks Table 4.3 lists the principal blocks in the MEA absorption process ?owsheet and the Aspen Plus? UOM with which it is modelled. The speci?cation of each block follows. Table 4.3: UOM's in MEA absorption process model Block UOM Absorber Stripper RateFrac? Blower COMPR Compressor MCOMPR H2O Pump Rich Pump PUMP Direct Contact Cooler RadFrac? Cooler HEATER Absorber and Stripper Selection of UOM Within Aspen Plus? there are two "general-purpose" UOM's in- dicated for simulating vapour-liquid absorption and stripping columns: RateFrac? and RadFrac?. RateFrac? takes as input the column type and some geometry information from which it computes the coef?cients, ?ow velocities, and hold-up times needed to water based cooling source in the Netherlands. Nanticoke obtains cooling water from Lake Erie and, maybe, similar average summer temperatures prevail. 12It should also be noted that, in this particular implementation, the ?ow rate of LEAN-ABS is the manipulated variable when a speci?c lean solvent CO2 loading is desired. 53 calculate mass transfer. RadFrac? treats separation as an equilibrium problem. Where this assumption is invalid, the departure from equilibrium can be described by assigning a tray or Murphree ef?ciency to each stage. Both RadFrac? and RateFrac? have been used in Aspen Plus? models of MEA absorption processes [55, 25, 26, 27, 20] but only RateFrac? is suitable for the devel- opment of this MEA absorption model. RadFrac? is ?ne in circumstances where tray and/or Murphree ef?ciencies are stable (e.g., column rating mode under constant operat- ing conditions). However, for the MEA absorption process model to be predictive under a wide variety of conditions, the more rigorous RateFrac? UOM is indicated. The decision to use RateFrac? versus RadFrac? creates additional challenges: 1. RateFrac? UOM is more computationally complex which means that simula- tions will solve more slowly and with more dif?culty (i.e., increased probability of non-convergence). This disadvantage is mitigated by intelligent problem initial- ization. 2. As mentioned previously, one of the points of emphasis of the MEA absorption model is to precisely determine the pressure pro?le of the Absorber and Stripper. In achieving this end, there is an important difference to be considered in the man- ner with which RadFrac? and RateFrac? treat column pressure. In RadFrac? the stage pressures can be included in the problem formulation as variables. Thus, in a RadFrac? solution, the pressures used to evaluate the col- umn performance are also outputs of the simulation. This is not the case for Rate- Frac? where segment pressures are constants. After the column performance is calculated using the pressure speci?cation given by the user, Aspen Plus? uses the results to estimate a pressure drop for each segment. There is thus a disconnect between the reported column pressure pro?le and the rest of the column results. It is possible to obtain estimates of actual column pressure pro?les using Rate- Frac? but at the cost of additional computation. Several iterations are required where the estimated pressure drops of one run are used to construct the input pres- sure pro?le of the subsequent run until convergence is achieved. Specifying RateFrac? In specifying the Absorber and Stripper, the model developer needs to make decisions regarding four different aspects of the units: column con?gura- tion, column type, internal geometry, and column pressure. 54 Column con?guration In the case of the Absorber, the inlets and outlets are con- nected to the top and bottom of the column. The Stripper will have both a partial condenser and a conventional reboiler. The feed enters the column above the mass-transfer region. The molar re?ux ratio is varied to achieve a speci?ed condenser temperature (typically 40? C); the bottoms-to-feed ratio is adjusted such that the desired molar ?ow of CO2 in the distillate is obtained (nominally 85% of the CO2 in the ?ue gas). Column type Both columns are modelled with sieve trays. There are other column types to choose from within Aspen Plus?. RateFrac? has built-in routines for bubble-cap and valve trays and for a plethora of random and struc- tured packings.13 Sieve trays are selected because they are commonly used and correla- tions exist for characterizing their hydrodynamic performance. They thus provide a good basis from which to compare more sophisticated column types. Internal con?guration The diameter of the column is an output of the model and is therefore not speci?ed. A diameter estimate (20 m), though, is required as is the number of trays. In addition, the approach to entrainment ?ooding, tray spacing, and weir height need to be given (or the default values of 80%, 24 in, and 2 in, respectively, used) in order to completely specify the tray geometry. The elucidation for the number of trays, tray spacing, and weir height used in the model is provided in Section 4.5.2. Column pressure Chakma et al. [14] originally hypothesized that increasing CO2 pressure in the Absorber would be a good thing because it increases reactivity of MEA with CO2. However, they discovered that any bene?ts accrued due to increased reactivity are more than offset by the increased cost of pressurizing the ?ue gas. Therefore, the pressure at the top of the Absorber is ?xed at 101.3 kPa. In the case of the Stripper, increasing the pressure, which raises the column temper- ature, has been shown to promote less energy-intensive solvent regeneration. However, above temperatures of 122? C, thermal degradation of 30 wt% MEA becomes intolera- ble. Therefore, in the process model, the pressure of the Stripper reboiler is set such that the reboiler temperature approaches, but does not exceed, 122? C. The actual pressure pro?le is determined using the iterative procedure mentioned above. At the beginning of each iteration, the input pressure pro?le of each RateFrac? 13A table listing the complete selection is given in the user documentation [12, p 17-34–17-35]. 55 block is constructed using the segmental pressure drops reported for that particular block from the previous run. The criteria for convergence is a difference between ? ?P? colof consecutive runs of less than 1 kPa or 3% of the total pressure drop: ? ?P? i col ? ?P? i¤ 1 col UWV 1 kPa 0 03 ? ?P? i¤ 1 col Calculations for tray-by-tray pressure drop and % downcomer ?ooding are taken from literature [29, p 14-24–14-34] and are implemented as a Fortran subroutine that is called during RateFrac? execution. RateFrac? contains a built-in routine for these same calculations but initial testing gave calculated pressure drops that were an order of magnitude greater than what was expected. Because the RateFrac? UOM is developed by a third-party, it was not possible to obtain documentation describing the routines and, thus, it was felt best to replace them with a well-known formulation. For reference, the exact calculations used are shown in Appendix B. Table 4.4 summarizes the parameters and stream properties that are required to size the column and evaluate its hydrodynamics. Table 4.4: Design parameters for sizing and hydrodynamic evaluation of tray columns symbol units typical value nominal value EFA % 60–85 75 TS mm 300–600 609.6 ε mm 0.046 0.046 dh mm 6.5–13 13 hc mm 25.4 25.4 hw mm 50 50.8 tt mm 2.0–3.6 3.6 Ah Aa 0.05–0.15 0.15 f 0.75 0.75 Blower The Blower is required to overcome the pressure drop in the cooler and the absorption column and is implemented in Aspen Plus? using the COMPR UOM. COMPR is used to change stream pressure when power requirement is needed and represents a single compressor stage. It requires that the stream pressure rise and the performance charac- teristics be speci?ed. 56 The pressure rise is initially set consistent with the initial pressure conditions in the Absorber. Then, at the beginning of each iteration, the pressure rise in this block is changed such that ? ?P? n Blower ? ?P? n Absorber The performance characteristics for a blower of the size needed to accommodate some 4 106m3 hr of ?ue gas are not readily available. Below are listed the design choices made by other researchers. CO2 Compressor The CO2 Compressor is required to compress the CO2 for transportation via pipeline and is implemented in Aspen Plus? using the MCOMPR UOM. Conceptually, MCOMPR is a series of COMPR blocks interspersed with heat exchangers and is therefore suitable for modelling a multi-stage compressor with inter-cooling. This block requires that the outlet pressure, compression performance, and interstage temperatures be speci?ed. The outlet pressure depends upon the pipelining requirements; the choice of condi- tions by previous researchers is varied and is shown in Table 4.5. Ultimately, the outlet pressure is determined by the pipeline length and design, the location and design of "booster" compressors, and ultimate end-use of the CO2. In this study, the CO2 is compressed to 110 bar at a temperature of 25? C. Table 4.5: Survey of CO2 delivery pressures used in MEA absorption studies Study CO2 conditions Iijima and Kamijo [33] 2000 psig (136 bar) Marion et al. [38] 2000 psig (136 bar), 82? F (28? C) Desideri and Paolucci [20] 140 bar, ambient temperature David Singh [55] 150 bar, 40? C Simmonds et al. [53] 220 bar Slater et al. [56] 220 bar H2O Pump and Rich Pump The H2O Pump and Rich Pump are both modelled with the PUMP UOM. In this work, PUMP only requires that the outlet pressure be speci?ed; by default, PUMP calculates 57 the power requirement using ef?ciency curves for water in a centrifugal pump [12] which provides suf?cient accuracy for this work. The outlet pressure of H2O Pump and Rich Pump are determined by the upstream units. For H2O Pump, the pressure rise is effectively that required to overcome the pres- sure drop of the Direct Contact Cooler and the Absorber. For Rich Pump, the rich sol- vent pressure is increased, if required, to equal the Stripper pressure at the feed segment. In both cases, values are updated along with the Absorber and Stripper pressure pro?les at the beginning of each iteration. Direct Contact Cooler The Direct Contact Cooler is modelled in the same manner as Desideri and Paolucci [20]: it is a two-stage, Rashig-ring packed column with a pressure loss of 0.1 bar. Cooler The Cooler cools the lean solvent to the desired Absorber inlet temperature (typically 40? C). It is modelled with the HEATER UOM. The outlet temperature of Direct Contact Cooler and Cooler is set at 40? C as this allegedly maximizes CO2 absorption. Aroonwilas et al. found that, from 20? C to 37? C, increasing temperature increased CO2 take-up due to an increase in the rate of the reac- tion between CO2 and MEA and, from 40? C to 65? C, increasing temperature decreased CO2 absorption because of Henry s constant increasing with temperature. 4.5 Model Parameter elucidation 4.5.1 Property method selection At the 2nd workshop of the International Test Network for CO2 capture,14 it was proposed that the Aspen Plus? "out of the box" could not accurately model the MEA absorption process. The assertion was made that Aspen Plus? does not ship with a physical property method or model capable of predicting VLE (vapour-liquid equilibrium) of the CO2- MEA-H2O system. 14This "network" is a collaborative effort amongst researchers from industry, academia, and government to develop technologies for capturing CO2 from power plant ?ue gases. It's inaugural meeting was held in Gaithersburg, USA in October 2000 and meetings have been held bianually since. 58 The experimental work of Jou et al. [36] has produced what is held to be the most accurate set of data of CO2 solubility in 30 wt% MEA solution. Jou et al. measured the solubility of CO2 in a 30 wt% solution of aqueous MEA at partial pressures of CO2 ranging from 0.001–20000 kPa and temperatures between 0–150? C. Their results are shown in Figure 4.6. CO2 solubility in aqueous MEA is a strong function of temperature and a moderate function of pressure. 150? C 120? C 100? C 80? C 60? C 40? C 25? C 0? C CO2 loading P CO 2 / kPa 1 4 1 2 1 0 0 8 0 6 0 4 0 2 0 0 10H 5 10H 4 10H 3 10H 2 10H 1 10H 0 10 ¤ 1 10 ¤ 2 10 ¤ 3 Figure 4.6: Solubility of CO2 in 30 wt% MEA solution (Jou et al. [36]) Using Aspen Plus?, CO2 solubility in 30 wt% MEA solution is estimated using rep- resentative property methods and models from the different classes shown in Table 4.2. This data is compared to the results of Jou et al. in two ways. 1. Figure 4.7 and Figure 4.8 contain plots of PCO2 versus α at 40? C and 120? C, re- spectively, for the entire range of PCO2 considered by Jou et al..15 2. In Figure 4.9 and Figure 4.10, CO2 solubility is revisited but, in these cases, only data points for which PCO2 U 2 bar are included in the graphs as the ability of As- pen Plus? to accurately predict high-pressure VLE of the CO2-MEA-H2O system 15 The temperatures of 40T C and 120T C are the low and high temperatures expected in the MEA absorp- tion process. The reader is referred to Appendix D for comparisons between experimental and simulation data at other temperatures. 59 ELECNRTL AMINES inserts Jou et al.[36] CO2 loading P CO 2 / kPa 1 4 1 2 1 0 0 8 0 6 0 4 0 2 0 0 10H 5 10H 4 10H 3 10H 2 10H 1 10H 0 10 ¤ 1 10 ¤ 2 10 ¤ 3 Figure 4.7: Comparison of calculated VLE with experimental values at 40? C ELECNRTL AMINES inserts Jou et al.[36] CO2 loading P CO 2 / kPa 1 4 1 2 1 0 0 8 0 6 0 4 0 2 0 0 10H 5 10H 4 10H 3 10H 2 10H 1 10H 0 10 ¤ 1 10 ¤ 2 10 ¤ 3 Figure 4.8: Comparison of calculated VLE with experimental values at 120? C 60 is not of immediate interest.16 The graphs show the percent difference between predicted CO2 partial pressures and experimental values plotted versus α at 40? C and 120? C. AMINES inserts CO2 loading P CO 2 % difference 0 60 0 50 0 40 0 30 0 20 0 10 0 00 900% 800% 700% 600% 500% 400% 300% 200% 100% 0% 100% Figure 4.9: Residual analysis of VLE data — ?PCO2 vs αlean at 40? C In Figures 4.7 and 4.8, Aspen Plus?, with the correct property method selected, ap- pears reasonably capable of modelling the solubility of CO2 in 30 wt% MEA. The fol- lowing observations regarding the property methods and models are worth noting: ? When developing the Aspen Plus? simulation with the user interface, it is recom- mended that the Electrolyte Wizard be used. This feature assists the development of the model by specifying an appropriate property method (i.e., ELECNRTL), adding any missing ionic components, de?ning the solution chemistry, retrieving binary interaction parameters, and inputting parameters for equilibrium constants. This last point is critical as equilibrium constants, unlike interaction parameters, will not be retrieved at run-time. The abysmal ELECNRTL curves in Figures 4.7 and 4.8 result from simulations for which the Electrolyte Wizard was not used. 16In the MEA absorption process, CO2 partial pressure can be expected not to exceed 2 bar; to do so would require Stripper pressures in excess of this which, in turn, would force reboiler temperatures to exceed 125T C — a temperature above which MEA thermal degradation is a show-stopper. 61 AMINES inserts CO2 loading P CO 2 % difference 0 45 0 40 0 35 0 30 0 25 0 20 0 15 0 10 0 05 0 00 100% 80% 60% 40% 20% 0% 20% 40% Figure 4.10: Residual analysis of VLE data — ?PCO2 vs αlean at 120? C ? The four "MEA" property inserts — mea, kmea, emea, kemea — all predicted identical VLE. This is also the same VLE generated using a simulation developed using the Aspen Plus? Electrolyte Wizard. Figures 4.9 and 4.10 allow one to more clearly observe the deviation between the experimental and predicted values. The horizontal line is provided as a point of reference; a perfectly behaved model would have its points evenly scattered around this line and, in the extreme case, the data points would be coincident with it. At 40? C, Aspen Plus? severely misstates the vapour phase concentration of CO2. The AMINES property model performs better than the property inserts, but as evidenced, AMINES can still overstate vapour phase CO2 concentration by factors of 2–4 . At 120? C, the ?t between the predicted and experimental results better than at 40? C but is still poor. At the higher temperature, the property inserts outperform the AMINES property model. 4.5.2 Absorber and Stripper internal con?guration A method of decomposing the MEA absorption process ?owsheet was developed as part of this work and has already been reported elsewhere [1]. It is applied to the particular ?owsheet shown in Figure 4.5 with the hope of obtaining: 62 ? a realistic indication of the internal con?guration for the Absorber and Stripper and ? 'good' initialization values for tear streams, Stripper re?ux ratio, Stripper bottoms- to-feed ratio, and Absorber and Stripper pressure pro?les. A synopsis of the decomposition concept is given below: 1. The total cost of CO2 capture is more sensitive to the operating costs than the annualized capital costs. 2. In regards to the operating costs, it is the cost of ful?lling Qreb that dominates. 3. For a particular recovery and αlean, Qreb will be minimized when the Stripper inlet ?ow rate is minimized. Well, the inlet ?ow rate is solely determined by the design of the Absorber. 4. As the number of trays in the Absorber is increased, the solvent ?ow rate needed for a particular recovery will decrease asymptotically to Fmin leanas NAbsorber approaches in?nity. It makes sense that at some NAbsorber U ∞, the reduction in solvent ?ow rate from adding an additional tray will be negligible and this N!Absorber will be the design number of trays for the Absorber. 5. At this minimum inlet ?ow rate, the reboiler heat duty is controlled by the design of the Stripper. 6. As the number of trays in the Stripper is increased, the reboiler heat duty will decrease. At some NStripper U ∞, the reduction in reboiler heat duty from adding an additional tray will be negligible and this N!Stripper will be the design number of trays for the Stripper. Absorber study Number of trays With a 'stand-alone' Absorber model, Flean required to achieve 85% recovery of CO2 is determined for 0 05 X αlean X 0 40. The number of trays in the Absorber is varied from one run to the next but the tray spacing and weir height values are not; they are kept constant at the RateFrac? default values. The results of this set of simulations is shown in Figure 4.11. The most signi?cant observations are: 63 Y ?P ` abs ?ow rate Number of trays Column pressure drop / kPa LEAN-ABS ?ow rate / kmol a s b 1 200 c 0 150 c 0 100 c 0 50 c 0 0 c 0 18 16 14 12 10 8 6 4 2 54 c 0 53 c 5 53 c 0 52 c 5 52 c 0 51 c 5 51 c 0 50 c 5 50 c 0 49 c 5 (a) αlean d 0 e 05 Y ?P ` abs ?ow rate Number of trays Column pressure drop / kPa LEAN-ABS ?ow rate / kmol a s b 1 200 c 0 150 c 0 100 c 0 50 c 0 0 c 0 18 16 14 12 10 8 6 4 2 62 c 0 61 c 0 60 c 0 59 c 0 58 c 0 57 c 0 56 c 0 (b) αlean d 0 e 10 Y ?P ` abs ?ow rate Number of trays Column pressure drop / kPa LEAN-ABS ?ow rate / kmol a s b 1 200 c 0 150 c 0 100 c 0 50 c 0 0 c 0 18 16 14 12 10 8 6 4 2 72 c 0 71 c 0 70 c 0 69 c 0 68 c 0 67 c 0 66 c 0 65 c 0 64 c 0 (c) αlean d 0 e 15 Y ?P ` abs ?ow rate Number of trays Column pressure drop / kPa LEAN-ABS ?ow rate / kmol a s b 1 200 c 0 150 c 0 100 c 0 50 c 0 0 c 0 18 16 14 12 10 8 6 4 2 86 c 0 84 c 0 82 c 0 80 c 0 78 c 0 76 c 0 74 c 0 (d) αlean d 0 e 20 Y ?P ` abs ?ow rate Number of trays Column pressure drop / kPa LEAN-ABS ?ow rate / kmol a s b 1 200 c 0 150 c 0 100 c 0 50 c 0 0 c 0 18 16 14 12 10 8 6 4 2 106 c 0 104 c 0 102 c 0 100 c 0 98 c 0 96 c 0 94 c 0 92 c 0 90 c 0 (e) αlean d 0 e 25 Y ?P ` abs ?ow rate Number of trays Column pressure drop / kPa LEAN-ABS ?ow rate / kmol a s b 1 200 c 0 150 c 0 100 c 0 50 c 0 0 c 0 18 16 14 12 10 8 6 4 2 136 c 0 134 c 0 132 c 0 130 c 0 128 c 0 126 c 0 124 c 0 122 c 0 120 c 0 118 c 0 116 c 0 114 c 0 (f) αlean d 0 e 30 Y ?P ` abs ?ow rate Number of trays Column pressure drop / kPa LEAN-ABS ?ow rate / kmol a s b 1 200 c 0 150 c 0 100 c 0 50 c 0 0 c 0 18 16 14 12 10 8 6 4 2 185 c 0 180 c 0 175 c 0 170 c 0 165 c 0 160 c 0 155 c 0 150 c 0 (g) αlean d 0 e 35 Y ?P ` abs ?ow rate Number of trays Column pressure drop / kPa LEAN-ABS ?ow rate / kmol a s b 1 200 c 0 150 c 0 100 c 0 50 c 0 0 c 0 18 16 14 12 10 8 6 4 2 280 c 0 270 c 0 260 c 0 250 c 0 240 c 0 230 c 0 220 c 0 210 c 0 200 c 0 (h) αlean d 0 e 40 Figure 4.11: Sensitivity of Flean to Absorber height 64 ? As is indicated by the increasing scale of the ordinate axes in sub-?gures 4.11(a) through 4.11(h), higher CO2 loadings require greater Flean to achieve the same level of CO2 recovery. ? Flean decreases asymptotically as more trays are added to the Absorber. ? ? ?P? Absorber is directly proportional to the number of trays in the Absorber. The criteria for selection of N!Absorber originally presented [1] was as follows: N!Absorber NAbsorber f f f f f Fi lean Fi H 1 lean Fi lean U 0 005 Why limit the number of trays in the initial Absorber design? Well, even without doing the complete economic analysis, at some point the marginal capital cost of an additional tray in the Absorber will trump the marginal bene?t that a larger column has in reducing Flean. So, adding trays to the Absorber until the reduction in lean solvent ?ow rate dropped below 5% seemed like a reasonable thing to do. The practical limit to the number of trays is tightened with the added consideration of column pressure drop. The bene?t of adding 'just one more tray' is further reduced in light of the fact that while there is a diminishing return from increasing tray number, the marginal cost associated with overcoming ? ?P? Absorber appears to be constant. This new reality spurred the modi?cation of the above selection criteria for N!Absorber: N!Absorber NAbsorber f f f ? ?P? i Absorber U 101 3kPa g 0 05 X αlean X 0 40 As it turns out, N!Absorber 10.17 Tray spacing and weir height With a 'stand-alone' Absorber model, Flean required to achieve 85% recovery of CO2 is determined for 0 05 X αlean X 0 40. N!Absorber trays is used in the Absorber and the tray spacing, weir height, and downcomer clearance are varied by adjusting, k,where TS k $ 24in hw k $ 2in hc k $ 1in 17For the case of αlean h 0¨ 05, a value of NAbsorber h 4 is used because, at this CO2 loading, it was impossible to routinely converge the Absorber with a greater number of trays. 65 The results of this set of simulations is shown in Figure 4.12. Of note is that: ? Using the default RateFrac? values for tray spacing and weir height (represented in Figure 4.12 by points lying on the ordinate axes), the Absorber downcomer ?ooding consistently exceeds the typical design value of 50%TS. ? As αlean increases, so does k!Absorber. ? Increasing the tray spacing such that hdc X 50%TS causes an accompanying in- crease in ? ?P? Absorber.18 An explanation of this observation is as follows. Pressure drop across a tray can be decomposed into the resistance to ?ow through the holes in the tray, hd, and the re- sistance to ?ow through the effective height of clear liquid on the tray, hL. Increas- ing the tray spacing reduces the liquid holdup on the tray (reducing hL) but turns out to decrease hd. This latter effect arises from the fact that increasing the tray spacing increases the gas velocity at which entrainment ?ooding occurs, UNF, and consequently, the gas phase design velocity through the column — the approach to entrainment ?ooding is a constant design parameter and UN EFA 100% UNF. And, of course, the faster the gas ?ows through the holes, the greater the resistance to ?ow. At higher values of k, hd dominates over hL. Values of k!Absorber for each αlean examined are given in Table 4.6. Also shown is the state and composition of the rich stream leaving the Absorber. This information is an input into the Stripper study. Table 4.6: Summary of results from Absorber study αlean N!Absorber k!Absorber Trich Prich Frich xH2O xMEA xCO2 [T C] [kPa] [kmol( s] 0.05 4 5 49.8 118.7 50.4 0.809 0.128 0.063 0.10 10 6 45.4 164.0 54.5 0.807 0.128 0.065 0.15 10 6 46.8 163.3 63.4 0.810 0.126 0.064 0.20 10 7 48.7 170.1 74.8 0.813 0.124 0.063 0.25 10 8 50.9 177.1 92.0 0.816 0.123 0.062 0.30 10 9 52.7 182.0 117.5 0.818 0.121 0.061 0.35 10 11 52.8 190.8 156.2 0.820 0.120 0.060 0.40 10 13 50.8 190.7 224.9 0.821 0.119 0.060 18There is one exception to this statement. With αlean h 0¨ 40 and ki h 13, p ?Pq Absorber h 98¨ 5kPa which is less than the 106¨ 1kPa observed with the RateFrac? default values. 66 Y ?P ` abs' ?ooding α r 0 c 05 Scale parameter Column pressure drop / kPa % downcomer ?ooding 100 c 0 80 c 0 60 c 0 40 c 0 20 c 0 0 c 0 12 10 8 6 4 2 550 c 0 500 c 0 450 c 0 400 c 0 350 c 0 300 c 0 250 c 0 200 c 0 150 c 0 100 c 0 50 c 0 0 c 0 (a) αlean d 0 e 05 Y ?P ` abs' ?ooding α r 0 c 10 Scale parameter Column pressure drop / kPa % downcomer ?ooding 100 c 0 80 c 0 60 c 0 40 c 0 20 c 0 0 c 0 12 10 8 6 4 2 800 c 0 700 c 0 600 c 0 500 c 0 400 c 0 300 c 0 200 c 0 100 c 0 0 c 0 (b) αlean d 0 e 10 Y ?P ` abs' ?ooding α r 0 c 15 Scale parameter Column pressure drop / kPa % downcomer ?ooding 100 c 0 80 c 0 60 c 0 40 c 0 20 c 0 0 c 0 12 10 8 6 4 2 1000 c 0 900 c 0 800 c 0 700 c 0 600 c 0 500 c 0 400 c 0 300 c 0 200 c 0 100 c 0 0 c 0 (c) αlean d 0 e 15 Y ?P ` abs' ?ooding α r 0 c 20 Scale parameter Column pressure drop / kPa % downcomer ?ooding 100 c 0 80 c 0 60 c 0 40 c 0 20 c 0 0 c 0 12 10 8 6 4 2 1400 c 0 1200 c 0 1000 c 0 800 c 0 600 c 0 400 c 0 200 c 0 0 c 0 (d) αlean d 0 e 20 Y ?P ` abs' ?ooding α r 0 c 25 Scale parameter Column pressure drop / kPa % downcomer ?ooding 100 c 0 80 c 0 60 c 0 40 c 0 20 c 0 0 c 0 12 10 8 6 4 2 1800 c 0 1600 c 0 1400 c 0 1200 c 0 1000 c 0 800 c 0 600 c 0 400 c 0 200 c 0 0 c 0 (e) αlean d 0 e 25 Y ?P ` abs' ?ooding α r 0 c 30 Scale parameter Column pressure drop / kPa % downcomer ?ooding 100 c 0 80 c 0 60 c 0 40 c 0 20 c 0 0 c 0 12 10 8 6 4 2 3000 c 0 2500 c 0 2000 c 0 1500 c 0 1000 c 0 500 c 0 0 c 0 (f) αlean d 0 e 30 Y ?P ` abs' ?ooding α r 0 c 35 Scale parameter Column pressure drop / kPa % downcomer ?ooding 100 c 0 80 c 0 60 c 0 40 c 0 20 c 0 0 c 0 12 10 8 6 4 2 4500 c 0 4000 c 0 3500 c 0 3000 c 0 2500 c 0 2000 c 0 1500 c 0 1000 c 0 500 c 0 0 c 0 (g) αlean d 0 e 35 Y ?P ` abs' ?ooding α r 0 c 40 Scale parameter Column pressure drop / kPa % downcomer ?ooding 100 c 0 80 c 0 60 c 0 40 c 0 20 c 0 0 c 0 12 10 8 6 4 2 8000 c 0 7000 c 0 6000 c 0 5000 c 0 4000 c 0 3000 c 0 2000 c 0 1000 c 0 0 c 0 (h) αlean d 0 e 40 Figure 4.12: Sensitivity of Absorber downcomer ?ooding to Absorber tray spacing 67 Stripper study Number of trays With a 'stand-alone' Stripper model, the molar re?ux and molar bottoms-to-feed ratios required for the removal of 2 85kmol s of CO2 (i.e., 85% of the CO2 fed into the bottom of the Absorber) are ascertained for 0 10 X αlean X 0 35.19 RICH-HX is speci?ed using the results from the Absorber study shown in Figure 4.6. The number of trays in the Stripper is varied from one run to the next but tray spacing and weir height are not; they are kept constant at the RateFrac? default values. The results from this set of simulations is presented in Figure 4.13. Some points worth mentioning: ? For αlean U 0 30, increasing αlean decreases Qreb. Where αlean 0 30, Qreb is insensitive to changes in αlean. ? For αlean X 0 25, Qreb decreases asymptotically as more trays are added to the Stripper. Where αlean 0 25, there is a point where Qreb is minimized with respect to NStripper. The '5% rule' criteria used to select N!Stripper presented in [1] is used here without modi?cation: N!Stripper NStripper f f f f f QN reb QN H 1 reb QN reb U 0 05 The value of N!Stripper for each αlean examined is given in Table 4.7. Tray spacing and weir height Once again, with a 'stand-alone' Stripper model, molar re?ux and molar bottoms-to-feed ratios are again ascertained for 0 10 X αlean X 0 35. However, in this case, for each αlean examined, the number of trays is kept constant at N!Stripper and it is k that is varied. The results for this set of simulation is given in Figure 4.14. In the main, the observations here are the same as those from the Absorber study: ? Using the default RateFrac? values for tray spacing and weir height (represented in Figure 4.14), the Stripper downcomer ?ooding consistently exceeds the typical design value of 50%TS. 19The range of CO2 loadings examined is narrowed because of the dif?culty converging the Stripper at very high and very low loadings. 68 Number of trays Q mathitreb / MW 10 9 8 7 6 5 4 3 2 1 5000s 0 4500s 0 4000s 0 3500s 0 3000s 0 2500s 0 2000s 0 1500s 0 1000s 0 500s 0 (a) αlean h 0¨ 10 Number of trays Q mathitreb / MW 10 9 8 7 6 5 4 3 2 1 2400s 0 2200s 0 2000s 0 1800s 0 1600s 0 1400s 0 1200s 0 1000s 0 800s 0 600s 0 400s 0 (b) αlean h 0¨ 15 Number of trays Q mathitreb / MW 10 9 8 7 6 5 4 3 2 1 1300s 0 1200s 0 1100s 0 1000s 0 900s 0 800s 0 700s 0 600s 0 500s 0 400s 0 (c) αlean h 0¨ 20 Number of trays Q mathitreb / MW 8 7 6 5 4 3 2 1 800s 0 750s 0 700s 0 650s 0 600s 0 550s 0 500s 0 450s 0 400s 0 (d) αlean h 0¨ 25 Number of trays Q mathitreb / MW 9 8 7 6 5 4 3 2 1 560s 0 540s 0 520s 0 500s 0 480s 0 460s 0 440s 0 420s 0 400s 0 (e) αlean h 0¨ 30 Number of trays Q mathitreb / MW 7 6 5 4 3 2 1 490s 0 480s 0 470s 0 460s 0 450s 0 440s 0 430s 0 420s 0 410s 0 400s 0 (f) αlean h 0¨ 35 Figure 4.13: Sensitivity of Qreb to Stripper height 69 heat duty downcomer ?ooding α t 0s 10 Scale parameter Q mathitreb / MW % downcomer ?ooding 1600s 0 1400s 0 1200s 0 1000s 0 800s 0 600s 0 400s 0 6 5 4 3 2 1 1100s 0 1000s 0 900s 0 800s 0 700s 0 600s 0 500s 0 400s 0 300s 0 200s 0 100s 0 0s 0 (a) αlean h 0¨ 10 heat duty downcomer ?ooding α t 0s 15 Scale parameter Q mathitreb / MW % downcomer ?ooding 850s 0 800s 0 750s 0 700s 0 650s 0 600s 0 550s 0 500s 0 450s 0 400s 0 7 6 5 4 3 2 1 1800s 0 1600s 0 1400s 0 1200s 0 1000s 0 800s 0 600s 0 400s 0 200s 0 0s 0 (b) αlean h 0¨ 15 heat duty downcomer ?ooding α t 0s 20 Scale parameter Q mathitreb / MW % downcomer ?ooding 540s 0 520s 0 500s 0 480s 0 460s 0 440s 0 420s 0 400s 0 8 7 6 5 4 3 2 1 3000s 0 2500s 0 2000s 0 1500s 0 1000s 0 500s 0 0s 0 (c) αlean h 0¨ 20 heat duty downcomer ?ooding α t 0s 25 Scale parameter Q mathitreb / MW % downcomer ?ooding 460s 0 450s 0 440s 0 430s 0 420s 0 410s 0 400s 0 9 8 7 6 5 4 3 2 1 4500s 0 4000s 0 3500s 0 3000s 0 2500s 0 2000s 0 1500s 0 1000s 0 500s 0 0s 0 (d) αlean h 0¨ 25 heat duty downcomer ?ooding α t 0s 30 Scale parameter Q mathitreb / MW % downcomer ?ooding 460s 0 450s 0 440s 0 430s 0 420s 0 410s 0 400s 0 12 10 8 6 4 2 6000s 0 5000s 0 4000s 0 3000s 0 2000s 0 1000s 0 0s 0 (e) αlean h 0¨ 30 heat duty downcomer ?ooding α t 0s 35 Scale parameter Q mathitreb / MW % downcomer ?ooding 470s 0 460s 0 450s 0 440s 0 430s 0 420s 0 410s 0 400s 0 12 10 8 6 4 2 8000s 0 7000s 0 6000s 0 5000s 0 4000s 0 3000s 0 2000s 0 1000s 0 0s 0 (f) αlean h 0¨ 35 Figure 4.14: Sensitivity of Stripper downcomer ?ooding to Stripper tray spacing 70 ? As αlean increases, so does k!Stripper. ? Increasing the tray spacing such that hdc X 50%TS has no signi?cant impact upon Qreb. Values of k!Stripper are given in Table 4.7. Summary of Absorber and Stripper studies' results The decomposition methodology has yielded a set of conditions which can be used to initialize the integrated process model. This data is given in Table 4.7. Table 4.7: MEA absorption process model initialization parameters αlean N!Absorber k!Absorber Flean N!Stripper k!Stripper L D B F [kmol( s] 0.10 10 6 56.0 9 6 6.75 0.928 0.15 10 6 64.6 9 7 3.05 0.939 0.20 10 7 75.7 8 8 1.22 0.951 0.25 10 8 92.4 6 9 0.68 0.961 0.30 10 9 117.5 5 11 0.59 0.970 0.35 10 11 155.0 5 12 0.54 0.978 4.6 Conclusions and recommendations 1. Generally speaking, Aspen Plus?, "out of the box", is not able to predict CO2 sol- ubility in 30 wt% MEA. At the most favourable conditions — moderate pressure (i.e., those of interest to MEA absorption processes) and higher temperatures (i.e., Stripper conditions) — agreement between experimental and predicted values is only mediocre. 2. For simulation of MEA absorption processes, on the basis of predicting CO2 sol- ubility, either the AMINES property model or the property inserts should be used. 3. For modelling an MEA absorption process, especially when handling ?ue gas vol- umes typically emitted from power plants and recovering substantial fractions of the CO2 contained therein, the RateFrac? default tray geometry is unsuitable. 71 4. After accumulating data like that shown in Table 4.7 and using it to initialize the MEA absorption model, solving said model is no longer dif?cult. 5. However, even with seemingly 'good' initialization values, convergence can still be dif?cult to achieve because of Aspen Plus?'s sensitivity to the initial conditions. For example, in initializing the Stripper, there were many occasions when, for example, initial ? L D? values of 0.4 and 0.5 are unsuccessful but ? L D?u 0 35 works. The reason for this behaviour is not well understood. 72 Chapter 5 Integration of Power Plant and MEA Absorption 5.1 Introduction Objective The objective of the work in this chapter is to intelligently integrate the combustion, steam cycle, and MEA absorption models such that heat and power from the power plant is used to satisfy the supplemental energy requirements of the CO2 capture process. Motivation Unifying the combustion, steam cycle, and CO2 capture models creates a platform from which the merits of steam extraction for process heating can be assessed. In addition, the model places the energy requirements of key unit operations — Blower, Stripper reboiler, and CO2 Compressor — on the same basis which allows different designs to be more easily compared. Merits of steam extraction for process heating Figure 32 contains the enthalpy-entropy curve for a unit at operating at base-load.1 I AB, vI CD, and wI DE represent expansion through the high-, medium-, and low-pressure sections 1The arguments presented in this section are in?uenced heavily by those presented by Mimura et al. [43]. 73 of the turbine, respectively.2 x I e f is the enthalpy change that occurs in the Condenser. Pre- CO2 capture, this heat is completely lost to the surroundings. If CO2 capture using MEA absorption is to be performed, signi?cant amounts of heat will be required by the Stripper reboiler. Assuming Treb 121? C, a 10? C hot-side temperature approach, and saturated inlet and outlet conditions, x Ixy represents the change in process steam enthalpy in the reboiler. Comparing x I e f and x Ixy , it appears possible to substantially mitigate the impact of large Qreb by diverting steam from the latter stages of the turbine. This, in effect, would translate much of the waste heat into useful energy. However, doing so would reduce the steam ?ow through the turbine thus de-rating the power plant. Obviously, the bene?t of extracting steam from the steam cycle for use in the CO2 capture plant depends upon the tradeoff between the recovery of waste heat and the accompanying reduction in electricity production. y x f e steam condensing line E D C B A Entropy / Btuy lb ??? R Enthalpy / Btu ? lb 2.00 1.95 1.90 1.85 1.80 1.75 1.70 1.65 1.60 1.55 1.50 1600 1400 1200 1000 800 600 400 200 0 A HP inlet B HP outlet C IP inlet D IP/LP crossover E Condenser outlet e hCondenser? in f hCondenser? out x hreb? out Y hreb? in Figure 5.1: Enthalpy-entropy curve for power plant Figure 5.2 better illustrates the inherent tradeoff mentioned above. Depicted is the utilization of steam internal energy through the steam cycle; Figure 5.2(a) re?ects nomi- nal steam cycle operation whereas Figure 5.2(b) represents a case where 50% of the LP section of the turbine is extracted.3. The upper three blocks in Figure 5.2(a) show the energy transfer as the steam expands in the turbine and the area of the lowest region is the energy released in the Condenser. 2The actual transitions would not necessarily appear as straight lines on the enthalpy-entropy diagram but this lack of precision does not adversely affect the discussion presented here. 3 See Appendix C for a discussion of the development of Figure 5.2. 74 1400 1200 1000 900 200 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 300 1100 1300 A B C D D E e f PSfrag replacements Steam ?ow rate / lb hr Enthalpy / Btu B lb (a) base 1400 1200 1000 900 200 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 300 1100 1300 A B C D D E e f x y PSfrag replacements Steam ?ow rate / lb hr Enthalpy / Btu B lb (b) w/ steam extraction Figure 5.2: Implication of steam extraction on steam cycle work and heat ?ows The split between the two 'sinks' is approximately 41% for the former and 59% for the latter. The additional shaded region in the adjacent ?gure, the one straddling vI CD and wI DE, is the ?ow of energy redirected from the turbine and Condenser once steam is extracted. What do the ?gures say in regards to the bene?t of extracting steam for reboiler heating? ? Some 49% of otherwise waste heat instead services the reboiler. This represents 86% of the reboiler heat duty. ? The caveat is that the remaining 14% of the reboiler duty is taken from energy that otherwise would go into generating power in the LP section of the turbine. This a little over 30% of the steam internal energy in the wI DE region. ? Post-CO2 capture, 70% of the total energy ?ow is going towards power generation or servicing the Stripper reboiler and only 30%, down from 59%, is "thrown out with the bath water". Given the above development, steam extraction does seem bene?cial. With the inte- grated model, it becomes possible to quantify the bene?ts/disadvantages of such a sys- tem. In particular, one will be able to ascertain: ? how much steam is required to satisfy the Stripper reboiler heat duty? ? and, by how much will this quantity of steam extraction de-rate the power-plant? 75 Comparison of different process designs and con?gurations In Chapter 4, it is stated that a motivation for a detailed and adaptable MEA absorp- tion model is the ?exibility it affords. More speci?cally, such a model allows changes in ?owsheet con?guration, equipment design, solvent selection, and process operating conditions of the process performance to be studied. However, comparing the results of different case studies can be dif?cult. As an example, consider the effect of changing the design CO2 loading in the lean solvent stream. The effect of manipulating this variable has been reported [26, 27, 1] but only in regards to its effect on Qreb. Increasing αlean increases the solvent ?ow rate through the Absorber which, in turn, increases the pressure drop across that column. So, on the one hand, increasing αlean reduces the process heating requirement but, on the other hand, it increases the need for compression power. These quantities are directly incomparable so how is one to truly ascertain the loading where the tradeoff is equal? There are other variables which create similar problems (e.g., Absorber height, Column height, Stripper pressure) as described above. With the uni?ed model, all process duties, be they work or heat, are ultimately re?ected in the plant's electricity output: a concept that is easy to grasp, is sensitive to design changes, and is relevant to the bigger question of "does CO2 capture at an existing coal-?red power plant make sense?" Thus, tradeoffs similar to the one described in the preceding paragraph, are more easily assessed with an integrated model. 5.2 Implementation The synthesis of the integrated ?owsheet required adding new units to simulate ?ue gas cleanup and the Stripper reboiler and, most importantly, deciding from which location in the steam cycle to extract steam and how best to re-inject the condensate. The simulation ?owsheet is given in Figure 5.3 and the details of its development are given below. 5.2.1 Location of steam extraction and condensate re-injection There are two considerations in regards to the identi?cation of the 'right' place to with- draw steam. 1. Steam needs to be at the right temperature. 76 BLOWER H2O_PUMP DCC CO2_COMP FLUE-BLO FLUE-DCC H2O-PUMP H2O-DCC H2O-OUT FLUE-ABS CO2 ST1 ST2 ST3 ABSORBER RICH_PUMP STRIPPER HEATX MIXER COOLER FLUE-ABS LEAN-ABS STACK RICH-PUM RICH-HX LEAN-HX RICH-STR LEAN-MIX MAKE-UP LEAN-COO CO2-COMP HP_SEP1 VALVE1 HP1 HP_SEP2 REHT IP_SEP1 IP2 IP1 IP_SEP2 IP3 IP4 IP_SEP3 IP_SEP4 IP_SEP5 IP_COMB VALVE2 LP_SEP1 LP_SEP2 LP1 LP2 LP_SEP3 LP3 LP_SEP4 LP6 LP5 LP_SEP5 LP4 LP_COMB1 LP_COMB2 FWPUMP2 FWP_C FWPUMP1 CONDENSE CND_COMB FPT1 FPT2 FPT_COMB BOIL H2O-BOIL ST_MAIN ST-FPT1 ST-HP HP_1X ST-REHT ST-FWPA ST-IPX ST-HPX ST-IP IP_03 IP_02 IP_2X IP_12 ST-FWPC IP-1LP IP_3X1 IP_34 IP_3X2 ST-FPT2 ST-FWPB IP-4LP ST-FWPD ST-LP LP_056 LP_012 LP_02 LP_01 ST-FWPF LP_2X LP_23 ST-2FWPG LP_3CR LP_06 LP_05 ST-FWPE LP_5X LP_45 ST-5FWPG LP_4CR ST-CNDR ST-FWPG H2O-FWPA STFWP_AB STFWP_BC H2O-FWPB FPT_12 H2O-FWPC H2O-FWPD ST-FWPE STFWP_EF H2O-FWPE H2O-FWPF STFWP_GC H2O-FWPG H2O-MAIN H2O-CNDR H2-PUMP ST-FPT1 IP_4X STFPT-CN IN-PUMP FPT_1X STFWP_FG STFWP_DE Q_FWPA Q_FWPB Q_FWPD Q_FWPE Q_FWPF Q_FWPG FWP_A FWP_B FWP_D FWP_E FWP_F FWP_G H2O-REB ST-REB EXTRACT HTRANS SEPARATE COAL-IN AIR FLUE-FGD SOLIDS Q_DECOMP Q_FURN COAL-OUT IN-BURN EXHAUST DECOMP BURN SCRUBBER Q_REHT Q_BOIL WASTE REBOIL Q-REB Figure 5.3: Power plant with integrated MEA absorption simulation ?owsheet 77 The consensus is that Treb X 122? C as, above this temperature, either thermal degradation of MEA or corrosion [61] becomes intolerable. Therefore, to maintain a "rule-of-thumb" 10? C hot-side temperature approach in the reboiler, the steam conditions must be such that Tsat 132? C. In addition, it is desirable to take the lowest quality steam that is available and meets this criteria; steam superheat is more valuable for power generation than heat transfer. 2. The extraction point must be both accessible and able to accommodate the needed steam ?ow rate. There is limited mention in the literature of steam being extracted from the steam cycle of a power plant for providing heat to the Stripper reboiler. ? Mimura et al. [43, 44] refer to an "optimum steam system for power plant ?ue gas CO2 recovery". In practice, this consists of extracting steam midway through the low-pressure section of the turbine. In the lone case in which they considered recovering CO2 from the coal-derived ?ue gas, some 3 25 106 lb hr of 1208Btu lb steam is extracted. As the nominal plant output is 900 MWe, this represents approximately 54% of the steam leaving the boiler. ? Desideri and Paolucci [20] extract steam for reboiler heating at the same po- sition as dearating steam is taken — midway through the LP section casing. 7 39 105 lb hr of steam at 5 bar pressure is removed from the turbine caus- ing the 320 MWe plant to be de-rated by about 17%. ? In their study, Marion et al. [38] extract steam from the IP/LP crossover pipe. 2 5 106 lb hr of steam, or 79% of that generated in the boiler, is extracted from the nominally 450 MWe plant. Inferred from these results is that it is not merely a "bleed" stream of steam that is required; access to large quantities of steam is necessary if this is going to work. The schematic of the steam turbine in Figure 5.4 is repeated from Figure 3.6; it shows the location of all the potential steam extraction points. The adjacent table gives the steam ?ow rate at each location and the steam saturation temperature. From the data, it is (almost?) obvious as to where steam can be taken. Consider the ?rst requirement that Tsat 132? C but as close to this cutoff as possible. Well, the steam in the IP/LP crossover pipe (position LP) and at position D have the same conditions and meet the ?rst set of criteria. These positions differ dramatically, though, in terms of accessibility and availability, the second requirement. 78 IP LP A C B D F G CNDR E HP high pressure intermediate pressure low pressure low pressure Location Flow rate Tsat [lb/hr] [6 C] HP 3358670 349 IP 2990122 248 LP 2494525 149 A 334659 255 B 143920 206 C 128853 175 D 136359 149 E 133578 120 F 89306 90 G 126171 70 CNDR 2143469 32 Figure 5.4: Base-load steam conditions in steam cycle The "bleed" stream D is used in the fourth feed water pre-heater. Steam at this loca- tion, and at any of the other extraction points for that matter, have several disadvantages that preclude their use for providing for the Stripper reboiler: ? They are situated on the underside of the turbine, restricting access. ? The ?ow paths would not permit signi?cantly increased ?ow rates than the nominal ones given in Figure 5.4[37]. Therefore, steam for the reboiler is taken from the IP/LP crossover pipe and, conse- quently, the condensate is re-injected into the cycle at the fourth feed water pre-heater. Doing so effectively splits the turbine into two parts: a base-load part consisting of the high- and intermediate-pressure sections and a part-load part consisting of the low- pressure section. In this manner, the correlations developed in Chapter 3 for predicting the performance of the turbine and feed water pre-heaters as a function of header and heat exchanger inlet ?ow rates is still applicable, even with steam extraction. Thus, the estimation of the power plant de-rate due to reduced steam ?ow through the LP section will be accurate. For illustration purposes, the turbine of one of the units at Nanticoke is depicted in Figures 5.8 through 5.7. The IP/LP crossover pipes are the double-pair of large, longitu- dinal pipes that come up from the middle of the section shown in Figure 5.6 and extend into Figure 5.7 and is from here that reboiler steam is to be taken.4 4Apparently, this extraction location is not just feasible on paper. A preliminary estimation is that appropriate access at this position could be added during planned shutdown periods [37]. 79 Figure 5.5: High-pressure section of Nanticoke turbine Figure 5.6: Intermediate-pressure section of Nanticoke turbine 80 Figure 5.7: Low-pressure section of Nanticoke turbine PSfrag replacements IP/LP crossover pipe HP section IP section Figure 5.8: Lengthwise view of Nanticoke turbine 81 5.2.2 Maximum available steam for Stripper reboiler heating Deciding to transfer power plant steam to the Stripper reboiler imposes a practical limit as to the magnitude of reboiler heat duty that can be accommodated. The speci?ed min- imum design load for Nanticoke is 25% [4] however, it should be operationally feasible to go down to 10% ?ow through the LP section of the turbine [37]. In any case, even if it were possible to extract all of the steam, there is a ?nite amount available and this dictates the reboiler heating possible. In order to ascertain this maximum heat duty, a series of simulations is performed where the fraction of steam extracted from the IP/LP crossover pipe is slowly increased. The extracted steam is condensed to the saturated liquid at 132? C and is re-injected into the steam cycle. The sensible and latent heat released is recorded and, along with the power plant terminal input, is shown in Figure 5.9. Heat output Net electricity output Fraction of steam extracted Power / MW 1.00 0.80 0.60 0.40 0.20 0.00 700 600 500 400 300 200 100 0 Figure 5.9: Sensitivity of power plant electricity output to steam extraction Cases with steam extraction up to 90% were examined. However, when more than 83% of steam was diverted to the reboiler, the ?ow through the lowest-pressure stages of the LP section of the turbine is reduced to zero. The corresponding maximum Qreb is approximately 625 MW and the terminal input is reduced from an initial 496.72 MWe to 360.02 MWe. 82 5.2.3 Flue gas pre-conditioning Background The ?rst attempts to capture CO2 from coal-derived ?ue gases were made in the 1980's at the Sundance Power Plant in Alberta [64] and, later, at the Boundary Dam Power Station in Saskatchewan [39]. While the efforts showed that large-scale capture of CO2 from coal power is feasible, operational problems abounded due principally to the presence of ?y ash, O2, NOx, and SOx in the ?ue gas. In the intervening years, practical limits for each of these components in the ?ue gas have evolved. Fly ash Fly ash causes foaming in the columns and plugging, scaling, corrosion, ero- sion in equipment and should be removed to 0.006 gr/dscf [15]. NOx NOx needs to be at or below 20 ppm [53]. O2 O2 is a problem because it oxidizes carbon steel and degrades MEA [15, 16, 28]. O2 is dealt with in Fluor Daniel's Econamine FG? by using oxygen inhibitors. Alterna- tive approaches include using oxygen-tolerant alloys, removal of all oxygen from the ?ue gas (near-stoichiometric combustion and/or catalytic reduction), and continuous addition of oxygen scavengers to the solvent [15]. SOx SOx is a problem because it reacts irreversibly with MEA to form heat-stable salts thus reducing the absorption capacity of the solvent[15, 64, 63]. In systems where 30 wt% MEA solution is used, solvent losses due to SOx become uneconomic when SOx is greater than 10 ppmv in the ?ue gas [15, 39, 40, 38, 53]. With the Kerr-McGee/ABB Lummus Crest process, SOx removal is necessary if the ?ue gas contains more than 100 ppmv. In the 50–100 ppmv range, upstream SOx removal is optional as it can be removed during reclaiming through the addition of caustic. The downside to SOx removal in the reclaimer is that some MEA loss occurs. Below 50 ppmv, SOx removal is not justi?ed [13]. Present implementation Removal of ?y ash is already accomplished in the Separate block that is part of the coal combustion model. A new block, Scrubber, modelled with the UOM's SEP2 and FLASH2, handles the removal of O2, NOx, and SOx from the ?ue gas as it ?ows between the coal combustion and MEA absorption parts of the ?owsheet. 83 5.2.4 Stripper reboiler The Stripper reboiler is modelled using the HEATER UOM. The outlet stream is satu- rated liquid. A zero pressure-drop is assumed across the unit. 5.2.5 Blower and CO2 Compressor It is assumed that electrical motors are used to drive the Blower and CO2 Compressor and motor ef?ciencies of 90% are assumed. 5.3 Process Simulation With an integrated process model, it is now possible to evaluate the feasibility of using the power plant as the source of Stripper reboiler heating. In Chapter 4 is given a number of different process design considerations that should in?uence the attractiveness of CO2 capture using MEA absorption. For clarity, these 'ideas' that are evaluated are given in Table 5.1. Table 5.1: Scope of MEA absorption sensitivity analysis Design variable Location αlean page 84 NAbsorber page 87 NStripper page 89 As a basis, initial column heights of NAbsorber 10 and NStripper 7 are used. 5.3.1 Sensitivity of CO2 capture to recycle CO2 loading In this study, the effect of changes in αlean to the net electric output of the plant is exam- ined. The results are shown in Figures 5.10 through 5.12. ? At lower CO2 loading, Qreb decreases quickly with increasing loading. With αlean 0 23, Qreb changes very little with CO2 loading, going though a shallow minimum at αlean 0 26 (Figure 5.10). 84 ? ?P? Absorber Qreb lean solvent CO2 loading Absorber pressure drop / kPa Stripper reboiler heat duty / MW th 100 95 90 85 80 75 0.32 0.30 0.28 0.26 0.24 0.22 0.20 600 580 560 540 520 500 480 460 440 420 Figure 5.10: Sensitivity of Absorber pressure drop and Stripper reboiler heat duty to CO2 loading CO2 Compressor Blower Stripper reboiler lean solvent CO2 loading Electric power / MW e 0.32 0.30 0.28 0.26 0.24 0.22 0.20 140 120 100 80 60 40 20 0 Figure 5.11: Sensitivity of capture plant's electricity demand to CO2 loading 85 terminal input pre-capture output lean solvent CO2 loading Electric power / MW e 0.32 0.30 0.28 0.26 0.24 0.22 0.20 550 500 450 400 350 300 250 Figure 5.12: Sensitivity of power plant electricity output to CO2 loading ? The Absorber pressure drop tends to increases as loading is increased. This is because Flean increases with αlean and this puts more resistance on the ?ow of vapour upwards through the column (Figure 5.10).5 ? EBLOWER ? ECO2 COMP (Figure 5.11). ? ? EBLOWER ECO2 COMP? ? ? ?E? reb 6 (Figure 5.11). ? With constant design considerations NAbsorber, NStripper, and % ?ooding, EBLOWER and ECO2 COMP are insensitive to αlean; EBLOWER experiences only a slight increase over the range observed (Figure 5.11). ? Power plant terminal output is less sensitive to changes in CO2 loading then Qreb. Over the interval 0 22 X αlean X 0 33, the change in Enet is never more than ? 5 MWe (Figure 5.12). 5The stepwise nature of p ?Pq Absorber curve arises from the constraint that the downcomer ?ooding must be less than 50% but that the tray spacing and weir height are only adjusted in whole number multiples of the Aspen Plus? default values of 24 and 2 inches, respectively. 6Just to be clear, p ?Eq reb represents the decrease in Eplant that occurs by the extraction of steam for Stripper reboiler service 86 5.3.2 Sensitivity of CO2 capture to Absorber height In this study, the effect of changes in Absorber height on the net electric output of the plant is examined. The optimum loading from the Section 5.3.1, αlean 0 25, is used. The results are shown in Figures 5.13 through 5.15. ? ?P? Absorber Qreb Absorber height / number of trays Absorber pressure drop / kPa Stripper reboiler heat duty / MW th 160 140 120 100 80 60 40 20 0 18 16 14 12 10 8 6 4 2 480 470 460 450 440 430 420 410 Figure 5.13: Sensitivity of Absorber pressure drop and Stripper reboiler heat duty to Absorber height ? Qreb decreases asymptotically as NAbsorber is increased. The overall effect is moder- ate; from the 'base case' at NAbsorber 10, it was only possible to obtain a reduction of 10 MWth in Qreb, about 2%, by moving to NAbsorber 18 (Figure 5.13). ? ? ?P? Absorber varies linearly with NAbsorber; every additional tray increases the col- umn pressure drop by about 8 kPa (Figure 5.13). ? The moderate reduction that increasing NAbsorber has on Qreb translates to even less impressive savings in electric power consumption. This is especially true in com- parison to the increases in required Blower power as the Absorber size is increased (Figure 5.14). ? As alluded to by Figure 5.17, the increased ? ?P? Absorber of making the Absorber bigger more than offsets any reductions in Qreb. The power plant de-rate grows in synchronization the Absorber (Figure 5.15). 87 CO2 Compressor Blower Stripper reboiler Absorber height / number of trays Electric power / MW e 18 16 14 12 10 8 6 4 2 120 100 80 60 40 20 0 Figure 5.14: Sensitivity of capture plant's electricity demand to Absorber height terminal input pre-capture output Absorber height / number of trays Electric power / MW e 18 16 14 12 10 8 6 4 2 550 500 450 400 350 300 250 Figure 5.15: Sensitivity of power plant electricity output to Absorber height 88 5.3.3 Sensitivity of CO2 capture to Stripper height In this study, the effect of changes in Stripper height on the net electric output of the plant is examined. The optimum loading from the Section 5.3.1, αlean 0 25, is used. The results are shown in Figures 5.16 through 5.18. ? ?P? Stripper Qreb Stripper height / number of trays Stripper pressure drop / kPa Stripper reboiler heat duty / MW th 50 45 40 35 30 25 20 15 10 5 9 8 7 6 5 4 3 2 1 750 700 650 600 550 500 450 400 Figure 5.16: Sensitivity of Stripper pressure drop and Stripper reboiler heat duty to Stripper height ? Following the Qreb curve from low to high values of NStripper, there is an imme- diate and strong bene?t to increasing Stripper height. This bene?t does taper off rather quickly, though; the 'base case' value, with just NStripper 7, has the lowest corresponding Qreb (Figure 5.16). ? ? ?P? Stripper increases with increasing NStripper but not as quickly as is the case with the Absorber. Here, each additional tray only caused an increase of about 5 kPa (Figure 5.16). ? Reductions in Qreb should translate directly into a reduced electric power con- sumption and that is indeed the case. In regards to ? ?P? Stripper, smaller values are preferred as this leads to higher pressures in the column distillate which means the CO2 Compressor has to work less hard. While technically, this effect is observed, the magnitude of the change is very small (Figure 5.17). 89 CO2 Compressor Blower Stripper reboiler Stripper height / number of trays Electric power / MW e 9 8 7 6 5 4 3 2 140 120 100 80 60 40 20 0 Figure 5.17: Sensitivity of capture plant's electricity demand to Stripper height terminal input pre-capture output Stripper height / number of trays Electric power / MW e 9 8 7 6 5 4 3 2 550 500 450 400 350 300 250 Figure 5.18: Sensitivity of power plant electricity output to Stripper height 90 ? The output of the plant tends to increase asymptotically as NStripper is increased (Figure 5.18). 5.4 Model validation Table 5.2 compares the best case selected from each study with results from literature. 91 Table 5.2: MEA absorption process energy duties Unit capacity Flue gas CO2 recovery Pout EBlower EComp Qreb Source [MWe] [mol% CO2] [tonne ? hr] [%] [tonne ? hr] [bar] [MWe] [MWe] [MWth] 500 13.6 2312 85 426 110 45 41 426 CO2 loading study 500 13.6 2312 85 426 110 10 41 474 Absorber study 500 13.6 2312 85 426 110 46 41 426 Stripper study 1000 13.2 3888 60 468 9 56 370 Morimoto et al. [45]7 450 15.0 2619 96 379 135 45 721 Marion et al. [38]8 400 14.6 1664 90 331 150 9 31 351 Singh [55]9 320 13.2 1205 90 234 140 9 20 234 Desideri et al. [20]10 300 11.6 1882 88 192 141 12 43 371 Mariz et al. [40]11 300 11.6 1797 92 192 141 6 43 285 Mariz et al. [40]12 300 333 54 245 Paitoon et al. [60]13 15.0 565 95 118 Chakma et al. [14] 7Only 2/3 of the "emit gas" from the power plant is processed by the capture plant; in actuality, 90% of the CO2 that enters the absorption process is removed. The blower and compressor are steam-driven so the duties given for these units represent shaft power and not electrical power 8 In estimating Qreb, the following assumptions are made: LIP d 3 e 1 ? 106 lb ? hr, ? ?P ? reb d 0, and the condensate leaving the reboiler is saturated liquid. The given compressor duty includes the energy required for the blower. 9The auxiliary energy equipment emits 84.45 tonne/year of CO2. Therefore, the CO2 abatement at the power plant is only 65%. 10The results reported by Desideri and Paolucci are of questionable quality. For starters, the base ef?ciency of the power plant in their study is 44.3% and their speci?c compression power and reboiler heat duty are much lower than observed elsewhere. 11 About 75% of the ?ue gas is from the power plant with the residual generated by the auxiliary coal-?red boiler. In this study, MEA absorption is based on Econamine FG? process. 12About 78% of the ?ue gas is from the power plant with the residual generated by the auxiliary coal-?red boiler. In this study, MEA absorption is based on MHI/KEPCO KS-1/KP-1 process. 13The study is based on capturing 8000 tonne/day from a 300 MW coal-?red power plant. A "back-of-the-envelope" calculation shows that, given the coal-composition given, even at 30% overall thermal ef?ciency, a 300 MW power plant would produce less than 8000 tonne/day of CO2. The given compressor duty includes the energy required for the blower. 92 First, in regards to the electricity consumption, while the CO2 Compressor duties obtained in this study are comparable to what has been observed elsewhere, the Blower duties obtained in the CO2 and Stripper studies are substantially higher than anything seen before. This is attributable to the fact that the MEA absorption model used in this work explicitly calculated pressure drop across the Absorber for a given column design. In other studies [54, 20], ? ?P? abs of approximately 0.2 bar is assumed irrespective of the height of the column or the type of packing used. For the moment, as most researchers have been apt to do, consider only Qreb. In Table 5.3 is required speci?c reboiler heat duty as reported in this study, and by others. Table 5.3: Stripper reboiler speci?c heat duty Source Qreb [kJ( kg CO2] CO2 loading study 1.00 Absorber study 1.11 Stripper study 1.11 Morimoto et al. [45] 0.37 Desideri and Paolucci [20] 0.73 Singh [55] 1.06 Mariz et al. [40]a 1.48 Marion et al. [38] 1.90 Mariz et al. [40]b 1.93 aIn this study, MEA absorption is based on MHI/KEPCO KS-1/KP-1 process. bIn this study, MEA absorption is based on Econamine FG? process. The values of Qreb obtained in this study fall in line with expectations regarding the effect of ?ue gas CO2 concentration and absorption solvent performance: 1. CO2 capture is facilitated by higher concentrations of CO2 in the ?ue gas. Consider the follow studies in which the CO2 concentration is ? 13%: ? Qreb 1 48 kJ kg CO2 for Mariz et al. versus 0.37 for Morimoto et al. (both studies use KS-1 as a solvent). ? Qreb 1 93 kJ kg CO2 for Mariz et al. versus 0.73 and 1.00–1.11 found by Desideri and Paolucci and in this study (all studies used 30 wt% CO2 as a solvent). 93 2. KS-1 (Morimoto et al., 0 37 kJ kg CO2) outperforms 30 wt% aqueous MEA (Desideri and Paolucci, 0 73 kJ kg CO2; this study, 1 00–1 11 kJ kg CO2; Singh, 1 06 kJ kg CO2) which outperforms 15 wt% aqueous MEA (Marion, 1 90 kJ kg CO2). 5.5 Conclusions and recommendations ? The integration of the coal combustion, steam cycle, and MEA absorption models is accomplished in a straight-forward manner. ? The IP/LP crossover pipe is the preferred extraction location from which to ex- tract steam for Stripper reboiler as it is easily accessible and furnishes steam at conditions relatively close to those required. ? The following table summarizes the best conditions observed in each of the sensi- tivity analyses performed to date: Table 5.4: Summary of best cases from sensitivity studies Study αlean NAbsorber NStripper xsteam Enet de-rate [MWe] [%] αlean 0.27 10 7 0.58 320 35.7 NAbsorber 0.25 2 7 0.64 342 31.2 NStripper 0.25 10 7 0.58 319 35.9 94 Chapter 6 Conclusion and Future Work 6.1 Conclusion As mentioned in the Introduction, the more conventional method of satiating the Stripper reboiler is by generating steam in a boiler dedicated to this purpose. This thesis seeks to evaluate the feasibility of obtaining the heat required for MEA absorption for the existing power plant which implicitly poses the question, "Is extracting steam from the existing power plant a superior alternative to disassociated units?" It seems a direct comparison is in order. . . The ef?cacy of the approaches can be compared using thermal ef?ciency de?ned by: ηth Enet Qb For the particular cases of concern here, ηth is evaluated in terms of model outputs as follows: ηth ? ?8? Enet QboilH Qreht integrated power plant w CO2 capture Enet QboilH QrehtH Qreb? ηaux con?guration w auxillary boiler The analyses from Chapter 5, with two modi?cations, are repeated; this time, a CO2 capture plant that produces its own steam, as required, using an auxiliary boiler is con- sidered and the measured variable is ηth. The results are tabulated and compared with 95 auxiliary integrated lean solvent CO2 loading Thernal ef?ciency 0.32 0.30 0.28 0.26 0.24 0.22 0.20 0.230 0.225 0.220 0.215 0.210 0.205 0.200 Figure 6.1: In?uence of CO2 loading on plant thermal ef?ciency auxiliary integrated Absorber height / number of trays Thernal ef?ciency 18 16 14 12 10 8 6 4 2 0.245 0.240 0.235 0.230 0.225 0.220 0.215 0.210 0.205 0.200 Figure 6.2: In?uence of Absorber height on plant thermal ef?ciency 96 auxiliary integrated Stripper height / number of trays Thernal ef?ciency 9 8 7 6 5 4 3 2 1 0.230 0.225 0.220 0.215 0.210 0.205 0.200 0.195 0.190 0.185 0.180 Figure 6.3: In?uence of Stripper height on plant thermal ef?ciency computed thermal ef?ciencies for the proposed capture plant with integrated CO2 capture in Figures 6.1 through 6.3. In all cases, extracting steam from the steam cycle is a Good Thing?. As a general rule, it can be said that doing so improves the plant's energy utilization by a full percent- age point. The conclusion of this work is that, for the case of adding CO2 capture using MEA absorption to the existing coal-?red power plant at Nanticoke Generating Station, satisfying the supplemental heat demand by using steam from the power plant is the way to go. 6.2 Future work The following are suggestions as to projects which build upon the successes of this work: ? Researchers at the University of Texas, at Austin have developed a rate-based model for the chemistry of the CO2-MEA-H2O system [25, 26, 27]. This is in con- trast to the equilibrium model used in this work. It would be interesting to incor- porate this kinetic model into the power plant with integrated CO2 capture model developed here and to measure any changes to the conclusions, if any, brought about by the increased rigour. 97 ? The sequential modular approach of the current MEA absorption model is a dis- advantage and complicates the issue of the extreme sensitivity of the RateFrac? UOM to changes in process conditions which makes convergence problematic. Newer versions of Aspen Plus? [11] now ship with an equation oriented solver. With such a solution method, iteration, a major source of trouble in this thesis, would no longer be necessary. The model should be re-implemented using the equation oriented approach and ef?cacy of solving the integrated ?owsheet using the two different techniques compared. ? Much of the discussion of improving the design of CO2 capture processes based on MEA absorption focuses on minimizing Qreb. As is clearly demonstrated in Section 5.3.2, minimizing Qreb does not necessarily optimize the design of the CO2 capture plant. For this reason, in this study, ηth is used as a metric for evaluating designs. Ultimately, though, it is the cost of each strategy that guides the decision as to what eventually to implement. While it is true that costs themselves can be misleading [49], they are required if apple-to-apple comparisons are to be made between different technology options (e.g., PCC versus NGCC). Therefore, the cost of CO2 capture of the best designs from this work should be costed out. ? A new power plant design with integrated CO2 capture should realize higher ther- mal ef?ciencies than any retro?t case whether steam extraction is implemented as part of the retro?t or not. A new design could preclude the extraction of super- heated steam for reboiler heating, make steam available at a variety of conditions by including extraction ports in the turbine casing, include additional auxiliary tur- bines to produce power for the Blower and Compressor, and increase the number and quality of heat-integration opportunities. To date, there is limited, if any, work being done in this area.1. One of the more useful outcomes would be a PCC case for comparison with new NGCC and IGCC power plants (both with and with- out CO2 capture) that are being proposed for construction in Ontario and North America at large. 1The only mention in the literature is of work done as part of a joint venture by the Japanese companies Mitsubishi Heavy Industries and Kansai Electric Power Company [43, 44] 98 Appendix A Conditions of steam at potential extraction locations Table A.1: Base and part-load conditions in Nanticoke steam cycle Stream 100% 75% 50% T P L T P L T P L [T F] [psia] [106 lb( hr] [T F] [psia] [106 lb( hr] [T F] [psia] [106 lb( hr] ST MAIN 1000 0 2365 0 3 36 1000 0 2365 0 3 26 1000 0 2365 0 3 24 ST-FPT1 1000 0 2365 0 0 01 1000 0 2365 0 0 01 1000 0 2365 0 0 01 ST-HP 994 6 2236 2 3 34 962 1 1631 5 3 24 930 0 1080 7 3 21 ST-REHT 646 7 622 4 2 99 629 4 460 8 2 94 611 7 309 9 2 95 ST-FWPA 646 7 622 4 0 33 629 4 460 8 0 29 611 7 309 9 0 24 ST-IP 1000 0 560 2 2 99 1000 0 414 5 2 94 1000 0 278 7 2 95 ST-FWPC 624 1 129 3 0 13 627 7 96 6 0 11 631 1 65 4 0 10 ST-FPT2 787 9 253 9 0 08 788 9 188 8 0 08 790 9 127 3 0 09 ST-FWPB 787 9 253 9 0 14 788 9 188 8 0 14 790 9 127 3 0 12 ST-FWPD 484 0 66 6 0 14 488 2 49 9 0 12 491 9 34 0 0 11 ST-LP 484 0 66 6 2 49 488 2 49 9 2 48 491 9 34 0 2 52 ST-FWPF 193 4 10 1 0 09 179 9 7 5 0 08 176 1 5 2 0 08 ST-FWPE 330 8 28 8 0 14 335 0 21 7 0 13 339 6 14 9 0 12 99 Base and part-load conditions in Nanticoke steam cycle cont. . . Stream 100% 75% 50% T P L T P L T P L [T F] [psia] [106 lb( hr] [T F] [psia] [106 lb( hr] [T F] [psia] [106 lb( hr] ST-CNDR 89 5 0 7 2 14 79 0 0 5 2 15 79 0 0 5 2 23 ST-FWPG 157 8 4 5 0 13 145 9 3 4 0 13 131 9 2 3 0 10 H2O-FWPA 400 6 3 36 379 1 3 26 350 7 3 24 H2O-BOIL 487 9 3 36 461 1 3 26 425 6 3 24 H2O-FWPB 351 2 2700 0 3 36 329 6 2550 0 3 26 303 9 2500 0 3 24 H2O-FWPC 293 2 2 74 276 8 2 71 255 2 2 75 H2-PUMP 345 4 3 36 324 2 3 26 297 7 3 24 H2O-FWPD 241 6 2 74 228 0 2 71 209 9 2 75 H2O-FWPE 186 4 2 74 175 0 2 71 160 3 2 75 H2O-FWPF 150 0 2 74 140 7 2 71 129 1 2 75 H2O-FWPG 90 2 2 74 79 9 2 71 80 3 2 75 STFPT CN 89 5 0 7 0 09 79 0 0 5 0 09 79 0 0 5 0 09 H2O-MAIN 89 5 2 74 79 0 2 71 79 0 2 75 100 Appendix B Sieve Tray Column Hydrodynamic Design Recipe Table B.1 summarizes the parameters and stream properties that are required to size the column and evaluate its hydrodynamics. B.1 Tower diameter The tower diameter is equal to the diameter of the largest tray. The following steps are required to calculate tray diameter: 1. Calculate constant FLG. FLG L G ? ρG ρL ? 0? 5 2. Calculate Csbf . Csbf 0 0105 8 127 10 ¤ 4 TS0? 755 exp ? 1 463F0? 842 LG ? TS is usually 300–600 mm. 101 Table B.1: Required input for sizing and hydrodynamic evaluation of tray columns Parameters Properties symbol nominal symbol units EFA 60–85% L kg s TS 300–600 mm G kg s ε 0.046 mm qL m3 s dh 6.5–13 mm qG m3 s g 9.8 m s2 ρL kg m3 hc 25.4 mm ρG kg m3 hw 50 mm σ dynes cm tt 2.0–3.6 mm ?L kg m $ s Ah Aa 0.05–0.15 f 0.75 3. Calculate the gas velocity through the net area at entrainment ?ooding, uNF. UNF Csbf d σ 20e 0? 2 ? ρL ρG ρG ? 0? 5 4. The design gas velocity through the net area, UN, is selected as a percentage of UNF. Perry's [29] says that prudent designs call for approaches to ?ooding of 75– 85%. Course notes [2] give typical design values of 60–80%. UN EFA 100% UNF This calculation is valid provided that the following conditions are met: ? system is low- or non-foaming ? hw U 0 15TS ? dh U 13mm ? Ah Aa f 0 1 5. The net area of the column is the portion through which gas ?ows. Therefore, the magnitude of this area, AN is the quotient of the gas volumetric ?ow rate and UN. 102 AN qG UN 6. The net area is the difference between the total cross-sectional area, Atotal, and the area under the downcomer, Ad. The area of the downcomer is determined by specifying the weir length which is speci?ed as a fraction f of the diameter, usually 75% [2]. Atotal AN 1 1 π d sin ¤ 1 f f g 1 f2 e 7. Calculate the diameter, d. d ih 4Atotal π B.2 Downcomer ?ooding The depth of liquid in a downcomer should be such that it is less than 50% full. The following recipe calculates the height of clear liquid in a downcomer, hdc. 1. Calculate the height due to the downcomer apron, hda. (a) Calculate the area for ?ow under the downcomer apron, Ada. Ada ? w hc As stated above, ? w f $ d. hc, as a rule of thumb, is 1787 [2]. (b) Then, calculate hda. hda 165 2 ? qL Ada ? 2 2. Calculate the height due to the hydraulic gradient across the tray, hhg. 103 (a) Calculate the gas velocity through the active area, Ua. Ua UN AN 2AN Atotal (b) Calculate Ks. Ks Ua ? ρG ρL ρG ? 0? 5 (c) Calculate the effective froth density on the plate, φe. φe exp ? 12 55K0? 91 s ? (d) Calculate the effective clear-liquid height (i.e., liquid holdup), hL. i. Calculate the constant C. C 0 0327 0 0286exp ? 0 1378hw? ii. Then use C to calculate hL. hL φe j hw 15330C ? qL φe ? 2 3 k hw is usually 50 mm and less than 15% of tray spacing [2]. (e) Calculate the froth height, hf . hf hL φe (f) Calculate the average width of the ?ow path, Df . Df Lw d 2 (g) Calculate the hydraulic radius of the aerated mass, Rh. Rh hf Df 2hf 1000Df 104 (h) Calculate the velocity of the aerated mass, Uf . Uf 1000qL hL Df (i) Calculate the Reynolds number for the ?ow, NRe h. NRe h RhUf ρL ?L (j) Calculate the Fanning friction factor, fF, for the ?ow. fF V 1 737ln l 0 269 ε Rh 2 185 NRe h ln ? 0 269 ε Rh 14 5 NRe h ?nmpo ¤ 2 ε is 0.046 mm for commercial steel [62]. (k) The length of the ?ow path across the plate, ? f , is given by Lf k $ d. Find k such that f $ k π 4 V 1 2 π q d sin ¤ 1 f f g 1 f2 e d sin ¤ 1 k kg 1 k2 esr o (l) Calculate hhg. hhg 1000 fF U2 f ? f gRh 3. Calculate the height caused by liquid pushing up in order to ?ow over the weir, how. how 664 ? qL ? w ? 2 3 4. Calculate the height caused by the pressure drop across plate, ht. (a) Calculate the pressure drop that would exist across the dry dispersion plate, hd. 105 i. Firstly, calculate the gas phase velocity through the tray perforations, Uh. Uh Ua ? Ah Aa ? ¤ 1 The hole area is usually 5–15% of the active area (i.e., Ah Aa ? 0 1 ). ii. Then calculate the constant Cv. Cv 0 74 ? Ah Aa ? exp l 0 29 ? tt dh ? 0 56 m tt is usually 2–3.6 mm. dh is 6.5–13 mm. iii. Finally, calculate hd. hd ? 50 8 C2 v ? ? ρG ρL ? U2 h (b) Calculate the pressure drop across the aerated liquid on the tray, h7L. The procedure of Bennet et al. as described in [29] is followed. i. Calculate the pressure drop for surface generation, h7σ. h7σ ? 472σ gρL ? l g ? ρL ρG? dh σ m 1 3 ii. Using the value of hL calculated during the determination of hhg, calcu- late h7L. h7L hL h7σ (c) Calculate ht. ht hd h7L 5. Calculate the height of liquid in the downcomer, hdc. hdc ht hw how hda hhg B.3 Tray pressure drop The total pressure drop across the tray, ?Pt, is expressed as a pressure head, ht, as follows: 106 ?Pt ht ρL g 1000 B.4 Downcomer seal The downcomer seal, hds must be great enough to prevent vapour from propagating up- wards along this channel. hds hw how 0 5hhg As rule of thumb, hds f ha 13–38 mm [2]. B.5 Weeping Weeping occurs when there is insuf?cient pressure to maintain a froth on the tray surface. Deleterious weeping occurs when a signi?cant amount of liquid ?ows through the tray, thereby diminishing contact between the vapour and liquid phases. Weeping is checked for the minimum expected ?ow rates for a particular column design using Figure 14-27 in [29]. The abscissa and ordinate values are calculated as follows: x hW how y 4σ dh 409σ ρL dh 107 Appendix C Steam Energy Calculations Expansion in turbine By de?nition,1 H U PV Taking the partial differential of both sides and rearranging gives, dH dU d ? PV ? dU dH d ? PV ? Assuming that the steam behaves ideally, i.e., PV nRT an expression for steam internal energy in terms of enthalpy and temperature is easily obtained 1Please note the following that in this Appendix the overline is used to distinguish between the absolute and mass-relative forms of internal energy and enthalpy. 108 dU dH nRdT dU m dH m RdT M Integrating both sides gives the ?nal expression: ?U ?H R?T M Condensing heat transfer Starting with the ?nal expression for speci?c internal energy calculated above ?U ?H R?T M it is apparent that, in the case of condensing heat transfer, the expression simpli?es to ?U ?H as this is a constant temperature process. 109 Changes in internal energy encountered in Nanticoke steam cycle Table C.1: Changes in steam internal energy in steam cycle Process Hin Hout Tin Tout ?U Btu( lb Btu( lb T F T F Btu( lb I AB 1462.83 1318.33 1000 646.7 -105.5 vI CD 1517.88 1274.65 1000 484.0 -186.3 wI DE 1274.65 1017.84 484 89.5 -213.3 x I e f 1017.84 57.52 -960.3 x Ixy 1169.93 236.70 -933.2 110 Appendix D Comparison of Calculated CO2 Solubility With Experimental Values Figures D.1 through D.8 compare the solubility of CO2 in 30 wt% aqueous MEA over the complete range of temperatures investigated by Jou et al. [36]. ELECNRTL AMINES inserts Jou et al.[36] CO2 loading P CO 2 / kPa 1 4 1 2 1 0 0 8 0 6 0 4 0 2 0 0 10H 5 10H 4 10H 3 10H 2 10H 1 10H 0 10 ¤ 1 10 ¤ 2 10 ¤ 3 Figure D.1: Comparison of calculated VLE with experimental values at 0? C 111 ELECNRTL AMINES inserts Jou et al.[36] CO2 loading P CO 2 / kPa 1 4 1 2 1 0 0 8 0 6 0 4 0 2 0 0 10H 5 10H 4 10H 3 10H 2 10H 1 10H 0 10 ¤ 1 10 ¤ 2 10 ¤ 3 Figure D.2: Comparison of calculated VLE with experimental values at 25? C ELECNRTL AMINES inserts Jou et al.[36] CO2 loading P CO 2 / kPa 1 4 1 2 1 0 0 8 0 6 0 4 0 2 0 0 10H 5 10H 4 10H 3 10H 2 10H 1 10H 0 10 ¤ 1 10 ¤ 2 10 ¤ 3 Figure D.3: Comparison of calculated VLE with experimental values at 40? C 112 ELECNRTL AMINES inserts Jou et al.[36] CO2 loading P CO 2 / kPa 1 4 1 2 1 0 0 8 0 6 0 4 0 2 0 0 10H 5 10H 4 10H 3 10H 2 10H 1 10H 0 10 ¤ 1 10 ¤ 2 10 ¤ 3 Figure D.4: Comparison of calculated VLE with experimental values at 60? C ELECNRTL AMINES inserts Jou et al. 80? C αCO2 P CO 2 / kPa 1 4 1 2 1 0 0 8 0 6 0 4 0 2 0 0 10H 5 10H 4 10H 3 10H 2 10H 1 10H 0 10 ¤ 1 10 ¤ 2 10 ¤ 3 Figure D.5: Comparison of calculated VLE with experimental values at 80? C 113 ELECNRTL AMINES inserts Jou et al.[36] CO2 loading P CO 2 / kPa 1 4 1 2 1 0 0 8 0 6 0 4 0 2 0 0 10H 5 10H 4 10H 3 10H 2 10H 1 10H 0 10 ¤ 1 10 ¤ 2 10 ¤ 3 Figure D.6: Comparison of calculated VLE with experimental values at 100? C ELECNRTL AMINES inserts Jou et al.[36] CO2 loading P CO 2 / kPa 1 4 1 2 1 0 0 8 0 6 0 4 0 2 0 0 10H 5 10H 4 10H 3 10H 2 10H 1 10H 0 10 ¤ 1 10 ¤ 2 10 ¤ 3 Figure D.7: Comparison of calculated VLE with experimental values at 120? C 114 ELECNRTL AMINES inserts Jou et al.[36] CO2 loading P CO 2 / kPa 1 4 1 2 1 0 0 8 0 6 0 4 0 2 0 0 10H 5 10H 4 10H 3 10H 2 10H 1 10H 0 10 ¤ 1 10 ¤ 2 10 ¤ 3 Figure D.8: Comparison of calculated VLE with experimental values at 150? C 115 Appendix E Aspen Plus Input ?le for Power Plant With Integrated MEA Absorption ; File: plant_w_capture_w_steam_extract.inp ; This file simulates the part-load performance of a nominal 500 MW ; power plant with CO2 capture. Steam is extracted from the IP/LP ; crossover pipe to supply the stripper reboiler. ; Report options STREAM-REPOR MOLEFLOW MASSFLOW PROPERTIES=ALL-SUBS CPCVMX ; Diagnostic specifications DIAGNOSTICS HISTORY SIM-LEVEL=4 CONV-LEVEL=4 MAX-PRINT SIM-LIMIT=9999 ; This paragraph specifies time and error limits. RUN-CONTROL MAX-TIME=84600 MAX-ERRORS=1000 ; This paragraph will cause AspenPlus to include FORTRAN tracebacks in the ; history file. SYS-OPTIONS TRACE=YES 116 ; Units IN-UNITS ENG POWER=KW OUT-UNITS SI PRESSURE=kPa TEMPERATURE=C PDROP=kPa ; Components COMPONENTS ; The property inserts component list contains: H2O, MEA, H2S, CO2, N2, ; HCO3-, MEACOO-, MEA+, CO32-, HS-, S2-, H3O+, and OH-. All other ; components need to be listed below: ; These components are involved in coal combustion. ; different types of coal COAL-IEA / COAL-PRB / COAL-USL / ASH / ; elements contained within coal C C / H2 H2 / CL2 CL2 / HCL HCL / S S / ; H2O H2O / ; components of air ; N2 N2 / O2 O2 / AR AR / NE NE / HE HE-4 / CH4 CH4 / KR KR / XE XE / ; combustion products CO CO / 117 ; CO2 CO2 / NO NO / NO2 NO2 / SO2 O2S / SO3 O3S ; This paragraph specifies the physical property method and model for each ; non-conventional component. NC-COMPS COAL-IEA ULTANAL SULFANAL PROXANAL NC-PROPS COAL-IEA ENTHALPY HCOALGEN 6 1 1 1 / DENSITY DCOALIGT NC-COMPS COAL-PRB ULTANAL SULFANAL PROXANAL NC-PROPS COAL-PRB ENTHALPY HCOALGEN 6 1 1 1 / DENSITY DCOALIGT NC-COMPS COAL-USL ULTANAL SULFANAL PROXANAL NC-PROPS COAL-USL ENTHALPY HCOALGEN 6 1 1 1 / DENSITY DCOALIGT NC-COMPS ASH PROXANAL ULTANAL SULFANAL NC-PROPS ASH ENTHALPY HCOALGEN / DENSITY DCOALIGT ; Properties ; This insert specifies property method and data for aqueous MEA-CO2 system. ; ELECNRTL becomes the default property method... INSERT MEA CEMEA H2O MEA H2S CO2 N2 NO ; Specify the property method to use in each section. PROPERTIES PR-BM COAL PROPERTIES STEAM-TA HP IP LP FPT FWP CNDR ; This section specifies which databanks to use. DATABANKS PURE11 / AQUEOUS / SOLIDS / INORGANIC / NOASPENPCD PROP-SOURCES PURE11 / AQUEOUS / SOLIDS / INORGANIC PROP-SET ALL-SUBS VOLFLMX MASSVFRA MASSSFRA RHOMX MASSFLOW & TEMP PRES UNITS='lb/cuft' SUBSTREAM=ALL ; "Entire Stream Flows, Density, Phase Frac, T, P" ; This paragraph specifies the gross calorific value for each type of 118 ; coal (Btu/lb) on a dry, mineral-matter free basis. PROP-DATA HEAT IN-UNITS SI MASS-ENTHALPY="KJ/KG" PROP-LIST HCOMB PVAL COAL-IEA 27060 ; 11632 PVAL COAL-PRB 27637 ; 11880 PVAL COAL-USL 31768 ; 13656 PROP-SET VFLOW VOLFLMX PROP-SET LPHASE MUMX RHOMX SIGMAMX VOLFLMX MASSFLMX PHASE=L & UNITS='KG/CUM' 'DYNE/CM' PROP-SET VPHASE RHOMX VOLFLMX MASSFLMX PHASE=V UNITS='KG/CUM' PROP-SET CPCVMX CPCVMX DEF-STREAMS MIXCINC COAL DEF-STREAMS CONVEN HP IP LP FPT FWP CNDR MEA ; BEGIN: flowsheet specification ; some globally defined blocks and streams FLOWSHEET GLOBAL BLOCK "SHAFT" IN="W_HP" "W_IP" "W_LP" OUT="P_INTERN" ; globally defined streams DEF-STREAMS WORK "P_INTERN" ; globally defined blocks BLOCK SHAFT MIXER ; COAL COMBUSTION ; Flowsheet 119 FLOWSHEET COAL BLOCK DECOMP IN=COAL-IN OUT=COAL-OUT "Q_DECOMP" BLOCK BURN IN=COAL-OUT AIR "Q_DECOMP" OUT=IN-BURN BLOCK HTRANS IN=IN-BURN OUT=EXHAUST "Q_FURN" BLOCK SEPARATE IN=EXHAUST OUT=FLUE-AHT SOLIDS BLOCK AIR-HEAT IN=FLUE-AHT OUT=FLUE-SCR BLOCK SCRUB1 IN=FLUE-SCR OUT=WASTE1 IN-SCRUB BLOCK SCRUB2 IN=IN-SCRUB OUT=FLUE-GAS WASTE2 ; Stream Specification ; specify the heat and work streams in the flowsheet DEF-STREAMS HEAT "Q_DECOMP" "Q_FURN" ; The composition of air is taken from Cooper et al., p 653. STREAM AIR TEMP=519 PRES=101.3 MOLE-FLOW=1.0 MOLE-FRAC H2 .000050 / N2 78.090 / O2 20.940 / AR .930 / CO2 .0360 / NE .00180 / HE .000520 / CH4 .000170 / KR .00010 / NO2 .000030 / XE 8.0000E-06 STREAM COAL-IN SUBSTREAM NC TEMP=160 PRES=101.30 MASS-FLOW=10 MASS-FRAC COAL-IEA 0.0 / COAL-PRB 0.5 / COAL-USL 0.5 ; PROXANAL ULTANAL ; water, moisture-included basis ash (dry-basis) ; fixed carbon (dry-basis) carbon (dry-basis) ; volatile matter (dry-basis) hydrogen (dry-basis) ; ash (dry-basis) nitrogen (dry-basis) ; chlorine (dry-basis) ; sulfur (dry-basis) ; oxygen (dry-basis) ; IEA tech specs coal... COMP-ATTR COAL-IEA ULTANAL ( 13.48 71.38 4.85 1.56 0.026 0.952 7.79 ) COMP-ATTR COAL-IEA PROXANAL ( 9.50 86.52 0.0 13.48 ) COMP-ATTR COAL-IEA SULFANAL ( 0.0 100 0.0 ) ; Powder River basin coal COMP-ATTR COAL-PRB ULTANAL ( 7.1 69.4 4.9 1.0 0.000 0.4 17.2 ) 120 COMP-ATTR COAL-PRB PROXANAL ( 28.1 49.95 42.92 7.13 ) COMP-ATTR COAL-PRB SULFANAL ( 0.0 100 0.0 ) ; US low-sulphur coal COMP-ATTR COAL-USL ULTANAL ( 10.4 77.2 4.9 1.5 0.000 1.0 5.0 ) COMP-ATTR COAL-USL PROXANAL ( 7.5 55.95 33.69 10.36 ) COMP-ATTR COAL-USL SULFANAL ( 0.0 100 0.0 ) ; Block Section BLOCK DECOMP RYIELD PARAM TEMP=298.15 PRES=0.0 MASS-YIELD MIXED H2O .30 / NC ASH .10 / CISOLID C .10 / MIXED H2 .10 / N2 .10 / CL2 .10 / S .10 / O2 .10 COMP-ATTR NC ASH PROXANAL ( 0.0 0.0 0.0 100 ) COMP-ATTR NC ASH ULTANAL ( 100 0.0 0.0 0.0 0.0 0.0 0.0 ) COMP-ATTR NC ASH SULFANAL ( 0.0 0.0 0.0 ) ; This block decomposes the coal into a stream of its component elements. CALCULATOR COAL-DEC DEFINE XC BLOCK-VAR BLOCK=DECOMP VARIABLE=YIELD SENTENCE=MASS-YIELD & ID1=CISOLID ID2=C DEFINE XH2 BLOCK-VAR BLOCK=DECOMP VARIABLE=YIELD SENTENCE=MASS-YIELD & ID1=MIXED ID2=H2 DEFINE XN2 BLOCK-VAR BLOCK=DECOMP VARIABLE=YIELD SENTENCE=MASS-YIELD & ID1=MIXED ID2=N2 DEFINE XCL2 BLOCK-VAR BLOCK=DECOMP VARIABLE=YIELD SENTENCE=MASS-YIELD & ID1=MIXED ID2=CL2 DEFINE XS BLOCK-VAR BLOCK=DECOMP VARIABLE=YIELD SENTENCE=MASS-YIELD & ID1=MIXED ID2=S DEFINE XO2 BLOCK-VAR BLOCK=DECOMP VARIABLE=YIELD SENTENCE=MASS-YIELD & ID1=MIXED ID2=O2 DEFINE XASH BLOCK-VAR BLOCK=DECOMP VARIABLE=YIELD SENTENCE=MASS-YIELD & ID1=NC ID2=ASH DEFINE XH2O BLOCK-VAR BLOCK=DECOMP VARIABLE=YIELD SENTENCE=MASS-YIELD & ID1=MIXED ID2=H2O DEFINE CIEA MASS-FLOW STREAM=COAL-IN SUBSTREAM=NC COMPONENT=COAL-IEA DEFINE CPRB MASS-FLOW STREAM=COAL-IN SUBSTREAM=NC COMPONENT=COAL-PRB 121 DEFINE CUSL MASS-FLOW STREAM=COAL-IN SUBSTREAM=NC COMPONENT=COAL-USL ; ultimate analyses of the three coals VECTOR-DEF UIEA COMP-ATTR STREAM=COAL-IN SUBSTREAM=NC & COMPONENT=COAL-IEA ATTRIBUTE=ULTANAL VECTOR-DEF UPRB COMP-ATTR STREAM=COAL-IN SUBSTREAM=NC & COMPONENT=COAL-PRB ATTRIBUTE=ULTANAL VECTOR-DEF UUSL COMP-ATTR STREAM=COAL-IN SUBSTREAM=NC & COMPONENT=COAL-USL ATTRIBUTE=ULTANAL ; proximate analyses of the three coals VECTOR-DEF PIEA COMP-ATTR STREAM=COAL-IN SUBSTREAM=NC & COMPONENT=COAL-IEA ATTRIBUTE=PROXANAL VECTOR-DEF PPRB COMP-ATTR STREAM=COAL-IN SUBSTREAM=NC & COMPONENT=COAL-PRB ATTRIBUTE=PROXANAL VECTOR-DEF PUSL COMP-ATTR STREAM=COAL-IN SUBSTREAM=NC & COMPONENT=COAL-USL ATTRIBUTE=PROXANAL ; Stupid fucking Aspen Plus fortran interpreter can't handle lines > ; 72 characters so I have to break up the arithmetic into bite-sized pieces... ; COAL => total coal mass flowrate F COAL = CIEA + CPRB + CUSL ; THE VECTOR U___ CONTAINS THE MASS FRACTIONS OF THE COAL CONSTITUENTS ; ON A DRY-BASIS WHEREAS THE COAL FLOW RATE ON A WET-BASIS. THE factor ; DRY___ is used to make this conversion. ; ; DRY___ => coal "dry" fraction (i.e. 1 - moisture fraction) ; P___(1) => coal moisture content, wt% F DRYIEA = (100 - PIEA(1)) / 100 F DRYPRB = (100 - PPRB(1)) / 100 F DRYUSL = (100 - PUSL(1)) / 100 F ASH1 = (UIEA(1) / 100) * DRYIEA * CIEA F ASH2 = (UPRB(1) / 100) * DRYPRB * CPRB F ASH3 = (UUSL(1) / 100) * DRYUSL * CUSL F XASH = (ASH1 + ASH2 + ASH3) / COAL F C1 = (UIEA(2) / 100) * DRYIEA * CIEA F C2 = (UPRB(2) / 100) * DRYPRB * CPRB F C3 = (UUSL(2) / 100) * DRYUSL * CUSL 122 F XC = (C1 + C2 + C3) / COAL F HYDRO1 = (UIEA(3) / 100) * DRYIEA * CIEA F HYDRO2 = (UPRB(3) / 100) * DRYPRB * CPRB F HYDRO3 = (UUSL(3) / 100) * DRYUSL * CUSL F XH2 = (HYDRO1 + HYDRO2 + HYDRO3) / COAL F FITRO1 = (UIEA(4) / 100) * DRYIEA * CIEA F FITRO2 = (UPRB(4) / 100) * DRYPRB * CPRB F FITRO3 = (UUSL(4) / 100) * DRYUSL * CUSL F XN2 = (FITRO1 + FITRO2 + FITRO3) / COAL F CHLOR1 = (UIEA(5) / 100) * DRYIEA * CIEA F CHLOR2 = (UPRB(5) / 100) * DRYPRB * CPRB F CHLOR3 = (UUSL(5) / 100) * DRYUSL * CUSL F XCL2 = (CHLOR1 + CHLOR2 + CHLOR3) / COAL F SULFR1 = (UIEA(6) / 100) * DRYIEA * CIEA F SULFR2 = (UPRB(6) / 100) * DRYPRB * CPRB F SULFR3 = (UUSL(6) / 100) * DRYUSL * CUSL F XS = (SULFR1 + SULFR2 + SULFR3) / COAL F OXYGN1 = (UIEA(7) / 100) * DRYIEA * CIEA F OXYGN2 = (UPRB(7) / 100) * DRYPRB * CPRB F OXYGN3 = (UUSL(7) / 100) * DRYUSL * CUSL F XO2 = (OXYGN1 + OXYGN2 + OXYGN3) / COAL F XH2O=(PIEA(1)*CIEA+PPRB(1)*CPRB+PUSL(1)*CUSL)/(COAL*100) C WRITE(NRPT, *) XH2O C WRITE(NRPT, *) XH2 C WRITE(NRPT, *) XN2 C WRITE(NRPT, *) XCL2 C WRITE(NRPT, *) XS C WRITE(NRPT, *) XO2 C WRITE(NRPT, *) XC C WRITE(NRPT, *) XASH EXECUTE BEFORE BLOCK DECOMP BLOCK BURN RGIBBS PARAM PRES=101.3 123 PROD H2O / C SS / H2 / N2 / CL2 / HCL / S / O2 / AR / CO / CO2 / NE / HE / CH4 / KR / XE / NO / NO2 / SO2 / SO3 ; This block adjusts the air flow rate such that there is 20 mol % ; excess oxygen present during the coal combustion. CALCULATOR AIR-FLOW DEFINE AIR STREAM-VAR STREAM=AIR SUBSTREAM=MIXED VARIABLE=MOLE-FLOW DEFINE O2COAL MOLE-FLOW STREAM=COAL-OUT SUBSTREAM=MIXED COMPONENT=O2 DEFINE C MOLE-FLOW STREAM=COAL-OUT SUBSTREAM=CISOLID COMPONENT=C DEFINE N2 MOLE-FLOW STREAM=COAL-OUT SUBSTREAM=MIXED COMPONENT=N2 DEFINE H2 MOLE-FLOW STREAM=COAL-OUT SUBSTREAM=MIXED COMPONENT=H2 DEFINE S MOLE-FLOW STREAM=COAL-OUT SUBSTREAM=MIXED COMPONENT=S F XS = 0.21 ; CMIXED IS THE MOLE FLOW OF CARBON IN THE COAL-OUT MIXED SUBSTREAM F AIR = ((C + 2*N2 + 0.5*H2 + S)* (1 + XS) - O2COAL) / 0.2094 EXECUTE BEFORE BLOCK BURN BLOCK HTRANS HEATER PARAM TEMP=320 PRES=0.0 NPHASE=2 ; Neill and Gunter ; PARAM TEMP=622 PRES=0.0 NPHASE=2 ; Boiler design data BLOCK SEPARATE SSPLIT FRAC MIXED FLUE-AHT 1.0 FRAC CISOLID FLUE-AHT 0.0 FRAC NC FLUE-AHT 0.0 ; The air heater outlet temperature is taken from the Neil and Gunter ; study. BLOCK AIR-HEAT HEATER ; PARAM TEMP=134 PARAM TEMP=247 BLOCK SCRUB1 SEP2 FRAC STREAM=IN-SCRUB COMPS=N2 CO2 H2O FRACS=1 1 1 FRAC STREAM=WASTE1 COMPS=H2 S O2 AR NE HE KR XE CO NO NO2 SO2 SO3 & FRACS= 1 1 1 1 1 1 1 1 1 1 1 1 1 BLOCK SCRUB2 FLASH2 PARAM TEMP=40 PRES=0 124 BLOCK CLCHNG1 CLCHNG ; HP turbine and FWP A ; Flowsheet FLOWSHEET HP BLOCK BOIL IN=H2O-BOIL OUT="ST_MAIN" "Q_BOIL" BLOCK "HP_SEP1" IN="ST_MAIN" OUT=ST-FPT1 ST-HPX BLOCK VALVE1 IN=ST-HPX OUT=ST-HP BLOCK HP1 IN=ST-HP OUT="HP_1X" "W_HP" BLOCK "HP_SEP2" IN="HP_1X" OUT=ST-REHT ST-FWPA BLOCK REHT IN=ST-REHT OUT=ST-IPX "Q_REHT" ; Streams ; specify the heat and work streams in the flowsheet DEF-STREAMS HEAT "Q_BOIL" "Q_REHT" DEF-STREAMS WORK "W_HP" STREAM H2O-BOIL TEMP=487.91 PRES=2700 MASS-FLOW=3358670 MOLE-FRAC H2O 1 ; Blocks BLOCK VALVE1 VALVE PARAM P-OUT=2236.19 ; This design spec maintains constant volumetric flow rate into HP section DESIGN-SPEC PRESOUT1 DEFINE F STREAM-PROP STREAM=ST-HP PROPERTY=VFLOW SPEC "F" TO "1.155e6" TOL-SPEC "0.001e6" 125 ; NB: @ 50% plant load, the ST-HP pressure is 1080.68 psia VARY BLOCK-VAR BLOCK=VALVE1 SENTENCE=PARAM VARIABLE=P-OUT LIMITS "900" "2365" BLOCK "HP_SEP1" FSPLIT MASS-FLOW ST-FPT1 7000 BLOCK "HP_SEP2" FSPLIT MASS-FLOW ST-FWPA 334659 CALCULATOR "C_HP_SEP" DESCRIPTION "Specify steam extracted for FW preheating from HP section" DEFINE FREF STREAM-VAR STREAM=ST-HP VARIABLE=MASS-FLOW DEFINE FA BLOCK-VAR BLOCK="HP_SEP2" SENTENCE=MASS-FLOW VARIABLE=FLOW & ID1=ST-FWPA F FA = 0.1231 * FREF - 0.7894e5 READ-VARS FREF WRITE-VARS FA BLOCK REHT HEATER PARAM TEMP=1000 ; This design spec maintains outlet temperature of 1000 F from VALVE2 DESIGN-SPEC TEMPOUT DEFINE T STREAM-VAR STREAM=ST-IP VARIABLE=TEMP SPEC "T" TO "1000" TOL-SPEC "0.5" VARY BLOCK-VAR BLOCK=REHT SENTENCE=PARAM VARIABLE=TEMP LIMITS "1000" "1100" BLOCK BOIL HEATER PARAM TEMP=1000 PRES=2365 BLOCK HP1 COMPR PARAM TYPE=ISENTROPIC PRATIO=0.282 SEFF=0.904 126 CALCULATOR "C_HP1_P" DESCRIPTION "Specify the pressure ratio of HP1" DEFINE FLOW STREAM-VAR STREAM=ST-HP VARIABLE=MASS-FLOW DEFINE PRATIO BLOCK-VAR BLOCK=HP1 SENTENCE=PARAM VARIABLE=PRATIO F PRATIO = -0.4820e-02 * (FLOW/1E6) + 0.2944 EXECUTE BEFORE HP1 ; IP turbine and FWP B, C, and D ; Flowsheet FLOWSHEET IP BLOCK VALVE2 IN=ST-IPX OUT=ST-IP BLOCK "IP_SEP1" IN=ST-IP OUT="IP_02" "IP_03" BLOCK IP2 IN="IP_02" OUT="IP_2X" "W_IP2" BLOCK "IP_SEP2" IN="IP_2X" OUT=ST-FWPC "IP_12" BLOCK IP1 IN="IP_12" OUT=IP-1LP "W_IP1" BLOCK IP3 IN="IP_03" OUT="IP_3X1" "W_IP3" BLOCK "IP_SEP3" IN="IP_3X1" OUT="IP_3X2" "IP_34" BLOCK IP4 IN="IP_34" OUT="IP_4X" "W_IP4" BLOCK "IP_SEP4" IN="IP_3X2" OUT="ST-FPT2" "ST-FWPB" BLOCK "IP_SEP5" IN="IP_4X" OUT=IP-4LP ST-FWPD BLOCK "IP_COMB" IN=IP-1LP IP-4LP OUT=ST-LPX BLOCK EXTRACT IN=ST-LPX OUT=ST-REB ST-LP BLOCK "IP_SHAFT" IN="W_IP1" "W_IP2" "W_IP3" "W_IP4" OUT="W_IP" ; Streams DEF-STREAMS WORK "W_IP1" "W_IP2" "W_IP3" "W_IP4" "W_IP" ; Blocks 127 BLOCK VALVE2 VALVE PARAM P-OUT=560.18 DESIGN-SPEC PRESOUT2 DEFINE F STREAM-PROP STREAM=ST-IP PROPERTY=VFLOW SPEC "F" TO "4.531e6" TOL-SPEC "0.009e6" ; NB: @ 50% plant load, the ST-IP pressure is 260 psia VARY BLOCK-VAR BLOCK=VALVE2 SENTENCE=PARAM VARIABLE=P-OUT LIMITS "250" "600" BLOCK "IP_COMB" MIXER BLOCK "IP_SEP1" FSPLIT FRAC "IP_02" 0.50 BLOCK "IP_SEP2" FSPLIT MASS-FLOW "ST-FWPC" 128853 BLOCK "IP_SEP3" FSPLIT MASS-FLOW "IP_3X2" 227662 ;sum of ST-FWPB and ST-FPT2 BLOCK "IP_SEP4" FSPLIT MASS-FLOW ST-FWPB 143920 BLOCK "IP_SEP5" FSPLIT MASS-FLOW ST-FWPD 136359 BLOCK EXTRACT FSPLIT FRAC ST-REB 0.00 CALCULATOR "C_IP_SEP" DESCRIPTION "Specify steam extracted for FW preheating from IP section" DEFINE FREF STREAM-VAR STREAM=ST-IP VARIABLE=MASS-FLOW DEFINE FBP BLOCK-VAR BLOCK="IP_SEP3" SENTENCE=MASS-FLOW VARIABLE=FLOW & ID1="IP_3X2" DEFINE FB BLOCK-VAR BLOCK="IP_SEP4" SENTENCE=MASS-FLOW VARIABLE=FLOW & ID1=ST-FWPB 128 DEFINE FD BLOCK-VAR BLOCK="IP_SEP5" SENTENCE=MASS-FLOW VARIABLE=FLOW & ID1=ST-FWPD F FB = 0.5389e-1 * FREF - 0.1685e5 F FP = 0.2684e-1 * FREF + 0.1948e4 F FBP = FB + FP F FC = 0.5095e-1 * FREF - 0.2440e5 F FD = 0.5236e-1 * FREF - 0.2077e5 READ-VARS FREF WRITE-VARS FB FBP FD DESIGN-SPEC "C_IPSEP2" DEFINE Q BLOCK-VAR BLOCK="FWP_C-C" SENTENCE=RESULTS VARIABLE=NET-DUTY SPEC "Q" TO "0" TOL-SPEC "1e4" VARY BLOCK-VAR BLOCK="IP_SEP2" SENTENCE=MASS-FLOW VARIABLE=FLOW & ID1=ST-FWPC LIMITS "50000" "150000" BLOCK IP1 COMPR PARAM TYPE=ISENTROPIC PRATIO=0.517 SEFF=0.902 NPHASE=2 BLOCK IP2 COMPR PARAM TYPE=ISENTROPIC PRATIO=0.233 SEFF=0.910 NPHASE=2 BLOCK IP3 COMPR PARAM TYPE=ISENTROPIC PRATIO=0.455 SEFF=0.895 NPHASE=2 BLOCK IP4 COMPR PARAM TYPE=ISENTROPIC PRATIO=0.265 SEFF=0.914 NPHASE=2 BLOCK "IP_SHAFT" MIXER ; LP turbine and FWP E, F, AND G 129 ; Flowsheet FLOWSHEET LP BLOCK "LP_SEP1" IN=ST-LP OUT="LP_012" "LP_056" BLOCK "LP_SEP2" IN="LP_012" OUT="LP_01" "LP_02" BLOCK LP1 IN="LP_01" OUT=ST-FWPF "W_LP1" BLOCK LP2 IN="LP_02" OUT="LP_2X" "W_LP2" BLOCK "LP_SEP3" IN="LP_2X" OUT="LP_23" ST-2FWPG BLOCK LP3 IN="LP_23" OUT="LP_3CR" "W_LP3" BLOCK "LP_SEP4" IN="LP_056" OUT="LP_05" "LP_06" BLOCK LP6 IN="LP_06" OUT=ST-FWPE "W_LP6" BLOCK LP5 IN="LP_05" OUT="LP_5X" "W_LP5" BLOCK "LP_SEP5" IN="LP_5X" OUT="LP_45" ST-5FWPG BLOCK LP4 IN="LP_45" OUT="LP_4CR" "W_LP4" BLOCK "LP_COMB1" IN="LP_3CR" "LP_4CR" OUT=ST-CNDR BLOCK "LP_COMB2" IN=ST-2FWPG ST-5FWPG OUT=ST-FWPG BLOCK "LP_SHAFT" IN="W_LP1" "W_LP2" "W_LP3" "W_LP4" & "W_LP5" "W_LP6" OUT="W_LP" ; Streams DEF-STREAMS WORK "W_LP1" "W_LP2" "W_LP3" "W_LP4" "W_LP5" "W_LP6" "W_LP" ; specify the material streams in the flowsheet ; Blocks BLOCK "LP_COMB1" MIXER BLOCK "LP_COMB2" MIXER BLOCK "LP_SEP1" FSPLIT FRAC "LP_012" 0.50 BLOCK "LP_SEP2" FSPLIT MASS-FLOW "LP_01" 89306 ; flow of ST-FWPF BLOCK "LP_SEP3" FSPLIT 130 MASS-FLOW "ST-2FWPG" 63085 ; half of ST-FWPG BLOCK "LP_SEP4" FSPLIT MASS-FLOW "LP_06" 135578 ; flow of ST-FWPE BLOCK "LP_SEP5" FSPLIT MASS-FLOW "ST-5FWPG" 63086 ; other half of ST-FWPG CALCULATOR "C_LP_SEP" DESCRIPTION "Specify steam extracted for FW preheating from LP section" DEFINE FREF STREAM-VAR STREAM=ST-LP VARIABLE=MASS-FLOW DEFINE FE BLOCK-VAR BLOCK="LP_SEP4" SENTENCE=MASS-FLOW VARIABLE=FLOW & ID1="LP_06" DEFINE FF BLOCK-VAR BLOCK="LP_SEP2" SENTENCE=MASS-FLOW VARIABLE=FLOW & ID1="LP_01" DEFINE FG2 BLOCK-VAR BLOCK="LP_SEP3" SENTENCE=MASS-FLOW VARIABLE=FLOW & ID1=ST-2FWPG DEFINE FG5 BLOCK-VAR BLOCK="LP_SEP5" SENTENCE=MASS-FLOW VARIABLE=FLOW & ID1=ST-5FWPG F FE = 0.6311e-1 * FREF - 0.2228e5 F FF = 0.4162e-1 * FREF - 0.1475e5 F FG = 0.6170e-1 * FREF - 0.2538e5 F FG2 = FG / 2 F FG5 = FG2 READ-VARS FREF WRITE-VARS FE FF FG2 FG5 BLOCK LP1 COMPR PARAM TYPE=ISENTROPIC PRATIO=0.151 SEFF=0.910 NPHASE=2 BLOCK LP2 COMPR PARAM TYPE=ISENTROPIC PRATIO=0.068 SEFF=0.907 NPHASE=2 BLOCK LP3 COMPR PARAM TYPE=ISENTROPIC PRES=0.686 SEFF=0.640 NPHASE=2 BLOCK LP4 COMPR PARAM TYPE=ISENTROPIC PRES=0.686 SEFF=0.640 NPHASE=2 131 CALCULATOR "C_LP_P" DESCRIPTION "Set the outlet P of LP3 and LP4 equal to the condenser" DEFINE PCOND BLOCK-VAR BLOCK=CONDENSE SENTENCE=PARAM VARIABLE=PRES DEFINE PLP3 BLOCK-VAR BLOCK=LP3 SENTENCE=PARAM VARIABLE=PRES DEFINE PLP4 BLOCK-VAR BLOCK=LP4 SENTENCE=PARAM VARIABLE=PRES F PLP3 = PCOND F PLP4 = PCOND EXECUTE BEFORE LP3 CALCULATOR "C_LP_EFF" DESCRIPTION "Use correlation to set LP3 and LP4 isentropic efficiency" DEFINE QOUT STREAM-PROP STREAM=ST-CNDR PROPERTY=VFLOW DEFINE SEFF3 BLOCK-VAR BLOCK=LP3 SENTENCE=PARAM VARIABLE=SEFF DEFINE SEFF4 BLOCK-VAR BLOCK=LP4 SENTENCE=PARAM VARIABLE=SEFF F ETA = -0.4016 * (QOUT/1e9) + 0.9867 F SEFF3 = ETA F SEFF4 = ETA EXECUTE BEFORE CONDENSE READ-VARS QOUT C WRITE-VARS SEFF3 SEFF4 BLOCK LP5 COMPR PARAM TYPE=ISENTROPIC PRATIO=0.068 SEFF=0.907 NPHASE=2 BLOCK LP6 COMPR PARAM TYPE=ISENTROPIC PRATIO=0.435 SEFF=0.901 NPHASE=2 BLOCK "LP_SHAFT" MIXER ; Feedwater pump turbine ; Flowsheet 132 FLOWSHEET FPT BLOCK FPT1 IN=ST-FPT1 OUT="FPT_1X" "W_FPT1" BLOCK "FPT_COMB" IN=ST-FPT2 "FPT_1X" OUT="FPT_12" BLOCK FPT2 IN="FPT_12" OUT=STFPT-CN "W_FPT2" BLOCK "FP_SHAFT" IN="W_FPT1" "W_FPT2" OUT="W_FPT" ; Streams DEF-STREAMS WORK "W_FPT1" "W_FPT2" "W_FPT" ; Blocks BLOCK "FPT_COMB" MIXER BLOCK FPT1 COMPR PARAM TYPE=ISENTROPIC PRES=100 SEFF=0.153 NPHASE=2 BLOCK FPT2 COMPR PARAM TYPE=ISENTROPIC PRES=0.686 SEFF=0.795 NPHASE=2 CALCULATOR "C_FPT_P" DESCRIPTION "Specifies the outlet pressure of FPT1 and FPT2" DEFINE PREF STREAM-VAR STREAM=ST-FPT2 VARIABLE=PRES DEFINE PCOND BLOCK-VAR BLOCK=CONDENSE SENTENCE=PARAM VARIABLE=PRES DEFINE PFPT1 BLOCK-VAR BLOCK=FPT1 SENTENCE=PARAM VARIABLE=PRES DEFINE PFPT2 BLOCK-VAR BLOCK=FPT2 SENTENCE=PARAM VARIABLE=PRES F PFTP1 = PREF F PFTP2 = PCOND READ-VARS PREF PCOND WRITE-VARS PFPT1 PFPT2 BLOCK "FP_SHAFT" MIXER 133 ; Feed water preheater train ; Flowsheet FLOWSHEET FWP BLOCK "FWP_A-H" IN=ST-FWPA Q-FWPA OUT="STFWP_AB" BLOCK "FWP_A-C" IN=H2O-FWPA OUT=H2O-BOIL Q-FWPA BLOCK "FWP_B-H" IN=ST-FWPB "STFWP_AB" Q-FWPB OUT="STFWP_BC" BLOCK "FWP_B-C" IN=H2O-FWPB OUT=H2O-FWPA Q-FWPB ; dearator and pump BLOCK "FWP_C" IN="STFWP_BC" ST-FWPC H2O-FWPC OUT=H2-PUMP BLOCK FWPUMP2 IN=H2-PUMP "W_FPT" OUT=IN-PUMP BLOCK "FWP_C-C" IN=IN-PUMP OUT=H2O-FWPB BLOCK "FWP_D-H" IN=ST-FWPD Q-FWPD OUT="STFWP_DE" BLOCK "FWP_D-C" IN=H2O-FWPD H2O-REB OUT=H2O-FWPC Q-FWPD BLOCK "FWP_E-H" IN=ST-FWPE "STFWP_DE" Q-FWPE OUT="STFWP_EF" BLOCK "FWP_E-C" IN=H2O-FWPE OUT=H2O-FWPD Q-FWPE BLOCK "FWP_F-H" IN=ST-FWPF "STFWP_EF" Q-FWPF OUT="STFWP_FG" BLOCK "FWP_F-C" IN=H2O-FWPF OUT=H2O-FWPE Q-FWPF BLOCK "FWP_G-H" IN=ST-FWPG "STFWP_FG" Q-FWPG OUT="STFWP_GC" BLOCK "FWP_G-C" IN=H2O-FWPG OUT=H2O-FWPF Q-FWPG ; Streams ; I need to define the heat streams in this flowsheet section DEF-STREAMS HEAT Q-FWPA Q-FWPB Q-FWPD Q-FWPE Q-FWPF Q-FWPG ; Blocks 134 ; feed water preheater "A" BLOCK "FWP_A-H" HEATER PARAM PRES=0 BLOCK "FWP_A-C" HEATER PARAM TEMP=487.91 CALCULATOR "T_FWPA" DESCRIPTION "Calculate the cold-side outlet temperature for FWPA" DEFINE FFWPA STREAM-VAR STREAM=H2O-FWPA VARIABLE=MASS-FLOW DEFINE TFWPA BLOCK-VAR BLOCK="FWP_A-C" SENTENCE=PARAM VARIABLE=TEMP F TFWPA = 0.8546e2 * dlog(FFWPA) - 0.7963e3 EXECUTE BEFORE "FWP_A-C" ; feed water preheater "B" BLOCK "FWP_B-H" HEATER PARAM PRES=0 BLOCK "FWP_B-C" HEATER PARAM TEMP=400.56 CALCULATOR "T_FWPB" DESCRIPTION "Calculate the cold-side outlet temperature for FWPB" DEFINE FFWPB STREAM-VAR STREAM=H2O-FWPB VARIABLE=MASS-FLOW DEFINE TFWPB BLOCK-VAR BLOCK="FWP_B-C" SENTENCE=PARAM VARIABLE=TEMP F TFWPB = 0.6840e2 * dlog(FFWPB) - 0.6272e3 EXECUTE BEFORE "FWP_B-C" ; feed water preheater "C" (dearator) and feed water pump BLOCK "FWP_C" MIXER BLOCK FWPUMP2 PUMP ; PARAM PRES=2700 BLOCK "FWP_C-C" HEATER PARAM TEMP=351.19 135 CALCULATOR "T_FWPC" DESCRIPTION "Calculate the cold-side outlet temperature for FWPC" ; using the outlet mass flow rate is easier than having to sum ; the three input mass flow rates DEFINE FFWPC STREAM-VAR STREAM=IN-PUMP VARIABLE=MASS-FLOW DEFINE TFWPC BLOCK-VAR BLOCK="FWP_C-C" SENTENCE=PARAM VARIABLE=TEMP F TFWPC = 0.6468e2 * dlog(FFWPC) - 0.6212e3 EXECUTE BEFORE "FWP_C-C" ; feed water preheater "D" BLOCK "FWP_D-H" HEATER PARAM PRES=0 BLOCK "FWP_D-C" HEATER PARAM TEMP=293.20 CALCULATOR "T_FWPD" DESCRIPTION "Calculate the cold-side outlet temperature for FWPD" DEFINE FFWPD STREAM-VAR STREAM=H2O-FWPD VARIABLE=MASS-FLOW DEFINE FREB STREAM-VAR STREAM=H2O-REB VARIABLE=MASS-FLOW DEFINE TFWPD BLOCK-VAR BLOCK="FWP_D-C" SENTENCE=PARAM VARIABLE=TEMP F TFWPD = 0.5537e2 * dlog(FFWPD + FREB) - 0.5274e3 EXECUTE BEFORE "FWP_D-C" ; feed water preheater "E" BLOCK "FWP_E-H" HEATER PARAM PRES=0 BLOCK "FWP_E-C" HEATER PARAM TEMP=241.55 CALCULATOR "T_FWPE" DESCRIPTION "Calculate the cold-side outlet temperature for FWPE" DEFINE FFWPE STREAM-VAR STREAM=H2O-FWPE VARIABLE=MASS-FLOW 136 DEFINE TFWPE BLOCK-VAR BLOCK="FWP_E-C" SENTENCE=PARAM VARIABLE=TEMP F TFWPE = 0.4602e2 * dlog(FFWPE) - 0.4405e3 EXECUTE BEFORE "FWP_E-C" ; feed water preheater "F" BLOCK "FWP_F-H" HEATER PARAM PRES=0 BLOCK "FWP_F-C" HEATER PARAM TEMP=186.37 CALCULATOR "T_FWPF" DESCRIPTION "Calculate the cold-side outlet temperature for FWPF" DEFINE FFWPF STREAM-VAR STREAM=H2O-FWPF VARIABLE=MASS-FLOW DEFINE TFWPF BLOCK-VAR BLOCK="FWP_F-C" SENTENCE=PARAM VARIABLE=TEMP F TFWPF = 0.3788e2 * dlog(FFWPF) - 0.3752e3 EXECUTE BEFORE "FWP_F-C" ; feed water preheater "G" BLOCK "FWP_G-H" HEATER PARAM PRES=0 BLOCK "FWP_G-C" HEATER PARAM TEMP=150.01 CALCULATOR "T_FWPG" DESCRIPTION "Calculate the cold-side outlet temperature for FWPG" DEFINE FFWPG STREAM-VAR STREAM=H2O-FWPG VARIABLE=MASS-FLOW DEFINE TFWPG BLOCK-VAR BLOCK="FWP_G-C" SENTENCE=PARAM VARIABLE=TEMP F TFWPG = 0.3033e2 * dlog(FFWPG) - 0.2996e3 EXECUTE BEFORE "FWP_G-C" ; Condensor specification 137 ; Flowsheet FLOWSHEET CNDR BLOCK "CND_COMB" IN="STFWP_GC" ST-CNDR STFPT-CN OUT=H2O-CNDR BLOCK CONDENSE IN=H2O-CNDR OUT=H2O-MAIN BLOCK FWPUMP1 IN=H2O-MAIN OUT=H2O-FWPG ; Blocks BLOCK "CND_COMB" MIXER BLOCK CONDENSE HEATER PARAM VFRAC=0 PRES=0.688 BLOCK FWPUMP1 PUMP PARAM PRES=128 ; MEA Absorption specification ; Flowsheet FLOWSHEET MEA BLOCK CLCHNG1 IN=FLUE-GAS OUT=FLUE-BLO BLOCK BLOWER IN=FLUE-BLO OUT=FLUE-DCC "P_BLOW" BLOCK "H2O_PUMP" IN=H2O-PUMP OUT=H2O-DCC P-H2OP BLOCK DCC IN=FLUE-DCC H2O-DCC OUT=FLUE-ABS H2O-OUT BLOCK ABSORBER IN=FLUE-ABS LEAN-ABS OUT=STACK RICH-PUM BLOCK "RICH_PUM" IN=RICH-PUM OUT=RICH-HX P-RICHP BLOCK STRIPPER IN=RICH-STR OUT=CO2-COMP LEAN-HX BLOCK "CO2_COMP" IN=CO2-COMP OUT=CO2 ST1 ST2 ST3 "P_COMP" BLOCK HEATX IN=RICH-HX LEAN-HX OUT=RICH-STR LEAN-MIX BLOCK MIXER IN=MAKE-UP LEAN-MIX OUT=LEAN-COO 138 BLOCK COOLER IN=LEAN-COO OUT=LEAN-ABS BLOCK POWER IN="P_BLOW" P-H2OP P-RICHP "P_COMP" OUT="P_DEMAND" BLOCK REBOIL IN=ST-REB OUT=H2O-REB "Q_REB" ; Stream Specification ; specify the heat and work streams in the flowsheet DEF-STREAMS WORK "P_BLOW" P-H2OP P-RICHP "P_COMP" "P_DEMAND" DEF-STREAMS HEAT "Q_REB" ; Cooling water temperature for Lake Erie is not given. 12C is summer ; mean temperature form IEA technical specifications document... STREAM H2O-PUMP TEMP=12 PRES=101.3 IN-UNITS SI PRESSURE=kPa TEMPERATURE=C PDROP=kPa MOLE-FLOW H2O 70 ; The mole flow of H2O and MEA are adjusted by the calculator block C_MAEKUP STREAM MAKE-UP TEMP=40 PRES=101.3 IN-UNITS SI PRESSURE=kPa TEMPERATURE=C PDROP=kPa MOLE-FLOW H2O 1 / MEA 1 ; tear streams ... ; Note: 12.6 M MEA is 30 wt% STREAM LEAN-ABS TEMP=40 PRES=101.3 MOLE-FLOW=87.1 MOLE-FRAC MEA 0.126 / H2O 0.874 / CO2 .0315 ; Note: F is obtained from absorber results STREAM LEAN-HX PRES=186 VFRAC=0 MOLE-FLOW=87.1 IN-UNITS SI PRESSURE=kPa TEMPERATURE=C PDROP=kPa MOLE-FRAC MEA 0.126 / H2O 0.874 / CO2 .0315 ; Block Specification ; BLOCK BLOWER COMPR PARAM TYPE=ISENTROPIC DELP=83.6 SEFF=0.90 ; 139 ; BLOCK "H2O_PUMP" PUMP PARAM DELP=83.6 ; ; This block cools the flue gas stream with water. BLOCK DCC FLASH2 IN-UNITS SI PRESSURE=kPa TEMPERATURE=C PDROP=kPa PARAM TEMP=40 PRES=0 ; BLOCK ABSORBER RATEFRAC IN-UNITS SI PRESSURE=kPa TEMPERATURE=C PDROP=kPa PARAM NCOL=1 TOT-SEGMENT=10 EQUILIBRIUM=NO & INIT-MAXIT=30 MAXIT=30 INIT-TOL=1E-2 TOL=9E-3 ;INIT-OPTION=CHEMICAL COL-CONFIG 1 10 CONDENSER=NO REBOILER=NO TRAY-SPECS 1 1 10 TRAY-TYPE=SIEVE DIAM-EST=20 & PERCENT-FLOOD=70 TRAY-SPACING=192 & WEIRHT=16 FEEDS FLUE-ABS 1 11 ABOVE-SEGMENT / LEAN-ABS 1 1 ABOVE-SEGMENT PRODUCTS STACK 1 1 V / RICH-PUM 1 10 L P-SPEC 1 1 101.3 / 1 10 176.9 SUBROUTINE PRESS-DROP=trayp COL-SPECS 1 MOLE-RDV=1 ; Provides information on proximity to flooding conditions and pressure drop ; on each nonequilibrium segment REPORT FLOOD-INFO ; The following line causes the Murphree efficiencies to be tabulated. SEGMENT-REPORT SEGMENT-OPTION=ALL-SEGMENTS FORMAT=PROFILE & COMP-EFF=YES PROPERTIES=LPHASE VPHASE WIDE=YES ; ; BLOCK "RICH_PUM" PUMP 140 PARAM DELP=0 ; ; BLOCK STRIPPER RATEFRAC IN-UNITS SI PRESSURE=kPa TEMPERATURE=C PDROP=kPa PARAM NCOL=1 TOT-SEGMENT=9 EQUILIBRIUM=NO INIT-MAXIT=45 & MAXIT=45 INIT-TOL=1E-2 TOL=9E-3 ;INIT-OPTION=CHEMICAL COL-CONFIG 1 9 CONDENSER=YES REBOILER=YES TRAY-SPECS 1 2 8 TRAY-TYPE=SIEVE DIAM-EST=20 & PERCENT-FLOOD=70 TRAY-SPACING=216 & WEIRHT=18 FEEDS RICH-STR 1 2 ABOVE-SEGMENT PRODUCTS CO2-COMP 1 1 V / LEAN-HX 1 9 L P-SPEC 1 1 101.3 / 1 9 186 SUBROUTINE PRESS-DROP=trayp COL-SPECS 1 MOLE-RDV=1 MOLE-RR=0.4 MOLE-B:F=0.95 DB:F-PARAMS 1 SPEC 1 MOLE-FLOW 2.45 COMPS=CO2 STREAMS=CO2-COMP VARY 1 MOLE-B:F COL=1 SPEC 2 TEMP 70 SEGMENT=1 COL=1 PHASE=L VARY 2 MOLE-RR COL=1 ; Provides information on proximity to flooding conditions and pressure drop ; on each nonequilibrium segment REPORT FLOOD-INFO ; The following line causes the Murphree efficiencies to be tabulated. SEGMENT-REPORT SEGMENT-OPTION=ALL-SEGMENTS FORMAT=PROFILE & COMP-EFF=YES PROPERTIES=LPHASE VPHASE WIDE=YES ; ; Shortcut heat exchanger calculation. ; 10 degree temperature approach at the hot stream outlet ; U = 1134 W / m?2 C (taken from Perry's for H2O-H2O liquid-liquid system) BLOCK HEATX HEATX 141 IN-UNITS SI PRESSURE=kPa TEMPERATURE=C PDROP=kPa PARAM DELT-HOT=10 FEEDS HOT=LEAN-HX COLD=RICH-HX PRODUCTS HOT=LEAN-MIX COLD=RICH-STR HEAT-TR-COEF U=1134 CALCULATOR "C_MAKEUP" DESCRIPTION "Set MEA and H2O flow rate in make-up stream" ; Streams for water balance DEFINE H2OFL MOLE-FLOW STREAM=FLUE-ABS COMPONENT=H2O DEFINE H2OAB MOLE-FLOW STREAM=STACK COMPONENT=H2O DEFINE H2OST MOLE-FLOW STREAM=CO2-COMP COMPONENT=H2O DEFINE MEAAB MOLE-FLOW STREAM=STACK COMPONENT=MEA DEFINE MEAST MOLE-FLOW STREAM=CO2-COMP COMPONENT=MEA DEFINE MEAMU MOLE-FLOW STREAM=MAKE-UP COMPONENT=MEA DEFINE H2OMU MOLE-FLOW STREAM=MAKE-UP COMPONENT=H2O F MEAMU = MEAAB + MEAST F H2OMU = H2OAB + H2OST - H2OFL EXECUTE BEFORE BLOCK MIXER BLOCK MIXER MIXER BLOCK COOLER HEATER IN-UNITS SI PRESSURE=kPa TEMPERATURE=C PDROP=kPa PARAM TEMP=40 PRES=101.3 BLOCK "CO2_COMP" MCOMPR IN-UNITS SI PRESSURE=kPa TEMPERATURE=C PDROP=kPa PARAM NSTAGE=4 TYPE=ISENTROPIC PRES=110 FEEDS CO2-COMP 1 PRODUCTS ST1 1 L / ST2 2 L / ST3 3 L / CO2 4 / "P_COMP" GLOBAL COMPR-SPECS 1 SEFF=0.90 COOLER-SPECS 1 TEMP=25 BLOCK REBOIL HEATER 142 IN-UNITS SI PRESSURE=kPa TEMPERATURE=C PDROP=kPa PARAM PRES=0 VFRAC=0 BLOCK POWER MIXER ; END: flowsheet specification ; Convergence options SIM-OPTIONS RESTART=YES CONVERGENCE "ABS_LOOP" BROYDEN DESCRIPTION "Converge Absorber-side recycle and set CO2 loading" TEAR LEAN-ABS SPEC ALPHA CONVERGENCE "STR_LOOP" BROYDEN DESCRIPTION "Converge Stripper-style recycle and set Stripper P" TEAR LEAN-HX SPEC "STR_PRES" CONVERGENCE "ST_CYCLE" BROYDEN DESCRIPTION "Converge steam cycle tear streams and BOILFLOW spec" TEAR Q-FWPA / Q-FWPB / Q-FWPD / Q-FWPE / Q-FWPF / Q-FWPG / H2O-BOIL SPEC "C_IPSEP2" SPEC BOILFLOW PARAM MAXIT=60 CONVERGENCE EXTRACT SECANT DESCRIPTION "Specifies parameters used to set steam extraction" SPEC EXTRACT ;SEQUENCE FLOW "ABS_LOOP" ABSORBER "RICH_PUM" & ; "STR_LOOP" HEATX STRIPPER & ; (RETURN "STR_LOOP") & ; "C_MAKEUP" MIXER COOLER & ; (RETURN "ABS_LOOP") 143 ; This paragraph specifies the convergence order for user-defined ; convergence blocks CONV-ORDER "STR_LOOP" "ABS_LOOP" "ST_CYCLE" EXTRACT ; Design specification: FCOAL ; This design specification adjusts adjusts the coal flow rate such that ; there is sufficient heat generated to satisfy the duties of BOIL and REHT. DESIGN-SPEC FCOAL DEFINE QBOIL INFO-VAR INFO=HEAT VARIABLE=DUTY STREAM="Q_BOIL" DEFINE QREHT INFO-VAR INFO=HEAT VARIABLE=DUTY STREAM="Q_REHT" DEFINE QFURN INFO-VAR INFO=HEAT VARIABLE=DUTY STREAM="Q_FURN" ; The boiler efficiency is 90% F EFF = 0.815 ; 1 kW = 3412.2 Btu/h F G = 3412.2 SPEC "QFURN" TO "-(QBOIL + QREHT) / EFF" TOL-SPEC "1*G" VARY STREAM-VAR STREAM=COAL-IN SUBSTREAM=NC VARIABLE=MASS-FLOW LIMITS "10" "793800" ; Design Spec: BOILFLOW ; Adjusts the flow rate of feed water until the desired value is achieved. DESIGN-SPEC BOILFLOW DEFINE FLOW STREAM-VAR STREAM=H2O-BOIL VARIABLE=MASS-FLOW ; 100% 3358670 ; 75% 2446607 ; 50% 1619896 SPEC "FLOW - 3358670" TO "0.0" TOL-SPEC "1" 144 VARY STREAM-VAR STREAM=H2O-BOIL VARIABLE=MASS-FLOW LIMITS "809948" "3400000" ; Calculator block: C_POWER ; Calculates mechanical losses of main and BFP turbines, generator losses, ; exciter power to generator, and station service. These are required to ; calculate the turbine and unit heat rates. ; MECH, GEN, EXC [=] MW; x [=] MW ; STA [=] MW; x [=] MW ; BFPM [=] kW; x [=] MW CALCULATOR "C_POWER" DEFINE PSUPP INFO-VAR STREAM="P_INTERN" INFO=WORK VARIABLE=POWER DEFINE PDMND INFO-VAR STREAM="P_DEMAND" INFO=WORK VARIABLE=POWER DEFINE PBLOW INFO-VAR STREAM="P_BLOW" INFO=WORK VARIABLE=POWER DEFINE PCOMP INFO-VAR STREAM="P_COMP" INFO=WORK VARIABLE=POWER DEFINE WFPT INFO-VAR STREAM="W_FPT" INFO=WORK VARIABLE=POWER ; 1 hp is equal to 0.745699 kW F F = 0.7456999 ; Convert power from units of kW to MW F PMAIN = -PSUPP / 1e3 F PBFPT = -WFPT / 1e3 F PREQD = PDMND / 1e3 F PMECH = 1.919 F PGEN = (0.1511e-1) * PMAIN + 0.7343 F PEXC = (0.3437e-2) * PMAIN - 0.4078 F PSTA = (0.1110e+2) * DEXP(PMAIN/1e3) - 0.3737e+1 F PBFPM = ((0.4534e+1) * PBFPT + 0.4244e2) / 1000 ; For the calculation of net electric power output, I'm assuming a generator ; efficiency of 90%. It's what David Singh used... F GEFF = 0.90 F EGRSS = PMAIN + PEXC - (PMECH + PGEN + PSTA) F EBLOW = -(PBLOW / GEFF) / 1e3 145 F ECOMP = -(PCOMP / GEFF) / 1e3 F ENET = EGRSS - (PREQD / GEFF) F WRITE(NRPT, '(A,F9.2,A3)') 'INTERNAL POWER ', PMAIN, 'MW' F WRITE(NRPT, '(A,F9.2,A3)') 'EXCITER POWER ', PEXC, 'MW' F WRITE(NRPT, '(A,F9.2,A3)') 'MECHANICAL LOSSES ', -PMECH, 'MW' F WRITE(NRPT, '(A,F9.2,A3)') 'GENERATOR LOSSES ', -PGEN, 'MW' F WRITE(NRPT, '(A,F9.2,A3)') 'STATION SERVICE ', -PSTA, 'MW' F WRITE(NRPT, '(A,F9.2,A3)') 'ELEC-TY, GROSS ', EGRSS, 'MW' F WRITE(NRPT, '(A,F9.2,A3)') 'ELEC-TY, BLOWER ', -EBLOW, 'MW' F WRITE(NRPT, '(A,F9.2,A3)') 'ELEC-TY, CO2_COMP ', -ECOMP, 'MW' F WRITE(NRPT, '(A,F9.2,A3)') 'ELEC-TY, NET ', ENET, 'MW' EXECUTE AFTER SHAFT ; Design specification: COOL-FLU ; This block sets the flow rate of cooling water needed to cool the flue gas ; to the temperature specified in the DCC block. DESIGN-SPEC COOL-FLU DEFINE QDCC BLOCK-VAR BLOCK=DCC SENTENCE=PARAM VARIABLE=QCALC ; 1 kmol/s = 7938 lbmol/h F F = 7938 ; 1 kW = 3412.2 Btu/h F G = 3412.2 SPEC "QDCC" TO "0" TOL-SPEC "1*G" VARY STREAM-VAR STREAM=H2O-PUMP VARIABLE=MOLE-FLOW LIMITS "1*F" "120*F" ; Calculator block: C_RECOV ; This block calculates the CO2 mole flow rate in the output stream ; that corresponds to a desired CO2 recovery. 146 CALCULATOR "C_RECOV" DEFINE CO2IN MOLE-FLOW STREAM=FLUE-GAS COMPONENT=CO2 DEFINE FCO2 BLOCK-VAR BLOCK=STRIPPER SENTENCE=SPEC VARIABLE=VALUE & ID1=1 ; CO2IN has units of lbmol/hr and FCO2 needs to be expressed in kmol/s. ; 1 kmol/s = 7938 lbmol/hr F FCO2 = CO2IN * 0.85 / 7938 EXECUTE BEFORE CONVERGENCE "ABS_LOOP" ; Design specification: ALPHA ; This block sets the CO2 loading of the recycle stream to a specified value. DESIGN-SPEC ALPHA DEFINE CO2 MOLE-FLOW STREAM=LEAN-ABS COMPONENT=CO2 DEFINE MEA MOLE-FLOW STREAM=LEAN-ABS COMPONENT=MEA F ALPHA = CO2 / MEA ; 1 kmol/s = 7938 lbmol/h F F = 7938 SPEC "ALPHA" TO "0.25" TOL-SPEC "0.0025" VARY STREAM-VAR STREAM=LEAN-ABS VARIABLE=MOLE-FLOW LIMITS "30*F" "250*F" ; Design specification: STR-PRES ; This block sets the Stripper reboiler pressure such that the reboiler ; temperature is 121C +- 1C. DESIGN-SPEC "STR_PRES" DEFINE TN STREAM-VAR STREAM=LEAN-HX VARIABLE=TEMP ; Temperature is in units of F; pressure is given in psi. 147 SPEC "TN" TO "250" TOL-SPEC "1.8" VARY BLOCK-VAR BLOCK=STRIPPER SENTENCE=P-SPEC VARIABLE=PRES ID1=1 & ID2=9 LIMITS "14.7" "32" ; Design specification: EXTRACT ; This design specification adjusts the amount of steam extracted from ; the IP/LP crossover pipe such that the reboiler heat duty is satisfied. DESIGN-SPEC EXTRACT DEFINE QREB BLOCK-VAR BLOCK=STRIPPER SENTENCE=RESULTS & VARIABLE=REB-DUTY ID1=1 DEFINE QEXT INFO-VAR STREAM="Q_REB" INFO=HEAT VARIABLE=DUTY ; 1 kW = 3412.2 Btu/h F G = 3412.2 SPEC "QEXT" TO "QREB" TOL-SPEC "10*G" VARY BLOCK-VAR BLOCK=EXTRACT SENTENCE=FRAC VARIABLE=FRAC ID1=ST-REB LIMITS "0.0" "0.83" 148 Glossary α CO2 loading An expression of the CO2 concentration in solution, it is the molar ratio of CO2 to MEA. 149 List of References [1] Colin Alie, Leslie Backham, Eric Croiset, and Peter L Douglas. Simulation of CO2 capture using mea scrubbing: a ?owhseet decomposition method. Energy Conversion and Management, 2004. in publication. 11, 62, 65, 68, 76 [2] W. A. Anderson. Che 045 process equipment sizing and selection course notes, 2001. 102, 103, 104, 107 [3] anon. Coal-?red electricity generation in ontario. Technical report, Ministry of the Environment, March 2001. 10 [4] Anon. Design basis document for reduction of mercury emissions at Nanticoke generating station. Internal report prepared for OPG, November 2002. Con?dential. 13, 15, 20, 82 [5] Anon. Something about coal. Technical report, American Society for Testing and Materials, 2002. 19 [6] anon. Canada's energy future: Scenarios for supply and demand to 2025. Technical report, National Energy Board, July 2003. 6 [7] Adisorn Aroonwilas, Amornvadee Veawab, and Paitoon Tontiwachuwuthikul. Be- haviour of the mass-transfer coef?cient of structured packings in CO2 absorbers with chemical reactions. Journal of Industrial Engineering Chemical Research, 38:2044–2050, 1999. 46, 47, 48 [8] Aspen Technology, Inc., Cambridge, MA, USA. Aspen Physical Property System: Physical Property Methods and Models 11.1, September 2001. 50 [9] Aspen Technology, Inc., Cambridge, MA, USA. Aspen Plus Version 11.1 Getting Started: Modeling Process with Solids, September 2001. 15 [10] Aspen Technology, Inc., Cambridge, MA, USA. Aspen Plus Version 11.1 Unit Operation Models, September 2001. 32 150 [11] Aspen Technology, Inc., Cambridge, MA, USA. Aspen Plus Version 11.1 User Guide, September 2001. 15, 98 [12] Aspen Technology, Inc., Cambridge, MA, USA. Aspen Plus Version 11.1 Service Pack 1 Input Language Guide, March 2002. 55, 58 [13] R Barchas and R Davis. The Kerr-McGee/ABB Lummus Crest technology for the recovery of CO2 from stack gases. Energy Conversion and Management, 33(5– 8):333–340, 1992. 43, 46, 83 [14] A. Chakma, A. K. Mehrotra, and B. Nielsen. Comparison of chemical solvents for mitigating CO2 emissions from coal-?red power plants. Heat Recovery Systems and CHP, 15(2):231–240, 1995. 8, 44, 47, 49, 55, 92 [15] Dan G. Chapel, Carl L. Mariz, and John Ernest. Recovery of CO2 from ?ue gases: Commercial trends. In Canadian Society of Chemical Engineers Annual Meeting, Saskatoon, Saskatchewan, October 1999. 44, 46, 83 [16] Susan Chi and Gary T. Rochelle. Oxidative degradation of monoethanolamine. In First National Conference on Carbon Sequestration, Washington, DC, May 2001. 83 [17] Chih-Chung Chou and Yen-Shiang Shih. A thermodynamic approach to the de- sign and synthesis of plant utility systems. Industrial & Engineering Chemistry Research, 26(6):1100–1108, June 1987. 26 [18] C. David Cooper and F. C. Alley. Air pollution control: a design approach. Wave- land Press, Inc., U.S.A., second edition, 1994. 1, 18 [19] Ralph Cos. Capabilities of hyprotech. In 3rd Workshop of International Test Network for CO2 capture, Apeldoorn, the Netherlands, May 2002. IEA Greenhouse Gas R&D Programme. 11 [20] Umberto Desideri and Alberto Paolucci. Performance modelling of a carbon dioxide removal system for power plants. Energy Conversion and Management, 40:1899–1915, 1999. 11, 42, 45, 49, 54, 57, 58, 78, 92, 93 [21] Energy Section, Manufacturing, Construction & Energy Division. Electric power generating stations 2000. Technical Report 57-206-XIB, Statistics Canada, Decem- ber 2001. 4, 7, 8 [22] Energy Section, Manufacturing, Construction & Energy Division. Electric power generation, transmission, and distribution. Technical Report 57-202-XIB, Statistics Canada, 2002. 3, 5 151 [23] M. R. Erbes and R. H. Eustis. A computer methodology for predicting hte design and off-design performance of utility steam turbine-generators. In Proceedings - American Power Conference, volume 48, pages 318–324. Illinois Institute of Tech- nology, 1986. 26 [24] Paul Feron. CO2 capture using membrane gas absorption - status report. In 5th Workshop of International Test Network for CO2 capture, Pittsburgh, U.S.A., June 2002. IEA Greenhouse Gas R&D Programme. 47 [25] Stefano Freguia and Gary T. Rochelle. Base conditions for modeling of CO2 cap- ture by mea absorption/stripping. In 3rd Workshop of International Test Network for CO2 capture, Apeldoorn, the Netherlands, May 2002. IEA Greenhouse Gas R&D Programme. 11, 42, 49, 54, 97 [26] Stefano Freguia and Gary T. Rochelle. Modeling of CO2 absorption/stripping with monoethanolamine. In 4th Workshop of International Test Network for CO2 capture, Kyoto, Japan, October 2002. IEA Greenhouse Gas R&D Programme. 11, 42, 46, 47, 49, 54, 76, 97 [27] Stefano Freguia and Gary T. Rochelle. Modeling of CO2 absorption/stripping with monoethanolamine. AIChE Journal, 49(7):1676–1687, 2003. 11, 42, 46, 47, 49, 54, 76, 97 [28] George S. Goff and Gary T. Rochelle. Oxidative degradation of aqueous mo- noethanolamine in CO2 capture systems under absorber conditions. In J. Gale and Y. Kaya, editors, Greenhouse Gas Control Technologies: Proceedings of the 6th International Conference on Greenhouse Gas Control Technologies, volume 1, pages 115–119. Elsevier Science Ltd., October 2002. 83 [29] Don W. Green, editor. Perry's chemical engineers' handbook. McGraw-Hill, sev- enth edition, 1997. 56, 102, 106, 107 [30] Ontario Hydro. Nanticoke generating station design heat balance, February 1972. 100%, 75%, and 50% load. 20, 26 [31] Ontario Hydro. Nanticoke generating station units 1–8 boiler design data, October 1978. 100% load. 20 [32] IEA Greenhouse Gas R&D programme. Technical and Financial Assessment Criteria, November 1999. Revision B. 10, 13, 15, 52 152 [33] Masaki Iijima and Takashi Kamijo. Flue gas CO2 recovery and compression cost study for CO2 enhanced oil recovery. In J. Gale and Y. Kaya, editors, Greenhouse Gas Control Technologies: Proceedings of the 6th International Conference on Greenhouse Gas Control Technologies, volume 1, pages 109–114. Elsevier Science Ltd., October 2002. 57 [34] Nobuo Imai. Flue gas CO2 capture: CO2 capture technology of KS-1. In 4th Workshop of International Test Network for CO2 capture, Kyoto, Japan, October 2002. IEA Greenhouse Gas R&D Programme. 46 [35] IPCC. Climate Change 2001: Mitigation. Cambridge University Press, 2001. Con- tribution of Working Group III to the Third Assessment Report of the Intergovern- mental Panel on Climate Change. 6 [36] Fang-Yuan Jou, Alan E. Mather, and Frederick D. Otto. The solubility of CO2 in a 30 mass percent monoethanolamine solution. The Canadian Journal of Chemical Engineering, 73:140–147, February 1995. 59, 60, 111, 112, 113, 114, 115 [37] Graeme Lamb, May 2003. Personal communication. 79, 82 [38] John Marion, Nsakala ya Nsakala, Carl Bozzuto, Gregory Liljedahl, Mark Palkes, David Vogel, J. C. Gupta, Manoj Guha, Howard Johnson, and Sean Plasynski. Engi- neering feasibility of CO2 capture on an existing US coal-?red power plant. In 26th International Conference on Coal Utilization & Fuel Systems, Clearwater, Florida, March 2001. 47, 57, 78, 83, 92, 93 [39] Carl Mariz. Carbon dioxide recovery: large scale design trends. Journal of Canadian Petroleum Technology, 37(7):42–47, July 1998. 83 [40] Carl Mariz, Larry Ward, Gar?led Ganong, and Rob Hargrave. Cost of CO2 recovery and transmission for EOR from boiler stack gas. In Pierce Riemer and Alexander Wokaun, editors, Greenhouse Gas Control Technologies: Proceedings of the 4th International Conference on Greenhouse Gas Control Technologies. Elsevier Sci- ence Ltd., April 1999. 8, 83, 92, 93 [41] Michael Lynn McGuire. STEAM67. World Wide Web, February 1985. http://www.winsim.com/steam/. 24 [42] Tomio Mimura, Takashi Nojo, Masaki Iijima, Ryuji Yoshiyama, and Hiroshi Tanaka. Recent developments in ?ue gas CO2 recovery technology. In J. Gale and Y. Kaya, editors, Greenhouse Gas Control Technologies: Proceedings of the 6th International Conference on Greenhouse Gas Control Technologies, volume 2, pages 1057–1062. Elsevier Science Ltd., October 2002. 46 153 [43] Tomio Mimura, S. Shimojo, T. Suda, M. Iijima, and S. Mitsuoka. Research and development on energy saving technology for ?ue gas carbond dioxide recovery and steam system in power plant. Energy Conversion and Management, 36(6– 9):397–400, 1995. 46, 73, 78, 98 [44] Tomio Mimura, Hidenobu Simayoshi, Taiichiro Suda, Masaki Iijima, and Sigeaki Mituoka. Development of energy saving technology for ?ue gas carbon dioxide recovery in power plant by chemical absorption method and steam system. Energy Conversion and Management, 38:S57–S62, 1997. 47, 78, 98 [45] Shinichirou Morimoto, Kotarou Taki, and Tadashi Maruyam. Current review of CO2 separation and recovery technologies. In 4th Workshop of International Test Network for CO2 capture, Kyoto, Japan, October 2002. IEA Greenhouse Gas R&D Programme. 92, 93 [46] Eriko Muramatsu and Masaki Iijima. Life cycle assessment for CO2 capture technology from exhaust gas of coal power plant. In J. Gale and Y. Kaya, edi- tors, Greenhouse Gas Control Technologies: Proceedings of the 6th International Conference on Greenhouse Gas Control Technologies, volume 1, pages 57–62. El- sevier Science Ltd., October 2002. 46 [47] Ken Olsen, Pascale Collas, Pierre Boileau, Dominique Blain, Chia Ha, Lori Hen- derson, Chang Liang, Scott McKibbon, and Laurent Moreal-` a-l'Huissier. Canada's greenhouse gas inventory: 1990–2000. Technical report, Environment Canada, June 2002. 2 [48] Alfred Ong'iro, V. Ismet Ugursal, A. M. Al Taweel, and G. Lajeunesse. Thermody- namic simulation and evaluation of a steam CHP plant using aspen plus. Applied Thermal Engineering, 16(3):263–271, 1996. 11, 26, 31 [49] Edward S. Rubin and Anand B. Rao. Uncertainties in CO2 capture and se- questration costs. In J. Gale and Y. Kaya, editors, Greenhouse Gas Control Technologies: Proceedings of the 6th International Conference on Greenhouse Gas Control Technologies, volume 2, pages 1119–1124. Elsevier Science Ltd., October 2002. 98 [50] Blair Seckington, May 2003. Personal communication. 4, 5 [51] Blair Seckington, April 2004. Personal communication. 20 [52] Ahmed Shafeen. Assessment of geological co2 sequestration in ontario. Master's thesis, University of Waterloo, 2003. 10 154 [53] M. Simmonds, P. Hurst, M. B. Wilkinson, C. Watt, and C. A. Roberts. A study of very large scale post combustion CO2 capture at a re?ning & petro- chemical complex. In J. Gale and Y. Kaya, editors, Greenhouse Gas Control Technologies: Proceedings of the 6th International Conference on Greenhouse Gas Control Technologies, volume 1, pages 39–44. Elsevier Science Ltd., October 2002. 57, 83 [54] David Singh, Eric Croiset, Peter L. Douglas, and Mark A. Douglas. Techno- economic study of CO2 capture from an existing coal-?red power plant: Mea scrubbing vs. O2/CO2 recycle combustion. Energy Conversion and Management, 44(19):3073–3091, November 2002. 8, 11, 42, 45, 49, 93 [55] David J. Singh. Simulation of CO2 capture strategies for an existing coal ?red power plant - MEA scrubbing versus O2/CO2 recycle combustion. Master's thesis, University of Waterloo, 2001. 8, 10, 11, 42, 45, 49, 54, 57, 92, 93 [56] M. Slater, E. West, and C. L. Mariz. Carbon dioxide capture from multiple ?ue gas sources. In J. Gale and Y. Kaya, editors, Greenhouse Gas Control Technologies: Proceedings of the 6th International Conference on Greenhouse Gas Control Technologies, volume 1, pages 103–107. Elsevier Science Ltd., October 2002. 57 [57] J. M. Smith, H. C. Van Ness, and M. M. Abbott. Introduction to Chemical Engineering Thermodynamics. McGraw-Hill Companies, Inc., sixth edition, 2001. 31 [58] Giorgio Soave and Josep A. Feliu. Saving energy in distillation towers by feed splitting. Applied Thermal Engineering, 22:889–896, 2002. 44 [59] R. C. Spencer, K. C. Cotton, and C. N. Cannon. A method for predicting the perfor- mance of steam turbine-generators. . . 16,500 kW and larger. Journal of Engineering for Power, 85:249–252, October 1963. 29, 30 [60] Paitoon Tontiwachwuthikul, Christine Chan, Weerapong Kritpiphat, David De- Montigny, David Skoropad, Don Gelowitz, Adisorm Aroonwilas, Frank Mourits, Malcolm Wilson, and Larry Ward. Large scale carbon dioxide production from coal-?red power stations for enhanced oil recovery: a new economic feasibility study. Journal of Canadian Petroleum Technology, 37(11):48–55, November 1998. 47, 92 [61] A. Veawab. Corrosion in CO2 capture unit for coal-?red power plant ?ue gas. In J. Gale and Y. Kaya, editors, Greenhouse Gas Control Technologies: Proceedings 155 of the 6th International Conference on Greenhouse Gas Control Technologies, vol- ume 2, pages 1595–1598. Elsevier Science Ltd., October 2002. 78 [62] James O. Wilkes. Fluid Mechanics for Chemical Engineers. Prentice-Hall PTR, Upper Saddle River, New Jersey, 1999. 105 [63] M. Wilson, P. Tontiwachwuthikul, A. Chakma, R. Idem, A. Veawab, A. Aroon- wilas, D. Gelowitz, J. Barrie, and C. Mariz. Test results from a CO2 extraction pilot plant at Boundary Dam coal-?red power station. In J. Gale and Y. Kaya, ed- itors, Greenhouse Gas Control Technologies: Proceedings of the 6th International Conference on Greenhouse Gas Control Technologies, volume 1, pages 31–36. El- sevier Science Ltd., October 2002. 83 [64] M. A. Wilson, R. M. Wrubleski, and L. Yarborough. Recovery of CO2 from power plant fule gases using amines. Energy Conversion and Management, 33(5–8):325– 331, 1992. 83 [65] Takahisa Yokoyama. Japanese R&D on CO2 capture. In J. Gale and Y. Kaya, ed- itors, Greenhouse Gas Control Technologies: Proceedings of the 6th International Conference on Greenhouse Gas Control Technologies, volume 1, pages 13–17. El- sevier Science Ltd., October 2002. 44 156