Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

NO[sub x] and N2O emissions from circulating fluidized bed combustion Sung, Lawrence Yu Jen 1995

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1995-0116.pdf [ 6.82MB ]
Metadata
JSON: 831-1.0058551.json
JSON-LD: 831-1.0058551-ld.json
RDF/XML (Pretty): 831-1.0058551-rdf.xml
RDF/JSON: 831-1.0058551-rdf.json
Turtle: 831-1.0058551-turtle.txt
N-Triples: 831-1.0058551-rdf-ntriples.txt
Original Record: 831-1.0058551-source.json
Full Text
831-1.0058551-fulltext.txt
Citation
831-1.0058551.ris

Full Text

NO x AND N 20 EMISSIONS FROM CIRCULATING FLUIDIZED BED COMBUSTION by LAWRENCE YU JEN SUNG B.Sc. Tunghai University, Taiwan, 1985 M.Sc. Sun Yet Sheng University, Taiwan, 1988 A THESIS SUBMITTED IN PARTIAL FULEFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES Department of Chemical Engineering We accept this thesis as comforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA February, 1995 ©Lawrence Yu Jen Sung, 1995 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of Ckp^lcvJi Tc*cJL The University of British Columbia Vancouver, Canada Date T-e.U , \<tt< DE-6 (2/88) Abstracts Parametric studies on NOx and N 20 emissions from circulating fluidized bed combustion were conducted by analyzing NOx and N 20 emission data from the UBC pilot plant facility. The 02, NOx and N 20 concentration profiles in the combustor were also investigated. A comprehensive NO/N20 model was developed to predict the NO and N 20 emissions from CFB under various operating conditions and the axial NO and N 20 concentration profiles. The parametric studies were carried out by combustiing four different fuels: Conoco coke, Poplar River lignite, CANMET pitch and Mt. Klappan anthracite, with the volatile content ranging from 6.5% (Mt. Klappan anthracite) to 67.7% (CANMET pitch) and the fuel nitrogen content varying from 0.9 wt % (Mt. Klappan anthracite and Poplar River lignite) to 1.9% (Conoco coke). The effects of fuel properties, temperature, excess air (0 2%) and limestone addition on the NxO and N 20 emissions were investigated with an emphasis on reducing both species simultaneously. Both axial and lateral 02, N0 X and N 20 concentration gradients existed in a CFBC riser, shedding some light on the formation and reduction mechanism of NxO and N 20 and indicating that a core-annulus structure existed in a CFB riser. Fifteen NO and N 20 formation and reduction reactions were included in the NO and N 20 model developed, with eight homogeneous and seven heterogeneous reactions. The reaction rate of each reaction was expressed in terms of global kinetics. Global kinetics for the gas phase reactions involved were derived from Hulgaard's (1991) experiments with HCN, NH3 and N 20 in a plug flow reactor. The NO/N20 model, based on a modified Senior and Brereton's model, could predict NxO and N 20 emissions as functions of operating conditions. u The NO/N20 model was applied to the Conoco coke combustion in the UBC pilot plant. With three unknown reaction rate constants in the NO and N 20 reaction scheme serving as fitting parameters, the predicted NO and N20 concentration profiles showed fair agreement with the experimental one. Sensitiviy analysis was performed on the reaction rate constants involved in the proposed reaction scheme, showing the relative importance of these reactions in the NO and N 20 reaction scheme. The effects of excess air, limestone addition and the height of secondary air port on the NO and N 20 emissions predicted by the NO/N20 model qualitatively agree with what have been observed from the UBC pilot plant. The NO/N20 model predicted that there was an optimized P/S ratio for the NO .and N 20 emission control for Conoco coke combustion and that the CFB riser with an abrupt exit would result in lower NO and N 20 emissions. iii Table of Contents Abstract ii List of Tables viii List of Figures x Acknowledgment xvi 1 Introduction 1.1 Circulating Fluidized Bed Combustion - (CFBC) 1 1.2 N O x and N 2 0 and their Environmental Impacts 3 1.2.1 Impacts of N O x • 5 1.2.2 Impacts of N 2 0 5 1.2.2.1 Ozone layer depletion 5 1.2.2.2 Greenhouse effect 6 1.3 Objectives 7 2 Literature Review 8 2.1 Effects of Operating Conditions 8 2.1.1 Temperature 8 2.1.2 Excess Air 9 2.1.3 Fuel Type 9 2.1.4 Staging 10 2.1.5 Limestone Additions 12 2.2 Nitrogen Chemistry 13 2.2.1 Pyrolysis 15 2.2.2 N O and N 2 0 Formation Reactions 15 iv 2.2.2.1 Oxidation of volatile nitrogen in gas phase 16 2.2.2.2 Catalytic oxidation of volatile nitrogen 16 2.2.2.3 Oxidation of char-N 17 2.2.1 NO and N 20 Reduction 18 2.2.3.1 Gas phase reaction 19 2.2.3.2 Gas-solid reactions 20 2.2.3.3 Catalytic reduction 20 2.2.3.4 Summary 21 2.3 Hydrodynamics of Circulating Fluidized Bed 22 3 Chemical Kinetics of Fuel-N 25 3.1 Homogeneous Reaction 26 3.1.1 Thermal Decomposition of N20' 27 3.1.2. NH 3 Oxidation 32 3.1.3.0xidationofHCN 40 3.2 Heterogeneous Reactions 46 3.2.1 NO Formation 46 3.2.2 N 20 Formation 47 3.2.3 Decomposition of NO and N 20 by Char 48 3.2.4 Decomposition of NO and N20 on the Surface of Calcined Lime 49 3.3 Summary 51 4 NO and N 20 Model Development 55 4.1 Hydrodynamic Model 55 4.2 Devolatilization Model 63 v 4.3 02 Model 67 4.3.1 Single Phase Primary Air Zone 69 4.3.2 Core-Annulus Primary Air Zone 73 4.4 NO/N20 Model 76 4.5 Summary 81 5. Multifuel Combustion Runs in the UBC C F B C Pilot Plant 82 5.1 UBC CFBC Pilot Plant Facility 82 5.2 NOx and N20 Emissions Results: 88 5.2.1 Fuel Types 92 5.2.2 Temperature Effect 94 5.2 3 Excess Air Effect 94 5.2.4 Effect of Limestone Addition 96 5.3 Gas Concentration Profiles 96 5.3.1 02 Concentration Profiles 98 5.3.2 NOx Concentration Profiles 99 5.3.3 N20 Concentration Profiles 105 5.4 Summary 107 6 NO and N2O Modelling Results 110 6.1 Reaction Rate Constants k n, k12 and k13 and Mass Transfer Coefficient 110 6.2 Sensitivity Analysis 115 6.3 Model Predictions 117 6.3.1 Effects of Fuel-N Distribution in Char and Volatile 117 6.3.2 Limestone Addition 119 vi 6.3.3 Excess Air Effects 122 6.3.4 Air Staging 127 6.3.5 Riser Exit Geometry 130 6.4 Summary 133 7 Overall Conclusions and Recommendations 7-1 Conclusions 138 7-2 Recommendations 139 Nomenclature 141 Bibliography 144 Appendices A Arrhenius Type Reaction Rate Constant Expressions 154 B Derivations of [NH3], [HCN], [NO] and [N20] Profiles 158 C Computer Programs for 0 2 Model and NO/N 2 0 Model 166 D 0 2 , N O x and N z O Profiles in the UBC C F B C Pilot Plant 188 vii List of Tables 1.1 Air Emission Limits for Coal-Fired Combustors (IEA, 1989) 4 3.1 Product Distribution of N 20 Decomposition 30 3.2 Predicted Global Kinetic Reaction Rate Constants &;and k2 30 3.3 Product Distribution of NH 3 Oxidation 35 3.4 Predicted Global Kinetic Reaction Rate Constants k3, k4 and k5 35 3.5 Product Distribution of HCN Oxidation 43 3.6 Predicted Global Kinetic Reaction Rate Constants k6 ,k7 and k% 43 3.7 Decomposition Rate Constants k12 and k13 for Chars (deSoete, 1990) 49 3.8 First Order Decomposition Rate constants for N 20 over Calcined Limestone, Johnsson (1991) 3.9 Summary of Reaction Equations and Rate Laws 3.10 Summary of Reaction Rate Constant Expressions 4.1 Model Equations for the Single Phase 02 Model 4.2 Model Equations for the Core-Annulus 02 Model 4.3 NFf3, HCN, NO and N 20 Mass Balance Equations used in the NO/N20 Model 77 4.4 Boundary Conditions for the NO/N20 Model 79 4.5 Summary of the Input Data for the NO/N20 Model 80 5.1 Summary of NOx and N 20 Emission Results 90 5.2 Proximate and Ultimate Analyses of Solid Fuels 91 5.3 Comparison of Average Bed Temperature and Secondary Cyclone Temperature. (Brereton et al., 1992) 108 5.4 Features of N 20 Concentration Profiles in CFBC Riser of Different Scales 108 viii 50 52 53 71 74 6.1 Predicted Reaction Rate Constants ku, k12 and k13 for Conoco Coke at 1155°K 112 6.2 Predicted Volume Fraction of Char in Solids (V c h a r) 127 B.l Summary of A; for i =1 to 26 164 D. 1 Index of 02, NOx and N 20 Concentration Profiles 188 ix List of Figures 1.1 Variations in Circulating Fluidized Bed System Configuration (Yang, 1992) 2 2.1 Fuel Nitrogen Conversion to NOx vs "VN" Product (Zhao, 1992) 11 2.2 Dynamic Change in Emissions of S02, NOx, N 20 for Medium Volatile Bituminous Coal Combustions with Impulsive and Continuous Feed of Limestone (Shimizu et al., 1993) 14 2.3 Axial Suspension Density Profiles in the UBC Pilot Plant (Grace et al., 1989) 23 3.1 Product Concentrations for N 20 Decomposition in a Homogeneous Plug Flow Reactor (Hulgaard etal., 1991) 28 3.2 Product Concentrations for NH 3 Oxidation in a Homogeneous Plug Flow Reactor with NO present at Reactor Inlet (Hulgaard et al., 1991) 33 3.3 Predicted Product Concentrations for NH 3 Oxidation in a Homogeneous Plug Flow Reactor (Hulgaard et al., 1991) 37 3.4 Predicted Product Concentrations for NH 3 Oxidation in a Homogeneous Plug Flow Reactor without NO Present at the Inlet(Hulgaard et al., 1991) 39 3.5 Product Concentrations for HCN Oxidation in a Homogeneous Plug Flow Reactor (Hulgaard et al., 1991) 41 3.6 Product Concentrations for HCN Oxidation in a Homogeneous Plug Flow Reactor with NO Present at the Inlet (Hulgaard et al., 1991) 45 3.7 Formation and Destruction Reaction Pathways of NO and N 20 in CFBC 54 4.1 Schematic Representation of Aa, Ac and Pa in the Core-annulus Structure (Senior and Brereton, 1992) 57 4.2 Predicted Axial Suspension Density Profile for the UBC Pilot Plant Unit: Run 17-1, Conoco Coke (Brereton et al., 1992) 60 x 4.3 Predicted Core Voidage Profiles for the UBC Pilot Plant Unit: Run 17-1, Conoco Coke (Brereton et al., 1992) 61 4.4 Predicted Core and Annulus Area Profile for the UBC Pilot Plant Unit: Run 17-1, Conoco Coke (Brereton et al., 1992) 62 4.5 Schematic Representation of the Single Phase 0 2 Model 69 4.6 0 2 Profile Predicted by the Single Phase Oxygen Model: Run 17-1, Conoco Coke (Brereton et al., 1992) 72 4.7 Schematic Representation of the Core-Annulus 0 2 Model 75 4.8 0 2 Profile Predicted by a Core-Annulus Oxygen Model: Run 17-1, Conoco Coke (Brereton et al., 1992) 76 4.9 Solution Algorithm for the NO/N20 Model 80 4.10 Schematic Representation of the Communication between NO and N 20 Model and Sub-Models 82 5.1 Simplified Schematic Diagram of the UBC CFBC Facility 84 5.2 View of the Refractory Lined Reactor Shaft. 85 5.3 Radial Gas Sampling Positions (top view) (Zhao, 1992) 87 5.4 Reactor Gas Sampling Train: (Zhao, 1992) 88 5.5 N 20 Concentrations Measured by FTIR for Run 22-3, CANMET Pitch (Brereton et al., 1992) 90 5.6 Fuel-N Conversions to N0 X and N 20 vs "VN" and Voaltile Content (Brereton et al., 1992) 94 5.7 Effect of Temperature on the Fuel-N Conversion to NOx and N 20 (Brereton et al., 1992) ' 96 xi 5.8 Effect of Limestone Addition on Fule-N Conversion to N 0 X and N 2 0 (Brereton et al., 1992) 5.9 0 2 Concentration Profiles for Conoco Coke Combustion: Run 17-2 (Brereton et al., 1992) 5.10 N O x Concentration Profiles for Conoco Coke Combustion: Run 22-3 (Brereton et al., 1992) 5.11 N O x Concentration Profiles for Conoco Coke Combustion: Run 17-1 (Brereton et al., 1992) 5.12 N O x Concentration Profiles for Conoco Coke Combustion: Run 17-3 (Brereton et al., 1992) 5.13 N 2 0 Concentration Profiles for Conoco Coke Combustion: Run 17-1 (Brereton et al., 1992) 6.1 Predicted NO and N 2 0 Concentrations Profiles for Conoco Coke Combustion: Run 17-1 113 6.2 Evolution of the Concentrations of CO, NO and N 2 0 for Char Combustion in a Fixed Bed Reactor (Amand et al, 1992) 114 6.3 Sensitivity Analysis of the Reaction Rate Constants 116 6.4 Predicted Effect of Fuel-Nitrogen Distribution on NO and N 2 0 Emissions 118 6.5 N 2 0 Formation for Different Reaction Conditions (Hulgaard, 1992) 120 6.6 Predicted Effect of Limestone Addition on NO Emissions 121 6.7 Sensitivity of the Predicted NO Emission to Reaction Rate Constant, k 1 4 123 xii 98 101 102 103 104 106 6.8 Predicted Effect of Limestone Addition on the N 20 Emission 124 6.9 Sensitivity of the Predicted N 20 Emission to Reaction Rate Constant, k 1 4 125 6.10 Predicted Effect of Excess Air on NO and N 20 Emissions 126 6.11 Predicted Effect of Excess Air on 0 2 profiles 128 6.12 Predicted Effect of P/S Ratio and Height of Secondary Air Injection on NO and N 20 Emissions 129 6.13 Predicted Effect of Height of Secondary Air Entry on 0 2 Profiles 131 6.14 Predicted Suspension Density Profiles for CFB Risers with Different Exit Geometries 132 6.15 Predicted 0 2 Profiles for Different Riser Exit Geometries 134 6.16 Predicted NO Profiles for Different Riser Exit Geometries 135 6.17 Predicted N 20 Profiles for Different Riser Exit Geometries 136 A l ln(ki) vs 1 0 ( y Q 0 fori-1 to 2 142 A.2 ln(k.) vs 1 0 ( y Q Q fori = 3to5 143 A3 ln(ki) vs [ Q Q^ Q Q Q fori = 6to8 144 D. 1 0 2 Concentration Profiles for Conoco Coke Combustion: Run 17-1 (Brereton et al., 1992) 189 D.2 0 2 Concentration Profiles for Conoco Coke Combustion. Run 17-2 (Brereton et al., 1992) 190 D.3 0 2 Concentration Profiles for Mt Klappan Anthracite Combustion: Run 14-1. (Brereton et al., 1992) 191 D.4 0 2 Concentration Profiles for Mt Klappan Anthracite Combustion: Run 19-1 (Brereton et al., 1992) 192 xiii D.5 02 Concentration Profiles for Polar River Lignite Combustion: Run 26-1 (Brereton et al., 1992) 193 D.6 0 2 Concentration Profiles for Polar River Lignite Combustion: Run 26-1 (Brereton et al., 1992) 194 D.7 NO x Concentration Profiles for Conoco Coke Combustion: Run 17-2 (Brereton etal., 1992) 195 D.8 NO x Concentration Profiles for Mt Klappan Anthracite Combustion:Run 14-1 (Brereton etal, 1992) 196 D.9 NOx Concentration Profiles for Mt Klappan Anthracite Combustion: Run 19-1 (Brereton etal., 1992) 197 D. 10 NOx Concentration Profiles for CANMET Pitch Combustion: Run 26-1 (Brereton etal., 1992) 198 D. 11 NOx Concentration Profiles for Polar River Lignite Combustion: Run 26-1 (Brereton et al., 1992) 199 D. 12 NOx Concentration Profiles for Polar River Lignite Combustion: Run 26-2 (Brereton etal., 1992) 200 D.13 N 20 Concentration Profiles for Conoco Coke Combustion: Run 17-2 (Brereton etal., 1992) 201 D.14 N 20 Concentration Profiles for Conoco Coke Combustion: Run 17-3 (Brereton etal., 1992) 202 D.15 N 20 Concentration Profiles for Mt Klappan Anthracite Combustion: Run 14-1 (Brereton et al., 1992) 203 D. 16 N 20 Concentration Profiles for Mt Klappan Anthracite Combustion: Run 19-1 (Brereton etal., 1992) " 204 D. 17 N 20 Concentration Profiles for Polar River Lignite Combustion: Run 26-1 (Brereton etal., 1992) 205 xiv D. 18 N 2 0 Concentration Profiles for Polar River Lignite Combustion: Run 26-1 (Brereton etal., 1992) 206 xv Acknowledgement I would like to greatly appreciate my supervisors, Dr. J. R. Grace, Dr. C. J. Lim and Dr. C. M. H. Brereton, whose continuous guidance and support direct me through the thesis. I am also grateful to Dr. J. Zhu and J. Zhang, giving me a lot of help at the beginning of this work. I sincerely thank my colleagues in the UBC CFBC group who have spent many sleepless nights in the highhead room working with heat and dust. A lot of my work was based on their experimental result, Finally, I am deeply indebed to my family, whose love and encouragement have always been the source of inspiration during the course of this thesis. xvi Chapter 1 Introduction CFB is a technology which provides favourable gas-solid contacting for continuous gas-solid reactions. Early applications were primarily in the fluidised catalytic cracking process in the oil industry. Not untill the 1970s when there was a need for new coal-fired power plants to replace many out-dated pulverized coal combustors subject to more and more demanding enviromental regulations, was attention devoted to applying CFB as a new combustion technology for utility-plant power generation and for industrial boilers. 1.1 Circulating Fluidized Bed Combustion - (CFBC) A CFBC facility has two major components - a riser and a solids return system. The riser is the reactor where most of the fuel is combusted. The solids return system is usually composed of a cyclone, a standpipe and a solids flow control device. These are important in considering the pressure and solids inventory balance in the whole CFBC loop (Bi and Zhu., 1992). There are many arrangements of circulating bed loops, with most variations being in the solids return system. Figure 1.1a shows a standpipe with a seal pot, Figure 1.1b a standpipe with a slide valve, Figure 1.1c a slow bubbling bed in combination with an L-valve, and Figure 1. Id a fluidized bed as an external heat exchanger. Compared with other coal combustion technologies, CFBC has the following advantages: 1. Fuel flexibility: Many different fuels have been demonstrated to burn very well in CFBC. The technology is especially noted for its ability to handle low grade fuels. 1 Chapter 1 Introduction 2 1.1a: CFB with Seal Pot 1.1b: CFB with Mechanical Valve 1.1c: CFB with Slow Bubbling Bed 1.1 d: CFB with External Heat and L-Valve Exchanger Figure 1.1: Variations in Circulating Fluidized Bed System Configuration (Yang, 1992) Chapter 1 Introduction 3 2. High combustion efficiency: The combustion efficiencies achieved in CFBC are usually more than 98%. 3. Ease of S0 2 removal and low NOx emissions: CFBC is capable of in-situ S0 2 removal by the addition of low-cost sorbents like limestone or dolomite. To optimize S0 2 removal by limestone, the CFBC unit is normally operated at temperatures between 800 and 900°C At these temperatures, negligible thermal NOx is formed from atmospheric nitrogen, and this leads to lower NOx emissions than from conventional pulverized coal combustors. In comparison with bubbling fluidized bed combustors operation with similar fuels under similar operating conditions, NOx emissions from CFBC have been found to be lower by about half (Leckner and Amand., 1987; Reh et al., 1980). 4. Good bed-to-immersed surface heat transfer: The heat transfer coefficient is typically 200-250 W/m2K (Grace et al, 1987), higher than for conventional pulverized coal combustors. This means less boiler volume and heat transfer surface are needed in boiler design for the same steam generation capacity. 5. Higher turndown ratio: Solid renewal rate and superficial velocity can be controlled independently, allowing a wider range of operation than in bubbling beds. Owing to the merits discussed above, CFB is gaining in popularity; according to the survey of Yang (1992), more than 223 CFB combustors were in operation or under construction worldwide at that time. 1.2 NOx and N 20 and their Environmental Impacts NOx has long been an air pollution concern in fossil fuel combustion. Utility generation accounts for about one-half of the man-made NOx. In 1980, 5.9 million tonnes of NOx were generated by utility boilers in North America (Whaley et al. 1989). Chapter 1 Introduction 4 Table 1.1 summarizes the emission limits of N O x for coal-fired combustors in various countries (PEA, 1989). Regulations are expected to become more severe in the future. It has been forcast (MacRae, 1991) that Canada's NO x . emission limitation will be lower than 100 mg/MJ by the year 2000. So far, no legislation has been enacted against N 2 0 . Before 1988, N 2 0 emissions measured from conventional fossil fuel combustion facilities were estimated to be of the following magnitude (Weiss and Craig, 1976; Hao et al., 1987): natural gas, 10 ppm; oil, 50ppm and coal 150 ppm. Most of the N 2 0 measured was formed in stainless steel containers where S 0 2 , 0 2 , NO and moisture coexisted while awaiting analysis (Muzio and Kramlich, 1988). With improved sampling technology, 0.2 ppm, 1 ppm and 2-5 ppm of N 2 0 have been reported for natural gas, oil and pulverized coal respectively (Linak et al., 1989). However, much higher values have been found for fluidized bed combustion, up to 250 ppm (Hulgaard, 1991). In the U B C CFBC pilot plant, N 2 0 concentrations have ranged from 20 to 230 ppm (Brereton et al, 1992). Table 1.1: Air Emission Limits for Coal-Fired Combustors (IEA, 1989) S0 2 Limit NO x Limit mg/Nm3 g/GJ lb/109 Btu mg/Nm3 g/GJ lb/109 Btu Australia 2000 840 1950 500 210 490 Austria 200 85 200 200 85 200 Belgium 400 150 350 650 240 560 Canada 615 258 600 615 258 600 Denmark na na na 610 260 600 Germany 400 170 400 200 85 200 Italy 400 150-160 350-370 650 240-275 630-640 Japan 223 95 220 411 175 410 Luxembourg 1700 640 1500 450 170 400 Netherland 400 170 400 400 170 400 Spain 2400 1010 2350 none none none Sweden 80-270 30-100 70-230 135-335 50-200 115-465 Switzerland 400 150 350 200 75 170 Turkey 1000 420 980 800-1800 335-755 780-1750 United State 1240 520 1210 475-620 200-260 465-605 Chapter 1 Introduction 5 1.2.1 Impacts ofNO x N 0 X is an abbreviation for NO and N 0 2 , which are produced in fluidized bed combustors with a ratio of about 20 : 1 (Sarofim and Flagan, 1976). NO is a colourless gas; its ambient concentration is usually lower than 0.5 ppm, too low to cause significant harm to human health. However, NO is a precursor to N 0 2 , which reacts with water to form nitric acid. It was estimated by Shaw (1984) that NOx accounts for 35% of annual acid rain precipitation. In metropolitan areas where there are high concentrations of NOx, it reacts with unburnt hydrocarbons to form smog in the presence of sunlight. This smog interferes with visibility and is detrimental to human health (Wark and Warner, 1981) 1.2.2 Impacts of N 2 0 NzO is the second most abundant nitrogen containing species in the atmosphere. It is a relatively inert gas in the atmosphere and does not participate in the normal nitrogen-cycle in the troposphere. It is not an acidifying agent, nor a contributor to the formation of nitrates, these being the most commonly reported detrimental properties of NOx. The concentration of N 2 0 in the atmosphere is not considered to cause health problems, nor is it damaging to vegetation, so N2O was not considered an air pollutant untill the late 1970's, when concern was aroused about its impact on the depletion of stratospheric ozone and its contribution to the greenhouse effect (Hayhurst et al., 1992). 1.2.2.1 Ozone layer depletion The most important catalytic cycle for ozone destruction involves NO and N 0 2 through Chapter 1 Introduction 6 the following cycle: NO + 0 3 -» N0 2 +0 2 N0 2 + O -> NO + 0 2 This sequence is responsible for about 70% of the destruction of 0 3 in the lower stratosphere. Because N 20 is the only significant source of NO in the stratosphere through the above sequence (Proffit, 1989; EPA, 1989), an increased concentration of N 20 in the stratosphere would increase the destruction rate of ozone through the following reactions. 0 3 + hv (XOlOnm) -> O^D) + 0 2 0(JD) + N 20 -> 2N0 1.2.2.2 Greenhouse effect In the atmosphere, C02, HjO, CH4, chlorofluorohydrocarbons and N 20 are known to absorb infrared light emitted from the surface of the earth, thereby increasing the earth's average temperature and causing climate change. This effect is known as the greenhouse effect. Although the concentration ofN 20 is 3 order of magnitude lower than that of C0 2 in the atmosphere, the greenhouse effect caused by N 20 is remarkable, about 1/6 the effect of C02. At some wavelengths, N 20 absorbs infrared light much more than C02. For instance, at X = 7.8 um the radiation absorbed by N 20 per molecule is 200 times stronger than by C02(Elkins, 1989). Chapter 1 Introduction 7 1.3 Objectives N0 X emissions from bubbling fluidized bed combustion has been intensely studied (Horio et al., 1977; Gibbs et al., 1980; Beer et al.; Lee et al., 1983). Much less information is available in the literature with respect to NOx in CFB combustion and even less for N20. No N 20 model has been published for either CFB or bubbling fluidized bed combustion. Zhao (1992) developed a CFBC NO model based on a scheme of homogeneous and heterogeneous reactions involving NOx. A simple core-annulus structure was assumed to represent the CFB riser, with the voidage in the core and the downflowing wall layer thickness taken as constants. In fact, the voidage in the core and the wall layer thickness vary with operating conditions. Without this linkage between operating conditions and the voidage in the core and wall layer thickness, it is difficult to predict how NOx varies with operating conditions. The objectives of this thesis are to provide insight into the NOx and N 20 profiles measured during the multi-fuel combustion runs in the UBC CFBC pilot plant and to develop a new N0 X and N 20 model based on a more realistic hydrodynamics structure. A literature review is given in Chapter 2. Chapter 3 outlines the key reactions involved in the formation of NO and N 20 and derives global kinetics expressions for the reactions of the key nitrogen-containing species, i.e., HCN and NH3. Chapter 4 develops a NOx and N 20 model, while Chapter 5 interprets N0 X and N 20 profiles from the UBC pilot plant runs. Chapter 6 presents the modelling results and compares these predictions with the experimental profiles measured from the UBC pilot plant runs. Chapter 7 provides conclusions and makes recommendations for NOx and N 20 abatement and future work. Chapter 2 Literature Review In this chapter, the effects of operating conditions on NOx and N 20 emissions in CFBC are reviewed first, followed by a summary of the fuel-N chemistry within the fluidized bed combustor, including the split of fuel-N during pyrolysis and the reaction paths leading to the formation and destruction of NO and N20. 2.1 Effects of Operating Conditions The effects of temperature, excess air, fuel type, staging, and limestone addition on N0 X and N 20 emissions have been widely reported for bubbling fluidized bed combustion. The available information in the literature for CFBC is generally similar to that for bubbling fluidized beds. Temperature, excess air and fuel type are the primary variables in influencing NOx and N 20 emissions which are amenable to generalization. Staging and limestone addition may have varying effects depending on system design and fuel type. 2.1.1 Temperature There is consensus in the literature for both bubbling and circulating fluidized beds that NOx increases with temperature while N 20 decreases (Brereton et al., 1991; Leckner and Amand, 1987, Moritomi et al., 1991; Wojtowicz et al., 1991, Shimizu et al., 1991). A similar trend was found in combustion runs in the UBC CFBC pilot plant unit (Brereton et al., 1992). Higher temperature leads to higher volatile-nitrogen release rates and faster oxidation in the gas phase (Pohl and Sarofim, 1976; Lee and Hiltunen, 1989) which, in turn, favours the formation of NOx. 8 Chapter 2 Literature Review 9 Tests in a bench scale reactor (Wojtowicz et al., 1991), a pilot plant (Brereton, et al., 1991), and a commercial CFBC unit (Amand and Leckner, 1991a) all showed that N 20 emissions decrease with increasing operating temperature. N 20 decomposes strongly at high temperature (Kilpinen and Hupa, 1991), with the extent of decomposition increased by the presence of some bed materials (Miettinen, 1991) 2.1.2 Excess Air Excess air is commonly expressed in terms of % 02 in the flue gas, with 3-5% 02 often employed in industry in order to achieve nearly complete combustion. All published data indicate that NOx and N 20 both increase with excess air in both bubbling and circulating fluidized beds (Moritomi et al, 1991; Wojtowicz et al, 1991, Shimizu et al, 1991). In the UBC CFBC pilot plant runs, NOx and N20 increased almost linearly with excess air in the range 2% to 7% 02 (Brereton et al, 1991). The increase of NOx and N 20 with increasing excess air arises from higher oxygen concentration and lower carbon loading in the reactor, thereby enhancing the formation of NOx and N 20 while reduction reactions are retarded. A trade-off exists between NOx and N20 emissions on the one hand, which are increased by increasing excess air, and combustion efficiency and ease of operation on the other, which also increase with increasing excess air. 2.1.3 Fuel Type In addtion to temperature and excesss air, NOx and N 20 emissions are closely related to the volatile matter and nitrogen content of the fuel Fig 2.1 shows that NOx emissions are well correlated to a VN number (Grace, et al, 1989), when the Chapter 2 Literature Review 10 25 Figure 2.1: Fuel Nitrogen Conversion to NOx vs "VN" Product (Zhao, 1992) ** To ensure that the effects are primarily caused by fuel type, the data are from experiments within a restricted temperature range, 876-911° K, and a limited excess air range, 3-4.3% 0 2 Chapter 2 Literature Review 11 VN number is defined as the product of the percentage volatile matter in the fuel from proximate analysis and the percentage of fuel nitrogen from the ultimate analysis. The conversion of fuel-N to NOx increases with "VN for the fuels tested in the UBC CFBC pilot plant. N 20 emissions correlate with coal rank (Amand and Leckner, 1991a) but do not appear to correlate with VN number. An increase of N 20 with coal rank was also reported by Wojwotcz et al. (1991) who found that low rank fuels with high volatile content (lignite and a sub-bituminous coal with 53-54% volatile matter on a dry ash-free basis) gave a lower fuel-N to N 20 conversion than two bituminous coal (32 and 40 % volatile matter on dry and ash-free basis). In qualitative agreement with Wojwotcz et al. (1992), Amand and Andersson (1989) found that in a 8 MWt CFBC unit, N 20 emissions are 10-30 mg / MJ (17-51ppm) from brown coal and 70- 100 mg / MJ (119-170 ppm) from bituminous coal or petroleum coke. In summary, NOx correlates well to-fuel properties by VN number, while N 20 emissions are closely correlated to coal rank . 2.1.4 Staging Staging is used in many types of combustion equipment to reduce NO x emissions. Staging is achieved by splitting air into a primary air stream introduced from the bottom of a combustor through the distributor, and a secondary air stream injected at a higher level. With staging, an environment of lower local oxygen concentration is created at the bottom, reducing the conversion of fiiel-N to NOx and N20. In a bubbling bed, a 60% reduction of NOx has been achieved compared to single stage combustion (Kunii et al., 1980). Chapter 2 Literature Review 12 Staged combustion is particularly suited to CFBC in comparison to bubbling beds because fuel and solids are distributed throughout a tall riser. In addition, the temperature rise in the freeboard region of bubbling bed combustor which suppresses NOx reduction does not happen in CFBC. Staging in CFBC units is controlled by the primary-to-secondary air ratio. The emissions then depend on the height of secondary air inlets and the method of secondary air injection. The primary-to-secondary air ratio affects the carbon/oxygen ratio in the primary and secondary air zones. The height of the secondary air inlet and the method of secondary air injection may also affect the solids distribution within the riser, especially around the secondary air port (Senior, 1992). The staging effect is not significant for either NOx or N 20 when the secondary air entry is close to the primary air inlet. In a 12 MW CFBC pilot plant of a height 13.5 m and square cross-section, when secondary air ports were raised from a lower level (2.2 m from the bottom) to a medium high level (5.5 m above the bottom), both the NOx and N 20 were found to decrease (Amand and Leckner., 1992). The same trend was observed in the UBC pilot plant with respect to NOx emissions during combustion runs of Highvale coal and anthracite (Brereton et al., 1992). 2.1.5 Limestone Additions It is generally accepted that limestone causes an increase in NOx emission in fluidized bed combustion. This has been true for most runs in the UBC CFBC pilot plant. However, the opposite trend was observed for Conoco delayed coke. As discussed below, limestone addition may catalyze both the formation of NOx from volatile nitrogen and the reduction of NOx by CO, so the net effect of limestone addition on NOx depends on the fuel nitrogen content and the distribution of nitrogen between char and volatiles. Chapter 2 Literature Review 13 In most reports (Hiltunen et al., 1991; Mjornell et al., 1991; Wojtowicz et al., 1991., Moritomi et al., 1991), limestone addition has a positive effect on N 2 0 reduction. However, some publications (Braun et al, 1991; Brown and Muzio, 1991; Cabrita, 1991; Gulyurtlu et al., 1991) have indicated that limestone addition has no effect on N 2 0 emissions. In a 16 MWt BFBC at 1120 °K (Amand and Anderson, 1989), N 2 0 emissions decreased more than 50% when the sand was replaced by sulphated limestone. Mjornell et al. (1989) found thatN 20 emission decreased with the Ca/S ratio, whereas Gulyurtlu et al. (1990) found that only a small change of N2O was observed when the Ca/S ratio in a bench scale bubbling bed was raised from 0 to 4.5 at 1073 and 1173 °K. The discrepancy regarding the effect of limestone addition has been attributed by Johnsson (1991) to the types and different degrees of sulphation of limestones. Figure 2.2 shows the transient effect of adding limestone on N O x and N 2 0 emissions in a bench scale CFBC (Shimizu et al., 1993), of which the inner diameter and height are 5.3cm and 4.3 m, respectively. The secondary air is introduced 1.3 m above the distributor. Here the N O x emission increased while the N 2 0 level decreased with increasing S 0 2 removal. In a 12 MWt CFBC (Amand et al., 1990), a 25% decrease of N 2 0 was seen, accompanied by a simultaneous increase of N O x after the addition of limestone. 2.2 Nitrogen Chemistry Fuel-N is the sole source of N O x and N 2 0 in fluidized bed combustion (Furusawa et al., 1978; Pereira et al., 1974). The transformation of fuel-N is divided into three stages in series: pyrolysis, N O / N 2 0 formation, and NO/N 2 0 destruction. The following sections review the fuel nitrogen chemistry related to the production and destruction of NO and N 2 0 in fluidized bed combustion. N 0 2 is neglected and N O x is denoted by NO hereafter. Chapter 2 Literature Review 14 (a) Impulsive addition of 150 g limestone cL 200 O c o o CM O 100 r -o z o E 200 D_ CL O C ' o o CM 8 100 d CM o z 0 • 1 1 1 1 1 • •• - • ;• °o 0.. CD-- O — A ' ' "• A A — i I . I . 1 0 2 4 6 time [hr] (b) Continuous feed of limestone. < 1 1 1 1 1 — — . • • — «• • - . • • • • -A ' ' : A A A - -Q A * ' " A ,2^'::co--o — 1 1 . 1 1 0 2 4 6 time [hr] Figure 2.2: Dynamic Change in Emissions of S02, NOx, N 20 for Medium Volatile Bituminous Coal Combustion with Impulsive and Continuous Feed of Limestone, (A) S02, (o) NOx, (•) N20. Reactor diameter: 5.3 cm; height: 4.3m. Secondary air is introduced 1.3 m above the distributor. Total air ratio = 1.25 ± 0.03 (Shimizu et al., 1993). Chapter 2 Literature Review 15 2.2.1 Pyrolysis Pyrolysis is a process whereby volatile matter in the solid fuel leaves the char matrix and undergoes decomposition reactions when heat is supplied in an inert atmosphere. During pyrolysis, fuel nitrogen divides into volatile-N and char-N. Most volatile-N exists in the form of hetrocyclic organic compounds, eg., pyridine, quinone, etc., a large portion of which are converted to H C N and N H 3 (Axworthy et al., 1978). These are regarded as the most important precursors leading to the formation of N O and N 2 0 (Hulgaard et al., 1991). Up to 340 ppm N H 3 and 40 ppm H C N were measured by Amand et al. (1990) at 1123 °K in a 12 MWt CFBC. The split of fuel nitrogen between char and volatile and the selectivity of H C N and N H 3 during pyrolysis depend on the fuel, temperature and solid mixing in the reactor. Collings (1993) measured the split of fuel-N for seven different fuels. The fraction of volatile nitrogen in fuel nitrogen ranges from 25% to 88%. Houser (1980) measured the kinetics of the formation of H C N during pyrolysis of pyridine in a quatz plug flow reactor. The selectivity of H C N at 900°C was 60% and increased with temperature, with 100% conversion to H C N occurring at temperatures, beyond 1000°C. 2.2.2 N O and N z O Formation Reactions NO and N 2 0 are formed from fuel-N mainly via three routes: (i) oxidation of volatile nitrogen in the gas phase; (ii) oxidation of volatile nitrogen over the surface of bed materials such as char, ash, and limestones; (iii) oxidation of char-N. As discussed previously, the major volatile nitrogen containing products are H C N and N H 3 , and volatile nitrogen is therefore represented by H C N and N H 3 in the following discussion. Chapter 2 Literature Review 16 2.2.2.1 Oxidation of volatile nitrogen in gas phase H C N and N H 3 are usually adopted as model compounds in the study of volatile nitrogen chemistry, e.g. see Hulgaard et al., (1991). The oxidations of H C N and N H 3 have been studied by state-of-the-art chemical kinetics technologies, free radical intermediates such as CN, NH, N H , NCO, HNO and the reaction rates to form N O between these species and some other free radicals H, O, O H have been measured. (Hayne et al., 1977; Perry, 1985). A scheme of free radical reactions in terms of these species was proposed by Miller and Bowman (1989). NCO has been identified as the most important precursor of N O formation. By doping a natural gas flame with 900 ppm of N H 3 (Kramlich et al., 1989), only N O was detected in the flue gas in the temperature range 1000 to 1600 °K, while N2O was detected at 1100 °K in an H C N flame, reaching a maximum of about 250 ppm 1250 0 K. Hulgaard (1991) investigated a series of H C N and N H 3 oxidation reactions in a quartz plug flow reactor. H C N was found to be a more important precursor than N H 3 to the formation of N 20; with a residence time around 56 ms, significant amounts of N 2 0 were produced from H C N oxidation above 1300°K in the presence of NO and HjO; CO addition caused the N 2 0 peak to shift to a lower temperature, between 1060 and 1250°K. 2.2.2.2 Catalytic oxidation of volatile nitrogen Catalytic oxidation of N H 3 over calcined and sulphated limestone was studied by Iisa et al. (1991) in a fixed bed reactor. NO was the major product at all conditions. Formation of N 2 0 was only detected on a calcined limestone surface with NO introduced in the inlet and the oxygen concentration more than 1 vol%. Chapter 2 Literature Review 17 deSoete and Nastoll (1991) studied the same reaction in a small fluidized bed reactor using calcined limestone as catalyst. With a lower inlet concentration of 0 2 , 3500 ppm. The conversion to N 2 0 was much lower than observed by Iisa et al. (1991). However, when the oxidation of H C N was carried out under the same conditions, 10% conversion to N 2 0 was observed. According to Furusawa (1985), limestone catalyzes the oxidation of NH3 : NH3+-02 - » NO+-H20 3 4 2 2 2 The catalytic oxidation of H C N and N H 3 over calcined limestone was examined by Shimizu (1992) in a fixed bed at 1123 °K; NO was the major product in this experiment. The rates of these two reactions were higher than the S0 2 removal rate. N 2 0 was detected only in the experiment with HCN. Unlike the homogeneous gas phase reaction of HCN, the selectivity to N 2 0 was greatly reduced from 50% to 4% by the presence of limestone. A similar conclusion was reached by Jensen et al (1993) who monitored the concentrations of NO and N 2 0 on-line while carrying out H C N oxidation in a packed bed of calcined limestone and then calculated the conversion to NO and N 2 0 by integration. When the limestone was replaced by char, the reaction still exhibited high NO (46%) and low N 2 0 selectivity. 2.2.2.3 Oxidation of char-N deSoete (1990) investigated the formation of NO and N 2 0 from char-N using different chars in a quartz fixed bed reactor. The conversion to NO and N 2 0 were accounted for Chapter 2 Literature Review 18 30 to 70% of the char nitrogen content. NO was higher than N 20 for all experimental conditions investigated. Wajtowitcz et al. ( 1991) compared NO and N 20 emissions from the combustion of a wide variety of coals and concluded that the oxidation of volatile and char nitrogen both play a role in both emissions. Neither contribution could be disregarded. NO is formed from char when 02 is in direct contact with char-N (deSoete, 1990; Tullin et al., 1993). Less agreement is achieved for the heterogeneous formation of N20. deSoete (1990) suggested that N20 is formed by the direct oxidation of char-N. Tullin et al. (1993) argued that, since the nitrogen-to-carbon ratio in char is around 1%, the probability of finding two char-N atoms in adjacent positions is small, N 20 formed from char-N more likely takes place via the reaction between NO and char-N. Tullin et al. (1993) illustrated this point by showing that the conversion of coal-N to N 20 doubled in a batch fluidized bed combustor as extra NO was introduced. This could, in part, account for the observations of Amand and Leckner (1991b) in a large circulating fluidized bed, who reported that when NO was injected at the bottom, the N 20 emission increased, while when char recycle from the cyclone was stopped, the concentration of NO increased while that of N 20 decreased. Others found that without 02, negligible N 20 was formed when NO reacted with char (Johnsson and Dam Johnsson, 1991). 2.2.3 NO and NzO Reduction Several decomposition paths of NO and N20, including gas phase reactions, gas-solid reactions and catalytic reactions have been addressed at FBC temperatures. The final emissions of NO and N 20 result from competition between the reduction reactions and the formation reactions discussed above. Chapter 2 Literature Review 19 2.2.3.1 Gas phase reaction In the presence of N H 3 , NO decomposes according to the following reaction (Beer et al., 1980): NO + —NH3 -+-N2+ H20 3 6 Thermal decomposition contributes to the destruction of N 2 0 at fluidized bed temperatures (Hulgaard et al., 1991; Moritomi and Suzuki, 1990; Zhang, 1992). Chemical kinetic modelling (Kramlich et al., 1989; Miller and Bowman, 1989) shows that the most important reaction for N 2 0 removal in fossil fuel combustion is: N20 + H-+N2 +OH Kilpinen and Hupa (1991) found that under FBC conditions, the reaction with O H was also important, i.e. N20 + OH -*N2+H02 Johnson (1991) noted that the free radical concentrations of H and O H may not be high enough at FBC temperatures for these reactions to be important and postulated that the following reaction is responsible for N 2 0 removal: N20+M-*N2 +0 + M where M may be any molecule which does not participate in the reactions, but serves as an energy carrier. Chapter 2 Literature Review 20 In addition to N2, NO was also detected as a product of the homogeneous reduction of N 20 (Hulgaard, 1991a, Zhang, 1992). According to Miller and Bowman (1989) , NO is mainly produced by: N20 + 0^> 2NO 2.2.3.2 Gas-solid reactions NO can decompose by reaction with char to form N 2 (Pohl and Sarofim, 1976; Song et al., 1982; deSoete, 1990). NO +(-C )^N2+ (-CO) Similarly, the reaction between N 20 and char is (deSoete, 1990): N20 + (-C ) ->. tf2 + (-CO) Measuring the decomposition rate of NO and N 20 over three types of chars, deSoete (1990) reported that, N20 is decomposed more readily than NO by chars. Details are given in Table 3.7. 2.2.3.3 Catalytic reduction Catalytic reduction of NO occurs over limestone in the presence of CO (Tsujimura et al., 1983): NO+CO^-N0+CO. 2 1 Chapter 2 Literature Review 21 Hansen et al. (1992) showed that the presence of C0 2 poisoned the catalytic sites on the surface of CaO during NO reduction by CO. The degree of poisoning appeared to be closely related to the inlet concentration of C0 2 and to the time of exposure to NO. Catalytic reduction of N 20 has also found over certain limestones (Johnsson, 1991), i.e. N20-*N2 +^02 The catalytic reactivity of limestone has been shown to decrease as the degree of sulphation increases, since CaS04 has a very low catalytic activity (Hansen et al., 1993). Ash and sands are potential catalysts for N 20 decomposition. Compared to an empty column, the presence of Ottawa sand (Zhang, 1992) and Peat ash (Hulgaard, 1991) significantly increases the degree of N 20 decomposition at FBC temperatures. Their catalytic capabilities are probably due to the mineral matter contained. MgO, Fe203, Fe304 have moderate to strong activity for N 20 decomposition; the activity of Fe203 is close to that of CaO, while Fe304 removes N 20 almost completely at temperature as low as 870°K (Miettinen et al, 1991). 2.2.3.4 Summary In fluidized bed combustion, fuel nitrogen splits between char and volatile during pyrolysis. Both char nitrogen and volatile nitrogen contribute to the formation of NO and N20. These species will decompose subsequently under FBC conditions. Considering the fact that volatile nitrogen mainly consists of HCN and NH3, the final emissions of NO and N 20 result from competition between a number of homogeneous and heterogeneous reactions: Chapter 2 Literature Review 22 (i) oxidation of HCN and NH 3 in the gas phase; (ii) oxidation of HCN and NH 3 over the surface of bed materials char, ash, and limestones; (iii) oxidation of char-N; (iv) homogeneous decomposition of NO and N20; (v) heterogeneous decomposition of NO and N20. The global kinetics of these reactions are discussed in Chapter 3. 2.3 Hydrodynamics of Circulating Fluidized Bed Hydrodynamics, as well as reaction kinetics, affect the pollutant emissions of a combustor. Solids and gas motion have direct impacts on the solids distribution, the residence time distributions of gas and solid, the extent of gas and solid mixing and gas-solid contact efficiency. In CFB, solids distribution in the axial direction is well represented by apparent suspension density profiles derived from pressure profiles, except in the region very close to distributor where particle acceleration plays an appreciable role (Senior, 1992). Figure 2.3 gives a typical apparent suspension density measured in the UBC pilot plant (Grace et al., 1989). The rise in density at the top does not occur for reactors with smooth exits, where only a small portion of the upflowing solids are reflected back down the riser. The radial voidage profiles measured by Bader et al. (1988) and solid flux profiles (Rhodes et al., 1988) indicate that a core-annulus structure exists in a CFB riser whereby dense sheets of solids flow downward at the wall while a dilute suspension travels upward in the center. Considerable mixing of solids has been measured (Bader et al., 1988; Kojima et al., 1989). Gas mixing in CFBC has been found to be much less than in fluidized beds operated in the bubbling or turbulent fluidization regimes (Yang, 1983). Lateral 02, NOx and N 20 concentration gradients also exist in the CFBC risers. Chapter 2 Literature Review 23 Figure 2.3: Axial Suspension Density Profiles in the UBC Pilot Plant. T = 838- 876°C, bed material: Ottawa sand (Grace et al, 1989). The profiles are derived from pressure profiles. Chapter 2 Literature Review 24 Many different CFB modelling approaches have been reported, including empirical correlations, random walk theory, energy minimization, two fluids Navier-Stokes equations, etc. Yang (1992) provided a review of the most salient features of these models. Two models have been developed at UBC (Senior, 1992; Senior and Brereton, 1992): (1) a fluid mechanical approach to explore the dynamics of particles in the core by considering particle-particle collisions and particle-turbulent interactions, and (2) a semi-empirical model accounting for particle exchange between the core and the annulus by assuming the exchange of solids in a CFB is analogous to exchange of droplets in gas-liquid annulur flow. The first model starts from the particle density and size distribution to calculate particle collision frequency, mean free path, eddy strength, the distribution of lateral particle velocity, etc. The second model considers the Sauter mean diameter to predict how the apparent density and core-annulus boundary vary with height. The latter is more readily related to reactor modelling and is much easier to calculate. Senior's semiempirical model (Senior and Brereton, 1992) is therefore employed in deriving the NO and N 20 model in this thesis. Chapter 3 Chemical Kinetics of Fuel-N In general, the concentrations of various pollutant species in the effluent gas from a combustor differ from their calculated equilibrium levels (Sarofim, 1992), indicating the importance of reaction kinetics in studying the pollutant emissions from combustion. The major objective of this chapter is to provide kinetics expressions for the reactions discussed in Chapter 2, including the chemical reaction equations, the reaction rate laws and the reaction rate constants. These reactions, depending on whether their chemical kinetics expressions are available or derivable from the literature, are divided into three categories: (i) chemical kinetics expressions previously reported in the literature; (ii) chemical kinetics expressions not available in the literature, but which may be derived from related publications; (iii) cases where chemical kinetics expressions are not available and where existing data do not suffice to enable chemical kinetics data to be derived. In the following discussions, a detailed reaction scheme consisting of 15 reactions is established to account for the formation and destruction of NO and N 20 in CFBC. The kinetics expressions for type (i) reactions are retained except where there are new experimental data which are preferable. Kinetics expressions are derived for type (ii) where required. A reaction rate law is proposed for the reaction between NO and char-N. Only two of the reactions discussed in Chapter 2 belong in category (iii). These are the HCN oxidation over limestone and char surface and the reaction between NO and char-N. The HCN oxidation over char and limestone is omitted in view of this and becuase it appears to be of secondary importance. 25 Chapter 3 Chemical Kinetics of Fuel-N 26 A rate law is assumed in this thesis for the reaction between NO and char-N, with the reaction rate constant treated as a one of the fitting parameters in the NO and N20 model. 3.1 Homogeneous Reaction The homogeneous reactions considered in this chapter for the NO and N20 model include NH 3 oxidation, HCN oxidation and the thermal decomposition of NO and N20. Hulgaard et al. (1991) performed a series of experiments on the homogeneous reactions of HCN, NH3, NO and N2O using a quartz plug flow reactor between 700 to 1200°C; moderate success was achieved in using a free radical scheme to explain the product distribution for these reactions. Free radical reactions have also been employed to explain the formation of NO in hydrocarbon flames. However, these have rarely been considered in fluidized bed combustion NO modelling, partly because the number of reactions and intermediates involved in a free radical reaction scheme would be large, requiring considerable memory capacity and computation time. Furthermore, the erosive environment in a CFBC unit restricts the sampling and measurement of the very short lifetime free radical species, preventing the corroboration of free radical mechanisms under real CFBC operating conditions. The chemical kinetics expressions for the homogeneous reactions considered in previous CFBC NO modelling approaches (Gibbs et al., 1980; Lee et al., 1985; Zhao, 1992) involve so-called global kinetics. A global reaction scheme consists of the reactions of molecules rather than free radicals. Fewer reactions must be considered in a global reaction scheme than in a free radical scheme, saving substantial memory capacity and computation time. Chapter 3 Chemical Kinetics of Fuel-N 27 The following sections, based on experiments of Hulgaard et al. (1991), derive global kinetics data for the homogeneous reactions involved in the formation and destruction of NO and N 20 for application in CFBC systems. 3.1.1 Thermal Decomposition of N 2 0 Hulgaard et al. (1991) studied the homogeneous chemistry of fuel-N by investigating HCN and NH3 oxidation and thermal decomposition of N2O in a plug flow reactor. To derive the kinetics expressions for these reactions, the kinetics expressions for N 20 decomposition must be derived first because thermal decomposition of N 20 occurs at FBC temperatures and the reaction cannot be "frozen" while HCN and NH3 are being oxidized. Figure 3.1 shows the result of N 20 decomposition (Hulgaard et al., 1991). N 20 is slightly decomposed below 1150°K, with only a very small portion converted to NO. This reaction is presumed to occur via . Reaction!: N20 —^—» NO + - N2 (3.1) To account for the nitrogen balance closure, in addition to eqn (3.1), there must be another reaction path accounting for the production of additional N 2 from N20. Here the reaction is assumed to proceed according to:.-Reaction 2: N20 -^>N2 + \o2 Chapter 3 Chemical Kinetics of Fuel-N 28 250-1 900 1000 1100 1200 1300 1400 1500 Temperature (K) Figure 3,1: Product Concentrations for N20 Decomposition in a Homogeneous Plug Flow Reactor. Residence time is 59 ms at 1200°K and is assumed to be reversely proportional to temperature (°K). Inlet concentration of N20 = 208 ppm. The symbols represent experimental data points (":N20, A: NO). and : free radical kinetics calculations results with and without water present, respectively (Hulgaard et al., 1991). : modelling results based on global kinetics discussed below,. Chapter 3 Chemical Kinetics of Fuel-N 29 Based on the reaction scheme, eqn (3:1) and (3.2), and the following assumptions, mass balance equations (3.4) and (3.5) predict the concentrations of NO and N 20 in Hulgaard's experiment ( Hulgaard et al., 1991) as a function of residence time (t). The boundary condition is based on the inlet NO and N 20 concentrations. These reactions and the following equations are based on the following assumptions: (i) The gas is in plug flow. (ii) There is negligible influence of water vapor. (iii) There is negligible volume change during the course of reaction. (iv) N 20 decomposition occurs via reactions 1 and 2 only, with assumed first order kinetics, i.e. d[N2Q] dt = -kt[N20], i = lor2 (3.3) We may write mass balance equations : d[NO] - ^ = k}[N20] (3.4) d[N20] dt (kx+k2)[N20] (3.5) Boundary conditions : at t = 0, [NO] = 0 ppm, [N20] = 208 ppm Chapter 3 Chemical Kinetics of Fuel-N 30 To determine the reaction rate constants k, and k2 in the above equations at a given temperature, a trial and error procedure was adopted with trials repeated until the calculated concentrations of NO and N 20 were both within one ppm from the experimental data (Figure 3.1). Table 3.1 gives the concentrations of NO and N 20 estimated from Figure 3.1. The corresponding values of k, and k2 are summarized in Table 3.2. Table 3.1: Product Distribution ofN 20 Decomposition (estimated from Figure 3.1) Temperature (°K) N,0 ( ppm ) NO (ppm) 1220 194 0.40 1280 170 0.85 1320 150 1.25 1370 96 4.20 Table 3.2 : Predicted Global Kinetic Rate Constants k1 and k2 Temperature (°K) *,(s-i) *, (s-1) 1220 3.600-10~2 1.186 1280 7.944-10"2 3.479 1320 1.315-10"1 5.955 1370 5.570-10"1 14.440 Chapter 3 Chemical Kinetics of Fuel-N 31 It is usually accepted that a chemical reaction rate constant follows an Arrhenius function : k = k0exp(-^r) (3.6) where k0 is the frequency factor and Eact is the activation energy. Referring to the data in Table 3.2, linear regression of ln(£. I = 1 o r 2) vs. (Appendix A, Figure Al) gives: 99790 kx (s_1) = 8.013-IO8 exp(--——) (3.7) 2743 3 k2 (s'x) = 6.886- 109exp(-^y^) (3.8) With &7 and k2 given by eqns (3.7) and (3.8), eqns (3.4) and (3.5) predict the concentrations of NO and N20 shown in Figure 3.1, compared to the other two curves obtained by a free radical kinetics calculation (Hulgaard et al, 1991), with and without the presence of water, respectively. The prediction by the global kinetics model gives good agreement with the experimental data over the entire temperature range examined. This indicates that approximating the chemical reaction rate constants as Arrhenius functions does not result in large errors in the model prediction. In summary, NO and N 2 are the major products of the thermal decomposition of N20. N20 is slightly decomposed at CFB temperatures, with the extent of thermal decomposition increasing with temperature. The global kinetics is capable of accounting for the thermal decomposition of N20. Chapter 3 Chemical Kinetics of Fuel-N 32 3.1.2. NH 3 Oxidation Figure 3.2 (Hulgaard et al., 1991) shows the product concentrations for NH 3 oxidation in a plug flow reactor with a significant amount of NO introduced at the reactor inlet. NH 3 is oxidized to form NO and N20; the formation of N20 starts at a higher temperature. If it is assumed that NH3, NO, N 20 and N 2 are the' only species including nitrogen and that all the hydrogen in NH 3 ends up as water, the relevant reactions are: Reaction 3: NH3 + 02 —^N20 +1H20 Reaction 4: NH, + -02 —^->M? +-H70 3 4 2 2 2 Reaction 5: NO +—NH3 —^—>-7V2 +H20 (3.11) 3 6 plus eqns (3.1) and (3.2) to account for the thermal decomposition of N20. The rate laws for Reactions 4 and 5 derived by deSoete (1975) are given in Eqns 3.12 and 3.13, respectively. (3.9) (3.10) d[NH3] dt K [NH3][02] d[NO] = dt k5 [NH3][NO] (3.12) (3.13) Chapter 3 Chemical Kinetics of Fuel-N 33 100 - 1 CL CL 80 H 60 H 40 H 20 H NH3/IO NO/10 A- - -900 1000 1100 1200 1300 Temperature (K) 1400 1500 Figure 3.2: Product Concentrations for N H 3 Oxidation in a Homogeneous Plug Flow Reactor with N O present at Reactor Inlet (Hulgaard et al., 1991). Inlet concentrations: [02] = 26,000 ppm, [NH 3 ] = 782 ppm, [NO] = 597 ppm, [N20] = 0 ppm. Residence time is 59 ms at 1200°K and is assumed to be inversely proportional to the temperature (°K). The symbols represent experimental data (*: N H 3 , A: NO, •: N20). The dashed lines are the concentrations calculated using a free radical reaction scheme (Hulgaard et al., 1991). Chapter 3 Chemical Kinetics of Fuel-N 34 If we assume that the reaction rate law for Reaction 3 is similar to that for Reaction 4, then for Reaction 3 d[NH3 ] - ^ - ^ = - [NH,][02] (3.14) Based on the same assumptions as employed in deriving kt and kj, we may write the following set of differential equations to predict the concentrations of N H 3 , N O and N 2 0 at the reactor outlet as functions of residence time (t) d[NH3] 2 d t =-(*3 + KWH3\02]--k5[NO][NH3] (3.15) d[NO] = k4 [NH 3 ][02 ] - k5 [NH3 ][NO] + kx [N20] (3.16) d[N20] — ^ - L = 0.5k3[NH3][O2]-(kl +k2)[N20] (3.17) = - ( * 3 +1.25k, )[NH3 ][02 ] + 0.5k2 [N20] (3.18) Boundary conditions at / = 0 (in ppm) are: [ NO] = 0 [N20] = 0 [NH3 ] = 782 [ 02] = 26,000 Chapter 3 Chemical Kinetics of Fuel-N 35 With kj and k2 known, eqns (3.15) to (3.18) can predict the concentrations of NH 3 , NO and N 20 at the reactor outlet if k3,k4 and k5 are also known. Table 3.3 gives the predicted concentrations of NH 3, NO and N20 corresponding to Fig 3.2 for various temperatures. The values of k 3,k 4andk 5 in Table 3.4 were determined by trial and error. By substituting those values into eqn (3.15) to (3.18), the concentrations of NH 3 , NO and N 20 calculated are all within a tolerance of one ppm from the experimental data in Table 3.3. Table 3.3: Product Distribution ofNH3 Oxidation (estimated from Figure 3.2) Temperature NH, (ppm) NO (ppm) N20(ppm) 1150 620 420 1 1170 470 305 3 1220 115 100 13 1270 12 106 17 1305 2 150 21 1370 0 255 22 Table 3.4: Predicted Global Kinetics Reaction Rate Constants k3,k4 andk5 Temperature k3 (ppm-'s-') . k 4 (ppm-'s'1) k5 (ppm~]s-]) 1150 5.0 xlO"7 2.40 x 10 s 9.50 xlO"3 1170 6.0 x IO"6 7.20X10 5 2.30 xlO' 2 1220 6.5 x IO"5 4.25 x 10"4 1.65 x 10 1 1270 1.9 xlO"4 1.20xl0"3 3.90 x l O 1 1305 3.9 xlO"4 1.80 xlO"3 4.50 x l O 1 1370 9.8X10"4 3.10 xlO"3 5.0 x l O 1 Chapter 3 Chemical Kinetics of Fuel-N 36 A linear regression of ln(Ar.,.=3to5) vs. y for the data in Table 3.4 ( Appendix A, Figure A2 ) yields Arrhenius type reaction rate constant functions for k3 to k5 : 50719 k3 (ppm's1) = 2.597 x 1013 exp(- =^—1) (3.19) kA (ppm's'1) = 5.797- 10s e x p ( - ^ ^ ) (3.20) 29378 ks(ppmlsA) = 2.234 x 109 e x p ( — ( 3 . 2 1 ) Due to the deviation from linearity in ln(& I I = 3 t o 5) vs. ^ , substituting the rate constant functions, eqn (3.19) to (3.21), for k 3, k4 and k5 into eqn (3.15) to (3.18) instead of the values given in Table 3.4 results in larger errors in model predictions. Nevertheless, Figure 3.3 shows that the global kinetics model prediction using Arrhenius type reaction rate constants shows favorable agreement with the experimental results, comparable to the free radical kinetics model of Hulgaard et al (1991). Chapter 3 Chemical Kinetics of Fuel-N 37 1 0 0 - , 900 1000 1100 1200 1300 1400 1500 Temperature (K) Figure 3.3: Predicted Product Concentrations for NH3 Oxidation in a Homogeneous Plug Flow Reactor (Hulgaard et al., 1991). Inlet concentrations: [02] = 26,000 ppm, [NH3 ] = 782 ppm, [NO] = 597 ppm, [N20] = 0 ppm. Residence time is 59 ms at 1200°K and is assumed to be inversely proportional to the temperature (°K). The symbols represent experimental data (*: NH3, •: NO, •: N20). The dashed lines are the calculated concentrations by a free radical reaction scheme (Hulgaard et al., 1991). The solid lines are obtained from the global kinetics model described above. Chapter 3 Chemical Kinetics of Fuel-N 38 The rate laws, eqns (3.12) to (3.14), and the Arrhenius functions for k3 to k5 , eqns (3.19) to (3.21), were also evaluated for another experimental condition, with no NO present at the reactor inlet while NH 3 was oxidized. The comparison between the model prediction, the experimental data and a free radical kinetics calculation result of Hulgaard et al. (1991) is shown in Figure 3.4. Both the global kinetics and the free radical kinetics model show large deviations from the experimental data at high tempeperatures, but the predicted trends match the experimental trends. This implies that the reaction pathways for NH 3 oxidation with or without excess NO may not be identical. NH 3 oxidation was studied long before N 20 caused concern in fluidized bed combustion. The reaction rate constants derived by deSoete (1975) are: k4(kmole m"3 s 1) = 3.28-106 7/exp(--^^) (3.22) k5(kmole m3 s"1) = 1.48• 107 Texp(--^^) (3.23) The above reaction rate constant functions were derived from combustion of NH 3 in hydrocarbon flame between 1900 and 2500°K, which is much higher than for fluidized bed combustor operation. N 20 was not included in order to close the mass balance of fuel-N in the experiments. Therefore, eqns (3.22) and (3.23) are not valid for CFBC NO/N20 modelling. Chapter 3 Chemical Kinetics of Fuel-N 39 N O 900 1000 1100 1200 1300 1400 1500 Temperature (K) Figure 3.4: Product Concentrations for NH3 Oxidation in a Homogeneous Plug Flow Reactor without NO Present at the Inlet (Hulgaard et al., 1991). Inlet concentrations: [02] = 26,000 ppm, [NH3 ] = 760 ppm, [NO] = 0. ppm, [N20] = 0 ppm. Residence time is 56 ms at 1200°K and is assumed to be inversely proportional to the temperature (°K). The symbols represent experimental data (A: NO, •: N20). The dashed line is the calculated concentration by a free radical reaction scheme (Hulgaard et al., 1991). The solid line is calculated based on the global kinetics expressions derived from Figure 3.3 Chapter 3 Chemical Kinetics of Fuel-N 40 The global kinetics expressions derived in this section are for the experiments at lower temperatures, 900 to 1500°K, covering the CFBC operating temperature range. The conversion of NFf3 to both NO and N 20 were considered in the experiments. Fair agreement has been achieved for a plug flow reactor under different conditions. Therefore, the global kinetics are employed to account for the formation of NO and N 20 from homogeneous NH 3 oxidation in the CFB combustor in this thesis. 3.1.3.Oxidation of HCN Figure 3.5 due to Hulgaard et al. (1991) shows experimental results for HCN oxidation in a plug flow reactor. The major products are NO, N20 and N2. The formation of NO and N 20 starts at around 1200°K, and increases-with temperature, 32 ppm being formed at 1350° K. To account for the product distribution at the reactor exit, here a parallel reaction scheme is proposed as below: Reaction 6: HCN +—02 —k-^>—N20 + C02 +—H20 (3.24) 2 2 2 Reaction 7: HCN + ^ 02 —NO +C02+^H20 (3.25) Reaction 8: HCN+ ^02-^C02+^N2+^H20 (3.26) together with eqn (3.1) and (3.2) to account for N 20 decomposition. The rate laws for Reaction 6, 7 and 8 are all assumed to be of the form: Chapter 3 Chemical Kinetics of Fuel-N 41 500 - i 400 -A 900 1000 1100 1200 . 1300 1400 1500 Temperature (K) Figure 3.5: Product Concentrations for HCN Oxidation in a Homogeneous Plug Flow Reactor without NO Present at the Inlet (Hulgaard et al., 1991). Inlet concentrations: [0 2] = 24,500 ppm, [HCN] = 320 ppm, [NO] = 0 ppm, [N20] = 0. Residence time is 56 ms at 1200°K and is assumed to be inversly proportional to temperature (°K). The symbols represent experimental data (•: HCN, A : NO, •: N 20). The dashed lines are the concentrations calculated using a free radical reaction scheme (Hulgaard et al., 1991). The solid lines represent the calculation results using the global kinetics scheme. Chapter 3 Chemical Kinetics of Fuel-N 42 d^HCN^ = . k [HCN][02], i = 6 to 8 (3.27) dt The following set of mass balance equations predicts the concentrations of HCN, NO and N 20 at the reactor exit with t equal to the residence time: d[IffN] = ~(K +k1+k% )[HCN][02 ] (3.28) ^^• = k1[HCN][02-\ + k,[N10\ (3.29) d[N2Q] dt = 0.5k6[HCN][O2]-(k] +k2)[N20] (3.30) ^dT = ~(2ke + 4 kl +\ks)[HCN][O2\ + 0.5k2[N20] (3.31) Boundary conditions at t = 0 are: [NO] = 0 ppm [N20] = 0 ppm [HCN] = 320 ppm [ 02 ] = 24,500 ppm Chapter 3 Chemical Kinetics of Fuel-N 43 The concentrations of HCN, NO and N 20 shown in Figure 3.5 at various temperatures are listed in Table 3.5. Table 3.6 summarizes the fitted values of k6,k1 and ks employed to make the model predictions of the exit concentrations of HCN, NO and N 20 all within a tolerance of one ppm from the values in Table 3.5. Table 3.5: Product Distribution of HCN Oxidation (estimated from Figure 3.5) Temperature (°K) HCN (ppm) NO (ppm) N20 (ppm) 1220 175 25 5 1270 100 30 10 1320 66 35 30 1350 33 40 32 Table 3.6: Predicted Global Kinetics Reaction Rate Constants k6,k7 and Jfc, Temperature(°K) k 6 (/?/wTV) k 7 (ppm') k 8 (ppm-xs-x) 1220 3.40 x IO'5 7.70 xlO"5 3.40 xlO"4 1270 9.20 x10 s 1.25 x Kr 4 6.85 xHT4 1320 3.70 xlO"4 1.76 xlO"4 7.30 x 10'4 1350 6.05 x IO"4 2.65 xlO"4 1.02 x 10"3 Arrhenius type functions for the reaction rate constants k6,k7 andk%, derived by applying linear regression of \n(k) vs. ^  for the data in Table 3.6 (Appendix A, Fig A3) lead to: Chapter 3 Chemical Kinetics of Fuel-N 44 k6(ppmlsl) = 7.983 * 108 exp(-37639 ) (3.32) T k^ppm's1) = 1.610 x 101 exp(-14960 ) (3.33) T k&(ppmlsl) = 1.172 x 101 exp(-12632 T ) (3.34) These global kinetics expressions have been derived for H C N oxidation without NO introduced at the reactor inlet. To see the applicability of the global kinetics expressions at different conditions, a calculation has been made for conditions where a substantial amount of NO is introduced into the reactor with H C N and 0 2 . Shown in Figure 3.6 are the global kinetics modelling result, experimental data and a free radical kinetics modelling result (Hulgaard et al., 1991). The prediction by the global kinetics calculation is in good agreement with experimental data at CFBC temperatures. However, large deviations arise at higher temperatures. This indicates that, as discussed above for N H 3 oxidation, the oxidation mechanism may be different with excess N O present. Large errors also occur in the free radical kinetics calculation. Since the global kinetics calculation shows better agreement with the experimental data at CFBC temperatures, this again indicates that the accuracy achieved by a free radical kinetics model does not justify the added complexity of calculation incurred. Chapter 3 Chemical Kinetics of Fuel-N 45 900 1000 1100 1200 1300 1400 1500 Temperature (K) Figure 3.6: Product Concentrations for HCN Oxidation in a Plug Flow Reactor with NO Present at the Inlet (Hulgaard et al., 1991). Inlet concentrations: [02] = 22,800 ppm, [HCN] = 303 ppm, [NO] = 441 ppm, [N20] = 3 ppm. Residence time is 56 ms at 1200°K and is assumed to be inversely proportional to the temperature. The symbols represent experimental data (•: HCN, A : NO, • : N20). The dashed lines give the concentrations calculated using a free radical reaction scheme (Hulgaard et al., 1991). The solid lines have been calculated based on the global kinetics modelling. Chapter 3 Chemical Kinetics of Fuel-N 46 3.2 Heterogeneous Reactions In a CFBC unit, heterogeneous reactions of fuel-N occur over the surfaces of char and limestone particles. Chemical kinetics expressions for some of these reactions are available in the literature. The next section compiles the chemical kinetics expressions for these heterogeneous reactions. These kinetics expressions, in combination with those derived for the homogeneous reactions in section 3.1, are employed for the reactions of fuel-N in circulating fluidized bed combustion. 3.2.1 NO Formation Char nitrogen oxidation and catalytic oxidation of NH 3 over limestone surfaces lead to the formation of NO in a CFBC unit. These two reactions are denoted (Pohl and Sarofim, 1976; Song et al., 1982) as: Reaction 9: N(s)+^02 -^^> NO (3.35) Reaction 10: NH3+^02 —NO+^H20 (3.36) In modelling, the nitrogen-to-carbon ratio in char is usually assumed to remain a constant during the combustion, indicating that carbon and nitrogen are oxidized in proportion to their concentration. Thus, the reaction rate for Reaction 9 is a product of the carbon combustion rate and the nitrogen-to-carbon ratio, i.e., d[NO] , „ r n • d t =k9Neat[02] (3.37) Chapter 3 Chemical Kinetics of Fuel-N 47 where k9 is the char combustion rate constant, ae is the available surface area per unit volume of char and 7VC is the nitrogen-to-carbon ratio in the char. The reaction rate expression is employed in this thesis, but with allowance for Nc to differ from that in inert pyrolyzed chars. The combustion rate constant is determined from the 0 2 sub-model by fitting the exit 0 2 concentration to the experimental data from the U B C pilot plant (see Chapter 4). Johnsson (1988) proposed a rate law for Reaction 10 as: =k,0[NH3][O2] (3.38) a t The reaction rate constant, after unit conversion, is 2.72-10 2 ppm ]s 'at 1073 °K, based on experimental data of Furusawa et al. (1985). 3.2.2 N z O Formation As discussed in Chapter 2, there is experimental evidence showing that the reaction between N O and char-N accounts for a significant portion of N 2 0 formation in CFBC. This reaction is denoted by : Reaction 11: N{s)+NO — N 2 0 (3.39) The reaction rate laws and the associated reaction rate constants are not available in the literature. Chapter 3 Chemical Kinetics of Fuel-N 48 It has been reported (Tullin et al., 1993, Horio et al., 1993) that the reaction is favoured when excessive 0 2 is present during char combustion. Since the reaction occurs between NO and char-N, the following rate expression is assumed. ^P--knaeNc[NO] (3.40) where ku is a model parameter to be determined from the combustion run data of the UBC pilot plant unit (Chapter 6). ae is the available char surface per unit volume of char, and Nc is the nitrogen-to-carbon ratio in char, equal to that in eqn (3.37). Note that in Tullin's experiment, no N 20 was formed when oxygen was absent; eqn (3.40) applies only when the molar ratio of 0 2 to carbon is larger than the stoichiometric ratio iof combustion reaction. 3.2.3 Decomposition of NO and N 20 by Char Char has a very strong ability to decompose NO and N 20 according to two reactions (deSoete, 1990): Reaction 12-.NO + C ( S ) —^N2 +CO (3.4 J) Reaction J3: N20 + C(s) k'3 > N2+CO (3.42) By assuming that these two reactions are first order in NO and N 20 respectively, i.e. d[NO] dt -k}2[NO] (3.43) Chapter 3 Chemical Kinetics of Fuel-N 49 Table 3.7: Decomposition Rate Constants k12 and k13 for Chars (deSoete, 1990) Fuel, (30-50 um) kl2(m'kg-'s-<) kn at 1123°K * I 3 0 » 3 * r , O k13 at 1123°K "Cedar Grove" char 275.W-7 0) 0.09 267.r.exp(-17°) 1.3 "Prosper" char 355.r.exp(-"^°) 0.151 M l - r . e x p C -1 ^ ) 4.3 "Eschweiler" char 1723T.=xp(-1™°) 0 .25 41.6.r.exp(-17°) 6.3 deSoete (1990) reported rate constants for three types of chars as shown in Table 3.7. 3.2.4 Catalytic Decomposition of NO and N 2 0 on the Surface of Calcined Lime Over the temperature range of 400 to 840°C, NO reacts with CO to form C0 2 over the limestone surface (Furusawa et al., 1980; Tsujimura et al., 1983). The reaction is denoted by: Reaction 14: NO + CO —-N2+C02 (3.45) The rate law and the rate constant suggested by Tsujimura (1983) are Chapter 3 Chemical Kinetics of Fuel-N 50 d[NO] dt = -ku[NO][CO] (3.46) ku (kmole'/w V ) - 2.13 • 1014 Texp(-^=^-) (3.47) Limestone has a catalytic effect on the decomposition of N 20 Reaction 15: N20 k'5 > N 2+^0 2 (3.48) Johnsson (1991) evaluated the reaction rate constant from Miettinen et al.(1991), Khan et al. (1991) and Moritomi et al.(1991) by assuming that the reaction is first order. d[N2Q] dt = -k,5[N20] (3.49) The results are summarized in Table 3.8. Table 3.8 : First Order Decomposition Rate Constants for N 20 over Calcined Limestone, Johnsson (1991).*15 = *15i0 exp(-^). Material ^15,0 (m 3kgV) R V ' ^15,1123°K (m3kgV) References CaO 1.3103 9850 0.2 Miettinen et al.(1991) Crystalline limestone NA NA 0.06 Amorphous limestone NA NA 0.01 Limestone 1.3 4000 0.04 Moritomi et al.(1991) Chapter 3 Chemical Kinetics of Fuel-N 51 3.3 Summary The production and destruction of NO and N 20 from CFBC have been accounted for by 15 reactions. Table 3.9 summarizes the reactions and the corresponding global rate laws, for these reactions, Table 3.10 summarizes the rate constants. Figure 3.7 gives a schematic representation of these reaction pathways. Except for Reaction 11, chemical kinetics expressions have been derived from experimental work, obtained here or by other researchers. The kinetics expressions for the homogeneous reactions have been derived from a very crude approach in which the rate laws are assumed to be first order with respect to each reactant, while the temperature dependence of the reaction rate constants has been approximated by Arrhenius expressions. Nevertheless, the global kinetics expressions derived for N 20 thermal decomposition, NH 3 oxidation and HCN oxidation appear to simulate the product distributions very well. The kinetics expressions for the heterogeneous reactions have been obtained from the work of other researchers. Reaction rate constant k u is an undetermined parameter to be determined by fitting the UBC CFBC pilot plant data. Besides the fifteen reactions, there may be other reactions which affect the formation and destruction of NO and N20, e.g. NO reduction by H^ in the presence of various catalysts, a reaction between NH 3 and HCN (Hulgaard, 1991), and enriched N 20 production in the presence of CO when HCN is oxidized (Hulgaard, 1991). Available information in the literature does not provide sufficient kinetics data for these reactions, so they are not included. To better understand the chemical kinetics of fuel-N in CFBC, more effort is required in the future to find other reaction pathways, and to derive the kinetic expressions more precisely at fluidized bed temperatures. . Chapter 3 Chemical Kinetics of Fuel-N Table 3.9: Summary of Reaction Equations and Rate Laws 52 No. Reactions Rate expressions 1. N,0—^NO+-N1 2 2 2. N20-^N2+±02 3. NH, + 02—^-N20 + -H20 3 2 2 2 2 2 4. NH, + 5 0. —±-^>NO + 3 tf,0 3 4 2 2 2 5. N0+-NH,-±->-N2+H20 3 3 6 2 2 6. HCN + -02 —k-^-N20 + C02 + -H20 2 2 • 2 2 2 2 2 df 7. HCN + -02 —i^NO+CO, +-H,0 4 2 2 2 2 r._AHCN] k A H C N ] l 0 ] dt 8. //CJV+ 5 0, — ^ C O , + 1 N2 + ' #,0 4 2 2 2 2 2 2 at 9. NM+±02-^N0 r , - -d ^ ~ - k M 0 2 ] 10. NH, + 5 02—!*->NO + 3 H,0 3 4 2 2 . 11. NO + N^—^N20 . r „ . W - ^ . . | A « l 12. NO + C„, —^-N2 +CO 2 „ = ^ 0 ] = - W [ ^ , 13 JV20 + C ( 0 — A f 2 + C 0 14. NO + CO -^-y -N2 +C0, 2 2 2 df 15. The reaction rates r9 to r ; 5 are based on the volume of char or limestone. Pchar» Pcao (kg/m3): apparent densities of char and calcined limestone used to convert the reaction rate expressions in which they are involved to the ones based on solid volume. ae (1/m): external surface per unit volume of char, Nc. nitrogen-to-carbon atomic ratio Chapter 3 Chemical Kinetics of Fuel-N Table 3.10: Summary of Reaction Rate Constant Expressions 53 Reaction rate constants Unit Source 1. 29290 kx = 8.013 x 108 exp( — ) Derived from Hulgaard(1991) 2. 27433 k2 = 6.886 x l0 9 exp(- ) Derived from Hulgaard(1991) 3. n , 50719, k3 = 2.597x 1013 exp( — ) ppm~'s~' Derived from Hulgaard(1991) 4. 8 , 34765, k, = 5.797 x 10s exp( — ) ppm~'s~' Derived from Hulgaard(1991) 5. 29378 k5 = 2.234 x 109 exp( — ) ppm~xs'x Derived from Hulgaard(1991) 6. . , 37639, k6 =7.983 xl0 8 exp( — ) ppm~'s~* Derived from Hulgaard(1991) 7. , 14960, jt7 = 1.610 xl0'exp( — ) ppm^s'1 Derived from Hulgaard(1991) 8. • , 12632, ks =1.172x10' exp( — ) ppmT^s'* Derived from Hulgaard(1991) 9. , 150732, k9 = 595x 7- xexp( — ) s'{m Field et al(1967) 10. kw = 2.72 xlO" 2 at 1073° K ppm'is'x Furusawa(1985) 11. a fitted parameter 12. , 16900, kn = 275-7- -exp( — ) m'kg-]s-] deSoete(1991) 13. , 13900, *„ = 2 6 7 - r - e x p ( ~ p - ) m3kg-ls'-1 deSoete(1991) 14. 2 8920, jfc14 =1.52xl0 2exp( p ) ppm~'s~* Tsujimura(1983) 15. , 4000, A 1 5 - 1.3-exp( ) m'kg-'s-' Jan E r i c Johnsson(1991) k 1 2 and k 1 3 listed in Table 3.10 are those determined for "Cedar Grove " char by deSoete (1991). In the NO/N 2 0 model presented in Chapter 4, k 1 2 k 1 3 are fitting parameters as well as k,,. Chapter 3 Chemical Kinetics of Fuel-N 54 Fuel-N N2 Figure 3.7: Formation and Destruction Reaction Pathways of NO and N 2 0 in CFBC. Reaction rate constants for the reactions denoted by the numbers i, i= 1-15 are k; defined in sections 3.1 and 3.2 . Chapter 4 NO/N20 Model Development NO and N 20 emission models are needed for improved design and operating strategies for FBC combustors. A number of NO models have been established for both bubbling and circulating fluidized bed combustions ( e.g. Horio et al., 1977; Gibbs, 1980; Beer et al., 1980; Zhao, 1992). These vary in the reactions and the hydrodynamic structures considered. N 20 model is not currently available in the literature. The original objective of this thesis was to develop an N 20 model for CFBC. However, as there is increasing evidence that NO and N 20 emissions in CFBC are interrelated (Amand and Leckner, 1991; Tulli'n et al, 1993), it was later decided to model NO and N 20 together in this thesis. A computer program package is developed in this chapter to calculate the concentration profiles of 02, NO and N 20 in a CFB riser based on a hydrodynamic component, a devolatization model, an 0 2 model and an NO/N20 model. 4.1 Hydrodynamic Model To model the NO and N 20 emissions from CFBC, it is important to know the distribution of particles in the riser as a function of the operating conditions. The CFB hydrodynamic model of Senior and Brereton (1992) is employed here as a basis for the NO/N20 model. Since the details of this model are available in the original paper, only the major assumptions and a summary of the input and output data are given here. ,55 Chapter 4 NO / N20 Model Development 56 The main assumptions for the model are that the solids exchange in the gas-solid annular flow is analogous to droplet exchange in gas-liquid annular flow. Equations used to model droplet motion in gas-liquid annular flow (Wallis, 1970 a, b), with some parameters re-derived from experiments in the UBC CFBC unit are used to predict the particle motion in a CFBC riser. Riser geometry, primary to secondary air ratio P/S, superficial gas velocity Ug, solid circulation rate Gs, and the terminal settling velocity Vt for particles of Sauter mean diameter Ds comprise the complete set of input data for this program. P/S and U g and Ds are documented for combustion runs in the UBC pilot plant ( Brereton et al. 1992). Vt can be calculated readily (e.g. Gift et al., 1978). Gs is determined by trial and error until the average suspension density throughout the riser, pave, predicted by the model matches the experimental data within a tolerance of 1 kg/m3. The Senior and Brereton model focuses on the secondary air zone. It calculates the cross-sectional averaged suspension density (y ), the voidage in the core (sc), the area of the core and annulus (Ac and Aa) and the total perimeter of the annulus phase in contact with the core (Pa) for all levels above the secondary air port. Ac, Aa and Pa are shown schematically in Figure 4.1. The solid distribution in the primary air zone is accounted for by an empirical correlation. If a single phase structure is assumed for the solid distribution in the primary air zone, i.e. the solids are distributed uniformly in a cross-section, knowing the axial profile of y in this region would enable us to characterize the solid hold-up, (1-s) throughout the region, i.e. Y=(l-e)Psand (4.1) where 6 is the average voidage in a cross-section. However, some experimental results (Senior, 1992; Zhu, 1994) indicated that, rather than a single phase structure, Chapter 4 NO / N20 Model Development 57 Figure 4.1: Schematic Representation of Area A a , A c and Perimeter P a in the Core-Annulus Structure (Senior and Brereton, 1992) Chapter 4 NO / N20 Model Development 58 a core-annulus structure is already present in the primary air zone, though the transition from the core to the annulus is not as abrupt as above the secondary air ports. As a first approximation, the solid distribution in the primary air zone can also be simulated by a core-annulus model. Given the cross sectionally averaged suspension density y in a cross-section of the primary air zone, which is calculated by an empirical correlation in Senior and Brereton's model, some assumptions must be introduced to simulate the solids distribution by a core-annulus structure so that ea, sc, Aa, Ac, and Pa can be derived as functions of height. Note that 8 a , the voidage in the annulus, has been assigned a value of 0.6 for the annulus region above the secondary air port (Senior and Brereton, 1992). This assumption is also applied to the primary air zone, except where the suspension density y is very high, i.e. where the cross-sectional average voidage calculated from eqn (4.1) is less than 0.6. In this case, which usually happens at the bottom of the riser, ea is assigned a value of 0.55. To calculate the voidage in the core, ec, it is assumed that the solids hold-up, 1 - ec, is proportional to the cross-sectional average suspension density y, i.e. (4.2) Y 1- S I sec c, sec where the subscript sec denotes the secondary air port. Since y s e c and ecsec can be calculated from the Senior and Brereton model, sc can be calculated from y using eqn (4.2). Aa and Ac are related to ec and sa by the following balance equations, where At is the total cross-sectional area of the riser. ( i - * j 4 . + ( i - f f e H = — 4 (4.3) P sand Chapter 4 NO / N20 Model Development 59 Aa +A=A, a c t (4.4) For simplicity, the core region is calculated as if the core region were a square coaxial with the axis of the riser, i.e. Pa is correlated to A c by Equations (4.1) to (4.5), supplementary to the correlations provided in the Senior and Brereton model, can predict the profiles of sa, sc, A a , A c , and Pa throughout the riser. Figure 4.2 shows the prediction of the suspension density for Run 17-1 of Brereton et al. (1992) where Conoco Coke was burned in the UBC CFBC pilot plant. Figure 4.3 plots the voidages in the core and the annulus, i.e. ec and ea. Figure 4.4 shows the predicted areas of the annulus and core, A a and A c . Figures 4.2 to 4.4 are typical of the model predictions for the UBC pilot plant runs. This hydrodynamic model has been tested against the suspension density profiles obtained from the UBC pilot plant and the Studsvic combustor. For the operating range, 1 < P/S < 1.1, 6 < U g < 10 m Is, 20 < Gs < 60 kg/m2 s(Senior and Brereton, 1992), the model predictions are quite accurate. However, large errors were found in simulating some high suspension density combustion runs in the UBC pilot plant. For example, for Run 6.2 for Highvale coal in the UBC pilot plant, where U g = 8.5 m Is, the highest possible value for p a v e predicted by the model is 150 kg/m3, 20% lower than measured experimentally (Brereton et al, 1992). The maximum corresponds to Gs = 120 kg / m2 s. After the maximum is reached, increasing Gs does not increase the predicted average suspension density but leads to difficulty in solving the model equations. This is probaly because many of the model parameters employed in the model were based on data derived for low suspension density conditions, so that they are not applicable for high suspension density operating conditions. (4.5) Chapter 4 NO / N20 Model Development 60 Figure 4.2: Predicted Axial Suspension Density Profile for the UBC Pilot Plant Unit: Run 17-1, Conoco Coke. For conditions and details, see Brereton et al (1992). Chapter 4 NO /N20 Model Development 61 1.0 -1 0.8 0.6 A <D O) CO •g £ 0.4-1 0.2 0.0 n—r 0 1 — C o r e - -Annulus 1 1 I 2 3 I 1 I 4 5 1 I ' 6 I ' I 7 8 Riser Height Figure 4.3: Predicted Core Voidage Profile for the UBC Pilot Plant Unit: Run 17-1, Conoco Coke. The solid line is the core voidage, ec, the dashed line shows the assumed annulus voidage, ea. For conditions and details, see Brereton et al. (1992). Chapter 4 NO /N20 Model Development 62 0.025 - i 0.000 A - - - - Rf = 0.75, Core - - - - Rf = 0.75, Annulus Rf = 0.86, Core Rf = 0.86, Annulus - p - r 4 I 1 I 1 I 1 I 5 6 7 8 Riser Height (m) Figure 4.4: Predicted Core and Annulus Area Profiles for the UBC Pilot Plant Unit: Run 17-1, Conoco Coke. The solid line gives the predicted core area, Ac, while the dashed line shows the annulus area, Aa. For details, see Brereton et al. (1992). Chapter 4 NO/N20 Model Development 63 In summary, with some modifications, the Senior and Brereton model (1992) provides axial profiles of y, ec, Aa, Ac and Pa, along the riser and predicts how they change with operating conditions. This hydrodynamic information is required in the devolatilization model, the 02 model, and the NO/N20 model. The hydrodynamics model provides good predictions when the combustor is operated at low suspension density, but large deviations occur for high suspension density conditions. 4.2 Devolatilization Model The generally accepted picture for solid fuel combustion is that fuel first undergoes changes into char and volatiles, followed by a series of homogeneous and heterogeneous reactions involving the C, H, N, S atoms in the fuel. The volatiles released during the devolatilization process contain low molecular weight hydrocarbons which affect the profiles of 02, HCN, NH3, NO and N20. Oxygen is consumed rapidly by the hydrocarbons, while NO and N 20 are formed from the oxidation reactions of HCN and NH 3 as discussed in Chap 3. To model the concentration profile of 02 for a CFBC unit, it is important to know the hydrocarbons release rate along the CFB riser. At the same time, the HCN and NH 3 release rates are essential for the NO and N 20 modelling. A model is established here to calculate the release rate of hydrocarbons, HCN and NH 3 in the CFBC riser during the devolatilization process. The following assumptions are employed in the devolatilization model: 1. The hydrocarbon compounds in the volatiles exist as H^  and CH 4 exclusively. The reactions between oxygen and H 2 and CH 4 occur instantaneously as follows: CHA + 202 -> C02 + 2H20 (4.6) Chapter 4 NO/N20 Model Development 64 H2+^02-+H20 (4.7) 2. The total volatile release rate in the riser is equal to the fuel feed rate, m, times the weight percentage of volatiles in the fuel. Volatiles are evolved from the char particles at a constant rate given the fuel feed rate. Char is thoroughly mixed with sand, the char-to-sand volume ratio can be regarded as the same throughout the CFBC riser (Zhao et al., 1994). 3. HCN accounts for 60% of the volatile-N released during the devolatization process (Houser et al., 1981), with NH 3 accounting for the remaining 40% . According to assumptions 1 and 2, if we model the primary zone in the CFB riser as a single phase, the total rate at which oxygen is consumed by hydrocarbon (kmole/s) in an cross-section is the product of the following expression and an incremental height dz: V - / i m W c g - A > r j - 1 m W » - A<? \ 1 where the first term in the brackets accounts for the consumption of oxygen by CH4, with 2 being the stochiometric coefficient of 02 in equation (4.6) and WCH4 is the weight fraction of CH4 in the fuel. The second term in the brackets accounts for the consumption of 02 by H2, with 1/2 being the stochiometric coefficient of 02 in equation (4.7). WH2 is the weight fraction of H2 in fuel, while the second 2 is the molecular weight of H2. H is the height of riser. The definitions of Aa, At, y, pme are the same as in section 4.1. The consumption rate of oxygen by the hydrocarbons in the core and the annulus are given by eqns (4.9) and (4.10) where a and P are the suspension densities in core and annulus respectively : Chapter 4 NO/N20 Model Development 65 m • W c„ 4 Aca | 1 m • WHi Aca J _ 16 'AtPme+2 2 ' AlPJ H <4-9) y _ N M • WCH4 AAB 1 m • WHI AAB 1 Following the same rationalization as above and based on assumptions 2 and 3, we may represent the HCN and NH 3 release rates in the primary air zone, the core and the annulus in the secondary zone by eqns (4.11) to (4.16) below, where N v o l is the weight fraction of Vol-N in the fuel, and 14 is the atomic weight of nitrogen: v -n « fm'N^ A<Y \ 1 V HCN, sp — U. O • ( ) / A j n 14 A p ' H (4-n) ti ave HCN, c — U.O-I J / A 14 - AtPJ H (4-12) K /fCAf, a — U . O ( ) / A , } l 14 4 P ™ # ^ Chapter 4 NO/N20 Model Development 66 V NHySp 04-14 (4.14) NH, OA 14 (4.15) V N H 3 , a =0.4-(- 14 ) H (4.16) The input data for the devolatilization model include the fuel feed rate m, the fuel composition, the average suspension sensity p a v e throughout the riser, the reactor height H, the suspension densities in the core and the annulus, a and P, and the area of core and annulus, Aa and Ac, as functions of height. These are available for the combustion runs in the UBC CFBC unit (Brereton et al., 1992), with a, p, Aa and Ac calculated by the hydrodynamic model. The output values include the profiles of the oxygen consumption rate (Vo2 j, i = sp or a or c) due to hydrocarbons and the HCN and NH 3 release rates ( VHCN.J and VnH3.i, i = sp or a or c). The oxygen consumption rate by hydrocarbons is imported to the 0 2 sub-model, while the HCN and NH 3 release rates are needed for the NO/N20 model. The calculations are conducted by Excel computer spread sheet. Chapter 4 NO/N20 Model Development 67 4.3 O z M o d e l A model is developed in this section to calculate the 0 2 concentration profile along the CFBC riser in the UBC pilot plant. The prediction of the 0 2 profile is important for the NO/N20 model because the local formation rates of NO and N20, from oxidation of both volatile-N and char-N, are proportional to the local oxygen concentration. Due to the small weight fraction of nitrogen in the fuel, 0 2 and NO are generally modelled separately, i.e. ignoring the effect of N oxidation on the 0 2 profile (Beer et al., 1980; Zhao, 1992). Again in this thesis 0 2 is modelled assuming that the consumption of oxygen by fuel nitrogen oxidation is negligible, with NO and N 20 then modelled based on the calculated oxygen profile. The 0 2 model developed here considers the combustion of both char and the hydrocarbons in the volatiles. For char combustion, with the assumptions that C0 2 is the sole product, that the reaction is controlled by chemical kinetics over the temperature range of interest, and that combustion occurs at the external surface of char particles only, the combustion rate of char is : rc = ~ ^ = Aekc[02] (4.17) where Ae is the external surface area of the char particle, and kc (m /s) is the combustion rate constant. Because the combustion rate constant depends on the fuel species, kc is a model parameter used to fit the 0 2 concentrations in the core measured at the top sampling probe. This also makes sure that the theoretical 0 2 profiles is close to the experimental one. The combustion of volatiles is discussed in the devolatilization model (section 4.2). The oxygen consumption rates by volatiles in the primary zone, the core and the annulus are given by eqns (4.8) to (4.10). Chapter 4 NO/N20 Model Development 68 The solid distribution in this model is simulated by the hydrodynamic model discussed previously. A distinct core-annulus structure is assumed to exist above the secondary air port, but not yet conclusive for the primary air zone. The solid distribution above the secondary air port is simulated by a core-annulus model. However, in the primary zone, both a single phase model and a core-annulus model are tested (see section 4.3.1 and 4.3.2. respectively.). Attempts are made to model the reaction in the riser only. Combustion of fuel in the cyclone and in the solid return loop are not considered. The assumptions common to the 0 2 models developed are as follows: (i) The increase in the volumetric flow rate of gas due to reaction is negligible. (ii) In the secondary zone, gas is in plug flow in the core phase, while there is negligible vertical movement of gas in the annulus. (iii) The gas exchange between the core and the annulus is characterized by a constant mass transfer coefficient, Kca. (iv) The primary and the secondary air are mixed thoroughly at the secondary air port level. The validity of assumption (iv) depends on the penetration of the secondary air jet into the riser, which is affected by the configuration of secondart air port and P/S ratio. 4.3.1 Single phase primary air zone. In this section, the solid distribution is simulated by a core-annulus model for the secondary air zone, while a single phase model is employed for the primary air zone. The gas flow in the primary air zone is assumed to be in plug flow or perfect mixing (see Figures 4.5a and 4.5b). Chapter 4 NO/N20 Model Development 69 (a) 2nd air (b) 2nd air £ Core £ T" f- -> <r- -> <--» ( - -> «-Annulus Primary air £ Core T <--> <--> «--> <--> <--> ArtrtulBi ~7T\ Primary air Figure 4.5: Schematic Representation of the Single Phase 0 2 Model. Solid distribution in the primary air zone is simulated by a single phase: (a) plug flow assumed in primary zone; (b) perfect mixing assumed in primary zone Chapter 4 NO/N20 Model Development 70 The mass balance equations for 02 are summarized in Table 4.1a for the model indicated in Figure 4.5a, and in Table 4.1b for the model portrayed in Figure 4.5b, where [02] represents the oxygen concentration in the single-phase zone, while [02]c and [02]a are the oxygen concentrations in the core and in the annulus in the two phase zone. Aa, Ac, Ap Pa, sa, eband ec are calculated by the hydrodynamic model, while Vo2,sp, Vo^c, and Vo2,a are obtained from the devolatilization model. Kca, the mass transfer coefficient between the core and the annulus is set equal to 0.1 m/s, close to that determined by fitting of gas mixing measurements in a CFB cold unit, 0.11 m/s (Brereton et al., 1988). P/S is the primary-to-secondary air ratio, Qlst and Q2nd, the volumetric flow rates of gas in the primary and the secondary zones respectively, ae the external surface area per unit volume of char calculated by char where Dchar is the Sauter mean diameter of char particles. Assuming that the char particles are spherical and of the same diameter throughout the riser, Dchar can be calculated given the particle size distribution in the fuels fed and the fuel conversion. vchar is the char-to-sand volume ratio, which is inferred from the analysis for the weight fraction of carbon in the bed materials (Brereton et al., 1992). The combustion rate constant kc is a model parameter which is adjusted to ensure that the predicted 02 concentration in the core at height 6.4 m meets the experimental data within 0.1% 02. Figure 4.6 compares the model predictions with the experimental data for run 17-1 for Conoco Coke in the UBC pilot plant. The experimental data were obtained from three lateral positions at each height. Because of the higher char concentration in the annulus, the concentrations of oxygen are higher in the core than in the annulus. The same trend was found in the modelling. However, the predicted 02 profiles deviate substantially from the experimental data. Chapter 4 NO/N20 Model Development 71 Table 4.1: Model Equations for the Single Phase Oxygen Model: (a) for plug flow in primary zone as shown in Figure 4.5a (b) for perfect mixing in primary zone as shown in Figure 4.5 b. The equations in the primary and the secondary air zone are based on different coordinates, z = 0 represents the primary or the secondary air entry. (a) Primary zone Single phase Q X s t ^ - + At(l-e Wcharae[O2] + Vo^=0 dz B.C.: [ 02 ]c = 210, OOOppm at z = 0 Secondary zone Core Annulus Qinll^^ + A(\-ec)Vch<,ra,[01}c +Vo2,c+PaKca([O2]c-[O2]a) = 0 •210000-l + [02]sec-P/S + . B.C.: [ 02 ]c = i + p/s P P m A (1 - ea )Vcharae [02 ]a + V0l.. + PaKca ([02 ]c - [02 )a) = 0 (b) Primary zone Single phase fi,,([02]-2ioooo)+£'"(4(1-* )K^A02] + v0j,SP)dz = o Secondary zone Core Annulus + Al(l-ec)Vchara,[02]c+ Vo2.« + PJC„([02], -[02].) = 0 210000 1 + [02L - i 5 / S B.C.: [ Oo ]r =• ppm at z = 0 2 c 1 + P/S AV-£a)Vcharae[02]a + Vo>,a + PaKca([02]e -[02]a) = 0 Chapter 4 NO/N20 Model Development 72 21 18 H ^ 1 5 - 1 Core, theoretical — Annulus, theoretical • Wall, experimental • Middle.experimental A Core, experimental 12 H Core T — n — r 2 2nd air -injection "<—r 6 Height (m) 8 Figure 4.6: 0 2 Profile Predicted by the Single Phase Oxygen Model: Run 17-1, Conoco Coke. The solid distribution in the secondary air zone is simulated by a core-annulus model. A single phase model is employed for the primary air zone. For experimental details see (Brereton et al., 1992) Chapter 4 NO/N20 Model Development 73 4.3.2 Core-Annulus Primary Air Zone The 0 2 model was re-derived to assume that the core-annulus structure extends throughout the CFB riser in this sub-section. Figure 4.7 is a schematic representation of this model. The model equations are summarized in Table 4.2, where the parameters have the same definitions as in section 4.3.1. Figure 4.8 compares the modelling and experimental results for combustion run 17-1 (Conoco Coke) in the UBC pilot plant. The profiles were obtained by adjusting the combustion rate constants until the predicted 0 2 concentration in the core at height 6.4 m matches what measured from experiment within 0.1% 02. The combustion rate constant obtained is 2.1 times the value by Field et al. (1967). In general, the 0 2 profile predicted here matches the experimental result better than those in the previous section. Compared with the experimental data in the core and middle at height 1.5 m and 2.5 m, the 0 2 concentration in the core are underpredicted. It is possibly because the secondary air backmixing is not considered in this model. By assuming 100% penetration of secondary air into the core, the calculated 0 2 concentration in the core at height 4.2 m is overpredicted. As mentioned previously, the assumption (iv) in p68 depends on the configuration of secondart air port and P/S ratio. Extreme cases were found in the incineration runs of SF6 and chloroform (Poly Wang, 1994), secondary air seemed to diffuse and disperse in the annulus only, and did not penetrate into the core. Experimental results show that the 0 2 concentration at the wall sometimes exceed that in the core, indicating that the dense phase is not completely restricted to the wall. Since the 0 2 model prediction shows significant deviations from the experimental data, questions may arise as to use empirical 0 2 profile rather than the theoretical one for NO and N 20 Chapter 4 NO/N20 Model Development J L 74 2nd air Core Core "TN:— -) e--> - 7 -> <-Primary air Figure 4.7: Schematic Representation of the Core-Annulus 02 Model. Solid distribution in the primary and the secondary air zone are both simulated by a core-annulus model Table 4.2: Model Equations for the Core-Annulus 02 Model Core Primary zone B.C.: [ 0 2 ] c = 210,000ppm at z = 0 Annulus 4, (i - lA 1. + ^ „ + PaKca ([02 ]c - [02 ]a) = 0 Core 2 w ^ § k + ^ ( i - ^ ) ^ « J 0 2 ] c + '^ .« + P „ * „ ( [ C U -[O 2].) = o Secondary zone ' 210000-l + [02]sec-P/S t . B.C.: [0 2] c = i + p/s ppmatz = 0 Annulus Aa (1 - ea )Vcharae [02 ]a + Vo2, a + PaKca ([02 ]e-[O2]a) = 0 The equations in the primary and the secondary air zone are based on different coordinates, z - 0 represents the primary or the secondary air entry. Chapter 4 NO/N20 Model Development 75 Height (m) Figure 4.8: Predicted 0 2 profile by the Core-Annulus Oxygen Model: Run 17-1, Conoco Coke. The solids distribution is simulated by a core-annulus model throughout the riser, Chapter 4 NO/N20 Model Development 76 concentration in the core equals: (i) the concentration measured from the "core"; (ii) the concentration measured from the "middle"; i.e. midway between the axis and wall (iii) the arithmetic average of the concentration measured from the "core" and the "middle" (iv) the integrated average of the concentrations measured from the "core", "middle" and the "wall". However, as indicated in Figure 4.8, approached (i) to (iii) lead to an 0 2 profile with the 0 2 concentration in the upper secondary zone (6.4 m) higher than that in the lower secondary zone (4.2 m). This is impossible without air introduced between these two heights. Approach (iv) was also abandoned because it is hard to determine the weight of the concentration measured from core and middle. Another reason to abandon the empirical approach is that with an attempt to develop a comprehensive NO, /N 20 model, the 0 2 model must allow for changes in the 0 2 profiles as a result of changes of operating conditions, e.g. air staging, different excess air ratio and riser exit geometry. This cannot be achieved if an empirical 0 2 profile is used. The 0 2 model is important for the N O x and N 20 modelling work; however, the oxygen model is not a major concern of this thesis. Since the 0 2 model presented here, which is based on the core-annulus structure throughout the riser, has provided an 0 2 profile in fair agreement with the experimental one, it is employed in the N O x and N 20 modelling. In this modelling work, the 0 2 sub-program could be replaced by any updated 0 2 model in the future as input to the NOj/N^O model, as long as they are based on the core and annulus hydrodynamic structure. In summary, when the solid distribution in the primary air zone is simulated by a core-annulus model, the predicted 0 2 profile is in better agreement with the experimental results than when a single phase model is employed in the primary air zone. This indicates that the core-annulus structure extends throughout the CFBC riser, not just in the secondary air zone. The 0 2 program is provided in Appendix C. Chapter 4 NO/N20 Model Development 77 4.4 NO/N20 Model Based on the reaction scheme in Table 3.8 and the hydrodynamic model shown schematically in Figure 4.7, the concentrations of NH3, HCN, NO and N 20 throughout the CFBC riser are predicted using the equations in Table 4.3 for the core and the annulus in the primary and the secondary air zones. Also we consider the balance equations for char-N in eqn (4.19). These equations have been derived from the mass balances for NH3, HCN, NO and N 20 over an incremental height in the core and the annulus. The boundary conditions for the primary and the secondary zones are summarized in Table 4.4. ru, x = a or c, represent the reaction rates in the annulus and the core defined in Table 3.9. Vcao is the volume fraction of limestone in the solids. In eqn (4.19) Nc0 is the nitrogen-to-carbon atomic ratio in the char obtained from ultimate analysis, while Nc is the ratio in the riser. The resulting equation is: C Nck9 ([02]aAa+[02]cAc)<fc + r^ek11-([NO].A.+[NO]eAe)dk J° (4.19) = 1 Nc0 kc ([02]aAa+[02]cAc)<fc After substitution of the expressions in Table 3.8 for ria, and after some algebraic manipulations, [NH3]a, [HCN]a, [NO]a, [N20]a can be expressed as functions of [NH3]e, [HCN]e, [NO]c, [N20]c (Appendix D). They are then substituted into the differential equations to eliminate [NH3 ]a, [HCN]a, [NO]a, [N20]a. The NO/N20 model is solved by a FORTRAN program (Appendix C). Figure 4.9 gives a schematic representation of the solution algorithm. It starts with guessed values for the reaction rate constants kg, k n and k13, which are believed to depend strongly on fuel species and should be determined specifically. An inner loop is included to calculate the N to C atomic ratio in the riser (Nc in eqn 4.19). The grid employed for the numerical integration is the same as employed in solving the hydrodynamic model and the 0 2 model. Chapter 4 NO/N20 Model Development 78 Table 4.3: NH3, HCN, NO and N 20 Mass Balance Equations used in the NO/N20 Model. N H 3 P.K„([NH3]e-[NH,].) + A.e,(r3Q+r4a+lrs.) + Aa(l-sa) VCaOrX0a + VA/H,.* = 0 (4.19) Annulus HCN Pa Kca ([HCN]C - [HCN]a ) + Aasa {r6a +rla +rSa)- VHCS.. = 0 (4.20) NO P. K. WO], - [NO] J + A. c,(-/•,„ -r 4 „ + r,a - r7.) + A.(1 - s.)Kcio (-r l 0. + r„ „ ) + 4 , 0 - 0 f ' . w ( ' - 9 . + ' - | i . + ' - i 2 . ) = 0 (4.21) N 2 0 ^ ( W I . - [ J V . O ] . ( r „ + -0.5r„ -0.5r„) + 4,(1 -ea) r„. + A (i- O ^ (-'•„» + ' • . 3 . ) = 0 (4.22) -N H 3 Qd{N!Li]'~p.K~ (lNHy],-[NH2l)-Ac ec (.r„+ri.+^ru)-A, (1-*,) K « r , f c aZ 3 - V«H,.« = 0 (4.23) Core HCN Q^p- ~ P° K~ (WON]. - [HCN]. )-Acsc (rfc + r7. + r„ ) - VHCN* = 0 (4.24) NO - 4 , ( 1 - 0 ^ ( - » • » . + ' , M . ) - 4 , 0 - O ^ ( ' V . + ' - u . + » • « . ) = 0 (4.25) N 2 0 Qd[Nf]c-P. K,([N20]a - [N20)c)~ Ac ec(rle + r2c - 0.5^ - 0.5r&) «z - ^ ( 1 - ^ ) ^ c o 0 ^ - ^ ( l - s . ) Vchar ( -r I l e + r 1 3 c ) = 0 (4.26) Chapter 4 NO/N20 Model Development 79 Table 4.4 Boundary Conditions for the NO/N20 Model. The subscript "sec" denotes the level of the secondary air port. Primary zone [NH3]e=0, [HCN]c=0, [NO]e=0, [N20]c = 0 Secondary zone [NH3]C = [ A ^ U y f ^ , [HCN]e = [HCN]c^j?^ [NO]c = [NO]c_sec [N20]c = [N20]CtSec Input data are imported from the output of the hydrodynamics model, the devolatization model and the 0 2 model, as summarized in Table 4.5. Note that Aa, Ac, sa, ec, Pa, Vm3,a, VNH3 ,C, VHCN ,a, VHCN ,c, [02 ]a, [02 ]c are the values at the nodal points of the grid. The riser geometry includes the riser height and the height of the secondary air port; the input operating conditions included the temperature, the superficial velocity of the primary air and the primary-to-secondary air ratio. The fuel composition includes the proximate and ultimate analysis of the fuel. The calculations are repeated by varying kg, k u and k 1 3 until the difference between the predicted [NO] and [N20] emissions and whatever the experimental data the NO/N2Omodelling is referred to are both less than 1 ppm. The NO/N20 model program is compatible with any hydrodynamic, devolatization and 0 2 model, as long as they are based on the core-annulus structure. Combustion runs in the UBC pilot plant are presented in Chapter 5 and compared with model predictions in Chapter 6. Chapter 4 NO/N20 Model Development 80 k „ , k r 2, k13 given N c guessed Grid set up Solve model equations| for 1 st and 2nd zones NH 3 , HCN \NO, N ? Q profiles \ ^ Input data \ ^ « — 1. hydrodynamic information 2. devolatilization information 3. oxygen profile 4. fuel composition 5. riser geometry Figure 4.9: Solution Algorithm for the NO/N20 Model Chapter 4 NO/N20 Model Development Table 4.5: Summary of the Input Data for the NO/N20 Model 81 Input data Source riser geometry operating conditions A, A, Hydrodynamics Model B„ P. fuel composition V NH} ,a VNH3 . c Devolatization Model V HCN , a VHCN ,c combustion rate constant 0 2 Model [02]a [ A L Chapter 4 NO/N20 Model Development 82 4.5 Summary A comprehensive NO/N20 model is established in this chapter as indicated in Figure 4.10. The hydrodynamic model calculates y, ea, ec, Afl and Ac profiles throughout the riser. These data are provided as input to the devolatilization model, the 0 2 model and the NO/N20 model to calculate the local release rates of NH3, HCN and the 0 2 concentration profiles, required by the NO/N20 model to calculate the profiles of NO and N 20 in a CFBC riser. X .'1. NH 3 , HCN 2. Fuel Analysis.' .t. .'1.CH4, H 2 2. Fuel Analysis.' 0 2 Model NO/N 2 0 Model 0 2 Profile NO & N 2 0 Profiles / Figure 4.10: Schematic Representation of the Communication between NO/N20 Model and Sub-Models. Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant During the multifuel combustion runs in the UBC CFBC pilot plant facility, NO x and N 20 emissions have been studied under a variety of conditions for Conoco coke, Poplar River lignite, CANMET pitch and Mt. Klappan anthracite. Both axial and radial concentration profiles have been obtained for some runs, showing the general trends in NOx and N 20 formation and reduction in the CFB riser. These profiles provide a database to validate the CFBC NO/N20 model developed in Chapter 4. This chapter introduces the basic components of the UBC CFBC facility, including the combustor, the gas sampling system and the analysis instrumentation, then discusses the experimental results. Results are presented on both: (i) the influence of the major operating parameters on NOx and N 20 emissions; (ii) NOx and N 20 concentration profiles. 5.1 UBC CFBC Pilot Plant Facility The UBC CFBC pilot plant facility is a 110 kW unit operated at atmospheric pressure. Figure 5.1 is a schematic of the major components. Full details are given by Grace and Lim (1987) and Brereton et al. (1992). The riser reactor is a refractory-lined chamber of 152 mm square cross-section and overall height 7.3 m, composed of five flanged sections as shown in Figure 5.2. To provide a high velocity region for particles, which prevents agglomeration and sintering problems in combustion, the bottom section is tapered on the inside from a 51 mm x 152 mm cross-section at the base to 152mm x 152 mm over a distance of 1.22 m. The primary air is introduced into the riser through a distributor, 83 Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant 84 secondary air I primary air 1. Reactor; 2. Windbox; 3. Primary cyclone; 4. Secondary cyclone; 5. Recycle hopper; 6. Standpipe; 7. Eductor; 8. Secondary air preheater; 9. Flue gas cooler; 10. Baghouse; 11. Induced draught fan; 12. Fuel hopper; 13. Sorber hopper; 14. Rotary values; 15. Secondary air ports; 16. Membrane wall; 17. Pneumatic feed line; 18. External burner; 19. Ventilation; 20. Calometric section Figure 5.1: Simplified Schematic Diagram of the UBC CFBC Facility er 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant Figure 5.2: View of the Refractory-Lined Reactor Shaft Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant 86 which has twenty 9.5 mm diameter drilled at 30 and 50 degrees to the horizontal axis. Located at the center of the base is a 38 mm tube which allows easy removal of small agglomerates of sintered solids, and other oversize material. Secondary air can be fed through secondary air ports at two levels, 0.9m and 3.4 m above the distributor. Solid fuels are fed pneumatically at a height of 540 mm above the distributor. The solids are entrained up to the top of the riser, captured by the primary cyclone and fall through a storage hopper into a 102 mm ID standpipe, then return to the riser via a non-mechanical solids flow L-valve. Fine particles which are not captured by the primary cyclone are separated from the flue gas by a secondary cyclone, then recycled back to the bottom of riser via an eductor by pneumatic transport air. This air is included with the secondary air when the lower level secondary air port is employed, or as primary air otherwise. Dusty gas leaving the secondary cyclone then passes through a water-cooled heat exchanger section to a baghouse. The riser is equipped with thermocouples and pressure taps at 610 mm maximum intervals to monitor the bulk temperature and suspension density profiles. The temperature and suspension density profiles within the riser and the flue gas composition are recorded every 5 minutes by an AT&T 6300 computer. Gas samples are withdrawn from four levels from the combustor 1.5, 2.7, 4.2 and 6.4 m above the air distributor, and in few runs at 0.6 m. At each level, the gas samples are taken from three lateral points shown schematically in Figure 5.3, denoted as "wall", "middle" and "center" respectively. The sampled gas is directed via a sampling train to gas analyzers, as indicated in Figure 5.4, where the solids in the sampled gas are removed by a sintered stainless steel filter, then the moisture is removed by a water-cooled heat exchanger and a packed Mg(C104)2 drier in series. Continuous on-line analyzers are used to provide 02, CO, C02, NO, N0 2 and S0 2 concentration measurement. An FTIR Figure 5.3: Lateral Gas Sampling Positions (top view): 1. Fuel Hopper; 2. Limestone Hopper; 3. Riser; 4. Primary Cyclone; 5. Secondary Cyclone (Zhao, 1992) Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant 88 Figure 5.4: Reactor Gas Sampling Train: 1. Combustor; 2. Sampling Port; 3. T-fitting; 4. Ball Valve; 5. Gas Probe; 6. Flexible Stainless Steel; 7. Stainless Steel Filter 8. Heat Exchanger; 9. Dryer; 10. Three-Way Valve (Zhao, 1992). Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant 89 analyzer is used for N 20 analysis. High accuracy can be achieved at wave numbers between 1250 and 1285 cm-1, where there is minimum interference from other gases. Due to the interference of CH4, it was usually impossible to determine the concentration of N 20 correctly in the bottom of the riser or, on occasion, in the wall region or the volatile plume using FTIR. For example, Figure 5.5 shows that the N 20 concentration is only available from the top sampling level for Run 22-3 with CANMET pitch. 5.2 NO, and NzO Emissions Results: Table 5.1 summarizes the emission results for NOx and N 20 for Conoco coke, Poplar River lignite, CANMET pitch and Mt. Klappan anthracite in the UBC CFBC pilot plant in terms of the fuel nitrogen conversions to NOx and N20, respectively. Table 5.2 contains the proximate and ultimate analyses for these fuels. The fuel-N conversions to N0 X were calculated by assuming that NOx consists of NO exclusively. Note that because different fuel feed rates and gas throughputs were employed in these runs, the relative order in the fuel nitrogen conversions to NOx and N 20 do not exactly follow the same order as the NOx and N 20 emissions. For example, Run 26-4 shows higher fuel-N-to-NOx conversion than Run 26-5, while the order reverses in emissions (110 ppm vs. 115ppm). The higher conversion in Run 26-4 results from the lower fuel feed rate and higher gas throughput. In Table 5-1, the Conoco coke runs show lower fuel-N-to-NOx conversion but higher fuel-N-to-N20 conversion than the others. However, these trends are not conclusive because of the wide range of operating parameters involved. These factors are discussed in depth below. The operating parameters significantly affecting the NOx and N 20 emissions include the fuel properties, temperatures, excess air (0 2%), height of Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant 90 CL CL 63 54 45 O ^ 36 | 27 co I 18 o c o O 9 -• — A — • • • • • I ' I 1 • • • . • • 1 1 ,T , T , , T, , ,T • • • T 0 3 4 Height (m) 8 \t Figure 5.5: N 2 0 Concentrations Measured by FTTR for Run 22-3, CANMET Pitch. Symbols are the experimental data measured from: •: Wall, •: Middle, A: Axis, *: Flue gas. The arrows indicate where the other sampling probes are located. Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant 91 Table 5.1 Summary of N0 X and N 20 Emission Results Detailed information for each run is available in Brereton et al. (1992) " ** " denotes the runs provided with profile measurement Fuel Run Temp U P/S Sec Air Density Ca/S o 2 Fuel-N Conversion to (NOv, N,0) °C m/s Level kg/m3 % (%, %) Mt Klappan Anthracite 14-r* 14-2 19- 1** 881 890 873 8.4 9.0 8.7 2.0 2.0 1.1 upper lower upper 112 89 123 0 0 0 3.8 4.6 5.4 ( 5.86, 7.13) (11.40, 2.91) ( 9.64, 4.00) Conoco 17-1** 17-2** 17-3** 882 848 861 7.0 6.9 7.1 1.6 1.6 1.6 upper upper upper 119 125 123 0 2.5 3.2 3.8 4.3 4.3 ( 8:96, 12.10) ( 1.97, 12.98) ( 2.26, 10.57) Delayed Coke 25-1 25-2 25-3 25-4 25-5 25-6 849 859 894 847 877 856 6.4 6.4 6.7 5.3 7.3 7.0 2.1 2.1 2.1 1.9 2.0 19 upper upper upper upper upper upper 105 98 96 100 145 106 0 2.5 2.8 2.4 2.4 1.1 4.9 5.2 5.0 3.8 3.0 4.2 ( 5.90, 16.41) ( 3.24, 14.17) ( 3.78, 10.07) ( 2.34, 11.90) ( 2.44, 10.42) ( 2.03, 12.96) CANMET Pitch 22-1 22-2 22-3** 909 839 841 9.6 8.9 9.0 2.5 2.5 2.5 upper upper upper 84 115 103 0 0 2.6 4.3 7.3 7.5 (14.85, 4.84) (17.27, 13.10) (18.10, 9.56) Poplar River Lignite 26-r* 26-2** 26-3 26-4 26-5 864 876 852 820 855 8.2 8.3 8.0 7.9 6.2 2.0 0.84 2.0 2.0 2.1 upper upper upper upper upper 173 114 74 135 128 0 0 0 0 0 4.3 3.4 5.9 2.7 2.1 (12.36, 6.96) (12.79, 5.22) (12.28, 8.41) ( 7; 16, 9.11) ( 5.72, 3.68) Temperatures indicated aboveare the arithmatic avereage of the riser temperatures. 0 2 represents the excess air, expressed by oxygen concentration in the flue gas. Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant 92 Table 5.2 Proximate and Ultimate Analyses of Solid Fuels Proximate Analysis wt % Poplar River Mt Klappan Conoco delayed CANMET Lignite Anthracite coke pitch Volatile Matter 26.6 6.5 10.0 67.7 Fixed Carbon 25.8 71.0 84.0 30.2 Ash 16.2 15.5 0.8 2.0 Moisture 31.4 7.0 5.2 0.1 Ultimate Analysis wt % (dry base) Poplar River Mt Klappan Conoco delayed CANMET Lignite Anthracite coke pitch Carbon 53.7 76.8 87.3 86.2 Hydrogen 4.1 2.5 3.8 7.1 Nitrogen 0.9 0.9 1.9 1.1 Chlorine 0.0 0.0 0.0 NA Sulphur 1.1 0.5 4.9 2.8 Oxygen (by 16.6 2.6 1.1 1.1 difference) Ash 23.6 16.7 0.9 18 Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant 93secondary air port and limestone addition. Little effect of the superficial gas velocity, suspension density and the P/S ratio have been found in the UBC pilot plant combustion runs (Zhao, 1992). Single factor experiments had been planned on the UBC pilot plant runs to explore the effects of these important operating parameters on NOx and N 20 emissions. However, given that there are so many parameters involved in the CFBC operation, the change in one parameters would inavoidably involve the change of others. For example, decreasing fuel feed rate to increase the excess air causes the bed temperature to decrease, even when heat removal by the heat transfer surface has been reduced to a minimum. The following sections illustrate the effects of fuel type, temperature, excess air and limestone on the NOx and N 20 emissions from the UBC pilot plant. Comparisons are made pairwise between runs where only one parameter was distinctly different, with small variations in other parameters compared to the fluctuations in a "steady state", e.g. temperature variation within ± 10°C and excess air varying within ± 0.5% (Zhao 1992). 5.2.1 Fuel Types Figure 5.6 shows the fuel-N conversions to NOx and N 20 for Run 14-1 (Mt Klappan anthracite), Run 17-1 (Conoco coke) and Run 26-2 (Poplar River lignite) as functions of volatile content and VN number (Grace et al, 1989). No limestones were added; the secondary air was introduced from the upper secondary air port; the temperature differences were within 6°C and the excess air, expressed by the 02 concentrations in the flue gas, were less than 0.4%. Fuel-N-to-NOx conversion increases with both the VN number and the volatile content. Fuel-N-to-N20 conversion is not simply correlated to either of these. When VN or the volatiles content increases, the Fuel-N-to-N20 conversion first increases then Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant 94 O cr c o o > c o O (D 3 0.22 0.20 0.18 0.16 0.14 0.12 0.10 0.08 0.06 0.04 0.0 — • - - N 0 X vs V N % -- • - - N 2 0 vs V N % — • - - N 0 X vs Volatiles Content -- o- - N 2 0 vs Volatiles Content Run 26-2 J_ _L 0.2 0.4 0.6 V N % or Volatiles Content 0.8 Figure 5.6 Fuel-N Conversions to N O x and N 2 0.vs " V N " and Volatile Content. The data are from Run 14-1, Mt Klappan anthracite, Run 17-1, Conoco coke and Run 26-2, Poplar River lignite. For details, see Brereton et al. (1992) Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant 95 decreases. The correlation between fuel-N conversions (to either NOx or N20) and the VN is similar to that between the conversions and the volatiles content. 5.2.2 Temperature Effect The arrows in Figure 5.7 connect the runs in which the difference in excess air (approximated by flue gas oxygen) was less than 0.8%. For each pair, the secondary air was introduced from the same secondary air port, similar Ca/S ratios were employed, and more than a 10°C decrease in the average temperatures throughout the riser occurred between runs connected by the arrows showing the temperature decrease. In general, the fuel-N conversions to N 20 are increased while the conversions to NO x are decreased along the arrows, which means that decreasing the temperature reduces NOx emissions but causes an increase in N20. There is one exception: when the operating condition was switched from Run 26-5 (877°C) to 26-4 (847°C), the fuel-N conversion to NOx and N 20 both increased. The reason is not clear yet. Note also that less NOx reduction was achieved for the Conoco coke than for the other fuels. 5.2.3 Excess Air Effect Table 5.1 shows that Runs 17-3 and 25-2 involved similar temperatures (861 and 859 °C), but distinctly different excess air ( 4.3% and 5.2% 02 in the flue gas). With a raised excess air, the fuel-N conversions to NOx and N 20 were both higher for Run 17-3 (2.26%, 10.57%) than for Run 25-2 (3.24%, 14.17%). There are examples where the excess air overrides the temperature effect. For example, although it has been reported that NO emissions increase with increased temperature, the temperature change from Run 25-2 (T = 859°C, 5.2% 02) to Run 25-5 (T = 877°C, 3.0% 02) was overshadowed by the major change in excess air involved. As regards N20, it is known that higher temperature Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant 96 0.16 ° c Z o 0.12 i 3 O c o \n CD > c o o 0.08 0.04 0.00 25-2 © Conoco Coke • • Lignite 1 7 - 2 . \ < A Anthracite - 2 5 - 4 • 17-3 * I • • 2 5 - 5 25-3 26-4 • • 14-1 A • • • • • • • 19-1 V A ^ " 26-2 • 26-5 X A - 14-2 I 1 l 1 0.00 0.04 0.08 0.12 0.16 Conversion of Fuel-N to NCL 0.20 Figure 5.7: Effects of Temperature on the Fuel-N conversion to NOx and N20. The data are taken from Table 5.1. Arrows denotes change where temperature (> 10°C) was decreased significantly with other conditions held relatively constant. Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant 97 correspond to lower N20, nevertheless, the fuel-N conversion to N 20 from Run 25-2 (T = 859°C, 5.2% 02) is larger than that from Run 25-4 (T =847°C, 3.8% 02) because of the higher excess air in Run 25-2. 5.2.4 Effect of Limestone Addition Limestone addition effects have been studied for the Conoco coke and the CANMET pitch runs, shown in Figure 5.8 with Ca/S ratios indicated for each run. The temperature difference between each pair of conditions connected by the arrows was less than 10°C and the difference in flue gas oxygen (1.1%) is close to to the criterion defined in section 5.2. Runs 22-2 and 22-3 for the CANMET pitch were operated at similar temperatures, 839 and 841°C, respectively. Limestone was added in Run 22-3, leading to higher fuel-N-to-NOx and lower fuel-N-to-N20 conversion. Tracing the arrows from Run 25-1 to Run 25-2, Run 25-4 and Run 25-6, respectively, shows that limestone reduced both the fiiel-N-to-NOx and fuel-N-to-N20 conversions for the Conoco coke. The conventional trend reported for CFB combustion is that limestone addition increases NOx by catalyzing NH 3 oxidation (Reaction 10 in Table 3.9). Since limestone can also catalyze NO reduction by CO at CFB temperatures (Reaction 14 in Table 3.9), the net effect depends on the competition between these two reactions. It seems that the catalytic reduction reactions dominate for Conoco coke combustion. In summary, limestone addition reduces N 20 emissions for Conoco coke and the CANMET pitch combustion. Substantial NOx reduction was also achieved in the Conoco runs. However, the reduction of N 20 emissions caused by adding limestone in the CANMET pitch led to increased NOx emissions. Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant 98 0.20 0.16 0.12 0.08 C D c 0.04 o O 0.00 • Conoco Coke 25-1, Ca/S = 0 J ^ 2 5 - 2 , Ca/S = 2.5 T CANMET Pitch 22-2, Ca/S = 0 _ \r 25-6, Ca/S = 1.1 • \ 25-4, Ca/S = 2.4 \ T -22-3, Ca/S = 2.6 i . l . l I 0.00 0.04 0.08 0.12 0.16 Conversion of Fuel-N to N O v 0.20 Figure 5.8: Effect of Limestone Addition on the Fuel-N conversion to NO x and N20. Data are taken from Table 5.1 for CANMET pitch and Conoco coke combustion. Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant 99 5.3 Gas Concentration Profiles 02, NOx and N 20 concentration profiles provide insight into the fuel distributions, the gas mixing and mechanism of NOx and N 20 formation and reduction in a CFBC riser. Both axial and lateral 02, NOx and N 20 concentration profiles were measured for Run 14-1, 19-1 (Mt Klappan Anthracite), Run 17-1, 17-2 and 17-3 (Conoco delayed coke), Run 22-3 (CANMET pitch) and Run 26-1 and 26-2 (Poplar River lignite). To discuss the general features, some profiles are presented below, while others are plotted in Appendix D. 5.3.1 0 2 Concentration Profiles Figure 5.9 shows 0 2 concentration profiles for Run 17-2 (Conoco coke). To distinguish axial profiles measured from different lateral positions, the data points are connected by smooth curves, which do not represent the modeling results. Lateral gradients are prevalent throughout the riser in Figure 5.9, as well as in the 0 2 concentration profiles given in Appendix D, where the 0 2 concentrations descend from the core to the wall. This indicates that, like the inert solids, the chars exhibit a core-annulus concentration distribution in the riser. The non-uniform cross-sectional char distribution combines with the poor lateral gas mixing to give strong lateral gradients. It has been shown (Senior, 1992; Yang, 1992) that suspended solids in the CFB riser suppress gas turbulence, causing the lateral gas mixing to be reduced thereby. Cross-overs occur in the profiles given in Appendix D, e.g. in Figure D.l for heights of 2.7 and 6.4 m and Figure D.3 at 1.5 m. These imply that the dense phase may not always be localized at the wall. The 0 2 concentrations measured from the lowest sampling probe (0.6 m) tend to be lower than those measured from the next (1.5m) level despite the fact that no air was supplied inbetween. As discussed in section 4.3.2, Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant 100 11.9 0 1 2 3 4 5 6 7 8 Height (m) Figure 5.9: 0 2 Concentration Profiles for Conoco Coke Combustion: Run 17-2. T = 1121 °K, U g = 6.9 m/s, P/S = 1.6 with secondary air introduced 3.4 m above the distributor. Symbols are the experimental data measured from: • : Wall, • : Middle, A: Axis, *: Flue gas. Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant 101 secondary air affects the 02 concentrations below the 2nd air port (1.5 m and 2.7 m) due to backmixing. However, the low 02 concentrations at the bottom (0.6 m) is anomalous. The anomalously low 02 concentration at the bottom do not occur for CANMET pitch (Figure 5.10), indicating that the low 02 concentrations for the other runs may be an artifact. Because of the large fraction of dense phase at the bottom of the riser (Figure 4.3), not necessarily localized at the wall, it is very likely that the dilute phase is not within reach of the gas sampling probe at the bottom. In the combustion of solid fuels where the distribution of fuels follow the distribution of inert solids, missing the dilute phase gives consistently low 02 concentrations at the bottom. Note that CANMET pitch was atomized as it was introduced into the combustor; the carbon content is low compared to that in Conoco coke combustion (0.1 W% vs. 3.8 W%) (Brereton et al, 1992). Thus burning CANMET pitch is like burning a liquified fuel. The fuels therefore follow the distribution of inert solids less readily and the 02 concentrations can be high in the dense phase. As shown in Figure 5.10, 02 concentrations at the wall exceed those in the core and the middle throughout the riser. Finally, it should be particularly noted that in Runs 17-3, 19-1 and 26-2, the 02 concentrations in the flue gas were higher than those measured by the top gas sampling probe regardless of the lateral position, indicating that there might be a leak between the top sampling level and the flue gas take-off for these runs. 5.3.2 NOx Concentration Profiles Figure 5.11 and Figure 5.12 show NOx profiles for combustion of Conoco coke in Runs 17-1 and Run 17-3, with and without limestone addition, respectively. The lateral NOx gradients are less than that of 02. The smaller lateral NOx gradient indicates that NO x is Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant 102 Height (m) Figure 5.10: 0 2 Concentration Profiles for CANMET Pitch Combustion: Run 22-3. T = 1114°K, U g = 9.0 m/s, P/S = 2.5 with secondary air introduced 3.4 m above the distributor. Symbols are the experimental data measured from: •: Wall, •: Middle, A: Axis, *: Flue gas. Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant 103 Figure 5.11: N0 X concentration Profiles for Conoco Coke Combustion: Run 17-1. T = 1155°K, U g = 7.0 m/s, P/S = 1.6, Ca/S = 0, with secondary air introduced 3.4 m above the distributor. Symbols are the experimental data measured from: •: Wall, •: Middle, •: Axis, *: Flue gas. Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant 104Figure 5.12: N 0 X Concentration Profiles for Conoco Coke Combustion: Run 17-3. T = 1134°K, U g = 7.1 m/s, P/S = 1.6 , Ca/S = 3.2, with secondary air introduced 3.4 m above the distributor. Symbols are the experimental data measured from: • : Wall, • : Middle, • : Axis, *: Flue gas. Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant 105 distributed more uniformly in a cross-section. This observation suggests that the mass transfer coefficients between the core and the annulus for 02 and NOx be distinct in modelling, with larger value for NOx and smaller value for 02. High NOx concentrations have been found at the bottom of the riser where NOx formation reactions dominate. The profiles decline higher up the riser because of the 02 concentration decay and because the reduction of NOx by char and limestones becomes important. A comparison of Figures 5.11 and 5.12 shows that the NOx concentration declines much more quickly with height in the presence of limestone. For Run 17-3 (Figure 5.12), which gave an anomalously high 02 concentration in the flue gas, the NOx concentration in the flue gas was higher than those measured from any lateral position of the top sampling level. NO x concentration profiles have been reported by other researchers for CFBC risers of different sizes. The findings are similar to ours. Ishikuzu (1988) measured axial NOx concentration profiles from aO. lmID x 5 m riser. High NOx concentrations were found at the bottom, which after reaching a maximum, declined higher up the reactor. Moritomi and Suzuki (1992) measured NOx in the same apparatus and found the same trend. Moritomi and Suzuki (1992) also replaced the silica sand by limestone and observed a monotonic increase in NOx. Berge (1988) measured the axial NOx concentration profiles from a 0.65 x 0.65 x 7 m 3 riser and found the same trend as above. The lateral positions of the gas sampling probe were not reported for any of these cases. 5.3.3 N zO Concentration Profiles Figure 5.13 shows N20 concentration profiles for Run 17-1. The concentrations are not available for certain locations due to the C H 4 interference. The lateral N20 gradients are larger than those for NOx, but are less than for 02. Both the lateral motion of char and Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant 106 Figure 5.13: N 20 Concentration Profiles for Conoco Coke Combustion: Run 17-1. T = 1155°K, U g = 7.0 m/s, P/S = 1.6, Ca/S = 0, secondary air introduced via the secondary air port 3.4 m above the distributor. Symbols are the experimental data measured from: •: Wall, •: Middle, A: Axis, *: Flue gas. Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant 107 the ability of char to decompose N O x and N 2 0 affects the lateral N O x and N 2 0 gradients. Chars usually exhibit a stronger ability to react with N 2 0 than N O x (Table 3.7). The core-annulus char concentration distribution and the strong ability of char to decompose N 2 0 probably cause a larger lateral N 2 0 gradient. The axial N 2 0 profiles generally follow the same trend as observed for the N O x profiles along the riser, but with much gentler slopes. The increase of N 2 0 concentration with height is less sharp than for N O x at the bottom, and it decreases gradually with height. In Figure 5.13 and all of the N 2 0 profiles in Appendix D, the N 2 0 concentrations in the flue gas are consistently higher than those measured at the top gas sampling level. This is presumably because of the temperature drop of gas after leaving the riser exit. Table 5.3 compares the average bed temperature and the secondary cyclone temperature for the runs covered in Table 5.1. The axial N 2 0 concentration profiles measured by other researchers are summarized in Table 5.3. Hulgaard (1991) measured profiles half-way between wall and axis of their riser. However, the radial sampling positions in the other cases were not specified. There is one thing in common for these profiles reported. The N 2 0 concentration increase at the bottom of the riser, however, change gradually with height higher up the riser. 5.4 Summary The N O x and N 2 0 emissions observed from combustion runs in the U B C pilot plants show that the fuel-N conversion to N O x increases with V N and volatile content. However, no simple correlation exists between fuel-N-to-N20 conversion and V N or volatile content. Decreasing the temperature decreases N O x but causes an increase in N 2 0 emissions. Decreased excess air decreases the conversion of fuel-N to both N O x and N 2 0 . Limestone Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant 108 Table 5.3: Comparison of Average Bed Temperature and Secondary Cyclone Temperature. For details, see Brereton et al. (1992) Fuels Run TBed(K°) T CK°} Secondary cyclone V / 17-1 17-1 882 839 17-2 17-2 848 843 17-3 17-3 861 858 14-1 14-1 881 817 19-1 19-1 873 847 22-3 22-3 841 767 26-1 26-1 864 851 26-2 26-2 876 844 Table 5.4: Features of N 20 Concentration Profiles in CFBC Risers of Different Scales Combustor Scale General Features Authors 1.70 * 1.70 * 12.5 m Gradual increase with height Amand et al, 1991 N 20 0.1m ID x 5 m Same as above Moritomi and Suzuki, 1992 A 2MWt combustor with riser height 10 m, cross section area undocumented Sharp increase at the bottom followed by a small decrease Hulgaard, 1991 Chapter 5 Multifuel Combustion Runs in the UBC CFBC Pilot Plant 109 addition causes a decrease in N 2 0 emissions, but the effect on N O x depends on the distribution of fuel-N between char and volatiles. Because of the lateral solids distribution and poor lateral gas mixing, lateral 0 2 , N O x and N 2 0 concentration gradients exist in the CFBC riser. These gradient are diminished by lateral char motion. The gradient is usually greater for 0 2 , followed by N 2 0 and then N O x . Lateral N 2 0 gradients tend to be larger than for N O x because N 2 0 is decomposed more readily by char. The 0 2 concentrations measured by the lowest gas sampling probe are anomalously low for most of the combustion runs except for C A N M E T pitch. This probably arises because of an artifact. The axial profiles of the N O x concentration show a sharp increase at the bottom of the riser and then decline higher up the riser. N O x concentrations decrease more sharply in the presence of limestone. N 2 0 concentration profiles follow the same trend but tends to be more gradual. Chapter 6 NO and N 2 0 Modelling Results This chapter presents the NO and N 2 0 modelling results for Conoco coke combustion (Run 17-1). The mass transfer coefficient (Kca) for NO and N 2 0 between the core and the annulus and some of the reaction rate constants (ku, k 1 2 and k13) in the proposed kinetic scheme (Table 3.9) are determined by fitting the predicted NO and N2O profiles to the experimental ones. Sensitivity analysis is performed on each reaction rate constant involved in the kinetic scheme. Finally, the effects of excess air, the relative fuel nitrogen distribution in char and volatile, limestone addition, air staging - P/S and the height of secondary air port and the riser exit geometry on the NO and N 2 0 emissions predicted by the model are presented. 6.1 Reaction Rate Constants k u , k 1 2 and k 1 3 and Mass Transfer Coefficient (K c a) Reaction rate constants are important parameters in the NO and N 2 0 model. The reaction rate constants k 1 2 and k13 in Table 3.9 differ among fuels (deSoete, 1990); k n is also believed to strongly depend on the specific fuel. Reliable values should be determined from separate kinetic experiments. Because these were not available for the Conoco coke, these constants were treated as fitting parameters in this model. The mass transfer coefficient is another fitting parameter in this model. As discussed earlier, the lateral NO and N 2 0 gradients are smaller than the 0 2 gradients, indicating that the mass transfer coefficient, Kca, for NO and N 2 0 may differ from that for 0 2 taken as 0.1 m/s. The values for k n, k12, k13 and Kca are determined as those which match the predicted NO and N 2 0 concentrations within ±1 ppm in the core at a height 6.4 m to the experimental data. 11.0 Chapter 6 NO and N20 Modelling Results 111 Two values are for K c a are explored, 0.1 m/s and 1 m/s, respectively, the values for ku, k 1 2 and k13 satisfying the above criterion are summarized in Table 6.1 These constants were determined by trial and error until the fitting criterion was met. Figure 6.1 shows the corresponding NO and N 20 calculated profiles. In general, K c o =1.0 m/s gives better prediction for both NO and N 20 profiles (Figure 6.1c and 6.Id), where the differences in NO and N 20 concentrations between the core and the annulus are small, comparable to those observed from experiments. Furthermore, the N 20 profile in Figure 6. Id shows much better agreement with the experimental data than in Figure 6.1b, indicating that the larger K c a is preferred for NO and N 20 modelling. Although the larger Kca gives more accurate predictions, the predicted NO profile (Figure 6.1c) still shows large deviations from the experimental data at height 1.5 m and 2.7 m. This is possibly because, without considering backmixing of the secondary air in the 0 2 sub-model, the 0 2 concentrations below the secondary air port have already been underpredicted (Figure 4.8). Values of kn, k12 and k13 corresponding to these two different values of K c a are given in Table 6.1. They do not depend much on Kca. k12 and k13 are close to those for "Eschweiler Char" (Table 3.7). It is interesting to find that k n is more than two hundred times greater than the NO formation constants kg. This indicates that NO strongly competes with O2 for char-N. The fraction of char-N consumed by NO (%N N O ) and O2 (xN02) in the CFBC riser are calculated by XN.NO ~ f\u ([NO]aAa+[NO]c)Acrfz J " [k9 ([0 2] aA a+[0 2] cA c) + k„ ([NO]aAa+[NO]cAc)]<fe (6.1) Chapter 6 NO and N20 Modelling Results 112 (a) 770 660 — 550 — :E 440 C L — C L 330 o z 220 — 110 (b) 2nd air f injection I I • I l I I 1 1 I 1 I 1 I 2 3 4 5 Height (m) I 1 I " I 6 7 8 (c) 735 — 630 — 525 — 420 D_ _ 0_ 315 _ O z 210 — 105 — n u (d) 2nd Air A Injection '* | ' | ' I ' I ' I ' I ' I ' I 1 2 3 4 5 6 7 8 Height (m) 2nd air A injection M l ' I 1 I ' M l 2 3 4 5 6 7 Height (m) 1 8 Q. fx z 315 - i 270 H 225 180 -135 -90 -45 2nd Air f Injection 0 I • M I 1 M l ' M M I ' 0 1 2 3 4 5 6 7 Height (m) Figure 6.1 Predicted NO and N 2 0 Concentrations Profiles for Conoco Coke Combustion: Run 17-1. (a) NO, K c a = 0.1 m/s; (b) N 2 0 , K c o = 0.1 m/s; (c). NO, K c a = 1.0 m/s; (d) N 2 0 , K c a =1.0 m/s. Solid lines are theoretical concentration in core, dashed lines are theoretical concentration in annulus. Discrete symbols are the experimental data measured from: • : Wall, • : Middle, • : Axisln the experiments, the secondary air was introduced 3.4 m above the distributor. T = 1155°K, U g = 7.0 m/s, G s = 65 kg/m2 s, P/S = 1.6. The 0 2 concentration at the riser exit, 4.7%, is based on the prediction by the 0 2 sub-model. Chapter 6 NO and N20 Modelling Results 113 Table 6.1: Predicted Reaction rate Constants ku, k 1 2 and k13 for Conoco Coke at 1155°K. Note that kg is the char combustion rate constant determined by the 0 2 sub-model (section 4.3) Values at 1155°K Note 0.1 (m/s) k n (ms-1) 31.3 299 times K k,, (m3 kg-1 sr1) 0.4 Eschweiler char*: 0.4 k^imng-1 s-1) 14.4 Eschweiler char*: 8.3 1.0 (m/s) k,, (m s-1) 27.0 258 times kg k„ (m3 kg-1 s-1) 0.4 Eschweiler char*: 0.4 k^Wkg-1 s-1) 14.6 Eschweiler char*: 8.3 *deSoete (1990) I k 9 ( [ 0 2 ] a A a + [ 0 2 ] c A c ) ^ %N,O2 = 771 " ( 6 - 2 > J 0 [k9 ( [ 0 2 ] a A a + [ 0 2 ] c A c ) + k„ ([NO] aA a+[NO] cA c)]<fc where A is the cross-section area and subscripts a and c represent the annulus and core, respectively. With Kca equal to 0.1 m/s, the fraction of char nitrogen consumed by NO is equal to 0.47, while it is equal to 0.41 when K c a is equal to 1 (m/s). The results are consistent with Tullin et al. (1993) who found that the reaction between NO and char-N is important for N 2 0 formation at FBC temperatures. They are also in qualitative agreement with the experimental results of Amand et al. (1992). The broad NO band in Figure 6.2a which occurred for char combustion in a fixed bed disappeared when NO was present at the reactor inlet (Figure 6.2b). Much more N 2 0 was generated instead. Chapter 6 NO and N20 Modelling Results 114 (a) 2000 1000 2000 3000 TIME (sec) 100 80 E CL CL O C o 60 o O 40 o 20 2 (b) CL Q. O C o o O o O O 2000 1500 1 \ — 1000 - - -v 500 \k\ NO . 0 1 1 ' M l 1 1 1 ' 1 ' 200 160 120 80 40 E CL CL d c o o Z 0 300 600 900 1200150018002100 TIME (sec) Figure 6.2: Evolution of the Outlet Concentrations of CO, NO and N 20 for Char Combustion in a Fixed Bed Reactor (Amand et al, 1992). Reactor conditions: (a) 800°C, 1.6% 0 2 and 1000 ppm CO in the inlet; (b) 800°C, 1.6% 02, 1000 ppm CO, 500 ppm NO and 100 ppm N 20 in the inlet. Chapter 6 NO and N20 Modelling Results 115 Since the NO and N 2 0 profiles predicted based on Kca = 1.0 are more accurate, the following discussion is based on Kca =1.0 and the corresponding reaction rate constants. 6.2 Sensitivity Analysis Fifteen reactions are involved in the kinetic scheme for this modelling work. The sensivity of the model predictions to the reaction rate constants kj to k15 were investigated by setting these constants equal to zero consecutively without altering the others. The results are shown in Figure 6.3. Since NO is reduced by CO only if limestone is present, the volume fraction of limestone in solids (V C a Q) is assumed to be 0.4 for the sensitivity analysis, different from that in modelling Run 17-1, where no limestone was added. NO is highly sensitive to kg and kn, while being moderately sensitive to k5, k8, k10, k 1 2 and k14. N 2 0 is strongly affected by kg, k n and k13, while it is hardly affected by the others. Setting kg and k n the formation rate constants for NO and N 2 0 from char-N, respectively equal to zero, dramatically changes NO and N 2 0. If the heterogeneous formation path of NO involving kg were blocked, negligible NO (2.86 ppm) and N 2 0 (less than 1 ppm) would be formed. NO emissions would be very high (3310 ppm) while very little N 2 0 would be formed (1.06 ppm) without the reaction between NO and char-N (k u). The catalytic formation and reduction of NO by limestone (k10 and k14), which have opposite effects on NO emissions, are both important. The ability of char to decompose NO and N 2 0 (k12 and k13) are also important determinants of NO and N 2 0 emissions. Both NO and N 2 0 predictions are insensitive to homogeneous HCN and NH 3 oxidation (k3, k4, k6 and k7) and to homogeneous N 2 0 decomposition (ky and kj), partly because only 20% of the fuel-N is contained in the volatiles. Note that the above sensitivity test applies to Conoco coke only and is only strictly correct if the various reaction rate constants are validated. Different chars and limestones Chapter 6 NO and N20 Modelling Results 116 400 350 300 250 ^ 200 c 1 150 g 100 c o U 50 CL CL 260 v73 A A (NO*100, N 2o) O ( NO/10, N 2o) O ( NO, N 2 ° ) Base condition: (286,127) w/7 74 A ^ A AA 12 280 300 320 NO (PPM) 340 360 Figure 6.3: Sensitivity Analysis of the Reaction Rate Constants kj to k15. The operating condition is based on Run 17-1 for Conoco coke in which the secondary air was introduced 3.4 m above the distributor. T = 1155°K, U g = 7.0 m/s, Gs = 65 kg / m2 s, P/S = 1.6. However, V C a Q is assumed to be 0.4 instead of 0. The 0 2 concentration at the riser exit is 4.7% based on the prediction by the 0 2 sub-model. The values of model parameters are as follows: Kca =1.0 m/s, k n = 27 m-s-', k12 = 0.4 m3-kg-1-s-1, k13 = 14.6 m3-kg-1 -s-J. Chapter 6 NO and N20 Modelling Results 117 can show different sensitivities with respect to these reaction rate constants because the relative magnitudes of the rate constants k^ k10, kn, k12 and k13 and k 1 4 depend strongly on the specific char and limestone. 6.3 Model Predictions The following sub-sections discuss the effects of fuel-N distribution in char and volatile, limestone addition, excess air, air staging and riser exit geometry. 6.3.1 Effects of Fuel-N Distribution in Char and Volatile Figure 6.4 shows the effect of the fuel-N distribution on the NO and N 20 emissions. N 20 emissions increase with increasing fraction of char nitrogen; NO first increases, then levels off. If the fuel nitrogen is completely contained in the volatile (i.e. Char-N/Fuel-N = 0), the N 20 emission is low, while the NO emission accounts for half of the emission when fuel-N is exclusively contained in the char. This means that both homogeneous and heterogeneous reactions are important for NO formation, but only the heterogeneous reactions are crucial for N 20 formation. Because of the specific reaction rate constants used, the above conclusions apply to Conoco coke only. It will be interesting to perform the calculation on differenr fuels in the future. As mentioned previously (Chapter 2), Zhao (1992) has correlated the NOx emissions to "VN". There might be interrelation between " VN" and the split of fuel-N between char and volatile. It is also of interest to see how the effcet of fuel-N split depend on the reaction rate constants which are characterist of fuels. Letting the decomposition constant for N 20 over the char surface, i.e. k13 equal to zero and asuming that the fuel nitrogen all exists in the form of HCN and NH3, barely 15 ppm N 20 is obtained, which indicates that the oxidation of HCN and NH 3 are not very active pathway for N 20 formation under FBC conditions. However, it is noteworthy that Chapter 6 NO and N20 Modelling Results 118 Char-N/Fuel-N Figure 6.4: Predicted Effect of Fuel-Nitrogen Distribution on NO and N 20 Emissions. In all the cases U g = 7.0 m/s, Gs = 65 kg/m2 s, P/S = 1.6, V c a o = 0 and the secondary air is introduced 3.4 m above distributor. The 0 2 concentration at the riser exit is 4.7% based on the prediction by the 0 2 sub-model. The values of model parameters are as follows: Kca = 1.0 m/s, k n = 27 m-s-J, k12 = 0.4 m3kg-J*-], k13 = 14.6 m3kg-'*-J. Chapter 6 NO and N20 Modelling Results 119 the effect of CO on N 2 0 formation from HCN oxidation is not taken into consideration in this thesis, and this might lead to under-prediction of the contribution of the homogeneous reaction to N 2 0 formation. According to Hulgaard (1991), CO is a strong reducing reagent which dramatically increases the yield of N 2 0 when HCN is oxidized at FBC temperature (Figure 6.5). 6.3.2 Limestone Addition As discussed in section 5.2.2.4, the formation of NO from NH 3 and the reduction of NO by CO (r 1 0 and r 1 4 in Table 3.9) are both catalyzed by limestone. The net effect of limestone addition on NO emissions results from competition between these two reactions. Because the 0 2 sub-model developed in Chapter 4 cannot predict CO concentration profiles, the CO concentration profiles employed in modelling the effects of limestone addition is an empirical one. For Run 17-1 (Conoco coke), approximately 2000 ppm CO was measured near the wall throughout the riser, while in the core, there was about 2000 ppm below and 1000 ppm above the secondary air port (Brereton et al, 1992). The predictions in Figure 6.6 illustrate that the effect of limestone addition on NO depends on the distribution of fuel nitrogen between the char and volatiles and the amount of limestone added. For the base case (Char-N/Fuel-N = 0.8, Conoco coke), NO is reduced as the fraction of limestone in solids (Vcao) increases. Limestone addition leads to less NO for fuels of lower char nitrogen content (Char-N/Fuel-N = 0.2 and 0.5). In these cases, concentration of NO goes through a maximum when V c a o is increased, more prominent with larger fraction of volatile-N (Char-N/Fuel-N = 0.2). The increasing portion of the curves are accounted for by the catalytic NO formation recation (r10). Since reaction rate of the catalytic reduction (r14) is proportional to the concentration of NO, it becomes more competitive as NO concentration is increased, leading to the predicted maximum. Chapter 6 NO and N20 Modelling Results 120 0 X CO 1 LL O O O o 60 50 40 30 20 10 A" T 1 1 1 1 1 r — A^- A_ _A H C N + N Q | A / /\ • / A / V / \ A A / / / / / / / / / / / A--^ / HCN + NO + CO "'A / \ / \ / A \ A HCN ' A / A / NH 3+ NO 'A- *r" X A-AA A-- A N H 3 T — r * £ - i — i — T i—1—i—1—r 900 1000 1100 1200 1300 1400 1500 TEMPERATURE (K) Figure 6.5: N 2 0 Formation for Different Reaction Conditions (Hulgaard, 1992). Experiments were carried out in a quartz reactor with a residence time of about 55 msec at 1200°K. The inlet conditions for H C N oxidations were: H C N : 310 ppm, NO: 440 ppm, CO: 1750 ppm, 0 2 : 2.3%, HjO: 2.7%. For N H 3 oxidations: N H 3 : 770 ppm, NO: 600 ppm, 0 2 : 2.6%, H 2 0 : 1.0%. Chapter 6 NO and N20 Modelling Results 121 CL CL 340 320 300 Char-N / Fuel-N • 0.8 • 0.5 A 0.2 g 260 c: o § 240 O 220 I— 0.0 0.1 0.2 0.3 V cao 0.4 0.5 Figure 6.6: Predicted Effect of Limestone Addition on NO Emissions. The calculations were based on the operating conditions for Run 17-1, Conoco coke (char-N/Fuel-N = 0.8) but with different distributions of fuel nitrogen between the char and volatiles. The secondary air was introduced 3.4 m above the distributor. T = 1155°K, U g = 7.0 m/s, Gs = 65 kg/m2s, P/S = 1.6, V C a Q = 0.4. The 0 2 concentration at the riser exit was 4.7% as predicted by the 0 2 sub-model. The values of model parameters are as follows: Kca =1.0 m/s, k u = 27 m-s-1, k12 = 0.4 m3-kg-1-s-1, k13 = 14.6 m3-kg-J-s-]. Chapter 6 NO and N20 Modelling Results 122 The above results roughly agree with the observations from the UBC pilot plant that the NO emission from the Conoco coke combustion is reduced by limestone addition while the reverese trend occurs for the CANMET pitch combustion. However, the NO reduction predicted in Figure 6.6 is much lower than was achieved in proceeding from Run 17-1 where T = 882, Ca/S = 0, NOx = 268 ppm to Run 17-3 where T = 848, Ca/S = 3.2, NOx = 66 ppm. The discrepancy may be because: (i) the temperature is assumed to be constant, while a drop in temperature of more than 40°C has occurred from Run 17-1 to Run 17-3; (ii) the rate constant for the decomposition of NO by CO over limestone, i.e. k 1 4 (Tsjujimura et al., 1983) is underpredicted. Figure 6.7 shows that the predicted concentration of NO is sensitive to k14. Larger value for k 1 4 corresponds to lower NO concentration. Limestone can catalyze the decomposition of N 2 0 (reaction 15 in Table 3.9). However, the rate is usually one to two order of magnitude less (Tables 3.7 and 3.8). Thus, only a very small fraction of N 2 0 is decomposed over the limestone surface 9 than . in CFB combustors, and k15 is not a significant parameter in this modelling work (Figure 6.3). Since limestone addition influences NO, which is a major source of N 2 0 in FBC, the impact of limestone addition on N 2 0 is primarily due to its impact on NO. The effect of limestone on N 2 0 in Figure 6.8 is smaller than the effect on NO in Figure 6.6. Similarly, N 2 0 reduction is less sensitive to V c a o in Figure 6.9 than the NO reduction in Figure 6.7. 6.3.3 Excess Air Effects As discussed in section 2.1.2, excess air is customarily approximated by 0 2 concentration in the flue gas in FBC combustion. The concentrations of 0 2 , NOx and N 2 0 in flue gas are influenced by reactions in both the riser and the secondary cyclone. The reaction in the secondary cyclone is beyond the scope of this work. Here excess air is therefore referred to 0 2 concentration at the riser exit. As shown in Figure 6.10, the model predicts that Chapter 6 NO and N20 Modelling Results 123 350 Figure 6.7: Sensitivity of the Predicted NO Emission to Reaction Rate Constant, k14. In all the cases T = 1155°K, U g = 7.0 m/s, Gs = 65 kg/m2 s, P/S = 1.6 with secondary air introduced 3.4 m above distributor. The 0 2 concentration at the riser exit is equal to 4.7% as per the prediction by the 0 2 sub-model. k 1 4 = 6.73E-2 (ppm-1 s_1) corresponds to the rate expression of Tsujimura et al. (1983). The values of model parameters are as follows: Kca = 1.0 m/s, k n = 27 m-s-', k12 = 0.4 m3-kg-J-s-], k13 = 14.6 m3-kg-J-s-]. Chapter 6 NO and N20 Modelling Results 124 CL CL £Z o i _ "E CD o c o o 200 i — 175 -150 -125 -100 -75 -50 -25 -0 — 0.0 ± Char-N / Fuel-N — • — 0.8 — • — 0.5 — A — 0.2 0.1 0.2 V 0.3 0.4 cao Figure 6.8: Predicted Effect of Limestone Addition on N 20 Emission. The calculation is based on Run 17-1 for Conoco coke, but the distribution of fuel nitrogen in char and volatile is varied. Secondary air was introduced 3.4 m above the distributor. T = 1155°K, U g = 7.0 m/s, Gs = 65 kg/m2 s, P/S = 1.6, V C a 0 = 0.4. The 0 2 concentration at the riser exit is 4.7% as predicted by the 0 2 sub-model. The values of model parameters are as follows: Kca = 1.0 m/s, k n = 27 ms-1, k12 = 0.4 m^kg-1^1, k13 = 14.6 m3-kg-J*-'. Chapter 6 NO and N20 Modelling Results 125 150 100 0.0 0.1 0.2 V 0.3 0.4 cao Figure 6.9: Sensitivity of the Predicted N 20 Emissions to Reaction Rate Constant, k14. In all the cases T = 1155°K, U g = 7.0 m/s, Gs = 65 kg/m2 s, P/S = 1.6 and secondary air is introduced 3.4 m above the distributor. The 0 2 concentration at the riser exit is equal to 4.7% based on the prediction from the 0 2 sub-model. k 1 4 = 6.73E-2 (ppm-1 s"1) corresponds to the rate expression given by Tsujimura et al. (1983). The values of model parameters are as follows: Kca =1.0 m/s, k n =27 ms-', k12 =0.4 m3-kg-]-s-J, k13 =14.6 m3-kg-'-s-1. Chapter 6 NO and N20 Modelling Results 126 400 Excess Air (O 2%) Figure 6.1.0: Predicted Effect of Excess Air on the NO and N 20 Emissions. In all the cases U g = 7.0 m/s, Gs = 65 kg/m2 s, P/S = 1.6 and the secondary air is introduced 3.4 m above the distributor. The values of model parameters are as follows: Kca =1.0 m/s, k u = 27 ms-], k 1 2 = 0,4 m3-kg-'-s-}, k13 = 14.6 m3-kg-'-s-]. Chapter 6 NO and N20 Modelling Results 127 both NO and N 2 0 emissions almost increase linearly with "excess air" . Figure 6.11 demonstrates how the predicted 0 2 profiles in the core and the annulus change with the "excess air". Note that in calculating these 0 2 profiles, the 0 2 consumption rate by the hydrocarbon (eqns 4.8 to 4.10) are assumed to be proportional to the difference between the 0 2 concentration at the riser inlet and outlet. Increasing "excess air" raises the 0 2 concentrations all along the riser. Furthermore, it decreases the volume fraction of char in solids as shown in Table 6.2, which means that fewer active sites are available for NOx and N 2 0 to be decomposed. The increased 0 2 and decreased V c h a r both contribute to the increased NOx and N 2 0 emissions. Table 6.2: Predicted Volume Fraction of Char in Solids (V c h a r) Excess Air (% 0 7 ) 2.0% 0.1210 3.0% 0.0870 4.0% 0.0665 4.7% 0.0565 6.3.4 Air Staging Figure 6.12 shows the effect of the height of secondary air port and the P/S ratio on NO and N 2 0 emissions by assuming that 0 2 concentration at the riser exit is kept constant while either of them is changed. Two heights are explored, 0.9 m and 3.4 m above the distributor, with P/S varying from 0.5 to 100. The model predicts that there is an optimum P/S giving the lowest NO and N 2 0 emissions. For small P/S (less than 1.5), NO is somewhat lower when secondary air is injected via the lower secondary air port, while at larger P/S ratios, use of the upper secondary air port gives better NO and N 2 0 emissions control. For very large P/S ratio (larger than 20), most air is supplied via the primary Chapter 6 NO and N20 Modelling Results 128 Figure 6.11: Predicted Effect of Excess Air on 0 2 Profiles. The calculation is based on Run 17-1 for Conoco coke, and assume that the volatile release rate is proportional to the difference between the 0 2 concentrations at the riser inlet and outlet. The secondary air is 3.4 m above the distributor. T = 1155°K, IT = 7.0 m/s, Gs = 65 kg/m2 s, P/S = 1.6. Chapter 6 NO and N20 Modeling Results 129 550 500 450 400 350 Z 300 "O ro 250 O Z 200 150 100 Q_ CL i; Secondary Air Port Position — • — 3.4 m — * — 0.9 m 1 | I r 1 0.1 1 10 100 P / S Figure 6.12: Predicted Effects of P/S ratio and Height of Secondary Air Injection on NO and N 20 Emissions. In all the cases T = 1155°K, U g = 7.0 m/s, Gs = 65 kg/m2 s. The 0 2 concentration at the riser exit is fixed at 4.7%. The values of model parameters are as follows: Kca = 1.0 m/s, k n = 27 m*-1, k12 = 0.4 m3-kg-J*-], k13 = 14.6 m3-kg-J-yJ. Chapter 6 NO and N20 Modeling Results 130 air inlet and NOand N 20 are hardly affected, no' matter where a small secondary air stream is introduced. The prediction that, for an intermediate value of P/S, NO and N 20 emissions are reduced when the secondary air is introduced at a higher level is consistent with observations of Brereton et al. (1992) and Amand et al.(1992). Air staging affects the fuel-N distribution and the oxygen profile in the CFB riser. For a small P/S ratio, both the solid suspension density profiles in the primary zone and the oxygen profiles change substantially with the height of the secondary air ports. It is difficult to interpret why the low secondary air port have a better performance in NO control. However, for larger P/S ratio, because the solid suspension density profiles are only slightly affected by the secondary air port height, the lower NO and N 20 emissions with the higher secondary air ports can be simply explained in terms of the oxygen profile. Figure 6.13 shows that with P/S = 1.6, when a higher secondary air port is used, there is a smaller area underlying the 0 2 concentration profile, suggesting that an oxygen-lean atmosphere is created in the combustor, retarding the formation of NO and N20. 6.3.5 Riser Exit Geometry Riser exit geometry has considerable impact -on solids distribution in CFBC risers, and is an important factor in riser design (Brereton and Grace, 1993). The riser exit has been correlated (Senior, 1992) by Rf, the fraction of upflowing solids reflected down the riser walls by the riser exit, with large Rf being close to one for abrupt exits. With two different values of Rf, 0.75 and 0.86, respectively, the solids suspension density profiles in the riser predicted by the modified Senior's model (section 4.1) are shown in Figure 6.14. The two profiles divert at the top of the reactor, where the suspension density rises more sharply Chapter 6 NO and N20 Modelling Results 131 21 Height (m) Figure 6.13: Predicted Effects of Height of Secondary Air Entry on 0 2 Profiles. The calculation is based on Run 17-1 for Conoco coke, but the fuel feed rate is varied. The secondary air is introduced 3.4 m above the distributor. T = 1155°K, U g = 7.0 m/s, Gs = 65 kg/m2 s, P/S = 1.6. The 0 2 concentration at the riser exit is fixed at 4.7%. Chapter 6 NO and N20 Modelling Results 132 Figure 6.14: Predicted Suspension Density Profiles for CFB Risers with Different Exit Geometry. The secondary air is introduced 3.4 m above the distributor. T = 1155°K, U g = 7.0 m/s, G s = 65 kg/m2 s. The sands are of Sauter's mean diameter 0.191mm. Chapter 6 NO and N20 Modelling Results 133 for Rf = 0.86. The assumptions that these two risers are operated with the same 0 2 concentration at the riser exit leads to values of V c h a r of 0.0595 and 0.0565 for Rf = 0.75 and 0.86, respectively. The abrupt exit then gives somewhat higher 0 2 concentrations throughout the riser as shown in Figure 6.15. Predicted NO profiles are shown in Figure 6.16. Corresponding to the jump in the suspension density profiles (Figure 6.14), both NO profiles show a sharp drop near the riser exit, which leads to a cross-over between them. Consequently, the riser with the more abrupt exit is predicted to give a somewhat lower NO emission. This is because the char has a strong ability to decompose NO and the abrupt exit causes more solids reflection, increasing the local char concentrations at the top of the riser, so that NO is reduced more sharply in the riser with the more abrupt exit. Besides the cross-over at the top of the reactor, there is another one at the bottom (0.5 m). NO formation reactions are dominant at the bottom of the riser, the larger V c h a r for Rf =0.75 leads to higher NO formation rates and hence to a higher NO concentration. The 0 2 profiles shown in Figure 6.15 explain why the NO concentrations are higher for Rf = 0.86 between the two cross-overs. The predicted N 2 0 profiles in Figure 6.17 shows trends similar to those for NO profiles. The above result indicates that an abrupt riser exit provides the advantage of lower NO and N 2 0 emissions. It also indicates that if the char concentration can be raised at the top of the riser, where the 0 2 concentration is low, even lower NO and N 2 0 emissions could be expected. This could possibly be achieved by staging of the fuel supply. Chapter 6 NO and N20 Modelling Results 134 injection Height (m) Figure 6.15: Predicted 0 2 Profiles for Different Riser Exit Geometries. The secondary air is introduced 3.4 m above the distributor. T = 1155°K, U g = 7.0 m/s, G s = 65 kg/m 2 s, P/S = 1.6. The core 0 2 concentration at the riser exit is fixed at 4.7%. Chapter 6 NOand N20 Modelling Results' 135 Figure 6.16: Predicted NO Profiles for Different Riser Exit Geometries. The secondary air is introduced 3.4 m above the distributor. T = 1155°K, U g = 7.0 m/s, Gs = 65 kg/m2 s, P/S = 1.6. The 0 2 concentration at the riser exit is fixed at 4.7%. The values of model parameters are as follows: Kca = 1.0 m/s, k n = 27 m-s-J, k12 = 0.4 m3-kg-1-s-1, k13 = 14.6 m3-kg-]-srJ. Chapter 6 NO and N20 Modelling Results 136 Figure 6.17: Predicted N 20 Profiles for Different Riser Exit Geometries. The secondary air is introduced 3.4 m above the distributor. T = 1155°K, U g = 7.0 m/s, Gs = 65 kg/m2 s, P/S = 1.6. The 0 2 concentration at the riser exit is fixed at 4.7%. The values of model parameters are as follows: Kca - 1.0 m/s, k n = 27 m-s-1, k12 = 0.4 m3-kg-J-s-1, k13 = 14.6 m3-kg-J-s-J. Chapter 6 NO and N20 Modelling Results 137 6.4 Summary NO and N 20 modelling has been performed for CFB combustion of Connoco coke. Better agreement with the experimental data is achieved when a larger mass transfer coefficient between the core and annulus is used for NO and N 20 than for 02. For the Conoco coke combustion, both homogeneous and heterogeneous reactions are important for NO formation. However, only heterogeous reactions appear to be important for N 20 formation. The reaction between NO and char-N is the most important reaction in this model, accounting for a significant portion of the char-N consumption. Without considering the effect of CO on the homogeneous reaction in this model, it is possible that the contribution of the homogeneous reaction to N 20 formation is under-predicted. More work is needed to allow the effect of CO on the homogeneous N 20 formation to be quantified. The model predicts that both NO and N 20 emissions increase with increasing excess air. The effect of limestone addition depends on the fuel-N distribution between the char and volatiles and the amount of limestone added. For the two heights of secondary air injection investigated, 0.9 m and 3.4 m above the distributor, there is predicted to be an optimum P/S giving both low NO and N 20 emissions. In general, better NO and N 20 control are achieved by using the upper secondary air port. An abrupt exit is predicted to help achieve low NO and N 20 emissions. Staging of fuel feed could be helpful in reducing NO and N 20 emissions. Chapter 7 Overall Conclusions and Recommendations 7-1 Conclusions Experimental and theoretical studies have been performed to improve the understanding of the formation and reduction of N0 X and N 20 in a CFBC riser. Effects of the major operating parameters on NOx and N 20 emissions, including the fuel properties, temperature, excess air and limestone addition, were analyzed and compared with the emission data measured from the UBC CFBC pilot plant for combustion of a petroleum coke, a pitch and two coals of different ranks. Insight was provided into the axial and lateral NOx and N 20 profiles. A comprehensive NOx and N 20 model has been developed which can predict the NOx and N 20 emissions under various operating conditions. For the fuels tested, the fuel-N-to-NOx conversion increased with both VN and volatile content. However, there is no simple correlation between fuel-N-to-N20 conversion and these two parameters. Decreasing the temperature decreases NOx emissions but causes an increase of N 20 emissions. A decrease in the excess air decreases the conversion of fuel-N to both N0 X and N26. Limestone addition is predicted to reduce N 20 emissions, while the effect of limestone on NOx depends on the fuel-N distribution between the char and volatiles. Lateral NOx and N 20 concentration gradients exist in a CFBC riser, arising from the core-annulus distribution of char particles. The gradients are smaller than for 0 2 because of lateral char motion. The axial NOx profile shows a sharp increase at the bottom of the riser due to NO formation, before declining higher up the riser, where the reduction 138 Chapter 7 Conclusions and Recommendations 139 reactions are dominant. The predicted N 2 0 concentration profiles show a similar trend. The NOx and N 2 0 model comprises four basic components, the main NOx and N 2 0 model, an 0 2 sub-model, a devolatilization model and the hydrodynamic model of Senior and Brereton (1992). Fifteen reactions are involved in the kinetic scheme. Volatile nitrogen compounds are assumed to exist as NH 3 and HCN exclusively, in the ratio 0.4:0.6 as per Houser et al. (1980). The global kinetics derived from Hulgaard et al. (1991) were employed for the NH3 and HCN oxidation and the homogeneous N 2 0 decomposition reactions. However, without considering the effect of CO (Hulgaard, 1991), the contribution of HCN to N 2 0 formation is very likely under-predicted. Some of the heterogeneous reaction rate constants are obtained by fitting the model predictions of NO and N 2 0 in the core to the experimental data. The NO/N20 model developed was tested against one of the Conoco combustion runs from the UBC pilot plant. Two values have been explored for the gas transfer coefficients between core and annulus, 0.1 and 1.0 m/s, with better agreement achieved using the larger one. The sensitivity analysis shows that the oxidation of char-N and the reaction between char-N and NO are the most important NO and N 2 0 formation pathways under CFBC condition. Gas phase homogeneous reactions produce a significant amount of NO, but show very little contribution to N 2 0 generation. The effects of excess air, limestone addition and the height of the secondary air port on the NOx and N 2 0 emissions predicted by the model are in agreement with experimental results (Brereton et al., 1992; Amand et al., 1992). The model forecasts that there is an optimum P/S ratio for NO and N 2 0 emission control and that a riser with an abrupt exit emits less NO and N 2 0 than one with less internal reflection of solids at the top. Staging of fuel feed is helpful in controlling NO and N 2 0 emissions from CFBC. Chapter 7 Conclusions and Recommendations 140 7-2 Recommendations Some of the reaction rate constants determined by fitting the model predictions to the experimental data should be validated by experiments. More study is also needed of the homogeneous reactions so that better global kinetic expressions can be derived. It is important to consider the effect of CO. Considering the toxicity of HCN, less hazardous model fuel-nitrogen compounds e.g. methacryonitrile, methylamine, pyridine and quinone can be used instead for kinetic studies. Since NH 3 and HCN are important for NOx formation and may be important in a revised N 2 0 reaction scheme, it is interesting to know the NH 3 and HCN profiles in riser. The NO and N 2 0 model is capable of predicting NH 3 and HCN profiles. These predictions can be compared with experimental NH3 and HCN concentration profiles if they can be obtained in the future. The effects of the many operating parameters on NOx and N 2 0 emissions should be studied based on a well-planned experimental design. The current 0 2 model should be improved by: (i) making it possible to predict CO profiles (ii) taking into consideration the back-mixing of secondary air. Finally, because the NO and N 2 0 model is based on a mechanistic hydrodynamic model (Senior and Brereton, 1992), in which the model parameters are independent of reactor size, the model can also be applied to the commercial combustors. Tests should be conducted to validate the model for combustors of different scale in the future. Nomenclature Symbol Definition ' Unit Aa area of annulus m2 Ac area of core m2 Ae external surface area of char m2 A, riser cross-sectional area m2 ae external surface area per unit volume of char m1 C(s) char carbon Ca/S calcium-to-sulphur molar ratio Dchar Sauter mean diameter of char particles in riser m Ds Sauter mean diameter of sand particles m Gs solids circulation rate kg/s-m H overall riser height ' m [HCN]a HCN concentration in annulus ppm [HCN]C HCN concentration in core ppm Kca mass transfer coefficient between core and m/s annulus kc char combustion rate constant m/s K i=i to is reaction rate constants defined in Table 3.10 m fuel feed rate kg/s Nc0 nitrogen-to-carbon atomic ratio in pyrolyzed char Nc nitrogen-to-carbon atomic ratio for char in reactor -N(s) char nitrogen N v o l weight fraction of volatile-N in fuel • 141 Nomenclature 142 Symbol Definition Unit NH 3 concentration in annulus ppm [NHJC NH 3 concentration in core ppm [N0]a NO concentration in annulus ppm [N0]c NO concentration in core ppm [N20]a N 2 0 concentration in annulus ppm [N20]c N 2 0 concentration in core ppm [0Ja 0 2 concentration in annulus ppm [0JC 0 2 concentration in core ppm Pa perimeter of the annulus in contact with core m P/S primary-to-secondary air ratio -Q M volumetric flow rate of air in the primary zone m3/s Q 2 n d volumetric flow rate of air in the secondary zone m3/s Vi, i=l to 15 reaction rates defined in Table 3.9 ppm/s T temperature K t time s Ug superficial gas velocity in the secondary zone m/s, VN fuel nitrogen parameter (Zhao, 1992) -terminal settling velocity m/s ^char volume fraction of char in solids -V cao volume fraction of limestone in solids -v V HCN, a HCN release rate in annulus kmole/ m v v HCN, c HCN release rate in core kmole/ m V ' HCN, sp HCN release rate in the single phase primary zone kmole/ m V N H j a NH 3 release rate in annulus kmole/m Nomenclature 143 Symbol Definition Unit V N H 3 C NH 3 release rate in core kmole/m-s V NH, sp NH 3 release rate in the single phase primary zone kmole m-s Vo2 a 0 2 consumption rate by volatiles in annulus kmole/m-s V o 2 c 0 2 consumption rate by volatiles in core kmole/m-s V o 2 8 P 0 2 consumption rate by volatile in the single phase kmole/ms primary zone WCH4 weight fraction of fuel released as CH4 W H , weight fraction of fuel released as H2 z axial coordinate m a suspension density in core kg /m3 P suspension density in annulus kg /m3 y cross-sectional average suspension density kg /m3 sa voidage in annulus ec voidage in core ec sec voidage in core at secondary air port p a v e average suspension density throughout riser kg /m3 p c a o density of limestone kg/m3 Pchar density of char kg /m3 Psand density of sand kg /m3 Bibliography Amand, L.-E. and Andersson, S., Emissions of Nitrous Oxides (N20) from Fluidized Bed Boilers, Proc. 10th Intern. Conf. on Fluidized Bed Combustion, ed. A. M. Manaker, ASME, New York, 49-56(1989). Amand, L.-E., Leckner, B. and Andersson, S., Formation of N 20 in Circulating Fluidized Bed Boilers, 2nd Nordic Conference on Control of S0 2 and NOx Emissions from Combustion of Solid Fuels, Lyngby, 1990. Amand, L.-E. and Leckner, B., Formation of N 20 in Circulating Fluidized Bed Boilers, Energy and Fuels, 5, 815-823(1991a). Amand, L.-E. and Leckner, B., Influences of Fuel on the Emissions of Nitrogen Oxide (NO and N 20) from an 8 MW Fluidized Bed Boilers, Energy and Fuels, 5, 815-823(1991b). Amand, L.-E. and Leckner, B., Influences of Air Supply on the Emissions of NO and N 20 from a Fluidized Bed Boilers, 24th Intern. Symp. on Combustion, Sydney, 1992. Axworthy, A. E., Dayan, V. H. and Martin, G. B., Reactions of Fuel-N Compounds under Conditions of Inert Pyrolysis, Fuel, 57, 29-35(1978). Bader, R., Findlay, J. and Knowlton, T. M., Gas/Solid Flow Pattern in a 30.5-cm-diameter Circulating Fluidized Bed, Circulating Fluidized Bed Technology II, ed. P. Basu and J. F. Large, Pergamon Press, Oxford, 123-137(1988). Beer, J. M., Sarofim, A. F., Lee, Y. Y., NO Formation and Reduction in Fluidized Bed Combustion of Coal, Proc. 6th Intern. Conf. on Fluidized Bed Combustion, Atlanta, 942-955(1980). Bi, H. and Zhu, J., Static Instability Analysis of Circulating Fluidized Beds and the Concept of High Density Riser, 39, 8, 1272-1280,(1992). 144 Bibliography 145 Bowman, C. T., Fossil Fuel Combustion, ed. Bartok, W. and Sarofim, A. F., John Wiley and Sons, New York, 215(1991) Brereton, C. M. Ff. and Grace, J. R., End Effects in Circulating Fluidized Bed Hydrodynamics, in Circulating Fluidized Bed Technology, ed. A. A. Avidam, AICHE, New York, pp 137-144(1993). Brereton, C, Grace, J. R., Lim, C. J., Zhu, J., Legros, R., Muir, J. R., Zhao, J., Senior, R. C, Luckos, A. Inumaru, N., Zhang, J., Hwang, I., Environmental Aspects, Control and Scale-up of Circulating Fluidized Bed Combustion for Application in Western Canada, Final Report to Energy, Mines, and Resources Canada under Contract 55SS 23440-8-9243, 1992. Braun, A., B, C, Renz, U. Drischel, J. and Koser, H.J.K., Emissions of NO and N 20 from a 4-MW Fluidized Bed Combustor with NO Reduction, Proc. 11th Intern. Conf. on Fluidized Bed Combustion, ed. E. J. Anthony, ASME, 709-719(1911). Brown, R. A. and Muzio, L., N 20 Emission from Fluidized Bed Combustion, Proc. 11th Intern. Conf. on Fluidized Bed Combustion, ed. E. J. Anthony, ASME, 719-724(1911). Cabrita, I, Costa, M. R., Esparteiro, H. and Gulyurtlu, I., Determination of Overall Fuel-N Balance during Fluidized Bed Coal Combustion, Proc. 11th Intern. Conf. on Fluidized Bed Combustion, ed. E. J. Anthony, ASME, 985-990(1911). Gift, R., Grace, J. R. and Weber, M. E., Bubbles, Drops and Particles, Academic Press, New York, 1978. Collings, M.E., Mann, M. D., Young, B. C. and Botros, P. E., The Effect of Coal-Fuel Properties on NO Emissions in Fluidized-Bed Combustion, Proc. 12th Intern. Conf. on Fluidized Bed Combustion, ed. L. N. Rubow, ASME, 619-627(1993). deSoete, G. G., Overall Reaction Rates of NO and N Formation from Fuel Nitrogen, Proc. 15th Intern. Symp. on Combustion, Combustion Institute, Reinhold Pub. Corp, New York, 1093-1102(1975). Bibliography 146 deSoete, G. G. and Nastoll, W., Catalytic Nitrogen Chemistry on CaO and CaC03 at Fluidized Bed Temperature Conditions. Report IFP (Institute Francais du Petrole) No. 39326(1989). deSoete, G. G., Heterogeneous N 20 and NO Formation from Bound Nitrogen Atoms during Coal Char Combustion, Proc. 23rd Intern. Symp. on Combustion, Combustion Institute, Reinhold Pub, Corp, New York, 1257-1264(1990). Elkins, J. W., State of the Research for Atmospheric Nitrous Oxide (N20), Paper contributed to the Intergovernmental Panel on Climate Change, Boulder, December, 1989. EPA, EPA/IFP European Workshop on the Emission of Nitrous Oxide from Fossil Fuel Combustion, EPA-600/9-89-089. Field, M. A., Gill, D. W., Morgan, B. B. and Hawsley, P. G. W., Combustion of Pulverized Coal, BCURA, Leatherhed, England(1967). Furusawa, T., Honda, T., Takamo, J. and Kunii, D., 2nd Intern. Conf. on Fluidization, Cambridge University Press(1978). Furusawa, T., Kunii, D., Oguma, A. and Yamada, N., Rate of Reduction of Nitric Oxide by Char, International Chemical Rngineering, 20, 239-244(1980). Furusawa, T., Tsjumura, M., Yasunaga, K. and Kojima, T., Fate of Fuel Bond Nitrogen within Fluidized Bed Combustor under Staged Air Firing, Proc. 8th Intern. Conf. on Fluidized Bed Combustion, ASME, 1095-110.4(1985). Gibbs, B. M., Beer, J. M. and Pereira, F. I., A Simplified Model for NO Formation from Fuel-Nitrogen in Fluidized Bed Combustion, Symp. Series No. 4, Inst, of Energy, V-4(l-12)(1980). Grace, J. R. and C. J. Lim, Circulating Fluidized Bed Combustion of Coal, Woodwaste and Pitch, Final Report to Energy, Mines, and Resources Canada, under Contract 24ST 23440-6-9007(1987). Bibliography 147 Grace, J. R., Lim, C. J., Brereton, C, Yu, W., Li, H, Zhao, J. and Wu, R. L., Multi-Fuel Combustion in a Circulating Fluidized Bed, Proc. 6th Canadian Bioenergy R & D Seminar, ed. Z. Stiasny, Elsevier Barking, UK, 438-442(1987). Grace, J. R., Brereton, C, Lim, C. J., Legro, R., Zhao, J., Senior, R. C, Wu, R. L., Muir, J. R. and Engman, R., Circulating Fluidized Bed Combustion of Western Canadian Fuels, Final Report to Energy, Mines and Resources Canada(1989). Hansen, P. F. B. and Dam-Johansen. K, Johnsson, J.-E. and Hulgaard, T., Catalytic Reduction of NO and N 20 on Limestone during Sulphur Capture under Fluidized Bed Combustion Conditions, Chemcal Engineering Science, 47, 2419-2424(1992). Hansen, P. F. B. and Dam-Johansen. K, Limestone Catalyzed Reduction of NO and N 20 Under Fluidized Bed Combustion Conditions, Proc. 12th Intern. Conf. on Fluidized Bed Combustion, ed. L. N. Rubow, ASME, 779-787(1993) Hao, W. M., Wofsy, S. C, McElory, M. B., Beer, J. M. and Toqan, M. A., Sources of Atmospheric Nitrous Oxide from Combustion, J. Geophysical Research, 92, 3098-3014(1987) Hayhurst A. N. and Lawreance A.D. Emissions of Nitrous Oxide from Combustion System, Progress in Energy and Combustion Science, 18, 529-552(1992). Hayne, B. S., Reactions of Ammonia and Nitric Oxide in the Burnt Gases of Fuel-Rich Hydrocarbon-Air Flames, Combustion AND FLAME, 28, 81-91(1977). Hiltunen, M., Kilpinen, P. Hupa, M. and Lee, Y.Y., N 20 Emission from CFB Boilers: Experimental Results and Chemical Interpretation, Proc. 11th Intern. Conf. on Fluidized Bed Combustion, ed. E.J. Anthony, ASME, 687-694(1991). Horio, M., Mochizuki and Koike, J., Emissions of NO from Fluidized Bed Char Combustion, The 4th Japan-China Symp. on Coal and Cl Chemistry Proc, Osaka, 249-254(1993) Bibliography 148 Horio, M., Mori, S., Muchi, I., A Model Study for the Development of Low NOx Fluidized-Bed Coal Combustion, Proc. 5th Intern. Conf. on Fluidized Bed Combustion, ASME, 605-624(1977). Houser, T. J., Hull, M., Alway, R. M. and Biftu, T., Kinetics of Formation of HCN during Pyridine Pyrolysis, Int. J. Chem. Kinetics, 12, 569(1980). Hulgaard, T., Glarbog, P. and Dam-Johansen, K., Hohogeneous Formation and Destruction of N 20 at Fluidized Bed Combustion Conditions,, Proc. 11th Intern. Conf. on Fluidized Bed Combustion, ed. E. J. Anthony, ASME, 991-998(1991). Hulgaard, T., Nitrous Oxide from Combustion, Ph.D. Dissertation, Technical University ofDenmark(1991). International Energy Agency, Coal Information 1989, OCED/IEA, OCED, Paris, 1989. Iisa, K., Santala, P. and Hupa, M., Heterogeneous Formation and Destruction of Nitrous Oxide under Fluidized Bed Condition, Proc. 11th Intern. Conf. on Fluidized Bed Combustion, ed. E.J. Anthony, ASME, 1027-1033(1991). Jensen A., Johnsson, J.-E. and Dam-Johansen, K., Formation of Nitric Oxide From Heterogeneous Oxidation of Hydrogen Cyanide at Fluidized Bed Combustion Conditions, Proc. 12th Intern. Conf. on Fluidized Bed Combustion, ed. L. N. Rubow, ASME, 447-454(1993). Johnsson, J. E., Modelling of NO Formation in Fluidized Bed Combustion, 16th IE A AFBC Meeting, Palo Alto(1988). Johnsson, J. E. and Dam-Johanson, K., Formation and Reduction of NOx in a Fluidized Bed Reactor. Proc. 11th Intern. Conf. on Fluidized Bed Combustion, ed. E. J. Anthony, ASME, 1389-1396 (1991). Johnsson, J. E., Nitrous Oxide Formation and Destruction in Fluidized Bed Combustion, 23rd IEA AFBC Meeting, Firenze, November(1991). Bibliography 149 Khan, T., Lee, Y. Y. and Young, L., Heterogeneous Decomposition of Nitrous Oxide in the Operating Temperature Range of Circulating Fluidized Bed Combustors. EPRI/EPA Joint Symposium on Stationary Combustion NOx Control, Washington D. C, March, 1991. Kilpinen, P. and Hupa, M., Homogeneous N 20 Chemistry at Fluidized Bed Combustion Conditions- a Kinetic Modelling Study, Combustion and Flame, 85, 94-104(1991). Kojima, T., Ishihara, K. I., Yang, G. and Furusawa, T., Measurement of Solids Behavior in a Fast Fluidized Bed, J. Chem. Eng. Japan, 72, 341(1987). Kramlich, J. C, Cole, J. A., McCarthy, J. M., Lanier, Wm. S., Mechanism of Nitrous Oxide Formation in Coal Flames, Combustion and Flame, 77, 375-384(1989). Kunii, D. Furusawa, T. and Wu K. T., NO Emissions Control by a Staged Fluidized Bed Combustor of Coal, Fluidization, ed. J.R. Grace and J. M. Matsen, Plenum, New York, 175-183(1980). Leckner, B. and Amand, L.-E, Emissions from a Circulating and a Stationary Fluidized Bed Boiler: A Comparison, Proc. 9th Intern. Conf. on Fluidized Bed Combustion, ed. J.P. Mustonen, ASME, 891-897(1987). Lee, Y. Y., Walsh, P. M., Dutta, A., Beer, J. M. and Sarofim, A. F., Spatial Distribution of Coal Nitrogen and Nitric Oxide in the Bed of a Fluidized Combustor, Proc. 7th Intern. Conf. on Fluidized Bed Combustion, 1, ASME, 264(A-I)(1983). Linak, W. P., McSorley, J. A., Hall, R. E., Ryan, J. V., Srivastava, R. K, Wendt, J. O. L. and Mereb, J. B., N 20 Emissions from Fossil Fuel Combustion, Proc. EPA/EPRI NOx Control Symposium, San Francisco, 1989. Lee Y. Y., and Hiltunen, M., The Conversion of Fuel-Nitrogen to NOx in Circulating Fluidized Bed Combustion, EPA/EPRI Symp. on NO control, San Francisco, 1989. Macrae, K. M., New Coal Technology and Electric Power Development, Can. Energy Res. Inst., Calgary, 1991. Bibliography 150 Miettinen, H., Stromberg, D. and Lindquist, O, The Influence of Some Oxide and Sulphate Surfaces on N 20 Decomposition, Proc. 11th Intern. Conf. on Fluidized Bed Combustion, ed. E.J. Anthony, ASME, 999-1003(1991). Miller, J. A. and Brown, C. T., Mechanism and Modelling of Nitrogen Chemistry in Combustion, Prog. Energy Comb. 287-338(1989). Mjornell, M., Leckner, B., Karlsson, M. and Lyngfelt, A., Emission Control with Additives in CFB Coal Combustion, Proc. 11th Intern Conf on Fluidized Bed Combustion, ed. E.J. Anthony, ASME, 655-664(1991). Moritomi, H., Suzuki, Y., Kido, N and Ogisu, Y., NOx Emission and Reduction from Circulating Fluidized Bed Combustor, Circulating Fluidized Bed Technology III, ed. P. Basu, M. Horio and M. Hastani, Pergamon Press, Oxford, 399-404(1991). Moritomi, H. and Suzuki Y., N 20 Formation during Coal Combustion, IEA/AFBC Meeting,, Belgrade, 1990. Muzio, L. J. and Kramlich, J. C, An Artifact in the Measurement of NO from Combustion Sources, Geophysics Research Letters, 15, 1369-1372(1988). Perry, R. A., Kinetics of the Reactions of NCO Radicals with H and NO using Laser Photolysis-Laser Induced Fluorescence, J. Chem. Phys., 12, 5755-5758(1985). Pereira, F. J., Beer, J. M., Gibbs, B. M. and Headley, A. B., 15th Intern. Symp. on Combustion, Combustion Institute, 1149-1156(1975). Pohl, J. H. and Sarofim, A. F., Devolatization and Oxidation of Coal Nitrogen, Proc. 16th Intern. Symp. on Combustion, Combustion Institute, Reinhold Pub. Corp, New York, 491-501(1976). Proffit, M. H., Fahey, D. H, Kelley, K. K, Loewenstein, M, Oodolske, J. R. and Chan, K. R., High Latitude Ozone Loss Outside the Antarctic Ozone Hole, Nature, 342, 233-237(1989). Bibliography 151 Reh, L., Schmidt, H-W., Daradimos, G. and Petersen, V., Circulating Fluidized Bed Combustion: an Efficient Technology for Energy Supply and Environmental Protection, Symp. Series No. 4, Inst, of Energy, VI-2(1-11)(1980). Rhodes, M. J., Laussmann, P., Villian, F. and Geldart, D., Measurement of Radial and Axial Solid Flux Variation in the Riser, Circulating Fluidized Bed Technology II, ed. P. Basu and J. F. Large, Pergamon Press, Oxford, 155-164(1988). Sarofim, A. F. and Flagan, R. C, NOx Control for Stationary Combustion Sources, Prog. Energy Combust. Sci., 2, 1-25(1976). Senior, R. C, Circulating Fluidized Bed Fluid and Particle Mechanics: Modelling and Experimental Studies with Application to Combustion, PhD Thesis, The University of British Columbia(1992). Senior, R. C. and Brereton C, Modelling of Circulating Fluidized Bed Solids Flow and Distribution, Chemcal Engineering Science, 47, 281-296(1992). Shaw, R. W., The Atmospheric Pathway for Oxides of Nitrogen, Report to Environment Canada, EPS 2/TS/2, 1984. Shimizu, T., Tachiyama, Y., Souma, M. and Inagaki, M., Emission Control of NOx and N 20 of Bubbling Fluidized Bed Combustor, Proc. 11th Intern. Conf. on Fluidized Bed Combustion, ed. E.J. Anthony, ASME, 695-700(1991). Shimizu, T., Tachiyama, Y., Fujita, D., Kumazawa, K., Wakayama, O., Ishizu, K, Kobayashi, S., Shikada S. and Inagaki, M, Effect of SO Removal by Limestone on NO and NO Emissions from a Circulating Fluidized Bed Combustor, 6, 753-757(1992). Shimizu, T., Fujita, D., Ishizu, K., Kobayashi, S. and Ingaki, M., Effect of Limestone Feed on Emissions of NOx and N 20 from a Circulating Fluidized Bed Combustor, Proc. 12th Intern. Conf. on Fluidized Bed Combustion, ed. L. N. Rubow, ASME, 611-616(1993). Song, Y. Ff., Beer, J. M. and Sarofim, A. F., Oxidation and Devolatilization of Nitrogen in Coal Char, Combustion Science and Technology. 28. 177-183(1982). Bibliography 152 Tsujimura, M., Furusawa, T. and Kunii, D., Catalytic Reduction of Nitric Oxide by Carbon Monoxide over Calcined Limestone, J. of Chem. Eng. Japan, 16, No. 2, 132-136(1983). Tullin, C. J., Sarofim, A. F. and Beer, J. M., Formation of NO and N 20 in Coal Combustion: The Relative Importance of Volatile and Char Nitrogen, 12th Intern Conf. on Fluidized Bed Combustion, ed. L. N. Rubow., ASME, 599-609(1993). Wallis, G. B., Annular Two-Phase Flow, Part 1: A Simple Theory, J. Basic Eng., Trans. ASME, 59-72(1970a). Wallis, G. B., Annular Two-Phase Flow, Part 2, Additional Effects, J. Basic Eng., Trans ASME, 73-82(1970b). Wark, K. and Warner, C. F., Air Pollution - Its Origin and Control, Harper & Rowe, New York, 1981. Whaley, H., Anthony, E.J. and Lee, G. K, Control of Acid Rain Precursors, CANMET Rept, Ottawa, ERL, 89-16, 1989. Weiss, R. E., and Craig, H., Production of Atmospheric Nitrous Oxide by Combustion, Geophys. Res. Lett., 3, 751-753(1976). Wojtowicz, M. A., Oude Louis, J. A., Tromp, P. J., and Moulijin, J. A., N 20 Formation in Fluidized Bed Combustion of Coal, Proc. 11th Intern. Conf. on Fluidized Bed Combustion, ed. E. J. Anthony, ASME, 1013-1020(1991). Wong Poly, Incineration of Industrial Organic Waste in a Circulating Fluiidized Bed Combustor, MSc Thesis, The University of British Columbia(1994). Yang, W. C, CHML 561b course hand out, 1992 Bibliography 153 Zhang, J. M., N 20 Emissions from Fluidized Bed Combustion, seminar presented to the UBC CFBC group(1992). Zhao, J., Nitric Oxides Emissions from Circulating Fluidized Bed Combustion, PhD Thesis, The University of British Columbia, 1992. Zhao, J., Brereton, C, Grace, J. R., Lim, C. J., and Legros, R., Gas Concentration Profiles and NO Formation in Circulating Fluidized Bed Combustion, To be submitted Zhou, I., Grace, J. R., Qin S., Brereton C. M. H., Lim, C. J., Zhu, J., Voidage Profiles in a Circulating Fluidized Bed of Square Cross-Section, Chemical Engineering Science, 49, 3217-3226(1994). Appendix A Arrhenius Type Reaction Rate Constant Expressions This Arrhenius type reaction rate constant expression is of the form: E. * = * 0exp(--^) (Al) where k0 is the frequency factor, Eact is the activation energy and R is the ideal gas constant. Taking the natural logarithm of both sides, then: ln(k) = \n(k0)-Eac' R T (A-2) = to(ko)-(_L_^L)l^ 0 100,000 R T Hence k0 can be calculated from the interception y0 in a plot of In (A:) vs ^ 0^000 according to: k0=exp(y0) (A3) 'act RT can be calculated from the slope m by: E = -100000- m (A.4) R Figures A.l to A.3 show plots of ln(k;) vs for i = 1 to 8, with k{ and T from Tables 3.2, 3.4 and 3.6 154 Appendix A Arrhenius Type Reaction Rate Constant Expressions 155 Figure A. 1: ln(£.) vs 1 0 0 ^ 0 0 0 f o r j = i to 2, with kt from Tables 3.2 Appendix A Arrhenius Type Reaction Rate Constant Expressions 156 0 -2 -4 -6 -8 -10 -12 -14 -16 -18 k5: slope = -.29378, interception = 21.527 J L k3: slope = -.50719, interception = 30.888 1 7.2 7.4 7.6 7.8 8.0 8.2 8.4 8.6 (1/T)*100000 8.8 Figure A.2: ln(&;.) vs for;' = 3 to 5, with ki from Tables 3.4 Appendix A Arrhenius Type Reaction Rate Constant Expressions Figure A.3: l n ( £ ( . ) vs 1 0 0 > Q Q Q f o r ; = 6 to 8, with kt from Tables 3.6 Appendix B Derivations of [NH3], [HCN], [NO] and [N20] Profiles From the four algebraic equations in Table 4.3, [NH3]a, [HCN]a, [NO]a and [N 20] a can be expressed as functions of [NH3]C, [HCN]C, [NO]c and [N20]c in order to eliminate [NH3]a, [HCN]a, [NO]a and [N 20] a in the differential equations. The four differential equations can then be solved simultaneously. The derivations are given below: First, we define : QA\ - A"£" Q42 = ^ a ^ ~ S" ^ cao CA3 - ^ ° ^ ~ S" ^char PaKca ' PaKca PaKca CA4 = ^ ^CA5 = ^ -PaKca P2Kca If eqn (4.19) is divided by PaKca and the rate expressions in Table 3.9 are substituted for r, eqn (4.19) becomes [NH31 - [NH3 l+CAh (-k3 [NH3 ]a [02 ]a - k4 [NH3 ]. [02 ]a -h5[NH3UNO]a) + CA2-(-kl0[NH3}a[O2]a) + CA4 = 0 (B.l) Eqn (B.l) can be rearranged to [NH3]c -(1 + CA1 -k3[02]a + CA2• k]0[02]a.+ CA 1 -k4 • [02]J-[NH3]a -^CAl-k5-[NH3]a[NO]a+CA4 = 0 (B.2) Similarly, eqns (4.21) and (4.22) can be rearranged as: 158 Appendix B: Derivations of [NHJ, [HCN], [NO] and [N20] Profiles 159 [N0]e - (1 + CA2• ku [C0]a + CA3 • knPchar + CA3-ku • Nc • ae • [02 ]a)• [N0]a +(C42 • k]0 • [02 ]a + CA\• k4 • [02 ]a)• [NH, ]a + CA\-kn • [02]a • [HCN\ (B.3) +CA\• k, • [N20]a - CA\• k5 • [NO] a[NH, ]a + CA3 • k9 • Nc • ae [O2]a=0 [N20]c +CA3-ku-Nc-ae-[N0]a • [02]a + 0.5• CAl• k3 • [NH3]a • [02]a +0.5-CAhk6-[O2]a[HCN]a-(l + CAhk2+CA2-k,5-pcao+CAhk, (B.4) +CA3-Pchar-ku)[N2O]a=0 With A\ = CA2- kw • [02 ]a+CA\-k4- [02 ]a, eqn (B.2) becomes [iV7Y31 -(1 + CAl• k3[02]a+Al)-[NH3]a ~^CA\-ks-[NH3]a • [NO]a +C/14-0 (B.5) For eqn (B.3), let A2 = 1 + CA2 • ku • [CO]a + CA3 • ku • pchar +CA3-ku- Nc • ae [02 ]a A3 = CAX-k1-[02]a A4 = CAl-k, A5 = CA3.k9-Nc:ae.[02]a Using Al, A2, A3, A4 and A5, eqn (B.3) is expressed as: [NO]c - A2• [NO]a + Al• [NH3 ]a+A3- [HCN]a + A4 • [N20]a -CAl-k5-[NO]a[NH3]a+A5 = 0 Equation (B.6) - (B.5)x 1.5 yields [NO]c -1.5• [NH,]c - A2• [NO]a + (1.5 +1.5• CAl k3[02]a + 2.5• Al)• [NH3]a +A3 • [HCN]a + A4 • [N20]a+A4 • [N20]a + A5 -1.5 • C44 = 0 Appendix B: Derivations of [NHJ, [HCN], [NO] and [N20] Profiles 160 Eqn (B.7) is rearranged to give A2-[N0]a -(1.5 + 1.5-G41-A:3[02]a + 2.5-Al)-[NH3]a-A4-[N20]a = [N0]c -1.5- [NH3 ]C+A3- [HCN]a + A5- 1.5• CA4 (B.8) For eqn (B.4), let A6 = CA3-ku -Nc • ae [02]fl Al = 0.5-CAl-k3[O2]a AS = 0.5-CAhk6[O2]a A9=l + CA2-kl5-pcao+CA3-pchar • kn + CA 1 • (k, + k2) Eqn (B.4) can then be written A6• [NO]a + Al• [NH3 ]a-A9- [N20]a =-AS- [HCN]a - [N20]c (B.9) For qn (B.8), let A\0=--l.5 + l.5-CAhk3-[O2]a+2.5-A\ Al 1 = [N0]c - 1.5• [NH3]C+A3-[HCN]a +A5-1.5-CA4 Eqn (B.8) may be written as A2-[NO]a-Al0[NH3]a-A4-[N2O]a = All (B.10) For eqn (B.9) let ^12 = -AS • [HCN]a - [N20]c, so that eqn (B.9) becomes A6 • [NO]a + Al • [NH3 ]a-A9- [N20]a = A12 (B.ll) Appendix B: Derivations of[NHJ, [HCN], [NO] and [N20] Profiles 161 AT. A10 A l l Dividing eqn (B.10) by A4 and letting A13 = — , ,414 = — , A15 = , we may then A4 A4 A4 write A\3-[NO]a-A\4-[NH3]a-[N20]a=A\5 (B.12) A6 Al A12 Dividing eqn (B.l 1) by A9 and letting A16 - — , .418 = — , A19 = , we obtain A9 A9 A4 A16-[NO]a+AlS-[NH3]a-[N20]a=Al9 (B.13) Substracting eqn (B. 13) from eqn (B. 12) gives (Al3-Al6)-[NO]a-(A14 + AlS)-[NH3]a=A15-A19 (B.14) T f i * 4^14 + ^18 A15-A19 Ifwelet^420 = , A21 = A13-A16 A13-A16 then eqn (B.14) becomes [NO]a-A20-[NH3]a=A2l i.e., [NO]a=A2l + A20-[NH3]a (B.15) For eqn (B.2), Let ^22 = 1 + CA2 • k3 • [02 ]a + Al Appendix B: Derivations offNHJ, [HCN], [NO] and [N20] Profiles 162 A23 = --CA2-L 3 5 Then eqn (B.2) may be written as [NH3]e-A22-[NH3]a+A23-[NH3]a[NO]a+CA4 = 0 (B.16) Substitution of eqn (B.15) into eqn (B.16) yields [NH3]c-A22-[NH3]a + A23-[NH3]a-(A2\ + A20-[NH3]a) + CA4 = 0 or A20 • A23 • ([NH3 ]a f + (A21-A23 - A22)[NH3 ]a + [NH3 }c + CA4 = 0 (B. 17) ^ 2 3 - ^ 2 2 [NH,].+C«4 420-423 ,420423 eqn (B.17) becomes ([NH3 ]a f + A24-[NH3]a+ 425 = 0 (B.18) Consider the positive root of eqn (B.18), then r x r r r n -424+ J(424)2 -4-425 [NH3]a= ^ — ^ (B.19) Substituting eqn (B.19) into eqn (B.15), we obtain Appendix B: Derivations of[NHJ, [HCN], [NO] and [N20] Profiles 163 -A24 + J(A24)2 -4-A25x [NO]a=A2l + A20-( ^ - —) (B.20) From eqn (B.12) -A24 + J(A24)2 -4-A25„ [N20]a = AU • [A21 + A20 • ( ^—1 )] -AU.CA24 + ^A24)2^^)-AIS (B.21) In eqns (B.19), (B.20), (B.21), A21 and A24 are functions of [NHJ^ [NO]c [N20]c and [HCN]a (This is easy to tell from the summary in Table B.l). Recall that the task here is to convert [NH 3 ] a , [NO] a [N 2 0] a and [HCN] a to functions of [NH 3] C , [NO] c [N 2 0] c and [HCN] C, The remaining work is to derive [HCN]^ Substituting the rate expressions into eqn (4.20) and dividing by PaKca, after sorting [HCN]a [02 ]a , we may obtain [ H C N ] ° = n + ™ ft A , n t m ^ C 7 ^ + C 4 5 > < B 2 2> If ,426 = ( l + C 4 1 - ( * 6 + * 7 + * l ) . { 0 2 ] a ) then [#C7V]fl = A26 • ([HCN]e + CA5) (B.23) Appendix B: Derivations of[NHJ, [HCN], [NO] and [N20] Profiles 164 The expression for [HCN] a is next substituted in eqn (B.23) back into Al 1 and A12. After that, we substitute A l l , A12 into A21 and A24, then substitute A21 and A24 into eqns (B.19), (B.20), (B.21). Finally the expressions for [NH 3 ] a , [HCN] a , [NO] a and [ N 2 0 ] a from eqns (B.19), (B.23), (B.20) and (B.21) are in terms of [NH 3] C , [HCN] C, [NO] c and [ N 2 0 ] c and other known parameters in the N O - N 2 0 model equations. The definitions of A,,, = j 26; involved in the derivations above are summarized in Table.B.l Table B.l: Summary of Ai, i =1, 26 Al CA2-kw-[02]a+CA\-k4[02]a A2 l + CA2-k]4- [CO] A+CA3-kX2- pchar +CA3-ku-Nc-ae- [02 ]A A3 CAhk7-[02]a A4 CA\-kx A5 CA3k9-Nc-ae-[02]a A6 CA3-ku.Nc-ae-[02]a A7 0.5CAhk3-[O2]a A8 0.5-C41-VIAL A9 l + CA2-ku -Pcao +CA3-pcharkl3 +CAh(k]+k2) A10 1.5 + 1.5-041^3 [02]a +2.5-41 A l l * * [NO]c -1.5 • [NH3 ]c + A3 • [HCN]a + A5 -1.5 • CA4 A12** -A%-[HCN]a-[N20]c • A13 A2 AA ** means [NH 3 ] a , [HCN] a , [N0] a and [N 2 0] a have not been completely eliminated from the expressions. Appendix B: Derivations offNHJ, [HCN], [NO] and [N20] Profiles 165 Table B.l: Summary of Aj for i =1, 26, continued from previous page A14 ,410 AA A15** A l l A4 A16 A6 A9 A17 A18 A9 A19** A12 A4 A20 ^14 + ^18 A13-A16 A21** A15-A19 A13-A16 A22 l + CA2-k3-[02]a+Al A23 --CA2-L 3 5 A24** ,421-,423-,422 A20-A23 A25 [NH3]C+CA4 A20-A23 A26 1 (l + C41.(* 6+* 7+* 8).[0 2U ***A17 is not involved in the above derivations Appendix C Computer Programs for 0 2 Model and NO/NzO Model The programs for 0 2 model and for the NO/N20 model are coded in FORTRAN, using Finite Element Galerkin solution algorithm. Details of the model development and the input and output data are given in Chapter 4. 166 Appendix C Computer Programs for 02 Model and NO/N20 Model 167 C 02 MODEL C ANNOTATIONS FOR VARIABLES: ccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc C AAS,ACS:AREA OF ANNULUS AND CORE (m2) C C AT: CROSS SECTIONAL AREA (lti2) C C AV:EXTERIOR SURFACE AREA PER UNIT VOLUME OF CHAR IN RISER(l/m) C C DS:SAUTER MEAN DIAMETER FOR CHARS IN RISER (m) C C KCA:GAS TRANSFER COEFFICIENT BETWEEN CORE AND ANNULUS (m/s) C C KCOM :COMBUSTIONATE CONSTANT (m/s) C C LENGTH:LENGTH FOR REACTOR CROSS-SECTION C C NEB,NET:NUMBER OF ELEMENTS EMPLOYED FOR 1st AND 2nd ZONES C C 02ABS,02ATS:02 CONCENTRATION IN ANNULUS IN l s t ( B ) AND 2nd(T) ZONE C C 02CBS,02CTS:02 CONCENTRATION IN CORE IN l s t ( B ) AND 2nd(T) ZONE C C QB,Q:VOLUMETRIC FLOW RATE IN THE PRIMARY AND SECONDARY ZONE (m3/s) C C SECPOS:SECONDARY AIR PORT HEIGHT (m) C C SPLIT:PRIMARY TO SECONDARY AIR RATIO C C T:TEMPERATURE (K) C C UG:SUPERFICIAL VELOCITY OF GAS IN THE 2nd ZONE (m/s) C C VCHAR:THE VOLUME FRACTION OF CHAR IN SOLIDS C C XB,XT:COORDINATES FOR THE GRID IN THE l s t ( B ) AND 2nd(T) Zone C ccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc C FORMAT INPUT FILES MUST ABIDE BY: cccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc c c C 02V0L.DAT:CONTAINS THE 02 CONSUMPTION BY THE HYDROCARBONS CONTAINED C C IN VOLATILE, FREE FORMAT; BUT MUST FOLLOW THE ORDER AS BELOW C C (X,02VOCS,02VOAS), WHERE X IS THE COORDINATES C C OF NODAL POINTS C C HYDRO.DAT:CONTAINS THE HYDRODYNAMIC INFORMATION FROM SENIOR AND BRERETON C C MODEL,INCLUDING THE VOIDAGE IN CORE AND ANNULUS THROUGHOUT THE C C RISER; FREE FORMAT; BUT MUST FLOLLOW THE ORDER AS BELOW C C (NE,9,X,EPSCS,EPSAS,AAS,PAS), WHERE X IS THE COORDINATES OF C C NODAL POINTS C C C cccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc DOUBLEPRECISION DOUBLEPRECISION DOUBLEPRECISION DOUBLEPRECISION DOUBLEPRECISION DOUBLEPRECISION DOUBLEPRECISION DOUBLEPRECISION DOUBLEPRECISION DOUBLEPRECISION DOUBLEPRECISION DOUBLEPRECISION AABS(30),AATS(30),AV EPSABS(30),EPSATS(30) EPSCBS(30),EPSCTS(30) LENGTH O2PMBC(30),O2PMBA(30),O2VOBC(30),O2VOBA(30) O2PMTC(30),O2PMTA(30),O2VOTC(30),O2VOTA(30) O2ABS(30),O2ATS(30) O2CBS(30),O2CTS(30) O2AVEB(30),O2AVET(30),02SECM PABS(30),PATS(30) SECPOS,UGB XB(30),XT(30) DOUBLEPRECISION AT,KCOM,O2C0,VCHAR COMMON/R1/AT,KCOM,O2C0,VCHAR Appendix C Computer Programs for 02 Model and NO/N20 Model 168 O P E N ( 5 , F I L E = ' C : H Y D R O . D A T ' ) O P E N ( 7 , F I L E = ' C : 0 2 V O L . D A T ' ) O P E N ( 8 , F I L E = ' C : 0 2 . D A T ' ) ccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc ccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc NEB=25 NET=2 5 V C H A R = . 0 5 6 5 0 K C O M = 1 . 0 5 E - 1 T = 1 1 5 5 . S P L I T = 1 . 6 S E C P 0 S = 3 . 4 U G = 7 . 0 D P = 2 . 4 5 8 8 E - 4 K C A = 0 . 1 L E N G T H = . 1 5 2 ccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc C C A L C U L A T E FLOW R A T E I N T H E PRIMARY AND SECONDARY ZONE Q = U G * A T Q B = S P L I T / ( 1 . + S P L I T ) * Q A V = 6 . / D P A T = L E N G T H * * 2 . C C MODEL BOTTOM AS C O R E - A N N U L U S C C C INPUT DATA C C C C C * C BOUNDARY C O N D I T I O N I N T H E 1 s t ZONE 0 2 C O = 2 1 0 0 0 0 . C S O L V E 02 B A L A N C E E Q U A T I O N I N 1ST ZONE C A L L S O L V E ( X B , N E B , Q B , A A B S , 0 2 C B S , 0 2 A B S ) C BOUNDARY C O N D I T I O N FOR T H E 2nd ZONE O 2 C 0 = S P L I T / ( l . + S P L I T ) * 0 2 C B S ( N E B + 1 ) + 1 / ( 1 . + S P L I T ) *2100 0 0 . C S O L V E 02 B A L A N C E E Q U A T I O N IN 2nd ZONE C A L L S O L V E ( X T , N E T , Q , A A T S , 0 2 C T S , 0 2 A T S ) C O U T P U T S E C T I O N DO 60 1 = 1 , N E B + 1 60 W R I T E ( 8 , * ) X B ( I ) , 0 2 C B S ( I ) , 0 2 A B S ( I ) Appendix C Computer Programs for 02 Model and NO/N20 Model 169 DO 80 I = 1 , N E T + 1 80 W R I T E ( 8 , * ) X T ( I ) , 0 2 C T S ( I ) , 0 2 A T S ( I ) END C * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * C * C F I N I T E E L E M E N T D I F F E R E N T I A L E Q U A T I A L E Q U A T I O N S O L V E R * C * C * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * S U B R O U T I N E S O L V E ( X , N E , Q , A A S , 0 2 C S , 0 2 A S ) D O U B L E P R E C I S I O N A A , A A S ( 3 0 ) , A C , A C S ( 3 0 ) D O U B L E P R E C I S I O N C T ( 1 2 0 ) , D X , I N D X ( 1 2 0 ) D O U B L E P R E C I S I O N E P S A / E P S A S ( 3 0 ) , E P S C , E P S C S ( 3 0 ) , E R R , E R R O R D O U B L E P R E C I S I O N G P ( 3 ) D O U B L E P R E C I S I O N O 2 C S ( 3 0 ) , O 2 A S ( 3 0 ) D O U B L E P R E C I S I O N 0 2 V O A , 0 2 V O A S ( 3 0 ) , 0 2 V 0 C , 0 2 V 0 C S ( 3 0 ) D O U B L E P R E C I S I O N P A , P A S ( 3 0 ) D O U B L E P R E C I S I O N P H I ( 2 ) , P H I X ( 2 ) , Q D O U B L E P R E C I S I O N S U B , S J ( 1 2 0 , 1 2 0 ) , S F ( 1 2 0 ) D O U B L E P R E C I S I O N V A C 1 , V A C 2 , V A C 3 , V A C 4 D O U B L E P R E C I S I O N V C C 1 , V C C 2 , V C C 3 D O U B L E P R E C I S I O N V O C 1 , V O C 2 , V O C 3 , V O C 4 , W ( 3 ) , X ( 3 0 ) D O U B L E P R E C I S I O N A T , K C O M , O 2 C 0 , V C H A R C O M M O N / R l / A T , K C O M , O 2 C 0 , V C H A R D A T A G P ( 1 ) / 0 . 1 1 2 7 0 1 6 6 5 4 / , G P ( 2 ) / 0 . 5 / , G P ( 3 ) / 0 . 8 8 7 2 9 8 3 3 4 6 / / , W ( 1 ) / 0 . 2 7 7 7 7 7 7 7 7 8 / , W ( 2 ) / 0 . 4 4 4 4 4 4 4 4 4 4 / / , W ( 3 ) / 0 . 2 7 7 7 7 7 7 7 7 8 / C N : N U M B E R O F NODAL P O I N T S * N=NE+1 C E P S C S , E P S A S : V O I D A G E I N T H E CORE (C) AND T H E A N N U L U S ( A ) C 0 2 V O C , 0 2 V O A : 02 CONSUMPTION R A T E BY V O L A T I L E I N C O R E ( C ) AND A N N U L U S ( A ) C 0 . 0 8 2 * T * 1 . 0 6 C O N V E R T U N I T FOR T H E C A L C U L A T I O N B A S E D ON ppm DO 14 1 = 1 , N R E A D ( 5 , * ) X ( I ) , E P S C S ( I ) , E P S A S ( I ) , A A S ( I ) , P A S ( I ) R E A D ( 7 , * ) X ( I ) , 0 2 V O C S ( I ) , 0 2 V O A S ( I ) 0 2 V O C S ( I ) = 0 2 V 0 C S ( I ) * 0 . 0 8 2 * T * 1 . 0 E 6 14 0 2 V 0 A S ( I ) = 0 2 V O A S ( I ) * 0 . 0 8 2 * T * 1 . 0 E 6 DO 11 1 = 1 , N A C S ( I ) = A T - A A S ( I ) 11 C T ( I ) = 1 . KK=0 12 KK=KK+1 P R I N T * , ' K K = ' , K K DO 13 1 = 1 , N S F ( I ) = 0 . Appendix C Computer Programs for 02 Model and NO/N20 Model 170 DO 13 J = 1 , N 13 S J ( I , J ) = 0 . DO 100 1 = 1 , N - l D X = 1 . / F L O A T ( N - l ) DO 100 J = l , 3 C A L L T F U N C T ( D X , G P ( J ) , P H I , P H I X ) A A = 0 . A C = 0 . E P S A = 0 . E P S C = 0 . P A = 0 . O 2 C = 0 . O 2 C X = 0 . O 2 V O A = 0 . O 2 V O C = 0 . DO 90 L = l , 2 L 1 = I + ( L - 1 ) A A = A A + A A S ( L l ) * P H I ( L ) A C = A C + A C S ( L l ) * P H I ( L ) E P S A = E P S A + E P S A S ( L l ) * P H I ( L ) E P S C = E P S C + E P S C S ( L l ) * P H I ( L ) P A = P A + P A S ( L l ) * P H I ( L ) 0 2 V O A = 0 2 V O A + 0 2 V O A S ( L l ) * P H I ( L ) 0 2 V 0 C = 0 2 P M C + 0 2 V 0 C S ( L l ) * P H I ( L ) 0 2 C = 0 2 C + C T ( L 1 ) * P H I ( L ) 90 0 2 C X = 0 2 C X + C T ( L 1 ) * P H I X ( L ) S U B = 1 . / ( P A * K C A + A A * A V * ( 1 . - E P S A ) * K C O M * V C H A R ) S U B V C A = - A A * A V * ( 1 . - E P S A ) * K C O M / / ( P A * K C A + A A * A V * ( 1 . - E P S A ) * K C O M * V C H A R ) * * 2 V O C l = l . / Q * ( A C * A V * ( 1 - E P S C ) * K C O M ) V O C 2 = l . / Q * ( P A * K C A * A A * A V * ( 1 - E P S A ) * K C O M ) V O C 3 = l . / Q * ( P A * K C A ) * ( 1 . / O 2 C 0 ) * 0 2 V O A V O C 4 = l . / Q * ( l . / O 2 C 0 ) * 0 2 V O C DO 100 L = l , 2 L 1 = I + ( L - 1 ) S F ( L 1 ) = S F ( L 1 ) - W ( J ) * D X * ( 0 2 C X + V 0 C 1 * V C H A R * 0 2 C + V 0 C 2 * V C H A R / * S U B * 0 2 C + V O C 3 * S U B + V O C 4 ) * P H I ( L ) DO 100 M = l , 2 M 1 = I + ( M - 1 ) 100 S J ( L l , M 1 ) = S J ( L l , M 1 ) + D X * W ( J ) * ( P H I X ( M ) + V 0 C 1 * V C H A R * P H I ( M ) + / V O C 2 * V C H A R * S U B * P H I ( M ) ) * P H I ( L ) C BOUNDARY C O N D I T I O N Appendix C Computer Programs for 02 Model andNO/N20 Model 171S F ( 1 ) = 0 . DO 150 I L = 1 , N 150 S J ( 1 , I L ) = 0 . S J ( 1 , 1 ) = 1 . C R E V E R S E M A T R I X C A L L L U D C M P ( S J , N , 1 2 0 , I N D X , D ) C A L L L U B K S B ( S J , N , 1 2 0 , I N D X , S F ) E R R = 0 . DO 38 1 = 1 , N 38 E R R = E R R + S F ( I ) * * 2 E R R O R = D S Q R T ( E R R / N ) P R I N T * , ' P R I N T * , ' E R R I N C A L C U L A T I N G M I D D L E = ' , E R R O R P R I N T * , ' I F ( E R R O R . L T . 1 . 0 E - 6 ) T H E N P R I N T * , ' E Q U A T I O N S S O L V E D ' E L S E I F ( K K . E Q . 1 0 ) THEN P R I N T * , ' C H E C K P R O G R A M ' , ' K K = ' , K K P A U S E E L S E DO 35 1 = 1 , N 35 C T ( I ) = C T ( I ) + S F ( I ) GO TO 12 E N D I F C A S S I G N T H E S O L U T I O N TO T H E CORRESPONDING ARRAY AND C A L C U L A T E C T H E C O N C E N T R A T I O N IN ANNULUS DO 55 1 = 1 , N 0 2 C S ( I ) = C T ( I ) * O 2 C 0 55 0 2 A S ( I ) = ( 0 2 C S ( I ) * P A S ( I ) * K C A - 0 2 V O A S ( I ) ) / / ( P A S ( I ) * K C A + A A S ( I ) * A V * ( l - E P S A S ( I ) ) * V C H A R * K C O M ) R E T U R N END C * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * C * S U B R O U T I N E T F U N C T , L I N E A R I N T E R P O L A T I O N F U N C T I O N C * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * S U B R O U T I N E T F U N C T ( D X , G P , P H I , P H I X ) D O U B L E P R E C I S I O N P H I ( 2 ) , P H I X ( 2 ) , G P , D X P H I ( 1 ) = 1 . 0 - G P P H I ( 2 ) = G P P H I X ( 1 ) = - 1 . / D X P H I X ( 2 ) = 1 . / D X R E T U R N END Appendix C Computer Programs for 02 Model and NO/N20 Model 172 c**************************************************************** T H E F O L L O W I N G TWO S U B R O U T I N E S C A L C U L A T E T H E I N V E R S E OF M A T R I X C * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * S U B R O U T I N E L U B K S B ( A , N , N P , I N D X , B ) D O U B L E P R E C I S I O N A ( N P , N P ) , I N D X ( N ) , B ( N ) 11=0 DO 12 1=1 , N L L = I N D X ( I ) S U M = B ( L L ) B ( L L ) = B ( I ) I F ( I I . N E . O ) T H E N DO 11 J = I I , I - 1 11 S U M = S U M - A ( I , J ) * B ( J ) E L S E I F ( S U M . N E . O ) T H E N 11=1 E N D I F 12 B ( I ) = S U M DO 14 I = N , 1 , - 1 S U M = B ( I ) DO 13 J = I + 1 , N 13 S U M = S U M - A ( I , J ) * B ( J ) 14 B ( I ) = S U M / A ( I , I ) R E T U R N END C * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * S U B R O U T I N E L U D C M P ( A , N , N P , I N D X , D ) P A R A M E T E R ( N M A X = 2 8 0 , T I N Y = 1 . O E - 2 0 ) D O U B L E P R E C I S I O N A ( N P , N P ) , I N D X ( N ) , W ( N M A X ) D = l . 0 , DO 12 I = 1 , N A A M A X = 0 . DO 11 J = 1 , N 11 I F ( A B S ( A ( I , J ) ) • G T . A A M A X ) A A M A X = A B S ( A ( I , J ) ) I F ( A A M A X . E Q . 0 . ) P A U S E ' S I N G U L A R M A T R I X ' 12 W ( I ) = 1 . / A A M A X DO 19 J = 1 , N DO 14 1 = 1 , J - l S U M = A ( I , J ) DO 13 K = 1 , I - 1 13 S U M = S U M - A ( I , K ) * A ( K , J ) 14 A ( I , J ) = S U M A A M A X = 0 . DO 16 I = J , N S U M = A ( I , J ) DO 15 K = 1 , J - 1 ,1-5 S U M = S U M - A ( I , K) * A ( K , J ) A ( I , J ) = S U M D U M = W ( I ) * A B S ( S U M ) I F ( D U M . G E . AAMAX) T H E N I M A X = I AAMAX=DUM E N D I F 16 C O N T I N U E I F ( J . N E . I M A X ) T H E N DO 17 K = 1 , N D U M = A ( I M A X , K ) A ( I M A X , K ) = A ( J , K ) Appendix C Computer Programs for 02 Model and NO/N20 Model J 73 17 A ( J , K ) = D U M D=-D W ( I M A X ) = W ( J ) E N D I F I N D X ( J ) = I M A X I F ( A ( J , J ) . E Q . 0 . ) A ( J , J ) = T I N Y I F ( J . N E . N ) T H E N D U M = 1 . / A ( J , J ) DO 18 I = J + 1 , N 18 A ( I , J ) = A ( I , J ) * DUM E N D I F 19 C O N T I N U E R E T U R N END Appendix C Computer Programs for 02 Model and NO/N20 Model 774 i C NO AND N20 MODEL C ANNOTATIONS FOR VARIABLES: ccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc C AAS,ACS:AREA OF ANNULUS AND CORE (m2) C C AT:CROSS SECTIONAL AREA (m2) C C AV:EXTERIOR SURFACE AREA PER UNIT VOLUME OF CHAR IN RISER(1/m) C C DS:SAUTER MEAN DIAMETER FOR CHARS IN RISER (m) C C EPSABS,EPSATS: ANNULUS VOIDAGE IN THE 1st AND 2nd ZONES C C EPSCBS,EPSCTS: CORE VOIDAGE IN THE 1st AND 2nd ZONES C C KCA:GAS TRANSFER COEFFICIENT BETWEEN CORE AND ANNULUS (m/s) C C Kl TO K15 :REACTION RATE CONSTANT (UNIT SEE THESIS TABLE 3.8) C C LENGTH:LENGTH FOR REACTOR CROSS-SECTION C C NEB,NET:NUMBER OF ELEMENTS EMPLOYED FOR 1st AND 2nd ZONES C C ($$)ABS,($$)ATS:$$ CONCENTRATION IN ANNULUS IN lst(B) AND 2nd(T) ZONE C C ($$)CBS,($$)CTS:$$ CONCENTRATION IN CORE IN lst(B) AND 2nd(T) ZONE C C $$$ DENOTE CO,NH3,HCN,NO AND; N20 UNIT: (ppm) C C PAS:THE PERIMETER OF ANNULUS IN CONTACT WITH CORE (m) C C QB,Q:VOLUMETRIC FLOW RATE IN THE PRIMARY AND SECONDARY ZONE (m3/s) C C SECPOS:SECONDARY AIR PORT HEIGHT (m) C C SPLIT:PRIMARY TO SECONDARY AIR RATIO C C T:TEMPERATURE (K) C C UG:SUPERFICIAL VELOCITY OF GAS IN THE 2nd ZONE (m/s) C C VCHAR:THE VOLUME FRACTION OF CHAR IN SOLIDS C C VCAO:THE VOLUME FRACTIONS OF LIMESTONE IN SOLIDS C C XB,XT:COORDINATES FOR THE GRID IN THE lst(B) AND 2nd(T) Zone C ccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc C FORMAT INPUT FILES MUST ABIDE BY: cccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc c c C 02.DAT:CONTAINS THE 02 CONCENTRATIONS IN CORE AND ANNULUS THROUGHOUT C C THE RISER; FREE FORMAT; BUT MUST FOLLOW THE ORDER AS BELOW C C (X,02CS,02AS), WHERE X IS THE COORDINATES OF NODAL POINTS C C C C CO.DAT:CONTAINS THE CO CONCENTRATIONS IN CORE AND ANNULUS THROUGHOUT C C THE RISER; FREE FORMAT; BUT MUST FOLLOW THE ORDER AS BELOW C C (X,COCS,COAS), WHERE X IS THE COORDINATES OF NODAL POINTS C C C C NVOL.DAT:CONTAINS THE NH3 AND HCN RELEASE RATES IN CORE AND ANNULUS C C THE RISER; FREE FORMAT; BUT MUST FOLLOW THE ORDER AS BELOW C C (X,VNH3CS,VNH3AS,VHCNCS,VHCNAS) , WHERE X IS THE COORDINATES C C OF NODAL POINTS C C C C HYDRO.DAT:CONTAINS THE HYDRODYNAMIC INFORMATION FROM SENIOR AND BRERETON C C . MODEL,INCLUDING THE VOIDAGE IN CORE AND ANNULUS THROUGHOUT THE C C RISER; FREE FORMAT; BUT MUST FLOLLOW THE ORDER AS BELOW C C (NE,9,X,EPSCS,EPSAS,AAS,PAS), WHERE X IS THE COORDINATES OF C C NODAL POINTS C C C cccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccccc DOUBLEPRECISION AABS(30),AATS(30) DOUBLEPRECISION AVNOB(30),AVNOT(30) DOUBLEPRECISION AVN2OB(30),AVN2OT(30) DOUBLEPRECISION CB1,CB2,CB3,CB4,CB5,CB6,CB7,CHAB DOUBLEPRECISION COABS(30),COATS(30) Appendix C Computer Programs for 02 Model and NO/N20 Model 175 D O U B L E P R E C I S I O N D O U B L E P R E C I S I O N D O U B L E P R E C I S I O N D O U B L E P R E C I S I O N D O U B L E P R E C I S I O N D O U B L E P R E C I S I O N D O U B L E P R E C I S I O N D O U B L E P R E C I S I O N D O U B L E P R E C I S I O N D O U B L E P R E C I S I O N D O U B L E P R E C I S I O N D O U B L E P R E C I S I O N D O U B L E P R E C I S I O N D O U B L E P R E C I S I O N D O U B L E P R E C I S I O N D O U B L E P R E C I S I O N D O U B L E P R E C I S I O N D O U B L E P R E C I S I O N D O U B L E P R E C I S I O N C O C B S ( 3 0 ) , C O C T S ( 3 0 ) , C T ( 1 2 0 ) E P S B A S ( 3 0 ) , E P S B C S ( 3 0 ) E P S T A S ( 3 0 ) , E P S T C S ( 3 0 ) ERRNCC H C N A B S ( 3 0 ) , H C N A T S ( 3 0 ) H C N C B S ( 3 0 ) , H C N C T S ( 3 0 ) I T G 0 2 B , I T G 0 2 T I T G N O B , I T G N O T L E N G T H N H 3 A B S ( 3 0 ) , N H 3 A T S ( 3 0 ) N H 3 C B S ( 3 0 ) , N H 3 C T S ( 3 0 ) N O A B S ( 3 0 ) , N O A T S ( 3 0 ) N O C B S ( 3 0 ) , N O C T S ( 3 0 ) N 2 O A B S ( 3 0 ) , N 2 O A T S ( 3 0 ) N 2 O C B S ( 3 0 ) , N 2 O C T S ( 3 0 ) O 2 A B S ( 3 0 ) , O 2 A T S ( 3 0 ) O 2 C B S ( 3 0 ) , O 2 C T S ( 3 0 ) X B ( 3 0 ) , X T ( 3 0 ) V N C N A B ( 3 0 ) , V H C N C B ( 3 0 ) , V N H 3 A B ( 3 0 ) , V N H 3 C B ( 3 0 ) D O U B L E P R E C I S I O N A T , A V D O U B L E P R E C I S I O N K l , K 2 , K 3 , K 4 , K 5 , K 6 , K 7 , K 8 , K 9 K 1 0 , K 1 1 , K 1 2 , K 1 3 , K 1 4 , K 1 5 , K C A , K N C C , N C C 1 D O U B L E P R E C I S I O N R O F U E L , R O C A O , R O P , T , V C A O D O U B L E P R E C I S I O N N H 3 C 0 , H C N C O , N O C O , N 2 0 C 0 , Y K K D O U B L E P R E C I S I O N VCHAR D O U B L E P R E C I S I O N D O U B L E P R E C I S I O N C O M M O N / R l / A T , A V C O M M O N / R 2 / K 1 , K 2 , K 3 , K 4 , K 5 , K 6 , K 7 , K 8 , K 9 C O M M O N / R 3 / K 1 0 , K 1 1 , K 1 2 , K 1 3 , K 1 4 , K 1 5 , K C A , K C O M M O N / R 4 / N C C , N C C 1 C O M M O N / R 5 / R O F U E L , R O C A O , R O P , T , V C A O C O M M O N / R 6 / N H 3 C 0 , H C N C O , N O C O , N 2 O C 0 C O M M O N / R 7 / V C H A R O P E N ( 6 , F I L E = ' C : 0 2 . D A T ' ) O P E N ( 7 , F I L E = ' C : C O . D A T ' ) O P E N ( 8 , F I L E = ' C : N V O L . D A T ' ) O P E N ( 9 , F I L E = ' C : H Y D R O . D A T ' ) C MANUAL I N P U T C * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * NEB=25 NET=2 5 N C C = 0 . 0 1 8 1 U G = 7 . 0 S E C P O S = 3 . 4 P S = 1 . 6 T = 1 1 5 5 . K C A = 0 . 1 V C A O = 0 . 4 V C H A R = 0 . 0 5 6 5 D P = 2 . 4 5 8 8 E - 4 ROCAO=18 0 0 . R O F U E L = 1 8 0 0 . L E N G T H = . 1 5 2 C * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Appendix C Computer Programs for 02 Model and NO/N20 Model 176 A T = L E N G T H * * 2 . A V = 6 . / D P Q = U G * A T Q B = Q * P S / ( l . + P S ) K l = 8 . 0 1 3 0 7 E 8 * E X P ( - 2 9 29 0 . / T ) K2 = 6 . 8 859 9 E 9 * E X P ( - 2 7433 . / T ) K 3 = 2 . 5 9 7 D 1 3 * E X P ( - 5 0 7 1 9 . / T ) K4 = 5 . 7 9 7 E 8 * E X P ( - 3 4 7 6 5 . / T ) K 5 = 2 . 2 3 4 E 9 * E X P ( - 2 9 3 7 8 . / T ) K 6 = 7 . 9 8 3 E 8 * E X P ( - 3 7 6 3 9 . / T ) K 7 = l . 6 1 0 E 1 * E X P ( - 1 4 9 6 0 . / T ) K 8 = l . 1 7 2 E 1 * E X P ( - 1 2 6 3 2 . / T ) K9 = l . 7 0*5 9 5 . * T * D E X P ( - 1 5 0 7 3 2 . D O / ( 8 . 3 1 4 D O * T ) ) K 1 0 = 2 . 4 D 6 * ( 1 / ( 1 . 0 D 6 * . 0 8 2 * T ) ) K l l = 2 5 8 . * K 9 K12=172 3 . * T * E X P ( - 1 7 8 0 0 / T ) K 1 3 = 4 1 . 6 * T * E X P ( - 1 0 0 0 0 / T ) * 1 . 7 3 K 1 4 = 1 . 5 2 E 2 * E X P ( - 8 9 2 0 / T ) K 1 5 = 1 . 3 * E X P ( - 4 0 0 0 . / T ) C I N I T I A T E N C C I T E R A T I O N S I N C C = 0 N C C 1 = 0 . 0 0 9 6 C N C C I T E R A T I O N COUNTER 117 I N C C = I N C C + 1 C I N C A S E TOO MANY I T E R A T I O N S I F ( I N C C . G E . 30) T H E N P R I N T * , ' W R O N G WITH NCC I T E R A T I O N S ' P A U S E E N D I F C BOUNDARY C O N D I T I O N S FOR 1 s t ZONE N H 3 C 0 = 0 . H C N C 0 = 0 . N O C 0 = 0 . N 2 O C 0 = 0 . C S O L V E D I F F E R E N T I A L EQUATIONS FOR 1 s t ZONE C A L L S O L V E ( N E B , X B , Q B , 0 2 C B S , 0 2 A B S , C O C B S , C O A B S , E P S C B S , E P S A B S / , N H 3 C B S , N H 3 A B S , H C N C B S , H C N A B S , N O C B S , N O A B S , N 2 0 C B S , N 2 0 A B S / , A A B S , I T G N O B , I T G 0 2 B ) C BOUNDARY C O N D I T I O N S FOR 2 n d ZONE N H 3 C O = Q B / Q * N H 3 B C S ( N E B + 1 ) H C N C O = Q B / Q * H C N B C S ( N E B + 1 ) N O C 0 = Q B / Q * NOBCS(NEB+1) N 2 0 C O = Q B / Q * N 2 0 B C S ( N E B + 1 ) Appendix C Computer Programs for 02 Model and NO/N20 Model 177 C A L L S O L V E ( N E T , X T , Q , 0 2 C T S , 0 2 A T S , C O C T S , C O A T S , E P S C T S , E P S A T S / , N H 3 C T S , N H 3 A T S , H C N C T S , H C N A T S , N O C T S , N O A T S , N 2 0 C T S , N 2 0 A T S / , A A T S , I T G N O T , I T G 0 2 T ) C SUMMING UP I N T E G R A T I O N S I N V O L V E D IN T H E S I S EQN ( 4 . 1 9 ) I T G 0 2 = I T G 0 2 B + I T G 0 2 T I T G N O = I T G N O B + I T G N O T C C A L C U L A T E NEW N - T O - C R A T I O I N CHAR I F NOT Y E T C C O N V E R G E D S T A R T ANOTHER I T E R A T I O N U S I N G NEW NCC N C C 2 = ( N C C * I T G 0 2 ) / ( I T G 0 2 + I T G N O ) E R R N C C = A B S ( N C C 2 - N C C 1 ) / N C C 1 I F ( E R R N C C . G T . 0 . 0 1 ) T H E N NCC1=NCC2 GO T O 117 E N D I F C C A L C U L A T E T H E C R O S S - S E C T I O N A L A V E R A G E C O N C E N T R A T I O N DO 1200 I = 1 , N E B + 1 A V N H 3 B ( I ) = N H 3 B C S ( I ) * ( A T - A A B S ( I ) ) / A T + N H 3 B A S ( I ) * A A B S ( I ) / A T A V H C N B ( I ) = H C N B C S ( I ) * ( A T - A A B S ( I ) ) / A T + H C N B A S ( I ) * A A B S ( I ) / A T A V N O B ( I ) = N O B C S ( I ) * ( A T - A A B S ( I ) ) / A T + N O B A S ( I ) * A A B S ( I ) / A T 12 00 A V N 2 0 B ( I ) = N 2 0 B C S ( I ) * ( A T - A A B S ( I ) ) / A T + N 2 0 B A S ( I ) * A A B S ( I ) / A T DO 1 3 0 0 I = 1 , N E T + 1 A V N H 3 T ( I ) = N H 3 T C S ( I ) * ( A T - A A T S A V H C N T ( I ) = H C N T C S ( I ) * ( A T - A A T S A V N O T ( I ) = N O T C S ( I ) * ( A T - A A T S 1300 A V N 2 0 T ( I ) = N 2 0 T C S ( I ) * ( A T - A A T S ( I ) ) / A T + N H 3 T A S ( I ) * A A T S ( I ) / A T ( I ) ) / A T + H C N T A S ( I ) * A A T S ( I ) / A T ( I ) ) / A T + N O T A S ( I ) * A A T S ( I ) / A T ( I ) ) / A T + N 2 0 T A S ( I ) * A A T S ( I ) / A T C O U T P U T S E C T I O N BELOW DO 1 5 0 0 I = 1 , N E B + 1 W R I T E ( 1 4 , 3 3 ) X B ( I ) , N O B C S ( I ) , N O B A S ( I ) , N 2 0 B C S ( I ) , N 2 0 B A S ( I ) / , A V N O B ( I ) , A V N 2 0 B ( I ) 1500 W R I T E ( 1 4 , 3 3 ) X B ( I ) , N H 3 B C S ( I ) , N H 3 B A S ( I ) , H C N B C S ( I ) , H C N B A S ( I ) / , A V N H 3 B ( I ) , A V H C N B ( I ) DO 1 6 0 0 I = 1 , N E T + 1 W R I T E ( 1 4 , 3.3) X T ( I ) , NOTCS (I ) , NOTAS (I) , N 2 0 T C S ( I ) , N 2 0 T A S ( I ) / , A V N O T ( I ) , A V N 2 0 T ( I ) 1600 W R I T E ( 1 4 , 3 3 ) X T ( I ) , N H 3 T C S ( I ) , N H 3 T A S ( I ) , H C N T C S ( I ) , H C N T A S ( I ) / , A V N H 3 T ( I ) , A V H C N T ( I ) 3 3 F O R M A T ( F 5 . 3 , ' ' , 6 ( F 5 . 1 , ' ' ) ) END C * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * C D I F F E R E N T I A L E Q U A T I O N S O L V E R . * C * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * S U B R O U T I N E S O L V E ( N E , X , Q , 0 2 C S , 0 2 A S , C O C S , C O A S , E P S C S , E P S A S , / , N H 3 C S , N H 3 A S , H C N C S , H C N A S , N O C S , N O A S , N 2 0 C S , N 2 0 A S / , I T G N O , I T G 0 2 ) Appendix C Computer Programs for 02 Model and NO/N20 Model 178 D O U B L E P R E C I S I O N A ( 5 0 ) D O U B L E P R E C I S I O N A A , A A S ( 3 0 ) , A C , A C S ( 3 0 ) D O U B L E P R E C I S I O N C A 1 , C A 2 , C A 3 , C A 4 , C A 5 , C A 6 D O U B L E P R E C I S I O N C D 1 , C D 2 , C D 3 ( C D 4 , C D 5 , C D 6 , C D 7 D O U B L E P R E C I S I O N C O A , C O A S ( 3 0 ) , C O C , C O C S ( 3 0 ) D O U B L E P R E C I S I O N C T ( 1 2 0 ) , D X , I N D X ( 1 2 0 ) D O U B L E P R E C I S I O N E P S A , E P S A S ( 3 0 ) , E P S C , E P S C S ( 3 0 ) , E P S A S ( 3 0 ) , E R R O R D O U B L E P R E C I S I O N G P ( 3 ) , 0 2 A , 0 2 A S ( 3 0 ) , 0 2 C , 0 2 C S ( 3 0 ) D O U B L E P R E C I S I O N H C N A S ( 3 0 ) , H C N C , H C N C S ( 3 0 ) D O U B L E P R E C I S I O N I T G 0 2 , I T G N O D O U B L E P R E C I S I O N N H 3 A S ( 3 0 ) , N H 3 C , N H 3 C S ( 3 0 ) D O U B L E P R E C I S I O N N O A S ( 3 0 ) , N O C , N O C S ( 3 0 ) D O U B L E P R E C I S I O N N 2 O A S ( 3 0 ) , N 2 0 C , N 2 0 C S ( 3 0 ) D O U B L E P R E C I S I O N P A , P A S ( 3 0 ) , P H I ( 2 ) , P H I X ( 2 ) , Q D O U B L E P R E C I S I O N S J ( 1 2 0 , 1 2 0 ) , S F ( 1 2 0 ) D O U B L E P R E C I S I O N V N H 3 A S ( 3 0 ) , V N H 3 C S ( 3 0 ) , V H C N A S ( 3 0 ) , V H C N C S ( 3 0 ) D O U B L E P R E C I S I O N V H C N A , V H C N C , V N H 3 A , V N H 3 C , W ( 3 ) D O U B L E P R E C I S I O N A T , A V D O U B L E P R E C I S I O N K l , K 2 , K 3 , K 4 , K 5 , K 6 , K 7 , K 8 , K 9 D O U B L E P R E C I S I O N K 1 0 , K l l , K 1 2 , K 1 3 , K 1 4 , K 1 5 , K C A , K D O U B L E P R E C I S I O N N C C , N C C 1 D O U B L E P R E C I S I O N R O F U E L , R O C A O , T , V C A O D O U B L E P R E C I S I O N N H 3 C 0 , H C N C O , N O C O , N 2 O C 0 D O U B L E P R E C I S I O N VCHAR C O M M O N / R l / A T , A V C O M M O N / R 2 / K l , K 2 , K 3 , K 4 , K 5 , K 6 , K 7 , K 8 , K 9 C O M M O N / R 3 / K 1 0 , K 1 1 , K 1 2 , K 1 3 , K 1 4 , K 1 5 , K C A , K C O M M O N / R 4 / N C C , N C C 1 C O M M O N / R 5 / R O F U E L , R O C A O , T , V C A O C O M M O N / R 6 / N H 3 C O , H C N C O , N O C O , N 2 O C 0 C O M M O N / R 7 / V C H A R D A T A G P ( l ) / 0 . 1 1 2 7 0 1 6 6 5 4 / , G P ( 2 ) / 0 . 5 / , G P ( 3 ) / 0 . 8 8 7 2 9 8 3 3 4 6 / / , W ( l ) / 0 . 2 7 7 7 7 7 7 7 7 8 / , W ( 2 ) / 0 . 4 4 4 4 4 4 4 4 4 4 / / , W ( 3 ) / 0 . 2 7 7 7 7 7 7 7 7 8 / C IMPORT I N P U T DATA C A L L I N P U T 1 ( N E , 6 , X , 0 2 C S , 0 2 A S ) C A L L I N P U T 1 ( N E , 7 , X , C 0 C S , C 0 A S ) C A L L I N P U T 2 ( N E , 8 , X , VNH3.CS, V N H 3 A S , V H C N C S , VHCNAS) C A L L I N P U T 2 ( N E , 9 , X , E P S C S , E P S A S , A A S , P A S ) C N E Q : N U M B E R OF D I F F E R E N T I A L E Q U A T I O N S NEQ=4 N=NE+1 N1=N*NEQ N 2 = N E * N E Q C U N I T C O N V E R S I O N S FOR T H E C A L C U L A T I O N S B A S E D ON ppm DO 15 1 = 1 , N I A C S ( I ) = A T - A A S ( I ) V N H 3 A S ( I ) = V N H 3 A S ( I ) * 1 . 0 E 6 * 0 . 0 8 2 * T V N H 3 C S ( I ) = V N H 3 C S ( I ) * 1 . 0 E 6 * 0 . 0 8 2 * T Appendix C Computer Programs for 02 Model and NO/N20 Model 179 V H C N A S ( I ) = V H C N A S ( I ) * 1 . 0 E 6 * 0 . 0 8 2 * T V H C N C S ( I ) = V H C N C S ( I ) * 1 . 0 E 6 * 0 . 0 8 2 * T 15 C T ( I ) = 0 . C T ( 1 ) = N H 3 C 0 C T ( 2 ) = H C N C 0 C T ( 3 ) = N O C 0 C T ( 4 ) = N 2 0 C O KK=0 160 KK=KK+1 DO 18 1 = 1 , N I S F ( I ) = 0 . DO 18 J = 1 , N 1 18 S J ( I , J ) = 0 . DO 1 0 0 I = 1 , N 2 , N E Q D X = 1 . / N E DO 100 J = l , 3 C A L L T F U N C T ( D X , G P ( J ) , P H I , P H I X ) A A = 0 . A C = 0 . C O A = 0 . C O C = 0 . E P S A = 0 . E P S C = 0 . O 2 A = 0 . O 2 C = 0 . P A = 0 . V H C N A = 0 . V H C N C = 0 . V N H 3 A = 0 . V N H 3 C = 0 . N H 3 C = 0 . N H 3 C X = 0 . H C N C = 0 . H C N C X = 0 . N O C = 0 . N O C X = 0 . N 2 O C = 0 . N 2 O C X = 0 . DO 190 L = l , 2 L 1 = I + ( L - 1 ) * N E Q L 2 = L 1 + 1 L 3 = L l + 2 L 4 = L l + 3 L I = L 4 / 4 A A = A A + A A S ( L I ) * P H I ( L ) Appendix C Computer Programs for 02 Model andNO/N20 Model 180 A C = A C + A C S ( L I ) * P H I ( L ) E P S A = E P S A + E P S A S ( L I ) * P H I ( L ) E P S C = E P S C + E P S C S ( L I ) * P H I ( L ) 0 2 A = 0 2 A + 0 2 A S ( L I ) * P H I ( L ) 0 2 C = 0 2 C + 0 2 C S ( L I ) * P H I ( L ) P A = P A + P A S ( L I ) * P H I ( L ) COA= C O A + C O A S ( L I ) * P H I ( L ) COC= C O C + C O C S ( L I ) * P H I ( L ) VHCNA=VHCNA+ V H C N A S ( L I ) * P H I ( L ) V H C N C = V H C N C + V H C N C S ( L I ) * P H I ( L ) VNH3A=VNH3A+ V N H 3 A S ( L I ) * P H I ( L ) VNH3C=VNH3C+ V N H 3 C S ( L I ) * P H I ( L ) N H 3 C = N H 3 C + C T ( L l ) * P H I ( L ) N H 3 C X = N H 3 C X + C T ( L 1 ) * P H I X ( L ) H C N C = H C N C + C T ( L 2 ) * P H I ( L ) H C N C X = H C N C X + C T ( L 2 ) * P H I X ( L ) N O C = N O C + C T ( L 3 ) * P H I ( L ) N O C X = N O C X + C T ( L 3 ) * P H I X ( L ) N 2 0 C = N 2 0 C + C T ( L 4 ) * P H I ( L ) 190 N 2 0 C X = N 2 0 C X + C T ( L 4 ) * P H I X ( L ) C A 1 = P A * K C A C A 2 = A A * E P S A / C A 1 C A 3 = A A * ( 1 - E P S A ) * V C A O / C A l C A 4 = A A * ( 1 - E P S A ) * V C H A R / C A 1 C A 5 = A A * V N H 3 A / C A 1 C A 6 = A A * V H C N A / C A 1 CD1=Q C D 2 = - P A * K C A / C D 1 C D 3 = - A C * E P S C / C D 1 C D 4 = - A C * ( 1 - E P S C ) * V C A O / C D l C D 5 = - A C * ( 1 - E P S C ) * V C H A R / C D 1 C D 6 = - A C * V N H 3 C / C D 1 C D 7 = - A C * V H C N C / C D 1 A ( l ) = C A 3 * K 1 0 * O 2 A + C A 2 * K 4 * O 2 A A ( 2 ) = 1 . + C A 3 * K 1 4 * C O A + C A 4 * K 1 2 + C A 4 * K 1 1 * N C C 1 * A V A ( 3 ) = C A 2 * K 7 * 0 2 A A ( 4 ) = C A 2 * K 1 A ( 5 ) = C A 4 * K 9 * N C C 1 * A V * 0 2 A A ( 6 ) = C A 4 * K 1 1 * N C C 1 * A V A ( 7 ) = . 5 * C A 2 * K 3 * 0 2 A A ( 8 ) = . 5 * C A 2 * K 6 * 0 2 A A ( 9 ) = 1 . + C A 3 * K 1 5 * R O C A O + C A 4 * K 1 3 * R O F U E L + C A 2 * ( K 2 + K 1 ) A ( 1 0 ) = 1 . 5 + 1 . 5 * C A 2 * K 3 * O 2 A + 2 . 5 * A ( l ) A ( 2 6 ) = l . / ( l . + C A 2 * K 6 * 0 2 A + C A 2 * K 7 * 0 2 A + C A 2 * K 8 * 0 2 A ) A ( l l ) = N O C - 1 . 5 * N H 3 C + A ( 3 ) * A ( 2 6 ) * ( H C N C + C A 6 ) / + A ( 5 ) - 1 . 5 * C A 5 A ( 1 2 ) = - A ( 8 ) * A ( 2 6 ) * ( H C N C + C A 6 ) - N 2 0 C A ( 1 3 ) = A ( 2 ) / A ( 4 ) Appendix C Computer Programs for 02 Model and NO/N20 Model 181 A ( 1 4 ) = A ( 1 0 ) / A ( 4 ) A ( 1 5 ) = A ( 1 1 ) / A ( 4 ) A ( 1 6 ) = A ( 6 ) / A ( 9 ) A ( 1 8 ) = A ( 7 ) / A ( 9 ) A ( 1 9 ) = A ( 1 2 ) / A ( 9 ) A ( 2 0 ) = ( A ( 1 8 ) + A ( 1 4 ) ) / ( A ( 1 3 ) - A ( 1 6 ) ) A ( 2 1 ) = ( A ( 1 5 ) - A ( 1 9 ) ) / ( A ( 1 3 ) - A ( 1 6 ) ) A ( 2 2 ) = 1 . + C A 2 * K 3 * 0 2 A + A ( 1 ) A ( 2 3 ) = - 2 . / 3 . * C A 2 * K 5 A ( 2 4 ) = ( A ( 2 1 ) * A ( 2 3 ) - A ( 2 2 ) ) / ( A ( 2 0 ) * A ( 2 3 ) ) A ( 2 5 ) = ( N H 3 C + C A 5 ) / ( A ( 2 0 ) * A ( 2 3 ) ) A ( 2 6 ) = l . / ( l . + C A 2 * K 6 * 0 2 A + C A 2 * K 7 * 0 2 A + C A 2 * K 8 * 0 2 A ) A ( 2 7 ) = 1 . / ( A ( 2 ) / A ( 4 ) - A ( 6 ) / A ( 9 ) ) A ( 2 8 ) = A ( 3 ) * A ( 2 6 ) / A ( 4 ) + A ( 8 ) * A ( 2 6 ) / A ( 9 ) A ( 2 9 ) = A ( 2 7 ) * A ( 2 8 ) / A ( 2 0 ) A ( 3 0 ) = - l . 5 * A ( 2 7 ) / ( A ( 4 ) * A ( 2 0 ) ) A ( 3 1 ) = ( A ( 2 4 ) * A ( 3 0 ) - 2 . / ( A ( 2 0 ) * A ( 2 3 ) ) ) / / D S Q R T ( A ( 2 4 ) * * 2 - 4 . * A ( 2 5 ) ) A ( 3 2 ) = A ( 2 7 ) / ( A ( 2 0 ) * A ( 4 ) ) A ( 3 3 ) = A ( 2 4 ) * A ( 3 2 ) / D S Q R T ( A ( 2 4 ) * * 2 - 4 . * A ( 2 5 ) ) A ( 3 4 ) = l . / A ( 2 0 ) * A ( 2 7 ) / A ( 9 ) A ( 3 5 ) = A ( 2 4 ) * A ( 3 4 ) / D S Q R T ( A ( 2 4 ) * * 2 - 4 . * A ( 2 5 ) ) A ( 3 6 ) = - 0 . 5 * ( A ( 2 9 ) - ( A ( 2 4 ) * A ( 2 9 ) ) / / D S Q R T ( A ( 2 4 ) * * 2 - 4 . * A ( 2 5 ) ) ) A ( 3 7 ) = ( A ( 2 7 ) * A ( 2 8 ) + A ( 2 0 ) * A ( 3 6 ) ) A ( 3 8 ) = - 0 . 5 * ( A ( 3 0 ) - A ( 3 1 ) ) A ( 3 9 ) = - 1 . 5 * A ( 2 7 ) / A ( 4 ) + A ( 2 0 ) * A ( 3 8 ) A ( 4 0 ) = - 0 . 5 * ( A ( 3 2 ) - A ( 3 3 ) ) A ( 4 1 ) = ( A ( 2 7 ) / A ( 4 ) + A ( 2 0 ) * A ( 4 0 ) ) A ( 4 2 ) = - 0 . 5 * ( A ( 3 4 ) - A ( 3 5 ) ) A ( 4 3 ) = ( A ( 2 7 ) / A ( 9 ) + A ( 2 0 ) * A ( 4 2 ) ) A ( 4 4 ) = l . / A ( 9 ) * ( A ( 6 ) * A ( 3 7 ) + A ( 7 ) * A ( 3 6 ) + A ( 8 ) * A ( 2 6 ) ) A ( 4 5 ) = l . / A ( 9 ) * ( A ( 6 ) * A ( 3 9 ) + A ( 7 ) * A ( 3 8 ) ) A ( 4 6 ) = 1 . / A ( 9 ) * ( A ( 6 ) * A ( 4 1 ) + A ( 7 ) * A ( 4 0 ) ) A ( 4 7 ) = l . / A ( 9 ) * ( A ( 6 ) * A ( 4 3 ) + A ( 7 ) * A ( 4 2 ) + 1 . ) N H 3 A = - . 5 * ( A ( 2 4 ) - D S Q R T ( A ( 2 4 ) * * 2 - 4 . * A ( 2 5 ) ) ) HCNA=A(2 6 ) * ( H C N C + C A 6 ) N O A = ( A ( 2 1 ) + A ( 2 0 ) * N H 3 A ) N 2 0 A = ( A ( 1 6 ) * N O A + A ( 1 8 ) * N H 3 A - A ( 1 9 ) ) DO 100 L = l , 2 L 1 = I + ( L - 1 ) * N E Q L 2 = L 1 + 1 L 3 = L l + 2 L 4 = L l + 3 S F ( L 1 ) = S F ( L 1 ) - W ( J ) * D X * / ( N H 3 C X + C D 2 * ( N H 3 A - N H 3 C ) / - C D 3 * ( ( K 3 + K 4 ) * 0 2 C + 2 . / 3 . * K 5 * N O C ) * N H 3 C / - C D 4 * K 1 0 * O 2 C * N H 3 C / + C D 6 ) * P H I ( L ) S F ( L 2 ) = S F ( L 2 ) - W ( J ) * D X * / ( H C N C X + C D 2 * ( H C N A - H C N C ) - C D 3 * ( K 6 + K 7 + K 8 ) * 0 2 C * H C N C / + C D 7 ) * P H I ( L ) Appendix C Computer Programs for 02 Model andNO/N20 Model 182 S F ( L 3 ) = S F ( L 3 ) - W ( J ) * D X * / ( N O C X + C D 2 * ( N O A - N O C ) - C D 3 * ( K 5 * N O C * N H 3 C / - K 7 * H C N C * 0 2 C / - K 4 * N H 3 C * 0 2 C / - K 1 * N 2 0 C ) / - C D 4 * ( K 1 4 * N O C * C O C / - K 1 0 * N H 3 C * O 2 C ) / - C D 5 * ( K 1 2 * N O C * R O F U E L / + N C C 1 * A V * ( K l l * N O C - K 9 * 0 2 C ) ) ) / * P H I ( L ) S F ( L 4 ) = S F ( L 4 ) - W ( J ) * D X * / ( N 2 0 C X + C D 2 * ( N 2 0 A - N 2 0 C ) - C D 3 * ( ( K 2 + K 1 ) * N 2 0 C / - . 5 * K 6 * H C N C * 0 2 C / - . 5 * K 3 * N H 3 C * 0 2 C ) / - C D 4 * K 1 5 * R O C A O * N 2 0 C ; / - C D 5 * R O F U E L * ( K 1 3 * N 2 0 C / - K l l * N C C l * A V / R O F U E L * N O C ) ) / * P H I ( L ) DO 100 M = l , 2 M 1 = I + ( M - 1 ) * N E Q M2=M1+1 M3=Ml+2 M4=Ml+3 S J ( L 1 , M 1 ) = S J ( L 1 , M 1 ) + D X * W ( J ) * / ( P H I X ( M ) + C D 2 * ( A ( 3 8 ) - 1 ) * P H I ( M ) / - C D 3 * ( ( K 3 + K 4 ) * 0 2 C + 2 . / 3 . * K 5 * N O C ) * P H I ( M ) / - C D 4 * K 1 0 * O 2 C * P H I ( M ) ) * P H I ( L ) S J ( L 1 , M 2 ) = S J ( L 1 , M 2 ) + D X * W ( J ) * C D 2 * A ( 3 6 ) / * P H I ( M ) * P H I ( L ) S J ( L 1 , M 3 ) = S J ( L 1 , M 3 ) + D X * W ( J ) * ( C D 2 * A ( 4 0 ) * P H I ( M ) / - C D 3 * 2 . / 3 . * K 5 * P H I ( M ) * N H 3 C ) / * P H I ( L ) S J ( L 1 , M 4 ) = S J ( L 1 , M 4 ) + D X * W ( J ) * C D 2 * A ( 4 2 ) * P H I ( M ) * P H I ( L ) S J ( L 2 , M 1 ) = 0 . S J ( L 2 , M 2 ) = S J ( L 2 , M 2 ) + D X * W ( J ) * ( P H I X ( M ) + C D 2 * ( A ( 2 6 ) - 1 ) * P H I ( M ) / - C D 3 * ( K 6 + K 7 + K 8 ) * 0 2 C * P H I ( M ) ) * P H I ( L ) S J ( L 2 , M 3 ) = 0 . S J ( L 2 , M 4 ) = 0 . S J ( L 3 , M 1 ) = S J ( L 3 , M 1 ) + W ( J ) * D X * ( C D 2 * A ( 3 9 ) * P H I ( M ) / - C D 3 * ( K 5 * N O C * P H I ( M ) / - K 4 * P H I ( M ) * 0 2 C ) / - C D 4 * ( - K 1 0 ) * P H I ( M ) * 0 2 C ) / * P H I ( L ) S J ( L 3 , M 2 ) = S J ( L 3 , M 2 ) + W ( J ) * D X * ( C D 2 * A ( 3 7 ) * P H I ( M ) / - C D 3 * ( - K 7 ) * P H I ( M ) * 0 2 C ) / * P H I ( L ) S J ( L 3 , M 3 ) = S J ( L 3 , M 3 ) + W ( J ) * D X * ( P H I X ( M ) + C D 2 * ( A ( 4 1 ) - 1 ) * P H I ( M ) / - C D 3 * K 5 * N H 3 C * P H I ( M ) / - C D 4 * K 1 4 * C O C * P H I ( M ) / - C D 5 * ( K 1 2 * R 0 F U E L + K 1 1 * N C C 1 * A V ) Appendix C Computer Programs for 02 Model and NO/N20 Model 183 I * P H I ( M ) ) * P H I ( L ) S J ( L 3 , M 4 ) = S J ( L 3 , M 4 ) + W ( J ) * D X * ( ( - C D 3 ) * ( - K l ) * P H I ( M ) / + C D 2 * A ( 4 3 ) * P H I ( M ) ) / * P H I ( L ) S J ( L 4 , M l ) = S J ( L 4 , M 1 ) + W ( J ) * D X * / * ( C D 2 * A ( 4 5 ) * P H I ( M ) / - C D 3 * ( - . 5 ) * K 3 * 0 2 C * P H I ( M ) ) / * P H I ( L ) S J ( L 4 , M 2 ) = S J ( L 4 , M 2 ) + W ( J ) * D X * / ( C D 2 * A ( 4 4 ) * P H I ( M ) / - C D 3 * ( - . 5 ) * K 1 * P H I ( M ) * 0 2 C ) / * P H I ( L ) S J ( L 4 , M 3 ) = S J ( L 4 , M 3 ) + W ( J ) * D X * / ( C D 2 * A ( 4 6 ) * P H I ( M ) / + C D 5 * K 1 1 * N C C 1 * A V * P H I ( M ) ) / * P H I ( L ) 100 S J ( L 4 , M 4 ) = S J ( L 4 , M 4 ) + W ( J ) * D X * ( P H I X ( M ) + C D 2 * ( A ( 4 7 ) - 1 ) * P H I ( M ) / - C D 3 * ( K 2 + K 1 ) * P H I ( M ) / - C D 4 * R O C A O * K 1 5 * P H I ( M ) / - C D 5 * K 1 3 * R O F U E L * P H I ( M ) ) / * P H I ( L ) C C BOUNDARY C O N D I T I O N S C S F ( 1 ) = 0 . S F ( 2 ) = 0 . S F ( 3 ) = 0 . S F ( 4 ) = 0 . DO 150 I L = 1 , N 1 S J ( 1 , I L ) = 0 . S J ( 2 , I L ) = 0 . S J ( 3 , I L ) = 0 . 150 S J ( 4 , I L ) = 0 . DO 149 I L = 1 , N 1 149 S J ( N 1 , I L ) = 0 . S J ( N 1 , N 1 ) = 1 . 0 E N D I F S J ( 1 , 1 ) = 1 . S J ( 2 , 2 ) = 1 . S J ( 3 , 3 ) = 1 . S J ( 4 , 4 ) = 1 . C A L L L U D C M P ( S J , N 1 , 1 2 0 , I N D X , D ) C A L L L U B K S B ( S J , N 1 , 1 2 0 , I N D X , S F ) Appendix C Computer Programs for 02 Model and NO/N20 Model 184E R R = 0 . DO 200 1 = 1 , N I 200 E R R = E R R + S F ( I ) * * 2 E R R O R = D S Q R T ( E R R / N 1 ) P R I N T * , ' P R I N T * , E R R O R P R I N T * , ' I F ( E R R O R . L T . 1 . 8 E - 0 ) T H E N P R I N T * , ' E Q U A T I O N S O L V E D ' E L S E I F ( K K . E Q . 3 0 ) T H E N P R I N T * , ' C H E C K P R O G R A M ' P A U S E E L S E DO 250 1 = 1 , N I 250 C T ( I ) = C T ( I ) + S F ( I ) GO T O 160 C A S S I G N T H E S O L U T I O N TO CORRESPONDING ARRAYS AND C A L C U L A T E T H E C O N C E N T R A T I O N S I N ANNULUS DO 270 1 = 1 , N N H 3 C S ( I ) = C T ( 4 * ( I - 1 ) + 1 ) H C N C S ( I ) = C T ( 4 * ( I - l ) + 2 ) N O C S ( I ) = C T ( 4 * ( I - 1 ) + 3 ) N 2 0 C S ( I ) = C T ( 4 * ( I - l ) + 4 ) C A 5 = A A * V N H 3 A S ( I ) / C A 1 . ' C A 6 = A A * V H C N A S ( I ) / C A 1 C D 6 = - A C * V N H 3 C S ( I ) / C D 1 C D 7 = - A C * V H C N C S ( I ) / C D 1 A ( 1 ) = C A 3 * K 1 0 * O 2 A S ( I ) + C A 2 *K4 * 0 2 A S ( I ) A ( 2 ) = 1 . + C A 3 * K 1 4 * C O A S ( I ) + C A 4 * K 1 2 + C A 4 * K 1 1 * N C C 1 * A V A ( 3 ) = C A 2 * K 7 * 0 2 A S ( I ) A ( 4 ) = C A 2 * K 1 A ( 5 ) = C A 4 * K 9 * N C C 1 * A V * 0 2 A S ( I ) A ( 6 ) = C A 4 * K 1 1 * N C C 1 * A V A ( 7 ) = . 5 * C A 2 * K 3 * 0 2 A S ( I ) A ( 8 ) = . 5 * C A 2 * K 6 * 0 2 A S ( I ) A ( 9 ) = l . + C A 3 * K 1 5 * R O C A O + C A 4 * K 1 3 * R O F U E L + C A 2 * ( K 2 + K 1 ) A ( 1 0 ) = 1 . 5 + 1 . 5 * C A 2 * K 3 * O 2 A S ( I ) + 2 . 5 * A ( 1 ) A ( 2 6 ) = l . / ( l . + C A 2 * K 6 * 0 2 A S ( I ) + C A 2 * K 7 * 0 2 A S ( I ) + C A 2 * K 8 * 0 2 A S ( I ) ) A ( 1 1 ) = N O C S ( I ) - 1 . 5 * N H 3 C S ( I ) + A ( 3 ) * A ( 2 6 ) * ( H C N C S ( I ) + C A 6 ) / + A ( 5 ) - 1 . 5 * C A 5 A ( 1 2 ) = - A ( 8 ) * A ( 2 6 ) * ( H C N C S ( I ) + C A 6 ) - N 2 0 C S ( I ) A ( 1 3 ) = A ( 2 ) / A ( 4 ) A ( 1 4 ) = A ( 1 0 ) / A ( 4 ) A ( 1 5 ) = A ( 1 1 ) / A ( 4 ) A ( 1 6 ) = A ( 6 ) / A ( 9 ) A ( 1 8 ) = A ( 7 ) / A ( 9 ) A ( 1 6 ) = A ( 6 ) / A ( 9 ) A ( 1 8 ) = A ( 7 ) / A ( 9 ) A ( 1 9 ) = A ( 1 2 ) / .A (9) A ( 2 0 ) = ( A ( 1 8 ) + A ( 1 4 ) ) / ( A ( 1 3 ) - A ( 1 6 ) ) A ( 2 1 ) = ( A ( 1 5 ) - A ( 1 9 ) ) / ( A ( 1 3 ) - A ( 1 6 ) ) A ( 2 4 ) = ( A ( 2 1 ) * A ( 2 3 ) - A ( 2 2 ) ) / ( A ( 2 0 ) * A ( 2 3 ) ) Appendix C Computer Programs for 02 Model andNO/N20 Model 185 A ( 2 5 ) = ( N H 3 C S ( I ) + C A 5 ) / ( A ( 2 0 ) * A ( 2 3 ) ) A ( 2 6 ) = l . / ( l . + C A 2 * K 6 * 0 2 A S ( I ) + C A 2 * K 6 * 0 2 A S ( I ) + C A 2 * K 8 * 0 2 A S ( I ) ) N H 3 A S ( I ) = - . 5 * ( A ( 2 4 ) - D S Q R T ( A ( 2 4 ) * * 2 - 4 . * A ( 2 5 ) ) ) H C N A S ( I ) = A ( 2 6 ) * ( H C N C S ( I ) + C A 6 ) N O A S ( I ) = ( A ( 2 1 ) + A ( 2 0 ) * N H 3 A S ( I ) ) 27 0 N 2 0 A S ( I ) = ( A ( 1 6 ) * N O A S ( I ) + A ( 1 8 ) * N H 3 A S ( I ) - A ( 1 9 ) ) C NCC I T E R A T I O N I N T E G R A T I O N COUNTERPART FOR 1 s t Z O N E ; C D X : I N T E G R A T I O N I N C R E M E N T D X = X ( 2 ) - X ( l ) C A L L I N T ( D X , A C S , A A S , 0 2 C S , 0 2 A S , I T G 0 2 ) C A L L I N T ( D X , A C S , A A S , N O C S , N O A S , ITGNO) E N D I F R E T U R N END C * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * C IMPORT 02 AND CO C O N C E N T R A T I O N S * C * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * S U B R O U T I N E I N P U T 1 ( N E , I I , X , C S , A S ) D O U B L E P R E C I S I O N X ( 3 0 ) , A S ( 3 0 ) , C S ( 3 0 ) DO 10 I = 1 , N E + 1 10 R E A D ( I I , * ) X ( I ) , C S ( I ) , A S ( I ) R E T U R N END C * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * C IMPORT D E V O L A T I L I Z A T I O N AND HYDRODYNAMIC I N F O R M A T I O N * C * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * S U B R O U T I N E I N P U T 2 ( N E , I I , X , C S , A S , A C S , C A S ) D O U B L E P R E C I S I O N X ( 3 0 ) , A S ( 3 0 ) , C S ( 3 0 ) , A C S ( 3 0 ) , C A S ( 3 0 ) DO 10 I=NE+1 10 R E A D ( I I , * ) X ( I ) , C S ( I ) , A S ( I ) , A C S ( I ) , C A S ( I ) R E T U R N END C * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * C * S U B R O U T I N E T F U N C T , L I N E A R I N T E R P O L A T I O N F U N C T I O N S * C * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * S U B R O U T I N E T F U N C T ( D X , G P , P H I , P H I X ) D O U B L E P R E C I S I O N P H I ( 2 ) , P H I X ( 2 ) , G P , D X P H I ( 1 ) = 1 . 0 - G P P H I ( 2 ) = G P P H I X ( 1 ) = - 1 . / D X P H I X ( 2 ) = 1 . / D X R E T U R N END C * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * C T H E F O L L O W I N G TWO S U B R O U T I N E F I N D T H E R E V E R S E OF M A T R I X * Appendix C Computer Programs for 02 Model and NO/N20 Model 186 c * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * S U B R O U T I N E L U B K S B ( A , N , N P , I N D X , B ) D O U B L E P R E C I S I O N A ( N P , N P ) , I N D X ( N ) , B ( N ) 11=0 DO 12 1 = 1 , N L L = I N D X ( I ) S U M = B ( L L ) B ( L L ) = B ( I ) I F ( I I . N E . O ) T H E N DO 11 J = I I , I - 1 11 S U M = S U M - A ( I , J ) * B ( J ) E L S E I F ( S U M . N E . O ) T H E N 11=1 E N D I F 12 B ( I ) = S U M DO 14 I = N , 1 , - 1 S U M = B ( I ) DO 13 J = I + 1 , N 13 S U M = S U M - A ( I , J ) * B ( J ) 14 B ( I ) = S U M / A ( I , I ) R E T U R N END C * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * S U B R O U T I N E L U D C M P ( A , N , N P , I N D X , D ) P A R A M E T E R ( N M A X = 2 8 0 , T I N Y = 1 . O E - 2 0 ) D O U B L E P R E C I S I O N A ( N P , N P ) , I N D X ( N ) , W ( N M A X ) D = 1 . 0 DO 12 I = 1 , N A A M A X = 0 . DO 11 J = 1 , N 11 I F ( A B S ( A ( I , J ) ) . G T . A A M A X ) A A M A X = A B S ( A ( I , J ) ) I F ( A A M A X . E Q . 0 . ) P A U S E ' S I N G U L A R M A T R I X ' 12 W ( I ) = 1 . / A A M A X DO 19 J = 1 , N DO 14 1 = 1 , J - l S U M = A ( I , J ) DO 13 K = 1 , I - 1 13 S U M = S U M - A ( I , K ) * A ( K , J ) 14 A ( I , J ) = S U M A A M A X = 0 . DO 16 I = J , N S U M = A ( I , J ) DO 15 K = 1 , J - 1 15 S U M = S U M - A ( I , K ) * A ( K , J ) A ( I , J ) = S U M D U M = W ( I ) * A B S ( S U M ) I F ( D U M . G E . AAMAX) T H E N IMAX=I AAMAX=DUM E N D I F 16 C O N T I N U E I F ( J . N E . I M A X ) . T H E N DO 17 K = 1 , N D U M = A ( I M A X , K ) A ( I M A X , K ) = A ( J , K ) 17 A ( J , K ) = D U M D=-D W ( I M A X ) = W ( J ) E N D I F Appendix C Computer Programs for 02 Model and' NO/N20 Model 187 I N D X ( J ) = I M A X I F ( A ( J , J ) . E Q . O . ) A ( J , J ) = T I N Y I F ( J . N E . N ) T H E N D U M = 1 . / A ( J , J ) DO 18 I = J + 1 , N 18 A ( I , J ) = A ( I , J ) * D U M E N D I F 19 C O N T I N U E R E T U R N END Appendix D 02 , JSOx and N20 Concentration Profiles in the UBC CFBC Unit 02, NOx and N 2 0 concentration profiles have been measured in Runs 14-1, 19-1 for Mt Klappan Anthracite, in Runs 17-1, 17-2 and 17-3 for Conoco delayed coke, in Runs 22-2 and 22-3 for C A N M E T pitch and in Runs 26-1 and 26-2 for Poplar River lignite. Some profiles are given in Chapter 5, while the others are plotted here. Table D . l addresses where each profile is located in this thesis. Note that there are five gas sampling ports along the riser at 0.6 m, 1.5 m, 2.7 m, 4.2 m, and 6.4 m above the distributor. Gas samples were withdrawn from three lateral position at each height, 0 mm, 38 mm and 76 mm from the wall. Due to the interference of C H 4 , the N 20 concentrations cannot usually be obtained at the lowest level, and are usually hot available for all the lateral positions at a given height. Table D . l : Index of 02, N O x and N 20 Concentration Profiles Fuels Run No 0, N O v N,0 Conoco Delayed coke 17-1 Figure D. l Figure 5-11 Figure 5.13 17-2 Figure 5.9 Figure D.7 Figure D.13 17.3 Figure D.2 Figure 5.12 Figure D.14 Mt Klappan Anthracite 14-1 Figure D.3 Figure D.8 Figure D.15 19-1 Figure D.4 Figure D.9 Figure D.16 C A N M E T 22-3 Figure 5.10 Figure D.10 Figure 5.5 Pitch Poplar River lignite 26-1 Figure D.5 Figure D. 11 Figure D.17 26-2 Figure D.6 Figure D. 12 Figure D.l8 188 Appendix D 0} N0X andN20 Concentration Profiles in the UBC CFBC Unit 189 9.1 Height (m) Figure D.l: 0 2 Concentration Profiles for Conoco Coke Combustion: Run 17-1. T = 1155 °K, U g = 7.0 m/s, P/S = 1.6, Ca/S = 0. The secondary air was introduced 3.4 m above the distributor. Symbols are the experimental data measured from: •: Wall, •: Middle, •: Axis, *: Flue gas. Appendix D 02, N0X and N20 Concentration Profiles in the UBC CFBC Unit 190 Riser Height (m) Figure D.2: 0 2 Concentration Profiles for Conoco Coke Combustion: Run 17-3. T = 1134 °K, U g = 7.1 m/s, P/S = 1.6, Ca/S = 3.2. The secondary air was introduced 3.4 m above the distributor. Symbols are the experimental data measured from: •: Wall, •: Middle, • : Axis, *: Flue gas. Appendix D 02, N0X and N20 Concentration Profdes in the UBC CFBC Unit 191 Figure D.3: 0 2 Concentration Profiles for Mt Klappan Anthracite Combustion: Run 14-3. T = 1154°K, U g = 8.4 m/s, P/S = 2.0, Ca/S = 0.0. The secondary air was introduced 3.4 m above the distributor. Symbols are the experimental data measured from: •: Wall, •: Middle, A: Axis, *: Flue gas. Appendix D 02, N0X andN20 Concentration Profiles in the UBC CFBC Unit 192 11.2 Height (m) Figure D.4: 0 2 Concentration Profiles for Mt Klappan Anthracite Combustion: Run 19-1. T = 1146°K, U g = 8.7 m/s, P/S = 1.1, Ca/S = 0.0. The secondary air was introduced 3.4 m above the distributor. Symbols are the experimental data measured from: • : Wall, • : Middle, A: Axis, *: Flue gas. Appendix D 02, N0X and N20 Concentration Profiles in the UBC CFBC Unit 193 Figure D.5: 0 2 Concentration Profiles for Poplar River Lignite Combustion: Run 26-1. T = 1137°K, U g = 8.2 m/s, P/S = 2.0, Ca/S = 0.0. The secondary air was introduced 3.4 m above the distributor. Symbols are the experimental data measured from: • : Wall, • : Middle, A : Axis, *: Flue gas. Appendix D 02, N0X andN20 Concentration Profiles in the UBC CFBC Unit 194 5.11 0.00 3 4 5 Height (m) 8 Figure D.6: 0 2 Concentration Profiles for Poplar River Lignite Combustion: Run 26-2. T = 1149°K, U g = 8.7 m/s, P/S = 0.84, Ca/S = 0.0. The secondary air was introduced 3.4 m above the distributor. Symbols are the experimental data measured from: •: Wall, •: Middle, •: Axis, *: Flue gas. Appendix D N0X and N20 Concentration Profiles in the UBC CFBC Unit 195 Figure D.7: N 0 X Concentration Profiles for Conoco Coke Combustion: Run 17-2. T 1121°K, U g = 6.9 m/s, P/S = 1.6, Ca/S = 2.5. The secondary air was introduced 3.4 above the distributor. Symbols are the experimental data measured from: • : Wall, Middle, • : Axis, *: Flue gas. Appendix D N0X and N20 Concentration Profdes in the UBC CFBC Unit 196 Figure D.8: N 0 X Concentration Profiles for Mt Klappan Anthracite Combustion:Run 14-1. T = 1154°K, U g = 8.4 m/s, P/S = 2.0, Ca/S = 0.0. The secondary air was introduced 3.4 m above the distributor. Symbols are the experimental data measured from: • : Wall, • : Middle, A : Axis, *: Flue gas. Appendix D 02, N0X andN20 Concentration Profiles in the UBC CFBC Unit 197 Figure D.9: N0 X Concentration Profiles for Mt Klappan Anthracite Combustion: Run 19-1. T= 1146°K, U g= 8.7 m/s, P/S = 1.1, Ca/S = 0.0. The secondary air was introduced 3.4 m above the distributor. Symbols are the experimental data measured from: •: Wall, •: Middle, •: Axis, *: Flue gas. Appendix D 0^ N0X and N20 Concentration Profiles in the UBC CFBC Unit 198 Figure D.10: NO x Concentration Profiles for CANMET Pitch Combustion: Run 22-3. T 1114°K, U g = 9.0 m/s, P/S = 2.5, Ca/S = 2.6. The secondary air was introduced 3.4 above the distributor. Symbols are the experimental data measured from: •: Wall, Middle, A: Axis, *: Flue gas. Appendix D Op N0X andN20 Concentration Profiles in the UBC CFBC Unit 199 Figure D . l l : N 0 X Concentration Profiles for Poplar River Lignite Combustion: Run 26-1. T = 1137°K, U g = 8.2 m/s, P/S = 2.0, Ca/S - 0 . 0 The secondary air was introduced 3.4 m above the distributor. Symbols are the experimental data measured from: • : Wall, • : Middle, • : Axis, *: Flue gas. Appendix D 02, N0X andN20 Concentration Profiles in the UBC CFBC Unit 200 3 4 5 Height (m) Figure D.12: N0 X Concentration Profiles for Poplar River Lignite Combustion: Run 26-2. T = 1149°K, U g = 8.3 m/s, P/S = 0.84, Ca/S = 0.0. The secondary air was introduced 3.4 m above the distributor. Symbols are the experimental data measured from: •: Wall, •: Middle, •: Axis, *: Flue gas. Appendix D 02, N0X cmdN20 Concentration Profdes in the UBC CFBC Unit 201 Figure D.13: N 20 Concentration Profiles for Conoco Coke Combustion: Run 17-2. T = 1121°K, U g = 6.9 m/s, P/S = 1.6, Ca/S = 2.5. The secondary air was introduced 3.4 n above the distributor. Symbols are the experimental, data measured from: •: Wall, • Middle, •: Axis, *: Flue gas. Appendix D N0X and N20 Concentration Profiles in the UBC CFBC Unit 202 CL CL c o ro t_ "c CD O c o O 3 4 5 Height (m) 8 Figure D.14: N 2 0 Concentration Profiles for Conoco Coke Combustion: Run 17-3. T 1134°K, U g = 7.1 m/s, P/S = 1.6, Ca/S = 3.2. The secondary air was introduced 3.4 above the distributor. Symbols are the experimental data measured from: • : Wall, « Middle, A: Axis, *: Flue gas. Appendix D 02, N0X and N20 Concentration Profiles in the UBC CFBC Unit 203 Figure D. 15: N 2 0 Concentration Profiles for Mt Klappan Anthracite Combustion: Run 14-1. T = 1154°K, U g = 8.4 m/s, P/S = 2.0, Ca/S = 0.0. The secondary air was introduced 3.4 m above the distributor. Symbols are the experimental data measured from: • : Wall, • : Middle, • : Axis, *: Flue gas. Appendix D 02, N0X and N20 Concentration Profiles in the UBC CFBC Unit 204 3 4 5 Riser Height (m) 8 Figure D . 16: N 2 0 Concentration Profiles for Mt Klappan Anthracite Combustion: Run 19-1. T = 1146°K, U g= 8.7 m/s, P/S = 1.1, Ca/S = 0.0. The secondary air was introduced 3.4 m above the distributor. Symbols are the experimental data measured from: •: Wall, •: Middle, •: Axis, *: Flue gas. Appendix D 0 2 , N0X andN20 Concentration Profiles in the UBC CFBC Unit 205 Figure D.17: N 2 0 Concentration Profiles for Poplar River Lignite Combustion: Run 26-1. T = 1137°K, U g = 8.2 m/s, P/S - 2.0, Ca/S = P.O. The secondary air was introduced 3.4 m above the distributor. Symbols are the experimental data measured from: • : Wall, • : Middle, • : Axis, *: Flue gas. Appendix D 02, NOx and N20 Concentration Profiles in the UBC CFBC Unit 206 Figure D.18: N 20 Concentration Profiles for Poplar River Lignite Combustion: Run 26-2. T = 1149°K, U g = 8.7 m/s, P/S = 0.84, Ca/S = 0.0. The secondary air was introduced 3.4 m above the distributor. Symbols are the experimental data measured from: •: Wall, •: Middle, •: Axis, *: Flue gas. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0058551/manifest

Comment

Related Items