Low NO
xEmissions from Fuel-Bound Nitrogen in
Gas Turbine Combustors
Proefschrift
ter verkrijging van de graad van doctor aan de Technische Universiteit Delft,
op gezag van de Rector Magnificus prof.dr.ir. J.T. Fokkema voorzitter van het College voor Promoties,
in het openbaar te verdedigen op dinsdag 20 juni 2006, om 10.00 uur
door
Belkacem ADOUANE
Dit proefschrift is goedgekeurd door de promotoren:
Prof.ir. J.P. van Buijtenen Prof.Dr.-Ing H. Spliethoff
Samenstelling promotiecommissie:
Rector Magnificus, voorzitter
Prof.ir. J.P. van Buijtenen Technische Universiteit Delft, promotor Prof.Dr.-Ing H. Spliethoff Technische Universiteit Delft, co-promotor Prof.dr.ir. D.J.E.M. Roekaerts Technische Universiteit Delft
Prof.dr.ir. T.H. van der Meer Universiteit Twente
Prof.dr.ir. C. Daey Ouwens Technische Uinversiteit Eindhoven Dr.ir. G.J. Witteveen Winnox Combustion Systems, Nijmegen Dr.ir. W. de Jong Technische Universiteit Delft
To my parents To my wife and children
Abstract
Biomass-derived LCV (Low Calorific Value) gas represents one of the best alternatives for fossil fuels. It is very attractive, because it is CO2 neutral as biomass consumes an
amount of CO2 when growing and releases almost the same amount when combusted.
However, the raw gasifier producer gas contains a high content of fuel-bound nitrogen (FBN), which results in high NOx emissions after its combustion. The NOx emissions
compromise the neutral aspect of the biomass-derived LCV gas.
Reducing the conversion of FBN to NOx has been one of the main challenges for
researchers working in the field of LCV gas combustion.
There are indeed three main ways to reduce the emission of NOx; upstream of
the combustor by scrubbing, downstream the combustor by SNCR (Selected Non Catalytic Reduction), or SCR (Selected Catalytic Reduction), or inside the combustor system by optimizing the combustion process to result in the lowest conversion of FBN to NOx, also called ”primary measures”.
For this research work the third approach was adopted, i.e. reducing the conversion of FBN to NOxby primary measures. A new combustor has been developed within the
framework of this thesis, the newly designed combustor was named ”Winnox-TUD”. The Winnox-TUD combustor has been developed after a series of first experiments using an available Winnox combustor. The Winnox combustor was designed for a Rover micro-gas turbine. It is composed of two stages, with an inserted depletion plate between the first and second stage. The Winnox combustor was fueled with natural gas doped with ammonia to simulate the FBN in the gas. Those first experiments consisted of investigating the effect of stoichiometry in the different stages on the conversion of ammonia to NOx.
Finding out the main factors controlling the conversion of FBN to NOx and
op-timizing those factors to result in the lowest possible conversion of FBN is the main goal throughout this research work.
To achieve the defined goal, the Winnox-TUD combustor was the subject of extensive experimental and modeling investigations.
The combustor was tested experimentally to define the effect of stoichiometry,
power, FBN concentration, gas composition, heating value and primary air tempera-ture on the conversion of FBN to NOx.
It was found that all those parameters affect the conversion of FBN to different extent, however, stoichiometry in the first stage, FBN content in the LCV gas in addition to natural gas (especially CH4) are the main factors controlling the conversion
of FBN.
Optimizing those factors can result in a very important reduction of FBN con-version to NOx thus a reduction in NOx emissions. In this thesis experiments are
described, the modeling of the reacting flow field using CFD (Fluent) and the mod-eling of the chemical kinetics (Chemkin).
The experiments were divided into two categories: experiments using natural gas diluted with nitrogen and experiments using LCV gas from a mixing station, where the main components of the real biomass-derived LCV gas were mixed to produce a synthetic LCV gas. Ammonia was added to the gas to simulate the presence of the fuel bound nitrogen in the fuel gas.
Furthermore, one of the other goals set for this research work, was to define clearly the limits of such primary measures in reducing the conversion of FBN to NOx, and
the reduction of the total NOx emitted from the combustion of biomass-derived LCV
gas in gas turbines, gas engines or boilers.
Samenvatting
Brandstofgas uit biomassa met een lage verbrandingswaarde (LCV: Low Calorific Value) vertegenwoordigt een van de beste alternatieven voor fossiele brandstoffen. Het is CO2 neutraal, daar biomassa tijdens de groei evenveel CO2 opneemt als het
emitteert tijdens verbranding. Echter, ruw gas, geproduceerd in een vergasser, bevat een hoog gehalte aan brandstofgebonden stikstof (Fuel-bound Nitrogen: FBN), dat resulteert in hoge emissies van NOx bij verbranding. Deze emissies van NOx kunnen
het voordeel van het CO2 -neutrale karakter tenietdoen.
De reductie van de omzetting van FBN in NOxis een van de belangrijkste
uitdagin-gen voor onderzoekers op het gebied van van de verbranding van biobrandstoffen. Er zijn drie wegen om de emissie van brandstof NOx te verlagen: verwijdering
van stikstofhoudende stoffen uit de brandstof d.m.v. scrubbing, verwijdering van NOx uit de rookgassen door SNCR of SCR, of door tijdens het verbrandingsproces de
conversie van FBN in NOx te minimaliseren: de zogenaamde primaire maatregelen.
Dit onderzoek behandelt de derde methode: namelijk de reductie van de omzetting van FBN in NOx door primaire maatregelen.
In het kader van dit onderzoek is een nieuwe verbrandingskamer ontwikkeld: de ”Winnox-TUD” verbrandingskamer. Deze verbrandingskamer is ontwikkeld na een reeks van experimenten met een beschikbare Winnox brander. Deze was ontworpen voor een kleine gasturbine van Rover, en bevat twee trappen, met een ontlastingsplaat ertussen. Als brandstof werd aardgas gebruikt met een toevoeging van ammonia om de aanwezigheid van FBN te simuleren. In deze eerste experimenten is de invloed van de stoichiometrie in beide trappen op de conversie van FBN in NOx onderzocht.
Het voornaamste doel van het gehele onderzoek is het vinden van de voornaamste parameters die de omzetting van FBN in NOx bepalen, en deze te optimaliseren
voor de laagst mogelijke conversie. Om dit doel te bereiken was de Winnox-TUD verbrandingskamer het onderwerp van intensieve modelvorming en experimenten.
De verbrandingskamer is getest om de effecten vast te leggen van stoichiometrie, belasting, concentratie van FBN, samenstelling, verbrandingswaarde en temperatuur van de verbrandingslucht op de omzetting van FBN in NOx.
Er kon worden vastgesteld dat al deze parameters de omzetting in verschillende mate benvloeden, maar dat de voornaamste factoren zijn: stoichiometrie in de eerste trap, het aandeel chemisch gebonden stikstof in de brandstof en de aanwezigheid van methaan. Optimalisatie van deze parameters resulteerde in een aanzienlijke reductie van de omzetting van FBN in NOx, en daarmee in een aanzienlijke reductie in de
emissie van NOx.
Voor de nieuw ontworpen Winnox-TUD verbrandingskamer is dit bereikt door een groep onderzoekers van de Sectie Energietechnologie van de Technische Universiteit in Delft. In dit proefschrift zijn de experimenten beschreven, de modelvorming van het stromingsveld in CFD (Fluent) en de chemische kinetiek (Chemkin).
De experimenten zijn verdeeld in twee categorien: experimenten met aardgas gemengd met chemisch gebonden stikstof (ammonia) en experimenten met ges-imuleerd productgas uit een mengstation.
Verder moesten in dit onderzoek duidelijk de grenzen worden aangegeven aan het effect van de primaire maatregelen voor de reductie van de conversie van FBN in NOx,
en de reductie in uiteindelijke emissie van NOx bij toepassing van de technologie in
gasturbines, gasmotoren en ketels.
Contents
Abstract v
Samenvatting vii
1 Introduction 1
1.0.1 NOx effect on health and environment . . . 1
1.1 World energy supply situation . . . 3
1.2 The need for renewable energy . . . 5
1.3 Barriers related to renewable energy application . . . 5
1.4 Biomass as a source for renewable energy . . . 6
1.5 This thesis . . . 7
1.6 Primary and secondary measures of reducing fuel-NOx . . . 8
1.7 Methodology . . . 9
1.7.1 Experimental . . . 9
1.7.2 Modeling . . . 11
2 Biomass gasification and subsequent combustion in gas turbines 13 2.1 Introduction . . . 13
2.1.1 Gasifier types - Advantages and Disadvantages . . . 14
2.1.2 Biomass derived LCV gas composition . . . 18
2.2 The application of biomass derived LCV gas in gas turbines . . . 19
2.2.1 Gasifier-gas turbine integrated process . . . 19
2.2.2 Problems related to the application of biomass derived LCV gas in gas turbines . . . 20
2.3 The approach of solving the problems related to the application of biomass derived LCV gas in gas turbines . . . 20
3 Combustor design for LCV gas 23 3.1 Introduction . . . 23
3.2 NOx types . . . 23
3.2.1 Thermal NOx . . . 24
3.2.2 Prompt NOx 27 3.2.3 Fuel-NOx . . . 27
3.3 Low NOx combustors for natural gas . . . 27
3.4 Low NOx combustors for LCV gas . . . 29
3.5 The newly designed combustor: Winnox-TUD combustor . . . 30
3.5.1 Design specifications . . . 30
3.5.2 Why is the newly designed combustor giving low NOx ? . . . 30
4 Experimental investigation 35 4.1 Introduction . . . 35
4.2 Related experimental work from literature . . . 36
4.2.1 Preliminary experiments and final experiments . . . 39
4.3 Description of the experiments using the newly designed combustor . 41 4.4 Experimental setup . . . 43
4.5 Measurement techniques . . . 44
4.5.1 NO or NOx? . . . 45
4.5.2 Corrections and assumptions . . . 45
4.5.3 NH3 conversion calculation . . . 47
4.6 Experiments with natural gas diluted with nitrogen and doped with ammonia . . . 47
4.6.1 Effect of NH3 concentration in the LCV gas . . . 47
4.6.2 Effect of stoichiometry in the first stage on the conversion of NH3 to NO . . . 49
4.6.3 Effect of NH3 concentration on the minimum of conversion . 49 4.6.4 Effect of power . . . 50
4.6.5 Effect of the fuel heating value . . . 53
4.6.6 Effect of secondary air . . . 55
4.6.7 LCV gas doped with NO instead of NH3, FTIR measurements 55 4.6.8 LCV gas doped with NO instead of NH3, a comparison with NH3 injcetion . . . 58
4.7 Experiments with LCV gas from the mixing station and doped with ammonia . . . 60
4.7.1 LCV gas containing natural gas . . . 60
4.7.2 LCV gas free of natural gas . . . 63
4.8 Comparison between LCV gas containing and not containing natural gas . . . 65
4.8.1 Effect of NH3 concentration in the LCV gas . . . 65
4.8.2 Effect of stoichiometry in the first stage . . . 66
4.9 Concluding summary . . . 67
5 Modeling theory applied to the Winnox-TUD combustor 69
5.1 Introduction . . . 69
5.2 Chemical rate expressions . . . 69
5.3 Homogeneous 0-D reactor . . . 71
5.4 Assumptions and limitations . . . 71
5.5 General equations . . . 71
5.6 Mass conservation and gas phase species equations . . . 72
5.7 Species conservation equation . . . 73
5.8 Residence time . . . 73
5.9 Gas Energy Equation . . . 73
5.10 Numerical Solution Methods . . . 76
5.10.1 Steady state solver for homogeneous systems . . . 76
6 Actual modeling and comparison with experiments 77 6.1 Introduction . . . 77
6.1.1 Strategy . . . 77
6.1.2 Model Implementation . . . 78
6.2 PART I Model validation . . . 78
6.2.1 Natural gas diluted with nitrogen . . . 79
6.2.2 Simulated LCV gas from the mixing station . . . 85
6.3 PART II Model extensive use . . . 85
6.3.1 Effect of heat loss in the primary zone . . . 87
6.3.2 Effect of primary air temperature . . . 87
6.3.3 Effect of flame temperature . . . 87
6.4 Adiabatic investigation using the model developed in Chemkin . . . . 88
6.5 Comparison between Kilpinen97 and SKG mechanisms . . . 96
6.6 Discussion . . . 96
7 Conclusions and recommendations 99 7.1 Conclusions . . . 99
7.1.1 Experimental investigation . . . 99
7.1.2 Modeling . . . 100
7.2 Recommendations . . . 100
Bibliography 103 A CFD simulations of the Winnox-TUD combustor 109 B Numerical Solution Methods 123 B.1 Steady state solver for homogeneous systems . . . 123
B.2 Starting estimates . . . 123
B.3 Modified damped Newtons method for 0-D reactors . . . 125
B.4 Jacobian matrix . . . 127
B.5 Pseudo time-stepping procedure . . . 127
C Kilpinen 97 mechanism 129
Acknowledgments 139
Curriculum Vitae 141
List of Figures
1.1 The cycle of nitrates and ozone production from NOx and hydrocarbons 4
1.2 Development of primary consumption . . . 4
1.3 Sketch of the methodology applied throughout this research work . . 10
2.1 Biomass gasification principle . . . 14
2.2 Updraft gasifier . . . 15
2.3 Downdraft gasifier . . . 15
2.4 Fluidised bed gasifier . . . 17
2.5 Circulating fluidised bed gasifier . . . 17
2.6 Entrained flow gasifier . . . 17
2.7 The approach adopted throughout this research work . . . 21
3.1 NOx formation as a function of time and temperature; p=1MPa, [1] . 25 3.2 Dependence of NOxon flame temperature for liquid and gaseous fuels, [1] 25 3.3 Effect of residence time and equivalence ratio on NOx in a premixed fuel-air system, [1] . . . 26
3.4 Prompt NO reaction path, [2] . . . 27
3.5 Reaction path diagram for oxidation of ammonia in flames, (Miller and Bowman, 1989). [3] . . . 28
3.6 Sketch of the Winnox combustor used in the preliminary experiments 29 3.7 Cross section of the newly designed combustor . . . 31
3.8 Pictures of the Winnox-TUD combustor . . . 31
3.9 Picture of the Winnox-TUD combustor mounted in the combustion setup at TU Delft . . . 32
4.1 NO emission and NH3 to NO conversion versus NH3 concentration, Al-Shaikhly et al. (1994) [4] . . . 37
4.2 NH3 conversion to NOx versus the combustor exit temperature, Bat-tista et al. (1996) [5] . . . 38
4.3 NO emission and NH3 to NO conversion versus NH3 concentration
without methane in the LCV gas, Kelsall et al. (1994) [6] . . . 38
4.4 NO emission and NH3 to NO conversion versus NH3 concentration with methane in the LCV gas, Kelsall et al. (1994) [6] . . . 39
4.5 NH3 to NO conversion versus methane in the LCV gas and pressure, Nakata et al. (1994) [7] . . . 40
4.6 The combustion set up . . . 43
4.7 Temperature measurement probe . . . 44
4.8 Water-cooled probe for species sampling . . . 44
4.9 FTIR measuring system . . . 45
4.10 Setup monitoring display . . . 46
4.11 NH3 conversion to NO versus NH3 in the fuel gas . . . 49
4.12 NH3 conversion to NO versus λ primary stage . . . . 50
4.13 NH3 conversion to NO versus λ primary stage at different NH3 con-centration in the fuel gas . . . 51
4.14 NH3 conversion to NO versus λ primary stage at different power settings 52 4.15 NH3 conversion to NO, versus λ primary stage for diffrent heating values of the LCV gas . . . 54
4.16 NH3 conversion to NO versus λ (primary + secondary) stages (NH3= 2% in the LCV gas, λ primary = 0.53) . . . 56
4.17 NO conversion and (NO+NH3+HCN)conversion to NO versus λ pri-mary stage . . . 58
4.18 NH3 and NO conversion to NO versus λ primary stage . . . . 59
4.19 NH3 conversion to NO, versus NH3 concentration in the LCV gas containing methane . . . 61
4.20 NH3 conversion to NO versus λ primary stage at different methane concentration in the LCV gas . . . 62
4.21 NH3 conversion to NO, versus NH3 concentration in the LCV gas . . 64
4.22 NH3 conversion to NO, versus Lambda primary stage . . . 64
4.23 NH3 conversion to NO versus λ (primary + secondary) stages (NH3= 3000 ppm, λprimary = 0.8) . . . 65
4.24 NH3 conversion to NO, versus NH3 concentration in the LCV gas and natural gas diluted with nitrogen . . . 66
4.25 NH3 conversion to NO, versus λ primary stage for LCV gas (NG free) and natural gas diluted with nitrogen . . . 67
5.1 Schematic representation of a well mixed reactor module . . . 72
6.1 Diagram of the Chemkin model adopted to the Winnox-TUD combus-tor . . . 78
6.2 different Chemkin Symbols used in the model . . . 79
6.3 NH3 conversion to NO versus NH3 concentration in the LCV gas, a comparison between experiment and model . . . 80
6.4 NO in the flue gas versus NH3 concentration in the LCV gas, a
com-parison between experiment and model . . . 81 6.5 Temperature at the exhaust of the Winnox-TUD combustor versus
NH3 concentration in the LCV gas, a comparison between experiment
and model . . . 81 6.6 O2 at the exhaust of the Winnox-TUD combustor versus NH3
concen-tration in the LCV gas, a comparison between experiment and model 82 6.7 CO2 at the exhaust of the Winnox-TUD combustor versus NH3
con-centration in the LCV gas, a comparison between experiment and model 82 6.8 NO emission versus λ in the first stage, a comparison between
exper-iment and model . . . 84 6.9 NH3 conversion to NO versus λ in the first stage, comparison between
experiment and model . . . 84 6.10 NH3 conversion to NO versus λ in the first stage at different power
setting from the model . . . 85 6.11 NH3 to NO conversion versus NH3 in the first stage for natural gas
free LCV gas,a comparison between the experiments and the model . 86 6.12 NH3 conversion and CO versus heat loss in the primary stage,
pri-mary air temperature as parameter from the chemkin model using Kilpinen97 mechanism . . . 86 6.13 NH3 conversion versus primary air temperature with lambda primary
stage as parameter from the Chemkin model using Kilpinen97 mechanism 87 6.14 Thermal NOx versus lambda primary stage with primary air
tempera-ture as parameter from the chemkin model using Kilpinen97 mechanism 88 6.15 Adiabatic flame temperature versus λ primary stage for an LCV gas
containing natural gas . . . 91 6.16 NH3 to NO conversion versus λ primary stage for an LCV gas
con-taining natural gas, taking primary air temperature as a parameter . 91 6.17 NH3 to NO conversion versus λ primary stage for an LCV gas
con-taining natural gas, taking the LCV gas heating value as a parameter 92 6.18 Species mole fractions versus λ primary stage for an LCV gas
contain-ing natural gas . . . 92 6.19 Species mole fractions versus λ primary stage for an LCV gas
contain-ing natural gas . . . 93 6.20 Adiabatic flame temperature versus λ primary stage for a natural gas
free LCV gas . . . 93 6.21 NH3 to NO conversion versus λ primary stage for a natural gas free
LCV gas, taking primary air temperature as a parameter . . . 94 6.22 NH3 to NO conversion versus λ primary stage for a natural gas free
LCV gas, taking the LCV gas heating value as a parameter . . . 94 6.23 Species mole fractions versus λ primary stage for a natural gas free
LCV gas . . . 95
6.24 Species mole fractions versus λ primary stage for a natural gas free
LCV gas . . . 95
6.25 NH3 to NO conversion versus λ primary stage for an LCV gas contain-ing natural gas, a comparison between the two mechanims: Kilpinen 97 and SKG . . . 96
A.1 Cross section and the dimensions of the TUD-Winnox combustor . . 112
A.2 Configuration of the computational domain for the Winnox-TUD com-bustor . . . 112
A.3 CH4 mole fraction for λprimarystage=1, at primary stage . . . 113
A.4 CH4 mole fraction for λprimarystage=0.5, at primary stage . . . 113
A.5 CH4 profile throughout the length of the Winnox-TUD combustor at λprimarystage=0.5 and λprimarystage=1 . . . 114
A.6 NH3 mole fraction at λprimarystage=1 . . . 114
A.7 NO mole fraction at λprimarystage=1 . . . 115
A.8 NH3 mole fraction at λprimarystage=0.5 . . . 115
A.9 NO mole fraction at λprimarystage=0.5 . . . 116
A.10 . . . 116
A.11 . . . 117
A.12 Temperature profile at λprimarystage=1 and with the inner cylinder . 117 A.13 Temperature profile at λprimarystage=1, and without the inner cylinder 118 A.14 velocity vectors at the end part of the Winnox-TUD combustor . . . 119
A.15 CH4 profile for λprimarystage=1, with the inner cylinder . . . 120
A.16 CH4 profile for λprimarystage=1, without the inner cylinder . . . 120
A.17 NO profile for λprimarystage=1, with the inner cylinder . . . 121
A.18 Velocity vectors for the Winnox-TUD combustor . . . 122
List of Tables
1.1 Renewables share in (%). [8] . . . 5
2.1 Advantages and disadvantages of different gasifiers. [9] . . . 18
2.2 Composition of gas from commercial wood and charcoal gasifiers [10] 19 4.1 Volumetric fuel composition (natural gas diluted with nitrogen) . . . 48
4.2 fuel and air flow rates . . . 48
4.3 Volumetric fuel composition (natural gas diluted with nitrogen, power effect) . . . 51
4.4 fuel and air flow rates . . . 51
4.5 LCV gas at different heating values . . . 53
4.6 fuel and air flow rates . . . 53
4.7 LCV gas composition, LHV and power . . . 55
4.8 fuel and air flow rates . . . 55
4.9 LCV gas composition, LHV and power . . . 57
4.10 Fuel and air flow rates, NO injected in the LCV gas and λprimary . . 57
4.11 LCV gas composition, LHV and power . . . 59
4.12 fuel and air flow rates . . . 59
4.13 Composition, higher heating value of the LCV gas and power setting 60 4.14 Composition, higher heating value of the LCV gas in addition to power 61 4.15 Fuel and air flow rates . . . 62
4.16 LCV gas composition, LHV and power . . . 63
4.17 fuel and air flow rates . . . 63
6.1 Volumetric fuel composition (natural gas diluted with nitrogen) . . . 79
6.2 LCV gas containing natural gas: composition, flow rates of the differ-ent streams and the fuel heating value . . . 89
6.3 Natural gas-free LCV gas: composition, flow rates of the different streams and The fuel heating value . . . 90
A.1 Simulation of the effect of NH3 concentration on the conversion rate 110
A.2 Simulation of the effect of heating value of the LCV gas on the con-version rate . . . 110 A.3 Simulation of the effect of capacity on the conversion rate . . . 111 A.4 CFD model boundary conditions . . . 111
Chapter 1
Introduction
Fossil fuels are dominating the energy sector until now, but the growth in global energy consumption, which will nearly double in the next 20 years, (IEA, 2001), will result in an increase in the share of renewable energy sources in the coming period. Rising oil prices and the necessity to meet the Kyoto targets for CO2 reductions in addition to
the fact that fossil-fuels are not coming from sustainable sources, all those reasons and more are driving forces towards the development of new technologies related to the application of sustainable energy sources.
Biomass derived LCV gas is one of the best alternatives for fossil fuels for the long term. Being sustainable, and renewable, this privileged place of biomass is reinforced by its neutral characteristic regarding CO2 emissions, as the same amount of CO2
consumed by the crops during their growth is released later during the combustion process. Nevertheless, there are other side effects of using biomass derived LCV gas which could compromise the neutral effect. One of the main problems related to the application of biomass-derived LCV gas in gas turbines or boilers is related to the high content of FBN (fuel-bound nitrogen) which leads to high NOx emissions. The
FBN is mainly composed of NH3 and to a lesser level of HCN.
NOx has direct local consequences as acid rain, and contributes to global warming
as well as CO2. [11]. Although, in the period 1990-1998, Europe was successful in
reducing nitrogen containing gases e.g. 21 % for NOx, this reduction is mainly a
result of the economic situation in the east of Europe, and no significant reduction is made from emissions control. The NOx emissions are expected to increase in the
period until 2010. [12]
The following section presents some of the harmful effects of NOx to the human
health and to the environment.
1.0.1
NO
xeffect on health and environment
Because this thesis is dealing mainly with NOx reduction it has been judged
inter-esting to give an overview of the threats that NOx emissions cause to the human
2 Chapter 1. Introduction
health, environment and vegetation [13]. NOx causes a wide variety of health and
environmental impacts because of various compounds and derivatives in the family of nitrogen oxides, including nitrogen dioxide, nitric acid, nitrous oxide, nitrates, and nitric oxide.
• Ground-level Ozone (Smog) - is formed when NOx and volatile organic
com-pounds (VOCs) react in the presence of heat and sunlight. Children, people with lung diseases such as asthma, and people who work or exercise outside are susceptible to adverse effects such as damage to lung tissue and reduction in lung function. Ozone can be transported by wind currents and cause health impacts far from original sources. Millions of Americans live in areas that do not meet the health standards for ozone. Other impacts from ozone include damaged vegetation and reduced crop yields
• Acid Rain - NOx and sulfur dioxide react with other substances in the air to
form acids which fall to earth as rain, fog, snow or dry particles. Some may be carried by wind for hundreds of miles. Acid rain damages; causes deterioration of cars, buildings and historical monuments; and causes lakes and streams to become acidic and unsuitable for many fish
• Particles - NOx reacts with ammonia, moisture, and other compounds to form
nitric acid and related particles. Human health concerns include effects on breathing and the respiratory system, damage to lung tissue, and premature death. Small particles penetrate deeply into sensitive parts of the lungs and can cause or worsen respiratory disease such as emphysema and bronchitis, and aggravate existing heart disease.
• Water Quality Deterioration - Increased nitrogen loading in water bodies, par-ticularly coastal estuaries, upsets the chemical balance of nutrients used by aquatic plants and animals. Additional nitrogen accelerates ”eutrophication,” which leads to oxygen depletion and reduces fish and shellfish populations. NOx
emissions in the air are one of the largest sources of nitrogen pollution in the Chesapeake Bay.
• Global Warming - One member of the NOx , nitrous oxide, is a greenhouse
gas. It accumulates in the atmosphere with other greenhouse gasses causing a gradual rise in the earth’s temperature. This will lead to increased risks to human health, a rise in the sea level, and other adverse changes to plant and animal habitat.
• Toxic Chemicals - In the air, NOx reacts readily with common organic chemicals
1.1. World energy supply situation 3
• Visibility Impairment - Nitrate particles and nitrogen dioxide can block the transmission of light, reducing visibility in urban areas and on a regional scale in our national parks.
Oxides of nitrogen can also:
• Seriously injure vegetation at certain concentrations. Effects include:
– Bleaching or killing plant tissue. Causing leaves to fall. Reducing growth rate.
– Deteriorate fabrics and fade dyes.
• Corrode metals (due to nitrate salts formed from nitrogen oxides). • Reduce visibility.
• Oxides of nitrogen, in the presence of sunlight, can also react with hydro-carbons, forming photochemical oxidants. Also, NOx is a precursor to acidic
precipitation, which may affect both terrestrial and aquatic ecosystems. NOx and the pollutants formed from NOx can be transported over long distances.
This means that problems associated with NOx are not confined to areas where NOx
are emitted. Therefore, controlling NOx is often most effective if done from a regional
perspective, rather than focusing on sources in one local area.
Trying to bring the LCV gas to the standards required at the input of gas turbines or gas engines will require intensive gas cleaning upstream of the combustor, and may be some extra measures (i.e. SCR or SNCR) downstream the combustion system will be necessary. This additional effort to get the LCV gas inline with the combustion system inlet requirements will add an extra burden to the competitiveness of the new gas. Reducing cleaning to the minimum possible (i.e. particulate and hot gas cleaning) will bring the LCV gas again to be competitive and will enhance it’s position as green energy.
The application of biomass derived LCV gas as a main or a secondary fuel in gas turbines and gas engines is a very promising and challenging area for green energy. Many research topics related to the combustion of biomass derived LCV gas are currently under investigation.
1.1
World energy supply situation
4 Chapter 1. Introduction
Figure 1.1: The cycle of nitrates and ozone production from NOx and hydrocarbons
1.2. The need for renewable energy 5
Table 1.1: Renewables share in (%). [8]
Year 2000 2010 2020 2030
IEA 6.5 6 5.5 7.5
EIA 9.7 9 8.3 6.4
EC 12 10 8.4 7
average 9.4 8.33 7.4 6.96
In the white paper on energy (1995), the Dutch government has set a goal to increase the share of renewable energy of the total energy production to 10% in 2020 (1.5% in 2001).This corresponds to 270 PJ of renewable energy in 2020, while 53 PJ was targeted in 2003. From the 270 PJ, 120 PJ (40%) is considered to be realized with energy from biomass and waste. Fiscal measures, green funds and an energy tax have been introduced to create a market for renewable energy. The production of electricity produced as green energy by the end of 2000 was 1500 GWhe [9].
Table 1.1 is made out of the data shown in figure 1.2, which shows clearly that despite the increase in the energy produced from renewable sources, the share is expected to decrease from 9.4 % in the year 2000 to 6.96 % in the year 2030. This is essentially a consequence of the consumption which will almost double in the coming 25 years. These figures show clearly that fossil fuel will remain the dominating source of energy in the coming three decades.
1.2
The need for renewable energy
It is a fact that we are relying on fossil fuel in our energy sources see figure 1.2, however it is also a fact that we are running out of our reserves of fossil fuel. Another fact is the global warming problem caused mainly by green house gases trapping the sun heat on the earth surface. Those green house gases like CO2 and NOx are mainly
produced from the combustion of fossil fuels.
By increasing the share of renewable energy, we will preserve our reserves of fossil fuel for the future generations and preserve our atmosphere by decreasing green house emissions.
From what has been said, it is clear that developing renewable sources of energy is not optional but a vital must. One day those renewable sources will replace totally fossil fuels.
1.3
Barriers related to renewable energy application
6 Chapter 1. Introduction
• Technical barriers • Economic barriers • Social barriers
The technical barriers are related to the technologies associated with those renew-able energies and their implementation for commercial purposes. A lot of research work is conducted all around the world to overcome those problems, and a good progress has already been achieved and more and more progress is expected in the near future. Regarding the economic barriers, it is a bit related to the first item, since the cost of energy produced from renewable sources stays above other conventional sources. This barrier is also being solved due to the increase of oil and gas prices and the progress in the new technologies. This makes the cost of kWh of electricity pro-duced from renewables already competitive in the time being. More competitiveness is expected in the future when special taxes would be applied on conventional fuels. Regarding the social barriers, a lot has to be done in this respect, since the public should understand what are the main challenges facing us in terms of energy supply and also environmental issues directly related to our consumption of different fuels. Of course renewable energy is not that developed compared to fossil fuels, but it is a matter of fact to know that those fossil fuels in addition to the pollution they produce they are not renewable; all our wells of oil and gas will run dry one day.
1.4
Biomass as a source for renewable energy
There are many renewable sources of energy and Biomass is one of them and has a privileged rank among them. Biomass has been and it is still used in many place of the world as primary energy sources. The up-coming of oil with low prices brought biomass out of the competition, but the increase in oil prices and the imminent hazards of climate change due to green house gases brought back biomass in addition to other green energy sources back to the scene of competition. Biomass has a very attractive feature regarding CO2, since biomass crops consume the same amount of CO2 when
growing up as the amount they deliver during their combustion, therefore it is a closed system which result in CO2 neutral source of energy.
Biomass is used in power and heat plant in different ways, which could summarized below:
1.5. This thesis 7
• Fermentation
In this research work, gasification and the use of producer LCV gas are of interest. There is a great advantage of using producer gas in existing fossil fuels equipments of course after some minor adjustments to meet the LCV gas requirements. This will take advantage of the large inheritance of fossil fuels plants and equipment.
Gasification is a non complete combustion, i.e. combustion with a deficiency of oxygen where the product is a mixture of combustible and noncombustible compo-nents. The main combustible components are CO, H2 and CH4. The remaining part
of air blown gasification is mainly composed of nitrogen and carbon dioxide and other contaminants. In chapter 3, an overview is given about the different gasification processes.
Biomass-derived LCV gas is characterized by its low calorific value, from where the name LCV gas, i.e. Low Calorific Value gas. This has a great effect on the design of the combustors using those gases.
Regarding the combustion of biomass-derived LCV gas, beside the attractive pic-ture of a neutral CO2, the producer gas contains fuel bound nitrogen (FBN). This
FBN is converted to NOx during the combustion process. This FBN problem destroys
the whole picture we built about biomass, that is because NOx is a green house gas
and it harms the environment,the human health and the vegetation to a comparable level of CO2. This forms the reasoning for this thesis.
1.5
This thesis
The work presented in this thesis is an approach towards reducing fuel-NOx, which
is generated from fuel-bound nitrogen. The aim is to investigate to what extent one can reduce fuel-NOx with primary measures alone. It is meant to understand and
define the main factors controlling the conversion of FBN to fuel-NOx within the
combustion process and in a second approach optimizing those factors to result in the lowest possible fuel-NOx emissions.
The work presented here is only a part of the total work conducted throughout the ”Low NOx” project, where extensive experiments have been conducted with different
combustors and different fuel gases in order to explore the main parameters affecting the conversion of FBN to fuel-NOx. The work starts with a set of so-called first
experiments. It is a kind of exploratory work. Building on those primary results, a new combustor called TUD-Winnox has been developed. The TUD-Winnox combustor is presented in chapter 3 of this thesis. The TUD-Winnox combustor was the subject of an extensive experiments and modeling.
This research work is presented in 7 chapters.
8 Chapter 1. Introduction
Chapter 2 presents the prospects and problems related to the application of biomass-derived LCV gas in gas turbines and giving the approaches adopted or might be adopted to overcome those barriers towards fueling not only gas turbines but also boilers and gas engines with biomass-derived LCV gas.
Chapter 3 gets closer the combustion system and an overview is given about low NOx combustors design and it gives the specifications of the TUD-Winnox combustor
as well.
Chapter 4 presents the experimental investigation. First a description of the experimental techniques and set up is given and afterwards, sets of experimental results are presented and discussed.
Chapter 5 is more about theory of the combustion process. It is all about the mathematical models, flow and chemical models governing the combustion process.
Chapter 6 presents the modeling work in Chemkin. The strategy adopted in the modeling part is to start with a validation of some experimental results giving enough confidence in the model to tackle our problem and then extend the application of the model to situations non-covered by the experiments.
Chapter 7 presents the main conclusions from all the work conducted, and as it is always said: a research work starts with one question and ends with many, this is why a set of recommendations is given at the end of this dissertation for the follow up of this work.
The results of the CFD investigation is presented in appendix A.
1.6
Primary and secondary measures of reducing
fuel-NO
xBy primary measures, it is meant all measures that can reduce fuel-NOx by optimizing
the combustion process. This means that the design characteristics of the combustor are investigated to result in the lowest possible fuel-NOx. Many parameters could
affect the fuel-NOx emissions, amongst others are:
• Stoichiometry • Swirl intensity • Residence time
• Air staging or fuel staging • Heat loss
• Flame type: diffusion or premixed flame
1.7. Methodology 9
• FBN concentration in the fuel
• Fuel heating value(HHV)
• LCV fuel composition (especially CH4 content)
On the other hand, secondary measures are taken downstream of the combustion process. Here two main procedures could be cited: [15]
• Selective catalytic reduction (SCR)
• Selective non catalytic reduction (SNCR).
In the current research, the aim is to reduce the FBN conversion to NO by primary measures, i.e. only the combustion process is optimized to reduce the conversion. Nevertheless, LCV fuel properties are also investigated to quantify to what extent can the fuel properties affect the conversion.
1.7
Methodology
To tackle the problem of reducing the conversion of FBN to NOxin the newly designed
combustor, both experiments and modeling are used. The sketch in figure 1.3 presents a good approach of the methodology applied throughout this work.
1.7.1
Experimental
The experimental work started at the early stage of this work. In order to situate the problem and understand the main combustion factors controlling the conversion of FBN to NO, a series of first experiments are performed using an old version of a Winnox combustor designed initially to fit on a Rover gas turbine. The combustor has two stages with a depletion plate between the two stages. Natural gas doped with ammonia was fired, and CO, O2, and NO were measured at the exhaust. In
those experiments the effect of stoichiometry on the conversion rate of FBN to NO is investigated. The results were used as a base for the design of the Winnox-TUD combustor. The newly designed combustor is extensively tested in an experimental setup at the Energy Technology Section of the TU Delft. The main objective of the experiment is to reach and optimal combustion regime with low NOx emissions and
high efficiency.
10 Chapter 1. Introduction
Figure 1.3: Sketch of the methodology applied throughout this research work
1.7. Methodology 11
1.7.2
Modeling
Modeling accompanies the experiments. Both CFD and kinetics modeling are per-formed using the Commercial codes Fluent and Chemkin respectively. Modeling is the strong tool used to understand the main parameters controlling the conversion of FBN, to test the dominance of a factor over others and at last to extrapolate data in order to answer ”what if” questions. More attention was given to the kinetic modeling, this because this study deals with NOx emissions reduction, i.e. with the
Chapter 2
Biomass gasification and subsequent
combustion in gas turbines
2.1
Introduction
In airblown gasification processes, wood, charcoal and other biomass materials are gasified to produce so-called ”producer gas”. The biomass-derived gas is characterized by its low calorific value compared to fossil fuels, therefore it is known as Low Calorific Value (LCV) gas.
In a gasifier, the wood particle is exposed to high temperatures generated from the partial oxidation of the carbon, primarily. As the particle is heated, the moisture is driven off. This could range from below 10 percent to over 50 percent of the incoming fuel weight. Further heating of the particle begins to drive off the volatile gases. For wood, this volatile content could be as much as 75 to 80 percent of the total dry weight. Discharge of these volatiles will generate in addition to CO a wide spectrum of hydrocarbons ranging from methane to long-chain hydrocarbons comprising tars, creosotes and heavy oils.
The quality of gas generated in a system is influenced by fuel characteristics, gasifier configuration, and the amount of air, oxygen or steam introduced. The output and quality of the gas produced is determined by the equilibrium established when the heat of oxidation (combustion) balances the heat of vaporization and volatilization plus the sensible heat (temperature rise) of the exhaust gases.
The quality of the outlet gas (MJ/nm3) is determined by the amount of volatile gases
(H2, CO, CH4, C2...) [19].
Figure 2.1, shows a simple sketch of the biomass gasification principle, where it shows that gasification is not more than a partial oxidation of the fuel. About half oxygen needed for combustion is supplied to the gasifier. However, oxygen used for gasification could be the oxygen of the air in the case of air-blown gasification, or it could be pure oxygen in the case of oxygen blown gasification.
The biomass-derived LCV gas can find application in the domain of heat and
14 Chapter 2. Biomass gasification and subsequent combustion in gas turbines
Figure 2.1: Biomass gasification principle
energy production.
In this chapter, an overview about different gasification processes is presented.More attention is given to biomass-derived LCV gas and its application in heat and power generation. The approach of solving problems related to the application of LCV gas in gas turbines is highlighted. however more emphasis is given to the problems re-lated to that application and approaches to be considered in this thesis to tackle and overcome those problems [4].
2.1.1
Gasifier types - Advantages and Disadvantages
The gasification process aims to produce gas out of solid fuels. However, this is done in different ways, therefore a series of gasifiers are found . Each kind of gasification process has its advantages and disadvantages, regarding the resulting LCV gas in addition to the pollutants. This diversification gives a good opportunity to choose the most convenient gasification process for a given application. Below a short definition of different gasification process in addition to their advantages and disadvantages.
• Updraft gasifier
An updraft gasifier, figure 2.2, has clearly defined zones for partial combustion, reduction, and pyrolysis. Air is introduced at the bottom and acts as counter-current to fuel flow. The gas is drawn at higher location
In the updraft gasifier, the LCV gas contains high concentration of tar vapor, which could cause serious problems for the engine.
• Downdraft gasifier
In a downdraft gasifier, 2.3, air is introduced downward the bed. The LCV gas is drawn off at the bottom.
2.1. Introduction 15
Figure 2.2: Updraft gasifier Figure 2.3: Downdraft gasifier
bring the gasifier to working temperature with good gas quality is shorter than updraft gas producer.
• Fluidized bed gasifier
In a fluidized bed gasifier, figure 2.4, the bed material can either be sand or char, or some combination. The fluidizing medium is usually air; however, oxygen and/or steam are also used. When a fuel particle is introduced into the FBG environment, its drying and pyrolyzing reactions proceed rapidly, driving off all gaseous portions of the fuel at relatively low temperatures. The remaining char is oxidized within the bed to provide the heat source for the drying and de-volatilizing reactions to continue [19].
In those systems using inert bed material, the wood particles are subjected to an intense abrasion action from fluidized sand. This etching action action tends to remove any surface deposits (ash, char, etc.) from the particle and expose a clean reaction surface to the surrounding gases. As a result, the residence time of a particle in this system is on the order of only a few minutes, as opposed to hours in other types of gasifiers.
16 Chapter 2. Biomass gasification and subsequent combustion in gas turbines
• Circulating Fluidized Bed
The CFB gasifier, figure 2.5, has no distinct interface between the dense phase of fluidized sand and the freeboard (dilute particle phase). In fact the higher ve-locity fluidization regime means that there is a particle density gradient from the bottom of the gasifier to the top. Entrained media and char fines are recycled back to the gasifier via a retention cyclone. The higher velocity regime gives an alternative approach to increasing char residence time to promote higher efficiency gasification. However, detailed studies show similar carbon conver-sion limits in a CFB Gasifier ( 92%) compared to a BFB Gasifier. Alternative design approaches are being considered at various research institutions, how-ever, to improve carbon conversion through staged oxidation within the reaction column.
• Entrained Flow
2.1. Introduction 17
Figure 2.4: Fluidised bed gasifier Figure 2.5: Circulating fluidised bed gasifier
18 Chapter 2. Biomass gasification and subsequent combustion in gas turbines
Table 2.1: Advantages and disadvantages of different gasifiers. [9]
Gasifier type advantages disadvantages
Updraft Mature for heat Feed size limits
Small scale applications High tar yields Can handle high moisture Scale limitations
No carbon in ash Producer gas
Slagging potential
Downdraft Small scale applications Medium tar yield
Low particles Scale limitations
Low tar Producer gas
Moisture sensitive
Fluidised bed Large scale applications Medium tar yield
Feed characteristics Higher particle loading Direct/indirect heating
Can produce syngas
Circulating fluidised bed Large scale application Medium tar yield Feed characteristics Higher particle loading Can produce syngas
Entrained flow Can be scaled Large amount of carrier gas
Potential for low tar Higher particle loading
Can produce syngas Potentially high S/C
Particle size limits
2.1.2
Biomass derived LCV gas composition
The biomass derived LCV gas is a mixture of combustible and non-combustible gases. The composition depends strongly on the the gasified fuel and the gasification con-ditions. The heating value is in the range of 3 to 7 MJ/nm3. Carbon monoxide
2.2. The application of biomass derived LCV gas in gas turbines 19
Table 2.2: Composition of gas from commercial wood and charcoal gasifiers [10]
Component Wood gas (vol. %) Charcoal gas (vol. %)
Nitrogen 50-54 55-65
Carbon monoxide 17-22 28-32
Carbon dioxide 9-15 1-3
Hydrogen 12-20 4-10
Methane 2-3 0-2
2.2
The application of biomass derived LCV gas in
gas turbines
One promising application of the biomass-derived LCV gas is combustion in a gas turbine. Small and medium scale gas turbines are good candidates for this application. The best way foreseen to apply syngas in gas turbine is by integration of the gasification process with a gas turbine, this because of the difficulties associated with the storage and transport of producer gas, due to its volume and low calorific value
2.2.1
Gasifier-gas turbine integrated process
The integration of a gasification system with a gas turbine is very promising. There-fore, pressurized gasifier would be the best option, since no compressor is needed to bring the LCV gas to the pressure of the gas turbine. A 1.5 MWth pressurized
fluidized bed gasifier integrated with a gas turbine combustor was operating at the section Energy Technology of the Delft University. One of the problems related to the PFBG was the hot gas cleaning. [20], [21].
Regarding the cleaning of the producer gas upstream of the gas turbine combustor; hot gas cleaning could be coupled with a wet cleaning system (scrubber), and this is valid for both atmospheric and pressurized gasifiers. In the case of an atmospheric gasifier, a wet gas cleaning followed by a compressor is needed upstream of the gas turbine. Scrubbers remove very efficiently ammonia from the LCV gas up to 99.5 %, however, this will result in the loss of heat, in addition to the polluted waste water. [22]
It is clear that the cleaning of the producer gas from FBN upstream of the gas turbine combustors compromises the overall efficiency of the integrated system and results in additional costs required to clean the waste water.
20 Chapter 2. Biomass gasification and subsequent combustion in gas turbines
2.2.2
Problems related to the application of biomass derived
LCV gas in gas turbines
Unfortunately, the application of biomass derived LCV gas in gas turbines is not straight forward, this because of the strict requested properties of the fuels that can be burnt in a gas turbine.
LCV gas, in addition to its low calorific value, which will affect the flame tem-perature and the whole design of the combustion system, has problems related to the contaminants it contains. Therefore biomass derived LCV gas should undergo a whole series of cleanup in order to bring the gas to the accepted conditions for gas turbine application. Those contaminants are ranging from particulates not captured by the filters, to alkali, tars, and fuel-bound nitrogen. Therefore, the additional costs needed to accustom the producer gas to gas turbines, increases the overall investment and lowers the competitiveness of those gases.
2.3
The approach of solving the problems related
to the application of biomass derived LCV gas
in gas turbines
In this research work, an effort is made to overcome one of the main problems related to the application of biomass derived LCV gas, which is fuel-bound nitrogen conversion to NOx. The producer gas contains an important amount of fuel bound nitrogen,
mainly ammonia depending strongly on the fuel being gasified, it ranges from about 1000 ppmv for wood, to about 2.5% for the case of chicken manure.
The FBN will result, if no measure is taken, in a very high NOx emissions. The
approach in this research work, is to try to reduce the conversion of FBN to NOx
by optimizing the combustion process in a newly designed combustor called Winnox-TUD . The idea is to define the parameters controlling the conversion of FBN to NOx,
and by optimizing those parameters, the conversion to NOx will be minimized. The
scope is to reduce the conversion to result in NOx emissions lower than the permitted
2.3. The approach of solving the problems related to the application of biomass derived LCV gas in gas turbines21
Chapter 3
Combustor design for LCV gas
3.1
Introduction
The design of a gas turbine combustor has to satisfy many conditions and to serve many purposes. The main objective of the combustion system is to achieve high combustion efficiency, i.e the burnout of all the hydrocarbons while maintaining low pollutant emissions, namely NOx, SOx, unburned hydrocarbons (UHC)and soot and
to achieve the desired turbine inlet temperature. Pressure drop in the combustor is another problem , which should always be minimized. [23]
In our research work, in addition to the conventional characteristics of the com-bustors, the focus is to achieve low conversion of the FBN to NOx inside the newly
designed combustor (Winnox-TUD ), to get low fuel NOx emissions.
To achieve this goal, one should take into account the specific character of the conversion of FBN during the combustion process, and this will need a deep study of the kinetics of those reactions responsible of NOx production, in order to define the
main factors controlling that conversion. By defining those factors, one can optimize both the design and the combustion process.
After showing the problems related to the combustion of biomass derived LCV gas in chapter 2, this chapter will present an overview over NOx types and the specific
design features required for LCV gas combustors, and moreover, the main design characteristics of the Winnox-TUD combustor developed within the present research work.
3.2
NO
xtypes
Nitrogen oxides or NOx is the generic term for a group of highly reactive gases all of
which contain nitrogen and oxygen in various proportions. Many nitrogen oxides are
24 Chapter 3. Combustor design for LCV gas
colorless and odorless. Two of the most common oxides of nitrogen are: NO - Nitric Oxide and NO2 - Nitrogen Dioxide. In stationary source, combustion approximately 90% of NOx formed is NO. After NO leaves, a stack, in the presence of sunlight,
ozone, and VOCs, it becomes NO2, which (in extreme cases) appears as a
reddish-brown plume. It can cause photochemical smog and/or acid rain
How is NOxformed? There are several ways that NOxis formed in a furnace or gas
turbine combustor. The thermal fixation of atmospheric nitrogen and oxygen in the combustion air produces thermal NOx ; while the conversion of the chemically bound
nitrogen in the fuel produces fuel NOx . Prompt NOx is produced by the breakdown
of CH portions of methane and other hydrocarbons in the fuel and their subsequent combination with nitrogen in the air.
Below is a short description of the different types of NOx.
3.2.1
Thermal NO
xIt is produced from the oxidation of the atmospheric nitrogen in regions of high temperature in the flame. The rate of thermal NOx production is significant above
1850 K. Thermal NOx is described by the well known Zeldovitch mechanism shown
below.
For natural gas and light-distillate-oil firing, nearly all NOx emissions result from
thermal fixation.
The formation rate of thermal NOx is dependent on the reaction temperature, the
local stoichiometry, and the residence time. The fuel NOx formation mechanism is
more complex depending upon fuel pyrolysis and subsequent reaction between many intermediate nitrogenous species and the oxidant species.
It is Y.B. Zeldovich, 1946, who postulated the mechanism of thermal NO. which is represented by these elementary reactions [2]:
N2 + O → NO + N k1 = 1.8.1014exp (−318kJ.mol−1/RT )cm3/(mol.s) (1)
N + O2 → NO + O k2 = 9.0.109exp (−27kJ.mol−1/RT )cm3/(mol.s) (2)
N + OH → NO + H k3 = 2.8.1013cm3/(mol.s) (3)
The exponential dependence of thermal NOx on flame temperature is
demon-strated in figure 3.1. It’s clear from the figure that NO decreases very rapidly as temperature decreases [1].
Thermal NO is very dependent on flame temperature, therefore an increase in inlet air temperature would increase NO. The results from Rink and Lefebvre shown in figure 3.1 confirm that. [1].
The combustor residence time can also influence the NOx emissions, as shown in
3.2. NOx types 25
Figure 3.1: NOx formation as a function of time and temperature; p=1MPa, [1]
26 Chapter 3. Combustor design for LCV gas
Figure 3.3: Effect of residence time and equivalence ratio on NOx in a premixed
3.3. Low NOx combustors for natural gas 27
Figure 3.4: Prompt NO reaction path, [2]
Low thermal NOx combustors are well developed, concepts like Rich-Quench-Lean
(RQL),short-flame/quick-quench or direct injection of N2 in the fuel result in very low
NOx emissions. [24–27].
3.2.2
Prompt NO
xThe mechanism of prompt or Fenimore NO was postulated by C.P. Fenimore (1979), who measured [NO] above a hydrocarbon flate flame and noted that the [NO] did not approach zero as the probe approached the flame from the downstream side, as the Zeldovich mechanism predicts. The additional mechanism that is promptly producing NO at the flame front is more complicated than thermal NO, because the prompt NO results from the radical CH, which was previously considered to be an unimportant transient species that is generated through a complex reaction scheme, which can be shortly described in this way:
The CH, which is formed as an intermediate at the flame front only, reacts with the nitrogen of the air, forming hydrocyanic acid (HCN), which reacts further to NO, as shown in the reaction path in figure 3.4 [2]:
3.2.3
Fuel-NO
xThe conversion of FBN to NO results in so-called fuel-NO. The FBN is generally in the form of NH3 or HCN. The scope of this research work is all about fuel-NO
and how to reduce it. Figure 3.5 presents the reaction path diagram for oxidation of ammonia in flames presented in 1989 by Miller and Bowman [3].
3.3
Low NO
xcombustors for natural gas
28 Chapter 3. Combustor design for LCV gas
Figure 3.5: Reaction path diagram for oxidation of ammonia in flames, (Miller and Bowman, 1989). [3]
negligible in the case of natural gas (fossil fuels), which leads to negligible fuel-NOx.
This in contrast with biomass-derived LCV gas, where fuel NOx resulting from FBN
represents the main source of NOx emissions
for combustion of natural gas, several techniques are currently available or under further development, like: [28, 29]
• Water or steam injection
• Flue gas recirculation
• Air or fuel Staged combustion
• Lean premixed combustion
• Selective catalytic reduction (SCR)
• Selective non catalytic reduction (SNCR)
As the most part of FBN will oxidize to NO if nothing is done to reduce its conversion, the aim for low NOx combustion here is to strive for the possibility to
3.4. Low NOx combustors for LCV gas 29
Figure 3.6: Sketch of the Winnox combustor used in the preliminary experiments
3.4
Low NO
xcombustors for LCV gas
For the case of LCV gas, the problem of NOx is essentially that of fuel NOx, that is
the conversion of FBN to NOx during the combustion process. For this kind of fuels,
and because of the low flame temperature, there is no problem with thermal NOx.
Therefore, the measures to be taken to decrease the conversion of FBN to NOx are
different from those adopted in the case of thermal NOx. Although some of those
measures could be applied with some special adjustments for fuel-NOx abatement, as
it is the case of staged combustion. [30]
Hasegawa et al. investigated the direct injection of N2 in the product gas of an
oxygen-blown IGCC (Integrated Gasifier Combined Cycle) in order to reduce both thermal and fuel NOx. There investigation resulted in about 50 % reduction in the
conversion rate of NH3 to NOx. [31]
For NOx reduction, there are two main possibilities; primary measures which
in-clude the optimization of the combustion process or secondary measures which inin-clude SNCR and SCR. In the current work, only first measures for reducing fuel-NO are in-vestigated. In this study, our goal is to optimize the combustion process in a newly designed combustor to result in low NOx emissions. As the main source of NOx in the
case of LCV gas is FBN, which is mainly ammonia, one should first understand the mechanisms undergoing the conversion of ammonia to NOx during the combustion
30 Chapter 3. Combustor design for LCV gas
3.5
The newly designed combustor: Winnox-TUD
combustor
Based on a set of experiments and kinetic and CFD modeling, a new combustor was designed and constructed. In those first experiments and modeling efforts, we were looking for the main factors controlling the conversion of FBN to NOx. In the
preliminary experiments, a Winnox combustor designed initially to be mounted on a 45 kWe Rover gas turbine was used, a cross section of this combustor is shown in
figure 3.6
3.5.1
Design specifications
Figure 3.7 shows a cross section of the Winnox-TUD combustor. Figure 3.8 shows two pictures of the Winnox-TUD combustor. In the newly designed combustor, primary , secondary and tertiary air are controlled independently, leading to a good control of the equivalence ratio in each stage. In addition to that, a cylinder has been inserted inside the annular combustor. This cylinder has a crucial role, in restricting the second stage of the combustor to an annulus zone, where mixing is enhanced, and which also let the primary zone be slightly under pressure (i.e. the primary zone’s pressure is slightly higher than the second stage) which will ensure that there is no air i.e. oxygen leakage from the second stage to the primary zone. A tertiary air arrangement in the combustor, can have one of two roles; as burnout air in case both primary and secondary stages are sub-stoichiometric (overall), or as a diluting air in case the burnout is finished at the end of the second stage. The newly designed combustor uses Winnoxr patented swirling burner head. This burner head ensures a
good mixing of air and fuel in the primary zone, and an extended residence time. The Winnox-TUD can reach a power of 100 kWth, and could be fueled with different
types of LCV gas at different heating values. The ignition is made of an electrical spark, and throughout the experimental work, the Winnox-TUD was igniting easily and smoothly. The flame is controlled by both temperature measurement at the exhaust and the ionization system.
Since the Winnox-TUD is meant for experiments, it would need some improve-ments in terms of material selection and insulation of some parts in order to suit the commercial applications.
3.5.2
Why is the newly designed combustor giving low NO
x?
This is may be the most important question in this thesis. What’s behind low NOx
from this combustion, what’s making it new in its domain?? The main novel thing in this combustor, is combining the optimization of the reacting flow and the combustion process, trying to assemble all the factors in favor of low FBN conversion to NOx;
3.5. The newly designed combustor: Winnox-TUD combustor 31
Figure 3.7: Cross section of the newly designed combustor
32 Chapter 3. Combustor design for LCV gas
3.5. The newly designed combustor: Winnox-TUD combustor 33
emissions.
This means, that this research work will not end up with producing a new type of combustors for LCV gas, which can simply be used and give low NOx. The job is a
combined one, and it needs really a NOx solution for different applications, using of
Chapter 4
Experimental investigation
4.1
Introduction
A diversity of experiments were performed throughout this research work. As the first goal was to define the factors controlling the conversion of ammonia to NOx during
the combustion process, the work was started by preliminary experiments using a two-stage combustor designed originally for a Rover gas turbine of 45 kWe. The first
experiments took an exploratory form, where natural gas doped with ammonia was used as fuel. Although the fuel used in those experiments was very far from biomass-derived LCV gas, it was still useful in terms of determining the effect of combustion parameters on the conversion of NH3 to NOx in addition to the effect of ammonia
concentration in the fuel gas on the conversion. Those experiments gave a good insight into the conversion of ammonia. That insight was used later in the refining of the design of the Winnox-TUD combustor, see figure 3.7.
Further experiments were done using the newly designed combustor (Winnox-TUD). In all the performed experiments, the goal was to determine and understand the main factors controlling the conversion of ammonia to NO, in other words, the fate of ammonia throughout the combustor was tracked.
The Winnox-TUD combustor was fueled with LCV gas composed of natural gas diluted with nitrogen and with LCV gas from a mixing station, where natural gas, CO, H2, CO2, and N2 are mixed in proportions simulating biomass-derived LCV gas,
as produced by a pressurized air blown gasifier. In fact the LCV gas from the mxing station contained the same important components of the producer gas, however it was at relatively higher heating values. There is an other way of producing LCV gas by natural gas or propane reforming, see reference [32] for more details.
The LCV gas was doped with ammonia at different concentrations. In a few experiments, the LCV gas was doped with NO instead of NH3. The aim was to clarify
the way ammonia is cracked and converted throughout the combustion chamber and at different combustion settings.
The architecture of this chapter is such that it starts with giving an overview
36 Chapter 4. Experimental investigation
of related work from literature in order to situate this work in its research field, followed by an overview of the preliminary experiments, where one gets an idea of the mechanisms behind fuel NOx formation in a combustion chamber.
Furthermore, the series of experiments performed in order to determine the effect of the main combustion parameters on the conversion of fuel bound nitrogen to NOx
will be reported.
Although some experimental investigations install the so-called NO/NO2
convert-ers in order to measure total NOx, [35], in the current research work, it was found
that N2O and NO2 represent a non significant proportion of the total NOx compared
to the amount of NO, therefore only NO is taken into account and the conversion rates are calculated in terms of NH3 to NO.
In this chapter, the following items are highlighted:
• Effect of stoichiometry in the first stage on the conversion of ammonia to NO. • Effect of ammonia concentration in the LCV gas on the conversion of ammonia
to NOx
• The difference between natural gas doped with ammonia and LCV gas from a mixing station doped with ammonia in terms of ammonia conversion to NO. • The effect of methane on the conversion of FBN to NO.
• The effect of heating value of the LCV gas on the conversion of FBN to NO.
Temperature was measured at the exit and on the wall. The temperature ranged from about 700 ◦C to about 950 ◦C depending on stoichiometry in different stages.
CO level was always very low throughout this investigation. This is why CO is not presented in the different figures as it is always less than 20 ppm. This gives an idea of the high combustion efficiency.
4.2
Related experimental work from literature
Al-Shaikhly et al. (1994), investigated the conversion rate of NH3 to NO as a function
of NH3 content in an LCV synthetic fuel gas. The LCV gas is doped with NH3 ranging
from 0 ppm to 3200 ppm. Figure 4.1, shows clearly that the conversion rate of NH3
to NO decreases as the NH3 content in the fuel increases, however on the left figure
of 4.1 the absolute concentration of NO increases as NH3 increases.
Figure 4.2, shows the results of the investigation performed by Battista et al. [5]. A gas turbine combustor at a pressure of 2.0 MPa was fueled by coal derived LCV gas with an NH3 content ranging from 2200 to 4600 ppm. Figure 4.2 presents the
conversion of NH3 to NO as a function of the combustor exit temperature with four
