Investigations_about_Co_firing_of_gasified_herbaceous_biomass_in_an_integrated_gasification_combined_cycle

This paper focuses on combustion tests which were conducted to investigate the flame stability when product gas of biomass gasification is mixed with natural gas (NG) and subsequently burned in gas turbine (GT) combustors. Especially the product gas component H2 can alter the combustion properties of such fuel gas mixtures significantly compared to NG. Mixtures of product gas and methane premixed with preheated air have been investigated for fuel lean conditions at elevated pressure up to 15 bars in order to study combustion characteristics relevant for gas turbine based IGCC systems. Operational limits (flashback, lean blow out), emission characteristics (NOX, CO) and fundamental combustion properties - such as turbulent flame speed data - are derived, which can be used for preliminary design guide lines for such co-firing applications. A major finding is that if the co-firing rate is kept below a certain limit (20% heat input), co-firing with biomass derived syngas seems to be feasible in gas turbine combustion systems with only minor modifications. NOx emissions in co-firing mode are (slightly) higher than for pure NG especially if N-species are not limited to very low concentrations by appropriate product gas cleaning steps prior to combustion.

INVESTIGATIONS ABOUT CO-FIRING OF GASIFIED HERBACEOUS BIOMASS IN AN INTEGRATED GASIFICATION COMBINED CYCLE Peter Jansohn Paul Scherrer Institut (PSI) CH-5232 Villigen PSI SWITZERLAND Abstract This paper focuses on combustion tests which were conducted to investigate the flame stability when product gas of biomass gasification is mixed with natural gas (NG) and subsequently burned in gas turbine (GT) combustors. Especially the product gas component H2 can alter the combustion properties of such fuel gas mixtures significantly compared to NG. Mixtures of product gas and methane premixed with preheated air have been investigated for fuel lean conditions at elevated pressure up to 15 bars in order to study combustion characteristics relevant for gas turbine based IGCC systems. Operational limits (flashback, lean blow out), emission characteristics (NOX, CO) and fundamental combustion properties - such as turbulent flame speed data - are derived, which can be used for preliminary design guide lines for such co-firing applications. A major finding is that if the co-firing rate is kept below a certain limit (20% heat input), co-firing with biomass derived syngas seems to be feasible in gas turbine combustion systems with only minor modifications. NOx emissions in co-firing mode are (slightly) higher than for pure NG especially if N-species are not limited to very low concentrations by appropriate product gas cleaning steps prior to combustion. 1 Introduction There is a growing interest in generating "green" electricity via gasification processes. With the use of biomass as feedstock a CO2 emission reduction potential is realized due to its CO2 neutral life cycle. The increasing energy price for fossil fuels and upcoming energy consumption prospects generate growing interest for the introduction o f biomass technologies into the market. Wood has - as any other biomass - a critical yield where further exploitation will show negative ecological impacts. Thus other biomass fuels must be also considered as part o f the energy mix. One option is grass or hay (herbaceous biomass), which is not available as standard fuel today but could be made available for power generation via small to medium scale gasification processes. Co­ firing o f NG and biomass based product gas (also referred to as syngas) can be a superior, simple option compared to pure syngas fuelled gas turbines or biomass fired boiler plants [1]. Syngas co-fired natural gas combined cycles offer the opportunity to reach a high electrical efficiency (>40%) already for reasonably low biomass supply rates. Figure 1: Proposed biom ass syngas cofired com bined cycle plant It is comprehensible that it would be a significant challenge to supply larger size (>50MWei) gas turbine (GT) plants with attractive electric efficiencies only with biofuel as the supply logistics for a distributed resource such as biomass are just prohibitive in most cases. Hence, either a mono-fueled plant having a small size aero derivative GT with a moderate efficiency or a cofiring plant having a medium size GT with a higher efficiency seems most appropriate. The first of the two options was set up in the Varnamo project. A smaller GT (4.2 MWel) was used there on mono fuel derived from biomass gasification. The burner and combustion chamber had to be modified to adopt the machine to the specific fuel properties [2][3]. A annual yield of 32 kt (kilo tons) of grass was set to be a realistic amount for our studies. With this feedrate biomass might contribute 10-20 % (based on HHV) to the thermal input of NG co-fueled CC systems, which would correspond to a CC plant with a power output of around 80 MWel and an electric efficiency of around 50 %. The gray bar in Fig. 2 & 3 depicts the selected 32 kt/a of biomass feed, the other biomass supply data points correspond to the amount needed to provide the equivalent energy input to 10 % HHV of natural gas in form of biomass derived fuel gas to the gas turbine. Figure 2 depicts efficiency data for various single cycle gas turbines with an electric efficiency close to 40 % found above 20 MW (aero derivatives) and above 200 MW (heavy duty). Aero derivatives can reach higher efficiencies but also have higher specific costs and are more sensitive to certain fuel compositions. The graph shows that with the selected amount of biomass the GT engines which can be used are either aero derivatives with a high efficiency around 40 % or heavy duty turbines with a somewhat lower efficiency around 33 %. The data points are simple cycle gas turbines or combined cycle configurations, as specified by the manufacturers/OEMs [4]. In Figure 3 the efficiencies of the corresponding combined cycle (CC) arrangements are depicted. Here again the combined cycles using aero derivatives show higher efficiencies though the effect is not as distinct as in Figure 2. Figure 2: GT efficiencies and fuel supply vs. ISO base rating (power output) [4]. The biomass supply was calculated for 10% energetic fraction of product gas in a mixture with natural gas. Figure 3: GT efficiencies and fuel supply vs. ISO base rating [4]. The biomass supply was calculated for 10% energetic fraction of product gas in a mixture with natural gas. (N.B.: all the electric efficiency values reported in Fig. 2 & 3 are exclusive of any parasitic losses linked to the production of biomass based fuel gas, i.e. gasification and product gas cleaning; such losses generally amount to an efficiency reduction in the range of 5-10% points in efficiency) In a case study [5] for the Swiss Federal Office of Energy (OFEN) the estimation of the potential of grass for energy usage was performed by the means of a GIS (Geographic Information System) analysis. A map of the Swiss land use statistics ("Arealstatistik") was the basis of the analysis. Upon these basic data various filters like the slope of the countryside, elevation, precipitation and sun light exposition were used to get the final result of availability and local potential for biomass. The filters are important tools to restrict the availability of grass in a way such that there is no competition with nutrition users of the biomass (i.e. cattle). This way grass with high nutrition and protein properties will not be considered available for power generation. Mountainous terrain (steep slopes, high altitudes) has also been excluded in a similar fashion. On top of the biomass availability map specific areas were highlighted where the infrastructure for power plant installations was already established on a basic level. Figure 4 shows the harvest area needed for 32 kt of grass around four suitable locations for a power plant. The decision factors for those sites are the presence of a natural gas grid, an electrical grid, a transportation infrastructure as well as the potential of biomass. Two of the 4 sites considered (Eclepens and Ruthi SG) require larger collection distances as they are located in the vicinity of the country borders outside of which biomass availability was set to zero (i.e. all the biomass is collected only within Switzerland). The mean distance via road from the perimeter to the center is 38 km (equivalent to a radius of 27 km). The result show that there is a biomass potential for up to four centralized biomass gasification sites (32 kt/a of biomass feed, 24 MWth) but the decision to build such plants is finally determined by the social acceptance of this biomass based technology and the economic value (price) assigned to the fuel by the local farmers. ■i ■ B ill* i ■ F ta U a ir 241 r a * Figure 4: Harvest areas for 32 kt grass (annual yield) for four suitable locations in Switzerland [5] Lean premixed combustion is considered the state-of-the art technology applied in stationary gas turbines for highly efficient, low-emission power generation using NG. Due to the increased interest in the integration o f power generation with gasification processes, for CO2 mitigation and general use o f a variety o f solid fuels, the fundamental combustion properties o f synthetic fuels like biomass derived syngas are being investigated. The general concern with co-firing o f syngas is the evaluation o f what kinds o f modifications are needed for a modern low emission NG-fueled engine in order to safely handle the new fuel. The impact of a new fuel is of course depending on its chemical composition and properties. The development o f new materials suitable for higher turbine inlet temperatures increased the efficiency o f the gas turbines but also made them less robust against impurities in the gas from biomass gasification. Thus there is need for gas cleaning with efficient removal o f all contaminants that are harmful for turbines and other parts of the plant. Limits for impurities are set by the gas turbine manufacturers [6] and are usually very low. For alkalines like K and Na the limit is less than 0.25 mg/m3N [7]. Older gas turbines can cope much better with these elements due to more robust materials but on the other hand lack the high efficiencies of the new turbine generations. This paper relates to results generated in the field o f a GIS based grass potential analysis and grass gasification in a fluidised bed gasifier, but focuses specifically on results from research into basic combustion parameters like operational limits (lean blow out, flashback), emissions (NOx, CO) and fuel consumption rates (turbulent flame speed) o f fuel gas mixtures o f biomass derived syngas with methane. 2 Conditions for product gas / syngas combustion in gas turbines Based on a typical product gas composition derived from gasification of grass pellets in a air-blown fluidized bed gasifier [8], certain mixing ratios with methane/natural gas have been considered. Depending on the mixing ratio the resulting fuel gas mixture will show different combustion properties such as flame stability, emissions limits and concentration o f (potentially corrosive) contaminants. With respect to emission characteristics, NOX emissions coming from the conversion of fuel nitrogen species (like NH3) are an important topic when combustion in a gas turbine at fuel lean/excess oxygen conditions is considered. CH4 -=-CO -^ H 2 -B -C 02 - c - N2 HHV 30 - 25 o U-j TO +C -» 0) o c o o <D CO TO _c Q. CO TO CD : 20 > X X - 15 10 Methan substitution, %HHV Figure 5: Gas composition of mixtures of syngas and natural gas Mixing 10 % (HHV) of product gas with methane would result in a gas mixture given in Figure 5. This corresponds to 47 % (by volume) of product gas and 53 % of natural gas. This mixture is an example for a specific gasification experiment and will vary with the concentration of the product gas components. In this case a high air to fuel ratio had to be chosen to reduce the char formation in the reactor. For this reason the resulting heating value is very low (4.6 MJ/m3). At different conditions this value can reach up to 6.5 MJ/m3and the resulting HHV of the mixture will be much higher. This in turn also leads to a lower nitrogen dilution. 2.1 Experimental set-up Different syngas mixtures have been studied experimentally and numerically at gas turbine relevant conditions adopting typical air temperatures up to 770 K, pressures values up to 15 bar and different inlet velocities. In order to specifically evaluate the impact of a co-fired gas turbine engine a mixture of CH4-H2-CO (60-20-20 %vol) has been selected as a test case. This fuel gas mixture neglects the dilution species (N2, CO2) of the real mixtures given in Figure 5 but has been chosen in order to show in a pronounced way the differences in combustion properties between pure natural gas (almost 100% CH4) and natural gas/syngas (CO, H 2) mixtures derived from different gasification processes. As the diluents species (N 2, CO2) have no/minor chemical kinetic effects, there main influence is given in the thermal balance of flames and can be accounted for by comparing the results at equivalent (adiabatic) flame temperature conditions. Fig. 6: High pressure combustor test rig for combustion of syngas - natural gas mixtures The test rig used to derive the combustion properties of syngas (CO, H2) mixtures with/without natural gas is shown in Figure 6. The combustor, specifically designed to study turbulent, lean premixed flames, is capable of operating at a pressure up to 30 bars at adiabatic flame temperatures up to 2000 K. The combustion air can be electrically preheated up to 820 K. The combustor has a maximum thermal power of 400 kW. A complete optical access is provided by the cylindrical liner consisting of two coaxial quartz glass tubes (inner quartz glass tube diameter D = 75 mm) which are convectively air-cooled. The fuel/air pipe inlet diameter d is 25 mm. The flame is stabilized by the recirculation of hot flue gases due to the sudden expansion geometry. The fuel is injected coaxially into the air stream at 40 cm distance upstream of the combustor head. From this point up to the combustion chamber no other devices are present in order not to provide any flow disturbance which could enhance the flashback risk. Based on NOX emissions analysis, the fuel air mixing quality is believed to be ideal. In order to start the rig, a hydrogen torch igniter is used to ignite the premixed fuel/air mixture. More detailed descriptions of the test facility can be found in [9][10]. 3 Results of product gas / syngas combustion at gas turbine conditions The presence of H 2 and CO in the fuel mixture leads to an improved stability of the combustion process due to the faster chemistry characteristic of these fuels. Although the improved stability represents a big advantage in several operational conditions - i.e. in case of part load operation of the engine or during fast load changes - it brings with itself the very critical aspect of a strongly enhanced flashback risk. Flashback happens when the flame propagation velocity locally overtakes the convective velocity of the incoming fuel/air mixture. The flame front starts to propagate upstream of the region where it is supposed to anchor into zones characterized by low flow velocity (shear layers) and high fuel concentration (injectors). The fuel injector and its vicinity are usually not designed to handle high temperatures and can rapidly be destroyed by the flame. 3.1 Reactivity of product gas / syngas fuels An important parameter for turbulent premixed flames which gives guidelines for combustor design engineering is the turbulent burning velocity (ST). It provides information directly related to flashback risk, flame stability and flame length. ST is the propagation velocity of a combustion wave in a turbulent flow field; it represents an overall property of the particular mixture adopted, operative conditions (pressure, temperature) and turbulence level all together. The corresponding parameter for laminar flow conditions is the so called laminar flame speed (SL) which can be calculated as fundamental property of the fuel mixture upon the operative conditions. As it is not depending on the turbulence it is independent on the combustion device used. In Figure 7, laminar and turbulent flame speed values for four fuel mixtures at two different adiabatic flame temperatures are reported. The two temperatures were selected as representative for the lower and the upper operational limits for a conventional GT engine. The SL values were calculated using Cantera [11], a public domain code using a freely propagating laminar flame routine and adopting GRI3.0 as reaction mechanism [12]; ST values come from own experiments and are based on flame front detection via laser induced fluorescence of the OH radical [10]. The resulting turbulent flame speed data derived from a post processing procedure based on the obtained flame front surface, resemble integral mass averaged fuel consumption rates. To highlight the strongly different combustion characteristics, the turbulent flame speed data are normalized by the corresponding CH4 reference value. Depending on their hydrogen content, the fuel mixtures have the ability to propagate faster and this is the case in a similar fashion for both the considered temperatures. Looking at the turbulent flame speed values, the trend for SL is still respected which means higher H2 fraction in the fuel mixture corresponds to a higher ST. For low flame temperature (1650 K) though, the higher H2 content brings disproportionate acceleration which leads to values of almost an order of magnitude higher with respect to the corresponding value for CH4. This additional effect is mostly due to the strongly different physical properties (e.g. diffusivity) of H2 and was further explained and analyzed in [10]. Fig. 7: Comparison of laminar and turbulent flame speed data for different methane/syngas mixtures The presented analysis highlights that the co-firing mixture shows only a moderate difference with reference to CH4. Still, when compared with NG/CH4 its ST values show a increase in flame speed by up to a factor of 2 (at the low flame temperature limit, T=1650K). This fact has to be carefully taken into account when adapting burner and combustion chamber for the new fuel. The considerations derived by the analysis of co-firing a mixture of CH4-H2-CO (60-20-20 %vol) remain valid for any other similar mixture diluted with a small amount of CO2 or N2 (up to 20 %vol). In fact, as it is evident from the analysis of SL and ST presented in Figure 7, diluting a mixture of H2-CO (50-50%vol) with 20 %vol of N 2 does not modify the behavior with respect to the chemical kinetics. Such a dilution with N 2 just has an influence on the adiabatic flame temperature not affecting in any appreciable way the chemical properties [9][10]. The selected co-firing mixture composition is already a "limiting case" in terms of high H 2 and CO content with respect to the mixtures previously discussed in this paper. It represents an upper limit estimation of what could happen from the combustion point of view when operating the GT with a mixture proposed in Figure 5. In general it can be stated that while the co-firing possibility looks achievable for a GT requiring only minor modifications, to have an engine which offers full flexibility, operating at different conditions and with pure syngas, still represents a major challenge. However in the Varnamo project the GT burner and combustion chamber have been modified to process pure syngas successfully (even though with high N2 dilution) [2][3]. 3.2 Emission characteristics of product gas / syngas fuels Besides flame stability and safe operation of gas turbine burners, the emission characteristics (of CO and NOX) is another major topic of interest. NOX emissions for a wide, representative set of measuring points are presented for different fuel gas mixtures in Figure 8. The measured points belonging to the "pure" syngas mixtures are contained in a narrow band which shows an exponential trend (note the logarithmic scale on the NOX axis), according to the characteristics of thermal NOX production. Compared to the "pure" syngas mixtures, Figure 8 shows that the data belonging to the CH4 containing fuel mixture (simulating co-firing of syngas and natural gas) do not fall in the same range: they are characterized by lower NOX emissions. This can be attributed to two factors: in residence time effect due to shorter syngas flames (characteristic of the particular test rig used for the experiments) and different chemical kinetic pathways for hydrocarbon oxidation. These two aspects are described in details in [9]. The NOX emissions presented are representative of ideal conditions: perfectly premixed fuel/air mixture and no nitrogen species in the fuel. Moving from these conditions to a "real" GT scenario, none ideal mixture and NOX formation due to fuel-bound nitrogen has to be taken into account. A d ia b a tic F la m e T e m p e ra tu re [K ] Figure 8: Thermal NOX emissions for various mixtures at several operating conditions. CO emission was less than 10ppm for all data points recorded. Non-perfect mixing leads to the presence of fuel rich pockets in the fuel/air mixture with respect to the desired average equivalence ratio. The contribution of these pockets can significantly modify the global NOX emissions. The degree of mixing depends on the particular burner and also on the operative conditions. In a previous work [13], differences up to 20 ppm were found when comparing different injectors (providing non perfect mixing) with the ideal case of perfect mixing. Modern ultralow-emission GT systems are able to offer single digit NOX emissions, operating at temperatures within the range presented in Figure 8. The other source for NOX emissions to take into account for syngas combustion is the NOX formation due to fuel-bound nitrogen. NHX compounds (like NH 3) or HCN are known to be present in syngas mixtures coming from various gasification processes [7][14]. The oxidation of fuel-bound nitrogen to NO is rapid and occurs on a time scale comparable to the heat release reactions; the NO yields are found to be nearly quantitative for low concentrations (few hundred ppm) and almost temperature independent suggesting a low-activation energy step [15][16][17]. Within this work, the quantity of fuel-bound nitrogen has not been considered, but following an estimation is given how many parts per million of NHX are allowed in the fuel in order to match the current acceptable NOX emissions limits of 25ppm NOx (corrected for 15%Vol. of O2 in the exhaust). For lean premixed systems there will always be enough O2 available to oxidize NHX species and all the N released will contribute to form NOX. A equivalence ratio of 0.5 (excess air ratio of 2) can be seen as representative for the operating conditions of a GT fueled with the presented co-firing fuel mixture CH4-H2-CO (60-20-20 %vol). Considering a preheating temperature of the fuel/air mixture of 670 K (due to compression of the mixture to 15 bar), the adiabatic flame temperature for the above considered condition is slightly higher than 1800 K, which would result in about 2 ppm NOX in perfectly premixed condition (Figure 8). Accounting for the air dilution factor (13.33) and the change in volume due to chemical conversion (0.94), the amount of NHX species in the fuel mixtures has to be kept below 460 ppm (100 % conversion of NHX to NOX), in order to match 25 ppm of NOX as maximum emission limit. According to literature data [14] this value is exceeded by the actual concentration o f ammonia in typical gasifier product gases. Nevertheless, it is o f course possible to stay below the required emission limits by additional fuel gas cleanup and combustion at moderate flame temperatures [2]. Conclusions Co-firing of natural gas and biomass based product gas (also referred to as syngas) can be a superior, simple option compared to pure syngas fuelled gas turbines or biomass fired boiler plants. Syngas co­ fired natural gas combined cycles offer the opportunity to reach a high electrical efficiency (>40%) already for reasonably low biomass supply rates. In a case study for the Swiss Federal Office of Energy (OFEN) the estimation of the potential of grass for energy usage was performed by the means of a GIS (Geographic Information System) analysis. A annual yield of 32 kt (kilo tons) of grass was set to be a realistic amount for our studies. With this feedrate biomass might contribute 10-20 % (based on HHV) to the thermal input of NG co-fueled CC systems, which would correspond to a CC plant with a power output of around 80 MWel and an electric efficiency of around 50 %. The GIS study showed that the main part of the available grass is located in the northwestern part of Switzerland. Even though Switzerland can not be considered a country with rich grass resources (due to the alpine region) the results show that there is a biomass potential for up to four centralized biomass gasification sites (each for 32 kt/a of biomass feed, 24 MWth). The decision to build such plants is finally determined by the social acceptance of biomass based power generation technologies and the economic value (price) assigned to the fuel by the local farmers. The general concern with co-firing o f syngas and natural gas is the evaluation o f what kinds o f modifications are needed for a modern low emission NG-fueled engine in order to safely handle the new fuel. Depending on the mixing ratio the resulting fuel gas mixture shows different combustion properties such as flame stability, emissions limits and concentration o f (potentially corrosive) contaminants. With respect to emission characteristics, NOX emissions coming from the conversion o f fuel nitrogen species (like NH3) are an important topic when combustion in a gas turbine at fuel lean/excess oxygen conditions is considered. Combustion of syngas/natural gas mixtures was shown to produce high turbulent flame velocities but as long as 20 %vol. of dilution compounds (N2, CO2) and 20 %vol of H2 and CO are not exceeded, combustion chambers and fuel nozzles as designed for natural gas/C H should be capable of handling such fuel mixtures. Increasing amounts of nitrogen dilution will lower the combustion temperature and will finally jeopardize flame stability, higher hydrogen content will excessively accelerate the fuel consumption rate (expressed as turbulent flame speed) and consequently increase the risk o f a flashback and subsequent overheating/destruction of burner components. NOx emissions in co-firing mode are only slightly higher than for pure methane but only if N-species like ammonia (NH3) are limited to very low concentrations by appropriate product gas cleaning steps prior to combustion. Acceptable NOx emission limits (25ppm @ 15% O2) will only be achievable if N-species concentration in the syngas used is less than a few hundred ppm. Taking all of the above points into consideration, co-firing of gasification derived syngas with natural gas should obviously be seen as a viable option towards the integration o f biomass into larger size combined cycle power plants (even as a retro fit solution). Acknowledgement Most of the results reported in this paper have been derived by PhD students S. Daniele and J.Judex during their work towards a doctoral degree. The author is grateful to S. Biollaz on his initiatives in gasification of grass for power generation applications and to J.-L. Hersener for performing the study on the potential o f grass for energy use in Switzerland. Financial support for this work was granted by the Swiss Federal Office of Energy (OFEN) and the Competence Center for Energy and Mobility (CCEM), Switzerland. References [1] Peter Jansohn, Kraftwerk 2020 - A Swiss Initiative for High Effciency/ Low CO2 Gas Turbine Plants, PowerGen Europe, Madrid, 2007 [2] Stahl, K., Neergaard, M.: Experiences from the Varnamo IGCC demonstration plant, Power Production from Biomass III, VTT Symposium, Espoo (Finland), 1998 [3] Stahl, K., Varnamo - demonstration programme 1996 - 2000, SYDKRAFT, Malmo, 2000 [4] GTW, Gas Turbine World 2008 Performance Specs, 2008 [5] J.-L. Hersener, S. Biollaz, J. Judex, Richard Meyer, Verfugbarkeit von Gras fur Kombi-Kraftwerke in der Schweiz, Final Report, 2009 [6] GE Power Systems, Specification for Fuel Gases for Combustion in Heavy Duty Gas Turbines, GEI 41040G, 2002 [7] Kaltschmitt, M.; Hartmann, H., Energie aus Biomasse. Grundlagen, Techniken und Verfahren, ISBN 3540648534, Springer, 2001 [8] J. Judex, S. Daniele, Jean-L. Hersener, S. Biollaz, P. Jansohn, Investigations about cofiring of herbaceous biomass in an Integrated Gasification Combined Cycle, 3rd Freiberg Conference on IGCC and XtL Technologies, Dresden, May 2009, [9] Daniele, S., Jansohn, P., Boulouchos, K., Lean Premixed Combustion of Undiluted Syngas at Gas Turbine Relevant Conditions: NOx Emissions and Lean Operational Limits, GT2008-50265, ASME Turbo Expo, Berlin, 2008 [10] Daniele, S., Jansohn, P., Boulouchos, K., Flame Front Characteristic and Turbulent Flame Speed of Lean Premixed Syngas Combustion at Gas Turbine Relevant Conditions, GT2009-59477, ASME Turbo Expo, Orlando, 2009 [11] www.cantera.org [12] www.me.berkeley.edu/gri-mech/ [13] Pascal Mueller, Dieter Winkler, Timothy Griffin, Salvatore Daniele, Peter Jansohn, Combustion o f syngases: Fundamental combustion studies at gas turbine conditions, 4th Int. Gas Turbine Conference, The Future o f Gas Turbine Technology, Brussels, 2008 [14] De Jong W. Nitrogen compounds in pressurised fluidised bed gasification o f biomass and fossil fuels. PhD Thesis, Technische Universiteit Delft, The Netherlands, 2005. [15] Fenimore, C.P., Formation of nitric oxide from fuel nitrogen in ethylene flames, Combustion & Flame 19, pp. 289-296 (1972) [16] Glassman, I., Combustion, 3rd edition, ISBN 0-12-285852-2, Academic Press, New York, 1996 [17] Jansohn, P., Bildung und Abbau N-haltiger Verbindungen, insbesondere von HCN, NH3 und NO, in turbulenten Diffusionflammen, PhD Thesis, Technical University of Karlsruhe, Germany, 1991