A counterflow diffusion flame for supercritical CO 2 combustion is investigated at various CO 2 dilution levels and pressures by accounting for real gas effects into both thermal and transport properties. The UCF 1.1 24-species mechanism is used to account the chemistry. The nature of important nonpremixed combustion characteristics such as Prandtl number, thermal diffusivity, Lewis number, stoichiometric scalar dissipation rate, flame thickness, and Damköhler number are investigated with respect to CO 2 dilution and pressure. The results show that the aforementioned parameters are influenced by both dilution and pressure; the dilution effect is more dominant. Further, the result shows that Prandtl number increases with CO 2 dilution and at 90% CO 2 dilution, the difference between the Prandtl number of the inlet jets and the flame is minimal. Also, the common assumption of unity Lewis number in the theory and modeling of nonpremixed combustion does not hold reasonable for sCO 2 applications due to large difference of Lewis number across the flame and the Lewis number on the flame drop significantly with an increase in the CO 2 dilution. An interesting relation between Lewis number and CO 2 dilution is observed. The Lewis number of species drops by 15% when increasing the CO 2 dilution by 30%. Increasing the CO 2 dilution increases both the flow and chemical timescales; however, chemical timescale increases faster than the flow timescales. The magnitudes of the Damköhler number signify the need to consider finite rate chemistry for sCO 2 applications. Further, the Damköhler numbers at 90% sCO 2 dilution are very small; hence, laminar flamelet assumptions in turbulent combustion simulations are not physically correct for this application. Also, it is observed that the Damköhler number drops nonlinearly with increasing CO 2 dilution in the oxidizer stream. This is a very important observation for the operation of sCO 2 combustors. Further, the flame thickness is found to increase with CO 2 dilution and reduce with pressure.

We report herein a computational study to characterize the effect of oxygenation on polycyclic aromatic hydrocarbons (PAHs) and soot emissions in ethylene diffusion flames at pressures 1–8 atm. Laminar oxygenated flames are established in a counterflow configuration by using N 2 diluted fuel stream along with O 2 -enriched oxidizer stream such that the stoichiometric mixture fraction ( ζ st ) is varied, but the adiabatic flame temperature is not materially changed. Simulations are performed using a validated fuel chemistry model and a detailed soot model. The primary objective is to enhance the fundamental understanding of PAHs and soot formation in oxygenated flames at elevated pressures. At a given pressure, as the level of oxygenation ( ζ st ) is increased, we observe a significant reduction in PAHs (benzene and pyrene) and consequently in soot formation. On the other hand, at a fixed ζ st , as pressure is increased, it leads to increased PAHs formation and thus higher soot emission. Both soot number density and soot volume fraction increase with pressure. The reaction path analysis indicates that at higher pressures, the C 2 /C 4 path becomes more significant for benzene formation compared to the propargyl recombination path. Results further indicate that the effectiveness of oxygenation in reducing the formation of pyrene and soot becomes less pronounced at higher pressures. In contrast, the effect of pressure on pyrene and soot formation becomes more pronounced at higher oxygenation levels. The behavior can be explained by examining the flame structure and hydrodynamics effects at different pressure and oxygenation levels.

In order to analyze the difference between the inverse diffusion flame (IDF) and normal diffusion flame (NDF) under various conditions, the emission spectra of OH* and CH* chemiluminescence in two dimensions measured by hyperspectral and ultraviolet (UV) cameras are described in this article. The results show that CH* mainly appears in the fuel side near the flame front, while OH* distribution can reflect the reaction region of flame. According to the OH* radial distributions in IDF and NDF, the flame can be divided into three parts: the core area of the flame, the transition region of the flame, and the developed region of flame. The peak intensity of CH* in IDF is higher than that in NDF. Moreover, the length of reaction region in NDF increases with O/C equivalence ratio ([O/C]e) until it reaches a steady value, while in IDF the length decreased with the increase of [O/C]e.

In this paper, we develop a linear technique that predicts how the stability of a thermoacoustic system changes due to the action of a generic passive feedback device or a generic change in the base state. From this, one can calculate the passive device or base state change that most stabilizes the system. This theoretical framework, based on adjoint equations, is applied to two types of Rijke tube. The first contains an electrically heated hot wire, and the second contains a diffusion flame. Both heat sources are assumed to be compact, so that the acoustic and heat release models can be decoupled. We find that the most effective passive control device is an adiabatic mesh placed at the downstream end of the Rijke tube. We also investigate the effects of a second hot wire and a local variation of the cross-sectional area but find that both affect the frequency more than the growth rate. This application of adjoint sensitivity analysis opens up new possibilities for the passive control of thermoacoustic oscillations. For example, the influence of base state changes can be combined with other constraints, such as that the total heat release rate remains constant, in order to show how an unstable thermoacoustic system should be changed in order to make it stable.

Lean-direct-injection (LDI) combustion is being considered at the National Energy Technology Laboratory as a means to attain low NO x emissions in a high-hydrogen gas turbine combustor. Integrated gasification combined cycle (IGCC) plant designs can create a high-hydrogen fuel using a water-gas shift reactor and subsequent CO 2 separation. The IGCC’s air separation unit produces a volume of N 2 roughly equivalent to the volume of H 2 in the gasifier product stream, which can be used to help reduce peak flame temperatures and NO x in the diffusion flame combustor. Placement of this diluent in either the air or fuel streams is a matter of practical importance, and it has not been studied to date for LDI combustion. The current work discusses how diluent placement affects diffusion flame temperatures, residence times, and stability limits, and their resulting effects on NO x emissions. From a peak flame temperature perspective, greater NO x reduction should be attainable with fuel dilution rather than air or independent dilution in any diffusion flame combustor with excess combustion air, due to the complete utilization of the diluent as a heat sink at the flame front, although the importance of this mechanism is shown to diminish as flow conditions approach stoichiometric proportions. For simple LDI combustor designs, residence time scaling relationships yield a lower NO x production potential for fuel-side dilution due to its smaller flame size, whereas air dilution yields a larger air entrainment requirement and a subsequently larger flame, with longer residence times and higher thermal NO x generation. For more complex staged-air LDI combustor designs, the dilution of the primary combustion air at fuel-rich conditions can result in the full utilization of the diluent for reducing the peak flame temperature, while also controlling flame volume and residence time for NO x reduction purposes. However, differential diffusion of hydrogen out of a diluted hydrogen/nitrogen fuel jet can create regions of higher hydrogen content in the immediate vicinity of the fuel injection point than can be attained with the dilution of the air stream, leading to increased flame stability. By this mechanism, fuel-side dilution extends the operating envelope to areas with higher velocities in the experimental configurations tested, where faster mixing rates further reduce flame residence times and NO x emissions. Strategies for accurate computational modeling of LDI combustors’ stability characteristics are also discussed.

Efficient turbulent combustion models are typically designed for the numerical simulation of two-stream problems, namely, the combustion of fuel in air. There are applications, however, where large amounts of a diluent such as water steam or recirculated exhaust gas is supplied to the combustor independent of fuel and air supplies. In such cases, classical approaches become quite time-consuming. In the present paper, a new three-stream flamelet model is presented, which is essentially an extension of the two-stream flamelet model for diffusion flames. Key points of the approach are the introduction of a second mixture fraction variable and the efficient establishment of the flamelet library. After presentation of the theory, the applicability of the new model is demonstrated by comparison with experimental results for the lift-off height of jet diffusion flames.

This paper describes the results of an experimental study to understand the influence of inlet flow disturbances on the dynamics of combustion process in bluff body stabilized diffusion flames of liquid petroleum gas and air. The results show the influence of weak disturbances created by the change in incoming pipe length on the amplitude of pressure oscillations and the phase angle between pressure and heat release. It is seen that the phase delay increases as the entry length increases. The rms value of pressure, however, generally falls with the increase in length. The phase angle is seen to be in the second quadrant, showing that the heat release oscillations damp the pressure oscillations. Therefore, the decrease in the phase angle results in the reduction in damping and hence an increase in pressure fluctuations. The dominant frequencies of combustion oscillations are found to be the low frequency oscillations, and the frequency of oscillations increases with a decrease in the inlet pipe length and an increase in the flow Reynolds number. It is suggested that such low frequency oscillations are driven by vortex shedding at the wake of the bluff body, which energizes the diffusion and mixing process.

A numerical investigation of the transient development of flame and soot distributions in a laminar axisymmetric coflowing diffusion flame of methane in air has been carried out considering the air preheating effect. The gas phase conservation equations of mass, momentum, energy, and species concentrations along with the conservation equations of soot mass concentration and number density are solved simultaneously, with appropriate boundary conditions, by an explicit finite difference method. Average soot diameters are then calculated from these results. It is observed that the soot is formed in the flame when the temperature exceeds 1300 K. The contribution of surface growth toward soot formation is more significant compared with that of nucleation. Once the soot particles reach the high temperature oxygen-enriched zone beyond the flame, the soot oxidation becomes important. During the initial period, when soot oxidation is not contributing significantly, some of the soot particles escape into the atmosphere. However, under steady condition the exhaust product gas is nonsooty. Preheating of air increases the soot volume fraction significantly. This is both due to more number of soot particles and the increase in the average diameter. However, preheating of air does not cause a qualitative difference in the development of the soot-laden zone during the flame transient period.

Sodium-water reaction (SWR) is a design basis accident of a sodium-cooled fast reactor (SFR). A breach of the heat transfer tube in a steam generator results in contact of liquid sodium with water. Typical phenomenon is that the pressurized water blows off, vaporizes, and mixes with the liquid sodium. It is necessary to quantify the SWR phenomena in the safety evaluation of the SFR system. In this paper, a new computer program has been developed and the SWR in a counterflow diffusion flame is studied by a numerical simulation and an experiment. The experiment is designed based on the numerical simulation so that the stable reaction flame is maintained for a long time and physical and chemical quantities are measured. From the comparison of the analysis and the experiment, there exist discrepancies that may be caused by the assumptions of the chemical reaction. Hence, a new experiment is proposed to enhance the measurement accuracy and to investigate the reason of the disagreement. The authors propose a depressurized experiment and show the preliminary result of the experiment. It is found that a stable chemical reaction flame is formed. With the depressurization, it is expected that the flame location can be controlled and the reaction region becomes thicker because of decrease in the reactant gas density.

A numerical study was carried out to understand the effect of CO enrichment on flame temperature and NO formation in counterflow C H 4 /air diffusion flames. The results indicate that when CO is added to the fuel, both flame temperature and NO formation rate are changed due to the variations in adiabatic flame temperature, fuel Lewis number, and chemical reaction. At a low strain rate, the addition of carbon monoxide causes a monotonic decrease in flame temperature and peak NO concentration. However, NO emission index first slightly increases, and then decreases. At a moderate strain rate, the addition of CO has negligible effect on flame temperature and leads to a slight increase in both peak NO concentration and NO emission index, until the fraction of carbon monoxide reaches about 0.7. Then, with a further increase in the fraction of added carbon monoxide, all three quantities quickly decrease. At a high strain rate, the addition of carbon monoxide causes increase in flame temperature and NO formation rate, until a critical carbon monoxide fraction is reached. After the critical fraction, the further addition of carbon monoxide leads to decrease in both flame temperature and NO formation rate.

Two-dimensional images of OH fluorescence, polycyclic aromatic hydrocarbons (PAHs) fluorescence, and laser-induced incandescence (LII) from soot were measured in a sooting diffusion flame. To obtain an accurate OH fluorescence image, two images were taken with the laser wavelength tuned to (“on”) and away from (“off”) the OH absorption line. An accurate OH fluorescence image was obtained by subtracting the off-resonance image from the on-resonance image. For the PAH fluorescence and LII measurements, temporally resolved measurements were used to obtain the individual images; the LII image was obtained by detecting the LII signal after the PAH fluorescence radiation had stopped and the PAH fluorescence image was obtained by subtracting the LII image from the simultaneous image of PAH fluorescence and LII. Based on the obtained images, the relative location of OH, PAH, and soot in the flame was discussed in detail. To investigate the PAH size distribution in a sooting flame using LIF, an estimation strategy for PAH size is proposed. Emission spectra were measured at several heights in the flame using a spectrograph. Since the emission wavelength of PAH fluorescence shifts toward longer wavelengths with increasing PAH size, the main PAH components in the emission spectra could be estimated. The results suggest that PAH grows and the type of PAH changes as the soot inception region was approached. Near the soot inception region, we estimated that the PAHs, which have over 16 carbon atoms, mainly constituted the emission spectrum.

A high-speed video camera was combined with a newly developed optical system to measure time resolved two-dimensional (2D) temperature distribution in flames. This diagnostics has been applied to measure the temperature distribution in an industrial size regenerative test furnace facility using highly preheated combustion air and heavy fuel oil. The 2D distributions of continuum emission from soot particles in these flames have been simultaneously measured at two discrete wave bands at 125 frames/sec. This allowed us to determine the temperature from each image on the basis of two-color 2D thermometry, in which the ratio of the 2D emission intensity distribution at various spatial position in the flame was converted into the respective 2D temperature distribution with much higher spatial resolution as compared to that obtainable with thermocouples. This diagnostic method was applied to both premixed and diffusion flames with highly preheated low oxygen concentration combustion air using heavy fuel oil. The results show that higher temperature regions exist continuously in the premixed flame as compared to the diffusion flame. This provided clear indication of higher NO emission from the premixed flame as compared to diffusion flames during the combustion of heavy fuel oil under high-temperature air combustion conditions. This observation is contrary to that obtained with normal temperature combustion air wherein diffusion flames result in higher NO x emission levels.

Thermal and chemical characteristics of the flames obtained from an industrial size regenerative combustion furnace have been obtained spectroscopically. The combustion characteristics of diffusion or premixed flames in the regenerative high-temperature air combustion facility have been examined using coal gas as the fuel. The fuel gas composition consisted of H 2 , hydrocarbon, CO, and N 2 . Monochromatic images of the flames have been observed in the emission mode using a CCD camera fitted with an optical band pass filter at the desired wavelength. The two-dimensional temperature distribution in the furnace has been determined using the two-line method by utilizing the Swan emission bands from within the flame. The emission intensity profiles of NO, as well as OH and CH radicals have also been observed spectroscopically. The results showed quite uniform two-dimensional temperature distribution and emission intensity of OH and CH radical species for the diffusion flame case as compared to the premixed case using high-temperature combustion air. The premixed flame case showed high local values and large fluctuations in the combustion zone for both emission intensity and temperature distribution. The temperature distribution of soot particles in the premixed flame was also determined using the two-color optical method. The results showed high local value of temperature, similar to that found for the gas temperature using signatures for C 2 species at two different wavelengths. In contrast the distribution of temperature for soot particles was different. The location of the maximum soot temperature shifted to downstream positions of the flame as compared to the maximum gas temperature regions measured from the C 2 species. The experimental results are discussed in conjunction with those obtained from the heat simulation analyses.

This paper describes the design and testing of a catalytically stabilized pilot burner for current and advanced Dry Low NO x (DLN) gas turbine combustors. In this paper, application of the catalytic pilot technology to industrial engines is described using Solar Turbines’ Taurus 70 engine. The objective of the work described is to develop the catalytic pilot technology and document the emission benefits of catalytic pilot technology when compared to higher, NO x producing pilots. The catalytic pilot was designed to replace the existing pilot in the existing DLN injector without major modification to the injector. During high-pressure testing, the catalytic pilot showed no incidence of flashback or autoignition while operating over wide range of combustion temperatures. The catalytic reactor lit off at a temperature of approximately 598 K (325°C/617°F) and operated at simulated 100% and 50% load conditions without a preburner. At high pressure, the maximum catalyst surface temperature was similar to that observed during atmospheric pressure testing and considerably lower than the surface temperature expected in lean-burn catalytic devices. In single-injector rig testing, the integrated assembly of the catalytic pilot and Taurus 70 injector demonstrated NO x and CO emission less than 5 ppm @ 15% O 2 for 100% and 50% load conditions along with low acoustics. The results demonstrate that a catalytic pilot burner replacing a diffusion flame or partially premixed pilot in an otherwise DLN combustor can enable operation at conditions with substantially reduced NO x emissions.

The prospects of reduced NO x emission, improved efficiency, stable, and oscillation-free combustion, and reduced construction costs achieved by an “Inverted Brayton Cycle” applied to midsize (0.5 to 5.0 MWe) power plants are discussed. In this cycle, the combustion products of an atmospheric pressure combustor are expanded in the gas turbine to subatmospheric pressure and following heat extraction are compressed back to slightly above the atmospheric, sufficient to enable a controlled fraction of the exhaust gas to be recirculated to the combustor. Due to the larger volume flow rate of the gas, the polytropic efficiency of both the turbine and compressor of this small machine is increased. Because of the low operating pressure and flue gas recirculation, both of which are instrumental to low NO x formation, the combustor can be operated in the diffusion flame mode; this, on the other hand, assures good flame stability and oscillation-free combustion over wide ranges of the operating variables. For the task of obtaining very low NO x formation, the well-tested multi annular swirl burner (MASB) is chosen. Recent computational and experimental development of the MASB by Siemens-Westinghouse as a topping combustor is discussed. It is shown that the MASB operated in rich-quench-lean mode is capable of single-digit NO x emission. The emissions are further lowered in the APGC by ambient pressure combustion, and by the injection of the recirculated gas in the quench zone of the combustor. Results of a computational optimization study of the ambient pressure gas turbine cycle (APGC) are presented.

Two new 14-step and 16-step reduced mechanisms for methane-air combustion were systematically developed by assuming the quasisteady state for 26–28 species in the starting mechanism. A series of comparison between the reduced mechanisms and the starting mechanism was carried out with the emphasis on their capabilities in predicting NO 2 formation and ignition delay. The two reduced mechanisms successfully capture the complex behaviors of NO 2 formation, which depends on the characteristic mixing time, pressure, and the contamination of hydrocarbon in air. The flame structure and NO x formation in diffusion flame were well predicted by the 16-step mechanism, while the 14-step showed less satisfactory performance on predicting prompt NO formation. The 16-step mechanism was shown accurate in predicting ignition delay over a wide range of equivalence ratio, temperature and pressure. The necessity of including CH 2 O , C 2 H 6 , C 2 H 4 , and HO 2 in the reduced mechanisms was discussed.

This paper describes reduced NO x diffusion flame combustors that have been developed for both simple cycle and regenerative cycle MS3002 and MS5002 gas turbines. Laboratory tests have shown that when firing with natural gas, without water or steam injection, NO x emissions from the new combustors are about 40 percent lower than NO x emissions from the standard combustors. CO emissions are virtually unchanged at base load, but increase at part load conditions. Commercial demonstration tests have confirmed the laboratory results. The standard combustors on both the MS3002 and MS5002 gas turbine are cylindrical cans, approximately 10.5 inches (27 cm) in diameter. A single fuel nozzle is centered at the inlet to each can and produces a swirl stabilized diffusion flame. The walls of the cans are louvered for cooling, and contain an array of mixing and dilution holes that provide the air needed to complete combustion and dilute the burned gas to the desired turbine inlet temperature. The MS3002 turbine is equipped with six combustor cans, while the MS5002 turbine is equipped with twelve combustors. The new, reduced NO x emissions combustors (referred to as a “lean head end,” or LHE, combustors) retain all of the key features of the conventional combustors; the only major difference is the arrangement of the mixing and dilution holes in the cylindrical combustor cans. By optimizing the number, diameter, and location of these holes, NO x emissions can be reduced considerably. Minor changes are also sometimes made to the combustor cap. The materials of construction, pressure drop, and fuel nozzle are all unchanged. The differences in NO x emissions between the standard and LHE combustors, as well as the variations in NO x emissions with firing temperature, are well correlated using turbulent flame length arguments. Details of this correlation are presented.

This paper describes a model used for the prediction of the formation of nitrogen oxides in modifications of an industrial diffusion flame, natural gas fueled can combustor. The flowfield inside the modified combustors is calculated using a Navier-Stokes solver. A fast chemistry assumption is used for modeling the heat release. Calculated turbulence parameters are then used for the calculation of the NO x formation rate in the post-processing mode with the aid of a flamelet model. The flamelet model permits the use of detailed kinetics with only minimal computational expense. The dependence of the NO x formation rate on the mixture fraction and scalar dissipation is calculated separately for each given condition. The validation of the model predictions is based on field test data taken earlier on several low NO x modifications recently applied to an industrial, reverse flow can type combustor. The reduced level of NO x emissions was achieved in these modifications by changes in the air distribution within the combustor liner. A comparison of the predicted and measured NO x emission levels shows good potential of the flamelet model.

This paper describes a reduced NO x diffusion flame combustor that has been developed for the MS5002 gas turbine. Laboratory tests have shown that when firing with natural gas, without water or steam injection, NO x emissions from the new combustor are about 40 percent lower than NO x emissions from the standard MS5002 combustor. CO emissions are virtually unchanged at base load, but increase at part load conditions. The laboratory results were confirmed in 1997 by a commercial demonstration test at a British Petroleum site in Prudhoe Bay, Alaska. The standard MS5002 gas turbine is equipped with a conventional, swirl stabilized diffusion flame combustion system. The twelve standard combustors in an MS5002 turbine are cylindrical cans, approximately 27 cm (10.5 in.) in diameter and 112 cm (44 in.) long. A small, annular, vortex generator surrounds the single fuel nozzle that is centered at the inlet to each can. The walls of the cans are louvered for cooling, and contain an array of mixing and dilution holes that provide the air needed to complete combustion and dilute the burned gas to the desired turbine inlet temperature. The new, reduced NO x emissions combustor (referred to as a “lean head end,” or LHE, combustor) retains all of the key features of the conventional combustor; the only significant difference is the arrangement of the mixing and dilution holes in the cylindrical combustor can. By optimizing the number, diameter, and location of these holes, NO x emissions were substantially reduced. The materials of construction, fuel nozzle, and total combustor air flow were unchanged. The differences in NO x emissions between the standard and LHE combustors, as well as the variations in NO x emissions with firing temperature, were well correlated using turbulent flame length arguments. Details of this correlation are also presented. [S0742-4795(00)01602-1]

In the numerical simulation of turbulent reacting flows, the high computational cost of integrating the reaction equations precludes the inclusion of detailed chemistry schemes, therefore reduced reaction mechanisms have been the more popular route for describing combustion chemistry, albeit at the loss of generality. The in situ adaptive tabulation scheme (ISAT) has significantly alleviated this problem by facilitating the efficient integration of the reaction equations via a unique combination of direct integration and dynamic creation of a look-up table, thus allowing for the implementation of detailed chemistry schemes in turbulent reacting flow calculations. In the present paper, the probability density function (PDF) method for turbulent combustion modeling is combined with the ISAT in a combustor design system, and calculations of a piloted jet diffusion flame and a low-emissions premixed gas turbine combustor are performed. It is demonstrated that the results are in good agreement with experimental data and computations of practical turbulent reacting flows with detailed chemistry schemes are affordable.