Search

A computational study was performed for ethylene/air non-premixed laminar co-flow jet flames using an axisymmetric CFD code to explore the effect of oxygenation on PAH and soot emissions. Oxygenated flames were established using N2 diluted fuel stream along with O2 enriched air stream such that the stoichiometric mixture fraction (Ζst) is varied but the adiabatic flame temperature is not materially changed. Simulations were carried out using a spatially and temporally accurate algorithm with detailed chemistry and transport. A detailed kinetic model involving 111 species and 784 reactions and a fairly detailed soot model were incorporated into the code. Two different approaches, one with constant flame height and other with constant inlet velocity are comprehensively examined to bring out the effects of changes in flame structure and residence time on soot emissions with respect to Zst. With increase in Ζst, a drastic reduction in the formation of soot precursors (acetylene and benzene) and thus in soot emissions are observed. In the present study, oxygenated flames with Ζst ≥ 0.424 are considered as blue flames or completely soot free. For various oxygenated flames a C/O ratio between 0.45 and 0.6 is found to be most favorable for soot formation.

The extinguishment of propane cup-burner flames by a halon-replacement fire-extinguishing agent C6F12O (Novec 1230) added to coflowing air in normal gravity has been studied computationally and experimentally. The time-dependent, axisymmetric numerical code with a detailed reaction mechanism (up to 141 species and 2206 reactions), molecular diffusive transport, and a radiation model, is used to reveal a unique two-zone flame structure. The peak reactivity spot (i.e., reaction kernel) at the flame base stabilizes a trailing diffusion flame, which is inclined inwardly by a buoyancy-induced entrainment flow. As the volume fraction of the agent in the coflow is increased gradually, the total heat release increases up to three times due to unwanted combustion enhancement by exothermic reactions to form HF and CF2O in the two-zone trailing flame; whereas at the base, the flame-anchoring reaction kernel weakens (the local heat release rate decreases) and eventually the flame blows off. A numerical experiment, in which the C6F12O agent decomposition reactions are turned off, indicates that for addition of inert C6F12O, the maximum flame temperature decreases rapidly due to its large molar heat capacity, and the blow-off extinguishment occurs at ≈1700 K, a value identical to that for inert gases previously studied, while the reaction kernel is still burning vigorously. The calculated minimum extinguishing concentration of C6F12O in a propane flame is 4.12 % (with full chemistry), which nearly coincides with the measured value of 4.17 ± 0.30 %.

Higher fuel-burn rate and, hence, thrust from a rotating detonation engine (RDE) may be achieved through the use of larger channel widths for the annular tubes in which combustion occurs. However, while it is known that presence of solid walls influences the propagation of a detonation wave in a planar geometry, it is not clear how an increase in channel width would affect the characteristics of a rotating detonation wave in an annular tube. An experimental and numerical study is performed to understand how a rotating detonation wave evolves when the channel width is increased. Three annular tubes with identical inner-wall diameters but with channel widths of 7.8, 16.2, and 22.8 mm are considered. Reacting flowfields in the annular tubes are simulated using a three-dimensional CFD code that solves unsteady Navier–Stokes equations. Hydrogen–air mixture is used as fuel and is modeled with a two-step chemical-kinetics mechanism. Simulations have resolved the detailed structures of the rotating detonation waves including the Kelvin–Helmholtz instabilities associated with the fuel-products interfaces. Comparison of simulations obtained for different annular tubes suggest that as the channel width increases 1) the detonation front on the inner wall moves ahead of that on the outer wall and causes an inclination to the detonation wave between the walls, 2) the detonation front on the outer wall gets stronger, and 3) an oblique shock system establishes between the inner and outer walls. An experimental campaign has been conducted for verifying these predictions. The structure of the rotating detonation wave is recorded using OH* chemiluminescence. Measurements have qualitatively confirmed the predictions. Simulation data are then analyzed for understanding the physics behind the observed changes to the rotating-detonation-wave structure. The compression and expansion occurring at the outer and inner walls, respectively, are found to influence the rotating detonation wave.

The interaction between a laminar flame and a vortex is an important study for understanding the fundamentals of turbulent combustion. In the past, however, flame-vortex interactions have been investigated only for high-temperature flames. In this study, the impact of a vortex on a premixed double flame, which consists of a coupled cool flame and a hot flame, is examined experimentally and computationally using dimethyl ether/oxygen/ozone mixtures. The double flame is first shown to occur near the extinction limit of the hot flame. The differences between steady-state cool flames, double flames, and hot flames are explored in a one-dimensional counterflow configuration. The transient interactions between double flames and impinging vortices are then investigated experimentally using a micro-jet and numerically in two-dimensional transient modeling. It is seen that the vortex can extinguish the near-limit hot flame locally, resulting in a lone cool flame. At higher vortex intensities, the cool flame may also be extinguished after the extinction of the hot flame. It is found that there can be three different transient flame structures coexisting at the same time: an extinguished flame hole, a cool flame, and a double flame. Moreover, flame curvature is shown to play an important role in determining whether the vortex weakens or strengthens the cool flame and double flame.

The two-stage combustion, local extinction, and flame-edge oscillations have been observed in single-droplet combustion tests conducted on the International Space Station. To understand such dynamic behavior of initially enveloped diffusion flames in microgravity, two-dimensional (axisymmetric) computation is performed for a gaseous n-heptane flame using a time-dependent code with a detailed reaction mechanism (127 species and 1130 reactions), diffusive transport, and a simple radiation model (for CO2, H2O, CO, CH4, and soot). The calculated combustion characteristics vary profoundly with a slight movement of air surrounding a fuel source. In a near-quiescent environment (≤ 2 mm/s), with a sufficiently large fuel injection velocity (1 cm/s), extinction of a growing spherical diffusion flame due to radiative heat losses is predicted at the flame temperature at ≈ 1200 K. The radiative extinction is typically followed by a transition to the “cool flame” burning regime (due to the negative temperature coefficient in the low-temperature chemistry) with a reaction zone (at ≈ 700 K) in close proximity to the fuel source. By contrast, if there is a slight relative velocity (≈ 3 mm/s) between the fuel source and the air, a local extinction of the envelope diffusion flame is predicted downstream at ≈ 1200 K, followed by periodic flame-edge oscillations. At higher relative velocities (4 to 10 mm/s), the locally extinguished flame becomes steady state. The present 2D computational approach can help in understanding further the non-premixed “cool flame” structure and flame-flow interactions in microgravity environments.

Liftoff and blowout behavior of nonpremixed syngas flames is investigated using a time-accurate CFD code with a detailed description of transport and chemistry. Lifted flames are established in coflowing laminar jets using N2 dilution in the fuel jet. Results focus on the effects of syngas composition and temperature on the liftoff, stabilization, and the edge (triple) flame structure. For a given syngas mixture, as the N2 dilution exceeds certain value, the flame lifts off from the burner rim and propagates along the stoichiometric mixture fraction line, and its structure changes from diffusion to double flame. With further dilution, the flame liftoff height increases rapidly, the base structure transitions from double to triple flame, and its stabilization involves a balance between the triple flame speed and local flow velocity. The temporal evolution of propagating jet flame also exhibits a similar behavior. The transition from diffusion to double and then to triple flame is examined using state relationships in mixture fraction coordinate. As H2 fraction in syngas and/or temperature is increased, the N2 dilution required for flame liftoff and blowout increases. The ratio of the triple flame speed to the unstretched premixed flame speed also increases with the increase in H2 fraction. For H2 fraction above 30%, the flame liftoff and blowout become less sensitive to syngas composition and temperature.

Low-temperature (∼700 K) “cool” flames formed with n-heptane fuel were observed in the experiments conducted under microgravity. Recently, Won et al. demonstrated experimentally that such “cool” flames could be established under normal gravity. However, as ozone was added to air for supporting the flames in that experiment, the fundamental question on the formation of a self-supported n-heptane/air “cool” flame remained unanswered. A numerical investigation is conducted for (1) finding an answer to this question and (2) determining a procedure for establishing a “cool” flame. A standard opposing-jet burner is considered. Investigations are performed using a well-tested CFD code. A reduced mechanism that successfully captured the “cool” flames in microgravity is used. Calculations have demonstrated that a “cool” diffusion flame can be established with n-heptane under normal gravity without resorting to flame-speed promoters such as ozone, pressure or temperature. Using velocities of 4 and 7 cm/s and temperatures of 400 and 300 K for the fuel and air jets, respectively, a stable “cool” flame was obtained. Flame was ignited through increasing the temperature of the air jet to a value >724 K and then a self-sustained “cool” flame was obtained by decreasing the temperature back to 300 K. Parametric studies are performed on the applied gravitational force such that the heavy n-heptane flowed upward from the bottom nozzle or downward from the top nozzle. In general, flame remained flat with or without the gravity. However, when the fuel jet was at the bottom, it expanded faster and moved the flame closer to the fuel nozzle. “Cool” flame could not be stabilized for the same velocity and temperature conditions when fuel flowed from the top. “Cool” and normal diffusion flames obtained with identical flow conditions are compared. Limiting stretch rate for extinguishing the “cool” flame is determined.

Computations of cup-burner flames in normal gravity have been performed using propane as the fuel to reveal the combustion inhibition and enhancement by the CF(3)Br (halon 1301) and potential alternative fire-extinguishing agents (C(2)HF(5), C(2)HF(3)Cl(2), and C(3)H(2)F(3)Br). The time-dependent, two-dimensional numerical code used includes a detailed kinetic model (up to 241 species and 3918 reactions), diffusive transport, and a gray-gas radiation model. The peak reactivity spot (i.e., reaction kernel) at the flame base stabilizes a trailing flame, which is inclined inwardly by a buoyancy-induced entrainment flow. As the volume fraction of agent in the coflow increases gradually, the premixed-like reaction kernel weakens, thus inducing the flame base detachment from the burner rim and blowoff-type extinguishment eventually. The two-zone flame structure (with two heat-release-rate peaks) is formed in the trailing diffusion flame. The H(2)O formed in the inner zone is converted further, primarily in the outer zone, to HF and CF(2)O through exothermic reactions most significantly with the C(2)HF(5) addition. The total heat release of the entire flame decreases (inhibiting) for CF(3)Br but increases (enhancing) for the halon alternative agents, particularly C(2)HF(5) and C(2)HF(3)Cl(2). Addition of C(2)HF(5) results in unusual (non-chain branching) reactions.

Phosphorus-containing compounds (PCCs) have been found to be significantly more effective than CF(3)Br for reducing burning velocity when added to stoichiometric hydrocarbon-air flames. However, when added to lean flames, DMMP (dimethylmethylphosphonate) is predicted to increase the burning velocity. The addition of DMMP to lean mixtures apparently increases the equivalence ratio (fuel/oxidizer) and the combustion temperature, as a result of hydrocarbon content of DMMP molecule. Premixed flames studies with added DMMP, OP(OH)(3), and CF(3)Br are used to understand the different behavior with varying equivalence ratio and agent loading. Decrease of the equivalence ratio leads to the decrease of inhibition effectiveness of PCCs relative to bromine-containing compounds. For very lean mixtures CF(3)Br becomes more effective inhibitor than PCCs. Calculations of laminar burning velocities for pure DMMP/air mixtures predict the maximum burning velocity of 10.5 cm/s at 4.04 % of DMMP in air and at an initial temperature of 400 K. Adiabatic combustion temperature is 2155 K at these conditions.

Thermodynamic equilibrium calculations, as well as perfectly-stirred reactor (PSR) simulations with detailed reaction kinetics, are performed for a potential halon replacement, C3H2F3Br (2-BTP, C3H2F3Br, 2-Bromo-3,3,3-trifluoropropene), to understand the reasons for the unexpected enhanced combustion rather than suppression in a mandated FAA test. The high pressure rise with added agent is shown to depend on the amount of agent, and is well-predicted by an equilibrium model corresponding to stoichiometric reaction of fuel, oxygen, and agent. A kinetic model for the reaction of C3H2F3Br in hydrocarbon-air flames has been applied to understand differences in the chemical suppression behavior of C3H2F3Br vs. CF3Br in the FAA test. Stirred-reactor simulations predict that in the conditions of the FAA test, the inhibition effectiveness of C3H2F3Br at high agent loadings is relatively insensitive to the overall stoichiometry (for fuel-lean conditions), and the marginal inhibitory effect of the agent is greatly reduced, so that the mixture remains flammable over a wide range of conditions. Most important, the flammability of the agent-air mixtures themselves (when compressively preheated), can support low-strain flames which are much more difficult to extinguish than the easy-to extinguish, high-strain primary fireball from the impulsively released fuel mixture. Hence, the exothermic reaction of halogenated hydrocarbons in air should be considered in other situations with strong ignition sources and low strain flows, especially at preheated conditions.

Summary The effectiveness of phosphorus-containing compounds as gas-phase combustion inhibitors varies widely with flame type. To understand this behavior, experiments are performed with dimethyl methylphosphonate (DMMP) added to the oxidizer stream of methane–air co-flow diffusion flames (cup-burner configuration). At low volume fraction, phosphorus (via DMMP addition) is shown to be about four times as effective as bromine (via Br2 addition) at reducing the amount of CO2 required for extinguishment; however, above about 3000 μL/L to 6000 μL/L, the marginal effectiveness of DMMP is approximately zero. In contrast, the diminished effectiveness does not occur for Br2 addition. To explore the role of condensation of active phosphorus-containing compounds to the particles, laser-scattering measurements are performed. Finally, to examine the behavior of the flame stabilization region (which is responsible for extinguishment), premixed burning velocity simulations with detailed kinetics are performed for DMMP addition to methane–air flames. Analyses of the numerical results are performed to understand the variation in the inhibition mechanism with temperature, agent loading, and stoichiometry, to interpret the loss of effectiveness for DMMP in the present experiments. Copyright © 2015 John Wiley & Sons, Ltd.

Advances in combustor technologies are driving aircraft gas turbine engines to operate at higher pressures, temperatures and equivalence ratios. A viable approach for protecting the combustor from the high-temperature environment is to inject air through the holes drilled on the surfaces. However, it is possible that the air intended for cooling purposes may react with fuel-rich combustion products and may increase heat flux. Air Force Research Laboratory (AFRL) has developed an experimental rig for studying the flames formed between the injected cold air and the cross flow of combustion products. Laser-based OH measurements revealed an upstream shift for the flames when the air injection velocity was increased and downstream shift when the fuel content in the cross flow was increased. As conventional understanding of the flame stability does not explain such shifts in flame anchoring location, a time-dependent, detailed-chemistry computational-fluid-dynamics model is used for identifying the mechanisms that are responsible. Combustion of propane fuel with air is modeled using a chemical-kinetics mechanism involving 52 species and 544 reactions. Calculations reveled that the flames in the film-cooling experiment are formed through autoignition process. Simulations have reproduced the various flame characteristics observed in the experiments. Numerical results are used for explaining the non-intuitive shifts in flame anchoring location to the changes in blowing ratio and equivalence ratio. The higher diffusive mass transfer rate of hydrogen in comparison to the local heat transport enhances H2–O2 mixing compared to thermal dissipation rate, which, in turn, affects the autoignition process. While increasing the blowing ratio abates the differences resulting from non-equal mass and heat transport rates, higher concentrations of hydrogen in the fuel-rich cross flows accelerate those differences.

Experimental results are presented for changes in soot volume fraction resulting from the addition of a JP-8 surrogate and each of its components to a N 2 -diluted, C 2 H 4 co-flow diffusion flame. The surrogate, which consists of 77% n-dodecane and 23% m-xylene by volume, was designed to match the threshold soot index of a nominal JP-8 fuel. Total carbon flow rate was constant for all experiments; pre-vaporized liquid fuel was added at two different levels: 2.5% and 5% of the total carbon flow rate. Tests were conducted at pressures from 1 to 5 atm. The use of relatively small amounts of carbon from the liquid fuel resulted in a linear relationship between the peak soot volume fraction and the amount of carbon from the liquid fuel. In these experiments, the peak soot volume fraction was found to be vary with pressure according to a power-law relationship, consistent with prior work on pressure effects on soot. The surrogate fuel showed very similar trends to the JP-8, but yielded lower soot volume fractions. Simulation results for the flames with m-xylene capture the trends of increasing soot volume fraction with increasing carbon from the liquid fuel and with increasing pressure. However, the simulations show smaller increases than were observed in the experiments.

Numerical simulations of cup-burner flames in normal Earth gravity have been performed to study the combustion inhibition and unwanted enhancement by fire-extinguishing agents CF 3 Br (Halon 1301) and some potential replacements (C 2 HF 5 , C 2 HF 3 Cl 2 , and C 3 H 2 F 3 Br). A propane–ethanol–water mixture, prescribed for a Federal Aviation Administration (FAA) aerosol can explosion simulator test, was used as the fuel. The time-dependent, two-dimensional numerical code, which includes a detailed kinetic model (up to 241 species and 3918 reactions), diffusive transport, and a gray-gas radiation model, revealed unique two-zone flame structure and predicted the minimum extinguishing concentration of agent when added to the air stream. Despite striking differences in the flame shape, the agent effects were similar to, but stronger than, those in microgravity flames studied previously (for two of the agents). The peak reactivity spot (i.e., reaction kernel) at the flame base stabilized a trailing flame, which was inclined inwardly by a buoyancy-induced entrainment flow. As the volume fraction of agent in the coflow ( X a ) increased gradually: (1) the premixed-like reaction kernel weakened; (2) the flame base detached from the burner rim, oscillated (particularly for CF 3 Br), until finally, blowoff-type extinguishment occurred; (3) the calculated maximum flame temperature remained nearly constant (≈1800 K) or mildly increased; and (4) the total heat release of the entire flame decreased (inhibited) for CF 3 Br but increased (enhanced) for the halon replacements. In the trailing flame with C 2 HF 5 , a two-zone flame structure (with two heat-release-rate peaks) developed: in the inner zone, H 2 O (a product of hydrocarbon–O 2 combustion and a fuel component) was converted further to HF and CF 2 O through exothermic reactions occurring in the outer zone, where exothermic reactions of the inhibitor also released heat; CO 2 was formed in-between. Thus, addition of C 2 HF 5 resulted in unusual (non-chain branching) reactions and increased total heat release (combustion enhancement) primarily in the trailing diffusion flame.

Spatially resolved quantitative measurements of methyl radicals (CH3) in CH4/air flames at atmospheric pressure have been achieved using coherent microwave Rayleigh scattering from Resonance enhanced multi-photon ionization, Radar REMPI. Relative direct measurements of the methyl radicals were conducted by Radar REMPI via the two-photon resonance of the $$ 3p^{2} A_{2}^{\prime \prime } 0_{0}^{0} $$ state and subsequent one-photon ionization. Due to the proximity of the argon resonance state of 2s 22p 54f [7/2, J = 4](4+1 REMPI by 332.5 nm) with the CH3 state of $$ 3p^{2} A_{2}^{\prime \prime } 0_{0}^{0} $$ (2+1 REMPI by 333.6 nm), in situ calibration with argon was performed to quantify the absolute concentration of CH3. The REMPI cross sections of CH3 and argon were calculated based on time-dependent quantum perturbation theory. The measured CH3 concentration in CH4/air flames was in good agreement with numerical simulations performed using detailed chemical kinetics. The Radar REMPI method has shown great flexibility for spatial scanning, large signal-to-noise ratio for measurements at atmospheric pressures, and significant potential to be straightforwardly generalized for the quantitative measurements of other radicals and intermediate species in practical and relevant combustion environments.

Recent measurements of Barlow et al. in a bluff-body burner revealed that local atom balances were not conserved. They attributed the observed higher C/H atom ratios to the preferential diffusion of certain species and longer residence times in the recirculation zones (RZs). For verifying this hypothesis, detailed calculations for the reacting flows in the bluff-body burner were made using UNICORN code incorporated with GRI version 3.0 chemical kinetics. A lean mixture (Φ = 0.77) of methane and air was used as fuel jet. Time-dependent laminar simulations resulted in steady RZs for fuel jet velocities up to 7.0 m/s. Similar to experiments, calculations also revealed nonuniform distributions for C/H atom ratio in the flame at all heights. Results were compared to the data obtained for an unstrained 1D flame. Simulations for the bluff-body flame were repeated after setting diffusion coefficients of every species equal to that of O2. In the absence of preferential diffusion, C/H atom ratio was found to be equal to 0.25 everywhere in the flame. For understanding the effects of flame stretch on C/H atom ratio, calculations were made for the double flames formed between the opposing jets of premixed methane–air mixtures at different velocities. Distributions of C/H atom ratio in the RZ-supported nonpremixed flame were investigated through performing calculations for the Air Force centerbody burner fueled with ethylene. Both the attached and lifted-flame conditions were studied for understanding the role of RZs in enhancing C/H atom ratio. It was found that the hypothesis proposed by Barlow et al. in general, explains the higher C/H atom ratios found in bluff-body flames. However, RZs are not found to enhancing the preferential diffusion effects. Instead, the sharp velocity gradients near the fuel jet in the vicinity of the bluff-body surface increased the removal of H2O.

The effects of fire-extinguishing agents CF3Br and C2HF5 on the structure and extinguishing processes of microgravity cup-burner flames have been studied numerically. Propane and a propane–ethanol–water fuel mixture, prescribed for a Federal Aviation Administration (FAA) aerosol can explosion simulator test, were used as the fuel. The time-dependent, two-dimensional numerical code, which includes a detailed kinetic model (177 species and 2986 reactions), diffusive transport, and a gray-gas radiation model, revealed unique flame structure and predicted the minimum extinguishing concentration of agent when added to the air stream. The peak reactivity spot (i.e., reaction kernel) at the flame base stabilized a trailing flame. The calculated flame temperature along the trailing flame decreased downstream due to radiative cooling, causing local extinction at

Computation has been performed for a methane-air co-flow diffusion flame, in the cup-burner configuration, with a phosphorus-containing compound (PCC), dimethyl methylphosphonate (DMMP) or phosphoric acid, added to the oxidizer stream. The effectiveness of compounds in gaseous flame inhibition depends upon the additive and flame types, which lead to different reaction environments. The time-dependent axisymmetric numerical code, which includes a detailed kinetics model (77 species and 886 reactions), diffusive transport, and a gray-gas radiation model (for CH4, CO, CO2, H2O, and soot), has revealed the interaction of the gas-phase mechanisms of PCCs with the flame structure. The PCCs behave similarly with regard to flame inhibition: both raise the maximum temperature in the trailing flame, lower radical concentrations, and lower the heat-release rate at the peak reactivity spot (i.e., reaction kernel) at the flame base where the flame is stabilized. The mechanism of lowered radical concentrations is primarily due to catalytic cycles involving phosphorus species in both regions of the flame. For DMMP, which contains three methyl groups, the flame exhibited higher temperature and combustion enhancement in the trailing flame, with unique two-zone flame structure.

Experiments of autoignitive jet flames in a hot and vitiated coflow have previously shown various flame behaviors, spanning lifted flames to moderate or intense low oxygen dilution (MILD) combustion. For better understanding the behavior of flames in this configuration, regime diagrams and ignition delay results are presented from well-stirred reactor calculations across a wide range of operating conditions for methane and ethylene fuels. In conjunction with two-dimensional calculations, the importance of flame precursors and oxygen penetration across the reaction zone is revealed. It is found that widely accepted definitions and regime diagrams are inadequate to classify and reconcile the different flame behaviors that are observed experimentally. For accurate prediction of the ignition process, it is necessary to incorporate boundary conditions that capture minor species in the oxidizer. The role of fuel type also has a major impact on the ignition process and flame appearance. Refereed/Peer-reviewed

Experimental data and model results are presented for the effects of m-xylene on aromatic species and soot in a nitrogen-diluted ethylene flame over a range of pressures from 1 to 5 atm. The experimental approach was designed to investigate the effects of m-xylene as a perturbation to a base flame by keeping the amount of carbon added as m-xylene to 5% or less. The experimental results indicate that the maximum soot levels and those of small (1 or 2 rings) and large (3 or more ring) aromatic species are roughly first order with respect to the amount of m-xylene added. A chemical kinetic model was formulated, integrated into a 2-D modeling code, and used to simulate the effects of m-xylene addition and pressure on aromatic species and soot. The modeling results capture the general trends in concentration of soot and small aromatics as m-xylene concentration and pressure are varied. However, the model under-predicts the effect of m-xylene concentration and pressure on soot compared to the experimental results.

Several agents are under consideration to replace CF3Br for use in suppressing fires in aircraft cargo bays. In a Federal Aviation Administration performance test simulating the explosion of an aerosol can, however, the replacements, when added at sub-inerting concentrations, have all been found to create higher pressure rise than with no agent, hence failing the test. Thermodynamic equilibrium calculations as well as perfectly-stirred reactor simulations with detailed reaction kinetics, are performed to understand the reasons for the unexpected enhanced combustion rather than suppression. The high pressure rise with added C2HF5 or C3H2F3Br is shown to be dependent upon the amount of added agent, and can only occur if a large fraction of the available oxidizer in the chamber is consumed, corresponding to stoichiometric proportions of fuel, oxygen, and agent. Conversely, due to the unique stoichiometry of CF3Br, this agent is predicted to cause no increase in pressure, even in the absence of chemical inhibition. The stirred-reactor simulations predict that the inhibition effectiveness of CF3Br is highly dependent upon the mixing conditions of the reactants (which affects the local stoichiometry and hence the overall reaction rate). For C2HF5, however, the overall reaction rate was only weakly dependent upon stoichiometry, so the fuel–oxidizer mixing state has less effect on the suppression effectiveness.

Steady-state global chemistry calculations for 20 different flames were carried out using an axisymmetric Computational Fluid Dynamics (CFD) code. Computational results for 16 flames were compared with flame images obtained at the NASA Glenn Research Center. The experimental flame data for these 16 flames were taken from Sunderland et al. [4] which included normal and inverse diffusion flames of ethane with varying oxidiser compositions (21, 30, 50, 100% O2 mole fraction in N2) stabilised on a 5.5 mm diameter burner. The test conditions of this reference resulted in highly convective inverse diffusion flames (Froude numbers of the order of 10) and buoyant normal diffusion flames (Froude numbers ∼0.1). Additionally, six flames were simulated to study the effect of oxygen enhancement on normal diffusion flames. The enhancement in oxygen resulted in increased flame temperatures and the presence of gravity led to increased gas velocities. The effect of gravity-variation and oxygen enhancement on flame shape a...

The centerbody burner was designed with the objective of understanding the coupled processes of formation, growth, and burn-off of soot through decoupling them using recirculation zones (RZs). Experimentally it was found that the sooting characteristics of the centerbody burner could be altered dramatically via simple changes in the operating conditions. One of the interesting operating regimes in which a flame lifts off and forms a column of soot was identified when oxygen in the annulus air jet was reduced sufficiently. This paper describes the numerical studies performed to aid the understanding of lifted flames in the centerbody burner. A time-dependent, axisymmetric, detailed-chemistry CFD model is used. Combustion and PAH formation are modeled using the Wang–Frenklach (99 species and 1066 reactions) mechanism, and soot is simulated using a two-equation model of Lindstedt. Calculations have predicted the structure of the lifted flame very well. Two RZs [outer (ORZ) and inner (IRZ)] are formed between the fuel and air jets. A diffusion flame that is lifted-off the centerbody plate anchors steadily to the outer periphery of the ORZ. A near-perfect match between the computed and measured flame lift-off heights is achieved. RZs transport soot that is formed in the flame toward the face of the centerbody and create the soot column. Ethylene and its lighter fuel fragments that are formed in the RZs diffuse toward the annulus air jet and establish a mixing layer with the inwardly diffusing oxygen. Heat diffusing away from the RZs initiates autoignition reactions in the mixing layer. A flame with a triple-flame-base structure becomes established at a location where the ignition-delay time matches the residence time. Soot that is transported into the RZs is found to have a significant effect on the flame lift-off height. Numerical experiments are performed to aid the understanding of the relationship between soot and flame lift-off. Radiation from the soot decreases the temperature, slows the autoignition process, and increases the lift-off height. Soot oxidation consumes O and OH radicals, slows the autoignition reactions, and increases the lift-off height.

The centerbody burner was designed with the objective of understanding the coupled processes of soot formation, growth, and burnout. Fuel that issues from the center of the burner establishes two flame zones – one associated with the recirculation zone (RZ) and the other, with the trailing jet. The sooting characteristics in these two flame zones can be quite different because of variations in residence time and transport of reactants and products. Calculations performed for this burner operating under a partially premixed fuel jet suggested that soot in the RZ decreases and that soot in the trailing jet flame increases with the amount of premixing. An experimental and numerical study is performed to aid the understanding of these differences. A time-dependent, axisymmetric, detailed-chemistry computational-fluid-dynamics (CFD) model known as Unsteady Ignition and Combustion using ReactioNs (UNICORN) is used for simulating flames under different equivalence-ratio conditions. Combustion and PAH formation are modeled using the Wang–Frenklach (99 species and 1066 reactions) mechanism, and soot is simulated using a two-equation model of Lindstedt. A Lagrangian-based particle-tracking model is used for understanding the evolution of soot-like particles. Flame and recirculation-zone structures and soot in the experiments are identified using direct photographs taken with and without Mie scattering from soot particles as well as laser-induced-incandescence (LII) measurements. Calculations predict the structures of the partially premixed centerbody flames for various equivalence ratios reasonably well. Experiments confirm the predicted soot suppression in the RZs and enhancement of soot in the trailing jet flame when air is added to the fuel jet. It is found that flame movement in the RZ increases soot-particle burnout and, thereby, reduces the amount of soot within the RZ. As the flame moves closer to the fuel jet, more soot becomes entrained into the inner vortex. Motion of soot-like particles explained the spiral rings observed in the experiment. Increased particle burnout with partial premixing leads to shrinkage of soot spirals.

During some exploratory experiments performed on a centerbody burner it was observed that the sooting behavior of the burner could be altered dramatically without changing the fluid dynamics. One of the interesting operating regimes, in which the flame lifts off and forms a column of soot, was identified when oxygen in the annular flow was sufficiently reduced. More interestingly, within a narrow window of flow conditions, an unusual toroidal flame was observed near the base of the lifted flame. This paper describes the numerical and experimental studies performed for understanding this peculiar toroidal flame tube. A time-dependent, axisymmetric, detailed-chemistry CFD model (UNICORN) is used. Combustion and PAH formation are modeled using Wang–Frenklach mechanism and soot is simulated using a two-equation model of Lindstedt. Calculations have accurately predicted the steady lifted flame that is anchored to the outer edge of the recirculation zone. Lift-off height of the computed flame matched well with that of the experiment. A dynamic lifted flame is then established through periodically oscillating the annular flow. The edge of the lifted flame is found to dance along the outer periphery of the recirculation zone while vortical structures establish downstream of it. However, none of the calculations made with varying flow conditions or perturbations yielded a toroidal-flame structure near the base of the lifted flame. Surprisingly, when time-averaging was performed for the CH-radical distributions of the dancing flame a toroidal flame-like structure, very similar to that observed in the experiment, appeared near the flame base. Based on these calculations and high-speed movies of the experimental flame it is concluded that the observed toroidal flame is an optical illusion created through the natural time-averaging process of the human eye. Detailed structures of the computed oscillating flame are compared with the thermal images of the flame obtained using an infrared camera.

Extinguishment of laminar coflow diffusion flames in microgravity (μg) have been studied experimentally and computationally. The μg experiments were conducted using a methane cup-burner flame aboard the NASA Reduced-Gravity Aircraft. Transient computations with full methane chemistry and a gray-gas radiation model were performed to reveal the flame structure and extinguishment processes. In μg, as an inert gas (N2, He, or Ar) was added incrementally to the coflowing O2–N2 mixture with the initial oxygen volume fraction ( X O 2 ,ox ) of 0.21 at 101 kPa or 0.3 at 70.3 kPa, the flame tip opened, and the flame base gradually lifted off the burner parallel to the axis until blowout. The predicted minimum extinguishing concentration (MEC) of each agent in the oxidizing stream was in a reasonable agreement with the measurement. The measured MEC was nearly independent of the mean oxidizer velocity. In μg, the MEC for each diluent was nearly at a critical air-inertization point of the Coward–Jones flammability-limit curve (the maximum diluent concentration that sustains premixed combustion); whereas in earth gravity, studied previously, the flame prematurely blew off after oscillations at ≈70% of the critical condition. The maximum oxygen volume fractions at extinguishment (converted from the MECs) were nearly the same for X O 2 ,ox = 0.21 and 0.3 despite the different atmospheric pressures. The computation of the lifted flame with a sufficiently long fuel-oxidizer mixing time (≈0.2 s) revealed that: (1) a peak reactivity spot (i.e., reaction kernel), formed in the flame base, broadened laterally, thereby supporting a super-lean reaction wing on the oxidizer side and a trailing diffusion flame on the fuel side; (2) the flammable mixture layer was, nevertheless, radially thin (≈0.4 mm); and (3) the unburned mixture velocity at the flame base was comparable to the stoichiometric laminar flame speed found in the literature.

This research is motivated by the necessity to improve the performance of ultracompact combustors, which requires flame stabilization in small cavities. An extensive computational investigation on the characteristics of cavity-stabilized flames is presented. A high-fidelity, time-accurate, implicit algorithm that uses a global chemical mechanism for JP8-air combustion and includes detailed thermodynamic and transport properties as well as radiation effects is used for simulation. Calculations are performed using both direct numerical simulation and standard k-e Reynolds-averaged Navier-Stokes model. The flow unsteadiness is first examined in large axisymmetric and small planar cavities with nonreactive flows. As with previous investigations on axisymmetric cavities, multiple flow regimes were obtained by varying cavity length (x/D o ) : wake backflow regime, unsteady cavity vortex regime, steady cavity vortex regime, and compressed cavity vortex regime. However, planar cavities only exhibit steady cavity vortex and compressed cavity vortex regimes. Two opposed nonaligned air jets were positioned in this planar cavity: the outermost air jet in coflow with the mainstream flow (i.e., normal injection). The fuel jet was injected either in coflow, crossflow, or counterflow with respect to the mainstream flow. Flow unsteadiness was observed to be relatively small for coflow- and crossflow-fuel-jet injection. By reversing the air jet positions (i.e., reverse injection), the flow unsteadiness is promoted regardless of fuel jet positioning. Finally, the effect of combustion and cavity equivalence ratio (φ CAV ) on flame unsteadiness is addressed. With normal injection (reverse injection), low and high φ CAV leads to low (high) and high (low) flame unsteadiness, respectively. Based on these results recommendations are provided to designers/engineers to reduce flame unsteadiness in these cavities.

The impact of a variety of non-metallic fuel additives on soot was investigated in a collaborative university, industry and government effort. The main objective of this program was to obtain fundamental understanding of the mechanisms through which blending compounds into a fuel affects soot emissions. The research team used a suite of laboratory devices that included a shock tube, a well-stirred reactor, a premixed flat flame, an opposed-jet diffusion flame, and a high pressure turbulent reactor. The work reported here focuses on the effects of ethanol addition to ethylene on soot. The addition of ethanol led to substantial reductions in soot in all of the devices except for the opposed-jet diffusion flame. Modeling of the premixed flame and opposed-jet diffusion flame was used to obtain insights into the mechanism behind the opposing effects of ethanol addition in these two flames.

The laminar-flamelet concept for turbulent non-premixed combustion is examined through a study of vortex/flame interactions in a hydrogen/air, opposing-jet non-premixed flame. Vortices with sizes ranging from centimeter to sub-millimeter are injected toward the flame surface. The vortex-injection velocity is selected such that every interaction results in (1) flame extinction and (2) the ratio between the vortex-turnaround and the chemical time scales falling in the laminar-flamelet regime of turbulent combustion. The dynamic changes that occur to the flame structure during its interaction with the vortex are mapped onto a scalar-dissipation-rate scale. It is found that not only the scalar-dissipation rate but also the size of the vortex (eddy) is required for describing the flame-stretching process. The large centimeter-size vortex, irrespective of the propagation velocity, wrinkles and strains the flame before causing local extinction, which represents typical laminar-flamelet behavior. On the other hand, the small sub-millimeter-size vortex replaces the local fluid in the flame zone with fresh air and destroys the flame structure without causing any wrinkling or stretching of the reaction zone, which represents non-flamelet behavior. Interactions with millimeter-size vortices are found to deviate gradually from the flamelet behavior as the vortex size decreases. Vortex/flame interactions that do not follow laminar-flamelet behavior produce very high heat-release rates that are not observed in stretched planar flames. Similar deviations from flamelet behavior are observed in methane and ethylene flames. Since turbulent eddies are two to three orders of magnitude smaller than a millimeter-size vortex, caution must be exercised when applying laminar-flamelet theory to turbulent-combustion modeling.

Extinguishment of methane and polymethylmethacrylate (PMMA) diffusion flames by gaseous and water-mist fire-extinguishing agents has been studied experimentally and computationally using a cylindrical burner inserted downwardly into an upward coaxial air stream. Transient computations were performed for methane flames with full chemistry and a simple water-mist model to reveal the flame structure and suppression processes. For methane, as a gaseous agent (CO 2 or N 2 ) or water mist (≈35 μm number mean diameter) was added incrementally to the air stream: (1) a diffusion flame enveloping a porous bottom surface of the cylinder extinguished first, resulting in a flame with its base (edge) anchored at the leading edge of the side of the porous cylinder; and (2) the flame base oscillated, detached, drifted downstream, and extinguished eventually. For PMMA, the flame base attached to the fuel surface more closely than methane flames until detachment led to blowoff. The volume fractions of CO 2 or N 2 in the oxidizer stream at detachment and extinguishment were independent (or mild functions) of the mean oxidizer velocity ( U ox ). In computations, the initial envelope flame extinction led to extinguishment at the limits between the measured detachment and extinguishing conditions. The measured water flow rate at extinguishment was independent of U ox ; thus, the water mass fraction decreased dramatically with increasing U ox , suggesting that the water mist was transported to the flame zone independently of the air flow. The calculated water mass fraction at extinguishment, by contrast, decreased only mildly with U ox because water mist was modeled as a gaseous species. The computation of the near-limit flames revealed that: (1) extinguishment occurred when the peak temperature decreased to ≈1600 K; (2) adding CO 2 or H 2 O exhibited chemical effects on the flame structure; and (3) the stagnation-point-flow structure was nearly identical to that of a near-extinction counterflow diffusion flame.

The impetus for the present study is a perceived anomaly in which a vertically mounted ethylene jet diffusion flame used to study soot is observed to be steady; whereas similar flames used in other studies have been observed to flicker or oscillate at a frequency of about 12 Hz. This difference raises a question as to why the sooting flame is steady. A well-validated, Navier–Stokes-based, time-dependent numerical code is used to provide an answer to this question. A prediction is made, and later confirmed by experiments, of a phenomenon that is new to the authors and perhaps to the combustion community. That is, soot radiation can influence flame flicker resulting from a buoyancy-induced instability to such a point that the oscillation is completely suppressed. Experiments are used to validate the prediction and the result of numerical studies of the impact of soot on flame flicker. It is found that in a jet diffusion flame, the magnitude of flame oscillations (flicker) decreases with the amount of soot generated in the flame; however, the frequency of the oscillations does not change. Good agreement is obtained between measurements and computations for the dynamic behavior of flames with different soot levels. Both experiments and calculations yield a steady flame when the generated soot is sufficiently high and then produce the initial dynamic flame when the soot formation is restricted. The impact of soot on flame flicker is further studied by performing a simulation for the sooty flame after neglecting radiation from soot.

Microgravity experiments and computations have been conducted to elucidate stabilization and extinguishment mechanisms of methane diffusion flames, in the cup-burner configuration, with CO2 added gradually to a coflowing air or oxygen-enriched stream. The minimum extinguishing concentration of CO2 under low oxidizer velocities ( 0 ⩽ C ⩽ 1 ). The fuel-lean peak reactivity spot (the so-called reaction kernel) at the flame base stabilized the trailing diffusion flame. The calculated temperature along the trailing flame decreased downstream due to radiative cooling, leading to local extinction at

A time-dependent, two-dimensional, detailed-chemistry computational fluid dynamics model, known as UNICORN (unsteady ignition and combustion using reactions), is used for solving complex flame problems. The unique features incorporated in UNICORN for handling extremely large chemical kinetics with ease and efficiency are discussed. A submixture concept that is used for evaluating transport properties is described. This concept increases the computational speed by a factor of five for a 208-species mechanism and is expected to have even higher efficiency with larger mechanisms. An implicit treatment for certain reaction-rate terms applied during the solution of species-conservation equations is described. Moving the reaction-rate source terms to the left-hand side of the partial differential equations eases the stiffness problem that is typically associated with combustion chemical kinetics. Computational speeds are further improved in UNICORN by completely integrating the chemical-kinetics mechanisms with the solution algorithm. A software-generated computational fluid dynamics approach is used to avoid the tedious and near-impossible task of manually integrating a large chemical-kinetics mechanism into a computational fluid dynamics code. Several calculations demonstrating the abilities of the UNICORN code are presented. Chemical-kinetics mechanisms up to 366 species and 3700 reaction steps are incorporated, and simulations for unsteady multidimensional flames are performed on personal computers. Making use of the robustness and efficiency of the UNICORN code, detailed chemical mechanisms developed for JP-8 fuel are tested for their accuracy, and a parametric study on the role of parent species of a surrogate mixture in predicting flame extinguishment is performed. Ease of changing chemical kinetics in the UNICORN code is demonstrated through the investigation of effects of additives in JP-8 fuel.

The effects of H{sub 2} enrichment on the propagation of laminar CH{sub 4}-air triple flames in axisymmetric coflowing jets are numerically investigated. A comprehensive, time-dependent computational model, which employs a detailed description of chemistry and transport, is used to simulate the transient ignition and flame propagation phenomena. Flames are ignited in a jet-mixing layer far downstream of the burner. Following ignition, a well-defined triple flame is formed that propagates upstream along the stoichiometric mixture fraction line with a nearly constant displacement velocity. As the flame approaches the burner, it transitions to a double flame, and subsequently to a burner-stabilized nonpremixed flame. Predictions are validated using measurements of the displacement flame velocity. As the H{sub 2} concentration in the fuel blend is increased, the displacement flame velocity and local triple flame speed increase progressively due to the enhanced chemical reactivity, diffusivity, and preferential diffusion caused by H{sub 2} addition. In addition, the flammability limits associated with the triple flames are progressively extended with the increase in H{sub 2} concentration. The flame structure and flame dynamics are also markedly modified by H{sub 2} enrichment, which substantially increases the flame curvature and mixture fraction gradient, as well as the hydrodynamic and curvature-induced stretchmore » near the triple point. For all the H{sub 2}-enriched methane-air flames investigated in this study, there is a negative correlation between flame speed and stretch, with the flame speed decreasing almost linearly with stretch, consistent with previous studies. The H{sub 2} addition also modifies the flame sensitivity to stretch, as it decreases the Markstein number (Ma), implying an increased tendency toward diffusive-thermal instability (i.e. Ma {yields} 0). These results are consistent with the previously reported experimental results for outwardly propagating spherical flames burning a mixture of natural gas and hydrogen. (author)« less

We report the application of electronic-resonance-enhanced coherent anti-Stokes Raman scattering (ERE-CARS) for measurements of nitric oxide concentration ([NO]) in three different atmospheric pressure flames. Visible pump (532 nm) and Stokes (591 nm) beams are used to probe the Q-branch of the Raman transition. A significant resonance enhancement is obtained by tuning an ultraviolet probe beam (236 nm) into resonance with specific rotational transitions in the (v’=0, v”=1) vibrational band of the A2Σ+–X2Π electronic system of NO. ERE-CARS spectra are recorded at various heights within a hydrogen-air flame producing relatively low concentrations of NO over a Hencken burner. Good agreement is obtained between NO ERE-CARS measurements and the results of flame computations using UNICORN, a two-dimensional flame code. Excellent agreement between measured and calculated NO spectra is also obtained when using a modified version of the Sandia CARSFT code for heavily sooting acetylene-air flames (φ=0.8 to φ=1.6) on the same Hencken burner. Finally, NO concentration profiles are measured using ERE-CARS in a laminar, counter-flow, non-premixed hydrogen-air flame. Spectral scans are recorded by probing the Q1 (9.5), Q1 (13.5) and Q1 (17.5) Raman transitions. The measured shape of the [NO] profile is in good agreement with that predicted using the OPPDIF code, even without correcting for collisional effects. These comparisons between [NO] measurements and predictions establish the utility of ERE-CARS for detection of NO in flames with large temperature and concentration gradients as well as in sooting environments.

Experiments and calculations have been performed for a methane–air coflow diffusion flame, in the cup-burner configuration, with CF3Br or Br2 added to the air stream. The time-dependent, two-dimensional numerical code, which includes a detailed kinetic model and diffusive transport, has predicted the flame extinction within 4 or 8% for each. Analysis of the flame structure has allowed the mechanisms of flame weakening in the base and trailing flame regions to be compared. The agents CF3Br and Br2 behave very similarly with regard to flame extinguishment: both raise the temperature in the flame everywhere, as well as lower radical volume fractions in the trailing diffusion flame and at the peak reactivity spot (the “reaction kernel”) at the flame base where the flame is stabilized. The mechanism of lowered radical volume fractions is shown primarily to be due to a catalytic cycle involving bromine species in both regions of the flame, with small contributions from radical trapping by fluorinated species in the trailing diffusion flame. In the reaction kernel, the radical volume fractions are reduced more, and the catalytic radical recombination cycles are shown to be more effective as compared to in the trailing diffusion flame. At the latter location, the effectiveness of the agents is reduced because the hydrocarbon species, which are necessary for the regeneration of HBr, are scarce at the location of the peak radical volume fraction (i.e., at the flame zone), a limitation which does not exist in the reaction kernel, where there is good upstream mixing of the fuel and oxidizer because the base is lifted. That is, the premixed character of the reaction kernel actually allows the HBr in the catalytic cycle to be more effective there because of the effective overlap between the Br and the hydrocarbon species, which allows efficient regeneration of HBr.

The unsteady characteristics of oscillating methane diffusion flames in coflowing air diluted with CO2 in earth gravity have been studied experimentally and computationally. The measured frequency of flame flickering due to buoyancy-driven large-scale vortices was bi-modal; it jumped from ≈11 Hz to ≈15 Hz with an increase in the air velocity (at ≈11 cm/s). As CO2 was added into coflowing air gradually, the base (edge) of the flame detached from the burner rim, oscillated at half the flickering frequency, and blew off eventually. Numerical simulations with full chemistry predicted the internal flame structure and unsteady flame behavior: flame flickering, tip separation, base detachment, oscillation, and blowoff, in good agreement with the experiment. The mechanism of the edge diffusion flame oscillation was due to a cyclic series of events: (1) flame-base detachment and drifting downstream as a result of weakening due to dilution and a momentary increase in the entrainment-flow velocity associated with the vortex evolution, (2) fuel–air mixing in widened, lower-speed, wake space between the flame base and the burner rim, and (3) flame-base propagation through the flammable mixture layer back to the burner rim. A peak reactivity spot (reaction kernel) at the edge diffusion flame controlled the unsteady behavior through its dramatic changes in characteristics from the passively drifting to (premixed-type) propagating phase during a cycle. Because a mixing time of approximately 100 ms was required before propagation was enabled, a subsequent vortex evolved and passed. Thus, the flame-base oscillation was strongly coupled with the buoyancy-driven vortex evolution and the oscillation frequency was locked-in to half the flame-flickering frequency. The results have implications in turbulent flame structure; more specifically, the local extinction–mixing–reignition processes, in that the slow molecular mixing can become rate-limiting and the edge diffusion flame structure can be significantly different, depending on the phase in the process.

The counterflow flame system established between lean-methane–air and lean-hydrogen–air streams is investigated experimentally and numerically. A two-dimensional model known as UNICORN was used for the simulation. Detailed measurements for temperature and species concentrations were obtained along the centerline using Raman spectroscopy. A double-state behavior for this flame system was identified in the numerical simulations, which was later confirmed by the experiments. For the given flow conditions, the flame system can have either a single-flame or a double-flame structure depending on the way those conditions were achieved. Detailed comparisons were made between measurements and calculations for the two flame structures. Calculations for various lean methane–air mixtures and stretch rates were performed to understand the double-state behavior of counterflow premixed flames. It was found that the flame system exhibits double-state behavior only for leaner ( ϕ CH 4 0.74 ) methane–air mixtures. Aerodynamic and chemical structures of the flames in different stretch-rate regimes were analyzed. When stretch rate on the flame system is increased, the flame transitions from a double-flame to a single-flame structure due to aerodynamic-cooling process. When stretch rate is decreased, the flame does not transition back to the double-flame structure due to stretch effects on molecular diffusion. However, for ( ϕ CH 4 > 0.81 ) , decrease in stretch rate increases flame temperature due to lack of stretch-induced cooling and returns the flame structure to a double-flame one. For a narrow range of equivalence ratios (0.74–0.81) counterflow premixed flames exhibit a hysteresis property.

The effect of O 3 on C 2 H 4 /synthetic-air flame propagation at sub-atmospheric pressure was investigated through detailed experiments and simulations. A Hencken burner provided an ideal platform to interrogate flame speed enhancement, producing a steady, laminar, nearly one-dimensional, minimally curved, weakly stretched, and nearly adiabatic flame that could be accurately compared with simulations. The experimental results showed enhancement of up to 7.5% in flame speed for 11 000 ppm of O 3 at stoichiometric conditions. Significantly, the axial stretch rate was also found to affect enhancement. Comparison of the flames for a given burner exit velocity resulted in the enhancement increasing almost 9% over the range of axial stretch rates that was investigated. Two-dimensional simulations agreed well with the experiments in terms of flame speed, as well as the trends of enhancement. Rate of production analysis showed that the primary pathway for O 3 consumption was through reaction with H, leading to early heat release and increased production of OH. Higher flame stretch rates resulted in increased flux through the H+O 3 reaction to provide increased enhancement, due to the thinning of the flame that accompanies higher stretch, and thus results in decreased distance for the H to diffuse before reacting with O 3 .

A swirl-stabilized combustor utilizes recirculation zones for stabilizing the flame. The performance of such combustors could depend on the fuel used as the cracked fuel products may enter the recirculation-zones and alter their characteristics. A numerical study is conducted for understanding the effects of fuel variation on the combustion and unburned-hydrocarbon-emission characteristics of a laboratory swirl combustor. A time-dependent, detailed-chemistry CFD model UNICORN is used. Six binary fuel mixtures formulated with n-dodecane and n-heptane, m-xylene, iso-octane or hexadecane are considered. A semi-detailed chemical-kinetics model (CRECK-0810) involving 206 species and 5652 reactions for the combustion of these fuels is incorporated into UNICORN code. Calculations are performed for a fuel-lean condition, which represents cruise operation of an aircraft. Combustor flows simulated with different fuel mixtures yielded nearly the same flowfields and flame structures. Production of the intermediate cracked fuel species that are key for the final flame structure and emissions seems to be independent of the fuel used. This finding could greatly simplify the detailed chemical kinetics used for obtaining cracked products. As the cracked fuel species are completely consumed with in the flame zone, no emissions are observed at the combustor exit for the considered fuel-lean condition.