Self-sustained Flame Research Articles

Laser-induced spark ignition for DME–air mixtures with low velocity

This work studied laser-induced spark ignition for dimethyl ether (DME)–air mixtures with low velocities ranging from 10 cm/s to 100 cm/s, with an emphasis on the minimum ignition energy (MIE) and flame kernel development. Part of ignition experiments were conducted in a microgravity environment that was simulated in a laboratory-scale drop tower to eliminate natural convective flow. MIE was significantly affected by flow velocity. MIE first decreased with increasing flow velocity but then increased with further increasing flow velocity. The role of the flowing mixture was to act as a fuel supplier in a low-velocity flow, and in contrast, to increase convective heat loss in a high-velocity flow. The competition between these roles determined MIE. The effect of flow direction on MIE was also investigated by inclining the combustion chamber. The minimum value of MIE was the same in all flow directions. Therefore, changing flow directions was merely intended to vary the flow velocity by adding or removing buoyant flow to or from the forced flow, which caused the MIE distribution to shift from side to side. The flame kernel hardly developed against a forced flow and was often locally quenched. Flame deformation was noticeable as flow velocity increased. Quenching is comparable to the heat loss from the flame kernel. Therefore, flame kernel development should have a significant effect on MIE. Flame stretch was also found at a high velocity. A high-velocity flow did not allow the flame kernel to stably develop into a self-sustaining flame. Thus, higher energy was required to form a large-volume of flame kernel to overcome both conductive heat loss and the effect of flame stretch. Accordingly, flame kernel behaviors were very much a concern regarding the mechanism via which to determine MIE.

High-Speed Imaging of Forced Ignition Kernels in Nonuniform Jet Fuel/Air Mixtures

This paper describes experimental measurements of forced ignition of prevaporized liquid fuels in a well-controlled facility that incorporates nonuniform flow conditions similar to those of gas turbine engine combustors. The goal here is to elucidate the processes by which the initially unfueled kernel evolves into a self-sustained flame. Three fuels are examined: a conventional Jet-A and two synthesized fuels that are used to explore fuel composition effects. A commercial, high-energy recessed cavity discharge igniter located at the test section wall ejects kernels at 15 Hz into a preheated, striated crossflow. Next to the igniter wall is an unfueled air flow; above this is a premixed, prevaporized, fuel–air flow, with a matched velocity and an equivalence ratio near 0.75. The fuels are prevaporized in order to isolate chemical effects. Differences in early ignition kernel development are explored using three synchronized, high-speed imaging diagnostics: schlieren, emission/chemiluminescence, and OH planar laser-induced fluorescence (PLIF). The schlieren images reveal rapid entrainment of crossflow fluid into the kernel. The PLIF and emission images suggest chemical reactions between the hot kernel and the entrained fuel–air mixture start within tens of microseconds after the kernel begins entraining fuel, with some heat release possibly occurring. Initially, dilution cooling of the kernel appears to outweigh whatever heat release occurs; so whether the kernel leads to successful ignition or not, the reaction rate and the spatial extent of the reacting region decrease significantly with time. During a successful ignition event, small regions of the reacting kernel survive this dilution and are able to transition into a self-sustained flame after ∼1–2 ms. The low-aromatic/low-cetane-number fuel, which also has the lowest ignition probability, takes much longer for the reaction zone to grow after the initial decay. The high-aromatic, more easily ignited fuel, shows the largest reaction region at early times.

Electric field influence on the stability and the soot particles emission of a laminar diffusion flame

ABSTRACTInfluence of electric fields on flames has been studied for many years and the ionic wind constitutes the main explanation of the observed effects on the flame structure and pollutant emissions. However, previous works have been limited to small flames. The interaction mechanisms of an electric field with longer flames, involving both ionic wind and buoyancy are not fully identified. In the present paper, the effects of a D.C. electric field on a laminar 88-mm-long ethylene diffusion flame burning in ambient air are investigated. Based on the calculated electric field configuration, the influence of both downward and upward electric field is compared via imaging, electrical diagnostic and soot measurements. The application of a negative (directed downstream) electric field triggers a flickering instability and an electric instability at higher field strength, in which self-sustained flame oscillations of flame length directly affect ion current. Conversely, the flame is stabilized by a positive electric field. In-situ soot volume fraction measurements show that the electric field decreases the average soot volume fraction measured on a stable flame axis, whereas flame oscillations lead to a sooting flame.

A thermodynamic model for the plasma kernel volume and temperature resulting from spark discharge at high pressures

In today’s internal combustion engines assisted ignition is predominantly achieved by conventional spark ignition. The spark ignition starts with the breakdown phase after which complex flow dynamics govern the energy deposition into the gas. The condition of the activated plasma kernel resulting from spark breakdown, among others, is important for the development of the flame kernel to a self-sustained flame. A simple thermodynamic model is formulated based on insight gained from available detailed plasma simulations and experiments. Under the assumption that the problem is cylinder symmetrical and that the breakdown energy is supplied quasi-instantaneously as a line source, the extent of the hot plasma kernel and the time for pressure equalization with the surrounding can be given by exploiting the properties of radially expanding blast waves. The so-defined control volume is treated as an open system, and a simple energy balance leads to a relation for the mean kernel temperature. The plasma volumes and mean temperatures predicted by this simple model are shown to correspond well with published results in the open literature based on full plasma simulations of the spark discharge available for pressures ranging from 1 to 8 bar. The model allows drawing some conclusions about the requirements for experimental methods, such as calorimetry, frequently used to characterize the efficacy of the spark discharge produced by ignition systems and the influence of pressure and breakdown energy on kernel volume and temperature.

Large-scale flame structures in ultra-lean hydrogen-air mixtures

The paper discusses the peculiarities of flame propagation in the ultra-lean hydrogen-air mixture. Numerical analysis of the problem shows the possibility of the stable self-sustained flame ball existence in unconfined space on sufficiently large spatial scales. The structure of the flame ball is determined by the convection processes related to the hot products rising in the terrestrial gravity field. It is shown that the structure of the flame ball corresponds to the axisymmetric structures of the gaseous bubble in the liquid. In addition to the stable flame core, there are satellite burning kernels separated from the original flameball and developing inside the thermal wake behind the propagating flame ball. The effective area of burning expands with time due to flame ball and satellite kernels development. Both stable flame ball existence in the ultra-lean mixture and increase in the burning area indicate the possibility of transition to rapid deflagrative combustion as soon as the flame ball enters the region filled with hydrogen-air mixture of the richer composition. Such a scenario is intrinsic to the natural spatial distribution of hydrogen in the conditions of terrestrial gravity and therefore it is crucial to take it into account in elaborating risk assessments techniques and prevention measures.

Convection-driven melting in an n-octane pool fire bounded by an ice wall

An experimental study on an n-octane pool fire bound on one side by an ice wall was carried out to investigate the effects on ice melting by convection within the liquid part of the fuel. Experiments were conducted in a square glass tray (9.6 cm × 9.6 cm × 5 cm) with a 3 cm thick ice wall (9.6 cm × 6.5 cm × 3 cm) placed on one side of the tray. The melting front velocity, as an indicator of the melting rate of the ice, increased from 0.04 cm/min to 1 cm/min. The measurement of the burning rates and flame heights showed two distinctive behaviors; an induction period from the initial self-sustained flame to the peak mass loss rate followed by a steady phase from the peak of mass loss rate until the manual extinguishment. Similarly, the flow field measurements by a 2-dimensional PIV system indicated the existence of two different flow regimes. In the moments before ignition of the fuel, coupling of surface tension and buoyancy forces led to a combined one roll structure in the fuel. After ignition the flow field began transitioning toward an unstable flow regime (separated) with an increase in number of vortices around the ice wall. The separated regime started with presence of a multi-roll structure separating from a primary horizontal flow on the top driven by Marangoni convection. As the burning rate/flame height increased the velocity and evolving flow patterns enhanced the melting rate of the ice wall. Experimentally determined temperature contours, using an array of finely spaced thermocouples in the liquid fuel, were used to further investigate the two layer temperature structure; an upper layer (∼8 mm thick) with steep temperature gradient in the vertical direction and a layer of low temperature in deeper regions. A hot zone with thickness of ∼3 mm was present below the free surface corresponding to the multi-roll location. The multi-roll structure could be the main reason for the transport of the heat received from the flame toward the ice wall which causes the melting.

Flammability limits of iso-butanol/iso-octane/n-heptane blends

A set of experiments was carried out to determine the flammability limits (FL) of blends of iso-butanol and a surrogate fuel for gasoline at 154±11°C and ≈91.4kPa. The surrogate gasoline was a binary PRF mixture of 87% iso-octane and 13% n-heptane (PRF 87). The volumetric fraction of iso-butanol in the liquid fuel was varied from 0 to 0.25 at a step of 0.05. Flammability tests with pure fuels were also performed to confirm the reliability of the applied experimental procedure. The homogeneous air/fuel mixtures were defined as flammable when formed a self-sustained flame able to travel upward a 0.3m long open combustion tube. The lower FL of the blends of iso-butanol and PRF 87 (0.80−0.98%) were estimated correctly with the mixing rule of Le Chatelier, but the same simplified model failed to reproduce the measured upper flammability limits (5.10−5.61%).

Aspects of experimental investigations of laser plug ignition use at spark ignition engine

Nowadays, the necessity of pollutant emissions reduction brings to the fore new technologies developed for a better control of the combustion process. Ignition is the process of starting radical reactions until a selfsustaining flame has developed and strongly affects the formation of pollutants. Among the new technologies used for ignition and combustion control, the Laser Plug Ignition system is defined as an innovative technology which overcomes several limitations of conventional spark ignition. Laser ignition technology presents many advantages for engine operating control, engine performance improvement and pollutant emissions reduction. The objective of the paper is the experimental research of the laser ignition used in the spark ignition engine. The laser plug ignition system was mounted on an experimental spark ignition engine and tested at the regime of 90% load and 2800 rpm and dosage, λ = 1. The experimental results present the influence of laser ignition on different engine parameters compared to electric spark ignition. The influence on in-cylinder pressure, heat release rate, engine efficiency and pollutant emissions level is analyzed. Compared to a conventional spark plug, a laser ignition system assures efficient engine operation at lean dosages.

Influence of spark ignition in the determination of Markstein lengths using spherically expanding flames

Constant pressure outwardly propagating flame experiments in a spherical bomb are performed to examine the duration and radius over which spark ignition effects persist. This is motivated by the need to properly account for such effects in the measurement of laminar burning velocity and Markstein length using the spark ignited expanding flame technique. Ignition energy was varied and its effects on flame propagation in methane-air and isooctane-air mixtures were studied. The Markstein length of the mixture proved critical in the ignition energy dependency of flame propagation. For relatively high values, an underlying common variation of self-sustaining flame speed with radius can be identified by the rapid convergence of curves for different ignition energies. As the Markstein length decreases, low energy spark ignition is found to give rise to a distorted and wrinkled flame kernel. For such mixtures, due to the weak effect of stretch, the kernel subsequently develops into a non-spherically propagating flame. In these cases the spark ignition effect persists up to large radius. It is shown that using low ignition energy leads to a flame speed, during the development phase, which is higher than that of a self-sustaining spherical flame. It is further shown that if this effect is not accounted for, measurements of Markstein length using standard fitting techniques results in a large error. This problem is found to worsen as the Markstein length decreases, such that its apparent measured value becomes increasingly influenced by any distortions of the flame kernel produced by the spark.

Oxy-combustion of Low-Volatility Liquid Fuel with High Water Content

Following our previous studies of swirl-stabilized spray combustion of fuels that have a high water content, in this work, we investigate low-volatility fuels, specifically glycerol. Results show that, to obtain a self-sustained flame, the concentration of glycerol in water is much higher than that of ethanol (a high-volatility fuel) in water. However, even for glycerol concentrations as high as 60 wt %, the self-sustained flame is lifted from the nozzle and not attached, as in the ethanol case. The calculated stoichiometric adiabatic flame temperature (Tad) for 60 wt % glycerol is 2602 °C, whereas for 15 wt % ethanol (which is a very stable flame), it is only 673 °C. The two systems differ in that ethanol is more volatile than water, whereas glycerol is less. Thus, in the former case, the fuel is preferentially vaporized, whereas in the latter case, water is preferentially vaporized. Thus, flame stabilization is particularly challenging for fuels where the volatility is less than that of water when the water content of the spray is high. To assist flame stability for glycerol–water solutions, tert-butanol and ethanol were added in low concentrations. Experimental results revealed that stable attached flames can be obtained for 30 wt % glycerol in water when 8.3 wt % tert-butanol (B8.3/G30) or 10 wt % ethanol (E10/G30) is added under oxy-fired conditions. Both tert-butanol and ethanol have been shown to be particularly volatile in water when present at low concentrations because of their high vapor pressures and very high activity coefficients. The flame stability for B8.3/G30 and E10/G30 was characterized under 100 and 85% swirl flow conditions. Temperature measurements were conducted for four stable flames, and temperature contours were generated. For the stable flames, a hot zone exists in the near burner region, because of the preferential vaporization of the high-volatility fuel component, which provides heat to robustly vaporize the fuel and stabilize the flame.

Influence of scalar dissipation on flame success in turbulent sprays with spark ignition

A Direct Numerical Simulation (DNS) study is performed to determine a quantitative indicator of imminent global extinction in spray flames ignited by a spark. The cases under consideration have Group Combustion numbers sufficiently small that each droplet has an individual flame form around it, which subsequently merge. The structure of the flames is examined, including identification of non-premixed behaviour in the core of the flame and premixed flame fronts except in the presence of droplets, which cause strong non-premixed behaviour. The reaction progress variable c is studied and its dissipation rate is identified as being a key indicator of whether a flame will globally extinguish after being ignited by the spark. Specifically, immediately after the spark is deactivated, the volume containing the end of the flame front and hot products is studied in detail with respect to c. For successful flames, it is observed that regions of zero dissipation of c were predominantly restricted to the highest reaction progress variable (c>0.98), with zero probability within the range 0.95<c<0.98 and low probability within 0.9<c<0.95. In contrast, cases which subsequently extinguished had substantial probability of zero dissipation for 0.95<c<0.98. This region was a secondary structure separate from the main flame kernel that was unable to evaporate sufficient liquid to create a self-sustaining flame and therefore contributed to the subsequent quenching of the flame. In the successfully-burning case under consideration, this region was part of the main flame structure. The low reaction rate contributed to a thickened flame structure near the hot core, which reduced the heat transfer to the flame front and prevented effective evaporation and preheating of the fluid ahead of the flame front. Calculation of the conditional probability of c for its dissipation rate being zero could provide a quantitative measure to determine whether a flame is likely to extinguish within a relatively short timeframe. This is equivalent to detecting that, for every value of 0.9<c<1, there are volumes of significant size where the value of c is uniform. Note that a successful flame must have a volume of substantial size with c=1. From a practical perspective, if each individual flame kernel is monitored, then extinction is imminent if secondary structures of incomplete reactions are present when the spark ceases adding energy.

Experimental analysis of laser-induced spark ignition of lean turbulent premixed flames: New insight into ignition transition

Recently, Shy and his co-workers reported a turbulent ignition transition based on measurements of minimum ignition energies (MIE) of lean premixed turbulent methane combustion for a wide range of equivalence ratios. Our study used a similar approach to present an additional and complete MIE data set for laser-induced spark ignitions in a flow configuration and under turbulent conditions (i.e., for different ranges of length and time scales) that differed from those chosen by Shy and his colleagues. To extend the previous analyses to very lean premixed methane/air flames, the study examined a characteristic chemical time (τCB), which was defined as the start of the chain-branching reactions. Physically, this chemical time corresponded to the time when the initial chemical reactions of the ignition process released enough heat to compensate for heat losses and to enable a self-sustained reaction in successful ignition. Values for τCB, which were experimentally obtained from the temporal evolution of the mean hot kernel emissions (in laminar flow), strongly decreased with the equivalence ratio Φ for very lean flames and remained nearly constant and equal to 150μs for Φ>0.7. We obtained a clear turbulent ignition transition on the MIE as a function of the rms velocity (u′) and reconfirmed the two modes of ignition. Temporal kernel emission recordings served to describe the transition through the study of the interaction between the turbulence and the hot kernel at the time t=τCB. For low turbulence, when the smallest time scales of the turbulence τK were larger than the τCB values, the hot kernel/turbulence interaction occurred after the initiation of the chemical reactions. The amount of deposited energy required to initiate the reactions and to attain a self-sustained flame kernel was consequently similar to that found under laminar conditions. After the transition, the smallest time scales of the turbulence were smaller than the τCB values. Turbulence may have affected and interacted with the hot kernel before the initiation of chain-branching reactions occurred. Larger amounts of deposited energy were therefore required to compensate for this turbulent dissipation and to attain a self-sustained flame. A Peclet number (PeCB) was introduced, which is equal to the ratio of the turbulent diffusivity at the time of the initiation of the chemical reactions and the thermal diffusivity. These very scattering MIE data (dependent on Φ and u′) are plotted as a function of PeCB and collapsed into a single curve with two drastically different increasing slopes showing the ignition transition.

Analysis of limit cycles sustained by two modes in the flame describing function framework

The Flame Describing Function (FDF) framework, developed for the nonlinear instability analysis of combustors, has been validated more recently in a generic configuration comprising an upstream manifold, an injection unit and a flame tube. This system featuring a wide variety of dynamical phenomena, is used here to explore a new range of self-sustained flame oscillations. Depending on the geometry, the system exhibits stable or variable amplitude limit cycles. In the first case, oscillations have an essentially constant amplitude and are well retrieved with the FDF framework, whereas in the second case, limit cycles feature different types of amplitude unsteadiness and require some further consideration. The present article is concerned with one type of unstable oscillation in which a regular period modification occurs in the presence of two modes, leading to frequency heterodyning. It is found from the FDF analysis that such oscillations sustained by two modes may occur when there is an overlap between modes corresponding to super and subcritical bifurcations. An additional condition which has to be fulfilled to obtain this behavior is inferred from experiments.

Full Chemical Kinetic Simulation of Biogas Early Phase Combustion in SI Engines

This paper presents computational work on the biogas early phase combustion in spark ignition (SI) engines using detailed chemical kinetics. Specifically, the early phase combustion is studied to assess the effect of various ignition parameters such as spark plug location, spark energy, and number of spark plugs. An integrated version of the KIVA-3V and CHEMKIN codes was developed and used for the simulations utilizing detailed kinetics involving 325 reactions and 53 species. The results show that location of the spark plug and local flow field play an important role. A central plug configuration, which is associated with higher local flow velocities in the vicinity of the spark plug, showed faster initial combustion. Although a dual plug configuration shows the highest rate of fuel consumption, it is comparable to the rate exhibited by the central plug case. The radical species important in the initiation of combustion are identified, and their concentrations are monitored during the early phase of combustion. The concentration of these radicals is also observed to correlate very well with the above-mentioned trend. Thus, the role of these radicals in promoting faster combustion has been clearly established. It is also observed that the minimum ignition energy required to initiate a self-sustained flame depends on the flow field condition in the vicinity of the spark plug. Increasing the methane content in the biogas has shown improved combustion.

Numerical Investigation of Edge Flame Propagation Behavior in an Igniting Turbulent Planar Jet

The effects of strain rate and curvature on the edge flame propagation characteristics in an igniting turbulent coflowing planar jet are studied based on 3-dimensional compressible Direct Numerical Simulation (DNS) with a modified single-step Arrhenius chemistry. In the present configuration the high-speed jet fluid is considered to be fuel-rich, whereas the slow-moving coflowing fluid is taken to be fuel-lean. Consistent with previous work with DNS without mean flow and shear, the resulting flame from localized forced ignition at the jet exhibits predominantly premixed edge flame structure where premixed flames are formed on both the fuel-rich and fuel-lean sides and the intersection between these two branches propagate on the stoichiometric mixture fraction isosurface. The edge flame propagation behavior has been studied in terms of the statistics of the edge flame displacement speed S d , which refers to the speed at which the fuel mass fraction Y F isosurface moves normal to itself, relative to an initially coincident material surface at the intersection between the fuel-rich and fuel-lean premixed flames on the stoichiometric mixture fraction isosurface. The probability density function of the density-weighted edge flame displacement speed shows nonzero probability of finding negative values of at later stages of self-sustained flame propagation. The mean value of decreases after the energy deposition is switched off but eventually settles to a value that does not change appreciably with time. The is found to be predominantly negatively correlated with curvature but the correlation between and tangential strain rate shows both positive and negative correlating trends with the positive correlating trend dominant at later stages of flame propagation. It has been found that strain rate and curvature dependences of |∇Y F | have significant influences on the statistical behavior of in response to strain rate and curvature. The observed strain rate and curvature dependences of have been explained in detail in terms of statistical behaviors of the reaction, normal diffusion, and the tangential diffusion components of (i.e., , , and ).

Laser-induced radical generation and evolution to a self-sustaining flame

Images and emission spectra of sparks produced by laser-induced breakdown of methane and propane air mixtures were investigated with a high degree of spatial and temporal resolution. The laser-induced breakdown was generated by focusing a 532-nm nanosecond pulse from a Q-switched Nd:YAG laser. The data were collected using an intensified high-speed camera and a single/multi-fiber Cassegrain optics system coupled to an ICCD spectrometer. Emission spectra of OH ∗, CN ∗, CH ∗, and C ∗ 2 radicals were also collected using spectra boxes. The results provided information about the different stages of the laser-induced breakdown, with a specific focus on the transition from a flame kernel to a self-sustaining flame. The plasma shape and emission spectrum were very reproducible. The differences in the size of the flame kernel and the evolution of the radical emissions were analyzed for mixtures that fired or misfired. The impact of the level of radicals in the flame kernel was a critical parameter for the firing process, starting around 1 μs after the laser-induced breakdown. The transition from plasma cooling to the classical chemical reactions in the combustion zone was analyzed. Even though the flame kernel size was directly linked to the spark energy, this was not a key parameter toward evolution to a self-sustaining flame. The Taylor blast wave theory was used to plot the location of the shock based on the evolution of the flame kernel size. The location was calculated using a laser-supported detonation model. A very good correlation was observed with the hot gas ignition process. Our results allowed us to obtain information about the process leading to firing or misfiring for similar environments, resulting in a better understanding of the laser breakdown phenomena and the means of utilizing this technique in an industrial context.

Direct numerical simulations of turbulent flame expansion in fine sprays

Direct Numerical Simulations of expanding flame kernels following localized ignition in decaying turbulence with the fuel in the form of a fine mist have been performed to identify the effects of the spray parameters on the possibility of self-sustained combustion. Simulations show that the flame kernel may quench due to fuel starvation in the gaseous phase if the droplets are large or if their number is insufficient to result in significant heat release to allow for self-sustained flame propagation for the given turbulent environment. The reaction proceeds in a large range of equivalence ratios due to the random location of the droplets relative to the igniter location that causes a wide range of mixture fractions to develop through pre-evaporation in the unreacted gas and through evaporation in the preheat zone of the propagating flame. The resulting flame exhibits both premixed and non-premixed characteristics.

EFFECTS OF TURBULENCE ON SPARK IGNITION IN INHOMOGENEOUS MIXTURES: A DIRECT NUMERICAL SIMULATION (DNS) STUDY

Spark ignition in inhomogeneous mixtures is numerically studied using three-dimensional compressible Direct Numerical Simulations (DNS) with simplified chemistry. The thermal effect of the spark is represented by a Gaussian power distribution in the energy transport equation. Success of the spark ignition and subsequent self-sustained flame propagation is shown to be heavily dependent on the turbulent velocity fluctuation. It has been found that high levels of turbulent velocity fluctuation have detrimental effects on the spark ignition event, which ultimately may lead to misfire. It is observed that the flame shows a tribrachial structure following successful ignition. The local flame propagation speed may attain values equal to several times the laminar burning velocity of an unstrained premixed stoichiometric flame S L , but may also be negative, in agreement with previous findings. The mean flame propagation speed decreases with increasing turbulent intensity.

A correlation function method of recovering the combustion law parameters for particles burning in optically thin dust flames

To determine the law describing how the rate of combustion of a particle depends on its diameter, we propose a correlation function (CF) method of measuring a particle's combustion time in a polydisperse dust flame. The method is based on observing the light intensity, I ( t ) , emitted by the flame, whose fluctuations are mostly governed by the random ignition and extinction of the particles. For an optically thin stationary dust flame, the CF for the intensity, I ( t ) , consists of correlation functions of radiation traces from each burning particle. To establish the relation between the CF for the intensity and the combustion law for a particle, the radiation trace was considered to be a pulse function, whose magnitude depends on the initial particle diameter, d 0 , according to α d 0 δ and whose duration depends on k d 0 γ . This model provides a measure for the linear decay of the CF with time (an effective correlation time, τ c ), which is experimentally accessible. It is found that the effective correlation time depends on moments of the initial size distribution, τ c = k 〈 d 0 2 δ + 2 γ − 1 〉 / 〈 d 0 2 δ + γ − 1 〉 , and can be used to evaluate the combustion law. The CF method has been applied both to single particles of Mg burning in air and to a self-sustained flame of polydisperse Mg dust particles with sizes in the range 30 μm ≲ d 0 ≲ 240 μm . The values of δ, γ, and the combustion constant k were estimated from experimentally measured values of τ c and the moments of the initial size distribution. For single burning particles, the CF method resulted in k = 0.51 ( ± 0.05 ) × 10 2 s / cm 2 , which was confirmed by direct measurement of the combustion time. For the self-sustained dust flame of particles with d 0 ≲ 120 μm , k was found to equal 1.27 ( ± 0.08 ) × 10 2 s / cm 2 . For dusts of larger particles, d 0 ≳ 120 μm , the value of k decreases with increasing d 0 towards the value for the single burning particle.

Effects of mixing on ammonia oxidation in combustion environments at intermediate temperatures

Ammonia often occurs in combustion gases as a product of fuel-nitrogen. Since either NO or N 2 may predominate as the product of ammonia oxidation, there is considerable interest in understanding the factors responsible for selecting the final products. In flames, the selectivity is known to depend on whether the reaction zones are premixed or nonpremixed. This paper reports on a combined experimental and modeling investigation of ammonia chemistry in a hot combustion environment that is below flame temperatures, such as in post-combustion gases. Experiments that mix highly diluted ammonia–methane and oxygen–water streams are interpreted in terms of a plug-flow model, a simplified mixing reactor model, and a two-dimensional direct numerical simulation. The study finds that the final products of ammonia oxidation remain sensitive to mixing even at temperatures below those of self-sustaining flames. At low temperatures, ammonia oxidation occurs in a premixed reaction zone, but at sufficiently high temperatures a nonpremixed reaction zone may develop that produces significantly less NO than the equivalent premixed system. A direct numerical simulation is required to predict the behavior over the full range of conditions investigated experimentally, while a simplified mixing reactor model captures the essential features as long as the radial gradients are not too steep.