Envelope Flame Research Articles

木質バイオペレットの熱分解・燃焼特性

From a viewpoint of environmental preservation and resource protection, the recycling of wastes has been promoting. Expectations to new energy resource are growing by decrease of fossil fuel. Biomass is one of new energies with prevent global warming. This study is an attempt to burn pelletized woody biomass (Bio-pellet) made from sawdust and logging residue in order to thermally recycle waste products of forestry and lumbering industry. The devolatilization property of Bio-pellet were observed by the thermogravimerty and differential thermal analysis (TG/DTA) to obtained fundamental data of Bio-pellet pyrolysis. The thermogravimetric analyzer was used to measure weight loss and temperature difference. It observed that the weight of Bio-pellet decreased under three stages with endothermic reaction during pyrolysis, and with exothermic reaction during combustion. The combustion behavior of Bio-pellets was observed in the electric furnace, where the video-recording and measurements of pellet temperature and weight were carried out at sequential steps of the combustion process. The effects of furnace temperature and pellet size were exam-ined in order to elucidate the combustion characteristics of Bio-pellets, such as ignition delay, burning period, char-combustion time and the change of weight decrease and temperature rise. The results indicated that they are influenced at each step of the combustion process such as devolatilization, ignition, visible envelope flame combustion and char combustion.

High pressure vaporization of burning droplet with flash vapor-liquid equilibrium calculation

Unsteady burning of n-pentane liquid droplet quiescent in high-pressure oxidizing air is investigated with emphasis placed on rigorous calculation of droplet vaporization. Fully coupled conservation equations in conjunction with global reaction kinetics for chemistry are preconditioned and time-integrated by means of a dual time-stepping technique with pertinent boundary conditions. Properties of equilibrated vapor-liquid interface are evaluated by the method of Soave-Redlich-Kwong equation of state along with consistent set of mixing rules for multi-component mixtures. Envelope flame propagates persistently with gradually increasing peak flame temperature as long as droplet remains in liquid phase. At all pressures, ignition is immediate and resultant steep temperature gradient at interfacial boundary aids in the activation of vaporization, but its extent decreases at later times as the droplet surface temperature arrives at its pseudo wet-bulb value. Ignition is immediate with continuous initial temperature distribution in the vicinity of interfacial boundary and renders an impulsive rise in droplet surface temperature. Other representative pressure effects on evolving droplet combustion are also presented and discussed.

Confined annular swirling jet combustion

Five different flame states are identified in a compact combustion chamber that is fired by a 30-kW swirl-stabilized partially premixed natural gas burner working at atmospheric pressure. These flame states include a nozzle-attached tulip-shaped flame, an unattached lifted ring-shaped flame suitable for single-digit-ppm NO x emission, and a Coanda flame that clings to the bottom wall of the burner. Flame-state transition is generated by changing the swirl number and may be further modified by premixing the combustion air with 70% of the natural gas flow. Detailed experimental results of two ring-shaped flames are reported. The reactants flow out of the burner in a conical sheet and enter the spread-out flame. Evidence of precessing and oscillatory behavior of the internal recirculation zone (IRZ) as well as the flame envelope boundary is discussed and explained by the bimodal probability density functions of the velocity distributions. Low-pressure regions in the corner of the tight combustion chamber and altered turbulent flame speed due to a change in swirl account for the flame transition and stabilization process. Finally, the strength of the IRZ is compared with data of other researchers revealing that both ring-shaped flames feature a large and much stronger IRZ than reported in literature although limited swirl numbers apply.

Ignition Phenomena of PMMA Rod in Highly Preheated Air Flow.

Ignition process and burning characteristics of a polymethylmethacrylate (PMMA) rod in highly preheated convective environment are investigated experimentally. The main parameters are O2 concentration, flow velocity, preheated air temperature and oxygen dilution gas. The convective ignition can be classified into three steps : blue flame generation, attached flame formation and envelope flame development. Convective ignition of PMMA suddenly exposed to high temperature air is initiated in downstream and dim blue premixed flame is formed there. When the pyrolysis of solid fuel increases enough for flame propagation, the blue flame propagates to upstream to form an attached flame or luminous envelope flame. The ignition delay time is almost independent on oxygen concentration because the initial blue flame formation process is mainly controlled by pyrolysis rate. However, the following attached flame and envelope flame formation time increases with decrease in the oxygen concentration since these processes are chiefly dominated by burning velocity.

Interactive combustion of two-dimensionally arranged quasi-droplet clusters under microgravity

To investigate the mutual interactions between droplets in the spray combustion, combustion of 2-dimensionally arranged quasi-droplet clusters is studied under microgravity. Quasi-droplet samples, which are solid in room temperature and change into liquid just after the ignition, consist of alcohol (propanol, butanol, pentanol, or hexanol) and polyethylene glycol with a volumetric ratio of 2:1. Seven samples sustained by glass rods form a 2-dimensional quasi-droplet cluster. Electrically heated nichrome wires ignite all samples in the cluster simultaneously. Single envelope flames that surround the clusters appeared. The results show that the sample spacing has a strong effect on the shape and movement of the flame. Sample clusters with large sample spacings come to the external group combustion through the scavenging combustion mode, whereas the small spacing clusters start directly with the external group combustion. At large sample spacings, the distance from the edge of the sample cluster to the flame (flame distance) increases to a maximum value and then decreases with time. The period of flame growth is prolonged with decreasing sample spacing and finally, at a small enough sample spacing, the flame distance keeps increasing until the flame disappears. This flame movement is attributed to the fuel vapor accumulation effect, which becomes more dominant with decreasing sample spacing. The burning lifetime decreases monotonically and approaches the value of the single flame with increasing sample spacing. The flame distance decreases monotonically and approaches the single flame radius with increasing sample spacing also. These results render important confirmations of the external group combustion phenomena and prove the importance of the two kinds of unsteadiness, that is, the scavenging combustion with large droplet interval and the fuel vapor accumulation effect with small droplet interval, in group combustion.

BRACHIAL BURNING AND GASIFICATION SPLIT OF A CONVECTING TWO-DROPLET SYSTEM

Brachial burning and gasification-split characteristics of two droplets situated in a forced convection are analyzed by means of numerical method. For the given droplet size ratios of 1.5 and 0.25, it is found that a tri-brachial burning in the former and a penta-brachial burning in the latter are exhibited, where the brachial burning and gasification split are defined according to the envelope flame formation and destruction. When the droplet size ratio is large, the two droplets are always enclosed together in a flame for low Reynolds numbers. As Reynolds numbers are higher, the flame is destroyed and brought to deposition behind the two droplets. Two or three gasification-split points are induced in accordance with the Reynolds number undergoing, thereby producing the tri-brachial burning. On the other hand, for a lower droplet size ratio such as 0.25, under certain conditions the flame surrounds only the trailing droplet, yielding a semienvelope flame. As a result, two extra brachial burn ing based on the variation of the semienvelope flame are defined, and the penta-brachial burning is obtained. Accordingly, the multiple burning states around two interactive droplets, including pure vaporization, wake flame, transition flame, envelope flame, and semienvelope flame, are established, except for the low Reynolds numbers in which only envelope flame develops. The new findings presented in the present article provide comprehensive insight for the development of droplet model in spray combustion, by intrinsically considering the variations of the Reynolds number around droplets.

Counterflow diffusion flame with polydisperse sprays

An experimental study was performed on counterflow diffusion flames, in which polydisperse n-heptane sprays carried by a nitrogen stream from the lower duct, burn in oxidizer streams from the upper duct. The droplet-size distributions were varied with the nitrogen flow rate for atomization. The burning behaviors of the spray were observed and the oxygen concentration at extinction was determined for different droplet-size distributions. The flat blue diffusion flame was observed between two ducts, which was supported by vaporization of small droplets. Relatively large droplets penetrated the flat blue flame and continued to burn on the oxidizer side with an envelope flame surrounding each single droplet. The burning behaviors for different droplet-size distributions were discussed based on the Stokes number, St0, and the ratio, ψ, of the vaporization time to the characteristic flow time. The extinction limit was also found to be affected by the droplet-size distribution. The oxygen concentration at extinction had a minimum value for a specific volume mean droplet diameter irrespective of the fuel flow rate. This dependence was explained as a result of a competitive effect of St0 and ψ. The extinction limit of an n-octane spray was also examined to verify the effect of St0 and ψ on the extinction limit of the spray.

Numerical Study for Interaction Between Water Mist and Counterflow Diffusion Flame Over Tsuji Burner

The interaction between a water mist and a counterflow diffusion flame over a Tsuji burner is studied numerically. The governing system is comprised of two parts. The gas phase combustion model adopts the one developed by Chen and Weng (1990). For droplets, their motions are described by Newton's second law. The corresponding droplet heat and mass transfer are considered by using empirical correlations. The PSI-Cell model is used to manage the two-phase coupling. The evaporation of droplets through heat absorption from the flame can generate vapor in the flow field. The vapor dilutes both the oxidizer and fuel concentrations, and the heat absorption by phase change lowers the flame temperature. For 50 p.m droplets, the envelope flame is extinguished from the forward stagnation area and becomes a wake flame as the mist flow rate increases to 7%. The wake flame no longer exists when the mist flow rate is greater than 15%. The critical values for transition of the envelope flame to a wake flame and complete flame extinction for 80 μn droplets were 4·5% and 10% separately. Under the same mist flow rate, the smaller the droplet sizes, the stronger the flame. The effect of droplet size is studied by changing the droplet size and the corresponding mist flows with the number flow rate constant. The critical sizes for transition of the envelope flame to a wake flame and for complete flame extinction were 55 μm and 75 μm respectively. Future work is also discussed.

Extinction Characteristics of Counterflow Diffusion Flame of n-Heptane Spray.

An experimental study was performed of the extinction characteristics of counterflow diffusion flames in which polydisperse n-heptane sprays carried by a nitrogen stream from the lower duct, burnt in oxidizer streams from the upper duct. The distance between the ducts was varied to vary the strain rate with constant fuel flow rate, so that the upstream spray characteristics were independent of the strain rate. The burning behavior of the spray is observed and the oxygen concentration at extinction was determined as a function of the strain rate. The flat blue diffusion flame was observed between two ducts, which is supported by vaporization of small droplets. Relatively large droplets penetrate the flat blue flame and continue to burn on the oxidizer side with an envelope flame surrounding each single droplet. The burning behaviors and extinction characteristics were discussed based on the characteristic time scales such as the droplet response time to the flow and the vaporization time.

Experiments on the ignition of monocomponent and bicomponent fuel droplets in convective flow

The ignition of free and suspended monocomponent (iso-octane) and bicomponent (mixtures of n-hexadecane with n-heptane, n-dodecane, or iso-octane) droplets was experimentally investigated in the postflame gases (gas temperature, T ∞, between 1080 and 1300 K; flow velocity U ∞ of 2.0–2.5 m/s) of a flat-flame burner. Initial droplet diameters, d 0, ranged from 300 to 1500 μm and droplet Reynolds numbers from 3.5 to 22. Comparing iso-octane and n-heptane droplets, both fuels with similar volatilities, iso-octane droplets had longer ignition delay times, larger ignition distances, and slower upstream flame propagation speeds, due to the lower reactivity of iso-octane. With decreasing initial diameter, ignition delay times decreased and became reaction-controlled. The ignition delay times of iso-octane droplets as small as 300 μm were, interestingly, longer than those for hexadecane droplets, despite their higher volatility. For bicomponent C 7/C 16 droplets of different proportions, the variation of the ignition location with gas temperature was similar to monocomponent droplets. For T ∞ < 1220 K, ignition was initiated far downstream as a dim blue premixed flame, followed by rapid upstream propagation to form an envelope flame. The ignition location was closer to the droplet for higher temperatures. For large droplets, when the ignition was evaporation-controlled, ignition delay times were strongly affected by the proportion of fuels having different volatilities. For C 7/C 16, i − C 8/C 16, and C 12/C 16, the ignition delay time was similar to that of a pure volatile fuel droplet when the volume fraction of the volatile component ( r ν) was larger than 20%–40%. If r ν was smaller, the ignition delay time resembled that of a pure nonvolatile fuel droplet. For small bicomponent droplets, ignition was reaction-controlled. The ignition delay times of C 7/C 16 droplets were independent of r ν due to the kinetic similarity of both components. However, those of i − C 8/C 16 droplets were longer for high iso-octane fractions, due to the lower reactivity of iso-octane.

Experiments on the instability modes of buoyant diffusion flames and effects of ambient atmosphere on the instabilities

Large scale dynamic behavior of buoyant diffusion flames were studied experimentally. It was found that buoyant diffusion flames originating from circular nozzles exhibit two different modes of flame instabilities. The first mode results in a sinuous meandering of the diffusion flame, characteristic of flames originating from small diameter nozzles. This instability originates at some distance downstream of the nozzle exit in the contraction region of the buoyant flame envelope and develops into a sinuous motion of the flame. The second mode is the varicose mode which develops very close to the nozzle exit as axisymmetric perturbations of a contracting flame surface. In this mode, flame oscillations result in the formation of toroidal vortical structures that convect through the flame and cause periodic burn out at the flame top resulting in the observed flame height fluctuations. The average flame heights are found to be typically shorter for these flames. The oscillation frequencies and their scaling for the two modes are also different with the sinuous mode having higher frequencies than the varicose mode. It was also observed that the instability can switch from one mode to the other and the probability of observing the varicose mode appears to increase with increasing Richardson number. Additionally, the feasibility of altering the behavior of buoyant diffusion flames was explored through variation of the oxidizer medium density. It was found that the flame oscillations can be completely suppressed for flames burning in helium rich helium–oxygen mixtures. At lower helium concentrations, the oscillation frequency can be significantly reduced. In order to enhance the buoyancy effect, CO2–O2 mixtures were also studied. However, the density increase and its effects on flame oscillation frequency were found to be small compared to those flames burning in air. These experiments point towards the feasibility of altering buoyant flame behavior under earth gravity and studying the large scale dynamical aspects of buoyant flames without the need of variable gravity environment.

Partially Premixed Flame Structure and Stability of Twin Droplets in Flows

The flame structure and stability, as well as the vaporization rates of twin droplets exposed to a high-temperature, partially premixed flow are investigated in the present study. Two important parameters of the Reynolds number and ambient equivalence ratio are taken into consideration to account for the influence of fuel vapor in the upstream far field on those of burning mechanisms around the twin droplets. When increasing the ambient equivalence ratio, the chemical reactivity in the upstream can be classified into three types; weakly, moderately, and obviously reactive flows, according to the distribution of the vaporization rate of the leading droplet versus the Reynolds number. In particular, if the flow is moderately reactive, say, ϕ=0.2, a double-peak profile is observed in the vaporization rate of the leading droplet, and it clearly depicts that by increasing the Reynolds number the vaporization is sequentially dominated by the envelope flame, reactive flow, and convective flow. With regard to the trailing droplet, because of the multiple effects stemming from the leading droplet, the impact of the ambient equivalence ratio on the vaporization rate distribution is similar, except for in the purely oxidizing environment in which the twin droplets behave as a single droplet. As a whole, the evaluated results illustrate that the partially premixed flow is conducive to promoting the vaporization and aides the flame stability in a twin-droplet system, while some burning characteristics in a counterflow system can also be obtained in front of the leading droplet. [S0022-1481(00)01504-8]

Aluminum droplet combustion in fluorine and mixed oxygen/fluorine containing environments

Experimental and numerical results are reported on the combustion of single aluminum (Al) droplets, 187 μm diameter, that were laser ignited in quiescent environments of SF 6 and SF 6 /O 2 or Ar mixtures. Combustion measurements consisted of luminosity versus time traces, photographic observations, and electron probe microanalysis (EPMA) and electron microscope inspection of condensed-phase products of combustion surrounding burn-intercepted droplets. Planar laser-induced fluorescence (PLIF) was employed to image the gaseous intermediated reaction species ALO during droplet combustion in SF 6 /O 2 mixtures. Aluminum droplets burning in pure SF 6 burned without condensed-phase products in the surrounding envelope flames. Accordingly, no condensed-phase products were found on surfaces of droplets captured by quenching into an inert environment. In SF 6 , classical d 2 law burning behavior was observed. Droplets burned in SF 6 /Ar mixtures exhibited similar behavior, with diminished burning rates. Disruptive burning was not observed for pure SF 6 or any of the SF 6 /Ar mixtures. Condensed-phase deposits on quench plates intercepting burning droplets showed the surrouunding gaseous flame structure to consist of two zones for burning in SF 6 and three zones in SF 6 /O 2 mixtures. For each system, near the surface where the gas phase was aluminum rich, significant amounts of sulfur were found in the condensed products. PLIF showed the species profile of AlO to shift closer to the surface with increasing fluorine content in SF 6 /O 2 mixtures. Experimental results were compared with a local equilibrium combustion model. The model predicted burn rates within 20% and qualitatively predicted radial profile speciation. As the O/F ratio was increased, the final equilibrium products were predicted to shift from AlF 3 to AlF 2 O to Al 2 O 3 . The formation of AlF 2 O reduced condensed-phase products at the reaction temperature for a wide range of O/F ratios.

Self-similarity, momentum scaling and Reynolds stress in non-premixed turbulent spray flames

An experimental study was conducted in a turbulent spray flame in which droplets were produced ultrasonically at low velocity relative to the host gas. In this fashion, injector-specific effects on the two-phase flow were minimized and a scenario generally characteristic of the far field of practical spray systems could be simulated. Close to the burner exit, the spray flame appeared as a dense column of drops burning with an envelope flame. Further downstream, it opened up slowly in the radial direction and developed a turbulent ‘brush’ appearance. Measurements of the size, velocity and concentration of the droplets, and of gas-phase velocity and temperature were made by combining a Phase-Doppler interferometric technique with Stokes/anti-Stokes Raman thermometry. The experimental data were used to derive scaling and self-similarity for the Reynolds-averaged continuity and momentum equations using suitable transformations.Results showed three distinct regions, on the basis of the behaviour of the gas axial velocity in the spray flame. In the lower part of the flame, the gas momentum increased because of vaporization. In the intermediate region of the spray flame, the axial velocity decayed along the centreline as an inverse power of the distance from the virtual origin, with exponents smaller than unity. In the upper part of the spray flame, the flow field recovered the axial velocity decay that is typical of incompressible jets, namely as an inverse of the axial distance. Self-similar behaviour held for the axial velocity throughout the intermediate region. The vapour source term in the gas continuity equation scaled approximately as an inverse power of axial distance, and exhibited self-similarity throughout the spray flame. As a result, a simple model of the Reynolds stress term could be formulated, in which two competing contributions appear: one, that is due to turbulent transport, tends to increase the value of the velocity correlation; another, that is due to the vaporization term, tends to reduce the value of the velocity correlation and can be construed as a vaporization-induced tendency towards relaminarization. The first term is modelled by a classic gradient-transport approach yielding an empirical mixing length relating the velocity correlation to the average velocity gradient. Model and experiments are found to be in good agreement, especially sufficiently far from the injector, where one-way coupling between the two phases holds.

Condensed-phase species distributions about Al particles reacting in various oxidizers

Experimental results on the combustion of single, isolated aluminum particles, laser ignited in quiescent environments consisting of pure N 2O, CO 2, CO and in mixtures of 21% O 2 / 79% N 2 and 21%O 2 / 79% Ar are reported. Combustion measurements consisted of photographic observations and electron probe microanalysis (EPMA) of the condensed-phase product composition and radial distribution. Aluminum particles in O 2, CO 2, and N 2O atmospheres were found to burn with envelope flames. Of these oxidizers, the largest flame envelope, as determined by the condensed-product distribution, occurred for Al combustion in the O 2/Ar mixture, followed by Al combustion in the O 2/N 2 mixture, the CO 2 atmosphere, and the N 2O atmosphere. Combustion in the CO atmosphere appeared to occur on (near) the particle surface with only a weak envelope reaction. Consistent with previous results in the literature, Al particle disruption was not observed in O 2/Ar environments, but was observed in O 2/N 2 environments. Although speculated in the literature, the present work confirms the existence of aluminum nitrides (oxy-nitrides) in the fuel-rich region near the particle surface for nitrogen-containing oxidizers (i.e., O 2/N 2 and N 2O). Equilibrium calculations indicate that near the surface, solid-phase AlN may exist to temperatures well above the melting temperature of aluminum oxide. Thus, its presence may affect the fragmentation process. Finally, condensed-phase carbon (possibly in the form of aluminum carbide) was found throughout the surrounding gas-phase for CO combustion.

Behavior of Ignition and Combustion for Single Refuse Derived Fuel Particles.

This study is an attempt to burn pulverized Refuse Derived Fuel (p-RDF) made from municipal solid waste (MSW) with a view point of recycling of waste resources from daily domestic and industrial processes. In order to obtain fundamental data on p-RDF combustion, the ignition and combustion behavior of the particles of RDF and its component materials such as plastics (polypropylene, polystyrene), fiber (nylon) and paper (newspaper) were observed in an electric furnace. The ignition delay and the burning period of particles both p RDF and its component materials were investigated under specified furnace temperatures and particle sizes of each material. The results indicated that each component material of p-RDF had different ignition and combustion mechanism in processes such as volatilization, ignition and visible envelope flame. The values of the ignition delay and the burning period of plastics and fiber were larger than those of paper, and RDF showed medium values between them. The estimated activation energies for plastics and fiber were larger than paper, and RDF showed medium values. Each component material of p RDF showed different activation energies which led to different ignition delay and burning period.

Numerical Simulation of Combustion and Extinction of a Solid Cylinder in Low-Speed Cross Flow

The combustion and extinction behavior of a diffusion flame around a solid fuel cylinder (PMMA) in low-speed forced flow in zero gravity was studied numerically using a quasi-steady gas phase model. This model includes two-dimensional continuity, full Navier Stokes’ momentum, energy, and species equations with a one-step overall chemical reaction and second-order finite-rate Arrhenius kinetics. Surface radiation and Arrhenius pyrolysis kinetics are included on the solid fuel surface description and a parameter Φ, representing the percentage of gas-phase conductive heat flux going into the solid, is introduced into the interfacial energy balance boundary condition to complete the description for the quasi-steady gas-phase system. The model was solved numerically using a body-fitted coordinate transformation and the SIMPLE algorithm. The effects of varying freestream velocity and Φ were studied. These parameters have a significant effect on the flame structure and extinction limits. Two flame modes were identified: envelope flame and wake flame. Two kinds of flammability limits were found: quenching at low-flow speeds due to radiative loss and blow-off at high flow speeds due to insufficient gas residence time. A flammability map was constructed showing the existence of maximum Φ above which the solid is not flammable at any freestream velocity.

Observations on the Combustion of Polymers (Plastics): From Single Particles to Groups of Particles

This investigation aimed at examining the combustion behavior of polymer particles in groups and at comparing it to that of single particles. Combustion observations were subsequently related to measurements on emissions of organic pollutants, such as PAHs and soot, which were previously measured in this laboratory under similar conditions. The following plastics (polymers), commonly found in municipal waste streams, were burned in powder form: poly(styrene) (PS), poly(ethylene) (PE), polyvinyl chloride) PVC, and poly(methyl methacry-late) (PMMA). Particles, 125-212um, were introduced to a drop-tube furnace, at steady-flow conditions, and burned in cylindrical streams (clouds), in 1250 K air. The particle mass loading in the furnace was varied to achieve either dilute clouds, where isolated (single) particle combustion occurred, or dense clouds where interactive particle combustion look place. Information on the combustion characteristics and flame temperatures of the particles was obtained by cinematography and three-color pyrometry. Results revealed striking differences in the combustion behavior of the four polymers burning in particle clouds. PVC particles burned with bright individual flames, resisting the formation of group flames, even at high particle number densities. In dense clouds, envelope flame diameters were generally smaller and their durations were shorter than in single particle combustion. Poly(ethylene) particles in very dilute clouds burned individually, surrounded by low-sooting, faint envelope flames. When particles burned in close proximity to each other, they abruptly exhibited localized flash-vaporization and formed highly-luminous, large, group flames, “puffs”. At high particle number densities the entire stream was engulfed in a cooler flame containing finely-dispersed soot. Poly(styrene) particles burned mostly as discrete entities at relatively low flowrates, and the progression to partial group flames was gradual as the particle mass loading increased. PS particles consistently formed flames with long wakes which, upon extinction formed strings of highly-agglomerated soot. Flame temperatures were lower in group combustion than in individual particle combustion. While single particle envelope flames were previously recorded to be in the neighborhood of 2000 K, those of group flames were found to be only a few hundred degrees above the gas temperature, at globally fuel-rich conditions.

Hysteresis effects of two interactive droplets burning in convective flows

The hysteresis characteristics of two equal-sized droplets in tandem burning in hot convective flows are investigated numerically. Similar to the results of convective single-droplet burning of Jiang et al. (1995), the present study shows that at certain flow conditions, dual solutions are also exhibited for the flame structure around the two droplets, depending upon the initial conditions assumed. For the lower-branch solution where the droplets are initially merged in a chemically frozen flow, four combustion modes, including pure vaporization, wake flame, transition flame, and envelope flame, are obtained sequentially as the ambient flow temperature varies from a low 600 K to a high 1400 K at Re=10. The upper-branch solution, where the droplets are initially enclosed by an envelope flame, however, possesses only two modes of wake and envelope flame under the equivalent conditions. The vaporization rates of both the leading and trailing droplets are hence found to be multi-valued at certain flow conditions where multiple flame configurations exist. The effects of droplet interaction are more profound on the trailing droplet so that its vaporization rate is generally lower than that of the leading droplet. However, as a transition flame occurs, the inverse exhibition in contrast to the foregoing result is observed, while the double reaction zones are excited. When the droplet spacing becomes large enough, the leading droplet behaves just like a single droplet in an equivalent flow condition, although it is not true for the trailing one.

Statistical investigation of the turbulent flame geometrical structures in a liquid oxygen/gaseous hydrogen shear-coaxial jet

Planar laser-induced fluorescence of oxygen has been performed to investigate a liquid oxygen/gaseous hydrogen turbulent flame. The hot vapor of oxygen is excited through a predissociative fluorescencescheme at 193 nm with a broad-band ArF excimer laser. The experimental configuration, where oxygen is within the flame envelope, implies that the laser beam is not absorbed before encountering the hot oxygen in the flame sheet. Since the fluorescence efficiency increases strongly with temperature [1], oxygen is detected with a maximum sensitivity in the flame where temperature is the greatest. Thus, such a configuration offers the opportunity of detecting the flame from the oxygen fluorescence images. Depending on the operating conditions, the flame shows different angles and develops structures of different sizes. The instantaneous 2-D images have been analyzed statistically to study the morphology of the flame in the near field. The increase in length due to wrinkling and bending of the flame is measured as well as the curvature of the flame. The presence probability distribution is obtained and leads to the proposition of a measurement method of flame surface density from the probability density function formation [2]. Such measurements provide validation tools for numerical modeling of complex turbulent flame in terms of constant-property surface for turbulent diffusion flame and give elements to calculate the mean reaction rate.