Search

The laser-induced incandescence (LII) technique has emerged as a viable non-intrusive technique for measuring soot concentration in flames. The incandescence of soot occurs when laser light heats the particles to the incandescent temperature (which is typically /spl sim/4,000 K). It has been shown that the resulting radiant emission intensity is proportional to the soot concentration. Studies compared the LII intensities to light extinction measurements of soot volume fraction to show this proportionality relationship. There are many practical benefits for using LII for soot concentration measurements. This technique is more accurate than the conventional light extinction method because it is expected to be less sensitive to light scattering by soot particles and absorption by large PAH molecules. With proper calibration, the LII technique is believed to be ideal for use in flames producing large gradients in soot concentrations (for which the light extinction-deconvolution may not be suitable) and for turbulent flames that do not possess symmetric characteristics.

ion reactions produce vinyl radicals from the ethylene that are then converted to acetylene via: C2H3 → C2H2 + H Reaction 2 Reaction 2 is very sensitive to ambient pressure at conditions relevant to the present experiments. Acetylene is generally accepted as a key species contributing to soot formation processes. Increased levels of acetylene promote aromatic hydrocarbon formation either through the C4 mechanism of Frenklach and co-workers or through subsequent formation of C3 species followed by the C3 ring formation mechanism of Miller and Melius. In an effort to increase the likelihood of forming soot in ethanol experiments, ambient pressure and the oxygen concentration were varied in conjunction. Figure 7 displays the backlit laser-backlit view of ethanol droplets burning in various oxygen concentrations in nitrogen at 2.2 atm. At 21 % and 25 % O2 in N2, there is no visible luminosity exhibited in the flame view and the attenuation of the laser beam in the backlit view was lacking. As the oxygen concentration is increased to 30% O2 in N2, the formation of a distinct sootshell and a luminous flame are observed. Another interesting behavior was noted in which the sooting propensity appears to decrease at 40 % O2 in N2 case compared to the 30% O2 in N2 case. Additional experiments and analysis are required to investigate this interesting behavior. From the experiments shown in figure 7, the maximum soot volume fraction, fv,max, was measured using the tomographic inversion technique. These measurements clearly bear out the interpretation from the visual observation – at 21% O2 in N2, there is no measurable soot concentration, while at 30% O2 in N2, the maximum soot volume fraction is approximately 13 ppm. The soot volume fraction was not measured for the 40% O2 in N2 case since the distribution of soot was not uniform. These measurements represent the first soot volume fraction data obtained for ethanol droplet combustion in microgravity environment. These experiments clearly demonstrate the strong dependence of sooting behavior of ethanol droplets on ambient pressure and oxygen concentration. Concluding Remarks This study provided the first detailed measurement of the spherically-symmetric burning and sooting behavior of isolated ethanol droplets burning in enhanced oxygen and high pressure conditions. The burning rate measurements are strongly influenced by ambient oxygen concentrations (21% to 50% O2 in N2) but are independent of pressure in the range studied (1.0 to 2.2 atm in air). Use of enhanced oxygen concentration combined with higher pressures resulted in distinct sootshell formation. Measurement of soot volume fraction indicates that the sooting propensity increases non-monotonically with oxygen concentration. The effective control of the sooting behavior of ethanol from a soot-free flame to a highly sooting flame by using pressure and oxygen concentration is important for its use as one of the primary fuels to investigate the influence of sooting and radiation influence on droplet combustion. Acknowledgments Support from NASA through Grant NCC3-822 is gratefully acknowledged. Opportunity to perform experiments at JAMIC 10 sec. dropshaft, facilitated through the Japan Space Utilization Program (Mr. T. Sakuraya) is greatly appreciated. References Vanderver T.A., Ed. Clean Air Law and Regulation. The Bureau of National Affairs. Washington, D.C., 1992. Poulopoulos,S.G., Samaras,D.P. , Philippopoulos, C.J., Atmospheric Environment 35(2001) 4399 -4406. Nag, P., Litzinger,T.A., Haworth,D.C. Eastern States Meeting of the Combustion Institute (2001) 4 Abu-Qudais, M., Haddad, O., Qudaisat, M., Energy Conversion and Management 41 (2000) 389-399. Kitamura T., Ito T., Senda J., Fujimoto H., JSAE Review 22 (2001) 139-145 6 Miyamoto,N. et al., SAE paper No.980506 (1998). Choi, M.Y., Dryer, F.L., Science Requirements Document for Experiments and Model Development for Investigation of Sooting and Radiation Effects in Microgravity Droplet Combustion, NASA, 2001 Godsave, G.A.E. (1953). Proc. Combust. Inst. 4: 818-830. 9 Okajima, S. and Kumagai, S. Proc. Combust. Inst. 15 (1975) 401-407. 10 Hara, H. and Kumagai, S. Proc. Combust. Inst. 23 (1991) 1605-1610. Lee, A., and Law, C. K., Combust. Sci. Technol. 86

Rotary kilns are used in the production of cement clinker, alumina, lime, and nickel. These kilns require high heat energy, with temperatures up to 1500°C, usually supplied from the combustion of fossil fuels. Rotary kilns are a relatively inefficient process and there is much room for improvement in their design. One method for improving the combustion within a rotary kiln is the use of the precessing jet nozzle developed at The University of Adelaide [2]. The precessing jet nozzle is licensed under the name Gyro-Therm T M and has been successfully applied to rotary kilns producing cement clinker, alumina, lime and zinc oxide, [1] and [5]. When firing natural gas in rotary kilns, a precessing jet provides a more luminous flame, lower peak temperatures, high flame stability and about a 50 percent less NOx emissions than comparable swirling and non-swirling flames.

Laser induced breakdown spectroscopy (LIBS) of a hydrocarbon flame and the rocket motor simulator plume is reported. The flame and plume were seeded with various types of elements. The LIBS signal of the seeded element goes down drastically in the presence of the flame in comparison to the signal obtained from the aerosols of the elements. Similarly the LIBS signal of the trace elements was very weak in the luminous flame and the simulator plume in comparison to the measurement outside it. The intensity of the LIBS signal from the trace elements present in the plume/flame was found to be dependent on the process of seeding, the transition probability and the decay time of the background emission. Ultimately LIBS was found to be more sensitive than atomic emission spectroscopy (AES) in detecting the trace elements in the simulator plume. AES was used to estimate the temperature of the plume by comparing the experimental and simulated emission spectra of OH radicals. This study establishes LIBS as an improved health monitoring system over AES for the plume studied here.

A comprehensive analysis and data-acquisition of the time-resolved numerical simulations of combustion dynamics of double-base homogenous propellant in a rocket motor is performed. The physicochemical processes occurring in the flame zone and their influence on the unsteady flow evolution in the combustion chamber are investigated in-depth using large eddy simulation (LES) techniques. A fivestep reduced-reaction mechanism is used to obtain a two-stage flame structure consisting of a primary flame, dark zone, and secondary luminous flame in the gas phase. It is observed that, for homogeneous propellant combustion, the chemical time scale is much higher than the smallest Kolmogorov time scale, rendering a highly stretched and thickened flame. The chemical reactions proceed at a slower rate than turbulent mixing and propellant combustion may be locally treated as a well-stirred reactor. The unsteady flowfield in the chamber indicates three regions of flow evolution: the upstream laminar region, transition zone, and fully developed turbulent regime further downstream. The oscillatory flame modifies the vorticity distribution through strong baroclinic effects causing mismatch of the radial density and axial pressure gradients. Volume dilatation also plays an important role in affecting the vorticity transport. A theoretical formulation exploring the coupling between the oscillatory heat-release, and ensuing entropy and vorticity waves in turbulent flows is established by deriving a generalized wave equation. The work done by Reynolds stresses and the interactions between vorticity and flame dynamics may cause resonance effects and excite eigenmode pressure oscillations leading to self-sustained unstable motions within the chamber. Nomenclature a = speed of sound

The purpose of the present paper is to develop a mathematical model for predicting thermal field and heat transfer inside short length/height ratio, rectangular and nonhomogeneous material pipe connecting a combustion chamber with outdoors, while the gas combustion temperatures are in the range of 2 000 to 3 000 K and the pipe is made of heat-resistant material. The used approach is based on a finite difference simulation and makes certain idealizations and approximations in order to get to an engineering workable and at the same time adequate, solution. Main assumptions are an equivalent temperature cavity to describe radiant heat transfer due to both the combustion chamber luminous flame and outdoors. The proposed model shows a very good agreement with experimental results coming from an experimental set-up used for materials testing, that is embodied in a refractory rectangular pipe, at high temperature.

Both fluctuations of local velocity and local temperature were measured in a steady turbulent diffusion flame of propane by using the semiconductor laser 2-focus velocimeter and the optical fiber thermometer and the optical fiber thermometer respectively. The flame temperature and the soot particle density were calculated by applying the IR two- color method to the measured radiant energy from the soot particles in the flame. In the analysis of the frequency power spectra of temperature and velocity fluctuations, the correlation-based slotting technique was adopted for those data with the nonuniform time interval. It is shown that the time mean value and the fluctuation of the flame temperature decrease gradually toward downstream in the luminous flame region, and those of the soot density increase due to decay of turbulence along the flame axis. On the other hand, both time mean and fluctuation of the flame temperature increase in the radial direction from the center to the periphery due to the effect of air entrainment marked in the peripheral region of the flame. Furthermore, the power spectrum of the velocity fluctuation is not always the same as that of the temperature fluctuation in the flame center.© (1997) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

When a column of droplets freely falling from an ultrasonic atomizer was ignited behind a reflected shock, no ignition occurred at a temperature below 1100 K, even if the pressure was as high as IMPa. Although, a higher temperature condition ensured ignition, no luminous flame was observable by high-speed photography, and even if a luminous flame lump appeared at an extremely high temperature, it disappeared without spreading over the entire column of droplets in this case. It is known however that, if a fuel is injected into a diesel cylinder or an electric furnace, ignition occurs even at a temperature as low as 650 K with a luminous flame spreading over the entire spray. These differences could be caused by the effects of turbulent mixing between fuel droplets and hot air, in fact, turbulence-generating rods were placed on the upstream side of the spray column. Experimental results indicates that the ignition limit was lowered to 840 K, and the ignition delay period was decreased by increasing the intensity of turbulence. Furthermore, the light emission of the flame was intensified, and normal spray combustion was maintained in the low-temperature atmosphere after the shock tube ceased its operation.

An endoscope-based image acquisition-and-processing camera system was used for diagnostics of pilot injection combustion in a single-cylinder heavy duty diesel engine. A study of the pilot injection or light load is of interest because the spray breakup, mixing and vaporization processes are less influenced by heat feedback from the flame than in full injection cases. This allows the spray process to be decoupled from the combustion process. The experimental cases were modeled using a version of the KIVA-II code that includes improvements in the turbulence, wall heat transfer, spray, ignition and combustion models. The combustion of the pilot injections was characterized by ignition sites located below the point where the spray impinges on the piston bowl surface for the injector configuration used in this study. Multiple ignition sites were observed and the majority of the combustion occurred at the bottom of the piston bowl, spreading along the bowl edge. As the start-of-injection timing was retarded some evidence of ignition above the impingement point was observed. However, these ignition sites did not develop into major combustion zones. Comparisons of the experimental results, which included pressure traces, heat release rates and the luminous flame images, and the numerical computations were made tomore » assess the performance of current models in the KIVA-II code. Good agreement was obtained for the timing and location of ignition. The penetration of the flame observed in the luminous flame images was also captured in the predictions. A feature of the pilot injection that was not captured by the prediction is the spread of the flame along the edge of the piston bowl. The comparisons indicate that recent improvements in KIVA-II allow better prediction of the combustion of pilot injections.« less

A droplet cloud of liquid fuel produced by an ultrasonic atomizer was ignited by a spark, and the flame ball propagating outward was observed in order to elucidate the mechanism of flame propagation and complicated group combustion behaviors of spray flames. For that purpose, the instantaneous images of droplet clusters, OH-radical chemiluminescence and C2-band flame luminosity were taken simultaneously. Furthermore, the light emission signals in OH- and CH-bands, Mie-scattering signal from droplets, and the size and velocity of droplets were monitored simultaneously in time series. It was found that a nonluminous flame propagated ahead of a luminous flame, and that droplets disappeared in the luminous flame zone due to rapid evaporation, where a number of small-scaled droplet clusters were burning in diffusion combustion mode associated with solid-body emissions.

Absorption spectroscopy, incorporated with multiparameter least-squares analysis, was used to determine profiles of @-phase temperature and NO and OH concentrations of RDX, XM39 and M43 solid propellants at pressures up to 3.55 MPa. The deduced RDX f d flame temperatures at different pressures are very close to results of equilibrium calculations; however, the deduced OH concentratiom are smaller than equilibrium solutions. The major source of measurement error in OH concentrations is in the determination of light path length Temperature and NO concentration profiles of M43 and XM39 showed plateaus when a dark zone existed between the burning surface and luminous flame regions. However, when IuminoUS flamelets were attached to the burning surface, line-of-sight averaged NO concentrations (where a uniform path length is used in data analysis) exhibited a rapid reduction for in~eased istance from burning surface This reduction is caused by shorter intercepted distance by the W light beam in the dark zone as the vertical distance from the burning surface increases, and the deduced NO concentration is thereby decreased if the whole propellant diameter is used as the path length in data analysis. Even though M43 and XM39 have simila~ dark-zone temperatures (around 1,050 K forM43 and 1,150 K for XM39), M43 has about 5% more of NO concentration than XM39 in the darkzone region (21% vs. 16%). d

A Precessing Jet (PJ) burner has been developed which enhances large scale turbulence by means of a naturally occurring fluid-mechanical phenomenon. A highly stable gas flame is produced without air swirl, quarl or a bluff body stabiliser. The resulting natural gas flame is lower in temperature and is more luminous than flames from many conventional burners so that it reduces NO x emissions while achieving almost total burnout of combustibles. The high luminosity indicates potential for increased fuel utilisation efficiency in process industries where radiation heat transfer is important. For one configuration of the PJ burner firing natural gas in an 18MW industrial cement kiln, NO x emissions are shown to be reduced by 75% with a simultaneous reduction in CO when compared with the original turbulent jet burner. Photographs of these flames show that the PJ burner produces a more rapidly spreading and more luminous flame. For another configuration, firing natural gas in a 25kW laboratory furnace, the radiation spectra of the flame produced by the PJ burner are compared with a high quality commercial swirl burner. The mean radiation intensity produced by the PJ burner is some twenty times that of the commercial burner in the visible spectrum and the normalised fluctuating intensity is about 50%, some five times that of the commercial burner in the visible range. Good potential for flame detectability is therefore apparent. Mean intensity is shown to increase by a factor of three with the addition of swirl to the air.

Soot formation process was examined by high speed photographs, using a single combustion diesel engine with a transparent swirl chamber. Fuel-air mixture and flames, and soot clouds were visualized by the schlieren method and the back-illuminated method, respectively. A three dimensional simulation program with soot formation and oxidation models was developed to clarify diesel soot formation processes. The models consist of several models previously proposed and partly improved in this study. Good agreement was obtained between calculated and experimental results. The following points were clarified through observation and numerical studies: The main soot area is considerably smaller than luminous flame area, especially in the initial soot formation process; the main soot cloud first appears in the tip region of fuel-air mixture, downstream of ignition position a few submilliseconds after the ignition. It is the soot carried down from the ignition position by a gas flow; and temperature is more influential in soot formation, rather than fuel vapor concentration.

A laminar methane-air diffusion flame has been extensively probed for its temperature, velocity, and stable species' concentration profiles. The data obtained were used to validate a theoretical model capable of characterizing flame macrostructure and useful in predicting the effects of flow rate, equivalence ratio, fuel and air preheat, and nozzle size on the thermal and aerodynamic fields established in confined, axisymmetric, laminar diffusion flames. The model utilizes the Burke and Schumann flame sheet concept to locate the stoichiometric fuel-oxygen interface and, hence, the points of heat release. The model eliminates the restrictions of the classical Burke-Schumann model in that allowances are made for natural convection effects and variable thermodynamic and transport properties. The stoichiometric fuel-oxygen interface was found to coincide with the outer edge of the primary reaction zone and represented a surface through which an amount of oxygen in stoichiometric proportion to the amount of fuel fired diffuses. The inner edge of the reaction zone was defined by the profile of the luminous flame core and was found to be well fitted by the theoretical predictions when the ratio of oxygen and fuel diffusing to the flame front was assigned a value 0.88 times the stoichiometric coefficient. The concentration and temperature profiles could be generalized by plotting the results as a function of the local-equivalence ratio.

Combustion wave structures of noncatalyzed and catalyzed double-base propellants were studied to clearly identify and explain the plateau burning mechanism with photographic observations and temperature profile measurements. The propellants used in this study were specially formulated to obtain a wide range plateau burning region that would occur at the lowest possible pressure and burning rate. The reaction processes in the dark zone were unchanged by the addition of the catalyst. While the burning rate remained constant in the plateau region, the luminous flame of the catalyzed propellant approached the burning surface more rapidly than the luminous flame of the noncatalyzed propellant, when the pressure was increased. The temperature gradient in the fizz zone was significantly different for the noncatalyzed and catalyzed propellants. The temperature gradient in the fizz zone of the noncatalyzed propellant increased monotonously with pressure, whenreas the temperature gradient in the fizz zone of the catalyzed propellant increased in the super-rate region and was constant in the plateau region. It was found that the change of the temperature gradient in the fizz zone with pressure is similar to the change of the burning rate with pressure for each propellant. The temperature gradient in the fizz zone of the catalyzed propellant remains unchanged in the plateau region, which indicates that the reaction rate is constant. Therefore, the conductive heat feedback from the fizz zone to the burning surface remains unchanged throughout the plateau pressure region. However, no special trends, in the burning surface temperature and the heat release in the subsurface reaction zone were observed in the plateau burning region, when compared with the normal burning of the noncatalyzed propellant. Thus, the pressure insensitive reaction in the fizz zone is considered to be directly responsible for plateau burning.

A shock tube technique was developed to observe the ignition process of sprays free from aerodynamic phenomena, such as atomization and mixing. In this technique, a premixed spray column, falling freely from an ultrasonic atomizer directed downwards, was ignited by a reflected shock. The ignition processes observed by this technique seem to be different in essence from those observed by the usual injection technique, in which fuel is injected into a hot atmoshere. The probability of normal ignition occurring reduces rapidly as the temperature is lowered below 1060 K for cetane and a pressure of 0.85 MPa. The luminosity of the flame varies widely up and down, and the period of appearance of a luminous flame observable by high-speed photography is only a part of the ignition process. The dependences of the ignition delay on the temperature and pressure also differ substantially from those obtained by the injection technique.

Certain metal organic salts (e.g., lead or copper salicylate) when used in double-base propellants induce desirable insensitivities of burning rate to pressure and initial temperature. To understand this, the combustion wave zones (luminous flame, dark, fizz, and surface reaction zones) were examined by means of photography and fine thermocouples (4fi bead). The metal salts significantly alter the surface and fizz zones. The surface zone accumulates carbonaceous material coincident with the appearance of an accelerated burning rate in the catalyzed case. No attendant change in surface heat release is detected. Coinciding with this carbonaceous layer occurrence are substantial (50-100%) increases in conductive feedback from the fizz zone. This latter effect is believed directly responsible for the altered burning behavior though its origin may lie in the altered surface chemistry.

Measurements of the combustion wave structure of solid propellants with and without cross-flow conditions were conducted to elucidate the mechanism of erosive burning. The propellants tested were three types of double-base propellants containing different energy levels: a lowenergy, a reference, and a high-energy propellant. The test results indicate that the erosive burning ratio decreases with increasing energy contained in the unit mass of propellant. The threshold velocity is lower for the low-energy propellant than for the high-energy propellant. Measurements with microphotographs and microthermocouples revealed that the dark zone which separates the luminous flame and the burning surface diminishes completely under cross-flow conditions. The heat flux transferred back from the fizz zone to the burning surface increases when the cross-flow velocity exceeds the threshold velocity. Erosive burning occurs when the fizz zone parameter Φ is increased by the turbulent mixing generated by the cross-flow. The analysis presented in this study indicates that the fizz zone parameter is correlated uniquely with the friction force τ, at the surface as Φ∼τ0.8.

A detailed mapping is presented of the temperature, CO, CO 2 and O 2 species concentrations, and monochromatic absorption coefficient at the base region of a kerosene pool fire. Comparative analysis of the isotherm and iso-concentration patterns indicates that the buoyant flame has an intermittent character at its base, contrary to its normally attributed persistent, or anchored, character. The temperature field shows a band of temperature maxima that approximately coincides with the luminous flame boundary. The temperature in this band, however, decreases toward the fire axis and is lower than that expected in a kerosene flame. This, in conjunction with the presence of high oxygen concentration in the fuel side of the temperature maxima band, indicates that the flame fluctuates around a mean position, and that as a consequence the measured temperatures are lower averaged values of the fluctuating temperature field. The CO and CO 2 iso-concentration fields show a region of elevated concentrations at the axis near the liquid surface that is displaced toward the fuel side with respect to the temperature maxima region, rather than coinciding with it. The lines of constant absorption coefficient, which can be associated with lines of constant soot concentration, also show a region of elevated values that approximately coincides in location with that of the CO and CO 2 concentrations. These results indicate the fluctuating character of the flame, and suggest the existence of a gas stagnation, or recirculation, zone at the axis near the liquid surface that permits the accumulation of combustion products in this region. From the observed characteristics of the flame structure it can be concluded that the flame at the pool fire base is better described as a fluctuating, laminar, diffusion flame, which later becomes a turbulent, intermittent one as it evolves upward along the fire plume.

The effects of radiant heat fluxes on the stability range and burning velocity of a laminar premixed flame were examined by inserting a water-cooled Bunsen burner into a radiant heat recirculation-type furnace. The intensity of the radiant heat flux in the furnace was altered by regulating the surface temperatures of the furnace wall and downstream permeable plate. It was found that the flame blowoff and flashback limits were hardly affected by a radiation field as intense as the one surrounded by a refractory whose temperature was between 1250K and 1450K, and that no perceptible changes appeared either in the laminar burning velocity or flame thickness.

Dependence of the height of turbulent wall flames and the wall flame heat transfer on the fuel conditions is studied by experiments using porous line burners against walls. The flame height is represented as a function of heat release rate and fuel width. According to the data of incident heat flux to the wall surface, the wall above a fire source appears to divide into four regions following the thickness and intermittency of a luminous flame. The distribution of the incident heat flux is represented as a function of the height normalized by a scale length representing the flame height. Exploratory analysis is made to explain this result from the heat transfer mechanism. The present analysis seems to be consistent with current experimental work on wall flames.

An experimental study on flame propagation through an interface between layered methane and air, propane and air, and methane and oxygen in inert baths has been made with a combustor having a length of 10 cm. Detailed flame structure was investigated by means of schlieren photography and a hot wire anemometer. The premixed flame, the diffusion flame, and the convection flame just like those propagating in line appeared in the limited combustor space, but one after another with some time lags in between. The diffusion flame appeared about 80 msec after the premixed flame passed, which equivalently corresponds to about 10 cm of space existing between the premixed flame and the forefront of the diffusion flame. An anomalous structure was observed in the premixed flame. That is, a post burning occurs immediately behind the conventional flame front which forms double flame fronts. The temperature in between these is flat, and the separated gap width increases towards lean and rich mixtures. Steady fla...

Research was made with double-base propellants with embedded metal wires to understand how the burning rate changes along the silver wire and to obtain what factors control the burning rate of the propellants. The burning rate along the wire is influenced greatly by the kind of metal wire, wire size, and propellant compositions. With double-base propellants having varied burning characteristics, burning rate along the silver wire was measured and the effects of the luminous flame, dark zone, and the fizz zone were studied. By means of microthermocouples, the temperature profiles in the propellant near the silver wire were measured. Based on the test results, it was clear that the factors which control the burning rate along the silver wire are the temperature gradient in the fizz zone and the temperature in the dark zone. The luminous flame zone that stands above the burning surface affects neither the burning rate of the propellant nor the burning rate along the silver wire.

Nitric oxide and hydrogen cyanide concentration distributions established in an unseeded, laminar, atmospheric methane-air diffusion flame have been analyzed in order to gain insight into the role nitrogen-containing species play in influencing NO formation in such flames. The experimental data indicate that HCN and NO formation are strongly dependent on the local stoichiometry within the flame region. Cyanide formation occurs only within the fuel-rich luminous flame core, presumably via reaction between hydrocarbon radicals and molecular nitrogen or nitric oxide. The cyanide species are subsequently partially converted to nitric oxide on the fuel-lean side of the primary reaction zone, contributing 15-30 percent of the total nitric oxide production by the flame. The nitric oxide concentration profiles peak on the fuel-lean side of the flame in the zone where the thermal production of NO by the extended Zeldovich mechanism occurs.

Swirl type combustor for study of spray combustion has been designed. Using this combustor luminous flame and non-luminos flame can be produced by shifting the spray nozzle position without changing the operating conditions. From the experimental studies the following results were obtained. (l) Supplying combustion air from tangential port, swirl number of S=0.99 was obtained. The circulation length was nearly equal to the combustion tube diameter. (2) Flame appearances were controlled by diffusion process of fuel spray and combustion air. If fuel spray was injected into a reverse flow in torroidal vortex, combustion process was so fast that non-luminous flame was obtained. (3) Paying attention to the primary combustion, combustion efficiency of swirl-type combustion was found better in the case of luminous flame, and it was shown combustion performances were controlled by the flow pattern and the spray directions.

The combustion mechanisms of nitramine composite propellants were investigated to determine the parameters that control the burning rate in combustion waves. The nitramines studied were RDX (cyclotrimethylene trinitramine) and HMX (cyclotetramethylene tetranitramine), which have similar chemical compositions, but different reaction schemes. Since RDX and HMX composite propellants consist of crystalline nitramines and a polymeric binder, the physical structure of these propellants is heterogeneous. From experiments, it was found that the combustion waves of these propellants are homogeneous. In addition, the combustion waves exhibited a two-stage flame structure. The luminous flame produced above the burning surface approached the burning surface when the pressure was increased. However, the burning rates of both RDX and HMX propellants were not affected by the luminous flame reaction. In the preparation zone separating the luminous flame and the burning surface, the gas phase reaction primarily consist of the exothermic reduction reaction of NO to N 2 which is similar to the reaction in the dark zone of double-base propellants. From quenched burning tests, it was found that the nitramines and the binder melted at the burning surface and formed an energetic mixture. Thus, the decomposed gases ejected from the burning surface consisted of a homogeneous mixture of oxidizer and fuel species. The temperature profile measurements in the combustion waves revealed that there is a large heat release at the burning surface. This exothermic decomposition reaction at the surface plays a major role in controlling the burning rate of RDX and HMX composite propellants. The higher burning rate of RDX propellants, as compared to HMX propellants, is due to the larger heat release at their burning surface.

Measurements of luminous flame extensions beneath ceilings under steady burning conditions are presented. Tests were conducted using both axisymmetric and corner-wall-ceiling configurations for a range of energy supply rates up to 400 kW and burner-to-ceiling heights up to 2·3 m. Flame length observations, which were made both visually and photographically, are expressed in dimensional and in nondimensional terms. Comparisons are made with previous measurements reported in the literature.

A technique was developed in which the light emission images of radicals, OH and CH, were visualized, even in a luminous flame or near-glowing bodies, by processing the monochromatic images of the flame at and in the neighborhood of the emission band of the radical. This was made by one or two interference optical filters and an image intensifier. The details of the technique were examined by spectroscopic observation of a Wolfhand-Parker flame, as well as by the computer simulation of the image processing. The accuracies of the emission images of radicals experimentally obtained were evaluated by comparing them with their counterparts estimated by simulation. It was found that this technique was applicable also to fluctuating flames, such as premixed spray flames or turbulent premixed flames, if time-averaging processes were involved.

The spectroscopic analysis of the radiation from luminous flames burning propane and methane fuels were carried out in the wavelength range from 0.4 ..mu..m to 5 ..mu..m and monochromatic spectra of the radiation from the soot particle cloud were measured. A large discrepancy between experimental and predicted k/sub lambda/ from Mie theory is observed. Calculations from Mie theory were performed to clarify the effect of the shape, size, and complex refractive index of soot particles.

The combustion wave structure and thermal decomposition process of HMX were examined in order to elucidate the burning rate characteristics of HMX. The combustion wave can be divided into three zones: nonreactive solid-phase, surface reaction, and gas-phase reaction zones. Measurements with micro-thermocouples revealed that the heat flux produced in the surface reaction zone is approximately equal to the heat flux transferred back from the gas phase to the burning surface. Accordingly, the reaction process in the suface reaction and the gas phase zones plays a dominant role in the burning rate of HMX. The gas phase reaction zone consists of a two-stage reaction process: the first stage is the exothermic rapid reaction process between NO2 and aldchydes, and the second stage is the exothermic slow reaction process between NO and N2O and remaining fuel species. The luminous flame zone which is determined to be the second stage reaction process approaches rapidly the burning surface as pressure increases. However, the luminous flame reaction appears to be little responsible for the burning rate of HMX. Examinations of the quenched burning surface of HMX samples revealed that the burning surface melts and forms a noncrystallized intermediate material. The surface structure appears to be different from the structure of thermally degraded HMX samples which were obtained by a thermogravnmetric analysis.

Planar visualization was employed to study flame structure and low frequency flame oscillation. Two distinct vortices were visualized in the flames studies: large toroidal vortices outside the luminous flame and small roll-up vortices inside the luminous flame. The flame oscillation frequency and the convective velocity of the toroidal vortices were measured for ethylene, methane, and propane diffusion flames over a wide range of test conditions. The frequency was typically in the range 10 to 20 Hz and the convective velocity was approximately, 0.8 m/s. The frequency of the toroidal vortices was found to correlate with the flame oscillation frequency. The potential effects of the toroidal vortices on the flame dynamics at low fuel flow rates are discussed, for example, low frequency flame oscillation, non-linear flame bulge motion, and quenching of the luminous flame surface.

The opposed jet assembly has been adapted for stabilizing a counter flow diffusion flame of ethanol. Data on the thermo-aerodynamic structure of the flame, obtained with the help of interferometric and particle track techniques, have been presented and analysed to yield the heat release rate. The maximum flame temperature is found to be close to adiabatic and the net heat release is in agreement with the theoretically expected value. The luminous flame is accompanied by endothermic zones on both the fuel and oxygen sides. Mass spectromctric analysis of the flame gases is found to confirm the presence of ethanol pyrolysis on the fuel side and to point out the possibility of significant chemical reactions on the oxygen side. The observations thus tend to strengthen the earlier findings (of endothermic zones) on gaseous diffusion flames.

During the combustion process, a thin, cool, fuel-rich core region is established near the fuel surface. In large-scale fires, where radiation is the dominant mechanism for energy feedback to the fuel surface, the attenuation of the radiant flux by the core gases significantly affects the mass consumption rate. A simple flame-radiation model is proposed to account for this blockage effect. It consists of two discrete, homogeneous, isothermal regions: a high temperature, luminous flame, and a cool, fuel-rich core. The thin core layer (about 7 percent of the flame height) has a surprisingly significant impact on the radiant feedback, as it attenuates the radiant flux to the surface by 25 to 35 percent. The simplified model is able to predict, with great accuracy, the mass consumption rate and total radiative power output of PMMA pool fires without rigorous specification of the temperature and concentration distributions within the flame.

Triaminoguanidine nitrate (TAGN: CH9N7O3) is a high-energy explosive which contains a large amount of hydrogen in the molecular structure. The oxidizer fragment (HNO3) is attached to the fuel fragment, by an ionic bond. In this study, the measurements of burning rate, combustion wave structure, and thermochemical properties were conducted. The burning rate of TAGN increases linearly with increasing pressure in a ln (pressure) vs In (burning rate) plot. The pressure exponent of burning rate is determined to be 0.78. The exothermic reaction at the first-stage decomposition process (488–498K) is considered to be caused by the breakage of N-NH2 bond. The heat produced in the surface reaction zone during TAGN burning is also determined to be 525 kJ/kg. The luminous flame stands some distance above the burning surface. The flame standoff distance decreases with increasing pressure. The heat flux fedback from the gas phase to the surface reaction zone increased with increasing pressure. The burning rate of TAGN depends mainly on the heat released in the surface reaction zone and the heat transferred back from the gas phase to the surface reaction zone.

For the gelled JP-4 of an oil-in-water emulsion type which includes 96∼99% oil concentration by volume as an internal phase and has the drainage temperature higher than 70°C in confined vessel, the flame spreading aspects with a pool combustion were observed by taking movie, and the measurement of temperatures, on and above the surface of the gel from the time when the flame is coming over to the time when it has passed away, was carried out with thermocouples fixed in place.While the flame is spreading on the gelled JP-4, an advance liquid flow ahead of the flame can be found in case of the initial gel temperatures lower than the apparent flash points (26∼35°C) measured by means of open cup method. This liquid flow invades on to the surface of the virgin gel, issued from the upstream region covered already with luminous flame, and is not formed by radiative and conductive heat transfer from the flame pillar to the virgin gel surface. In view point that the temperatures of foregoing liquid flow is 10∼20°C at the initial level of gel surface and the flash point of the original JP-4 is -23°C, a large part of the low hydrocarbon fractions seem to have burned in the flame zone before streaming out on the virgin gel surface. The behavior of liquid flow was observed in detail with fine aluminum powders sowed in the vicinity of the ignition position. As the result, it is found that in the foregoing liquid flow there is a strong convection due to temperature gradient. Thus, the covering of the liquid flow on the virgin gel plays a role to reduce the drainage rate of the gel, protecting the gel surface from being exposed in air in the place close to hot flame, and the flame spreading velocity can be suppressed by one-tenth to one-twentieth, compared to that of the original JP-4.

Particulate matter in axisymmetric laminar diffusion flames produced by burning mixtures of either CO and trimethylborate (TMB) or CH4 and TMB with air were investigated using laser light-scattering techniques. Boron oxide particle sizes and number densities were determined at various heights in the flames using polarization ratio and relative intensity measurements, respectively. In the CO/TMB flames, two distinct particle-laden regions were found. The first region was located on the rich side of the luminous flame zone and initially appeared as a narrow annulus, which grew in width downstream until the particles filled the core. A second thin annular zone appeared on the air side of the flame zone, starting approximately at the height of the luminous green flame tip and continuing to grow downstream. Particle sizes did not vary significantly with location in the flames, with diameters of approximately 0.09 and 0.15 μm in the 95% CO/5% TMB and 90% CO/10% TMB flames, respectively. Corresponding peak number densities were approximately 1.5 × 1010 and 6 × 109 cm−3. The CH4/TMB flames were considerably different than the CO/TMB flames. The presence of significant quantities of water vapor presumably contributed to the formation of HBO2(g) in favor of condensed-phase B2O3. At locations where oxide particles did form, they were closer to the flame centerline than the soot-containing regions. Computations of equilibrium yields of condensed-phase oxide were in qualitative agreement with the experimental results.

各種燃料の燃焼形態は大別して輝炎, 不輝炎である. 単一燃焼器でこれら火炎が任意に得られる旋回流形式による噴霧燃焼器の試作を示す. 燃焼に関与する諸条件(燃料量, 噴霧特性, 空気量など)を同一としたままで, 旋回流の流れ場と噴霧主流との相対位置を選択することによって燃焼を制御, それぞれの燃焼形態での燃焼特性を明らかにした. これらデータは噴霧燃焼のモデリング, あるいは数値計算による予測に利用が期待される.

A simple expression for the total emissivity of a luminous flame has been established on the basis of a flame-radiation model that consists of gray soot and non-gray gases. The gray-soot approximation is demonstrated to be an accurate characterization of soot emission from both the analytical formulation and the experimental observations. This simple model also allows a comparative analysis of soot emission characteristics in flames produced by various common solid and gaseous fuels. The present calculations for the luminous-flame emissivity show excellent agreement with existing experimental data and exact analysis.

The pressure dependences of carbon particle cloud radiation and of nunluminous gas radiation are examined individually in order to predictradiation from luminous flame of liquid fuel at high pressures. For carbon particle cloud radiation, the equation of carbon formation is determined in terms of pressure and excess air factor with reference to the experiments on the soot formation from vaporized fuel at low excess air factor and on the combustion of a single droplet of fuel in a large chamber. For nonluminous gas radiation, the expressions obtained by Edwards are adopted. A method is developed to calculate the luminous flame radiation at high pressures using the above expressions and the experimental results at an atmospheric pressure. The calculated results agree well with the measured values. The maximum, average and endzone absorption coefficients in the luminous flame can be easily calculated.

Luminous radiation from the soot particles of a flame propagating vertically downward on a polymethyl methacrylate (PMMA) rod is investigated by using exact and approximate analyses. Investigation shows that for vertically downward burning of a 0.5 in. diam PMMA rod, the luminous radiative heat transfer from the flame to the fuel rod is well within the optically thin limit. The approximate analysis is extended to conventional hydrocarbon fuels. A simple optically thin approximation in terms of flame temperature T f and mean beam length L m is obtained for the luminous radiation from soot particles. The thin approximation gives errors of less than 5% from exact values for L m T f

Flame structures and burning rate characteristics of composite modified double-base (CMDB) propellants are studied by photographic observations and temperature profile measurements of the combustion zones. The oxidizers studied for CMDB propellants were various particle sizes and concentrations of AP (ammonium perchlorate) and HMX (cyclotetramethylene-tetranitramine), each of which has a different reaction scheme. AP particles decompose at the burning surface and diffuse into the decomposition gas of the DB (double-base propellant) matrix in the fizz and dark zones. The diffusing decomposition gases shift in equivalence ratio toward stoichiometric and produce an AP/DB luminous flame of increased reaction rate and flame temperature. Increased temperature gradient in the fizz zone and large temperature fluctuations in the dark zone were observed. The increased temperature gradient increases the heat feedback from the gas phase to the burning surface, and increases the burning rate. On the other hand, HMX particles in CMDB propellants gasify at the burning surface, but the flame structure and the temperature profile in the fizz and dark zones are not altered. The burning rate was slightly decreased by the addition of HMX. Based on the experimental observations, simplified burning rate models' were presented. The models describe fairly well the observed effects, of the concentration and particle size of oxidizers on burning rate characteristics.

In 1863 Julius Thomsen communicated to the “Naturforscherversammlung” at Stockholm an account of some determinations he had made of “The Mechanical Equivalent of Light,” and an abstract of this paper appeared in ‘Poggendorff’s Annalen’ (vol. 125, 1865, p. 348). He allowed the radiation from a sperm candle, a moderator lamp, and a gas flame, to fall on the face of a thermopile, and noted the deflection of the needle of a galvanometer in the thermoelectric circuit, when the radiation fell directly on the pile, and when it did so after passing through 20 cm. of water contained in a glass cell.

A shock tube method is described for studying the combustion of a cloud of coal particles under pressure. Two important types of reaction are observed. First, the preflame reaction, in which particle heating rates are controlled by conduction from the gas together with the exothermal release of heat at the particle surface. This stage always precedes the second type of reaction which is characterized by a highly luminous flame. Here solid particles of less than 80 μ m and their products of volatilization appear to burn simultaneously and completely in less than 2 ms. The kinetics of the two types of reaction are analysed by means of thermal explosion theory and oxygen transport rates respectively.

Further development and testing of the zone method of analysis is described. In the zone method, the temperature and heat-flux distributions in a furnace are obtained by dividing the system into zones, and solving simultaneously energy balances on each zone. The present state of development of the method is discussed, and avenues for further development are considered. A method of including the effect of luminous radiation is described. For this the emissivity ofa cloud of soot particles is represented as the weighted sum of grey gases. It is also shown that the emissivity of a gas-soot mixture can be represented as the weighted sum of grey gases. The validity of this emissivity equation is examined by comparing calculated and measured distributions of radiation intensity through a luminous flame. Good agreement between calculated and measured traverses is found. With luminous flames, large variations in gas-absorption coefficient occur in furnaces.Methods of allowing for variations of gas-absorption coefficient when evaluating radiative “exchange areas” are discussed, and an approximate method is presented. A method of allowing for different types of heat sink in a surface zone is described. Themethod allows for the replacement of the composite surface with an equivalent grey surface having an effective emissivity and temperature. Also, surface energy factors are derived which enable the radiant energy at a surface to be divided in terms of the amount of heat absorbed by each type of sink in the zone. Experimental testing of the zone method of analysis is described. The zone method incorporatingthe abovementioned improvements was used to predict temperature and heat-flux distributions in the IJmuiden furnace. The initial conditions of flow pattern, heat release, and absorption coefficient were obtained from the experimentally measured properties in the furnace. The predictions were compared with measured values of gas temperature and surface heat flux, and good agreement was found. It is concluded from the agreement shown, that the zone method of analysis is a realisticmathematical model which can be used with some confidence for the calculation of heat transfer in furnaces.

Luminous-flame firing of natural gas in glass furnaces has been accomplished inefficiently and with, difficulty for many years in an attempt to secure the advantages of the luminosity characteristics. Luminosity in the gas flame is now secured by retardation of combustion between overlying layers of air and gas, the resulting flame being particularly adapted for glassmelting. A successful method for controlling the length and shape of the luminous flame in a glass furnace involves the mixing of spent gases from the flue with the incoming fuel gas.