Robert J. Santoro

Spatially resolved measurements of soot volume fraction using laser-induced incandescence

Laser-induced incandescence is used to obtain spatially resolved measurements of soot volume fraction in a laminar diffusion flame, in which comparisons with laser scattering/extinction data yield excellent agreement. In addition, the laser-induced incandescence signal is observed to involve a rapid rise in intensity followed by a relatively long (ca. 600 ns) decay period subsequent to the laser pulse, while the effect of laser fluence is manifest in nonlinear and near-saturated response of the laser-induced incandescence signal with the transition occurring at a laser fluence of approximately 1.2 × 108 W/cm2. Spectral response of the laser-induced incandescence involves a continuous spectrum in the visible wavelength range due to the blackbody nature of the emission. Simultaneous measurements of laser-induced incandescence and light scattering yield encouraging results concerning the mean soot particle diameter and number concentration. Thus, laser-induced incandescence can be used as an instantaneous, spatially resolved diagnostic of soot volume fraction without the need for the conventional line-of-sight laser extinction method, while potential applications in two-dimensional imaging and simultaneous measurements of laser-induced incandescence and light-scattering to generate a complete soot property characterization are significant.

Aerosol dynamic processes of soot aggregates in a laminar ethene diffusion flame

Abstract Laser scattering/extinction tests on a coannular ethene diffusion flame were analyzed using cross sections for polydisperse aggregates. Using an improved experimental arrangement that allowed simultaneous measurement of light scattering at multiple angles, it was possible to determine the fractal dimension of the aggregates in the flame. The analysis also yields the mean-square radius of gyration, the aggregate number concentration, the average number of primary particles per aggregate, as well as the volume average of the volume-mean diameter as a function of height or residence time along the particle path of maximum soot concentration in this flame. These results lead to the conclusion that soot aerosol dynamic processes in the laminar ethene flame are partitioned into four regions. Low in the diffusion flame there is a region of particle inception that establishes the number of primary particles per unit volume that remains constant along a prescribed soot pathline. In the second region, there is sustained particle growth through the combined action of cluster-cluster aggregation (CCA) accompanied by heterogeneous reactions contributing to monomer-cluster growth. Oxidation processes occur in the third region where CCA continues. If aggregate burnout is not complete in the oxidation region, then smoke is released to the surrounding in the fourth region where reactions cease but clusters continue to grow by CCA. The experiments yield the CCA growth rate within the flame which compares favorably with the theoretical value. The similarities and differences between this data reduction and the traditional analysis based on the use of cross sections for Rayleigh spheres and Mie theory spheres is discussed.

Two-dimensional imaging of soot volume fraction by the use of laser-induced incandescence.

A recently developed laser-induced incandescence technique is used to make novel planar measurements of soot volume fraction within turbulent diffusion flames and droplet flames. The two-dimensional imaging technique is developed and assessed by systematic experiments in a coannular laminar diffusion flame, in which the soot characteristics have been well established. With a single point calibration procedure, agreement to within 10% was found between the values of soot volume fraction measured by this technique and those determined by conventional laser scattering-extinction methods in the flame. As a demonstration of the wide range of applicability of the technique, soot volume fraction images are also obtained from both turbulent ethene diffusion flames and from a freely falling droplet flame that burns the mixture of 75% benzene and 25% methanol. For the turbulent diffusion flames, approximately an 80% reduction in soot volume fraction was found when the Reynolds number of the fuel jet increased from 4000 to 8000. In the droplet flame case, the distribution of soot field was found to be similar to that observed in coannular laminar diffusion flames.

The oxidation of soot and carbon monoxide in hydrocarbon diffusion flames

Abstract Quantitative OH · concentrations and primary soot particle sizes have been determined in the soot oxidation regions of axisymmetric diffusion flames burning methane, methane/butane, and methane/1-butene in air at atmospheric pressure. The total carbon flow rate was held constant in these flames while the maximum amount of soot varied by a factor of seven along the centerline. Laser-induced fluorescence measurements of OH · were placed on an absolute basis by calibration against earlier absorption results. The primary size measurements of the soot particles were made using thermophoretic sampling and transmission electron microscopy. OH · concentrations are greatly reduced in the presence of soot particles. Whereas large super-equilibrium ratios are observed in the high-temperature reaction zones in the absence of soot, the OH · concentrations approach equilibrium values when the soot loading is high. The diminished OH · concentrations are found to arise from reactions with the soot particles and only to a minor degree from lower temperatures due to soot radiation losses. Analysis of the soot oxidation rates computed from the primary particle size profiles as a function of time along the flame centerlines shows that OH · is the dominant oxidizer of soot, with O2 making only a small contribution. Higher collision efficiencies of OH · reactions with soot particles are found for the flames containing larger soot concentrations at lower temperatures. A comparison of the soot and CO oxidation rates shows that although CO is inherently more reactive than soot, the soot successfully competes with CO for OH · and hence suppresses CO oxidation for large soot concentrations.

Atomization characteristics of impinging liquid jets

The atomization characteristics of sheets formed by both laminar and turbulent impinging jets were experimentally studied as a function of flow and injector geometric parameters. In particular, sheet breakup length along the sheet centerline, distance between adjacent waves apparent on the sheet, and drop-size distributions were measured over a Weber number range between 350-6600 and a Reynolds number range between 2.8 x 103 to 2.6 x 10 4. A linear stability-based model was used to determine the most unstable wave number and the corresponding growth rate factor on two-dimension al thinning inviscid and viscous sheets. These wave characteristics were used to predict both the sheet breakup length and the resulting drop sizes. A second model, applicable for a low Weber number regime, in which sheet disintegration is controlled by stationary antisymmetric waves, was used to predict the shape of the sheet formed by two impinging liquid jets. The linear stabilitybased theory predictions of breakup length did not agree in trend or magnitude with experimental measurements. However, for Weber numbers less than 350, the measured breakup length for laminar impinging jets was within 50% of that predicted by the stationary antisymmetric wave-based model. Finally, drop-size predictions based on linear stability theory agreed in trend, but not in magnitude, with the measured drop sizes. The contrast between the sheet atomization characteristics of laminar vs turbulent impinging jets suggest that the initial conditions of the impinging jets significantly influence the sheet breakup mechanism. Also, the comparison between experimental results and theoretical predictions indicates that the impact wave generation process at the jet impingement point needs to be incorporated in the theoretical models for sheet atomization. Nomenclature d = diameter F = thickness distribution h = sheet thickness k = wave number L = length of injection element / = length r = radial distance from impingement point Re = Reynolds number, Ujdjv/, based on liquid properties, jet velocity, and orifice diameter Rex = Reynolds number, Ushlvi, based on liquid properties, sheet velocity, and sheet thickness 5 = ratio of gas density to liquid density t = time U = velocity W = maximum width of sheet We = Weber number, piUjdJcr, based on liquid properties, jet velocity, and orifice diameter Wes = Weber number, p,£/;/z/cr, based on liquid properties, sheet velocity, and sheet thickness x = axial distance from impingement point y = coordinate perpendicular to x in the plane of the sheet a = fan inclination angle ft = complex growth rate factor, pr + //3, 77 = disturbance amplitude 6 = impingement half-angle A = wavelength ju = dynamic viscosity

Modeling and measurements of soot and species in a laminar diffusion flame

A model of laminar, soot-laden ethene diffusion flames has been developed and compared with measurements in nonsooting and sooting flames. Concentrations of stable gas-phase species were measured with mass spectrometry; laser-induced fluorescence was used to measure the OH concentrations. A system of elementary reactions was used to describe the gas-phase chemistry. The model incorporated a simple description of the growth of soot which assumed that acetylene was the only growth species. Soot formation was coupled with the flame chemistry via the loss of acetylene and OH on soot and the production of CO during soot oxidation. The model predicted most of the gas-phase species quite well, with the exception of OH. The loadings of soot in the flames were reproduced adequately, although less success was achieved in predicting the transition from nonsooting to sooting conditions. Details concerning the products of soot oxidation by OH were found to be important with regard to the flame chemistry. The inclusion of soot in the flame model had a significant impact on the predicted structure of the flame as seen in changes to the formation and destruction rates of OH on the fuel side of the flame. The rate of OH reaction with soot in the midregion of the flame was small compared with the rate of reaction of OH with CO. However, the two rates became comparable in the soot burnout zone.

AN EXPERIMENTAL ESTIMATION OF MEAN REACTION RATE AND FLAME STRUCTURE DURING COMBUSTION INSTABILITY IN A LEAN PREMIXED GAS TURBINE COMBUSTOR

The interactions between the local flame structure produced by a premixed swirl-stabilized injector with combustion instabilities were experimentally studied for a model gas turbine combustor operating at high pressure and temperature. The model gas turbine combustor studied utilizes a sudden-expansion dump combustor with a single swirler and bluff body for enhancing mixing rate and flame stabilization, respectively. Laser-based measurements were made for both stable and unstable operating conditions. The local flame front structure was visualized using planar laser-induced fluorescence (PLIF) from the OH⊙, and the global heat release zone was interpreted from flame emission measurements. For stable combustion conditions, the mean reaction rate estimated independently from both OH-PLIF and OH * chemiluminescence measurements showed good agreement, thereby indicating confidence in the use of OH-PLIF measurements for extracting the local mean reaction rate. For unstable combustion conditions, the flamefront characteristics, including flame surface density and mean reaction rate, were evaluated together with the information from the OH * chemiluminescence measurements to identify the boundary of the heat release region at discrete phases of the unstable flame. Analysis of the flame structures during combustion instability indicated significant variations during different phases of the instability. The heat release flow field, particularly in the recirculation regions appearing at the corner and inner face of the dump plane, varied substantially. Rayleigh index information indicated that the recirculation zones play an important part in driving the instability. In contrast, the high shear layer formed along the interface between reactants and hot products produced a region where the instability was damped due to a lowering of the heat release.

Benchmark Wall Heat Flux Data for a GO2/GH2 Single Element Combustor

Wall heat flux measurements in a 1.5 in. diameter circular cross-section rocket chamber for a uni-element shear coaxial injector element operating on gaseous oxygen (GOz)/gaseous hydrogen (GH,) propellants are presented. The wall heat flux measurements were made using arrays of Gardon type heat flux gauges and coaxial thermocouple instrumentation. Wall heat flux measurements were made for two cases. For the first case, GOZ/GHz oxidizer-rich (O/F=l65) and fuel-rich preburners (O/F=1.09) integrated with the main chamber were utilized to provide vitiated hot fuel and oxidizer to the study shear coaxial injector element. For the second case, the preburners were removed and ambient temperature gaseous oxygen/gaseous hydrogen propellants were supplied to the study injector. Experiments were conducted at four chamber pressures of 750, 600, 450 and 300psia for each case. The overall mixture ratio for the preburner case was 6.6, whereas for the ambient propellant case, the mixture ratio was 6.0. Total propellant flow was nominally 0.27-0.29 Ibm/s for the 750 psia case with flowrates scaled down linearly for lower chamber pressures. The axial heat flux profile results for both the preburner and ambient propellant cases show peak heat flux levels a t axial locations between 2.0 and 3.0 in. from the injector face. The maximum heat flux level was about two times greater for the preburner case. This is attributed to the higher injector fuel-to-oxidizer momentum flux ratio that promotes mixing and higher initial propellant temperature for the preburner case which results in a shorter reaction zone. The axial heat flux profiles were also scaled with respect to the chamber pressure to the power 0.8. The results at the four chamber pressures for both cases collapsed to a single profile indicating that at least to first approximation, the basic fluid dynamic structures in the flow field are pressure independent as long as the chamber/njector/nozzle geometry and injection velocities remain the same.

Suppression of soot formation in ethene laminar diffusion flames by chemical additives

The effects of methane, methanol, ethanol, 1-propanol, 2-propanol, carbon dioxide, and carbon disulfideaddition on soot formation in ethene laminar diffusion flames were examined. In this study, one-dimensional (ID) laser-induced incandescence (LII) and fluorescence measurements were used to determine soot volume fraction and relative soot precursor concentration, respectively, in the ethene and (ethene + additive) flames. Up to 60% reductions in soot volume fraction were found with the addition of 25% methanol to an ethene diffusion flame. More significant soot reductions were observed with the addition of carbon disulfide. It has been shown that a more than 50% reduction in soot volume fraction was achieved by adding 9.5% CS 2 . Experimental results strongly suggest suppression of soot formation by methanol and carbon disulfide to be mainly a chemical effect.

Experimental and analytical characterization of a shear coaxial combusting GO2/GH2 flowfield

Detailed measurements of the flow velocities and major species distributions downstream of a uni-element shear coaxial injector in a gaseous oxygen/hydrogen (GO2/GH2) rocket chamber are presented and compared with CFD predictions from a reacting Navier-Stokes solution. The velocity profiles are obtained by means of laser Doppler velocimetry, while the species (H2, O2, and H2O) are measured by Raman spectroscopy. The CFD predictions are in good agreement with the measurements, although they moderately overpredict the spreading rate of the reaction zone at the downstream location. Both experimental and numerical results indicate that the shear coaxial injector mixes slowly but that combustion is completed on a length scale representative of that tested in actual combustors. The experiments show high c* efficiency, and the computations show the flame combusts all the GO2, both suggesting complete combustion. Detailed CFD resolution shows that the base region behind the GO2 post tip acts as a flameholder for the GO2/GH2 flame. Overall, the results demonstrate that advanced diagnostics and CFD predictions can be used in concert as tools for gas/gas injector design. (Author)