Application of spontaneous Raman and Rayleigh scattering and 2D LIF for the characterization of a turbulent CH4/H2/N2 jet diffusion flame

4 /H₂/N₂ diffusion flame. Important aspects of the measuring technique, such as accuracy, cross talk between different Raman bands, and the correction procedure for background from laser-induced fluorescence are discussed. In addition, a 2D LIF and Rayleigh imaging system were used to study the structures of OH, CH, NO, and temperature distributions in the flame. A comparison between two different CH detection schemes is presented. A main goal of the investigations was a detailed and accurate characterization of the investigated flame as well as the study of experimental techniques. Joint pdfs of the temperature and major species concentrations were determined at nearly 100 measuring locations covering the complete flame. Parts of the results are presented in the paper in order to discuss effects of differential diffusion, flame extinction, and interaction between flow field and chemistry. The measured data sets which are available on the Internet are well suited for testing and validating mathematical flame models. Received: 8 July 1997/Revised version: 23 October 1997     

Cavity ring-down measurements of seeded NO in premixed atmospheric-pressure H2/air and CH4/air flames

Quantitative aspects of using cavity ring-down absorption spectroscopy near 226 nm for measurements of NO mole fractions in premixed atmospheric-pressure flames are discussed. Measurements in methane–air flames showed strong broadband absorption near 226 nm by hot CO₂ molecules, precluding using the cavity ring-down method in these flames at atmospheric pressure. In hydrogen–air flames, the broadband absorption at this wavelength was substantially lower. Absorption cross sections derived from non-seeded cavity ring-down spectra suggest that absorption by water is the major contribution to the background in these flames. The detectability limit for NO by cavity ring-down measurements in hydrogen–air flames using the current setup is estimated to be 10 ppm. Effects of the cold boundary layer on the measured NO mole fraction were accounted for by measuring the radial distributions of temperature and NO mole fraction using coherent anti-Stokes Raman scattering and laser-induced fluorescence (LIF), respectively. Measurements performed in seeded stoichiometric and lean hydrogen–air flames showed no reburning at temperatures above 1750 K, demonstrating the adequacy of using these flames for calibration of LIF measurements. At lower temperatures, the mole fraction of NO in the hot gases was up to 30% lower than that expected from the degree of seeding in the cold gases. PACS 42.62.Fi; 42.68.Ca; 82.33.Vx     

Simultaneous multiple-line Raman/Rayleigh/LIF measurements in combustion

It is demonstrated that multiple 1D Raman scattering, Rayleigh scattering, and laser-induced fluorescence (LIF) measurements can be performed simultaneously. This can be used for quasi-2D (or quasi-3D) single-shot measurements of multiple species and the temperature in turbulent reacting and non-reacting flows. The technique has the potential to yield more precise information than most competitive planar imaging approaches in combustion. For example, it can be used to overcome Raman/LIF interference problems in technical flames. This is achieved by a new optical set-up that makes use of an imaging spectrograph combined with fiber optics. Received: 13 October 1999 / Revised version: 26 November 1999 / Published online: 27 January 2000     

Thermal and chemical effects of differential diffusion in turbulent non-premixed H2 flames

As a renewable fuel, hydrogen (H₂) may play an increasingly important role in the development and control of piston and gas turbine engines to achieve zero carbon emissions. Predictive modeling of H₂-fueled combustion processes requires a clear understanding of differential diffusion (DD) due to the high diffusivity of H₂. On the assumption that turbulent mixing is a far more dominant process than molecular mixing, DD effects are typically neglected in turbulent combustion simulations to reduce modeling complications. While this assumption is reasonable for hydrocarbon fuels, it is less valid for H₂ combustion, where DD is significant. In this work, two three-dimensional direct numerical simulations of temporally evolving turbulent H₂ jet flames with and without considering DD are performed and compared with laminar flamelet solutions to assess DD effects under turbulent conditions. The emphasis is placed on assessing the suitability of classical mixture fraction Z and Bilger mixture fraction Z_Bilger as conditioning variables for non-premixed turbulent combustion modeling through analyzing DD effects on flame structure, chemical reactions, and tangential diffusion (TD). Furthermore, the persistence of DD effects under turbulent conditions and the suitability of a conventional DD parameter are investigated by comparing the turbulent flames to laminar flamelet solutions. It is found that conditioning the thermochemical state on Z_Bilger helps to capture DD effects and mitigate the relative contribution of TD, which gives Z_Bilger advantages over Z when employing flamelet modeling. Due to close coupling between DD and local chemical reactions, DD can affect the turbulent/laminar flames in the form of thermal effects due to the change in flame temperature, chemical effects due to the change in chemical reactions, and transport effects due to multiple species with varying diffusivities that could result in the difference between Z and Z_Bilger. While the transport effects are suppressed, significant chemical and thermal effects of DD still persist under turbulent conditions, which indicates that the DD parameter is probably unsuitable for comprehensively characterizing and assessing DD effects on the structure of turbulent non-premixed flames.     

Linewise Raman-Stokes/anti-Stokes temperature measurements in flames using an unintensified charge-coupled device

One-dimensional, spatially resolved (linewise), temporally averaged Stokes/anti-Stokes temperature measurements with 200 m spatial resolution and 6% precision uncertainty are demonstrated in an opposed-jet-diffusion flame of air vs N₂-diluted H₂ using the Raman-Stokes/anti-Stokes technique. The Stokes/anti-Stokes temperature profile is compared to temperature measurements obtained using all major species'Q-branches and differences between the two sets of results are discussed. In addition, single-pulse, linewise measurements with 710 m spatial resolution and 19% precision uncertainty are also demonstrated. To the authors' knowledge this is the first report of linewise Stokes/anti-Stokes temperature measurements in flames. Simultaneous measurements of the N₂ Stokes and anti-Stokes vibrationalQ-branches are obtained by using an unintensified Charge-Coupled Device (CCD) that is gated by a ferro-electric liquid crystal light valve and/or a low duty cycle optical chopper to suppress detected flame luminosity. This detection system provides increased detection efficiency, dynamic range, and imaged spectral range over an intensified CCD, allowing spatially resolved Stokes and anti-Stokes signals to be imaged onto one detector.     

Prediction of NO in turbulent diffusion flames using Eulerian particle flamelet model

The combustion characteristics for the turbulent diffusion flames using the unsteady flamelet concept have been numerically investigated. The Favre-averaged Navier–Stokes equations are solved by a finite volume method of SIMPLE type that incorporates the laminar flamelet concept with a modified k ? ε turbulence model. The NO formation is estimated by solving the Eulerian particle transport equations in a postprocessing mode. Two test problems are considered: CH₄/H₂/N₂ jet flame and CH₄/H₂ stabilised bluff body flame. The temperature and species profiles are well captured by the flamelet model. Two different chemical mechanisms (GRI 2.11 and 3.0) give nearly identical results for temperature and species except NO. The GRI 3.0 gives significantly higher NO levels compared to the GRI 2.11. This is mainly attributed to the difference in NO formation by the prompt mechanism. The NO formation is sensitive to the number of flamelet particles. The NO levels for two test flames do not change when the flamelet particle number exceeds six.     

Investigations in the TECFLAM swirling diffusion flame: Laser Raman measurements and CFD calculations

A standard burner for confined swirling natural gas flames is presented which was developed within the German TECFLAM cooperation. The aims of the TECFLAM research program are the establishment of an extensive experimental database from selective flames and the validation and improvement of mathematical combustion models. In this paper, results from joint PDF measurements of temperature, mixture fraction, and major species concentrations in a turbulent diffusion flame with 150 kW thermal load, equivalence ratio 0.833, and swirl number 0.9 are presented. Major aspects of the investigation are the general quantitative characterization of the flame and the study of the thermochemical state, e.g. effects of turbulence–chemistry interactions. Scatterplots of temperature, CH₄, and CO mole fractions as well as mean mixture fraction and temperature fields are presented and discussed. Furthermore, CFD calculations have been performed using the code Fluent 5 as an example of a commercially available code that is frequently used for technical applications. The comparison between the calculated and measured results reveals some significant deviations which are discussed with respect to the applicability of this code to swirling turbulent flames. Received: 19 April 2000 / Revised version: 15 June 2000 / Published online: 5 October 2000     

Uncertainty assessment of species measurements in acetone counterflow diffusion flames

To quantitatively understand the uncertainty of intrusive species sampling measurements using a microprobe, velocity and speciation profiles of acetone counterflow diffusion flames have been experimentally investigated with cross validations using non-intrusive particle image velocimetry (PIV) and laser induced fluorescence (LIF) measurements. It is shown that the separation distance between the fuel and oxidizer nozzles needs to be sufficiently large to achieve uniform radial velocity profiles at the nozzle exit and accurate measurements of fuel concentration distributions in flames. The impacts of the diffusion flame location relative to the stagnation plane and the diffusion flame thickness on quantitative species sampling are investigated by varying the fuel to oxygen ratio as well as nitrogen and helium as fuel diluents. The results show that the diffusion flame needs to be located on the fuel side far from the stagnation plane in order to obtain reliable speciation measurements of fuel oxidation-related species. For helium dilution in the fuel side, a significant deviation from the model prediction is found due to the excessively fast diffusion velocity of helium. The impact of the intrusive probe on the flow field and the structure of the counterflow diffusion flame are identified by acetone and OH LIF measurements. The uncertainty in the speciation measurement associated with flow perturbations by the probe is quantified and found to be comparable to the outer diameter of the probe, ±0.3 mm. A simple Reynolds number analysis shows that the flow near the probe is just on the outskirts of the Stokes regime. Finally, the structure of the acetone diffusion flame is measured quantitatively with species measurements of ethane, ethylene, and acetylene. The comparison between predictions and measurements indicate that the current C₂ kinetic mechanism needs to be improved for quantitative prediction of the acetone flame structures.     

Determination of the Li/Nb ratio in lithium niobate by means of birefringence and Raman measurements

Two optical methods for the determination of the Li/Nb ratio in lithium niobate (LiNbO₃) are discussed. Data for the optical birefringence method are presented, they cover the entire composition range in which LiNbO₃ can be fabricated and a wide spectral region. The line-widths of lattice modes have been measured by means of Raman scattering as a function of the Li/Nb ratio for single crystals with polished and rough surfaces.     

Lewis number effects on laminar and turbulent expanding flames of NH3/H2/air mixtures at elevated pressures

This study clarifies the effects of Lewis number (Le) on laminar and turbulent expanding flames of NH₃/H₂/air mixtures. The laminar burning velocity (S_L) and turbulent burning velocity (S_T) were measured using a medium-scale, fan-stirred combustion chamber with ammonia/hydrogen molar ratio (NH₃/H₂) of 50/50 and 80/20 under the maximum pressures of 5 atm. The lean laminar flame with NH₃/H₂ = 50/50 is significantly accelerated by the diffusional–thermal instability, which dominated the trend of S_T,c=0.1 with the equivalence ratio (ϕ). The lean normalized turbulent burning velocity (S_T/S_L) increases with the decrease of hydrogen content due to the weakening effects of S_L. However, the S_T/S_L reaches peak with hydrogen volumetric content less than 20% due to effects made by diffusional–thermal instability than S_L did. The turbulent flame of NH₃/H₂/air mixtures is characterized by self-similar acceleration propagation, and propagation with Le < 1 is faster. A modified correlation considering the effects of Le was proposed, as (d<r>/dt)/σS_L = 0.118(Re_T,flameLe⁻²)^0.57, which was able to predict not only the self-similar propagation of NH₃/H₂/air but also the previous syngas/air flames. The Kobayashi correlations modified by three kinds of Le power exponents were used to clarify the effects of Le by comparing their fitting parameters and predictive powers on experimental data and literature data. Similar pre-factors, power exponents and the goodness of fit (R²) were obtained with Le ranging from 0.58 to 1.62, which suggested that the determination of Le power exponent had no significant effect on the prediction accuracy of the S_T/S_L trend with data of Le near unity. This might be attributed to the fact that the variation ranges of the dimensionless number that characterizes the experimental conditions is much larger than that of the Le.     

Stochastic simulation of transport and chemical kinetics in turbulent CO/H2/N2 flames

The one-dimensional turbulence (ODT) model is applied to the study of turbulent jet CO/H₂/N₂ flames. The ODT model retains the full range of length scales and no assumption of scale separation is required. In the present case, the ODT model describes the evolution of the entire flowfield; the ability of ODT to model the flowfield evolution is discussed. Predictions of the conditional means and fluctuations of temperature and the species mass fractions are compared with measurements and the level of closure required to address certain phenomena is discussed. The ODT model is shown to be useful for studying extinction/re-ignition phenomena and differential diffusion. The streamwise evolution in multidimensional flows is affected by dilatation in a manner not captured by the ODT model, and this results in discrepancies between the ODT mixing rates and the mixing rates evident in the measurements.     

Laser-induced fluorescence measurements and modeling of absolute CH concentrations in strained laminar methane/air diffusion flames

We have applied linear laser-induced fluorescence to obtain spatially resolved profiles of CH radicals in laminar methane/air and methane/nitric oxide/air counterflow diffusion flames at atmospheric pressure. Excitation and detection of transitions in the A–X band and calibrating the optical detection efficiency via Rayleigh scattering allowed the determination of absolute radical concentrations. Flames at strain rates from 59 to 269 s⁻¹ were studied to characterize the strain rate dependence of the CH concentration. The work shows that CH concentrations increase with increasing strain rate. Comparisons have been made with predicted CH levels obtained using two different chemical kinetic mechanisms (Lindstedt et al. and GRI-Mech. 3.0). Computed concentrations are shown to be in good agreement with experimental data. It was furthermore found that the addition of up to 600 ppm NO to the fuel did not have a measurable effect on the CH radical concentration. This is also in agreement with predictions from both mechanisms. The current work has shown that measurements of absolute CH radical concentrations are possible in non-premixed flames without the need for spatial temperature or quenching corrections.     

Raman and point contact current-voltage characterization of laser-induced diffusion in GaAs

Raman scattering and point contact current-voltage (PCIV) measurements were used as characterization tools of tin-diffused GaAs layers. Diffusion was induced by irradiating GaAs substrates covered with thin tin layers single pulses of a ruby laser. Samples processed with the lowest energies show strong damage and incomplete electrical activation as deduced from Raman and PCIV measurements, respectively. Raman microprobe in depth analysis and PCIV profiles also suggest the presence of a damaged region with incomplete electrical activation at the boundary between the molten layer and the solid substrate.     

Soot and NO formation in counterflow ethylene/oxygen/nitrogen diffusion flames

Formation of soot and NO in counterflow ethylene/oxygen/nitrogen diffusion flames was numerically investigated. Detailed chemistry and complex thermal and transport properties were used. A simplified two-equation soot model was adopted. The results indicate that NO emission has negligible influence on soot formation. However, soot formation affects the emission of NO through the radiation induced thermal effect and the reaction induced chemical effect. When the oxygen index of the oxidant stream is lower, the relative influence of chemical reaction caused by soot on NO emission is more important, while the relative influence of the radiation induced thermal effect becomes more important for the flame with a higher oxygen index in the oxidant stream.     

Dilution effects analysis of opposed-jet H2/CO syngas diffusion flames

This paper reported the analysis of dilution effects on the opposed-jet H₂/CO syngas diffusion flames. A computational model, OPPDIF coupled with narrowband radiation calculation, was used to study one-dimensional counterflow syngas diffusion flames with fuel side dilution from CO₂, H₂O and N₂. To distinguish the contributing effects from inert, thermal/diffusion, chemical, and radiation effects, five artificial and chemically inert species XH₂, XCO, XCO₂, XH₂O and XN₂ with the same physical properties as their counterparts were assumed. By comparing the realistic and hypothetical flames, the individual dilution effects on the syngas flames were revealed. Results show, for equal-molar syngas (H₂/CO = 1) at strain rate of 10 s^?1, the maximum flame temperature decreases the most by CO₂ dilution, followed by H₂O and N₂. The inert effect, which reduces the chemical reaction rates by behaving as the inert part of mixtures, drops flame temperature the most. The thermal/diffusion effect of N₂ and the chemical effect of H₂O actually contribute the increase of flame temperature. However, the chemical effect of CO₂ and the radiation effect always decreases flame temperature. For flame extinction by adding diluents, CO₂ dilution favours flame extinction from all contributing effects, while thermal/diffusion effects of H₂O and N₂ extend the flammability. Therefore, extinction dilution percentage is the least for CO₂. The dilution effects on chemical kinetics are also examined. Due to the inert effect, the reaction rate of R84 (OH+H₂ = H+H₂O) is decreasing greatly with increasing dilution percentage while R99 (CO+OH→CO₂+H) is less affected. When the diluents participate chemically, reaction R99 is promoted and R84 is inhibited with H₂O addition, but the trend reverses with CO₂ dilution. Besides, the main chain-branching reaction of R38 (H+O₂→O+OH) is enhanced by the chemical effect of H₂O dilution, but suppressed by CO₂ dilution. Relatively, the influences of thermal/diffusion and radiation effects on the reaction kinetics are then small.     

Imaging of flame behavior in flickering methane/air diffusion flames

During this study, flow visualization through the use of imaging provided visual data of the events that occurred as the flame oscillated. Imaging was performed in two different ways: 1) the first method was phase-locked imaging to capture a detailed history by simply advancing the phase angle during each image capture, 2) the second method involved high-speed imaging to gather visual image data of a natural or forced oscillating flame. For visualization, two items were considered. The first one was the shape of the flame envelope as it evolved during one oscillation cycle. From the data gathered, it was confirmed that the flame stretched in the vertical direction before quenching in the region near its center. The second consideration was imaging of the oxidizer (air) in the region immediately outside the flame. This was done by imaging the laser light reflected from particles seeded into the flow, which revealed formation of vortical structures in those regions where quenching had occurred. It was noted that quenching took place primarily by the entrainment of fresh non-reacting air into the flame. The quenching process was in turn responsible for the oscillatory behavior.     

Multiplex CARS measurements in supersonic H2/air combustion

2 and O₂ multiplex coherent anti-stokes Raman spectroscopy (CARS) employing a single dye laser has been explored to simultaneously determine the temperature and concentrations of H₂ and O₂ in a hydrogen-fueled supersonic combustor. Systematic calibrations were performed through a well-characterized H₂/air premixed flat-flame burner. In particular, temperature measurement was accomplished using the intensity ratio of the H₂ S(5) and S(6) rotational lines, whereas extraction of the H₂ and O₂ concentrations was obtained from the H₂ S(6) and O₂ Q-branch, respectively. Details of the calibration procedure and data reduction are discussed. Quantification of the supersonic mixing and combustion characteristics applying the present technique has been demonstrated to be feasible. The associated detection limits as well as possible improvements are also identified. Received: 1 July 1997/Revised version: 29 September 1998