Detailed numerical modeling of PAH formation and growth in non-premixed ethylene and ethane flames

A chemical kinetic mechanism for C₁ and C₂ fuel combustion and PAH growth, previously validated for laminar premixed combustion, has now been modified and applied to opposed flow diffusion flames. Some modifications and extensions have been made to the reaction scheme to take into account recent kinetic investigations, and to reduce the stiffness of the reaction model. Updates have been made to the cyclopentadienyl reactions, indene formation reactions, and aromatic oxidation and decomposition reactions. Reverse reaction rate parameters have been revised to account for numerical stiffness. Opposed flow diffusion flame simulation data for ethylene and ethane flames with the present mechanism are compared to data computed using two other mechanisms from the literature and to experimental data. Whereas the fuel oxidation chemistry in all three mechanisms are essentially the same, the PAH growth pathways vary considerably. The current mechanism considers a detailed set of PAH growth routes, and includes hydrogen atom migration, possible free radical addition schemes, methyl substitution/acetylene addition pathways, cyclopentadienyl moiety in aromatic ring formation, and numerous reactions between aromatic radicals and molecules. It is shown that while bulk flame properties and major species profiles are the same for the three mechanisms, the enhanced PAH growth routes in the present mechanism are necessary to numerically predict the correct order of magnitude of PAHs that were measured in the experimental studies. In particular, predicting concentrations of naphthalene, phenanthrene, and pyrene, to within the correct order of magnitude with the present mechanism show a significant improvement over predictions obtained using mechanisms in the literature. Sensitivity and production rate analyses show that this improvement is attributable to the enhanced PAH growth pathways and updated reaction rates in the present mechanism. The overreaching goal of this research is to generate and fully validate a detailed chemical kinetic mechanism, with as few fitted rates as possible, that can be applied to premixed or diffusion systems, and used with any type of soot model. To that end, in recently published works, the present mechanism has been used to simulate premixed flames, while coupled to a method of moments to determine soot formation, and to simulate diffusion flames, while coupled to a sectional representation for soot formation. The present work extends the validation of the mechanism by applying it to counterflow diffusion flames, for which measurements of large PAH molecules are uniquely available. The validation of PAH growth predictions are of key interest to soot modeling studies as soot inception from PAH combination and PAH condensation are often major constituents of soot production.     

Size and charge of soot particles in rich premixed ethylene flames

A nano differential mobility analyzer (DMA) is used to measure both the size and electrical charge distributions of soot particles generated during rich premixed combustion. The size distributions are bimodal. One mode peaks at diameters below the 3 nm lower limit of the nano DMA and falls off nearly exponentially with increasing particle diameter. The intensity of this mode persists with increasing height above the burner suggesting that it represents the continued formation of new particles. The second mode is lognormal in shape. Its intensity decreases and the mean diameter increases with increasing height above the burner due to coagulation and surface growth as the particles rise in the flame. The DMA measurements show that a substantial fraction of the soot particles are electrically charged in the flame, predominantly with a single charge per particle and with essentially equal numbers of positive and negative particles. These charged particles belong solely to the upper mode, whereas the lower mode remains charge neutral, suggesting that ions do not act as soot nuclei. Following soot inception, the fraction of charged particles quickly increases with height above the burner and stabilizes at ∼30% of the upper mode for each polarity.     

Soot predictions in premixed and non-premixed laminar flames using a sectional approach for PAHs and soot

A new soot model is presented, which has been developed for CFD applications, combining accuracy and efficiency. While the chemical reactions of small gas phase species are captured by a detailed chemical kinetic mechanism, polycyclic aromatic hydrocarbons (PAHs) and soot particles are represented by sectional approaches. The latter account for important mechanisms such as the formation of sections, their oxidation, the condensation of acetylene, and the collisions between sections. The model has been designed to predict soot for a variety of fuels with good accuracy at relatively low computational cost. Universal model parameters are applied, which require no tuning in dependence of test case or fuel. Soot predictions of ethylene, propylene, kerosene surrogate, and toluene flames are presented, which show good agreement with the experimental data. Furthermore, the importance of the correct choice for thermodynamic data of PAHs and soot is highlighted and the impact of heat radiation is discussed.     

Chaotic map models of soot fluctuations in turbulent diffusion flames

In this paper, we introduce a methodology to characterize time-dependent soot volume fraction fluctuations in turbulent diffusion flames via chaotic maps. The approach is based on the hypothesis that fluctuations of properties in turbulent flames are deterministic in nature, rather than statistical. Our objective is to develop models of these fluctuations to be used in comprehensive algorithms to study the nature of turbulent flames and the interaction of turbulence with radiation. To this end we measured the time series of soot scattering coefficient in an ethylene diffusion flame from light scattering experiments and fit these data to linear combinations of chaotic maps of the unit interval. Both time series and power spectra can be modeled with reasonable accuracy in this way. © 1998 Elsevier Science Ltd. All rights reserved.     

A soot particle surface reactivity model applied to a wide range of laminar ethylene/air flames

The effect of soot surface reactivity, in terms of the evolution of sites on the soot particles’ surface available for reaction with gas phase species, is investigated via modeling numerous ethylene/air flames, using a detailed combustion and sectional soot particle dynamics model. A new definition of a particles’ age is introduced. A methodology has been developed to study soot particle surface reactivity. Subsequently, it is investigated if the surface reactivity can be correlated with the particle age. An exponential function giving a smooth transition of surface activity with particle age is employed to model a variety of ethylene/air flames, which differ in fuel stream dilution levels, fuel stream premixing, and burner configurations. Excellent agreement with measured soot volume fractions of a variety of flames, burners, and datasets could be obtained with this approach. The newly developed function based on particle age eliminates the need to fit soot surface growth parameters to each experimental condition. Finally, the applicability and limitation of the new surface reactivity function for use in detailed soot formation models is discussed.     

Characteristics of non-premixed oxygen-enhanced combustion: II. Flame structure effects on soot precursor kinetics resulting in soot-free flames

A detailed computational study was performed to understand the effects of the flame structure on the formation and destruction of soot precursors during ethylene combustion. Using the USC Mech Version II mechanism the contributions of different pathways to the formation of benzene and phenyl were determined in a wide domain of Z_st values via a reverse-pathway analysis. It was shown that for conventional ethylene-air flames two sequential reversible reactions play primary roles in the propargyl (C₃H₃) chemistry, namely

(1)     

Implementation of two-equation soot flamelet models for laminar diffusion flames

The two-equation soot model proposed by Leung et al. [K.M. Leung, R.P. Lindstedt, W.P. Jones, Combust. Flame 87 (1991) 289-305] has been derived in the mixture fraction space. The model has been implemented using both Interactive and Non-Interactive flamelet strategies. An Extended Enthalpy Defect Flamelet Model (E-EDFM) which uses a flamelet library obtained neglecting the soot formation is proposed as a Non-Interactive method. The Lagrangian Flamelet Model (LFM) is used to represent the Interactive models. This model uses direct values of soot mass fraction from flamelet calculations. An Extended version (E-LFM) of this model is also suggested in which soot mass fraction reaction rates are used from flamelet calculations. Results presented in this work show that the E-EDFM predict acceptable results. However, it overpredicts the soot volume fraction due to the inability of this model to couple the soot and gas-phase mechanisms. It has been demonstrated that the LFM is not able to predict accurately the soot volume fraction. On the other hand, the extended version proposed here has been shown to be very accurate. The different flamelet mathematical formulations have been tested and compared using well verified reference calculations obtained solving the set of the Full Transport Equations (FTE) in the physical space.     

Conditional moment closure prediction of soot formation in turbulent, nonpremixed ethylene flames

Results obtained from incorporating a semiempirical soot model into a first-order conditional moment closure (CMC) approach to modeling turbulent nonpremixed flames of ethylene and air are presented. Soot formation is determined via the solution of two transport equations for soot mass fraction and particle number density, with acetylene and benzene employed as the incipient species responsible for soot nucleation, and the concentrations of these species calculated using a detailed gas-phase kinetic scheme involving 463 reactions and 70 species. The study focuses on the influence of differential diffusion of soot particles on soot volume fraction predictions. The results of calculations are compared with experimental data for three sooting ethylene flames and, in general, predictions of mixing and temperature fields within the three flames show good agreement with data. Soot volume fraction predictions are found to be in significantly better accord with data when differential diffusion is accounted for in the CMC-based soot model, supporting the importance of such effects in sooting flames, as previously noted by Kronenburg et al. in relation to methane combustion. Overall, the study demonstrates that the CMC-based soot model, when used in conjunction with a model of differential diffusion effects, is capable of accurately predicting soot formation in turbulent nonpremixed ethylene-air flames.     

Effect of a uniform electric field on soot in laminar premixed ethylene/air flames

The effect of a nominally uniform electric field on the initially uniform distribution of soot has been assessed for laminar premixed ethylene/air flames from a McKenna burner. An electrophoretic influence on charged soot particles was measured through changes to the deposition rate of soot on the McKenna plug, using laser extinction (LE). Soot volume fraction was measured in situ using laser-induced incandescence (LII). Particle size and morphologies were assessed through ex situ transmission electron microscopy (TEM) using thermophoretic sampling particle diagnostics (TSPD). The results show that the majority of these soot particles are positively charged. The presence of a negatively charged plug was found to decrease the particle residence times in the flame and to influence the formation and oxidation progress. A positively charged plug has the opposite effect. The effect on soot volume fraction, particles size and morphology with electric field strength is also reported. Flame stability was also found to be affected by the presence of the electric field, with the balance of the electrophoretic force and drag force controlling the transition to unstable flame flicker. The presence of charged species generated by the flame was found to reduce the dielectric field strength to one seventh that of air.     

Formation,growth, and transport of soot in a three-dimensional turbulent non-premixed jet flame

The formation, growth, and transport of soot is investigated via large scale numerical simulation in a three-dimensional turbulent non-premixed n-heptane/air jet flame at a jet Reynolds number of 15,000. For the first time, a detailed chemical mechanism, which includes the soot precursor naphthalene and a high-order method of moments are employed in a three-dimensional simulation of a turbulent sooting flame. The results are used to discuss the interaction of turbulence, chemistry, and the formation of soot. Compared to temperature and other species controlled by oxidation chemistry, naphthalene is found to be affected more significantly by the scalar dissipation rate. While the mixture fraction and temperature fields show fairly smooth spatial and temporal variations, the sensitivity of naphthalene to turbulent mixing causes large inhomogeneities in the precursor fields, which in turn generate even stronger intermittency in the soot fields. A strong correlation is apparent between soot number density and the concentration of naphthalene. On the contrary, while soot mass fraction is usually large where naphthalene is present, pockets of fluid with large soot mass are also frequent in regions with very low naphthalene mass fraction values. From the analysis of Lagrangian statistics, it is shown that soot nucleates and grows mainly in a layer close to the flame and spreads on the rich side of the flame due to the fluctuating mixing field, resulting in more than half of the total soot mass being located at mixture fractions larger than 0.6. Only a small fraction of soot is transported towards the flame and is completely oxidized in the vicinity of the stoichiometric surface. These results show the leading order effects of turbulent mixing in controlling the dynamics of soot in turbulent flames. Finally, given the difficulties in obtaining quantitative data in experiments of turbulent sooting flames, this simulation provides valuable data to guide the development of models for Large Eddy Simulation and Reynolds Average Navier Stokes approaches.     

Different chemical effect of hydrogen addition on soot formation in laminar coflow methane and ethylene diffusion flames

Previous studies showed that adding hydrogen (H₂) can have an opposite chemical effect on soot formation: its chemical effect enhances and suppresses soot formation in methane (CH₄) and ethylene (C₂H₄) diffusion flames, respectively. Such opposite chemical effect of H₂ (CE-H₂) remains unresolved. The different CE-H₂ is studied numerically in the two laminar coflow diffusion flames. A detailed chemical mechanism with the addition of a chemically inert virtual species FH₂ is used to model the gas-phase combustion chemistry in this study. Particularly, a reaction pathway analysis was performed based on the numerical results to gain insights into how H₂ addition to fuel affects the pathways leading to the formation of benzene (A₁) in CH₄ and C₂H₄ flames. The numerical results show that the CE-H₂ in CH₄ diffusion flame to prompt soot formation is ascribed that the higher mole fraction of H atom promotes the formation of A₁ and Acetylene (C₂H₂) and leads to higher nucleation rate and eventually higher soot surface growth rate. In contrast, adding H₂ to C₂H₄ diffusion flames decreases soot nucleation and surface growth rate. The lower soot nucleation rate is due to the lower mole fractions of pyrene (A₄), while the lower soot surface growth rate is due to the lower mole fractions of H atom and C₂H₂, higher mole fraction of H₂ and lower soot nucleation rate. Furthermore, the CE-H₂ in C₂H₄ diffusion flames promotes the formation of A₁, but suppresses the formation of A₄.     

Hydrogen-enriched non-premixed jet flames: Analysis of the flame surface,flame normal,flame index and Wobbe index

A non-premixed impinging jet flame is studied using three-dimensional direct numerical simulation with detailed chemical kinetics in order to investigate the influence of fuel variability on flame surface, flame normal, flame index and Wobbe index for hydrogen-enriched combustion. Analyses indicate that the fuel composition greatly influences the H₂/CO syngas combustion, not only on the important local stoichiometric iso-mixture fraction surface distribution but also on the vortical structures in the flow field. As a result of CO addition to hydrogen-rich combustion, changes of the reaction zone in the flammable layer, shift of peak flame surface density distribution, shift of non-premixed regions, formation of widely populated scalar dissipation distribution rate with respect to tangential strain and reduction of global heat release are all found to appear. In particular, the CO addition induces a micromixing process which appears to be an important factor for the modelling investigation of turbulence/chemistry interaction especially for combustion modelling of H₂-rich syngas fuels.     

Extinction and near-extinction instability of non-premixed tubular flames

Tubular non-premixed flames are formed by an opposed tubular burner, a new tool to study the effects of curvature on extinction and flame instability of non-premixed flames. Extinction of the opposed tubular flames generated by burning diluted H₂, CH₄ or C₃H₈ with air is investigated for both concave and convex curvature. To examine the effects of curvature on extinction, the critical fuel dilution ratios at extinction are measured at various stretch rates, initial mixture strengths and flame curvature for fuels diluted in N₂, He, Ar or CO₂. In addition, the onset conditions of the cellular instability are mapped as a function of stretch rates, initial mixture strengths, and flame curvature. For fuel mixtures with Lewis numbers much less than unity, such as H₂/N₂, concave flame curvature towards the fuel suppresses cellular instabilities.     

Effect of free stream turbulence on NOx and soot formation in turbulent diffusion CH4-air flames

A two-dimensional axisymmetric RANS numerical model was solved to investigate the effect of increasing the turbulence intensity of the air stream on the NOx and soot formation in turbulent methane diffusion flames. The turbulence–combustion interaction in the flame field was modelled in a k − ε/EDM framework, while the NO and soot concentrations were predicted through implementing the extended Zildovich mechanism and two transport equations model, respectively. The predicted spatial temperature gradients showed acceptable agreement with published experimental measurements. It was found that the increase of free stream turbulence intensity of the air supply results in a significant reduction in the NO formation of the flame. Such phenomenon is discussed by depicting the spatial distribution of the NO concentration in the flame. An observable reduction of the soot formation was also found to be associated with the increase of inlet turbulence intensity of air stream.     

Measurements of soot, OH, and PAH concentrations in turbulent ethylene/air jet flames

This paper presents results from an investigation of soot formation in turbulent, non-premixed, C₂H₄/air jet flames. Tests were conducted using a H₂-piloted burner with fuel issuing from a 2.18 mm i.d. tube into quiescent ambient air. A range of test conditions was studied using the initial jet velocity (16.2-94.1 m/s) as a parameter. Fuel-jet Reynolds numbers ranged from 4000 to 23,200. Planar laser-induced incandescence (LII) was employed to determine soot volume fractions, and laser-induced fluorescence (LIF) was used to measure relative hydroxyl radical (OH) concentrations and polycyclic aromatic hydrocarbons (PAHs) concentrations. Extensive information on the structure of the soot and OH fields was obtained from two-dimensional imaging experiments. Quantitative measurements were obtained by employing the LII and LIF techniques independently. Imaging results for soot, OH, and PAH show the existence of three soot formation/oxidation regions: a rapid soot growth region, in which OH and soot particles lie in distinctly different radial locations; a mixing-dominated region controlled by large-scale motion; and a soot-oxidation region in which the OH and soot fields overlap spatially, resulting in the rapid oxidation of soot particles. Detailed quantitative analyzes of soot volume fractions and OH and soot zone thicknesses were performed along with the temperature measurement using the N₂-CARS system. Measurements of OH and soot zone thicknesses show that the soot zone thickness increases linearly with axial distance in the soot formation region, whereas the OH zone thickness is nearly constant in this region. The OH zone thickness then rapidly increases with downstream distance and approximately doubles in the soot-oxidation region. Probability density functions also were obtained for soot volume fractions and OH concentrations. These probability density functions clearly define the spatial relationships among the OH, PAH concentrations, the soot formation, and oxidation processes.     

Field measurements of soot volume fractions in laminar partially premixed coflow ethylene/air flames

The conditions under which soot is formed vary widely and depend upon several factors, including pressure, temperature, fuel type, combustor geometry, and extent of premixing. Although it is known that partially premixed flames (PPFs) can become either more or less sooting than their nonpremixed or premixed counterparts, the impact of partial premixing on soot formation across a large equivalence ratio and flow range is still inadequately understood. Comprehensive experimental data are relatively sparse for this important configuration. Herein, we report on soot formation in various ethylene/air PPFs utilizing full-field light extinction. The dimensionless extinction coefficient Kext is an important calibrated constant for the determination of the soot volume fraction for this measurement technique. We find that a value of Kext=7.1 provides results that are in good agreement with benchmark literature data for a nonpremixed flame. We examined the soot microstructures for two flames established at ?=∞ (i.e., nonpremixed) and 5. In both cases, the primary particles were found to be nearly spherical. In case of the nonpremixed flame the average primary soot particle diameter was ∼35 nm, but for the ?=5 flame it was ∼20 nm. However, the parameter responsible for the value of Kext is the average aggregate size and not that of the primary particles. The aggregate sizes are similar for the two flames. We consider this as verification of a constant Kext value over the entire equivalence ratio range. The addition of air to the fuel stream produces an initial increase in the flame height. Further air addition gradually decreases the flame height, which is followed by a more rapid decrease with larger premixing. Likewise, the peak soot concentration first increases with small amounts of air addition (or partial premixing of the fuel stream) and reaches a maximum value at ?∼24. With further air addition, as ? decreases below a value of 20, the soot volume fraction considerably decreases.     

Evolution of soot size distribution in premixed ethylene/air and ethylene/benzene/air flames: Experimental and modeling study

The effect of benzene concentration in the initial fuel on the evolution of soot size distribution in ethylene/air and ethylene/benzene/air flat flames was characterized by experimental measurements and model predictions of size and number concentration within the flames. Experimentally, a scanning mobility particle sizer was used to allow spatially resolved and online measurements of particle concentration and sizes in the nanometer-size range. The model couples a detailed kinetic scheme with a discrete-sectional approach to follow the transition from gas-phase to nascent particles and their coagulation to larger soot particles. The evolution of soot size distribution (experimental and modeled) in pure ethylene and ethylene flames doped with benzene showed a typical nucleation-sized (since particles do not actually nucleate in the classical sense particle inception is often used in place of nucleation) mode close to the burner surface, and a bimodal behavior at greater height above burner (HAB). However, major features were distinguished between the data sets. The growth of nucleation and agglomeration-sized particles was faster for ethylene/benzene/air flames, evidenced by the earlier presence of bimodality in these flames. The most significant changes in size distribution were attributed to an increase in benzene concentration in the initial fuel. However, these changes were more evident for high temperature flames. In agreement with the experimental data, the model also predicted the decrease of nucleation-sized particles in the postflame region for ethylene flames doped with benzene. This behavior was associated with the decrease of soot precursors after the main oxidation zone of the flames.     

A comparison of soot size and charge distributions from ethane, ethylene, acetylene, and benzene/ethylene premixed flames

The size and electrical charge distributions of soot particles generated in rich premixed flames are examined using a nano-differential mobility analyzer (nDMA). The aim is to investigate how these distributions vary with the choice of fuel, diluent, and flame gas velocity. The measured soot size distributions are typically bimodal. The dynamics of the upper size mode is qualitatively independent of the fuel. At increasing heights above the burner this mode increases in diameter and volume fraction, but decreases in number concentration, as expected from surface growth and coagulation. At about 6 mm above the burner a small fraction of the particles, <10%, acquire a bipolar charge. The charge is associated with the upper mode, where the fractions of positive and negative particles evolve to a Boltzmann distribution, whereas the lower mode remains electrically neutral. The existence of the lower mode depends sensitively on the choice of fuel, the flame gas velocity, and the diluent. Comparison to model calculations of flame structure reveals that as each of these are varied, the lower mode exhibits an inverse correlation with temperature, hydrogen atom, and pyrene concentrations.     

Effect of different turbulence models on combustion and emission characteristics of hydrogen/air flames

This paper aims to present modeling results of hydrogen/air combustion in a micro-cylindrical combustor. Modeling studies were carried out with different turbulence models to evaluate performance of these models in micro combustion simulations by using a commercially available computational fluid dynamics code. Turbulence models implemented in this study are Standard k-ε, Renormalization Group k-ε, Realizable k-ε, and Reynolds Stress Transport. A three-dimensional micro combustor model was built to investigate impact of various turbulence models on combustion and emission behavior of studied hydrogen/air flames. Performance evaluation of these models was executed by examining combustor outer wall temperature distribution; combustor centerline temperature, velocity, pressure, species and NO_x profiles. Combustion reaction scheme with 9 species and 19 steps was modeled using Eddy Dissipation Concept model. Results obtained from this study were validated with published experimental data. Numerical results showed that two equation turbulence models give consistent simulation results with published experimental data by means of trend and value. Renormalization Group k-ε model was found to give consistent simulation results with experimental data, whereas Reynolds Stress Model was failed to predict detailed features of combustion process.     

The effect of stretch on cellular formation in non-premixed opposed-flow tubular flames

Cellular formation in non-premixed flames is experimentally studied in an opposed-flow tubular burner. This burner allows independent variation of the global stretch rate and overall flame curvature. In opposed-flow flames formed by 21.7% hydrogen diluted in carbon dioxide versus air, cells are formed near extinction with a low fuel Lewis number and a low initial mixture strength. Using an intensified CCD camera, the flame chemiluminescence is imaged to study cellular formation from the onset of cells to near extinction conditions. The experimental onset of cellular instability is found to be at or at a slightly lower Damköhler number than the numerically determined extinction limit based on a two-point boundary value solution of the tubular flame. For fuel Lewis numbers less than unity, concave curvature towards the fuel retards combustion and weakens the flame and convex curvature towards the fuel promotes combustion and strengthens the flame. In the cell formation process, the locally concave flame cell midsection is weakened and the locally convex flame cell ends are strengthened. With increasing stretch rate, the flame breaks into cells and the cell formation process continues until near-circular cells are formed with no concave midsection. Further increase in the stretch rate leads to cell extinction. With increasing stretch rate, the flame thickness at the cell midsection decreases similar to a planar opposed-flow flame while the flame thickness at the cell edges is unchanged and can even increase due to the strengthening effect of convex curvature at the flame edges toward the low Lewis number fuel. The results show the existence of cellular flames well beyond the two-point boundary value extinction limit and the importance of local flame curvature in the formation of flame cells.