C.J. Sun

Dynamics of weakly stretched flames: quantitative description and extraction of global flame parameters

Generalized expressions for the flame response to weak stretch rate variations were derived based on an integral analysis. Together with values of the laminar flame speed, laminar flame thickness, and the one-step overall reaction order and activation energy determined from the computational results of the one-dimensional planar flame, these expressions for the stretched flames were then used to correlate the computational results of the spherical outwardly propagating, spherical inwardly propagating, and counterflow hydrogen/air and propane/air flames. These correlations yielded the laminar flame speeds through linear extrapolation to zero stretch rate, the Markstein lengths representing the sensitivity of the flame response to stretch rate, and the flame Lewis number. Furthermore, it is shown that the extracted Markstein lengths and Lewis numbers from the three flame configurations are largely the same for given equivalence ratio and system pressure, and that these Lewis numbers also agree well with those predicted from the two-reactant flame theory of Joulin and Mitani. The feasibility of a priori quantitative determination of stretch effects on laminar premixed flames is suggested.

On the structure of nonsooting counterflow ethylene and acetylene diffusion flames

The structures of ethylene/oxygen/nitrogen and acetylene/oxygen/nitrogen diffusion flames in the counterflow configuration were investigated experimentally and computationally. The temperature and major species concentration profiles were measured with spontaneous Raman scattering. The experimental situations were computationally simulated with detailed reaction mechanisms and transport properties. The kinetic mechanism was based on GRI-Mech, with modifications to predict more closely the adiabatic flame speeds of ethylene/air and acetylene/air mixtures, and with additional description of higher hydrocarbon formation and oxidation up to C6 species. The numerical predictions were found to be in reasonably good agreement with the experiment. Both experimental and computational results indicate that acetylene is the major intermediate species in the ethylene flame, having a significant influence on the heat release, overall fuel destruction, and molecular mass growth. The reaction pathways leading to benzene formation in these flames were examined computationally, with the goal of achieving a better understanding of soot nucleation in diffusion flames.

Mild oxidation regimes and multiple criticality in nonpremixed hydrogen-air counterflow

Abstract This study investigates experimentally and computationally the existence of mild oxidation regimes and multiple ignition and extinction states in a system of nonpremixed, counterflowing hydrogen against heated air. Spontaneous Raman spectroscopy measurements of the water concentration show that up to three stable stationary states can be achieved for identical boundary conditions. Computationally, up to five steady-state solutions can be found, although only three are likely to be stable. This multiplicity is the result of combined thermokinetic and transport effects on the behavior of critical ignition and extinction states. To understand these effects, the system response was simulated using detailed kinetics and transport properties, and S-curve sensitivity was employed to identify the dominant chemistry near the critical states and to simplify the kinetic mechanism. The response to changes in the fuel concentration and system pressure was investigated experimentally by measuring the air temperatures corresponding to ignition and extinction, for fuel concentrations in the range of 6–38% H 2 in N 2 by volume, and pressures between 0.3 and 8 atm, at a constant pressure-weighted strain rate of 300 s −1 . The experimental results were found to agree well with the computational results. The experimental triple-solution multiplicity disappears for fuel concentrations in excess of ∼25% or below ∼7% H 2 in N 2 , and was only found in the pressure range between ∼1.5 and 7 atm, at 9% H 2 in N 2 and a pressure-weighted strain rate of 300 s −1 . In addition, the response to changes in the strain rate was studied computationally, for strain rates between 10 and 40,000 s −1 and for air boundary temperatures ranging between 950 and 1100 K. The same features of up to five steady-state multiplicities and up to two ignition and extinction states can be obtained by changing the flow strain rate. In the strain rate space, the computational quintiple-solution multiplicity extends from ∼ 100–10,000 s −1 , at 9% H 2 in N 2 and 4 atm.

Response of counterflow premixed and diffusion flames to strain rate variations at reduced and elevated pressures

The thermal structure of counterflow premixed and diffusion flames was experimentally and computationally studied to examine the response of flame structure to strain rate and pressure variations. The temperature profiles were experimentally measured as a function of strain rate at reduced and elevated pressures by using spontaneous Raman scattering, and were found to agree well with the independently computed profiles using detailed reaction mechanisms and transport properties. For both near-equidiffusive and non-equidiffusive premixed flames, results show that variation of the thermal structures is much smaller than that of the strain rate, and the structural sensitivity decreases with increasing pressure due to the reduced flame thickness. For diffusion flames, results show that the flame structure at different pressures largely scales with the density-weighted strain rate instead of the strain rate alone. Further analysis was conducted to extract the global flame parameters from the thicknesses of stretched non-equidiffusive flames in response to strain rate and pressure variations. It is demonstrated that the effects of stretch vary linearly with the imposed strain rate for the flames studied herein, and that the stretch effects can be predicted with good accuracy by knowing the laminar flame speeds of the one-dimensional planar flame as a function of the system pressure and equivalence ratio and by knowing the mixture Lewis number as a function of the equivalence ratio.

On the nonlinear response of stretched premixed flames

The nonlinear response of stretched premixed flames was studied analytically and computationally, with emphasis on the turning point behavior associated with flame extinction. Extending our previous study of the weakly stretched flames exhibiting linear responses, expressions for the nonlinear flame response were derived by using an integral method. For adiabatic flames, the extinction stretch rates were determined by using the global flame parameters extracted from the linear flame response. The predicted flame responses were then compared with computed results for counterflow and inwardly propagating spherical flames, and reasonably good agreements were obtained. The agreement was further improved by taking into account the increase of the effective activation energy with decreasing flame temperature as the flame approaches the extinction state. For nonadiabatic flames, the effect of volumetric heat loss was investigated via the one-dimensional planar flame subject to radiation loss, while the effect of conductive loss was studied via the counterflow flame against an isothermal wall. The formulation holds potential utility in predicting quantitatively accurate flame responses and in its implementation in the modeling of turbulent combustion.

On the consumption of fuel pockets via inwardly propagating flames

The extinction mechanisms of fuel pockets and the formation of uncreacted residue fuels through flamevortex interaction in turbulent flames were investigated via a computational study of inwardly propagating spherical flames (IPF) in lean and rich hydrogen/air mixtures. Results obtained assuming quasi-steady propagation show that though extinction of the Le Le >1, rich flame is induced by the depletion of the deficient reactant ahead of the flame and yields practically no unconsumed reactant. Results obtained for the more realistic, transient propagation show that flame propagation can actually persist almost to the center for both the lean and rich flames, and, as such, all deficient reactants are consumed upon flame extinction. Extending these results to include hydrocarbon flames within the context of stretch and nonequidiffusion, it is suggested that the formation of unreacted fuel pockets via flamevortex interaction is not expected to be a matter of serious concern because of the potential of complete consumption of the deficient reactant upon extinction of the IPF.

On adiabatic stabilization and geometry of bunsen flames

Two aspects of stretched flame dynamics are investigated via the model problem of the stabilization andgeometry of Bunsen flames. Specifically, the possibility of stabilizing a Bunsen flame without heat loss to the burner rim is experimentally investigated by examining the temperature of the rim, the temperature gradient between the rim and the flame base, and the standoff distance of the flame base in relation to the flame thickness. Results show that, while heat loss is still the dominant stabilization mechanism for flames in uniform flows and for strong flames in parabolic flows, adiabatic stabilization and, subsequently, blowoff are indeed possible for weak flames in parabolic flows. The adiabatically stabilized flame is then modeled by using the scalar field formulation and by allowing for the effects of curvature and aerodynamic straining on the local flame speed. The calculated flame configuration agrees well with the experiment for the adiabatically stabilized flame but not for the nonadiabatic flame. Results further show that active modification of the flame curvature is the dominant cause for the flame to maintain adiabatic stabilization. Implications of the present results on turbulent flame modeling are discussed.

Stretch effects in counterflow and propagating spherical flames

Generalized expressions for the flame response to stretch rate variations were derived based on an integral analysis. Together with values of the laminar flame speeds, laminar flame thicknesses, and the one-step overall reaction order and activation energy determined from the computational results of the one-dimensional planar flame, these expressions for the stretched flames were then used to correlate the computational results of the spherical outwardly-propagating, spherical inwardlypropagating, and counterflow hydrogen/air and propane/air flames. Such correlations yielded the laminar flame speeds through linear extrapolation to zero stretch rate, the Markstein lengths representing the sensitivity of the flame response to stretch rate, and the flame Lewis number. Furthermore, it was shown that the extracted Markstein lengths and Lewis numbers from the three flame configurations are largely the same for given equivalence ratio and system pressure, and that these Lewis numbers also agree well with those predicted from the two-reactant flame theory of Joulin and Mitani. The feasibility of a priori determination of stretch effects on laminar premixed flames is emphasized. Introduction Since the late seventies, significant advance has been made on the theoretical description of stretched flames, and as such there now exists a reasonably mature qualitative understanding on the structure and response of laminar premixed flames to stretch rate variations. Specifically, it is well established that stretch effects can be manifested through aerodynamic straining, flame curvature, and flame motion, and that these influences are particularly strong in the presence of mixture nonequidiffusion because of the resulting modification of the flame temperature. Since these formulations invoke such simplifying assumptions as one-step overall reaction and constant transport properties, it is obviously not justifiable to conduct quantitative comparisons between the theoretical predictions and results obtained through either experimentation or computational simulation with detailed reaction and variable transport properties. The rational approach towards enabling the analytical results quantitatively useful is thus to extract certain bulk flame parameters characterizing the stretch expressions through comparisons between the analytical and experimental/computational results. Such an extraction would be especially meaningful if the bulk parameters extracted are shown to be applicable to flames subjected to diverse modes of stretching. As a first step towards such an approach, in Ref. 1 the response of the symmetrical counterflow methane/air and propane/air flames to strain rate and pressure variations were studied. The extraction was based on the analytical result that the sensitivity of the flame response to stretch rate is a function of several parameters characterizing the reference, one-dimensional unstretched adiabatic flame in the doubly-infinite limit. These include the laminar flame speed (

Stationary state multiplicities in diffusive-reactive systems - Hydrogen versus heated air

„) and flame thickness (Sf), which are directly determinable from the flame solution, the Zeldovich number (Ze) and hence the one-step overall activation energy (Ea), which can be extracted from the dependence of S^ on the adiabatic flame temperature (T^), and the mixture Lewis number (Le), which has been conventionally * Graduate Student ** Research Staff Member *** Robert H. Goddard Professor, Fellow AIAA Copyright

Analytic description of the evolution of two-dimensional flame surfaces

This study investigates the existence of multiple stationary states in a system of nonpremixed, counterflowing hydrogen against heated air. Raman spectroscopy measurements of the water concentration show that up to three stable stationary states can exist for identical boundary conditions. The system is simulated using detailed kinetics and transport properties, and S-curve sensitivity is employed to identify the dominant chemistry. The response of the multiplicity regime to changes in the fuel concentration and system pressure is investigated by measuring the air temperatures corresponding to ignition and extinction for concentrations between 6-38 percent H2 in N2 (vol) and pressures between 0.3-8 atm. The triple-solution multiplicity is found between about 7-25 percent H2 in N2 at 4 atm, and between about 1.5-7 atm at 9 percent H2 in N2 and a constant pressure-weighted strain rate of 300/s. In addition, the response to changes in the strain rate is studied computationally for strain rates between 10 and 40,000/s and for air boundary temperatures ranging between 950 and 1100 K. The experimental results are found to agree well with the computational results. (Author)