SESSION 6

FLAME DYNAMICS AND TURBULENCE

ABSTRACT

The demand for efficient, compact and clean combustion systems have spurred research into the fundamental mechanisms governing their performance and means of interactively changing their performance characteristics. Thermoacoustic instability which is frequently observed in combustion systems with high power density, when burning close to the lean flammability limit, or using exhaust gas recirculation to meet more stringent emission regulations, etc. Its Occurrence and/or means to mitigate them passively lead to performance degradation such as reduced combustion efficiency, high local heat transfer rates, increase in the mixture equivalence ratio or system failure due to structural damage. This paper reports on our study of the origin of thermoacoustic instability, its dependence on system parameters and the means of actively controlling it.

We have developed an analytical model of thermoacoustic instability in premixed combustors. The model combines a heat release dynamics model constructed using the kinematics of a premixed flame stabilized behind a perforated plate with the linearized conservation equations governing the system acoustics. This formulation allows model based controller design. In order to test the performance of the analytical model, a numerical solution of the partial differential equations governing the system has been carried out using the principle of harmonic separation and focusing on the dominant unstable mode. This leads to a system of ODEs governing the thermo fluid variables: Analytical predictions of the frequency and growth rate of the unstable mode are shown to he in good agreement with the numerical simulations as well as with those obtained using experimental identification techniques when applied to a laboratory combustor. We use these results to confirm the validity of the assumptions used in formulating the analytical model. A controller based on the minimization of a cost function using the LQR technique has been designed using the analytical model and implemented on a bench top laboratory combustor. We show that the controller is capable of suppressing the pressure oscillations in the combustor with a settling time much shorter than what had been attained before and without exciting secondary peaks.

ABSTRACT

Modeling of a lean premixed propane/air turbulent flame stabilized by a bluff body is presented. The laminar flamelet library approach, which has been successfully used for modeling of nonpremixed turbulent flames, is applied to premixed combustion. In this approach the mean flame location and mean flame thickness are modeled by a level-set G-equation and the variance of G, along with empirical expressions. The detailed species, temperature and density are calculated using a presumed probability density function together with the laminar flamelet library. The sensitivity of the results to the model components is illustrated. Comparison of calculations and experimental data is presented.

ABSTRACT

Turbulent combustion rates in lean hydrogen-air mixtures (f=0.39, 0.26, and 0.21) diluted with carbon dioxide were measured in a closed vessel with central ignition at room temperature and pressures ranging from 0.1 to 0.5 MPa. Turbulence was produced by four identical fans with rotational speeds up to 6000 r.p.m. The turbulent intensity ranged up to about 10 m/s and measured by a thermoanemometer. Inhomogeneity of the turbulent intensity did not exceed 20%.

The methods of high-speed schlieren photography and pressure transient measurements were used for the studies. Results show that turbulent premixed flames in the mixtures diluted with CO₂ exhibit turbulent quenching. The flame speed increases initially with turbulence up to a point where this speed decreases just prior to quenching of the flame. The extinction limits in turbulent flames of hydrogen-air-carbon dioxide mixtures at the concentrations of hydrogen and carbon dioxide close to the limiting concentrations obtained in the experiments with laminar flames. The variation of initial pressure from 0.1 to 0.5 MPa does not affect the turbulent combustion rate.

ABSTRACT

The a-parameter is a variable appearing in the formulation of the Bray number, a dimensionless term that can be used to predict the behaviour of the turbulent scalar flux as counter gradient or gradient in premixed combustion. However, the successful application of the Bray number is hampered by poor characterisation of the a-parameter. The results of previous laser-imaging studies of conditional mean velocities in turbulent premixed methane/air and propane/air Bunsen-flames are used to formulate an empirical description of a. The resulting description of a suggests that the behaviour of the scalar flux is dependent on the heat release parameter and the lengthscale ratio, l_I/d_L, rather than the velocity ratio, u'/S_L.

ABSTRACT

We calculate the two-dimensional unsteady reactive flow to study the influence of body-force instability on the temporal motion of cellular flame fronts at low Lewis numbers. The equation used is the compressible Navier-Stokes equation including an exothermic one-step irreversible chemical reaction, where the hydrodynamic effect caused by thermal expansion is taken into account. We superimpose the disturbance with the peculiar wavelength on a stationary plane flame and. simulate the evolution of the disturbed flame front. The disturbance superimposed grows initially with time and then the flame front changes from a sinusoidal to a cellular shape. After the formation of a cellular flame, cells on the flame move laterally when the Lewis number is lower than unity. Because, the diffusive-thermal effect and the nonlinear effect of the flame front play a primary role in the appearance of the lateral movement of cells. Body-force instability has a great influence on the lateral velocity of cells and on the structure of flame fronts. As the acceleration increases, the lateral velocity becomes smaller and the cell depth becomes larger. The former is due to the augmentation of the high-temperature region at a convex flame front toward the unburned gas, and the latter is due to the increase in the instability level.

ABSTRACT

Key Words: Methane-air flames, Electric field effects on flames, Ionic wind

The effects of an electric field on the behavior of methane-air flames have been studied. Co-flowing diffusion flames and candle-type flames under the electric field effect have been observed experimentally. A numerical model, which considers various physical and chemical phenomena associate with the flame-field interaction process, has been developed to explain the experimental observations. The model employs a two-dimensional cylindrical coordinate system and assumes axial symmetry. A simplified chemical reaction scheme for a methane- air mixture which contains 19 chemical species and 31 reactions is employed. The mass, momentum, species and energy conservation equations are solved numerically by an integrated version of the PHOENICS and CHEMKIN computer codes. It is concluded that the effects of an electric field on the flame behavior are mainly due to ionic wind effects.

ABSTRACT

The reduction of the rate of the increase of burning velocity U_t with r.m.s turbulent velocity u' or even the reduction of burning velocity with u' are well-known phenomena (the so-called bending effect) which are still badly predicted by the premixed turbulent combustion theory. The goal of the paper is to discuss certain geometrical and transient mechanisms which can contribute to the phenomena side by side with the processes of flamelet stretching and quenching, commonly invoked to explain them. For these purposes, the propagation of planar one-dimensional and statistically spherical, premixed turbulent flames is simulated by employing the Turbulent Flame Speed Closure Model. The results of the simulations reveal two effects. First, the reduction in the flame speed by the flame curvature, well known but weak under laminar conditions, is substantial in the turbulent case. The reduction increases with u' and this can lead to the bending or even a decrease in U_t(u'). Second, when the ignition energy provided by a spark is close to a critical value, a regime of kernel development, characterized by reduced burning velocity, can occur even in relatively large, spherical turbulent flames. The physical cause of this "memory effect" consists in the formation of a highly dispersed kernel followed by slow afterburning. When the spark energy is kept, constant, the increase of u' increases the critical ignition energy and the aforementioned regime occurs. As a result, a decrease in U_t with u' can be observed. The two mechanisms affect statistically curved turbulent flames only.

ABSTRACT

Experimental and numerical studies on the laminar burning velocity of hydrogen- air mixtures were performed. Measurements of laminar burning velocities were conducted using a new technique based on a particle tracking velocimetry (PTV) and image processing for burner-stabilized flames in a high-pressure chamber. Equivalence ratios of the mixtures were varried from 0.6 to 3.0 for the pressure range from 0.1 MPa to 0.5 MPa. A numerical simulation was conducted by solving Navier-Stokes equations considering detailed reaction mechanisms and transport properties. At high pressures, the experimental and numerical results agreed reasonably with each other for mixtures of equivalence ratios of 1.0 and 2.0, while discrepancies were seen for equivalence ratio 3.0. These discrepancies can be diminished effectively by modifying the rate-coefficient expressions of recombination reactions.

ABSTRACT

The structure of premixed methane-air flames is analyzed using the "laminar flamelet concept". A new model based on one-dimensional set of transformed coupled second order differential conservation equations describing the species concentrations of CO₂, CO, O₂, CH₄, H₂O, H₂ and N₂ and the sensible enthalpy are presented in the present work. The equations are rigorously derived and solved numerically. In these equations, a reaction progress variable (c) is taken as the independent variable that varies from zero to one. A three-step chemical kinetic mechanism is adopted. This was deduced in a systematic way from a detailed chemical kinetic mechanism. The rates for the three steps are related to the rates of the elementary reactions of the full reaction mechanism. Calculations are made for different fixed values of the scalar dissipation rate (c) until the flamelet eventually reaches the extinction limit at different levels of pressure. Moreover, simultaneous and instantaneous 1-D measurements of CO₂, O₂, CO, N₂, CH₄, H₂O, H₂, OH and temperature have been carried out in a premixed laminar methane-air flame. A one-dimensional UV Raman-Rayleigh and Laser Induced Predisposition (LIPF) technique has been applied in the present work. The spatial and temporal resolutions are limited to the signal-to-noise ratio and the laser pulse duration. The results of the calculations are assessed against the measurements and previous predictions based on the asymptotic approach of Cl mechanism and a 4-step reduced mechanism. The burning velocity at different equivalence ratios was also deduced from the flamelet properties and assessed against available data.

ABSTRACT

Local scalar flame properties of freely-propagating turbulent premixed flames including the flame curvature h, and local displacement flame speed S_d have been measured simultaneously for methane, propane and hydrogen/air flames at Lewis numbers varying from 0.33 to 1.4 and u'/S_L° varying from 0 to 3.1. An advanced field imaging technique based on high speed laser tomography was used to measure the temporal evolution of local flame properties. Local flame curvature and local displacement speed were calculated from flame front contours. Local flame speed S_d is controlled by the mean flame stretch since the temporal evolution of Sa repartition on flame contours indicates a strong variability of mean and local flame speed with flame expansion. Flame response in terms of displacement speed to curvature is found to be statistically depend and a linear relationship is observed. For propane/air flames the local displacement speed can be assumed to be independent of local flame curvature at each stages of flame propagation, whereas very strong increasing of displacement speed with positive curvatures can be observed along the wrinkled flame contour for hydrogen/air flames. Spatially averaged values of S_d and h for low turbulent flames, computed at each stages of flame propagation, indicate a good agreements with the asymptotic relation, with signs and values Markstein length depending of value of Lewis numbers.

ABSTRACT

An experimental and numerical study of the turbulent flow structure of three rectangular jet flows was conducted. Measurements were performed using a hot- wire anemometer having an x-type probe. Predictions were obtained using a finite-difference based computer program that solves the parabolic governing equations of the mean flow along with the turbulence kinetic energy and its dissipation rate. Measurements of the streamwise mean velocity, kinetic energy and shear stress are presented and compared with the numerical results. The achieved good agreement between the experimental and numerical results proves the ability of the mathematical model to predict successfully the flow field of multiple rectangular jet flows.

Keywords: Experimental data, numerical predictions, multiple rectangular jets