Flame-vortex Interaction Research Articles

Experiments on the interaction between a vortex and a strained diffusion flame

Flame-vortex interactions constitute a basic problem in the analysis of turbulent combustion. Vortex rollup is also found to be one of the major driving mechanisms of combustion instabilities. While there are many analytic and numerical studies of the process, the number of detailed experiments is relatively limited. In particular, the nonpremixed case has not been explored, apparently because the experimental configuration is less easily designed. It is shown here that this case may be examined by employing a counterflow burner. A steady diffusion flame is established in this geometry and a vortex ring is generated from a cylindrical tube installed in one of the combustor nozzles. The vortex impinges on the flame from the oxidizer side entraining the reactive layer and producing a hole in the initial flame. The apparatus is described and initial results of visualizations and velocity measurements by laser doppler anemometry are discussed. It is found that the interaction leads to different outcomes. Strong vortices produce flame extinction and a subsequent blowout. In contrast, the flame is recovered after an interaction with a weak vortex.

On the structure of a stretched/compressed laminar flamelet—Influence of preferential diffusion

The local Lewis number on the fuel side of a H 2/air diffusion flame is knwn to be lower than unity. When a high-Damkohler-number hydrogen flame is stretched, the flame temperature changes—not because of incomplete chemical reactions, but due to the preferential diffusion resulting from the nonunity Lewis number. These changes are examined using a Computational Fluid Dynamics with Chemical reactions (CFDC) code that is third-order accurate. Detailed chemical kinetics including NO reactions is used to investigate the effects of stretch on the flame structure. A low-speed H 2/air diffusion flame having a fuel-jet velocity of 3.26 m/s is subjected to stretch using two vortices—one located on the fuel side of the flame and the other on the air side. Both vortices create positive (stretch) and negative (compression) flame-stretch factors when interacting with the flame. The temperature of the positively stretched flamelet was found to decrease when the vortex is located on the air side and to increase when it is on the fuel side. Similarly, compression of the flamelet resulting from the air-side vortex increases the flame temperature, whereas fuel-side vortex decreases the flame temperature. Among the eleven species considered in this model, production of NO in the flame zone appears to be the most sensitive to flame stretching or compression. Concentration of NO is found to be maximum in the compressed flamelet formed during the outside vortex-flame interaction. Finally, a comparison of the mean-NO-concentration profiles obtained by averaging over several cycles of outside vortex-flame interactions and the time-averaged measurements yielded favorable agreement.

Effect of unsteady stretch rate on OH chemistry during a flame-vortex interaction: To assess flamelet models

Effect of unsteady stretch rate on OH chemistry during a flame-vortex interaction: To assess flamelet models

Effect of unsteady stretch rate on OH chemistry during a flame-vortex interaction: To assess flamelet models

Some basic assumptions of flamelet models are assessed by comparing profiles of OH mole fraction measured during an unsteady flame-vortex interaction to the OH profiles computed for a steady, planar counterflow flame (SPCF) with full chemistry. It is important to make such comparisons for both the same local three-dimensional stretch rate, which is measured instantaneously at locations along the flame front, and the same heat loss, as characterized by the product temperature T 2. The fundamental experimental procedure consists of interacting a laminar, premixed flame with an impinging vortex ring of reactants. The OH flame chemistry was quantified using planar laser-induced fluorescence (PLIF) techniques, while the three-dimensional stretch rate measurements were made possible by the use of particle-imaging velocimetry (PIV) diagnostics on the repeatable, axisymmetric experiment. Whereas previous comparisons have been limited to steady counterflow flame experiments, the present study considers a flame which is unsteady, freely-propagating, curved, far from walls, and has realistic heat losses; thus, it contains the physical processes present in turbulent premixed flames. It was found that there are significant (25%) differences between measurements and steady counterflow flame computations of peak OH mole fractions and OH reaction zone widths. Even where the stretch rate was constant along the flame, the OH profiles showed variations, indicating that the OH profile is not a unique function of the instantaneous local stretch, but depends on the time history of the flowfield. Such history effects may be better modeled using unsteady counterflow flame simulations. Large differences (a factor of two on centerline) occur between the measured three-dimensional and two-dimensional stretch rates, indicating the importance of experimentally determining the full three-dimensional stretch rate for meaningful comparisons with models. Sensitivity analysis shows that heat losses must be realistically modeled, especially if flame extinction is to be simulated. The present type of comparison represents a first step in the assessment of flamelet models.

Applications of direct numerical simulation to premixed turbulent combustion

This paper describes recent progress in the field of Direct Numerical Simulations (DNS) for premixed turbulent combustion. A classification of Computational Fluid Dynamics techniques used in reacting flows is first proposed. Numerical methods and problems specific to DNS of combustion like grid resolution or boundary conditions treatment are discussed. The presentation of DNS results is done in the framework of flamelet modeling. The domain of validity of the flamelet assumptions is studied with simulations of flame vortex interactions. The local structure of flamelets in premixed turbulent combustion is investigated using two and three-dimensional DNS with simple chemistry models. Effects of complex chemistry are considered through recent two-dimensional DNS performed with realistic chemical schemes. Flame surface densities, flame strain and curvature statistics obtained from various three dimensional simulations are given and implications for modeling are presented. DNS is also exploited to deal with some specific problems like turbulent flame ignition or flame-wall interactions. This review is concluded with a discussion of current limitations and a list of future challenges.

Numerical experiments on the vortex-flame interactions in a jet diffusion flame

Dynamic behavior of a laminar diffusion flame in response to an artificial vortex that issues radially from the fuel-jet core toward the flame zone has been studied numerically to illustrate essential physics of a naturallyforming shear layer vortex that has some radial velocity. A time-dependent, axisymmetric, implicit, third-order accurate numerical model is used with the infinitely-fast chemistry and unity Lewis number assumptions. A packet of fuel (methane) is ejected as a single-pulsed jet at given initial and boundary conditions (the ejection velocity and period), which cover an order of magnitude in the time scale of the vortex-flame interaction. A vortex system with a pair of counter-rotating toroidal vortex rings is generated and penetrated into the hightemperature (highly viscous) layer with solid-body rotation. If the Peclet number Pe defined as the ratio of the characteristic diffusion to convection times, is small (the order of 10 or less), the vortex system pushes out the flame surface. If Pe is large (the order of 100), the vortex system nearly cuts through the high-temperature layer with a minimal flame movement, thus creating a significantly thin diffusive-thermal layer. In both cases, as the vortex approaches the flame surface, the net radial velocity of the incoming oxidizer stream crossing the flame and the reactants' diffusive fluxes into the flame increase. As a result, the flame structure similar to that of a strained counterflow diffusion flame is formed. A periodically-p ulsed side has also been studied; the fuel packets ejected (at 500 Hz) induces the development of a large-scale vortex train in the shear layer of the primary jet, interacting with the flame zone.

Effects of curvature and unsteadiness in diffusion flames. Implications for turbulent diffusion combustion

Direct numerical simulations (DNS) of flame-vortex interaction are used to study the validity limits ofthe laminar flamelet assumption (LFA) in diffusion flames. Flame structures measured from DNS are compared with laminar flamelet library predictions obtained from asymptotic analysis performed with similar chemical and transport models (finite rate chemistry and variable density are fully accounted for). Domains where the LFA applies for flame-vortex interaction cases are presented in a (length scale-velocity scale) spectral diagram. Three main effects limit the LFA domain: (1) unsteady effects (the flame structure does not respond fast enough to hydrodynamic excitation), (2) curvature effects (nonlocal effects modify the one-dimensional structure of the flame), and (3) quenching. All these effects are evidenced through numerical simulations and presented on the spectral diagram. As a result, quantitative information is obtained on the exact limits of the LFA for turbulent diffusion combustion. Results show, for example, that quenching is obtained in conditions that are beyond the LFA domain and that, in certain cases, quenching strain rates in turbulent flames may not be obtained from simple laminar flamelet concepts.

Large eddy simulation and direct simulation of compressible turbulence and combusting flows in engines based on the BI-SCALES method

Physical features of turbulence and vortex-flame interaction in engines are investigated by performing large eddy simulation and direct numerical simulation of compressible flows. A BI-SCALES (Boundary-Inner Smoothly Coupled, Alternating multi-Level Equations System) is proposed, which is a mathematical formulation of the system of governing equations suitable for computations of compressible flows. A numerical method based on the proposed formulation is developed. Using the method, the transition to turbulence in the compression process and the cyclic variations are examined. Then, the vortex-flame interaction is studied, mainly on the relation between the flame structure and the Kalvoritz- and turbulent Reynolds number effect. Finally, the large wrinkle simulation (LWS) of engine combusting flow is performed.

OH fluorescence images of the quenching of a premixed flame during an interaction with a vortex

Local quenching of a premixed laminar flame has been imaged using OH planar laser induced fluorescence as the flame interacts with the well-defined strain field of a single laminar vortex. This flame-vortex interaction represents a fundamental “building block” of premixed turbulent combustion. Since the OH concentrations reach a maximum in the flame-front which is 10 times larger than the value in the burnt gas, the measured [OH] profiles indicate the instantaneous shape of the reaction zone. Local quenching is confirmed by the disappearance of the reaction zone. Four different vortex strengths were studied; the weakest vortex causes flame wrinkling only; stronger vortices create pockets of unburned gas. Local quenching occurs at the leading edge of the vortex when the Karlovitz number (based upon the vortex diameter and rotational velocity) exceeds 0.17, which is significantly less than the value 1 predicted previously. Quenching is believed to result from a combination of flame stretch and radiative heat loss in the burnt gas. Effects of flame curvature are quantified; the highest OH fluorescence intensities occur when the flamefront is convex towards the reactants, as expected for these flames for which Lewis number

Stretching and quenching of flamelets in premixed turbulent combustion

The stretch rate of flamelets in premixed turbulent combustion is computed using (1) detailed numerical simulations of vortex-flame interactions and (2) a model for intermittent turbulence taking into account all possible turbulence scales acting on the flame front. Simulations of interactions between isolated vortices and a laminar flame front are used to obtain a relation between the characteristics of a given vortex and the actual flame stretch generated by this structure. Quenching conditions and quenching times are also given by these simulations. A net rate of stretch is then defined in the case of a complete turbulent flow field as the difference between the total rate of flame stretch and the quenching rate due to scales that have a high enough energy and a long enough lifetime to quench locally the flame front. The net rate of stretch is computed for a variety of parameters of interest in practical applications. It is a function of the large-scale turbulence parameters and the laminar flame speed and flame thickness and may be used as an input in most flamelet models for premixed turbulent combustion. Different criteria for total flame quenching in premixed turbulent combustion are derived and compared (1) to the classical Klimov-Williams theory, (2) to a criterion proposed by Poinsot et al. [8, 9], who studied quenching according to the presence near the flame front of a single eddy able to locally quench combustion, and (3) to the experimental results of Abdel-Gayed and Bradley [6, 7].

Simulations of flame-vortex interactions

Full numerical simulations are used to study the interaction of a vortex and a premixed flame as a model problem for turbulent flames. A vortex structure is constructed that avoids boundary condition difficulties. The effects of heat release and the importance of the relative length and time scales of the vortex and flame are examined. Changes in the internal structure as well as the overall shapes of the flame and vortex are studied. Frozen flame and frozen vortex cases are studied to isolate various effects and aid in understanding the full interaction. The principal effect of the frozen flame is to expand the vorticity normal to the flame surface. A model using the baroclinic torque is developed to explain this effect. The principal effect of the frozen vortex on the flame is to increase the flame area. This is modeled using the total strain rate associated with the vortex. Important changes in internal flame structure were found to occur only at relatively low Damköhler numbers (Da < 0.5). The full flame-vortex interaction is a combination of the two frozen cases, with additional effects due to flame curvature and to the simultaneous evolution of the flame and vortex. The former results in flame generated vorticity, an effect directly related to the heat release. The latter altered the vortex structure resulting in changes in the pressure field and increased dissipation.

A new spectral method for calculation of the time-varying area of a laminar flame in homogeneous turbulence

A novel approach to flame area simulation is presented, based on the physical processes of chaotic mixing. The “wrinkled” flame front is represented by a length-scale distribution of distortions of the surface, bounded by the chamber or flame kernel size, and the laminar flame thickness. The turbulent fluid in which the flame propagates is modelled as a length scale spectrum of vortices, with the contribution of each scale to the total turbulent kinetic energy being obtained from a prescribed energy spectrum function. Vortex-flame interaction is modelled by: (a) flame area generation by the vortices at their length scale, (b) stretching and folding by generation of flame area at larger scales. The effect of flame propagation on the radius of curvature of the flame constitutes the third mechanism for altering the “wrinkle” spectrum. The resulting integro-differential equation is solved numerically by a time-marching procedure to yield the evolution of surface area and asymptotic (equilibrium) limit for given conditions of turbulence and prescribed laminar flame speed. Good agreement with experimental results is obtained for the fractal dimension of passive scalar and flame surfaces, and also with turbulent flame speed data over a wide range of conditions. The model also makes predictions of flame-speed trends in the near-wall region in agreement with observed behaviour, and also shows that there is a weak dependence of turbulent flame speed on S l additional to that on u′ /S l used in the Bradley correlations. The effects of turbulent flame stretch and vortices smaller than the laminar flame thickness are not modelled, leading to reduced agreement in the high u′ /S l limit. The physical nature of the modelling means that these, and other phenomena such as holes in the flame front caused by rapid local stretching, can be easily included in future work.

Propagation of a premixed flame in a divided-chamber combustor

A study of the propagation of lean premixed ethyleneair flames in a divided-chamber combustion vessel was conducted to compare experimental observations with a numerical simulation based on a flame sheet-vortex dynamics model in axisymmetric coordinates. The temporal and spatial development of flames exiting a small cylindrical prechamber were observed in the experiments using laser-schlieren videography. The equivalence ratio of the ethyleneair mixture was varied from 0.5 to 0.65. The corresponding flame speeds S u for this range of equivalence ratios were measured in a complementary study and were shown to increase from 10 to 24 cm/s, respectively, while the ratio of unburned to burned gas densities varied by less than 20 %. The associated increase in gas velocity with equivalence ratio was used to change the estimated Reynolds number in the experiment from 1870 to 8090. Good agreement between the predicted and measured prechamber flame propagation rates was obtained. For the lowest Reynolds number experiment the calculated spatial and temporal development of the flame in the main combustion chamber agreed with the experimental observations. For all Reynolds numbers the numerical results scale with the characteristic time scale obtained by normalizing with the burned gas flame speed S u. In the experimental results this normalized time scale is not an accurate measure of similar flame development. As the Reynolds number was increased in the experiment the flame development was more rapid. At the highest Reynolds number conditions, a combination of the simplifying assumptions in the flame sheet-vortex dynamics model results in a slower development of the flame-vortex interaction without the experimentally observed shear layer instabilities.

Flame Generation of Vorticity: Vortex Dipoles from Monopoles

Abstract Numerical simulation of premixed flame propagation through a two-dimensional vorticity distribution exhibits the coupling between combustion heat release and fluid dynamics. Effects of both thermal expansion and vorticity generation via baroclinic torques are considered. The rotational part of the velocity field is described by discrete vorticity. The irrotational velocity. given by a velocity potential. is determined by a Poisson equation where the chemical reaction determines the spatial distribution of volume expansion. The transport of a single reacting scalar is computed on an Eulerian mesh with a small-Mach-number flow assumption so that the density gradient is nonzero only within the reaction zone. The vector cross product of density gradient and pressure gradient defines the baroclinic generation of vorticity. With a single initial vortical region-the monopole interacting with a premixed flame-the baroclinic effect produces opposite circulation, and thus. creates a dipole configuration. We use the dipole description in a loose sense. because in some solutions the two signs of vorticity are spatially intermingled. However. in comparison to the monopole. the "dipole" solution has stronger local velocity fluctuations but weaker long-range velocities. Thus, the turbulence in the burnt gas is more intense with smaller length scales than the configuration before the flame-vortex interaction.

An experimental investigation of two dimensional flame-vortex interactions

Experiments have been performed to study the interaction between a flame and a single vortex impulsively generated by the flame. The objectives are to investigate the amplitude of the cylindrically diverging acoustic wave produced by the interaction and to develop a quantitative approach to determine the degree of violence of the flame-vortex interaction. With the Reynolds number, defined as the ratio of the strength of the vortex to the kinematic viscosity of the unburned gas, characterizing the unburned gas flow condition quantitatively, the acoustic pressure Δ p(t) generated by the interaction has been measured at distances 1.5, 1.75 and 2.0 meters away from the interaction zone using a microcomputer-based data acquisition system and a storage oscilloscope. The dimensionless time rate of change in the heat addition rate, Q( t ), of the flame-vortex interaction, is deduced from the acoustic pressure data via the line source equation in linear acoustic theory. The Fast Fourier Transform has been used to perform the analysis. For the thirty experiments performed, the amplitudes of Q( t ) and Δ p(t) have been found to be dependent on Re, burning speed and two dimensionless geometric parameters. The results show that Q max increases with Re when the geometric parameters are fixed. An empirical logarithmic equation has been obtained to describe the variation of Q max for Re between 10 5 and 4.5×10 5 . The maximum Δ p max recorded was 72 mbar at 1.5 m from the interaction zone. The increase in the “violence” of the interaction has been found to be associated with an increase in the heat addition rate during the period of flame-vortex interaction.

Gas turbine engine emissions—Problems, progress and future

The present article reports the investigation on the effect of inlet flow turbulence on flame–vortex dynamics during the thermo-acoustically induced flame flashback in a premixed dump combustor. The inlet turbulence intensity can be varied independently of the Reynolds number by using plates of different slot widths and is placed significantly upstream of the nominal flame-stabilization zone. The study aims to experimentally address the flame flashback behavior and the conventional combustion dynamics mechanisms involving flame–vortex interaction. The reacting flow-field is characterized using high-speed simultaneous measurement of dynamic pressure fluctuation with the time-resolved/phase-resolved visualization of flame and flow-field utilizing CH*/OH* chemiluminescence and particle image velocimetry. We note that higher turbulence intensity delays the onset of combustion dynamics and significantly influence the flame flashback behavior. In the present work, we observed two types of flame flashback behavior with large amplitude oscillations — (a) intermittent flame flashback, which relates to upstream flame displacement during certain phases of the acoustic pressure cycle and (b) permanent flame flashback, which results in flame quenching. The results reveals that the recirculation zone length plays a key role in determining the flame flashback. The role of enhanced turbulence is seen in two effects- (1) Dictating the length of the recirculation zone and (2) The modulation of the flame by the wake vortex, which correlates with the tendency of flame to flashback directly.

On the Construction and Comparison of Difference Schemes

On the Construction and Comparison of Difference Schemes