Turbulent Premixed Flame Propagation Research Articles

Effects of position and frequency of obstacles on turbulent premixed propagating flames

This paper studies the effects of the number and location of solid obstacles on the rate of propagation of turbulent premixed flames. A vented explosion chamber is constructed where controlled premixed flames are ignited from rest to propagate past grids or baffles plates as well as other solid obstacles strategically positioned in the chamber. Laser Induced Fluorescence (LIF) is used to image OH which is used as an indicator of the reaction zone while pressure transducers are used to obtain pressure–time traces. Single grids or baffle plates located at different distances from the ignition source are tested. Two as well as three baffle plates are also investigated in varying configurations. It is found that while the peak overpressure increases with increasing number of grids or baffle plates, a limit is reached where the pressure starts to decrease. The location of the obstacles is found to have a significant effect on the overpressure and the flame structure. Higher overpressures are obtained when the baffle plates and obstacles are stacked closer together hence not allowing turbulence to decay. LIF images for OH show that the reaction zones become more contorted with increasing number of baffle plates in the flame path.

Large eddy simulation of vortex breakdown/flame interaction

The dynamics of a swirl-stabilized premixed flame is studied using large eddy simulation (LES). A filtered flamelet model is used to account for the subgrid combustion. The model provides a consistent and robust reaction-diffusion expression for simulating the propagation of turbulent premixed flames correctly. The numerical results were found to be relatively insensitive to small changes in the inflow boundary conditions and to the numerical mesh employed. Furthermore, the results were found to agree well with the available experimental data both for velocity and scalar fields. In addition, unsteady flame features [i.e., precessing vortex core (PVC)] were identified and compared with experimental data. The agreement between LES results and experimental data, in terms of flame dynamics, was also good. Increasing swirl did not affect the flame strongly but a decrease of swirl number was shown to change the flame shape and suppress the PVC. The PVC and flame dynamics were studied using proper orthogonal decomposition (POD) allowing us to identify and isolate the PVC from smaller-scale turbulence. The POD results indicate that the PVC corresponds to a helical wave consisting of two counter-rotating helices. A dynamical reduced model was also derived do describe the flame response to the PVC.

FLAME SURFACE DENSITIES IN PREMIXED COMBUSTION AT MEDIUM TO HIGH TURBULENCE INTENSITIES

The surface densities of flame fronts in turbulent premixed propane/air flames were determined experimentally. The instantaneous flame fronts were visualized using laser induced fluorescence (LIF) of OH on two Bunsen type burners of 11.2 and 22.4 mm diameters. Non-dimensional turbulence intensity, u′/S L, was varied from 0.84 to 15, and the Reynolds number, based on the integral length scale, varied from 34 to 467. These flames are in the flamelet combustion regime as defined by the most recent turbulent premixed combustion diagrams. From 100 to 800 images were recorded for each experimental condition. Flame surface densities were obtained from the instantaneous maps of the progress variable, which is zero in the reactants and unity in the products. These flame surface densities were corrected for the mean direction cosines of the flame fronts, which had a typical value of 0.69 for the Bunsen flames. In the non-dimensional turbulence intensity range of up to 15, it was found that the maximum flame surface density and the integrated flame surface density across the flame brush do not show any significant dependence on turbulence intensity. This was discussed in the framework of a flame surface density-based turbulent premixed flame propagation closure model. The implication is that the conceptual increase in flame surface density with turbulence may not be the dominant mechanism for flame velocity enhancement in turbulent combustion in the region specified as the flamelet combustion regime by the current turbulent premixed combustion diagrams. Small-scale transport of heat and species may be more important and chemistry may not be decoupled from turbulence. Further, the applicability of the flamelet approach may be limited to a much smaller range of conditions than presently believed.

LES modelling of an unconfined large-scale hydrogen–air deflagration**This paper was presented at the Second International Synposium on Nonequilibrium Processes, Combustion, and Atmospheric Phenomena (Dagomys, Sochi, Russia, 3–7 October 2005).

This paper describes the large eddy simulation modelling of unconfined large-scale explosions. The simulations are compared with the largest hydrogen–air deflagration experiment in a 20 m diameter hemispherical polyethylene shell in the open. Two combustion sub-models, one developed on the basis of the renormalization group (RNG) theory and another derived from the fractal theory, were applied. Both sub-models include a sub-grid scale model of the turbulence generated by flame front itself based on Karlovitz's theory and the observation by Gostintsev et al on a critical distance for transition from laminar to self-similar flame propagation regime. The RNG sub-model employs Yakhot's formula for turbulent premixed flame propagation velocity. The best fit flame propagation dynamics is obtained for the fractal sub-model with a fractal dimension D = 2.22. The fractal sub-model reproduces the experimentally observed flame acceleration during the whole duration of explosion, accurately simulating the negative phase of the pressure wave but overestimating by 50% the positive phase amplitude. The RNG sub-model is closer to the experiment in predicting the positive phase but under-predicts by 30% the negative phase amplitude. Both sub-models simulate experimental flame propagation up to 20 m and pressure dynamics up to 80 m with reasonable accuracy.

Effects of initial radius on the propagation of premixed flame kernels in a turbulent environment

The effects of mean curvature on the propagation of turbulent premixed flames have been investigated using three-dimensional direct numerical simulations (DNS) with single step Arrhenius-type chemistry in the thin reaction zones regime. A number of spherical flame kernels with different initial radius have been studied under identical conditions of turbulence and thermochemistry. A statistically planar turbulent back-to-back flame has been simulated as a special case of a spherical kernel in the limit of infinite kernel radius. Statistical analysis in terms of standard and joint probability density functions (pdfs) clearly indicates that the mean curvature of the flame kernel configuration has a major influence on the propagation behavior of the flame. For the planar flame configuration the density-weighted displacement speed is found to be fairly constant throughout the flame brush, in good agreement with previous DNS results. By contrast, for the flame kernel configuration the density-weighted displacement speed is found to vary strongly through the flame brush, changing from values on the order of the corresponding laminar flame speed near the fresh gas side to considerably smaller values near the burned gas side. The joint pdfs of displacement speed and its components with curvature are extensively studied, allowing for an explanation of the observed phenomena in terms of local flame geometry and its interaction with the turbulent flow field.

Influence of Lewis number on strain rate effects in turbulent premixed flame propagation

The effects of tangential strain rate on the displacement speed of turbulent premixed flames in the thin reaction zones regime are studied for three different Lewis numbers ( Le = 0.8, 1.0 and 1.2) using three-dimensional compressible direct numerical simulation (DNS) of statistically planar flames. For non-unity Lewis numbers it is shown that the variations of temperature and scalar gradient in response to tangential strain rate on a given reaction progress variable isosurface have a profound influence on displacement speed behaviour. In the case of Le = 0.8, temperature and tangential strain rate are found to be positively correlated at locations of zero curvature whereas the opposite behaviour is apparent for the case of Le = 1.2. It is demonstrated that the effects of the temperature-curvature and tangential strain rate-curvature correlations are implicitly present in the response of the temperature to local strain rate. The temperature-curvature correlation and strain rate-curvature correlation are found to be in agreement with previous experimental results. Displacement speed and strain rate are found to be weakly correlated in general, but their conditional joint pdf at zero curvature locations shows a negative correlation.

Influence of Lewis number on curvature effects in turbulent premixed flame propagation in the thin reaction zones regime

The influence of differential diffusion on the statistical behavior of the local displacement speed (Sd) in relation to flame curvature is studied based on three-dimensional compressible direct numerical simulations (DNS) of statistically planar flames with single-step Arrhenius-type chemistry. Three different Lewis number cases (Le=0.8, 1.0, and 1.2) are considered. In order to study the influence of differential diffusion on curvature effects in flame propagation, temperature statistics are presented in terms of standard probability density functions (pdfs) and also joint pdfs with curvature for the nonunity Lewis number cases. Temperature statistics are found to be consistent with previous incompressible combustion DNS studies. It is found that both dilatation and tangential strain rate are negatively correlated with curvature. The relative strength of these two correlations determines the nature of the correlation between surface density function (SDF) (∣∇c∣) and curvature. It is also found that the variations of temperature and SDF on an isosurface of reaction progress variable (c) have significant influence on local displacement speed behavior. Displacement speed statistics are presented in terms of standard pdfs as well as joint pdfs with curvature for the three Lewis number cases. The curvature response of the displacement speed and its different components is found to be nonlinear which is consistent with previous two-dimensional DNS with detailed chemistry. The observed nonlinear behavior in the present study in the absence of a detailed chemical mechanism is explained through the influence of differential diffusion.

A time‐accurate pseudo‐compressibility approach based on an unstructured hybrid finite volume technique applied to unsteady turbulent premixed flame propagation

AbstractA preconditioning approach based on the artificial compressibility formulation is extended to solve the governing equations for unsteady turbulent reactive flows with heat release, at low Mach numbers, on an unstructured hybrid grid context. Premixed reactants are considered and a flamelet approach for combustion modelling is adopted using a continuous quenched mean reaction rate. An overlapped cell‐vertex finite volume method is adopted as a discretisation scheme. Artificial dissipation terms for hybrid grids are explicitly added to ensure a stable, discretised set of equations. A second‐order, explicit, hybrid Runge–Kutta scheme is applied for the time marching in pseudo‐time. A time derivative of the dependent variable is added to recover the time accuracy of the preconditioned set of equations. This derivative is discretised by an implicit, second‐order scheme. The resulting scheme is applied to the calculation of an infinite planar (one‐dimensional) turbulent premixed flame propagating freely in reactants whose turbulence is supposed to be frozen, homogeneous and isotropic. The accuracy of the results obtained with the proposed method proves to be excellent when compared to the data available in the literature. Copyright © 2004 John Wiley & Sons, Ltd.

On the high-resolution modelling of a turbulent premixed open v-flame

The clustering of marker nodes in regions of high curvature and the topological change of flame fronts have long been the main difficulties that limit conventional Lagrangian front-tracking methods from wide use in the modelling of turbulent premixed flame propagation. In this work, we introduce a robust numerical method called contour advection with surgery (CAS) to simulate, in wholly Lagrangian terms, the propagation and advection of the flame front in ambient turbulent flows. With a novel technique known as contour surgery, CAS can properly treat topological changes of the flame front. We apply CAS to study the problem of the two-dimensional turbulent premixed open V-flame. The computed velocity statistics and the estimated flame surface density are compared with corresponding laboratory measurements, and show satisfactory agreement for the trends and basic features. In addition, this method is found to be able to capture the counter-gradient diffusion effects and to reproduce the trends of distributions of local flame front curvature observed in experiments. (Some figures in this article are in colour only in the electronic version; see wwww.iop.org)

Effect of the Darrieus-Landau instability on turbulent flame velocity.

The propagation of turbulent premixed flames influenced by the intrinsic hydrodynamic flame instability (the Darrieus-Landau instability) is considered in a two-dimensional case using the model nonlinear equation proposed recently by Bychkov [Phys. Rev. Lett. 84, 6122 (2000)]. The nonlinear equation takes into account both the influence of external turbulence and the intrinsic properties of a flame front, such as small but finite flame thickness and realistically large density variations across the flame front. Dependence of the flame velocity on the turbulent length scale, turbulent intensity, and density variations is investigated in the case of weak nonlinearity and weak external turbulence. It is shown that the Darrieus-Landau instability influences the flamelet velocity considerably. The obtained results are in agreement with experimental data on the turbulent burning of moderate values of the Reynolds number.

A study of the propagation of turbulent premixed flame using the flame surface density model in a constant volume combustion chamber

Three-dimensional numerical analysis of the turbulent premixed flame propagation in a constant volume combustion chamber is performed using the KIVA-3V code (Amsden et. al. 1997) by the flame surface density (FSD) model. A simple near-wall boundary condition is eaployed to describe the interaction between turbulent premixed flame and the wall. A mean stretch factor is introduced to include the stretch and curvature effects of turbulence. The results from the FSD model are compared with the experimental results of schlieren photos and pressure measurements. It is found that the burned mass rate and flame propagation by the FSD model are in reasonable agreement with the experimental results. The FSD combustion model proved to be effective for description of turbulent premixed flames.

Markstein numbers of negatively stretched premixed flames: Microgravity measurements and computations

The effect of flame stretch, composed of strain and curvature, plays a major role in the propagation of turbulent premixed flames. Although all forms of stretch (positive and negative) are present in turbulent conditions, little research has been focused on the stretch due to curvature. The present study quantifies the Markstein number (which characterizes the sensitivity of the flame propagation speed to the imposed stretch rate) for an inwardly-propagating flame (IPF). This flame is of interest because it is negatively stretched, and is subjected to curvature effects alone, without the competing effects of strain. In an extension of our previous work, microgravity experiments were run using a vortex-flame interaction to create a pocket of reactants surrounded by an IPF. Computations using the RUN-1DL code of Rogg were also performed in order to explain the measurements. It was found that the Markstein number of an inwardly-propagating flame, for both the microgravity experiment and the computations, is significantly larger than that of an outwardly-propagating flame. Further insight was gained by running the computations for the simplified (hypothetical) cases of one step chemistry, unity Lewis number, and negligible heat release. Results provide additional evidence that the Markstein numbers associated with strain and curvature have different values.

Simultaneous laser raman-rayleigh-lif measurements and numerical modeling results of a lifted turbulent H 2/N 2 jet flame in a vitiated coflow

An experiment and numerical investigation is presented of a lifted turbulent H2/N2 jet flame in a coflow of hot, vitiated gases. The vitiated coflow burner emulates the coupling of turbulent mixing and chemical kinetics exemplary of the reacting flow in the recirculation region of advanced combustors. It also simplifies numerical investigation of this coupled problem by removing the complexity of recirculating flow. Scalar measurements are reported for a lifted turbulent jet flame of H2/N2 (Re = 23,600, H/d = 10) in a coflow of hot combustion products from a lean H2/Air flame ((empty set) = 0.25, T = 1,045 K). The combination of Rayleigh scattering, Raman scattering, and laser-induced fluorescence is used to obtain simultaneous measurements of temperature and concentrations of the major species, OH, and NO. The data attest to the success of the experimental design in providing a uniform vitiated coflow throughout the entire test region. Two combustion models (PDF: joint scalar Probability Density Function and EDC: Eddy Dissipation Concept) are used in conjunction with various turbulence models to predict the lift-off height (H(sub PDF)/d = 7,H(sub EDC)/d = 8.5). Kalghatgi's classic phenomenological theory, which is based on scaling arguments, yields a reasonably accurate prediction (H(sub K)/d = 11.4) of the lift-off height for the present flame. The vitiated coflow admits the possibility of auto-ignition of mixed fluid, and the success of the present parabolic implementation of the PDF model in predicting a stable lifted flame is attributable to such ignition. The measurements indicate a thickened turbulent reaction zone at the flame base. Experimental results and numerical investigations support the plausibility of turbulent premixed flame propagation by small scale (on the order of the flame thickness) recirculation and mixing of hot products into reactants and subsequent rapid ignition of the mixture.

An experimental and numerical investigation of premixed flame deflagration in a semiconfined explosion chamber

A combined experimental and numerical study of turbulent premixed flame propagation over multipleobstacles mounted in a semiconfined explosion combustion chamber is reported. The experimental method used a high-speed laser sheet flow visualization technique to record the progress of the flame front in a stoichiometric methane/air mixture. This allowed calculation of flame speed. Pressure was measured at two locations within the combustion chamber. For the simulations, transient Favre-averaged equations were solved using recent developments of a flamelet model. This model formulates the mean rate of reaction as a function of a transport equation for the flamelet surface density model. Both linear and nonlinear eddy viscosity turbulence models have been used and investigated for the closure of the Reynolds stresses. Excellent agreement of flame structure and pressure impulse was obtained with the use of a nonlinear eddy viscosity turbulence model. It was also found that, irrespective of the eddy viscosity turbulence model, quantitative results of pressure and flame speed were found to be in good agreement with the experimental results. Regimes of combustion covered by the present study have been identified and found to reside in the wrinkled and corrugated (reaction sheets) flamelet regimes.

On the Development of a Distributed Reaction Zones Model for Turbulent Premixed Flame of Lean Mixtures in a Closed Vessel

Turbulent premixed flame propagation in a closed spherical vessel with a homogeneous and isotropic turbulent flow field is modelled using the one-point transport equation for the probability density function (pdf) in the thermodynamic space. The pdf equation is discretized using the control volume formulation on a one-dimensional spherical grid. Then the discretized equation is solved using the Monte Carlo method to follow particles with variable volumes on a grid moving with the expansion radial velocity. The mixing term is modelled using the Curl model, and the modified Curl model. The source term is simulated delerministically using a one-step chemical reaction mechanism. Transient solutions are obtained for spherical benzene-air flames using frozen turbulent flow. The pressure rise, the mass fraction burned, the flame speed, the entrainment speed, the burning speed, and the burned gas speed are calculated. The calculated entrainment speeds are compared with the ones reported in Bradley et al (1992) for KLc>5.3 where distributed reaction zones are assumed to govern the flame propagation. The effects of different turbulent flow field (uvS1) and the effect of the product (KLe) are discussed.

Premixed flame–wall interaction in a turbulent channel flow: budget for the flame surface density evolution equation and modelling

Turbulent premixed flame propagation in the vicinity of a wall is studied using a three-dimensional constant-density simulation of flames propagating in a channel. The influence of the walls is investigated in terms of the flamelet approach, where flamelet speed and flame surface density transport are used to describe the flame. The walls have constant temperature and lead to flamelet quenching for sufficiently small wall–flame distances. Starting from the exact evolution equation for the surface density of propagating interfaces (Trouvé & Poinsot 1994; Candel & Poinsot 1990; Pope 1988), a budget for the flame surface density equation is presented before, during, and after the interaction with the wall. Before the flame interacts with the wall, flame propagation is controlled by a balance between surface production and annihilation. During the interaction, high flame surface density gradients near the wall are responsible for the predominance of the transport terms. Closures of all terms of the flame surface density equation are proposed. These models are based on flamelet ideas and take into account wall effects. Enthalpy loss through the wall affects flamelet speed, flamelet annihilation and flame propagation. Decrease of turbulent scales near the wall affects turbulent diffusion and flame strain. This model is compared to DNS results using two types of tests: (i) a priori tests, where individual terms of the modelled flame surface density equation are compared to the terms of the exact interface density propagation equation, calculated with the DNS; (ii) a posteriori tests, where the final model is used to obtain total reaction rate, mean fuel mass fraction, heat flux at the wall and fuel mass fraction at the wall in the configuration used in the DNS. For both types of tests the model compares well with the DNS results.

A novel flow reactor for the study of heat-loss effects on turbulent flame propagation

A novel flow reactor is proposed for the study of heat-loss effects on the propagation of turbulent premixed flames. This Taylor-Couette flow reactor allows the stabilization of a premixed flame in a stationary, high-intensity turbulent flow and the examination of the role of heat transfer from the flame (to the walls of the apparatus) on the experimentally observed quenching of the premixed flame at high turbulence intensities. The key advantages and design parameters of the flow reactor are identified and discussed.

Turbulent premixed flame propagation in a cylindrical vessel

An experimental and theoretical study of premixed flame propagation in a cylindrical vessel, containing turbulence inducing rings is described. The study was undertaken to provide further understanding of flame propagation and the generation of overpressure in explosions and to assess a mathematical model of explosions through comparison with detailed experimental data. Data gathered on flame location and shape, flow velocities, and overpressure during the course of an explosion are used to elucidate the dynamics of the various combustion processes occurring in the different chambers of the vessel. In particular, flame propagation through the vessel, until the flame front vents from the tube, is found to be substantially laminar, with significant overpressure only being generated in the later stages of the explosion due to rapid turbulent combustion in the shear layers and recirculation zones induced by the obstacles. Comparisons between measurements and numerical predictions obtained using an explicit, second-order accurate, finite-volume integration scheme, coupled to an adaptive grid algorithm, demonstrate that the mathematical model used provides a reasonable simulation of combustion within the vessel In particular, good qualitative agreement with observations is obtained for flame speed and shape, with rapid turbulent combustion being predicted with sufficient accuracy to yield reliable results for flow velocities and the overpressures generated.

Temporal evolution of turbulence/chemistry interactions in lifted, turbulent-jet flames

Flame ignition, extinction, and reignition that control flame stability are inherently time-dependentphenomena. To better understand and quantify these important characteristics, two-pulse images of the instantaneous CH/CH 4 distributions, with a variable time delay between images, were used to study the temporal evolution of turbulence/chemistry interactions in lifted, turbulent CH 4 -jet flames. The jet-exit velocity was 21 m/s, corresponding to a jet-exit Reynolds number of 7000. The results are presented to emphasize the interactions between turbulence structure and the flame that lead to the large fluctuations in flame liftoff height observed experimentally. The results show that the flame envelope curves around semiorganized, large-scale vortical structures formed along the periphery of the jet. Interactions between these vortical structures and the flame cause excessive stretch that leads to local flame extinction. Extinction occurs on both the downstream edges of the structures and on the upstream sides. Once the flame is extinguished by a vortical structure, the flame stabilization point can be carried significant distances downstream by the vortex. The return to an upstream stabilization location occurs through turbulent premixed flame propagation. The images clearly show upstream propagation of the flame within the plane of the image. Evidence also exists for the importance of flame propagation in the azimuthal direction normal to the image plane.

Turbulent premixed flame propagation models for different combustion regimes

Conceptual models of turbulent premixed flame propagation speeds for the three combustion regimes have been proposed and tested by comparing to the available well-defined data obtained on a variety of experimental rigs. The model for the wrinkled flame regime has been based on the Tennekes and Kuo-Corrsin concept of isotropic turbulence which proposes that dissipative eddies are most probably consist of vortex tubes with a diameter of the order of Kolmogorow scale and a spacing of the order of the Taylor length scale. Using this concept of turbulence structure, it was shown that the normalized turbulent flame propagation speed, u t /u l , is proportional to the square root of the normalized turbulence intensity, u′/u l and the turbulent Reynolds number based on Kolmogorov length scale. R n . In the intermediate regime, assuming that the effective flame front surface area grows exponentially with the turbulence intensity, u t /u l . is linearly dependent on R ν and exponentially dependent on the square root of u′/u l . In the distributed reaction zones regime, starting with the Damköhler's formulation, dimensionless numbers contributing to the flame propagation have been selected, and it was found that u t /u′ is proportional to the 3/4th power of u l /u′ . Comparison of the predictions of the models and the experimental data from a variety of rigs shows excellent agreement despite the uncertainties involved in the measurement of the turbulence parameters and propagation velocities.