CO2 addition and pressure effects on laminar and turbulent lean premixed CH4 air flames

An experimental study on CH₄–CO₂–air flames at various pressures is conducted by using both laminar and turbulent Bunsen flame configurations. The aim of this research is to contribute to the characterization of fuel lean methane/carbon dioxide/air premixed laminar and turbulent flames at different pressures, by studying laminar and turbulent flame propagation velocities, the flame surface density and the instantaneous flame front wrinkling parameters. PREMIX computations and experimental results indicate a decrease of the laminar flame propagation velocities with increasing CO₂ dilution rate. Instantaneous flame images are obtained by Mie scattering tomography. The image analysis shows that although the height of the turbulent flame increases with the CO₂ addition rate, the flame structure is quite similar. This implies that the flame wrinkling parameters and flame surface density are indifferent to the CO₂ addition. However, the pressure increase has a drastic effect on both parameters. This is also confirmed by a fractal analysis of instantaneous images. It is also observed that the combustion intensity S_T/S_L increases both with pressure and the CO₂ rate. Finally, the mean fuel consumption rate decreases with the CO₂ addition rate but increases with the pressure.     

Flame front analysis of high-pressure turbulent lean premixed methane–air flames

An experimental study on lean turbulent premixed methane–air flames at high pressure is conducted by using a turbulent Bunsen flame configuration. A single equivalence ratio flame at Φ = 0.6 is explored for pressures ranging from atmospheric pressure to 0.9 MPa. LDA measurements of the cold flow indicate that turbulence intensities and the integral length scale are not sensitive to pressure. Due to the decreased kinematic viscosity with increasing pressure, the turbulent Reynolds numbers increase, and isotropic turbulence scaling relations indicate a large decrease of the smallest turbulence scales. Available experimental results and PREMIX code computations indicate a decrease in laminar flame propagation velocities with increasing pressure, essentially between the atmospheric pressure and 0.5 MPa. The u′/S_L ratio increases therefore accordingly. Instantaneous flame images are obtained by Mie scattering tomography. The images and their analysis show that pressure increase generates small scale flame structures. In an attempt to generalize these results, the variance of the flamelet curvatures, the standard deviation of the flamelet orientation angle, and the flamelet crossing lengths have been plotted against which is proportional to the ratio between the integral and Taylor length scales, and which increases with pressure. These three parameters vary linearly with the ratio between large and small turbulence scales and clearly indicate the strong effect of this parameter on premixed turbulent flame dynamics and structure. An obvious consequence is the increase in flame surface density and hence burning rate with pressure, as confirmed by its direct determination from 2D tomographic images.     

Large scale effects in weakly turbulent premixed flames

In this study we numerically investigate large scale premixed flames in weakly turbulent flow fields. A large scale flame is classified as such based on a reference hydrodynamic lengthscale being larger than a neutral (cutoff) lengthscale for which the hydrodynamic or Darrieus–Landau (DL) instability is balanced by stabilizing diffusive effects. As a result, DL instability can develop for large scale flames and is inhibited otherwise. Direct numerical simulations of both large scale and small scale three-dimensional, weakly turbulent flames are performed at constant Karlovitz and turbulent Reynolds number, using two paradigmatic configurations, namely a statistically planar flame and a slot Bunsen flame. As expected from linear stability analysis, DL instability induces its characteristic cusp-like corrugation only on large scale flames. We therefore observe significant morphological and topological differences as well as DL-enhanced turbulent flame speeds in large scale flames. Furthermore, we investigate issues related to reaction rate modeling in the context of flame surface density closure. Thicker flame brushes are observed for large scale flames resulting in smaller flame surface densities and overall larger wrinkling factors.     

Large eddy simulation of turbulent premixed flames using level-set G-equation

Level-set G-equation and stationary flamelet chemistry are used in large eddy simulation of a propane/air premixed turbulent flame stabilized by a bluff body. The aim was to study the interaction between the flame front and turbulent eddies, and in particular to examine the effect of sub-grid scale (SGS) eddies on the wrinkling of the flame surface. The results indicated that the two types of turbulence eddies—the resolved large scale eddies and the unresolved SGS eddies—have different effects on the flame. The fluctuation of the flame surface, which is responsible for the broadening of the time averaged mean flame brush by turbulence, depends on the large resolved turbulence eddies. Time averaged mean flow velocity, temperature, and major species concentrations mainly depend on the large scale resolved eddies. The unresolved SGS eddies contribute to the wrinkling at the SGS level and play an important role in the enhancement of the propagation speed of the resolved flame front. In addition, the spatially filtered intermediate species, such as radicals, and the spatially filtered reaction rates strongly depend on the small SGS eddies. The asymptotic behavior of flame wrinkling by the SGS eddies, with respect to the decrease in filter size and grid size, is investigated further using a simplified level-set equation in a model shear flow. It is shown that to minimize the influence of the SGS eddies, fine grid and filter size may have to be used.     

Stratified turbulent Bunsen flames: flame surface analysis and flame surface density modelling

In this paper it is investigated whether the Flame Surface Density (FSD) model, developed for turbulent premixed combustion, is also applicable to stratified flames. Direct Numerical Simulations (DNS) of turbulent stratified Bunsen flames have been carried out, using the Flamelet Generated Manifold (FGM) reduction method for reaction kinetics. Before examining the suitability of the FSD model, flame surfaces are characterized in terms of thickness, curvature and stratification.

All flames are in the Thin Reaction Zones regime, and the maximum equivalence ratio range covers 0.1?φ?1.3. For all flames, local flame thicknesses correspond very well to those observed in stretchless, steady premixed flamelets. Extracted curvature radii and mixing length scales are significantly larger than the flame thickness, implying that the stratified flames all burn in a premixed mode. The remaining challenge is accounting for the large variation in (subfilter) mass burning rate.

In this contribution, the FSD model is proven to be applicable for Large Eddy Simulations (LES) of stratified flames for the equivalence ratio range 0.1?φ?1.3. Subfilter mass burning rate variations are taken into account by a subfilter Probability Density Function (PDF) for the mixture fraction, on which the mass burning rate directly depends. A priori analysis point out that for small stratifications (0.4?φ?1.0), the replacement of the subfilter PDF (obtained from DNS data) by the corresponding Dirac function is appropriate. Integration of the Dirac function with the mass burning rate m=m(φ), can then adequately model the filtered mass burning rate obtained from filtered DNS data. For a larger stratification (0.1?φ?1.3), and filter widths up to ten flame thicknesses, a β-function for the subfilter PDF yields substantially better predictions than a Dirac function. Finally, inclusion of a simple algebraic model for the FSD resulted only in small additional deviations from DNS data, thereby rendering this approach promising for application in LES.     

Steady large-scale modulation of a moderately turbulent co-flow jet

The effects of a spatial modulation acting at the inflow of a moderately turbulent planar jet surrounded by a faster co-flow are investigated using direct numerical simulation of the Navier–Stokes equations. We adopt a superposition of spatially filtered small-scale random perturbations and a structured large-scale flow modulation. The large-scale modulation is characterised in terms of a Beltrami flow, specified by a wavenumber K. These large-scale modulations are steady and spatially periodic, while the random small-scale perturbations fluctuate in time and in space. The flow configuration studied in this paper is agitated by this combined large- and small-scale agitation at the inflow plane of a rectangular domain of size L × L × 2L in the x-, y- and streamwise z-directions. The inflow perturbation is focused on a strip of size L × D in the x- and y-directions. A parametric variation is carried out considering different choices for the wavenumber of the large-scale modulation. We focus on effects that the inflow modulation has on global characteristics of the flow, e.g. the width of the mixing region formed between the two streams and the dissipation rate, ?. Results show that the width of the mixing region increases faster compared to the case without the large-scale perturbation, when the flow is agitated by structures of size comparable to the integral scales of the flow. For the dissipation rate, results show the presence of a maximum response at a certain wavenumber K in case we apply a large-scale modulation. This maximum is attained at modulation scales that vary locally with respect to the distance from the inflow plane. Close to the inflow, the maximum response occurs at small modulation scales, while further into the domain a maximum response is present at comparably large modulation scales.     

Measurements of three-dimensional mean flame surface area ratio in turbulent premixed Bunsen flames

A data processing scheme with particular emphasis on proper flame contour smoothing is developed and applied to measure the three-dimensional mean flame surface area ratio in turbulent premixed flames. The scheme is based on the two-sheet imaging technique such that the mean flame surface area ratio is an average within a window covering a finite section of the turbulent flame brush. This is in contrast to the crossed-plane tomograph technique which applies only to a line. Two sets of Bunsen flames have been investigated in this work with the turbulent Reynolds number up to 4000 and the Damköhler number ranging from less than unity to close to 10. The results show that three-dimensional effects are substantial. The measured three-dimensional mean flame surface area ratio correlates well with a formula similar to the Zimont model for turbulent burning velocity but with different model constants. Also, the mean flame surface area ratio displays a weak dependency on turbulence intensity but a strong positive dependency on the turbulence integral length scale.     

Strain rates,flow patterns and flame surface densities in hydrodynamically unstable,weakly turbulent premixed flames

Recent numerical and experimental studies have unveiled a potentially marked difference between the laminar as well as turbulent propagation of premixed flames exhibiting Darrieus–Landau (DL) (or hydrodynamic) instabilities from flames for which instabilities are inhibited. In this study we utilize two-dimensional numerical simulations of slot burner flames as well as experimental Propane–Air Bunsen flames to analyse differences in turbulent propagation, strain rate and induced flow patterns of hydrodynamically stable and unstable flames. We also investigate the effects of hydrodynamic instability on quantities which are directly related to reaction rate closure models, such as flame surface density and stretch factor. A clear enhancement of turbulent flame speed can be observed for unstable flames, generally mitigated at higher turbulence intensity, which is attributed to a flame area increase induced by the characteristic cusp-like DL-induced corrugation, absent in stable flames, which occurs concurrently and in synergy with turbulent wrinkling. Unstable flames also exhibit, both numerically and experimentally, a different correlation between strain rate and flame curvature and are observed to give rise to a channeling of the induced flow in the fresh mixture. Conditionally averaged flame surface density is also observed to attain smaller values in unstable flames, as a result of the thicker turbulent flame brush, indicating that closure models should incorporate instability-related parameters in addition to turbulence-related parameters.     

Modelling of the subgrid scale wrinkling factor for large eddy simulation of turbulent premixed combustion

We propose a model for assessing the unresolved wrinkling factor in the large eddy simulation of turbulent premixed combustion. It relies essentially on a power-law dependence of the wrinkling factor on the filter size and an original expression for the ‘active’ corrugating strain rate. The latter is written as the turbulent strain multiplied by an efficiency function that accounts for viscous effects and the kinematic constraint of Peters. This yields functional expressions for the fractal dimension and the inner cut-off length scale, the latter being (i) filter-size independent and (ii) consistent with the Damköhler asymptotic behaviours at both large and small Karlovitz numbers. A new expression for the wrinkling factor that incorporates finite Reynolds number effects is further proposed. Finally, the model is successfully assessed on an experimental filtered database.     

Evolution and scaling of the peak flame surface density in spherical turbulent premixed flames subjected to decaying isotropic turbulence

The peak flame surface density within the turbulent flame brush is central to turbulent premixed combustion models in the flamelet regime. This work investigates the evolution of the peak surface density in spherically expanding turbulent premixed flames with the help of direct numerical simulations at various values of the Reynolds and Karlovitz number. The flames propagate in decaying isotropic turbulence inside a closed vessel. The effects of turbulent transport, transport due to mean velocity gradient, and flame stretch on the peak surface density are identified and characterized with an analysis based on the transport equation for the flame surface density function. The three mechanisms are governed by distinct flow time scales; turbulent transport by the eddy turnover time, mean transport by a time scale related to the pressure rise in the closed chamber, and flame stretch by the Kolmogorov time scale. Appropriate scaling of the terms is proposed and shown to collapse the data despite variations in the dimensionless groups. Overall, the transport terms lead to a reduction in the peak value of the surface density, while flame stretch has the opposite effect. In the present configuration, a small imbalance between the two leads to an exponential decay of the peak surface density in time. The dimensionless decay rate is found to be consistent with the evolution of the wrinkling scale as defined in the Bray-Moss-Libby model.     

How fast can we burn, 2.0

We review the state of the art in measurements and simulations of the behavior of premixed laminar and turbulent flames, subject to differential diffusion, stretch and curvature. The first part of the paper reviews the behavior of premixed laminar flames subject to flow stretch, and how it affects the accuracy of measurements of unstrained laminar flame speeds in stretched and spherically propagating flames. We then examine how flow field stretch and differential diffusion interact with flame propagation, promoting or suppressing the onset of thermodiffusive instabilities. Secondly, we survey the methodology for and results of measurements of turbulent flame speeds in the light of theory, and identify issues of consistency in the definition of mean flame speeds, and their corresponding mean areas. Data for methane at a single operating condition are compared for a range of turbulent conditions, showing that fundamental issues that have yet to be resolved for Bunsen and spherically propagating flames. Finally, we consider how the laminar flame scale response of flames to flow perturbations interacting with differential diffusion leads to very different outcomes to the overall sensitivity of the burning rate to turbulence, according to numerical simulations (DNS). The paper concludes with opportunities for future measurements and model development, including the perennial recommendation for robust archival databases of experimental and DNS results for future testing of models.     

湍流燃烧自适应求解技术在低排放燃烧室计算中的应用

基于各向异性非结构网格生成技术, 开发了面向复杂几何和复杂湍流燃烧问题的自适应求解算法, 并进行了程序代码的可靠性验证工作, 展示了各向异性网格自适应算法在降低问题求解规模、提高火焰面和流场计算精度等方面的优势.应用该自适应求解技术准确捕捉到了一维预混层流火焰、二维对冲火焰和三维本生灯湍流火焰的流场信息, 火焰面附近的温度、速度、组分等物理量与实验值吻合很好.对一款富油-快速混合-贫油(rich-burn, quick-mix, lean-burn, RQL)低排放发动机燃烧室进行了计算分析, 发现了燃烧室内的热声不稳定现象.      

LES/PDF for premixed combustion in the DNS limit

We investigated the behaviour of the composition probability density function (PDF) model equations used in a large-eddy simulation (LES) of turbulent combustion in the direct numerical simulation (DNS) limit; that is, in the limit of the LES resolution length scale Δ (and the numerical mesh spacing h) being small compared to the smallest flow length scale, so that the resolution is sufficient to perform a DNS. The correct behaviour of a PDF model in the DNS limit is that the resolved composition fields satisfy the DNS equations, and there are no residual fluctuations (i.e. the PDF is everywhere a delta function). In the DNS limit, the treatment of molecular diffusion in the PDF equations is crucial, and both the ‘random-walk’ and ‘mean-drift’ models for molecular diffusion are investigated. Two test cases are considered, both of premixed laminar flames (of thickness δ_L). We examine the solutions of the model PDF equations for these test cases as functions of Δ/δ_L and h/δ_L. Each of the two PDF models has advantages and disadvantages. The mean-drift model behaves correctly in the DNS limit, but it is more difficult to implement and computationally more expensive. The random-walk model does not have the correct behaviour in the DNS limit in that it produces non-zero residual fluctuations. However, if the specified mixing rate Ω normalised by the reaction timescale τ_c is sufficiently large (Ωτ_c ? 1), then the residual fluctuations are less than 10% and the observed flame speed and thickness are close to their laminar values. Away from the DNS limit (i.e. h/δ_L ? 1), the observed flame thickness scales with the mesh spacing h, and the flame speed scales with Ωh. For this case it is possible to construct a non-general specification of the mixing rate Ω such that the flame speed matches the laminar flame speed.     

Direct evaluation of the subgrid scale scalar flux in turbulent premixed flames with conditioned dual-plane stereo PIV

With the dual-plane stereo PIV technique the instantaneous three-dimensional resolved rate-of-strain tensor is directly measured in turbulent premixed flames. Simultaneously, also the instantaneous subgrid scale (SGS) scalar flux is measured with fine resolution, where for the latter term the conditioned particle image velocimetry (CPIV) technique is applied. The subgrid resolution reaches 118 μm, allowing a 9 × 9 resolution of a subgrid filter with width Δ = 1 mm. This combined measurement approach allows the a-priori comparison of models for the SGS scalar flux term with direct measurements which is important for large eddy simulation methods in turbulent premixed flames. Two different flame conditions of a premixed V-shaped turbulent flame are investigated where the turbulence intensity is varied by a factor of nearly three. The instantaneous radial and axial SGS fluxes are compared with the following three models: gradient model with Smagorinsky approach for the turbulent viscosity, Clark model, and extended gradient model with an anisotropy term. None of these models shows a good correlation with the directly measured flux. The anisotropy term alone (being nearly similar to the Clark model) shows, however, a right trend behaviour. An analysis of the data indicates a significant dependency of the experimentally determined SGS flux on the Favre averaged reaction progress (spatially averaged over the SGS area). A relatively simple closure for the SGS flux, which describes the dilatation due to the gasdynamic expansion, and which is a function proportional to , shows a rather good correlation with direct measurement for some of the components. A successful SGS scalar flux model for premixed turbulent flames most likely needs to include at least two different effects.     

Curvature and wrinkling of premixed flame kernels—comparisons of OH PLIF and DNS data

The effects of curvature and wrinkling on the growth of turbulent premixed flame kernels have been investigated using both 2D OH planar laser-induced fluorescence (PLIF) and 3D direct numerical simulation (DNS). Comparisons of results between the two approaches show a high level of agreement, providing confidence in the simplified chemistry treatment employed in the DNS, and indicating that chemistry may have only a limited influence on the evolution of the freely propagating flame. This is in contrast to previous studies of the very early flame development where chemistry may be dominant. Statistics for curvature and wrinkling are presented in the form of probability density functions, and there is good agreement with previous findings. The limitations of 2D PLIF measurements of curvature are quantified by comparison with full 3D information obtained from the DNS. The usefulness of PLIF in providing data over a wide parameter range is illustrated using statistics obtained from both CH₄/air and H₂/air mixtures, which show a markedly different behaviour due to their different thermo-diffusive properties. The results provide a demonstration of the combined power of PLIF and DNS for flame investigation. Each technique is shown to compensate for the weaknesses of the other and to reinforce the strengths of both.     

A LES-CMC formulation for premixed flames including differential diffusion

A finite volume large eddy simulation–conditional moment closure (LES-CMC) numerical framework for premixed combustion developed in a previous studyhas been extended to account for differential diffusion. The non-unity Lewis number CMC transport equation has an additional convective term in sample space proportional to the conditional diffusion of the progress variable, that in turn accounts for diffusion normal to the flame front and curvature-induced effects. Planar laminar simulations are first performed using a spatially homogeneous non-unity Lewis number CMC formulation and validated against physical-space fully resolved reference solutions. The same CMC formulation is subsequently used to numerically investigate the effects of curvature for laminar flames having different effective Lewis numbers: a lean methane–air flame with Le_eff = 0.99 and a lean hydrogen–air flame with Le_eff = 0.33. Results suggest that curvature does not affect the conditional heat release if the effective Lewis number tends to unity, so that curvature-induced transport may be neglected. Finally, the effect of turbulence on the flame structure is qualitatively analysed using LES-CMC simulations with and without differential diffusion for a turbulent premixed bluff body methane–air flame exhibiting local extinction behaviour. Overall, both the unity and the non-unity computations predict the characteristic M-shaped flame observed experimentally, although some minor differences are identified. The findings suggest that for the high Karlovitz number (from 1 to 10) flame considered, turbulent mixing within the flame weakens the differential transport contribution by reducing the conditional scalar dissipation rate and accordingly the conditional diffusion of the progress variable.     

Flame propagation in a small-scale parallel flow

We consider the propagation of laminar premixed flames in the presence of a parallel flow whose scale is smaller than the laminar flame thickness. The study addresses fundamental aspects with relevance to flame propagation in narrow channels, to the emerging micro-combustion technology, and to the understanding of the effect of small scales in a (turbulent) flow on the flame structure. In part, the study extends the results of a previous analytical study carried out in the thick flame asymptotic limit which has in particular addressed the validity of Damköhler's second hypothesis in the context of laminar steady parallel flows. Several new contributions are made here.

Analytical contributions include the derivation of an explicit formula for the effective speed of a premixed flame U _T in the presence of an oscillatory parallel flow whose scale ? (measured with the laminar flame thickness δ _L ) is small and amplitude A (measured with the laminar flame speed U _L ) is (1). The formula shows a quadratic dependence on both the amplitude and the scale of the flow. The validity of the formula is established analytically in two distinguished limits corresponding to (1) frequencies of oscillations (measured with the natural frequency of the flame U _L /δ _L ), and to higher frequencies of (A/?) (the natural frequency of the flow). The analytical study yields partial support of Damköhler's second hypothesis in that it shows that the flame behaves as a planar flame (to leading order) with an increased propagation speed which depends on both the scale and amplitude of the velocity fluctuation. However our formula for U _T contradicts the formula given by Damköhler in his original paper where U _T has a square root dependence on the scale and amplitude.

Numerical contributions include a significant set of two-dimensional calculations which determine the range of validity of the asymptotic findings. In particular, these account for volumetric heat loss and differential diffusion effects. Good agreement between the numerics and asymptotics is found in all cases, both for steady and oscillatory flows, at least in the expected range of validity of the asymptotics. The effect of the frequency of oscillation is also discussed. Additional related aspects such as the difference in the response of thin and thick flames to the combined effect of heat loss and fluid flow are also addressed. It is found for example that the sensitivity of thick flames to volumetric heat loss is negligibly affected by the parallel flow intensity, in marked contrast to the sensitivity of thin flames. Interestingly, and somewhat surprisingly, thin flames are found to be more resistant to heat loss when a flow is present, even for unit Lewis number; this ceases to be the case, however, when the Lewis number is large enough.     

Comparison of a spectral model for premixed turbulent flame propagation to DNS and experiments

A recently developed spectral model for premixed turbulent combustion in the flamelet regime (based on the EDQNM turbulence theory) has been compared with both direct numerical simulations (DNS) and experimental data. The 128³ DNS is performed at a Reynolds number of 223 based on the integral length scale. Good agreement is observed for both single- and two-point quantities (i.e. ratio of the turbulent to laminar burning velocities, scalar autocorrelation, dissipation and scalar-velocity cross correlation spectra) for the two different values of u′/s _L0 considered. The model also predicts the rapid transient behaviour of the flame at early times. An experimental set-up is then described for generating a lean methane-air flame and measuring two-point spatial correlations along the midpoint of the flame brush (i.e. along the C¯=0.5 contour). The experimental measurements in the flamelet regime take the form of a discontinuous or ‘telegraph’ signal. The EDQNM model, in contrast, describes an ‘ensemble’ of flames, and thus is based solely on continuous variables. A theoretical relationship between the correlation obtained from the EDQNM model and the equivalent correlation for a discontinuous (experimental) flame is derived. The relationship is used to enable a meaningful comparison between experimentally observed and model correlations. In general, the agreement is good for the three different cases considered in this study, with most of the error occurring at the lowest Reynolds number (Re _L =22). Furthermore, it is shown that considerably more error would result if no attempt is made to convert the ensemble representation in the model to an equivalent single-flame or ‘telegraph’ signal.     

Detailed measurements of local scalar front structures in stagnation-type turbulent premixed flames

Local scalar front structures of OH mole fraction, reaction progress variable, and its three-dimensional gradient have been measured in stagnation-type turbulent premixed flames. The reaction progress variable front is observed to change with increasing turbulence from parallel iso-scalar contours but reduced progress variable gradients, called the lamella-like front, to disrupted non-parallel iso-contours that deviate substantially from those of wrinkled laminar flamelets, called the non-flamelet front. This transition is attributed to the different scales of interaction between the flame internal structure and a spectrum of turbulence extending from the integral scale to the Kolmogorov scale. The lamella-like front pattern occurs when the length scales of interaction are smaller than the laminar flame thickness but the time scales are greater than the flame residence time. The non-flamelet front pattern occurs when the length scales of interaction are greater than the laminar flame thickness but the time scales are smaller than the flame residence time. This difference corresponds to the change of combustion regime from complex-strain flame front to turbulent flame front on a revised regime diagram. A correlation is also proposed for the turbulent flame brush thickness as a function of turbulent Reynolds number and heat release parameter. The heat release parameter is considered to arise from the non-passive effects of flame-surface wrinkling.     

Modelling of turbulent lifted jet flames using flamelets: a priori assessment and a posteriori validation

This study focuses on the modelling of turbulent lifted jet flames using flamelets and a presumed Probability Density Function (PDF) approach with interest in both flame lift-off height and flame brush structure. First, flamelet models used to capture contributions from premixed and non-premixed modes of the partially premixed combustion in the lifted jet flame are assessed using a Direct Numerical Simulation (DNS) data for a turbulent lifted hydrogen jet flame. The joint PDFs of mixture fraction Z and progress variable c, including their statistical correlation, are obtained using a copula method, which is also validated using the DNS data. The statistically independent PDFs are found to be generally inadequate to represent the joint PDFs from the DNS data. The effects of Z–c correlation and the contribution from the non-premixed combustion mode on the flame lift-off height are studied systematically by including one effect at a time in the simulations used for a posteriori validation. A simple model including the effects of chemical kinetics and scalar dissipation rate is suggested and used for non-premixed combustion contributions. The results clearly show that both Z–c correlation and non-premixed combustion effects are required in the premixed flamelets approach to get good agreement with the measured flame lift-off heights as a function of jet velocity. The flame brush structure reported in earlier experimental studies is also captured reasonably well for various axial positions. It seems that flame stabilisation is influenced by both premixed and non-premixed combustion modes, and their mutual influences.