Simulations Of Flames Research Articles

Large eddy simulation of OME3 and OME4 spray combustion under heavy-duty conditions

The synthetic diesel fuel (Poly-)oxymethylene ether (OME) has various interesting properties for the transport sector. OME reduces emissions in direct injection (DI) diesel engines, e.g., soot and NOx, and is considered CO2 neutral. While most studies have examined a mixture of OMEs with different chain lengths, recent studies have focused on neat OME3 and OME4. In the literature similar chemical ignition delay times were shown under homogeneous conditions and significant differences were reported in their thermophysical properties. Furthermore, differences between OME3 and OME4 in terms of the mixture formation were shown on a heavy-duty injector under non-reactive conditions. The question of how this mixture formation affects the ignition and flame characteristics of neat OME3 and OME4 spray flames remains unresolved. To address this topic, this study conducts a combined experimental and numerical investigation under inert and reactive spray conditions for OME3 and OME4 on a heavy-duty injector. Large Eddy Simulations (LESs) and high-pressure, high-temperature spray chamber measurements are conducted. The distinct differences between neat OME3 and OME4 are analyzed along the cause–effect chain comprising the mixture formation, ignition behavior and its effects on the overall heat release rate. Based on the differences in the thermophysical properties, e.g., density and vapor pressure, the investigations in inert conditions reveal differences in the liquid penetration and mixture formation. The reactive spray flame measurements and simulations show a significant offset in the ignition delay time for OME3 and OME4. An explanation is found in the significantly different distribution of the scalar dissipation rate due to the differences in the mixture formation. Despite the differing ignition delay times, similar local flame structures are found for both fuels.

A pairwise mixing model with kernel constraint and its appraisal in transported PDF simulations of ethylene flames

Modeling micro-mixing remains a primary challenge for the transported probability density function (TPDF) method. In this study, the pairwise mixing with kernel constraint (KerM) model is proposed and evaluated in TPDF simulations of a turbulent non-premixed ethylene flame. The new model features controllable localness by selecting mixing particle pairs with a probability that depends on their distance in composition space, normalized by a characteristic kernel size. To evaluate the model in a controlled setting, TPDF simulations were carried out with the mean velocity, turbulent diffusivity and scalar mixing frequencies extracted from a direct numerical simulation (DNS) dataset. The predictions of the mean statistics and the conditional PDFs illustrate that the KerM model exhibits asymptotic behavior to the modified Curl (MC) and the Euclidean minimum spanning trees (EMST) model, respectively, depending on the kernel size. For the flame considered, results for KerM approach those for EMST when the kernel size is less than 0.01, while they approach those for MC when the kernel size is larger than 1.0. Given a kernel size of 0.1, KerM yields a notably better prediction of the overall combustion process than the conventional MC model. The predictions of the mean temperature and species mass fractions by KerM are similar to those by EMST, while KerM better predicts the conditional PDF of temperature than EMST that overdamps the conditional fluctuations. In addition, its complexity with the number of particles per cell is instead of O(Np2) for EMST. By incorporating different mixing timescales among species, the KerM-DD model better predicts the mean temperature field at the time instant close to local extinction than the KerM and EMST models. This work demonstrates that the KerM model, apart from being simple and robust in terms of implementation, features the advantages of being flexible in localness and being capable of incorporating differential diffusion.

High-order methods for the simulation of unsteady counterflow flames subject to stochastic forcing of large amplitude

Unsteady counterflows are employed to understand and model the effect of turbulence on flames. We present a novel numerical approach for the simulation of one-dimensional unsteady counterflow flames with fourth order spatial discretization and up to fourth order time discretization. The approach couples a three-stage Lobatto IIIa formula for boundary value problems and variable-order, variable time step size Backward Differentiation Formulas for time integration. The framework is explained in detail, its computational performance is analysed, and its use is demonstrated for the case of stochastic forcing of premixed counterflow flames, whereby the imposed rate of strain is a multi-scale lognormal discrete random process with exponential autocorrelation. High-order spatial and temporal discretization make the approach well-suited for the accurate and computationally efficient simulation of the effect of turbulence on flames, which are characterised by large amplitude stochastic fluctuations of the local rate of strain.

A computationally efficient approach for soot modeling with discrete sectional method and FGM chemistry

A novel approach for the prediction of soot formation in combustion simulations within the framework of discrete sectional method (DSM) based univariate soot model and Flamelet Generated Manifold (FGM) chemistry, referred to as FGM-CDSM, is proposed in this study. The FGM-CDSM considers the clustering of soot sections derived from the original soot particle size distribution function (PSDF) to minimize the computational cost. Unlike conventional DSM, in FGM-CDSM, governing equations for soot mass fractions are solved for the clusters, by using a pre-computed lookup table with tabulated soot source terms from the flamelet manifold, while the original soot PSDF is re-constructed in a post-processing stage. The flamelets employed for the manifold are computed with detailed chemistry and the complete sectional soot model. A comparative assessment of FGM-CDSM is conducted in laminar diffusion flames for its accuracy and computational performance against the detailed kinetics-based classical sectional model. Numerical results reveal that the FGM-CDSM can favorably reproduce the global soot quantities and capture their dynamic response predicted by detailed kinetics with a good qualitative agreement. Furthermore, compared to detailed kinetics, FGM-CDSM is shown to substantially reduce the computational cost of the complete reacting flow simulation with soot particle transport. Primarily, the use of FGM reduces the overall calculation by about two orders of magnitude compared to detailed kinetics, which is advanced further with the clustering of sections at a low memory footprint. Therefore, the present work demonstrates the promising capabilities of FGM-CDSM in the context of computationally efficient soot calculations and provides an excellent framework for extending its application to the simulations of turbulent sooting flames.

Skeletal kinetic mechanism for predicting formation of non-fuel hydrocarbons and soot in ethylene flames – A CFD approach

Since existing reduced mechanisms of ethylene oxidation are either too large or lack of sufficient chemical reactions to describe aromatics formation, the study aims at developing a skeletal mechanism with extended predictive capability and lowered computational cost in computational fluid dynamics modeling of ethylene combustion. The newly assembled and refined kinetic mechanism composed of 664 species and 3582 reactions is minimized to 79 species and 538 reactions without empirically adjusting kinetic parameters. In the 1-D simulation for a premixed flame of ethylene, the ethylene-PAH mechanism generally well predicts experimental profiles of five hydrocarbons and nine aromatics. Integrated into a 2-D axisymmetric laminar finite-rate model, the ethylene-PAH mechanism reproduces experimental data of seven C2–C5 hydrocarbons, five aromatic hydrocarbons and soot in a diffusion flame of ethylene. In comparison with the 158-species (Narayanaswamy) mechanism, the newly proposed ethylene mechanism leads to a speedup of approximately 5.5 times in the 2-D laminar diffusion flame simulation and improved accuracy for predicting species formation in both premixed and nonpremixed flames. Moreover, we unveil the correlation between reaction pathways and intermediate formation from ethylene oxidation at the low temperature regime.

Measurement and Simulation of Autoignition-Assisted Flame Speed of N-Heptane Mixture at Elevated Gas Temperature

Laminar flame speed and autoignition are two fundamental characteristics of a hydrocarbon’s combustion. These two characteristics are extensively studied due to their importance in designing and developing combustion systems. However, a regime in which these two characteristics simultaneously affect the combustion process is not well understood. Thus, the primary focus of this investigation is to understand the autoignition-assisted flame regime prior to the first-stage heat release. The premixed mixture of n-heptane in oxygen–nitrogen–diluent at an equivalence ratio of 1.0, a compressed gas pressure of 6.85 bar, and compressed gas temperatures of 621 K (when using argon) and 616 K (when using helium) were studied in this work. At these initial conditions, the mixture ignition delays are 32.31 ms (using argon) and 53.92 ms (using helium). The flame-propagating images were recorded using a high-speed camera at different spark ignition times to measure the flame locations. The Rapid Compression Machine─Flame (RCM-Flame) apparatus was used to perform experiments, and one-dimensional modeling and three-dimensional numerical modeling were performed. The simulations provided data to understand the effect of stretch on the measured spherical burning velocity. The measured and simulated autoignition-assisted laminar flame speeds were in excellent agreement, the linear method can be used to remove the stretch from the spherical burning velocity, and the autoignition-assisted flame speeds show a negligible dependency on the spark ignition times when the first-stage heat release is negligible.

Data reduction considerations for the burning velocity of spherical constant volume flames of R32 (CH2F2) with air

The present work explores data reduction techniques for the measurement of the laminar burning velocities of R32(CH2F2)-air mixtures using a constant volume combustion device, in which the pressure-time history is the only measured parameter. To allow clear assessment of the accuracy of the data reduction methods, the pressure-time histories used for analysis are synthetically generated via a detailed numerical simulation employing full kinetics and with and without an optically-thin radiation model. Various data reduction models are employed, including a two-zone model and two multi-zone models, and these are compared with the results from the burning velocity obtained from the output of the numerical simulation. The data reduction schemes are shown to be accurate if the same radiation model is employed in the data reduction as was used in the flame simulation to generate the pressure trace used for post-processing. If the incorrect radiation model is employed, however, the errors can be quite large. The effects of stretch, radiation, and different data post-processing methodologies are explored and the errors quantified. Stretch is shown to be important for the early stages and the selected data range that is used for extrapolation has a significant effect on the extrapolated burning velocity. However, with an appropriate choice of data considered for extrapolation, the prediction of the unstretched burning velocity can be quite accurate.

Stretch Effects in Large Eddy Simulation of Turbulent Premixed Bunsen Flames

ABSTRACT The stretch effects in large eddy simulation (LES) of a turbulent lean propane-air premixed Bunsen flame in the wrinkled flamelet regime are investigated. A simple approach to model the subgrid-scale flame stretch is proposed. It is based on the analysis which shows that, when the surface-filtered strain rate in a subgrid-scale stretch model is approximated using the volume-filtered velocity field solved for in LES, heat release and associated gas expansion in the filtered flame brush tend to artificially alter the wrinkling of resolved flame fronts. To minimize such artifacts, it is suggested that the volume-filtered strain rate on the unburned side of the filtered flame brush be used to approximate the surface-filtered strain rate and projected through the filtered flame brush. The results show the importance of mitigating the artificial heat release effects when considering the strain effects in LES. The relative importance of the curvature stretch is also investigated in terms of the mean and local effects. For a Bunsen flame with a positive Markstein number, the mean flame curvature effect tends to increase the total burning rate by enhancing the laminar flame speed near the flame tip, while the local flame curvature effect tends to decrease the total burning rate by suppressing the wrinkling of the resolved flames. It is found that the competition between the two makes the overall curvature effects not influence the total burning rate much for the flame investigated here, as compared with the strain effects.

Pressurised combustion in hot and low-oxygen environments: Chemiluminescence imaging and modelling

Turbulent flames in a confined and pressurised jet-in-hot-coflow combustor, under a range of pressures and coflow conditions, are investigated in this study. The investigation involves both experimental and numerical analyses, with the central focus on extending the use of an atmospheric-pressure computational fluid dynamics (CFD) model to describe hot and low-oxygen combustion at elevated pressures. A series of CFD simulations are compared against experimental images, in terms of both flame structure and chemiluminescence behaviour. The imaging of OH* and CH* is performed experimentally for flames with Reynolds numbers of 10,000 and 15,000, issuing into coflows with O2 concentrations ranging from 3%–9%. The experimental results show a reduction in OH* and CH* intensity with increasing pressure under the various coflow conditions, while the CFD results display some key differences in the trends with pressure and O2. The computational analysis is complemented by a series of one-dimensional laminar flame simulations at a range of pressures and oxidant O2 concentrations. These simulations enable the changes in chemical kinetics with pressure and O2 concentration to be studied in greater detail, and several key differences in comparison with the CFD results are observed. Ultimately, the results highlight the importance of accurately predicting both the flow-field and finite-rate chemistry to reproduce the trends observed under hot and low-oxygen combustion conditions at elevated pressures.

Large eddy simulations of ammonia-hydrogen jet flames at elevated pressure using principal component analysis and deep neural networks

The combustion of ammonia/hydrogen is currently gaining importance in the power generation sector as an alternative for hydrocarbon fuels and improved fundamental insights will facilitate its application. To investigate the complex interactions between turbulence and chemistry for ammonia-hydrogen jet flames under high-pressure conditions, large eddy simulation (LES) computations are conducted using the PC-transport model, which is based on Principal Component Analysis (PCA), coupled with nonlinear regression that utilizes deep neural networks (DNN) to enhance the size-reduction potential of PCA. Classical statistics-based nonlinear regression methods are inefficient at fitting highly nonlinear manifolds and when large data sets are involved. These two drawbacks can be overcome by utilizing DNN tools. Several DNN architectures composed of fully connected layers of different depths and widths, batch normalization, and various activation functions coupled with various loss functions (mean squared error, absolute error, and R2) are explored to find an optimal fit to the thermo-chemical state-space manifold. The ability to achieve highly accurate mapping through DNN-based nonlinear regression with an R2-score >0.99 is shown by employing a single graphical processing unit (Nvidia RTX 3090). Furthermore, the proposed PC-DNN approach is extended to include differential diffusion based on a rotation technique and utilization of the mixture-averaged transport model for the training data set. To demonstrate the potential of the PC-DNN approach in modeling turbulent non-premixed combustion, LES results are compared with the recent Raman/Rayleigh scattering measurements that were obtained at the KAUST high-pressure combustion duct (HPCD). Results show that the PC-DNN approach is able to capture key flame characteristics with reasonable accuracy using only two principal components. The inclusion of differential diffusion leads to improved predictions, although some discrepancies are observed in fuel-lean regions.

Investigation of Very Large Eddy Simulation Method for Applications of Supersonic Turbulent Combustion

The very large eddy simulation (VLES) method was investigated for supersonic reacting flows in the present work. The advantages and characteristics of the VLES model and the widely used improved delayed detached eddy simulation (IDDES) method were revealed through a supersonic ramped-cavity cold flow. Compared to the IDDES model, the VLES model transformed from RANS mode to LES mode faster, resulting in a smaller gray region caused by the mode transition. However, the original volume-averaging truncation length scale could lead to poor predictions of the velocity profiles and wall pressure distribution. By introducing a hybrid truncation length scale combining the maximum grid length and the shear layer adaptive (SLA) length with different coefficients, the accuracy of the VLES method was significantly improved, and the issue of the low shear layer position was solved. Moreover, owing to the resolution control function, the VLES method could adaptively model more turbulent kinetic energy and maintain a good accuracy in a coarser mesh. Finally, the modified VLES method was applied in conjunction with a hybrid combustion model constructed by the partially stirred reactor (PaSR) model and the Ingenito supersonic combustion model (ISCM) in simulations of the supersonic flame in the DLR scramjet combustor. After introducing the correction of the molecular collision frequency by the ISCM, the results obtained by the hybrid combustion model were more consistent with the experimental results, especially for the time-averaging temperature profile in the ignition zone.

Combining flamelet-generated manifold and machine learning models in simulation of a non-premixed diffusion flame

Flamelet Generated Manifold (FGM) is an example of a chemistry tabulation or a flamelet method that is under attention because of its accuracy and speed in predicting combustion characteristics. However, the main problem in applying the model is a large amount of memory required. One way to solve this problem is to apply machine learning (ML) to replace the stored tabulated data. Four different machine learning methods, including two Artificial Neural Networks (ANNs), a Random Forest (RF), and a Gradient Boosted Trees (GBT), are trained, validated, and compared in terms of various performance measures. The progress variable source term and transport properties are replaced with the ML models. Particular attention was paid to the progress variable source term due to its high gradient and wide range of its value in the control variables space. Data preprocessing is shown to play an essential role in improving the performance of the models. Two ensemble models, namely RF and GBT, exhibit high training efficiency and acceptable accuracy. On the other hand, the ANN models have lower training errors and take longer to train. The four models are then combined with a one-dimensional combustion code to simulate a counterflow non-premixed diffusion flame in engine-relevant conditions. The predictions of the ML-FGM models are compared with detailed chemical simulations and the original FGM model for key combustion properties and representative species profiles.

Prediction of flashback limits for laminar premixed hydrogen-air flames using flamelet generated manifolds

Development of models that can help predict flashback limits of premixed flames at an affordable computational cost is essential for the safe and efficient design of combustion chambers. For flames with strong preferential diffusion effects, usually the focus has been on the development of at least a three dimensional flamelet database that can predict the enthalpy and mixture fraction mapped on to the reaction progress variable. However, in this study, we show that a 3D FGM table is sufficient to predict flashback limits for lean laminar methane-air flames but is not sufficient to predict the same for lean hydrogen flames and an over-prediction of 100% could occur in the calculation of the flashback limits. We trace the root cause of this over-prediction to be related to the thickness of the reaction zone in the progress variable for hydrogen flames. This results in the development of a novel correction factor for the progress variable source term using 1D flame simulations where the flame experiences strong enthalpy gradients. In the end, we successfully show for the first time that the flashback limits for hydrogen flames can be predicted accurately using flamelet generated manifolds with a source term corrector function.

Simultaneous single-shot imaging of H and O atoms in premixed turbulent flames using femtosecond two-photon laser-induced fluorescence.

A method based on femtosecond two-photon excitation has been developed for simultaneous visualization of interference-free fluorescence of H and O atoms in turbulent flames. This work shows pioneering results on single-shot simultaneous imaging of these radicals under non-stationary flame conditions. The fluorescence signal, showing the distribution of H and O radicals in premixed CH4/O2 flames was investigated for equivalence ratios ranging from ϕ = 0.8 to ϕ = 1.3. The images have been quantified through calibration measurements and indicate single-shot detection limits on the order of a few percent. Experimental profiles have also been compared with profiles from flame simulations, showing similar trends.

On the incorporation of conservation laws in machine learning tabulation of kinetics for reacting flow simulation

Tabulation of chemical mechanisms with artificial neural networks (ANNs) offers significant speed benefits when computing the real-time integration of reaction source terms in turbulent reacting flow simulations. In such approaches, the ANNs should be physically consistent with the reaction mechanism by conserving mass and chemical elements, as well as obey the bounds of species mass fractions. In the present paper, a method is developed for satisfying these constraints to machine precision. The method can be readily applied to any reacting system and appended to the existing ANN architectures. To satisfy the conservation laws, certain species in a reaction mechanism are selected as residual species and recalculated after ANN predictions of all of the species have been made. Predicted species mass fractions are set to be bounded. While the residual species mass fractions are not guaranteed to be non-negative, it is shown that negative predictions can be avoided in almost all cases and easily rectified if necessary. The ANN method with conservation is applied to one-dimensional laminar premixed flame simulations, and comparisons are made with simulations performed with direct integration (DI) of chemical kinetics. The ANNs with conservation are shown to satisfy the conservation laws for every reacting point to machine precision and, furthermore, to provide results in better agreement with DI than ANNs without conservation. It is, thus, shown that the proposed method reduces accumulation of errors and positively impacts the overall accuracy of the ANN prediction at negligible additional computational cost.

Characterisation of hydrogen jet flames under different pressures with varying coflow oxygen concentrations

The elevated temperature of hydrogen combustion increases the formation of thermal NOx. Moderate or intense low oxygen dilution (MILD) combustion is known to reduce NOx emissions and increase thermal efficiency. Pressure is often also used for increasing thermal efficiency. The impact that pressure has on fluid dynamics and chemical kinetics is especially relevant in MILD combustion conditions. Hydrogen jet flames issuing into a hot and vitiated coflow were imaged using OH∗ chemiluminescence at different pressures (1–7 bar) and oxygen levels (3–9% by vol.). Laminar flame simulations complemented the experiments. The observed mean radial OH∗ width increased with increased pressure, but only at O2 content less than 9%, suggesting that pressure has greater influence on kinetics when oxygen is reduced. The integrated OH∗ signal strength remained constant at 3% coflow O2, despite an apparent increase in flame width, suggesting a spatial broadening of the flame with pressure. Numerical results indicate that at 3–6% O2, conditions for MILD combustion of H2 are met across a wide range of strains and pressures, supporting the experimental observations for 3% O2.

Mechanisms of air breakdowns for aluminum conductors steel reinforced wire-wire gaps induced by propane jet flames

In recent years, events of transmission line tripping caused by wildfires have occurred frequently in many mountain areas. In order to understand the mechanism of such events, an experimental system for simulating breakdown discharge processes of model wires induced by a simulated fire source is designed. In this experiment, propane jet flame (somewhat like eruptive fires in real wild fires) is used as a simulated flame source. By setting three different gas flow rates and five different wire-wire gaps, discharge breakdown behaviors and characteristics between two ACSR (aluminum conductors steel reinforced) segments in propane flames with three flame zones are explored. The influences of these different factors on the breakdown characteristics and mechanisms of wire-wire air gaps are analyzed. Results show that breakdown voltages increase almost linearly with the increase of wire-wire gaps in continuous zone of the flames for three gas flow rates. For a short wire-wire gap like 5.0 cm, the pilot discharge channels can be formed more easily, stably and completely under a lower voltage and weaker external ionization environment. The propane jet flame conditions are complex and affected by multiple actions like initial jet force (affected by gas flow rate), temperature distribution, soot particle, ion wind. As a result, local electric field will be distorted, a streamer channel is established dependently in different cases. These factors play individual, competitive or synergistic roles at the same time in breakdowns of real ACSR wire-wire gaps.

Numerical simulation of microwave-enhanced methane-air flames I: Modeling and one-dimensional premixed laminar flames

This study considers the numerical modeling of sub-breakdown microwave-enhanced combustion in one-dimensional laminar methane-air flames over a range of stoichiometries at standard temperature and pressure. The effect of uniform microwave fields on the laminar flame speed, post-flame temperature and composition is of primary interest here.A model for the non-thermal plasma generated by the microwave field in such flames is constructed based on fitting of electron properties and electron impact reaction rates to pre-computed solutions of the Boltzmann equation, combined with ambipolar diffusion transport for charged species. This model is used with a selection of existing and novel methane-air kinetics schemes to compute laminar flame profiles at a range of microwave field strengths and varying stoichiometry. The trends for laminar flame strength and temperature, peak- and outlet species fractions are compared over this parameter space. While all schemes demonstrate increasing flame speed and temperature with higher field strength and richer flames, significant variations in the form and magnitude of this increase are observed for the various kinetics schemes considered.The laminar flame-speed increase at sub-breakdown electric field strengths is found to be due primarily to electron heating of the gases in both the pre-heat region and across the flame. Analysis of the electron energy transfer shows that the majority of the energy at moderate field strengths is transferred via excitement of vibrational states of nitrogen, followed by water and carbon monoxide. In the pre-heat regime, vibrational excitement of methane is also seen to play an important role for fuel-rich flames. Particular attention is paid to the modification of the flame speed due to the relaxation time of the vibrational N2 states, which is demonstrated to result in a moderate reduction of the flame speed for lean and stoichiometric flames.A driving concern here is the development of models suitable for use within large-scale three-dimensional numerical simulations and the knowledge gained in comparing the existing schemes is used to derive a range of simplified mechanisms more suited for this purpose. The ability of these reduced models to capture the flame enhancement is demonstrated over a range of flame stoichiometries and sub-breakdown field strengths.

Experimental support for a new NOx formation route via an HNNO intermediate

Achieving minimal levels of nitrogen oxides (NOx) during combustion is a major constraint in the design of advanced high-efficiency engines. NOx can be formed during combustion of any fuel—including those without fuel-bound nitrogen—in air, where radicals can attack molecular nitrogen (N2) present in air to break the strong NN bond to ultimately form NOx. Paramount to the goal of minimizing NOx formation is knowledge of the fundamental routes by which the strong NN bond in N2 can be broken. Historically, there have been four known routes for breaking the strong NN bond in N2 to ultimately form NOx. We have recently posited that another route—mediated by an HNNO intermediate—may also play a role, particularly at the high pressures and low peak temperatures relevant to high-efficiency, low-NOx engines. Our previous theoretical and modeling studies show HNNO to be a major product of the N2O + H reaction at high pressures and low temperatures; once formed, HNNO is likely to react with radicals in barrierless reactions that would occur quickly and with high NOx yields. In the present paper, we report measurements of H2, O2, H2O, N2O, NO, NOx, and NH3 in jet-stirred reactor experiments for an H2/O2/N2O/NO/N2/Ar mixture that specifically target HNNO pathways. Importantly, we observe significant formation of NO and NH3—both of which provide signatures of the HNNO mechanism that are not predicted by previous models without it. Flame simulations using a new sub-model describing pressure-dependent formation and consumption of HNNO show these pathways to be among the most prominent formation routes at high pressures and low peak temperatures. However, exact quantification of the role of HNNO in NOx formation and quantitative predictions of NOx in general require more accurate rate constants for both HNNO pathways and mixture rules for pressure-dependent reactions.

A species-weighted flamelet/progress variable model with differential diffusion effects for oxy-fuel jet flames

A flamelet/progress variable (FPV) model accounting for the differential diffusion (DD) effects is proposed and applied to large eddy simulation (LES) of turbulent oxy-fuel flames (Sevault et al. 2012). Based on tabulations with the detailed molecular diffusion and the equal diffusion (ED) assumption, a species-weighted flamelet (SWF) model is developed to represent the DD effects. The influences of combustion progress, mixture fraction, and turbulence on variable Lewis numbers are incorporated into the model. The model assessments are first conducted on a laminar coflow flame, showing that the effect of DD on temperature and species is accurately captured, in that the importance of DD is seen in the laminar flame. Subsequently the model is implemented in a fully coupled LES of oxy-fuel jet flames. The LES results show that DD plays an important role in the reaction zone and regions near the fuel nozzle, and its effect decreases farther downstream, consistent with the experimental observations. The LES with the SWF model yields good predictions on the mean temperature and major species at the fuel-rich side while slight deviation (less than 6%) at the fuel-lean side. In comparison, LES with the original unity Lewis numbers flamelet model (Pierce et al. 2004) predicts a full extinction because of neglecting the DD characteristics, while the variable Lewis number flamelet model provides the maximum deviation of 26% because of ignoring the ED characteristics produced by turbulent disturbance. The trend and location of localized extinction of oxy-fuel flame are also well predicted using the SWF model.