LES/CMC Simulations of Swirl-Stabilised Ethanol Spray Flames Approaching Blow-Off

Large Eddy Simulations (LES) with the Conditional Moment Closure (CMC) combustion model of swirling ethanol spray flames have been performed in conditions close to blow-off for which a wide database of experimental measurements is available for both flame and spray characterization. The solution of CMC equations exploits a three-dimensional unstructured code with a first order closure for chemical source terms. It is shown that LES/CMC is able to properly capture the flame structure at different conditions and agrees reasonably well with the measurements both in terms of mean flame shape and dynamic behaviour of the flame evaluated in terms of local extinctions and statistics of the lift-off height. Experimental measurements of the overall (liquid plus gaseous) mixture fraction, performed using the Laser-Induced Breakdown Spectroscopy technique, are also included allowing further assessment and validation of the numerical method. The sensitivity of the simulation results to the various boundary conditions is discussed.     

Heat Transfer Modeling in the Context of Large Eddy Simulation of Premixed Combustion with Tabulated Chemistry

Tabulated chemistry models like the Flamelet Generated Manifolds method are a good approach to include detailed information on the reaction kinetics in a turbulent flame at reasonable computational costs. However, so far, not all information on e.g. heat losses are contained in these models. As those often appear in typical technical applications with enclosed flames in combustion chambers, extensions to the standard FGM approach will be presented in this paper, allowing for the representation of non-adiabatic boundaries. The enthalpy as additional control variable for the table access is introduced, such that the chemistry database becomes three-dimensional with mixture fraction, reaction progress variable and enthalpy describing the thermo-chemical state. The model presented here is first validated with a two-dimensional enclosed Bunsen flame and then applied within the Large Eddy Simulations of a turbulent premixed swirl flame with a water-cooled bluff body and a turbulent stratified flame, where additional modeling for the flame structure using artificially thickened flames was included. The results are encouraging, as the temperature decrease towards the bluff body in the swirl flame and the cooling of the pilot flame exhaust gases in the stratified configuration can be observed in both experiments and simulation.     

Modeling of Turbulent Natural Gas and Biogas Flames of the Delft Jet-in-Hot-Coflow Burner: Effects of Coflow Temperature,Fuel Temperature and Fuel Composition on the Flame Lift-Off Height

The structure of a turbulent non-premixed flame of a biogas fuel in a hot and diluted coflow mimicking moderate and intense low dilution (MILD) combustion is studied numerically. Biogas fuel is obtained by dilution of Dutch natural gas (DNG) with CO₂. The results of biogas combustion are compared with those of DNG combustion in the Delft Jet-in-Hot-Coflow (DJHC) burner. New experimental measurements of lift-off height and of velocity and temperature statistics have been made to provide a database for evaluating the capability of numerical methods in predicting the flame structure. Compared to the lift-off height of the DNG flame, addition of 30 % carbon dioxide to the fuel increases the lift-off height by less than 15 %. Numerical simulations are conducted by solving the RANS equations using Reynolds stress model (RSM) as turbulence model in combination with EDC (Eddy Dissipation Concept) and transported probability density function (PDF) as turbulence-chemistry interaction models. The DRM19 reduced mechanism is used as chemical kinetics with the EDC model. A tabulated chemistry model based on the Flamelet Generated Manifold (FGM) is adopted in the PDF method. The table describes a non-adiabatic three stream mixing problem between fuel, coflow and ambient air based on igniting counterflow diffusion flamelets. The results show that the EDC/DRM19 and PDF/FGM models predict the experimentally observed decreasing trend of lift-off height with increase of the coflow temperature. Although more detailed chemistry is used with EDC, the temperature fluctuations at the coflow inlet (approximately 100K) cannot be included resulting in a significant overprediction of the flame temperature. Only the PDF modeling results with temperature fluctuations predict the correct mean temperature profiles of the biogas case and compare well with the experimental temperature distributions.     

Effects of Spray and Turbulence Modelling on the Mixing and Combustion Characteristics of an n-heptane Spray Flame Simulated with Dynamic Adaptive Chemistry

Accurate modelling of spray combustion process is essential for efficiency improvement and emissions reduction in practical combustion engines. In this work, both unsteady Reynolds-averaged Navier-Stokes (URANS) simulations and large eddy simulations (LES) are performed to investigate the effects of spray and turbulence modelling on the mixing and combustion characteristics of an n-heptane spray flame in a constant volume chamber at realistic conditions. The non-reacting spray process is first simulated with URANS to investigate the effects of entrainment gas-jet model on the penetration characteristics and fuel vapor distributions. It is found that the droplet motion near the nozzle has significant influence on the fuel vapor distribution, while the liquid penetration length is controlled by the evaporation process and insensitive to gas-jet model. For the case considered, both URANS with the gas-jet model and large eddy simulations can properly predict the vapor penetration. For the combustion characteristics, it is found that LES yields better predictions in the global combustion characteristics. The URANS with gas jet model yields a comparable flame length and lift-off-length (LOL) to LES, but results in a larger ignition delay time compared to the experimental data. Another focus of this work is to qualify the convergence characteristics of the dynamic adaptive chemistry (DAC) method in these transient combustion simulations, where DAC is applied to reduce the mechanism locally and on-the-fly to accelerate chemistry calculations. The instantaneous flame structures and global combustion characteristics such as ignition delay time, flame lift-off length and emissions are compared between simulations with and without DAC. For URANS, good agreements are observed both on instantaneous flame structures and global characteristics. For LES, it is shown that the errors incurred by DAC are small for scatter distributions in composition space and global combustion characteristics, while they may significantly affect instantaneous flame structures in physical space. The study reveals that for DAC application in transient simulations, global or statistic information should be used to assess the accuracy, such as manifolds in composition space, conditional quantities and global combustion characteristics. For the cases investigated, a speed-up factor of more than two is achieved by DAC with a 92-species skeletal mechanism with less than 0.2 % and 3.0 % discrepancy in ignition delay and LOL, respectively.     

A 5-D Implementation of FGM for the Large Eddy Simulation of a Stratified Swirled Flame with Heat Loss in a Gas Turbine Combustor

Numerical simulations are foreseen to provide a tremendous increase in gas-turbine burners efficiency in the near future. Modern developments in numerical schemes, turbulence models and the consistent increase of computing power allow Large Eddy Simulation (LES) to be applied to real cold flow industrial applications. However, the detailed simulation of the gas-turbine combustion process remains still prohibited because of its enormous computational cost. Several numerical models have been developed in order to reduce the costs of flame simulations for engineering applications. In this paper, the Flamelet-Generated Manifold (FGM) chemistry reduction technique is implemented and progressively extended for the inclusion of all the combustion features that are typically observed in stationary gas-turbine combustion. These consist of stratification effects, heat loss and turbulence. Three control variables are included for the chemistry representation: the reaction evolution is described by the reaction progress variable, the heat loss is described by the enthalpy and the stratification effect is expressed by the mixture fraction. The interaction between chemistry and turbulence is considered through a presumed beta-shaped probability density function (PDF) approach, which is considered for progress variable and mixture fraction, finally attaining a 5-D manifold. The application of FGM in combination with heat loss, fuel stratification and turbulence has never been studied in literature. To this aim, a highly turbulent and swirling flame in a gas turbine combustor is computed by means of the present 5-D FGM implementation coupled to an LES turbulence model, and the results are compared with experimental data. In general, the model gives a rather good agreement with experimental data. It is shown that the inclusion of heat loss strongly enhances the temperature predictions in the whole burner and leads to greatly improved NO predictions. The use of FGM as a combustion model shows that combustion features at gas turbine conditions can be satisfactorily reproduced with a reasonable computational effort. The implemented combustion model retains most of the physical accuracy of a detailed simulation while drastically reducing its computational time, paving the way for new developments of alternative fuel usage in a cleaner and more efficient combustion.     

A Comparison of the Blow-Off Behaviour of Swirl-Stabilized Premixed, Non-Premixed and Spray Flames

Confined short turbulent swirling premixed and non-premixed methane and heptane spray flames stabilized on an axisymmetric bluff body in a square enclosure have been examined close to the blow-off limit and during the extinction transient with OH* chemiluminescence and OH-PLIF operated at 5 kHz. The comparison of flames of different canonical types in the same basic aerodynamic field allows insights on the relative blow-off behaviour. The flame structure has been examined for conditions increasingly closer to blow-off. The premixed flame was seen to change from a cylindrical shape at stable burning condtions, with the flame brush closing across the flow at conditions close to blow-off. The PLIF images show that for the gaseous non-premixed flame, holes appear along the flame sheet with increasing frequency as the blow-off condition is approached, while the trend is less obvious for the spray flame. Non-premixed and spray flames showed randomly-occurring lift-off, which is further evidence of localised extinction. The mean lift-off height increased with increasing fuel jet velocity and decreased with increasing air velocity and approaches zero (i.e. the flame is virtually attached) just before the blow-off condition, despite the fact that more holes were evident in the flame sheet as extinction was approached. It was found that the average duration of the blow-off event, when normalised with the characteristic flow time d/U _b (d being the bluff-body diameter and U _b the bulk velocity) was in the range 9–38 with the spray flame extinction lasting a shorter time than the gaseous flames. Finally, it was found that correlations based on a Damköhler number collapse the blow-off velocity data for all flames with reasonable accuracy. The results can help the development of advanced turbulent combustion models.     

Large-Eddy Simulation of Turbulent Flow Around a Finite-Height Wall-Mounted Square Cylinder Within a Thin Boundary Layer

The paper describes the results of a computational study of the auto-ignition of a fuel spray under Exhaust Gas Recirculation (EGR) conditions, a technique used to reduce the production of NO_x. Large Eddy Simulation (LES) is performed, and the stochastic field method is used for the solution of the joint sub-grid probability density function (pdf) of the chemical species and energy. The fuel spray is n-heptane, a diesel surrogate and its chemical kinetics are described by a reduced mechanism involving 22 species and 18 reaction steps. The method is applied to a constant volume combustion vessel able to reproduce EGR conditions by the ignition of a hot gas mixture previously introduced into the chamber. Once the prescribed conditions are reached the fuel is then injected. Different EGR conditions in terms of temperature and initial ambient chemical composition are simulated. The results are in good overall agreement with measurements both regarding the ignition delay times and the lift-off heights.     

Large Eddy Simulation of Stratified and Sheared Flames of a Premixed Turbulent Stratified Flame Burner Using a Flamelet Model with Heat Loss

This paper presents large eddy simulations (LES) of the Darmstadt turbulent stratified flame burner (TSF) at different operating conditions including detailed heat loss modeling. The target cases are a non-reacting and two reacting cases. Both reacting cases are characterized by stratification, while one flame additionally features shear. In the regime diagram for premixed combustion, the studied flames are found at the border separating the thin reaction zones regime and the broken reaction zones regime. A coupled level set/progress variable model is utilized to describe the combustion process. To account for heat loss, an enthalpy defect approach is adopted and reformulated to include differential diffusion effects. A novel power-law rescaling methodology is proposed to integrate the enthalpy defect approach into the level set/progress variable model which is extensively validated in two validation scenarios. It is demonstrated that the LES with the newly developed model captures the influence of heat loss well and that the incorporation of heat loss effects improves the predictions of the TSF-burner over adiabatic simulations, while reproducing the experimentally observed flame lift-off from the pilot nozzle.     

Autoignition of n-decane Droplets in the Low-, Intermediate-, and High-temperature Regimes from a Mixture Fraction Viewpoint

Detailed numerical simulations of isolated n-decane droplets autoignition are presented for different values of the ambient pressure and temperature. The ignition modes considered included single-stage ignition, two-stage ignition and cool-flame ignition. The analysis was conducted from a mixture fraction perspective. Two characteristic chemical time scales were identified for two-stage ignition: one for cool-flame ignition, and another for hot-flame ignition. The appearance and subsequent spatial propagation of a cool flame at lean compositions was found to play an important role in the ignition process, since it created the conditions for activating the high-temperature reactions pathway in regions with locally rich composition. Single-stage ignition was characterized by a single chemical time scale, corresponding to hot-flame ignition. Low-temperature reactions were negligible for this case, and spatial diffusion of heat and chemical species mainly affected the duration of the ignition transient, but not the location in mixture fraction space at which ignition first occurs. Finally, ignition of several cool flames of decreasing strength was observed in the cool-flame ignition case, which eventually lead to a plateau in the maximum gas-phase temperature. The first cool flame ignited in a region where the fuel / air mixture was locally lean, whereas ignition of the remaining cool flames occurred at rich mixture compositions.     

Combustion Modeling Including Heat Loss Using Flamelet Generated Manifolds: A Validation Study in OpenFOAM

In numerical combustion applications the Flamelet Generated Manifolds technique (FGM) is being used at an increasingly number of occasions. This technique is an approach to reduce the chemistry efficiently and accurately. In the present work FGM is coupled to an OpenFOAM-based CFD solver. The multidimensional flame is described by an ensemble of 1D laminar flames generated through a 1D detailed chemistry solver, by taking into account both convective and diffusive contributions as well as the required source terms. The flame structure is parameterized as function of a progress variable and few controlling variables such as the variance of the progress variable and the enthalpy. A manifold, which collects the 1D flame properties, is built from the 1D flame solutions. For the progress variable and each controlling variable, a transport equation is added to the standard flow conservation equations. During runtime, key quantities are retrieved from the manifold by interpolation. The resulting FGM-CFD coupled code has two significant features: the ability to treat heat loss effects and the adoption of turbulence level to describe the flame structure, providing high quality numerical results within practical industrial configurations. In the present work, a backward-facing step configuration with a methane/air mixture is investigated. Some key aspects of reactive phenomena in standard industrial burner configurations, such as the recirculation region development and the flame stabilization, are considered here. Numerical simulations are performed comparing results with experiments available in literature (Banhawy et al. Combust. Flame 50:153–165, 18). Both RANS and LES approaches are adopted: improvements with respect to prior available works are highlighted. Moreover, LES data, available for the first time within this configuration, are used to provide a deeper insight of turbulence/combustion interaction.     

Study of a Filtered Flamelet Formulation for Large Eddy Simulation of Premixed Turbulent Flames

Despite significant advances in the understanding and modelling of turbulent combustion, no general model has been proposed for simulating flames in industrial combustion devices. Recently, the increase in computational possibilities has raised the hope of directly solving the large turbulent scales using large eddy simulation (LES) and capturing the important time-dependant phenomena. However, the chemical reactions involved in combustion occur at very small scales and the modelling of turbulent combustion processes is still required within the LES framework. In the present paper, a recently presented model for the LES of turbulent premixed flames is presented, analysed and discussed. The flamelet hypothesis is used to derive a filtered source term for the filtered progress variable equation. The model ensures proper flame propagation. The effect of subgrid scale (SGS) turbulence on the flame is modelled through the flame-wrinkling factor. The present modelling of the source term is successfully tested against filtered direct numerical simulation (DNS) data of a V-shape flame. Further, a premixed turbulent flame, stabilised behind an expansion, is simulated. The predictions agree well with the available experimental data, showing the capabilities of the model for performing accurate simulations of unsteady premixed flames.     

Microgravity experiments and numerical studies on ethanol/air spray flames

Spray flames are known to exhibit amazing features in comparison with single-phase flames. The weightless situation offers the conditions in which the spray characteristics can be well controlled before and during combustion. The article reports on a joint experimental/numerical work that concerns ethanol/air spray flames observed in a spherical chamber using the condensation technique of expansion cooling (based on the Wilson cloud chamber principle), under microgravity.We describe the experimental set-up and give details on the creation of a homogeneous and nearly monosized aerosol. Different optical diagnostics are employed successfully to measure the relevant parameters of two-phase combustion. A classical shadowgraphy system is used to track the flame speed propagation and allow us to observe the flame front instability. The complete characterization of the aerosol is performed with a laser diffraction particle size analyser by measuring the droplet diameter and the droplet density number, just before ignition. A laser tomography device allows us to measure the temporal evolution of the droplet displacement during flame propagation, as well as to identify the presence of droplets in the burnt gases. The numerical modelling is briefly recalled. In particular, spray-flame propagation is schematized by the combustion spread in a 2-D lattice of fuel droplets surrounded by an initial gaseous mixture of fuel vapour and air.In its spherical expansion, the spray flame presents a corrugated front pattern, while the equivalent single-phase flame does not. From a numerical point of view, the same phenomena of wrinkles are also observed in the simulations. The front pattern pointed out by the numerical approach is identified as of Darrieus–Landau (DL) type. The droplets are found to trigger the instability. Then, we quantitatively compare experimental data with numerical predictions on spray-flame speed. The experimental results show that the spray-flame speed is of the same order of magnitude as that of the single-phase premixed flame. On the other hand, the numerical results exhibit the role played by the droplet radius in spray-flame propagation, and retrieve the experiments only when the droplets are small enough and when the Darrieus–Landau instability is triggered. A final discussion is developed to interpret the various patterns experimentally observed for the spray-flame front.     

A-Priori Validation of Scalar Dissipation Rate Models for Turbulent Non-Premixed Flames

The modelling of scalar dissipation rate in conditional methods for large-eddy simulations is investigated based on a priori direct numerical simulation analysis using a dataset representing an igniting non-premixed planar jet flame. The main objective is to provide a comprehensive assessment of models typically used for large-eddy simulations of non-premixed turbulent flames with the Conditional Moment Closure combustion model. The linear relaxation model gives a good estimate of the Favre-filtered scalar dissipation rate throughout the ignition with a value of the related constant close to the one deduced from theoretical arguments. Such value of the constant is one order of magnitude higher than typical values used in Reynolds-averaged approaches. The amplitude mapping closure model provides a satisfactory estimate of the conditionally filtered scalar dissipation rate even in flows characterised by shear driven turbulence and strong density variation.

     

Large eddy simulation of turbulent premixed and stratified combustion using flame surface density model coupled with tabulation method

Large eddy simulations (LESs) are performed to investigate the Cambridge premixed and stratified flames, SwB1 and SwB5, respectively. The flame surface density (FSD) model incorporated with two different wrinkling factor models, i.e., the Muppala and Charlette2 wrinkling factor models, is used to describe combustion/turbulence interaction, and the flamelet generated manifolds (FGM) method is employed to determine major scalars. This coupled sub-grid scale (SGS) combustion model is named as the FSD-FGM model. The FGM method can provide the detailed species in the flame which cannot be obtained from the origin FSD model. The LES results show that the FSD-FGM model has the ability of describing flame propagation, especially for stratified flames. The Charlette2 wrinkling factor model performs better than the Muppala wrinkling factor model in predicting the flame surface area change by the turbulence. The combustion characteristics are analyzed in detail by the flame index and probability distributions of the equivalence ratio and the orientation angle, which confirms that for the investigated stratified flame, the dominant combustion modes in the upstream and downstream regions are the premixed mode and the back-supported mode, respectively.     

Direct Numerical Simulations of Premixed Turbulent Flames with Reduced Chemistry: Validation and Flamelet Analysis

Direct numerical simulation is a very powerful tool to evaluate the validity of new models and theories for turbulent combustion. In this paper, direct numerical simulations of spherically expanding premixed turbulent flames in the thin reaction zone regime and in the broken reaction zone regime are performed. The flamelet-generated manifold method is used in order to deal with detailed reaction kinetics. The computational results are analyzed by using an extended flame stretch theory. It is investigated whether this theory is able to describe the influence of flame stretch and curvature on the local burning velocity of the flame. It is found that if the full profiles of flame stretch and curvature through the flame front are included in the theory, the local mass burning rate is well predicted. The influence of several approximations, which are used in other existing theories, is studied. When flame stretch is assumed constant through the flame front or when curvature of the flame front is neglected, the theory fails to predict the local mass burning rate. The influence of using a reduced chemistry model is investigated by comparing flamelet simulations with reduced and detailed chemistry.     

Flow Physics of a Bluff-Body Swirl Stabilized Flame and their Prediction by Means of a Joint Eulerian Stochastic Field and Tabulated Chemistry Approach

In the frame of this work a transported joint scalar probability density function (PDF) method is combined with the flamelet generated manifolds (FGM) tabulated chemistry approach for large eddy simulation (LES) modeling of a three-dimensional turbulent premixed swirl burner. This strategy accounts for the turbulence-chemistry interaction at reasonable computational costs. At the same time, it allows the usage of detailed chemistry mechanisms for the creation of the chemical database. The simulation results obtained are comparatively assessed along with complementary measurements. Furthermore, transient and time-averaged data are used to provide insight into the flow physics of the bluff-body swirl stabilized flame considered. The sensitivity of the results to different modeling approaches regarding the predicted flame shape and its dynamics is also investigated, where the implemented approach is compared with the well-established artificially thickened flame (ATF) combustion model. Consequently, the investigation conducted in this work aims to provide a complete picture on the ability of the proposed combustion model to reproduce the flow conditions within complex bluff-body swirl stabilized flames.     

Large Eddy Simulation of Reactive Two-Phase Flow in an Aeronautical Multipoint Burner

Because of compressibility criteria, fuel used in aeronautical combustors is liquid. Their numerical simulation therefore requires the modeling of two-phase flames, involving key phenomena such as injection, atomization, polydispersion, drag, evaporation and turbulent combustion. In the present work, particular modeling efforts have been made on spray injection and evaporation, and their coupling to turbulent combustion models in the Large Eddy Simulation (LES) approach. The model developed for fuel injection is validated against measurements in a non-evaporating spray in a quiescent atmosphere, while the evaporation model accuracy is discussed from results obtained in the case of evaporating isolated droplets. These models are finally used in reacting LES of a multipoint burner in take-off conditions, showing the complex two-phase flame structure.     

3D Numerical Simulation of a Laminar Experimental SWQ Burner with Tabulated Chemistry

Flame-wall interaction (FWI) plays an important role in enclosed combustion systems. For avoiding the complexity of close to reality combustors, in this study an atmospheric premixed V-shaped flame interacting with an isothermal cold wall in a side wall quenching (SWQ) configuration is investigated. A stoichiometric methane/air mixture is used as fuel. A three-dimensional (3D) numerical simulation, which resolves all flow structures is combined with a tabulated chemistry approach (flamelet generated manifold, FGM). Results are compared with experimental data and two-dimensional simulations. The FGM approach is a suitable trade-off between computationally expensive detailed chemistry simulations and over simplified single step mechanisms. 2D simulations are used to investigate the influence of the uncertainty of the wall temperature, to show that the resolution in 3D is sufficient and that the influence of the flame thickening on the wall heat fluxes can be determined. Our results show that the 3D FGM approach is in close agreement to experimentally obtained flow and temperature fields. The dimensionless wall heat flux and Péclet number matches the expected values of 0.16 and 7, respectively. However, during FWI the measured CO mole fractions are not reproduced accurately showing that the transported variables in the present approach of tabulated chemistry do not recover premixed flame structures near walls.