Search

Recent investigations on a novel multi-regime burner (MRB) configuration showed significant deviations in the CO flame structure compared to the limiting cases of premixed and non-premixed flames. However, a prior analysis revealed that major species and temperature are captured by both limiting cases (Butz et al., Combust. Flame, 2019 [1]). In the present work, large eddy simulations using an artificial thickened flame approach and tabulated chemistry are performed for the MRB configuration. Simulation results are compared to experimental Raman/Rayleigh/CO-LIF and PIV measurements, confirming the applicability of the modeling approach. Further, simulation results are consistent with the aforementioned prior analysis. Special attention is paid to predicting CO by analyzing the conditional flame structure and the effects of local residence time. This combined CO and residence time analysis reveals that local convective and diffusive transport processes should be resolved simultaneously with the unsteady flow, instead of being tabulated. Substantial improvements in CO are achieved when local transport is considered.

Gradient free regime identification (GFRI) is applied to 1D Raman/Rayleigh/LIF measurements of temperature and major species from the intermediate velocity case of the Sydney piloted inhomogeneous jet flame series to better understand the structure of reaction zones and the downstream evolution of multi-regime characteristics. The GFRI approach allows local reaction zones to be detected and characterized as premixed, dominantly premixed, multi-regime, dominantly non-premixed, or non-premixed flame structures, based on flame markers (mixture fraction, chemical mode, and heat release rate) derived from the experimental data. The statistics of chemical mode zero-crossings, which mark premixed reaction zones, and the relative populations of flame structures are shown to be sensitive to the state of mixing in the near field of the flame and to the level of local extinction farther downstream. Multi-regime structures, where premixed and non-premixed reaction zones occur in close proximity and both contribute to overall heat release, account for nearly half the total population at streamwise locations within the first several jet diameters. There is a rapid transition within the near field whereby the relative population of non-premixed and dominantly non-premixed structures grows from 0.05 to nearly 0.5, and the population of premixed and dominantly premixed structures decreases correspondingly as fluid entering the reaction zone becomes progressively fuel-rich. Local extinction and re-ignition bring a resurgence in premixed-type structures, many of which occur at fuel-lean conditions. There are also modest populations of multi-regime structures, having chemical mode zero-crossings at lean conditions, which would not exist in a fully burning jet flame.

In this work, the flame structures of the recent Darmstadt turbulent premixed/stratified flame series with 20% volume fraction H2 addition (Schneider et al., PCI, 2019) are analyzed using flamelet tabulated manifolds. Three different methane/hydrogen flames (MHFs) with increasing complexity are studied. The effects of differential diffusion, stretch and stratification on the applicability of the flamelet model are investigated in detail. Specifically, to investigate the differential diffusion effects, three different modeling approaches are considered, in which the diffusion fluxes are calculated using the multi-component (MC), mixture-averaged (MA) and unity Lewis number (Le1) approaches. To investigate stretch effects, a strained premixed flamelet (SPF) model is proposed, in which the flame’s internal response to stretch is characterized with an additional tabulation coordinate. A double-conditioning analysis is conducted for different flame series. The dataset is conditioned on both the local equivalence ratio and local stratification to analyze the coupling effects of stratification and differential diffusion on the flame structure. Overall, the flamelet predictions are consistent with the experimental findings for all flames.

International audience; Interactions between premixed and non-premixed reaction zones can lead to complex mixed combustion regimes, here denoted as multi-regime combustion, which pose challenges to many conventional combustion modeling approaches. Such conditions occur in most practical combustors and can originate from partial premixing, mixture inhomogeneities/stratification, hot product recirculation, or local flame extinction and re-ignition. Therefore, novel equations are derived for modeling multi-regime combustion which are formulated with respect to a twodimensional composition space spanned by mixture fraction and reaction progress variable. Contrary to previous works, the dependency of the progress variable on the mixture fraction is considered in the new model. This is achieved by splitting the progress variable gradient into an aligned and an orthogonal component with respect to the mixture fraction gradient and the latter is used to define the second coordinate. In the theory that follows, a balance equation for the progress variable on mixture fraction iso-surfaces is formulated. Using this balance equation together with the orthogonal coordinate system, the transformation of species and temperature equations to the 2D composition space yields a novel set of equation without so-called cross-terms. This is advantageous since cross-terms obtained with previous approaches lack a general closure and it is uncertain if it exists at all. Furthermore, the approach allows to naturally distinguish between non-premixed and premixed combustion regimes, auto ignition, and it covers multi-regime combustion characteristics. The theory is validated and discussed by means of a fully resolved solution of a laminar triple flame using detailed chemistry. At first, regions which exhibit premixed, non-premixed or multi-regime combustion characteristics are identified. The triple flame solution then serves as a database from which all relevant theoretical relations are post-processed and validated. In comparison to budgets of conventional 1D flamelet equations for premixed and non-premixed combustion it is shown that only the full set of transport terms considered in the 2D equations accurately balances chemical source terms everywhere in the triple flame, especially in regions where

To study combustion fundamentals of complex fuels under well-defined boundary conditions, a novel Temperature Controlled Jet Burner (TCJB) system is designed that can stabilise both gaseous or pre-vaporised liquid fuels. In a first experimental exploratory study, piloted turbulent jet flames of pre-vaporised methanol, ethanol, 2-propanol and 2-butanol mixtures are compared to methane/air as a reference fuel. Complementary one-dimensional laminar flame calculations are used to provide flame parameters for comparison. Blow-off and flame length as global flame characteristics are measured over a wide range of equivalence ratios. For fuel rich conditions, blow-off limits correlate well with extinction strain rate calculations. Differing flame lengths from lean to rich conditions are explained partly by different flame wrinkling that is assessed using planar laser-induced fluorescence imaging of the hydroxyl radical (OH-PLIF). A study of Lewis-number effects indicates that they have substantial influence on flame wrinkling. Lean alcohol/air flames, opposed to methane/air, have a Lewis-number greater than unity. This impedes curvature development, which promotes relatively large flame lengths. In contrast, across stoichiometric conditions, all alcohol/air mixture Lewis-numbers decrease significantly. At such conditions, alcohol/air flames show alike or even larger wrinkling compared to methane/air flames. However, quantitatively, the differences in flame length and wrinkling observed among the flames can neither be explained alone by Lewis-number differences, nor other global mixture parameters available from 1D laminar flame calculations. This study shall therefore emphasise the need for more detailed experimental analyses of the full thermochemical state of laminar and turbulent flames fuelled with complex fuels.

In practical applications, partial premixing of fuel and oxidizer, as well as recirculation of combustion products, result in complex combustion scenarios where multi-regime effects arise and a numerical representation of local reaction zones by purely premixed or purely non-premixed flame structures may not hold. Here, a novel burner system is introduced to investigate the fundamental characteristics of multi-regime combustion and to provide a basis for validating numerical models. This multi-regime burner (MRB) is specifically designed to produce flames with multi-regime characteristics while maintaining well-defined boundary conditions. Thermochemical data from Raman/Rayleigh/CO-LIF scattering experiments are provided for two selected operating conditions. The experimental investigation focuses on the overall flame structure by examining radial profiles of temperature and mixture fraction, as well as scatter plots of temperature, CH4, and CO versus mixture fraction. In order to assess the relative importance of different flame regimes, the gradient-free regime identification (GFRI) approach is extended to allow for an automated classification of local reaction zone structures. Classification criteria are defined, based on the ratio of local heat release rate peaks associated with premixed and non-premixed reaction zones located in close spatial proximity, and an automated process is implemented to classify 1D Raman/Rayleigh sample lines as premixed, dominantly premixed, multi-regime, dominantly non-premixed, or non-premixed flame zones. The importance of different flame zones, indicated by their population fractions, are found to evolve with downstream distance and show distinct differences between the two selected flames. Further, a prior analysis is used to test the applicability of 1D flame structure assumptions for the underlying combustion regime.

Obtaining information about burning characteristics and flame structures by analyzing experimental data is an important issue for understanding combustion processes and pursuing combustion modeling approaches. It has been shown that Raman/Rayleigh measurements of major species and temperature can be used to approximate the local heat release rate and the chemical explosive mode, and that these results are sufficiently accurate for a qualitative assessment of the relative importance of different heat release zones within the same overall flame structure in laminar and mildly turbulent partially premixed flames [1] , [2] . The present study uses data from direct numerical simulation (DNS) to extend and quantify the understanding of the approximation method with respect to premixed and stratified-premixed flames with significant turbulence–chemistry interaction (high Karlovitz number). The accuracy of the approximation procedure is assessed as previously applied, using just major species and temperature, as well as with the OH radical included as an additional experimentally accessible species. The accuracy of the local chemical explosive mode and the local heat release rate results from the approximation are significantly improved with OH included, yielding quantitative agreement with the DNS results. Further, a global sensitivity analysis is applied to identify the sensitivity of the heat release rate and chemical explosive mode to experimental uncertainties imprinted upon the DNS data prior to the approximation procedure.

Premixed flame characteristics play a crucial role in most advanced combustion applications, such as gas turbines, aircraft combustors or internal combustion engines. The laminar burning velocity is one key parameter to determine the stabilisation and propagation of premixed flames. This study examined the influence of high pressures on the laminar burning velocity of biofuels and the correlation between laminar burning velocity and pressure. The investigated biofuels were pure ethanol as the state-of-the-art surrogate, pure iso-butanol as a possible alternative and several blends of ethanol/iso-octane. Due to the usage of two methods to measure the laminar burning velocity – the Heat Flux burner and the closed combustion vessel – a comparison was added for atmospheric conditions. The measured laminar burning velocities were conducted for equivalence ratios from 0.7 to 1.3, a pressure range from 1 bar(a) to 15 bar(a) and 373 K. The results were further compared to existing numerical mechanisms and literature data. The numerical mechanisms reveal prediction performances in the range of 0.2 cm/s (0.4%) and 5.2 cm/s (16.2%) for dedicated mechanisms and equivalence ratios between 0.9 and 1.1. The two methods show deviations of the laminar burning velocities between 2.3 cm/s (4.9%) and 6.7 cm/s (11.4%).

Understanding and quantifying the relative importance of premixed and non-premixed reaction zones within turbulent partially premixed flames is an important issue for multi-regime combustion. In the present work, the recently-developed method of gradient-free regime identification (GFRI) is applied to instantaneous 1D Raman/Rayleigh measurements of temperature and major species from two turbulent lifted methane/air flames. Local premixed and non-premixed reaction zones are identified using criteria based on the mixture fraction, the chemical explosive mode, and the heat release rate, the latter two being calculated from an approximation of the full thermochemical state of each measured sample. A chemical mode (CM) zero-crossing is a previously documented marker for a premixed reaction zone. Results from the lifted flames show strong correlations among the mixture fraction at the CM zero-crossing, the magnitude of the change in CM at the zero-crossing, and the local heat release rate at the CM zero-crossing compared to the maximum heat release rate. The trends are confirmed through a comparable analysis of numerical simulations of two laminar triple flames. These newly documented trends are associated with the transition from dominantly premixed flame structures to dominantly non-premixed flames structures. The methods introduced for assessing the relative importance of local premixed and non-premixed reactions zones have potential for application to a broad range of turbulent flames.

International audience; A combustion regime identification based on convolutional neural networks (CNNs) is developed using the recently proposed gradient-free regime identification (GFRI) approach applied to two turbulent CH 4 /air jet flames featuring multi-regime characteristics. The training and the subsequent application of the CNN rely on the processing of one-dimensional Raman/Rayleigh line measurements of species mass fractions and temperature (CNN input). The combustion regime index is then readily predicted at every point along the measured line (CNN output). For training the neural network, the combustion regime index is first determined using the GFRI method (Hartl et al., 2018) based on the chemical explosive mode analysis (CEMA). Six classes of combustion regimes, including premixed (P), dominantly premixed (DP), multi-regime (MR), dominantly non-premixed (DNP), non-premixed (NP), and lean back-supported (LBS), are well detected by the trained CNN, with a pixel-wise accuracy of more than 85% for burner operating conditions unseen during training (different free-stream equivalence ratios). The quasi instantaneous neural network response provides a perspective towards real-time global combustion regime identification for pollutants emission control. From the results, it is also concluded that introducing physical insight, by combining advanced experimental (Raman/Rayleigh line measurements) and numerical analysis (GFRI), allows for reducing the amount of data needed to train neural networks.

Laminar dimethyl ether (DME) coflow jet flames, defined by different levels of partial premixing in the fuel stream, provide canonical test cases for studying conditions in between perfectly premixed and non-premixed characteristics. This work investigates the structure of DME/air flames and the evaluation of suitable flamelet modeling approaches. Three laminar, partially premixed DME/air flames are studied using fully resolved numerical simulations and the flamelet/progress variable (FPV) approach. The flame structure of the partially premixed DME/air flames is discussed and selected slices are compared to each other as well as to Raman/Rayleigh data. Overall, the experimental data is well predicted and the local flame structure is represented by the fully resolved simulations. In the context of the FPV approach, an a priori analysis of the underlying tabulated manifold is carried out. Lookup tables based on strained counterflow flames were identified as the most suitable candidate for the fully coupled FPV simulations of all three partially premixed DME/air flames. The FPV results are further compared a posteriori to the fully resolved simulation data. The flame structure of the partially premixed DME/air flames was reproduced and good results were obtained for the temperature and the species mass fractions. This numerical investigation contributes to the understanding of local flame structures in partially premixed DME/air flames and has the potential to support model selection in complex combustion processes.

Current methods for combustion regime characterization, such as the flame index, rely on 3D gradient information that is not accessible with available experimental techniques. Here, a method is proposed for reaction zone detection and characterization, which can be applied to instantaneous 1D Raman/Rayleigh line measurements of major species and temperature as well as to results of laminar and turbulent flame simulations, without the need for 3D gradient information. Several derived flame markers, namely the mixture fraction, the heat release rate, and the chemical explosive mode, are combined to detect and characterize premixed versus non-premixed reaction zones. The methodology is developed and evaluated using fully resolved simulation data from laminar flames. The fully resolved 1D simulation data are spatially filtered to account for the difference in spatial resolution between experiment and simulation. Then, starting from just temperature and major species, a constrained homogeneous constant pressure, constant temperature reactor calculation gives an approximation of the full thermochemical state at each sample location along the line. Finally, the chemical explosive mode and the heat release rate are calculated from this approximated state and compared to those calculated directly from the simulation data. As a further test, experimental uncertainty is superimposed onto the filtered numerical data to produce a Raman/Rayleigh equivalent state before running the constrained homogeneous reactor, and results are again compared. After successful tests using the numerical data, the approach is applied to Raman/Rayleigh line measurements from laminar counterflow flames and a mildly turbulent lifted flame. The results confirm that the reaction zones can be reliably detected and characterized using experimental data. Furthermore, the relative importance of premixed and non-premixed reaction zones within the same flame can be qualitatively assessed as demonstrated in the results.

Partial oxidation (POX) reactors are characterized by different reaction regimes: the near burner zone is mainly governed by the mixing of the initially non-premixed reactants and by the fast oxidation zone, while the post-flame zone is mainly governed by the slow reforming reactions under almost fully premixed conditions. Advanced turbulence–chemistry interaction models are required to describe different regimes in POX reactors. This work investigates the use of the flamelet/progress variable (FPV) approach in laminar and turbulent POX reactors. The FPV approach allows different times and length scales to be captured, using a progress variable to describe the advancement of the reactions in reacting mixtures. The solution of the chemical reactions is generally separated from the solution of the flow and mixing by pre-calculating solutions of one-dimensional canonical flames. The solution of the chemistry is stored in look-up tables and linked to the solution of the flow using the mixture fraction and the progress variable scalars. In this work, different methods for generating the flamelet look-up tables are analyzed for a laboratory reference laminar POX flame. The results are compared to the direct numerical simulation with the goal of identifying the most suitable flame structure for tabulation in POX systems. These results are then used to build a FPV table. Finally, a pilot-scale high-pressure POX reactor was investigated and the numerical results are validated using the measured composition of the syngas at the exit of the reactor.

In the present study the flame structure of a piloted partially-premixed dimethyl ether flame (DME-D), which is based on the Sydney/Sandia piloted jet burner flame series, is investigated using an LES-flamelet-progress variable approach (LES-FPV). Simulation results are used together with a comprehensive experimental data set including multi-scalar measurements of temperature and major species from Raman/Rayleigh scattering, intermediate species CH2O and OH from laser induced fluorescence (LIF) and velocity data from particle image velocimetry (PIV). The comparison between numerical and experimental data includes the mean and the root mean square (RMS) radial profiles for velocity, mixture fraction, temperature and mole fractions of major species at different downstream locations. Furthermore, species distributions conditioned on the experimentally accessible mixture fraction are compared and differences between DME and methane flames are discussed. In addition to this comparison, the computation of CH2O-LIF and OH-LIF signals as well as the effective Rayleigh cross-section was incorporated into the flamelet-progress variable approach. The filtered and time-averaged numerical results are then directly compared with the corresponding experimental signals at different axial positions. The characteristic separation of instantaneous CH2O and OH fields, which was previously observed in simultaneous LIF measurements, is discussed and analyzed based on the underlying flamelet structures. Finally, modeling assumptions from the experimental post-processing for the effective Rayleigh cross-section, which were introduced to account for the experimentally inaccessible intermediate hydrocarbons, are evaluated using the detailed species composition from the numerical simulations.

Flame structure look-up table generating strategies, e.g. the flamelet-progress variable approach, are based on the definition of a suitable progress variable. This variable is usually obtained from a linear combination of the species’ mass fractions contained in the system. The most important issue is the unique mapping between the progress variable and the independent variable, e.g. space or time. This is ensured by claiming the monotonicity of the progress variable. For simple fuel-oxidiser compositions, this can be performed by analysing the flame structure. However, in the case of complex chemical systems, finding such a progress variable is a non-trivial task, since a lack of monotonicity in the main species mass fractions often exists. In this article it is investigated how a valid progress variable can be found by an automated procedure. Some recent investigations use the fact that finding a monotonous progress variable is equivalent to solving a constrained optimisation problem. The new approach is to construct an algorithm to find a valid progress variable, which is also optimal in the sense that the gradients of the tabulated quantities are minimized. This leads to smoother data and reduces the interpolation effort when reading from the table. The method solves a nonlinear optimal control problem, where the monotonicity of the progress variable is ensured as an inequality constraint by using a genetic algorithm. Finally the performance of the new algorithm is evaluated for homogeneous reactor calculations and laminar diffusion flamelet look-up tables.

The substitution of gasoline with bio-ethanol is an intended way to reduce the climate impact of the traffic sector. To extend the knowledge of fundamental flame properties of ethanol/iso-octane flames and improve the numerical predictions for the effect of ethanol blending in internal combustion engines, measurements of the laminar burning velocity of established blend ratios of ethanol and iso-octane were carried out and compared to existing numerical mechanisms. The measurements were carried out with the Heat Flux burner, with thermocouples of type E at the burner plate, which was adapted with an evaporation unit based on direct vaporization to investigate the liquid fuels. The preheating temperature ranges from 298 K to 373 K and the pressure is atmospheric. First measurements of the laminar burning velocity of ethanol/air and iso-octane/air flames were carried out for validating the system. A good agreement with available literature data could be achieved for the investigated equivalence ratios from 0.7 to 1.4. Furthermore laminar burning velocities of different iso-octane/ethanol/air blends, namely E10, E24, E40 and E85, are presented. Through variation of the preheating temperatures (298 K, 323 K, 348 K and 373 K) the temperature dependency could be analyzed. The uncertainty analysis of the measurements has been revealed. Numerical simulations were carried out using different chemical mechanisms for ethanol/air flames, iso-octane/air flames as well as various fuel blends and preheating temperatures and are compared to the experimental data. The agreement is evaluated through a classification of the discrepancy between both.