Turbulent Flame Speed Correlations Research Articles

A computational assessment of flame speed correlation in an ultra-lean pre-chamber engine

Predictive modeling of pre-chamber combustion engines relies primarily on the correct description of laminar and turbulent flame speeds. For engineering applications, the correlations of the flame speeds with physical variables involve empirical constants that are valid for a limited range of operating conditions. The current work aims at assessing the significance of laminar flame speed prediction in the simulation of ultra-lean pre-chamber engine combustion operated with methane. Gülder’s empirical correlation for laminar flame speed was chosen as a reference and was further modified for equivalence ratio, pressure and temperature ranges beyond what it was originally derived for, in order to confirm the original hypothesis; the pressure and temperature dependence were adopted as a power-law correlation. Based on the computational results using the skeletal reaction mechanism, the correlation was modified better represent the flame speeds at ultra-lean engine conditions, using GRI 3.0 as a reference. The modified correlation for methane was implemented in CONVERGE, a three-dimensional computational fluid dynamics (CFD) solver, and the results were validated against the experimental data. In all cases, the original formulation of Peters’s turbulent flame speed correlation was used and was found to have insignificant effect on the conditions under study, confirming the importance of the accurate determination of the laminar flame speed that dominates over any turbulence corrections for high Karlovitz number effects. The flame topology was also investigated to provide insights into the observed pressure behavior among the tested cases. Finally, the relevant turbulent combustion regimes encountered in the pre-chamber combustion engine conditions were examined in the Borghi-Peters diagram, further confirming the findings of the study.

Turbulent flame speed measurement of NH3/H2/air and CH4/air flames and a numerical case study of NO emission in a constant volume combustion chamber (C.V.C.C.)

Ammonia combustion is a meaningful method to retrieve stored amounts of excess variable renewable energy. The practical combustors like engines and gas turbines fired by the common hydrocarbon fuels need to be operated by the carbon-free like ammonia or ammonia/hydrogen blends shortly. Thus, in this study, the turbulent flame speed (ST) and NO emission potentials of the NH3/H2 blend (NH3/H2 = 40/60, 50/50, and 60/40 vol%) are investigated in comparison with the methane/air flames in a constant volume combustion vessel. The effects of Lewis number (Le), turbulent intensity (u′=0.78–2.34 m/s), and pressures (1, 5 bar) on normalized ST/SL are considered and validated against literature proposed correlations. Flame morphologies show that the turbulent flame wrinkling characteristics are mainly dominated by the turbulent intensity (u′/SL), turbulent length scale (lT/lF) at different pressures, and differential-diffusion term (Le). The differential-diffusion and turbulent stretch, jointly determine the ST/SL of NH3/H2/air and CH4/air mixtures. The scaling parameter ReT,flame/Le2 can describe the self-similar propagation characteristics of ST/SL of spherical flame and consider the differential-diffusion effect. The further Damköhler number, Da modification on the general turbulent flame speed correlation proposed as: ST/SL ∼ ADaB·(ReT,flame/Le2)0.5 unifies different hydrogen content cases by considering the turbulent stretch effect. When literature experimental data are also correlated together, the scaling law becomes: ST/SL ∼ Re0.5/Le ∼ (u'/SL)0.5(lT/lF)0.5Le-1 with equal weighting factors of u'/SL and lT/lF terms, this follows the Damköhler's second hypothesis at highly turbulent regimes. The impact of turbulent intensity and residence time on NO emissions in the flue gas of NH3/H2/air and CH4/air flames are independent of turbulent intensity due to the long residence time. Simulated results show that residence time has contrary effects on NO emissions of NH3/H2/air and CH4/air flames due to the contrary role of the thermal-NOx pathway. Practical utilization of NH3/H2/air in gas turbine combustors could increase the flame residence time by increasing the back-flow zone in the post-flame zone and swirl number to suppress the NOx emissions.

Impact of Flame-Generated Turbulent Intensity and Flame Speed on the Low-Order Modelling of Light-Round

A previously-developed low-order Lagrangian stochastic model for ignition of premixed and non-premixed flames is modified in this paper to improve the numerical prediction of the light-round process in premixed annular combustors. The model refinements take into account Flame-Generated Turbulent Intensity (FGTI) and impose a turbulent flame speed correlation to the flame particles using expressions from the literature. For this, using RANS CFD results as an input, the model was applied to simulate the ignition transient in a premixed, swirled bluff body stabilised annular combustor to characterise the light-round time, both in stable conditions and close to the stability limits. Several cases were analysed, where flame speed and fuel were varied and light-round times were compared to experimental results. The proposed modifications improved the accuracy of the light-round time predictions, suggesting that FGTI may be an important phenomenon to be modelled. This modified model coupled with dilatation and the Peter’s assumption for the turbulent flame speed resulted in considerable improvement for the light-round time calculation for the explored range of parameters. This is an attractive feature considering the low computational cost of these simulations, which can be run in a single core of a local workstation. The improved model can help gas turbine engineers assess the ignition behaviour of annular combustors early in the design process.

Turbulent Flame Speed of a Gasoline surrogate in conditions representative of modern downsized Spark-Ignition engine

The downsized spark ignition (SI) engine is considered as one promising technology to reduce pollutants and greenhouse gas emissions. By boosting the intake air pressure, this operating mode achieves higher efficiency, thus reducing pollutant emissions. Modeling the combustion process in these drastic conditions (high pressure, high temperature and high dilution rate) remains challenging, however, since most of models of premixed turbulent combustion seem insufficiently robust because the turbulent combustion regime might be out of the classical flamelet regime. Two different setups, a spherical vessel and a one-shot engine, were used to study turbulence-flame interaction in conditions representative of a modern downsized SI engine, i.e. under high pressure, high temperature and high dilution and controlled turbulence. Turbulent flame speed were measured in a wide range of Damkohler and Karlovitz number. Several turbulent flame speed correlations from the literature are also tested and compared to the complete experimental dataset gathered on both setups.

An Optimized Correlation for Turbulent Flame Speed for C1–C3 Fuels at Engine-Relevant Conditions

Abstract The aim of this work is to examine the state-of-the-art turbulent flame speed (ST) correlations and optimize their adjustable parameters to best match a wide range experimental turbulent premixed combustion results. Based on previous work, four correlations have been selected for this study. Using a matlab-based Nelder–Mead simplex direct search method, each correlation's adjustable parameters are optimized such that their mean absolute percentage error (MAPE) is minimized. In addition to the literature correlations, a new empirical correlation is developed using the same search method to define constants and powers in the expression. Two sets of optimized parameters are proposed to account for atmospheric and elevated (0.2–3.0 MPa) pressure flames. Each correlation is tested further, examining their ability to match ST trends for varying equivalence ratio (φ) and turbulent velocity ratio (u′/SL). It was found that a minimum of two correlations and two sets of adjustable parameters are required to accurately account for the entire range of data, thus showing that there is currently no turbulent flame speed correlation that is applicable across all engine-relevant conditions.

Modelling turbulent premixed flame–wall interactions including flame quenching and near-wall turbulence based on a level-set flamelet approach

A level-set flamelet model for turbulent premixed combustion that accounts for the main effects of flame–wall interaction, quenching and near-wall turbulence, is proposed based on the G-equation. The structure of laminar unquenched and quenched flames is analysed and a consistent G-equation valid for both flame types is derived. For the modelling of the turbulent quenching process, it is argued that the state of individual flamelets can be described as either quenched or unquenched. This binary mechanism leads to a kinematic description of turbulent flame–wall interactions in which the fraction of unquenched flames is described by the unquenched factor Q. It is shown that Q allows for a general and appropriate scaling of the turbulent burning velocity due to the fact that only unquenched flamelets contribute to the overall propagation speed. For the modelling of turbulent premixed combustion, a unified G-equation valid for unquenched and quenched flames is derived. Modelling closures accounting for near-wall turbulence as well as unsteady flame development effects are introduced. The modelling approach is analysed a priori as well as a posteriori using a turbulent channel flow. With regard to the turbulent burning velocity, it is found that the effect of quenching is dominant compared to the effect of wall-bounded turbulence. The latter becomes important for a proper estimation of the turbulent flame length. Moreover, the model is compared against available DNS data of flame–wall interaction in a turbulent channel flow. It is shown that the turbulent burning velocity, the turbulent flame thickness as well as the reactive flame surface density near the wall are correctly reproduced. The present modelling approach thus allows for a consistent modelling of flame–wall interactions, which can also be transferred to other combustion models that are based on turbulent flame speed correlations.

Experimental study on turbulent expanding flames of lean hydrogen/air mixtures

The aim of this paper is to report new experimental results on the effect of turbulence on the combustion properties of lean to stoichiometric H2/air mixtures in a closed, fan-stirred, spherical vessel. To do so, a new experimental setup, spherical bomb, has been used to investigate the effect of a given and well-characterized turbulence intensity on the increase of hydrogen/air flame speed and on the combustion pressure. The initial hydrogen molar percent of H2 in air was comprised between 16% and 28%. The initial turbulence created in the vessel prior to ignition was varied between 0.56 and 2.81m/s for an integral length scale of around 50mm. It was shown that the turbulent speed increases drastically when the turbulence is increased but the maximum combustion pressure remains the same. A correlation was proposed in order to the turbulent speed. The proposed turbulent flame speed correlation was able to predict not only the present data but also the results of the literature (Kitagawa et al., 2008).

Laboratory Study of Premixed H2-Air and H2–N2-Air Flames in a Low-Swirl Injector for Ultralow Emissions Gas Turbines

The objective of this study is to conduct laboratory experiments on low-swirl injectors (LSIs) to obtain the basic information for adapting LSI to burn H2 and diluted H2 fuels that will be utilized in the gas turbines of the integrated gasification combined cycle coal power plants. The LSI is a novel ultralow emission dry-low NOx combustion method that has been developed for gas turbines operating on natural gas. It is being developed for fuel-flexible turbines burning a variety of hydrocarbon fuels, biomass gases, and refinery gases. The adaptation of the LSI to accept H2 flames is guided by an analytical expression derived from the flow field characteristics and the turbulent flame speed correlation. The evaluation of the operating regimes of nine LSI configurations for H2 shows an optimum swirl number of 0.51, which is slightly lower than the swirl number of 0.54 for the hydrocarbon LSI. Using particle image velocimetry (PIV), the flow fields of 32 premixed H2-air and H2–N2-air flames were measured. The turbulent flame speeds deduced from PIV show a linear correlation with turbulence intensity. The correlation constant for H2 is 3.1 and is higher than the 2.14 value for hydrocarbons. The analysis of velocity profiles confirms that the near field flow features of the H2 flames are self-similar. These results demonstrate that the basic LSI mechanism is not affected by the differences in the properties of H2 and hydrocarbon flames and support the feasibility of the LSI concept for hydrogen fueled gas turbines.

Turbulence Effects on Flame Speed and Flame Structure

Turbulence effects on methane-air flames stabilized in grid turbulence were investigated through measurements of flame speed and mean and fluctuating flame temperature profiles. Published turbulent flame speed correlations were able to correlate the experimental flame speed data but were contradictory in indicating flame structure and combustion mechanisms. A simple, one-dimensional, wrinkled laminar flame model was used to predict characteristic flame temperature fluctuation levels. Comparison of these predictions with measured temperature fluctuations indicated that the majority of the flames studied were wrinkled laminar flames. However, a wrinkled laminar flame structure was inappropriate for the most intensely turbulent flames examined.

A rational correlation of combustion noise results from open turbulent premixed flames

The results are presented of 103 tests on the noise generation capability of bunsen burner type premixed hydrocarbon-air turbulent flames. The fuels used have been propane, propylene, acetylene and ethylene, all burning at 1 atmosphere in an anechoic chamber.The results for sound power output, reacting volume and spectral content are correlated with the burner diameter, flow velocity, fuel mass fraction and laminar flame speed; these variables are sufficient to produce correlations with respect to the Reynolds number, Mach number, fuel mass fraction and the first Damköhler similarity group. Introducing a new simplified version of the theory of combustion noise, known turbulent flame speed correlations, and the theory of isotropic turbulence decay downstream of the burner lip, it is shown that the correlations for frequency content and radiated sound power have a rational basis.