Session LV: Reacting Flows III

3D Flame and Eddy Structures of Turbulent Plane Jet Premixed Flame

3D DNS of hydrogen-air turbulent plane jet premixed flames, which are composed of high-speed unburnt gas and surrounding burnt gas, have been conducted. Fully-developed homogeneous isotropic turbulence is superimposed on the high speed mean flow under the assumption that a turbulence gird is installed in the upstream. A detailed kinetic mechanism including 12 species and 27 elementary reactions is considered and 0.228 billion grid points are used. Eddy structures which have large scale in space and streamwise rotating axis are produced along the outer edge of OH layer in burnt gas. These streamwise eddies are induced by velocity difference due to strong expansion of the burnt gas. Although combustion condition of the present DNS is classified into corrugated flamelets regime, unburnt mixture islands frequently appear behind the main flame. The creation of these islands is closely related to the fine scale eddies in the unburnt turbulence and the separated unburnt mixture is consumed rapidly by the heating from surrounding burnt gas.   

Three-dimensional DNS of turbulent premixed flames in a constant volume vessel

Clarification of flame behaviors in a vessel is of great importance for high efficiency of combustors, especially in SI engines. Direct numerical simulation of turbulent hydrogen-air premixed flames in a constant volume rectangular vessel at relatively high Reynolds number has been conducted by considering detailed kinetic mechanism. At first, flame ignites and propagates from the ignition kernel. When the flame approaches a wall, the flame displacement speed normal to the wall decreases gradually. After the flame impingement on the wall, the flame propagates along the wall and the flame displacement speed parallel to the wall becomes higher than that of freely propagating flames. The flame is also strongly affected by internal pressure rise in the vessel. Since the pressure increase makes flame thickness thin, heat release rate of each flame element is augmented. The local pressure rise due to dilatation also enhances turbulence and finer scale vortices appear, which makes flame surface more complicated and results in increase of the flame surface area.   

Analysis of Turbulent Premixed Flames using Gradient Trajectories

A new idea of analysing turbulent premixed flames was inspired by dissipation element analysis developed by Wang and Peters. Starting from every point on the flame surface in the direction of ascending and descending gradient of the underlaying curvature field, while not leaving the flame surface itself, one ends in a local maximum, respectively minimum point. The surface decomposes into fragments from which trajectories reach the same minimum and maximum. These fragments are surface-filling and do not overlap. Their geometry is non-arbitrary, since defined by the flame surface only. The linear distance between maximum and minimum of each fragment defines a length scale and the difference of curvature a scalar difference. The surface area of each fragment can then be estimated using the linear length and curvature difference. Therefore, estimating the whole flame surface area results in summing the areas of all fragments. Knowing the distribution function of the linear distance and the curvature difference can lead to a fine scale model of the flame surface area and the turbulent flame speed. DNS of the extended level-set G-equation in homogeneous isotropic forced turbulence have been performed in a cubic domain of $2\pi$ side length and $1024^3$ grid points.   

Mixture fraction and its Dissipation in a turbulent flame by using Krypton PLIF

Simultaneous conserved scalar and temperature imaging was performed in turbulent non-premixed jet flames (TNF DLR-A and -B) by using two-photon PLIF of Krypton and planar Rayleigh scattering, respectively. The motivation is to study the scalar and thermal dissipation characteristics of the flames. The experiments were performed at axial distances of 10 and 20 tube diameters (D) from the jet exit. The mixture fraction (f) field was calculated from the Kr LIF and the temperature images using a combination of the state relationship and an iterative technique. The mean and \textit{RMS} of the f and T compared very well with those of the point measurements made by previous researchers. The mixture fraction and thermal dissipation field appeared very similar away from the flame and less similar close to the flame. Mixture fraction diffusion and thermal diffusion conditioned on f and T reveal a linear trend, whose slope is very similar to the time scale estimated by the ratio of the respective dissipation to the variance.   

PDF-Based Simulations of Nonpremixed Turbulent Jet Flames With Advanced Radiation Models

Transported probability density function (PDF) methods have been adopted as the basis for simulating turbulent reacting flows in both Reynolds-averaged and spatially filtered (large-eddy simulation) contexts. Skeletal gas-phase chemical mechanisms have been implemented for oxidation and pollutant formation in hydrocarbon-air flames, and detailed soot models have been implemented using a method-of-moments for soot aerosol dynamics. Photon Monte Carlo and high-order spherical harmonics radiative transfer equation (RTE) solvers have been coupled with the PDF method to deal with participating-medium radiation and turbulence/radiation interactions. Line-by-line and advanced $k$-distribution methods have been developed for spectral radiation properties. The result is a comprehensive framework for simulating luminous and nonluminous turbulent flames that accurately captures complex turbulence/chemistry/soot/radiation interactions. Quantitative comparisons with experimental measurements have been made. Systematic parametric studies have been performed to explore the relative importance of the choice of RTE solver versus the spectral radiation model.   

Large-eddy simulation/probability density function modeling of local extinction and re-ignition in Sandia Flame E

A large-eddy simulation (LES)/probability density function (PDF) code is developed and applied to the study of local extinction and re-ignition in Sandia Flame E. The modified Curl mixing model is used to account for the sub-filter scalar mixing; the ARM1 mechanism is used for the chemical reaction; and the in- situ adaptive tabulation (ISAT) algorithm is used to accelerate the chemistry calculations. Calculations are performed on different grids to study the resolution requirement for this flame. Then, with sufficient grid resolution, full-scale LES/PDF calculations are performed to study the flame characteristics and the turbulence-chemistry interactions. Sensitivity to the mixing frequency model is explored in order to understand the behavior of sub-filter scalar mixing in the context of LES. The simulation results are compared to the experimental data to demonstrate the capability of the code. Comparison is also made to previous RANS/PDF simulations.   

Lagrangian and Eulerian PDF Simulations of Nonpremixed Turbulent Flames with Moderate-to-Strong Turbulence/Chemistry Interactions

Transported probability density function (PDF) methods offer compelling advantages for modeling chemically reacting turbulent flows, and Lagrangian particle-based Monte Carlo algorithms have become the predominant method for solving modeled PDF transport equations. Significant progress has been made in Lagrangian particle methods to facilitate their implementation into conventional Eulerian computational fluid dynamics (CFD) codes. Still, it would be desirable to realize the advantages of PDF methods using more conventional numerical algorithms and/or at lower computational cost. Toward these ends, Eulerian field methods have been proposed as alternatives to particle-based methods for solving modeled PDF transport equations. Here we apply Lagrangian particle and Eulerian field (both stochastic and deterministic) PDF methods to a hierarchy of three piloted methane-air nonpremixed turbulent jet flames where turbulence/chemistry interactions become progressively more important. Both accuracy and computational efficiency are assessed. Based on these results, we provide recommendations on how to apply each method most effectively, and we identify outstanding issues requiring future research.   

Simulations of Turbulent Spray Combustion in a Constant-Volume Chamber for Diesel-Engine-Like Conditions

In-cylinder aero-thermal-chemical processes in piston engines are rich and complex, and modern engines are already at a high level of refinement. Further increases in performance, reductions in fuel consumption and emissions, and accommodation of nontraditional fuels will require the effective use of high-spatial-and-temporal-resolution optical diagnostics and numerical simulations. In this research, computational fluid dynamics tools are being developed to explore the influences of fuel properties on autoignition, combustion, and pollutant emissions in compression-ignition engines. The modeling includes a transported probability density function method to account for turbulent fluctuations in composition and temperature, detailed soot models with a method of moments for soot aerosol dynamics, a stochastic photon Monte Carlo method for participating-medium radiation heat transfer, and line-by-line spectral properties for mixtures of molecular gases and soot. The models are applied to a constant-volume spray combustion bomb where measurements are available for a range of thermochemical conditions and for a variety of fuels. Parametric studies of the influences of key physical and numerical parameters are performed to determine sensitivities and to establish best practices to be carried forward into subsequent modeling studies of real engines.   

Mechanism for AC Electric Field Deflection of Flames

Effects of electric fields on flames have been observed and studied since the 19th century. It is well known that the presence of an electric field can modify the shape of a burner or candle-like diffusion flame. Most experimental observations and theoretical analyses focused on DC fields. Recent experiments show that a flame can also be bent and even put out by an AC field. To explain how a zero time average cause can give rise to a net effect on the flame we develop a perturbation theory of the combustion equations modified to allow for the presence of the field and completed by Maxwell's equation. Theoretical and numerical analyses of the equations indeed show that the AC field creates a force whose magnitude is comparable to gravity for high enough fields (1e5 V/m). The dependency of this critical field on the frequency and the effect on the flame shape are also obtained and compared to experimental results.   

Wildfire simulation using a chemically-reacting plume in a crossflow

Water tunnel experiments reveal the flame length of a chemically-reacting plume in a crossflow. Salt water containing a pH indicator and a base is slowly injected from above into the test section of a water tunnel containing an acidic solution. The flame length is measured optically as a function of the buoyancy flux, crossflow speed, and volume equivalence ratio of the chemical reaction. Based on earlier work of Broadwell with the transverse jet, a simple dilution model predicts the flame length of the transverse plume. The plume observations are in accord with the model. As with the jet, there is a minimum in the flame length of the plume at a transition between two self-similar regimes, corresponding to the formation of a pair of counter-rotating vortices at a certain crossflow speed. At the transition, there is a maximum in the entrainment and mixing rates. In an actual wildfire with variable winds, this transition may correspond to a dangerous condition for firefighters.