Session G17: Reacting Flows: Flame Dynamics

Large-activation-energy analysis of gaseous reactive flow in pipes

Frank-Kamenetskii's analysis of thermal explosions is applied, using also a single-reaction model with an Arrhenius rate having a large activation energy, to describe the evolution of an initially cold gaseous mixture flowing along a circular pipe with constant wall temperature for moderately large values of the relevant Reynolds number. The analysis shows two modes of combustion. There is a flameless slowly reacting mode for low wall temperatures or small pipe radii, when the temperature rise resulting from the heat released by the reaction is kept small by the heat-conduction losses to the wall, so as not to change significantly the order of magnitude of the reaction rate. In the other mode, the slow reaction rates occur only in an initial ignition stage, which ends abruptly when very large reaction rates cause a temperature runaway, or thermal explosion, at a well-defined ignition distance. The analysis determines the slow streamwise evolution for the flameless mode of combustion as well as the ignition distance for the explosive mode.   

Propagation Limits of High Pressure Cool Flames

The flame speeds and propagation limits of premixed cool flames at elevated pressures with radiative heat loss are numerically modelled using dimethyl ether mixtures. The primary focus is paid on the effects of pressure, mixture dilution, flame size, and heat loss on cool flame propagation. The results showed that cool flames exist on both fuel lean and fuel rich sides and thus dramatically extend the lean and rich flammability limits. There exist three different flame regimes, hot flame, cool flame, and double flame. A new flame flammability diagram including both cool flames and hot flames is obtained at elevated pressure. The results show that pressure significantly changes cool flame propagation. It is found that the increases of pressure affects the propagation speeds of lean and rich cool flames differently due to the negative temperature coefficient effect. On the lean side, the increase of pressure accelerates the cool flame chemistry and shifts the transition limit of cool flame to hot flame to lower equivalence ratio. At lower pressure, there is an extinction transition from hot flame to cool flame. However, there exists a critical pressure above which the cool flame to hot flame transition limit merges with the lean flammability limit of the hot flame, resulting in a direct transition from hot flame to cool flame. On the other hand, the increase of dilution reduces the heat release of hot flame and promotes cool flame formation. Moreover, it is shown that a smaller flame size and a higher heat loss also extend the cool flame transition limit and promote cool flame formation.   

Flame speeds and curvature of premixed, spherically expanding flames advecting in a turbulent channel flow

A new facility has been developed at the Georgia Institute of Technology to study sub- and supersonic combustion, which is based on classical flame bomb studies but incorporates a mean flow, allowing for a wider variety of turbulent conditions and the inclusion of effects like compressibility, while supporting shear-free spherical flames. Homogeneous, isotropic turbulence is generated via an active vane grid. Methane-air flame kernels advecting with the mean flow are generated using Laser Induced Breakdown ignition. The facility is accessing the thin reaction zone regime with $u'_{RMS}/S^0_L = 6.9-22$, $L_{11}/\delta_F = 44-68$ and $Re_\lambda = 190 - 550$. The flame kernels are probed with OH-Planar Laser Induced Fluorescence (PLIF). To validate the facility, results at $\overline{U} = 30$ m/s are compared to existing data using a scaling derived from a spectral closure of the G-equation. This indicates the reacting flow remains Galilean invariant under the given conditions. The differences between global and local turbulent consumption speeds derived from OH-PLIF results are discussed with a focus on modeling efforts. The curvature of flame wrinkles is evaluated to examine the impact of different turbulent scales on flame development.   

Non linear dynamics of flame cusps: from experiments to modeling

The propagation of premixed flames in a medium initially at rest exhibits the appearance and competition of elementary local singularities called cusps. We investigate this problem both experimentally and numerically. An analytical solution of the two-dimensional Michelson Sivashinsky equation is obtained as a composition of pole solutions, which is compared with experimental flames fronts propagating between glass plates separated by a thin gap width. We demonstrate that the front dynamics can be reproduced numerically with a good accuracy, from the linear stages of destabilization to its late time evolution, using this model-equation. In particular, the model accounts for the experimentally observed steady distribution of distances between cusps, which is well-described by a one-parameter Gamma distribution, reflecting the aggregation type of interaction between the cusps. A modification of the Michelson Sivashinsky equation taking into account gravity allows to reproduce some other special features of these fronts.   

Role of Molecular Diffusion in Turbulent Flames: Two Examples.

In modeling turbulent flames, especially for high turbulent Reynolds number, it is often believed that diffusion is primarily controlled by turbulent diffusivity; and as such the effects of molecular diffusion can be ignored by artificially assuming unity Lewis number (Le) defined as the ratio of thermal diffusivity to mass diffusivity. Based on our recent experiments with expanding turbulent flames, we will present two examples where \textit{Le} significantly alters the flame dynamics even in strong turbulent environments. In the first example, we show that ignition of a combustible mixture by a high-energy kernel can be facilitated by turbulence for \textit{Le\textgreater 1} mixture while it is commonly believed to be more difficult in turbulence due to the increased dissipation rate of the deposited energy. In the second example, we show that nominally nonflammable mixtures with low adiabatic flame temperatures can burn strongly in turbulence for \textit{Le\textless 1} mixtures. In both cases, turbulence morphs the positively stretched spherical flame into a multitude of wrinkled flamelets subjected to both positive and negative stretches. Mechanistically, these effects are consequences of the coupling between differential molecular diffusion and these positively or negatively stretched flamelets.   

Representation of the Essential Flame-Turbulence Dynamics using Specific Flame-Vortex Interactions

Many engineering applications involve turbulent reacting flows, where nonlinear, multi-scale turbulence-combustion couplings are important. Directly resolving the complex fluid dynamics involved in these applications is associated with prohibitive computational costs, which makes it necessary to employ turbulent closure models and turbulent combustion models to account for the effects of unresolved scales on resolved scales. Most of these existent closure models rely on some assumptions about the turbulence dynamics and the scale separation between turbulence and the different combustion processes. A better understanding of the turbulence-combustion interactions is required for the development of more accurate, physics-based sub-grid-scale models for turbulent reacting flows. Instead of developing an extreme-resolution, high Reynolds number turbulent flame simulation that is limited to a localized part of the regime diagram, in this work, we propose to develop a series of numerical experiments of simplified interactions between a laminar premixed flame and specified vortex distributions of varying strengths and scales to capture the essential flame-turbulence dynamics over distinct premixed turbulent combustion regimes. The response of the laminar flame to different vortex time and length scales is investigated and the physical relevance of each dataset to practical turbulent premixed flames is discussed.   

Turbulent/Non-Turbulent Interface in a Reacting Compressible Shear Layer

Since entrainment occurs across the turbulent/non-turbulent interface (TNTI), DNS data is used to study the characteristics of this interface in shear layers. Several cases are considered ranging from a low compressible non-reacting to highly compressible reacting flows. As the compressibility level increases, the average size of the structures that form the TNTI increases, however, as the heat release level increases, the average size of the structures that form the local shape of TNTI decreases. The geometrical shape of the turbulent/non-turbulent interface looking from the turbulent region is examined. It is observed that in non-reacting cases the TNTI is dominated by the concave shaped surfaces. As the level of compressibility increases, the probability of finding highly curved concave shaped surfaces on the TNTI decreases, while the probability of finding flatter concave and convex shaped surfaces increases. In reacting flows with high heat release level, the TNTI is dominated by the convex shaped surfaces. As the heat release level increases the probability of finding highly curved convex shaped surfaces on the TNTI increases, whereas the probability of finding flatter concave and convex shaped surfaces decreases.   

Entrainment in a Reacting Compressible Shear Layer

DNS of reacting turbulent shear layer is performed to study the entrainment of the irrotational flow into the turbulent region. The effects of heat release and compressibility on the flow are examined. Infinitely fast chemistry approximation is used to model the one-step global reaction of hydrogen in air. Entrainment is studied via two mechanisms; nibbling, considered as the vorticity diffusion across the turbulent/non-turbulent interface, and engulfment, the drawing of the pockets of the outside irrotational fluid into the turbulent region. As the level of compressibility or heat release increases, the total entrained mass flow rate into the shear layer decreases. It is observed that nibbling is a viscous dominated mechanism in non-reacting cases, whereas it is essentially inviscid in reacting flows with high heat release level. It is shown that the contribution of the engulfment to entrainment is small for the non-reacting flows, while mass flow rate due to engulfment can constitute up to forty percent of total entrainment in reacting cases. This increase is primarily related to a decrease of mass flow rate due to nibbling while the mass flow rate due to engulfment does not change significantly in reacting cases.