Session OG: Reacting Flows III

Blow out and flashback in axisymmetric lean premixed combustion with swirl

Modern ground-based lean premixed gas turbine combustors achieve flame stabilization by the use of inlet swirl and a chamber expansion into a larger domain. This predisposes the premixed inlet vortex flow to transition to a vortex breakdown state, with stagnation and recirculation zones near the dump plane. This creates a compact turbulent and highly mixed reaction zone, which enables the use of leaner combustion to reduce emissions drastically. However, when the swirl level is too low, the flame cannot be stabilized and blows out. On the other hand, when swirl is too high, flame flashback may occur into the injection region. This work describes numerical simulations of non-reacting and combusting flows, which provide insight into the characteristics of the axisymmetric vortex breakdown in an expanding pipe. Results show good agreement with available theoretical predictions for the first appearance of breakdown downstream of the pipe expansion plane as well as its first appearance in the inlet pipe section as upstream swirl level is increased. It is demonstrated that the theoretical criteria can be used to predict the occurrence of flame blow out and flashback. The effects of inlet flow temperature, Mach number and global pressure on the reacting flow are also investigated.   

NO$_x$ formation in a premixed syngas flame

Reduction of NO$_x$ is a subject of significant current interest in stationary gas turbines. The objective of this study is to examine the effects of turbulence on non-thermal NO$_x$ formation in a syngas flame. This is archived by a detailed parametric study via PDF simulations of a partially stirred reactor and a dumped axisymmetric premixed flame. Several different detailed and reduced kinetics schemes are considered. The simulated results demonstrate the strong dependence of combustion process on turbulence. It is shown that the amount of NO$_x$ formation is significantly influenced by the inlet conditions. That is, the turbulence intensity can be tweaked to attain optimal ultra-low NO$_x$ emissions at a given temperature.   

Flow-combustion interactions in ducted flameholder-stabilized premixed flames

Turbulent premixed combustion is present in many power generation and propulsion systems due to its large energy conversion rate (as compared to non-premixed combustion) and its potential for reduced emissions (at the lean limit). As a result, the study of turbulent premixed flames has received substantial attention in the past through experiment, analysis and simulation. In the recent past, unsteady Computational Fluid Dynamics (CFD) based models have been increasingly leveraged towards the in depth study of the physics of turbulent premixed flames. The bulk of this effort focuses on the response of the flame to turbulence. In contrast, we focus on the opposite problem, i.e. the modification of the turbulent flowfield by the flame. This topic has also received some attention but with a strong emphasis on planar (in the mean), flames propagating normal to the flow. Instead, we focus on flameholder-stabilized ducted flames, i.e. ones in which the flame is confined and substantially inclined to the incoming flow. The fundamental mechanisms by which the flame impacts the flow, i.e. dilatation, baroclinic vorticity generation and molecular diffusion enhancement are discussed in detail and their relative impact quantified. Limitations of modeling these mechanisms in current state of the art CFD models are also addressed.   

Measurement of Local Burning Velocity in Turbulent Premixed Flames by Simultaneous CH-DPPLIF and Stereoscopic PIV

Simultaneous CH double-pulsed planar laser induced fluorescence (CH-DPPLIF) and stereoscopic particle image velocimetry (SPIV) measurements have been developed to investigate dynamics of flame elements in turbulent premixed flames. The CH-DPPLIF measurement, which provides displacement speed of the flame fronts directly, and SPIV measurement were conducted simultaneously in relatively high Reynolds number turbulent premixed jet flames. By selecting appropriate time interval of successive CH PLIF, movement of the flame front was captured clearly. In addition, by applying cross-correlation method for successive CH images obtained with minute time interval, displacement speed of the local flame element is measured, where the flame front is determined from mixed-average of CH fluorescence intensity in corresponding interrogation regions. Furthermore, by comparing the flame displacement speed with fluids velocity near the flame front, local burning velocity of each flame element is estimated and its relations with curvature and strain rate are discussed.   

Detailed analysis of premixed turbulent flame properties

Clean and efficient combustion is a hot topic. Many industrial applications are based on turbulent premixed combustion and it is very interesting to study these flames. Here we study statistically flat flames, because these can be regarded as essential parts of any complex turbulent flame. The flow is solved using Direct Numerical Simulation. The chemical kinetics are tackled using a chemical reduction technique, Flamelet Generated Manifolds. Two cases with different turbulence intensities are investigated. Different turbulence intensities result in different flame stretch rates, causing different flame structures. The effects of flame stretch on a turbulent flame are twofold: 1) the local flame speed changes and 2) the flame surface area changes. We present a detailed analysis of turbulent flame properties, including turbulent flame speed, mass consumption rate, actual flame surface area and Karlovitz integral. The Karlovitz integral describes the dimensionless flame stretch rate, integrated through the flame. These properties are valuable when testing new combustion models.   

Slow Invariant Manifolds in Chemically Reactive Systems

The scientific design of practical gas phase combustion devices has come to rely on the use of mathematical models which include detailed chemical kinetics. Such models intrinsically admit a wide range of scales which renders their accurate numerical approximation difficult. Over the past decade, rational strategies, such as Intrinsic Low Dimensional Manifolds (ILDM) or Computational Singular Perturbations (CSP), for equilibrating fast time scale events have been successfully developed, though their computation can be challenging and their accuracy in most cases uncertain. Both are approximations to the preferable slow invariant manifold which best describes how the system evolves in the long time limit. Strategies for computing the slow invariant manifold are examined, and results are presented for practical combustion systems.   

Effective Use of Storage/Retrieval-Based Chemistry Acceleration in CFD

{\em In situ} adaptive tabulation (ISAT) $\lbrack$S.B. Pope {\em Combust. Theory Modell}. 1:41-63, 1997$\rbrack$ has proven to be an efficient strategy for incorporating multiple-step chemical kinetics into CFD-based combustion modeling. Most ISAT applications reported in the literature have been for probability density function (PDF)-based simulations of statistically stationary flames, and have been limited to relatively small chemical mechanisms (fewer than 20 independent composition variables). Effective use of ISAT requires judicious specification of several control parameters, including error tolerances and scale factors. Here an ISAT- based algorithm $\lbrack$I. Veljkovic, P. Plassmann and D. Haworth, in {\em Computational Science and Its Applications – ICCSA 2003, Lecture Notes in Computer Science (LNCS 2667)}, Part I, pp. 643-653, Springer Verlag, 2003$\rbrack$ is applied to statistically stationary (nonpremixed jet flames) and nonstationary (direct-injection IC engines) combustion systems, using chemical mechanisms that range from fewer than 20 to more than 100 chemical species, with PDF-based and non-PDF-based modeling approaches. Modifications to the original ISAT algorithm are proposed to enhance its effectiveness in nonstationary systems and for larger chemical mechanisms, and recommendations for control-parameter specification are offered for different situations.   

A particle formulation for treating differential diffusion in filtered density function methods

We present a new approach for treating the molecular diffusion term in filtered density function (FDF) methods for modeling turbulent reacting flows. The diffusion term accounts for molecular `transport' in physical space and molecular `mixing' in scalar space. Conventionally, the FDF is represented by an ensemble of fluid particles and transport is modeled by a random walk in physical space. There are two significant shortcomings in this transport model: (1) the resulting composition variance equation contains a spurious production term, and (2) because the random walk is governed by a single diffusion coefficient, the formulation cannot account for differential diffusion, which can have a first-order effect in reacting flows. In our new approach transport is simply modeled by a mean drift in the particle composition equation. The resulting variance equation contains no spurious production term and differential diffusion is treated easily. Hence, the new formulation reduces to a DNS in the limit of vanishing filter width, a desirable property of any LES approach. We use the IEM model for mixing. It is shown that there is a lower bound on the specified mixing rate such that the boundedness of the compositions is ensured. We also present a numerical method for solving the particle equations which is second-order accurate in space and time, obeys detailed conservation, enforces the boundedness constraints, and is unconditionally stable.