Session M17: Reacting Flows: Ignition and Extinction

Premixed autoignition in compressible turbulence

Prediction of chemical ignition delay in an autoignition process is critical in combustion systems like compression ignition engines and gas turbines. Often, ignition delay times measured in simple homogeneous experiments or homogeneous calculations are not representative of actual autoignition processes in complex turbulent flows. This is due the presence of turbulent mixing which results in fluctuations in thermodynamic properties as well as chemical composition. In the present study the effect of fluctuations of thermodynamic variables on the ignition delay is quantified with direct numerical simulations of compressible isotropic turbulence. A premixed syngas-air mixture is used to remove the effects of inhomogeneity in the chemical composition. Preliminary results show a significant spatial variation in the ignition delay time. We analyze the topology of autoignition kernels and identify the influence of extreme events resulting from compressibility and intermittency. The dependence of ignition delay time on Reynolds and turbulent Mach numbers is also quantified.   

Dynamics of autoignitive DME/air coflow flames in oscillating flows

The structure and dynamics of laminar nonpremixed dimethyl ether (DME)/air coflow flames were investigated at elevated temperatures and pressures, conditions at which autoignition times become competitive with flame times. Computations with detailed chemistry were performed for DME and heated coflow air at 30 atm with uniform but sinusoidally oscillating inlet velocities. These unsteady cases were compared with steady flames to elucidate the effect of oscillation frequency on the flame dynamics. In the oscillating reacting flow, periodic but hysteretic transition occurs between a multibrachial autoignition front that locates downstream at high inlet velocity and a tribrachial flame that locates upstream at low inlet velocity. The finite induction time for autoignition results in this hysteretic behavior, which diminishes at lower oscillation frequency as there is more time for chemistry to respond to the hydrodynamic changes and consequently approach steady state.   

Fluid-Plasma-Combustion Coupling Effects on the Ignition of a Fuel Jet

We analyze the effect of plasma-combustion coupling on the ignition and flame supported by a DBD interacting with a jet of H$_2$ in a air cross-flow. We propose that plasma-combustion coupling is due to the strong temperature-dependence of specific collisional energy loss as predicted by the Boltzmann equation, and that e$^-$ transport can be modeled by assuming a form for the E-field pulse in microstreamers. We introduce a two-way coupling based on the Boltzmann equation and the charged species conservation. The addition of this mechanism to a hydrogen combustion scheme leads to an improvement of the ignition prediction and of the understanding of the effect of the plasma on the flow. The key points of the analysis are 1) explanation of the mechanism for the two-stage ignition and quenching observed experimentally, 2) explanation of the existence of a power threshold above which the plasma is beneficial to the ignition probability, 3) understanding of the increase in power absorbed by the plasma in burning conditions and the reduction in power absorbed with an increase in the cross velocity, 4) explanation of the non-symmetric emissions and the increase in luminescence at the rotovibrational H$_2$O band. The model is validated in part against air-H$_2$ flow experiments.   

A Computational and Experimental Study of Ignition Behavior of Gasoline Surrogate Fuels Under Low-Temperature Combustion Conditions

One strategy for next-generation engines is low-temperature compression ignition of gasoline. Reaction pathways that are not relevant for high-temperature flame propagation are activated under these conditions, and the ignition behavior of these fuels under low-temperature conditions has not been widely explored. Here the ignition behavior of gasoline and two- and three-component surrogates has been studied experimentally and computationally over a range of operating conditions of interest for low-temperature engine combustion. Experiments were performed in a single-cylinder research engine. For each fuel blend, the critical compression ratio (lowest compression ratio at which the main ignition occurs) was determined over a range of operating conditions, by varying one parameter at a time with all other parameters held fixed. A simplified CFD model that considers detailed chemical kinetics was used to simulate the experiment. The focus of the study is to determine which surrogate fuel mixtures and chemical mechanisms are able to capture the ignition behavior of gasoline under these conditions. For example, different ignition behavior is found for different surrogate mixtures that all have the same Research Octane Number, and it is important to capture this behavior in CFD models.   

Spark ignition of aviation fuel in isotropic turbulence

Turbulent spark ignition occurs in combustion engines where the spark must establish a viable flame kernel that leads to stable combustion. A competition exists between kernel growth, due to flame propagation, and kernel attenuation, due to flame stretch and turbulence. This competition can be measured by the Karlovitz number, Ka, and kernel viability decreases rapidly for $\text{Ka} \gg 1$. In this study, the evolution of an initially spherical flame kernel in a turbulent field is investigated at two cases: $\text{Ka}_{-}$ ($\text{Ka}=25$) and $\text{Ka}_{+}$ ($\text{Ka}=125$) using direct numerical simulation (DNS). A detailed chemical mechanism for jet fuel (Jet-A) is used, which is relevant for many practical conditions, and the mechanism includes a pyrolysis sub-model which is important for the ignition of large hydrocarbon fuels. An auxiliary non-reacting DNS generates the initial field of isotropic turbulence with a turbulent Reynolds number of 500 ($\text{Ka}_{-}$) and 1,500 ($\text{Ka}_{+}$). The kernel is then imposed at the center of the domain and the reacting DNS is performed. The $\text{Ka}_{-}$ case survives and the $\text{Ka}_{+}$ case is extinguished. An analysis of the turbulence chemistry interactions is performed and the process of extinction is described.   

Optimal ignition placement using nonlinear adjoint looping

Spark ignition of a turbulent mixture of fuel and oxidizer is a highly sensitive process. Traditionally, a large number of parametric studies are used to determine the effects of different factors on ignition and this can be quite tedious. In contrast, we treat ignition as an initial value problem and seek to find the initial condition that maximizes a given cost function. We use direct numerical simulation of the low Mach number equations with finite rate one-step chemistry, and of the corresponding adjoint equations, to study an axisymmetric jet diffusion flame. We find the $L-2$ norm of the temperature field integrated over a short time to be a suitable cost function. We find that the adjoint fields localize around the flame front, identifying the most sensitive region of the flow. The adjoint fields provide gradient information that we use as part of an optimization loop to converge to a local optimal ignition location. We find that the optimal locations correspond with the stoichiometric surface downstream of the jet inlet plane. The methods and results of this study can be easily applied to more complex flow geometries.   

Adjoint-based sensitivity of flames to ignition parameters in non-premixed shear-flow turbulence

The adjoint of the linearized and perturbed compressible flow equations for a mixture of chemically reacting ideal gases is used to assess the sensitivity of ignition in non-premixed shear-flow turbulence. Direct numerical simulations are used to provide an initial prediction, and the corresponding space-time discrete-exact adjoint is used to provide a sensitivity gradient for a specific quantity of interest (QoI). Owing to the ultimately binary outcome of ignition (i.e., it succeeds or fails after some period), a QoI is defined that both quantifies ignition success and varies smoothly near its threshold based on the heat release parameter in a short-time horizon during the ignition process.~ We use the resulting gradient to quantify the flow properties and model parameters that most affect the initiation of a sustained flame. A line-search algorithm is used to identify regions of high ignition probability and map the boundary between successful and failed ignition. The approach is demonstrated on a non-premixed turbulent shear layer and on a reacting jet-in-crossflow.   

Numerical solution of an edge flame boundary value problem

We study edge flames for modeling extinction, reignition, and flame lifting in turbulent non-premixed combustion. An adaptive resolution finite element method is developed for solving a strained laminar edge flame in the intrinsic moving frame of reference of a spatially evolving shear layer. The variable-density zero Mach Navier-Stokes equations are used to solve for both advancing and retreating edge flames. The eigenvalues of the system are determined simultaneously (implicitly) with the scalar fields using a Schur complement strategy. A homotopy transformation over density is used to transition from constant- to variable-density, and pseudo arc-length continuation is used for parametric tracing of solutions. Full details of the edge flames as a function of strain and Lewis numbers will be discussed.   

Real fuel effects on flame extinction and re-ignition

Flame-vortex interactions have significant implications in studying combustion in practical aeronautical engines, and can be used to facilitate the model development in capturing local extinction and re-ignition. To study the interactions between the complex fuel and the intense turbulence that are commonly encountered in engines, direct numerical simulations of the interactions between a flame and a vortex pair are carried out using a recently-developed 24-species reduced chemistry for n-dodecane. Both non-premixed and premixed flames with different initial and inlet thermochemical conditions are studied. Parametric studies of different vortex strengths and orientations are carried out to induce maximum local extinction and re-ignition. Chemical-explosive-mode-analysis based flame diagnostic tools are used to identify different modes of combustion, including auto-ignition and extinction. Results obtained from the reduced chemistry are compared with those obtained from one-step chemistry to quantify the effect of fuel pyrolysis on the extinction limit. Effects of flame curvature, heat loss and unsteadiness on flame extinction are also explored. Finally, the validity of current turbulent combustion models to capture the local extinction and re-ignition will be discussed.   

Localized flame extinction and re-ignition in turbulent jet ignition assisted combustion

Direct numerical simulations (DNS) of turbulent jet ignition (TJI)-assisted combustion of ultra-lean fuel-air is performed in a three-dimensional planar jet configuration. TJI is a novel ignition enhancement method which facilitates the combustion of lean and ultra-lean mixtures by rapidly exposing them to high temperature combustion products. Fully compressible gas dynamics and species equations are solved with high order finite difference methods. The hydrogen-air reaction is simulated with a detailed chemical kinetics mechanism consisting of 9 species and 38 elementary reactions. The interesting phenomena involved in TJI combustion including localized premixed flame extinction/re-ignition and simultaneous premixed/non-premixed flames are investigated by using the flame heat release, temperature, species concentrations, and a newly defined TJI progress variable.