Session G2: Reacting Flows: Extinction and Ignition

Dynamics of Near Limit Diffusion Flames at Elevated Pressures and Temperatures

The effects of cool flame chemistry and thermal radiation on the flame regimes and flammabilility limits of diffusion flames at elevated pressures and temperatures on dimethyl ether/air mixtures have been examined numerically. Three different flame regimes, cool flames, mild flames, and hot flames, are found. For near limit flames, the hot flames are bounded by a radiation extinction limit and a stretch extinction limit. The mild flame has a stretch extinction limit at high stretch rate. The cool flames have a reignition limit at low stretch rates and a stretch extinction limit at high stretch rate. Both pressure and temperature have significant impact on the flame regimes and their flammable regions. A new near limit diffusion flame flammability diagram, which includes all three flame regimes, is obtained.   

Ignition and Extinction Dynamics in Turbulent Nonpremixed “Cool” Flames

“Cool” flames result from the coupling of low-temperature chemistry with molecular transport. These flames have been experimentally and computationally observed under laminar flow conditions but have not been isolated under turbulent flow conditions. In this work, a skeletal n-heptane chemical mechanism including low-temperature chemistry is used to conduct detailed numerical simulations of nonpremixed “cool” flames subjected to unsteady, two-dimensional flow initialized from a plane of isotropic turbulence. Like conventional “hot” flames, under high Damköhler number conditions, these “cool” flames are found to be adequately described with a steady flamelet model. However, unlike conventional “hot” flames, “cool” flames exhibit two limit phenomena: extinction to a non-burning state at large scalar dissipation rate and ignition to a conventional “hot” flame at small scalar dissipation rate. The latter phenomenon allows for the possibility of ignition and re-extinction in addition to well-known extinction and re-ignition. The detailed simulation databases are analyzed to determine the relative contributions of nonpremixed (aligned with mixture fraction gradient) and premixed (normal to mixture fraction gradient) processes to this ignition and re-extinction phenomenon.   

Effects of scalar alignment on flame structure in multi-modal combustion

In practical systems, combustion does not occur in the asymptotic limits of nonpremixed and premixed combustion. Furthermore, when coupled with autoignition at elevated temperatures and pressures, complex multi-modal processes occur. An example of the complex interactions between combustion modes is the stabilization of lifted nonpremixed jet flames, which can occur kinematically via a classical “triple” flame, kinetically via autoignition, or via a combination of the two. In these lifted coflow jet flames, the direction of premixed (“triple”) flame propagation and autoignition front propagation is perpendicular to the mixture fraction gradient. Conversely, in equivalent multi-modal counterflow flames, the premixed flame or autoignition front propagation would be aligned with the mixture fraction gradient. In this work, the effect of this alignment on the multi-modal flame structure is assessed by comparing detailed numerical simulations of laminar DME/air flames in both lifted coflow jet and counterflow configurations under otherwise equivalent conditions. The thermochemical structures of both configurations are compared to a recently proposed reduced-order manifold for multi-modal combustion.   

DNS of flame stabilization assisted by auto-ignition at reheat conditions

Staged gas turbines with two sequential combustion chambers are being developed for power generation for their ability to achieve low emissions within a wide operational range while conserving high thermal efficiency. A particular implementation of the sequential combustion concept is characterized by a "reheat" combustion stage downstream of a first premixed-type combustor. Hot exhaust gases from the first stage are mixed with fuel in a mixing section, which provides the inlet conditions for the second-stage reheat combustor. DNS of flame stabilization regimes in the reheat burner, i.e. the combustor including the mixing section (duct-in-a-duct), at idealized conditions is performed using a detailed hydrogen-air mechanism. Results show that combustion occurs in two distinct modes. The first mode is an auto-ignition mode, whereby the vitiated oxidant facilitates the auto-ignition of the fuel in the mixing section. The second mode combines both auto-ignition and flame propagation, with auto-ignition occurring at and around the centerline of the combustor while flame propagation is stabilized at the recirculation zones near the corners. Chemical explosive mode analysis is employed to quantify the contribution of auto-ignition to the combustion rate relative to flame propagation.   

Direct numerical simulations of premixed autoignition in compressible uniformly-sheared turbulence

High-speed combustion systems, such as scramjet engines, operate at high temperatures and pressures, extremely short combustor residence times, very high rates of shear stress, and intense turbulent mixing. As a result, the reacting flow can be premixed and have highly-compressible turbulence fluctuations. We investigate the effects of compressible turbulence on the ignition delay time, heat-release-rate (HRR) intermittency, and mode of autoignition of premixed Hydrogen-air fuel in uniformly-sheared turbulence using new three-dimensional direct numerical simulations with a multi-step chemistry mechanism. We analyze autoignition in both the Eulerian and Lagrangian reference frames at eight different turbulence Mach numbers, $Ma_t$, spanning the quasi-isentropic, linear thermodynamic, and nonlinear compressibility regimes, with eddy shocklets appearing in the nonlinear regime. Results are compared to our previous study of premixed autoignition in isotropic turbulence at the same $Ma_t$ and with a single-step reaction mechanism. This previous study found large decreases in delay times and large increases in HRR intermittency between the linear and nonlinear compressibility regimes and that detonation waves could form in both regimes.   

Vorticity generation and jetting caused by a laser-induced optical breakdown

A focused laser can cause optical breakdown of a gas that absorbs energy and can seed ignition. The local hydrodynamics are complex. The breakdown is observed to produce vorticity that subsequently collects into a jetting flow towards the laser source. The strength and the very direction of the jet is observed to be sensitive to the plasma kernel geometry. We use detailed numerical simulations to examine the short-time ($<1\,\mu$s) dynamics leading to this vorticity and jetting. The simulation employs a two-temperature model, free-electron generation by multi-photon ionization, absorption of laser energy by inverse \textit{Bremsstrahlung}, and 11 charged and neutral species for air. We quantify the early-time contributions of different thermodynamic and gas-dynamic effects to the baroclinic torque. It is found that the breakdown produces compression waves within the plasma kernel, and that the mismatch in their strengths precipitates the involution of the plasma remnants and yields the net vorticity that ultimately develops into the jet. We also quantify the temperature distribution and local strain rates and demonstrate their importance in seeding ignition in non-homogeneous hydrogen/air mixtures.   

Ignition Prediction in a Hydrogen Jet in Turbulent Crossflow by a Laser-Induced Breakdown

We use large-scale simulation to predict the ignition (or not) of a hydrogen jet in turbulent air crossflow. The round jet issues from the wall of a low-speed wind tunnel into a turbulent boundary layer; an upstream laser-induced optical breakdown (LIB) is used to ignite it. The LIB hotspot also introduces a locally elevated oxygen radical concentration. A detailed hydrogen chemical mechanism is used to model the radicals. Additionally, ignition is augmented via actuation with a dielectric-barrier discharge that generates body forces and additional radicals. Comparison is made with corresponding experiments. We focus particularly in the ignition process. Traditional ignition identification involves a lengthy simulation run, until either the hotspot dissipates (unsuccessful ignition) or a sustained high temperature is observed (successful ignition). To avoid the computational cost in mapping the sustained-ignition threshold, a short-time criterion is developed based on detailed observations. It evaluates key radicals near the stoichiometric surface a short time after the LIB. This criterion allows for low-cost estimation of the approximate ignition boundary location, which can then be further refined via the traditional process.   

Numerical study of the ignition behavior of a post-discharge kernel injected into a turbulent stratified cross-flow

The reliable initiation of flame ignition by high-energy spark kernels is critical for the operability of aviation gas turbines. The evolution of a spark kernel ejected by an igniter into a turbulent stratified environment is investigated using detailed numerical simulations with complex chemistry. At early times post ejection, comparisons of simulation results with high-speed Schlieren data show that the initial trajectory of the kernel is well reproduced, with a significant amount of air entrainment from the surrounding flow that is induced by the kernel ejection. After transiting in a non-flammable mixture, the kernel reaches a second stream of flammable methane-air mixture, where the successful of the kernel ignition was found to depend on the local flow state and operating conditions. By performing parametric studies, the probability of kernel ignition was identified, and compared with experimental observations. The ignition behavior is characterized by analyzing the local chemical structure, and its stochastic variability is also investigated.