Session QU: Reacting Flows V

Detonation propagation through methane-air mixtures with fuel concentration gradients

The complex structure of a multidimensional detonation front consists of constantly changing, multiply intersecting incident shocks and Mach stems followed by growing and shrinking regions of reacted and unreacted gases. Because these flow structures change in time, the energy release in the shocked and compressed gases varies in space and time. Trajectories of triple points formed at shock intersections create cellular patterns whose size and structure are characteristic of the particular material and the background condition. In high-activation-energy fuel-air mixtures, such as methane in air, cellular patterns are relatively large, very irregular, and have complex and changing substructures. Here we use numerical simulations to study the behavior of detonations propagating through methane-air mixtures with a spatial gradient of fuel concentration. When the mixture stoichiometry varies from stoichiometric, the detonation propagation speed slows and sizes of cellular structures grow. In partially premixed systems with a nonuniform concentration of fuel -- a condition that can occur, for example, naturally in sealed underground coal mine tunnels -- both the propagation speed and the characteristic detonation cell size vary spatially.   

The role of hydrodynamic instability in the turbulent propagation of premixed flames

We investigate the role of hydrodynamic instability in the wrinkled flamelet regime of turbulent combustion, where the turbulence scale is large compared to the flame thickness and the intensity small compared to the laminar flame speed. To do so we use the Michelson-Sivashinsky equation suitably forced by a noise term, representing the turbulent field, whose intensity and spatial auto-correlation function can be fully controlled. We study the effects on the turbulent propagation speed of turbulence intensity, integral scale and of a parameter measuring the degree of hydrodynamic instability. We find two different behaviors depending on the stability of the planar front. If the planar front is hydrodynamically stable we find a quadratic dependence of the turbulent speed on intensity, modulated by a linear dependence on the instability parameter. If the planar flame is unstable, the basic state is corrugated and the scaling law is more complex, revealing a certain resilience of the flame to turbulent perturbation. We also observe, both for stable planar and corrugated flames, the existence of an intermediate integral scale at which the turbulent speed is maximized. For vanishing integral scales the flame surface becomes fractal and a limiting fractal dimension was established. Two-dimensional flames exhibit polyhedral-cellular structures similar to those observed in experiments.   

Pinning of reaction fronts by moving vortices

We present experimental studies of the effects of moving vortices on a propagating reaction front. A vortex moving through the front in the same direction as the front pins and drags the front, depending on the translation speed of the vortex and the maximum component of the vortex velocity in the same direction. The relation between these two velocities when the front is pinned is determined experimentally for a single vortex forced magnetohydrodynamically. Lateral motion of this vortex is achieved with a translation stage that moves both a magnet underneath and a central electrode through the fluid. Studies are also done of front-pinning with a random vortex flow. As the translation speed of the vortex pattern increases, pinning drops out at the weaker vortices in the flow. We discuss generalizations of these front-pinning ideas to front propagation in more general, vortex-dominated flows.   

Burning manifolds and burning lobes

We present experimental studies of the propagation of a reaction front in a fluid flow composed of a chain of alternating vortices. We propose that the tools used to describe the transport of a {\em passive} impurity in a flow can be expanded to account for the behavior of a reaction front. In particular, we propose that motion of a reaction front from one region to another in the flow is determined by {\em burning manifolds} and {\em burning lobes}. These ideas are tested experimentally for both the time-independent and time-dependent vortex chain. For a time-independent flow, the time that it takes for a triggered reaction to propagate from one vortex to the next is the minimum time $\tau$ for the stable burning manifold $B_S(\tau)$ to envelope the original trigger point. For a time-dependent (oscillatory) vortex chain, we use the burning manifold/lobe framework to explain mode-locking behavior seen in earlier studies.\footnote{M.S. Paoletti and T.H. Solomon, Europhys. Lett. {\bf 69}, 819 (2005); Phys. Rev. E {\bf 72}, 046204 (2005).}   

Numerical Investigation of the Hydrogen Jet Flammable Envelope Extent with Account for Unsteady Phenomena

An important aspect of safety analysis in hydrogen applications is determination of the extent of flammable gas envelope in case of hydrogen jet release. Experimental investigations had shown significant disagreements between the extent of average flammable envelope predicted by steady-state numerical methods, and the region observed to support ignition, with proposed cause being non-steady jet phenomena resulting in significant variations of instantaneous gas concentration and velocity fields in the jet. In order to investigate the influence of these transient phenomena, a numerical investigation of hydrogen jet at low Mach number had been performed using unsteady Large Eddy Simulation. Instantaneous hydrogen concentration and velocity fields were monitored to determine instantaneous flammable envelope. The evolution of the instantaneous fields, including the development of the turbulence structures carrying hydrogen, their extent and frequency, and their relation with averaged fields had been characterized. Simulation had shown significant variability of the flammable envelope, with jet flapping causing shedding of large scale rich and lean gas pockets from the main jet core, which persist for significant times and substantially alter the extent of flammability envelope.   

Water Mist Interaction with Flame Spreading Against Gravity

Water mist fire-suppression systems have gained importance since chemical agents like Halons are being phased out for environment preservation. The present study focuses on the effect of water mist droplet size and concentration in inhibiting the flame spreading downward over thin solid fuel at different gravity levels. The water droplets are introduced into the air stream at pre-specified concentration and droplet size. An Eulerian-Eulerian two phase model is used for this particular study. The polydisperse spray is modeled using the moments of the droplet size distribution function. The gas phase is modeled by full Navier-Stokes equations for laminar flow along with the conservation equations of mass, energy {\&} species. A one-step Arrhenius reaction between fuel vapor and oxygen is assumed. The gas radiation equation is solved using DOM. The solid fuel considered is assumed to burn ideally to form fuel vapors without melting. The thin solid fuel is modeled by equations of continuity and energy. The pyrolysis of fuel is modeled as one-step, zeroth-order Arrhenius kinetics. For the dilute sprays, droplet sizes below 100\textit{$\mu $m} are increasingly effective in reducing the flame temperature.   

Vorticity, Strain Rate, and Scalar Gradient Dynamics in Premixed Reacting Flows

The interactions between turbulence and flames in premixed, stoichiometric hydrogen-air combustion are studied as a function of turbulence intensity by analyzing the coupled dynamics of the vorticity, strain rate, and scalar (reactant mass fraction) gradient. The analysis is based on fully compressible numerical simulations of statistically planar flames at a range of intensities, where the intensity is characterized by the turbulent \emph{rms} velocity in the unburned mixture with respect to the laminar flame speed. The simulations have been carried out using the reactive-flow code Athena-RFX, and high numerical resolution allows the dynamics within the flame to be studied using conditional diagnostics based on the local, instantaneous value of the scalar. Particular emphasis is placed on the magnitudes and relative alignments of the vorticity, strain rate, and scalar gradient, which give insights into the interactions between these quantities in the presence of heat release effects. The analysis shows that there are substantial variations in the dynamics with both the turbulence intensity and position in the flame. The implications of these results for understanding the structure and evolution of premixed flames, particularly when the turbulence intensity is large, are discussed.   

Characteristics of Edge Flames in Microcombustors

Two streams, one containing fuel and the other oxidizer, are flowing into a relatively narrow channel where they mix and support an edge flame at some distance downstream. Our analysis is based on two models; one that fully couples the fluid dynamics and transport equations, used to determine the flame shape and location, and the other that assumes a constant-density flow, used to test the steady solutions for stability. It is found that in relatively wide channels the flame has a premixed, rounded edge with a trailing diffusion flame, but when the channel width decreases the flame is located further away from the supply and has a broader edge that can span the entire channel, when its width becomes comparable to the characteristic flame thickness. The effect of thermal expansion is to relocate the edge flame closer to the reactant supply. Heat losses at the channel walls cause a drop in the overall temperature and, as a result, the edge flame is confined to the center of the channel and the trailing diffusion flame is shortened significantly. Depending on the Lewis number, the flow rate, and the extent of heat loss, the edge may either remain steady, oscillate, or be blown off by the flow. With appreciable heat losses, residual fuel and oxidizer are observed at the end of the channel, so that under appropriate conditions, they could re-ignite and support a streak of diffusion flamelets, as seen experimentally.