Alessandro Gomez

The present study experimentally investigates non-flamelet behavior leading to extinction in a NonPremixed Flame (NPF) and a Partially Premixed Flame (PPF) in a turbulent counterflow configuration. Both flames at Re t ∼ 900 are indistinguishable in terms of the turbulence properties that are imposed at the cold boundaries and persist up to the mixing layer. For each extinction event, the perturbation that leads to the first breach of the OH layer is tracked back in time in a Lagrangian manner using high-speed stereoscopic PIV and OH-PLIF imaging techniques, allowing the reconstruction of the time sequence of strain rate and vorticity that leads to flame extinction. The NPF is found to be more prone to extinction than the PPF, which is at odds with the computed laminar extinction strain rate, which is 23 % larger in the NPF than in the PPF, implying a greater resistance to strain for the NPF. Extinction appears to be caused by a combination of relatively intense strain rate and/or vorticity pockets interacting with the flame and causing a tear of the OH layer. The maximum strain rate norm exceeds the computed extinction limit in >85 % of the cases for the PPF, whereas it does so in approximately 75 % of the cases for the NPF, revealing a more pronounced strain rate effect on the extinction process for the PPF. Vorticity plays multiple roles in extinction. It manifests itself as: i) a pair of counterrotating vortices approaching the flame from either or both sides, creating a region of high strain rate; ii) a single vortex interacting with the flame by diluting the reactants with inert, thereby, weakening the flame; and iii) a vortex penetrating the oxidizer layer, making it thicker and, possibly, disrupting the laminar flame structure. The last two scenarios may explain extinction without a history of remarkably high strain rates. A 35 % reduction in the overall strain rate of the flames resulted in nearly complete suppression of extinction events in both PPF and NPF.

An experimental investigation was performed on self-sustained, axis-symmetric, laminar diffusion flames in which the fuel, heptane, was injected as a monodisperse spray in a co-flow of air. The appearance of such flames was in all respects similar to that of a candle. The inner core of the flame was a relatively dark region, in which the droplets simply evaporated at relatively low temperatures, but did not burn. This inner core was enveloped by a higher temperature region where soot was present. The flame was located immediately outside of the sooty region and was characterized by a soft, deep blue luminescence. The droplet life-history along the centerline of three flames was examined by phase Doppler anemometry and sizing, as well as thermocouple measurements. It was found that the D-square law with constant evaporation coefficient was satisfied over a significant portion of the droplet lifetimes. The evaporation coefficients extracted from such measurements range from 0.53 to 0.60 mm²/s, values that are on average 35 % to 45% lower than those calculated for an isolated droplet in a convective atmosphere of N₂. It was also shown that an average evaporation coefficient for the droplets in the flame could be determined by simply measuring the height of the dark, vaporization region and the liquid flow rate, without the need for in-situ measurements. No evidence of individual droplet burning was found. Rather, conditions were more akin to what is often labeled sheath combustion, with a continuous flame enveloping the entire droplet region and a relatively cooler core in which droplets are evaporating but not burning. Droplet-droplet interactions, although certainly present near the flame surface, do not seem to be significant enough in the flame core to cause a departure from the D-square law.