Experimental and numerical studies on the premixed oscillating flame have been conducted in order to examine the kinematic behavior of the premixed flame and the flame/flow interaction under the flow oscillation.
The flame front behavior under the oscillating flow field is observed experimentally. Lean propane/air flame with conical shape is stabilized inside the tube. Flow oscillation is driven by a loud speaker and oscillation frequency is 30 Hz. The flame motion at each phase of oscillation is captured by the triggered ICCD camera system to explain the kinematic response of the flame. Streak line of the flow field is also visualized by the Mie-scattering method. Upstream velocity oscillates sinusoidal with the same order of magnitude as mean velocity. The flame is also oscillating with the same frequency as that of the flow oscillation, while its conical shape is maintained during the oscillating period. The magnitude of flame surface oscillation near the tube wall is about 1/6 to the initial flame height, but the magnitude at the tube center is nearly zero. It is a very notable phenomenon that the center of conical flame does not respond while the upstream flow is entirely oscillating across the tube. Velocity measurements show that the upstream axial velocity oscillation is largely attenuated and become zero at the flame zone along the tube center. Streak line visualization and radial velocity measurement show that this axial velocity attenuation results from the flame surface oscillation. From the above experimental observation, it is known that in case of the curved flame front the upstream flow characteristics are greatly influenced by the flame motion.
Wrinkled flame front in oscillating flow field is numerically simulated to investigate the flame/flow interaction. A hydrodynamic model is used to simulate the reaction front behavior and the flame/flow interaction in oscillating channel flow. The amplitude of the oscillating velocity is the same order of magnitude as that of mean velocity. Initially conical flame of which curvature is negative at the center is located in the 2-dimensional channel. Flame is regarded as an interface which has an appropriate flame speed. Markstein relation Su=Su˚(1-LK)$for flame speed is used to decouple the flame structure from the flow field. General velocity decomposition principle, according to which a flow field is comprised of volumetric(due to volume source on the flame surface), rotational(due to vorticity production across the flame), and potential(due to upstream oscillating velocity) components, naturally leads to a model in which the flame is regarded as a surface source of volume. The flame surface is determined by conventional G-equation which is the level-set approach to track the interface. Upstream flow characteristics, which would be addressed without combustion are modified by the volume expansion and by the vorticity production across the flame through interaction mechanism. Hayes''s expression (1959, JFM) for vorticity jump across the hydrodynamic discontinuity is adapted in this work. In this expression, vorticity is produced across the flame by the tangential gradient of local flame speed and by the unsteady motion of the flame surface. The numerical simulation clarifies that the flame surface is also oscillating as upstream flow oscillates, and oscillation frequency of the flame surface is the same as that of flow field. But the flame surface oscillation is remarkably attenuated at the center. It is found that the attenuation of the flame surface oscillation is mainly caused by the rotational velocity field due to the vorticity production across the flame. At the phases where oscillating velocity is increased, rotational velocity field due to vorticity production across the flame accelerates the flow near the wall but decelerates the flow near the center. At the phases where oscillating velocity decreases, rotational velocity field decelerates the upstream flow near the wall but accelerates the upstream flow near the center.
From the results of numerical simulation and the experimental observation, it is clearly seen that the upstream flow characteristics are modified by the motion of the curved flame surface and that this hydrodynamic interaction between flame and flow results from the vorticity production across the flame front. Consequently, it is concluded that the hydrodynamic model including the volume expansion and the vorticity production across the flame surface describes the flame/flow interaction reasonably, and that the vorticity production in this model would be the crucial parameter to explain the flame/flow interaction.
