Flamelet Motion in Premixed Turbulent Axisymmetric Rod-Stabilized and Bunsen Flames

Abstract

A four-element electrostatic probe was employed to examine flamelet motion in axisymmetric propane-air flames at normal atmospheric pressure and temperature. Besides a number of turbulent (rim-stabilized) Bunsen flames, a turbulent V-shaped flame, stabilized by a vertical rod placed along the axis of the Bunsen-type burner, was investigated. Measurements were conducted at three different positions in the Bunsen flames (i.e., one on the centerline and two off axis, the lowest being at half the height of the centerline position) and at two positions in the V flame (roughly corresponding to the off-axis positions of the Bunsen flames). For the Bunsen flames, measurements were made at three different propane-air equivalence ratios, two of them (one lean and the other rich) having the same laminar burning velocity. To aid in interpretation of the propane-air results for these last two measurements, Bunsen-flame measurements also were made for two (lean and rich) pairs of methane-air flames having the same laminar burning velocities. Fuel flow rates and turbulence intensities and scales were essentially the same in all experiments (fully developed turbulent pipe flow at a Reynolds number of 7000, with a ratio of mean velocity to laminar burning velocity on the order of ten). In all cases, scatterplots for flamelet velocity vectors were obtained when flamelets passed the probe in the direction of fresh-to-burnt mixtures and, separately, of burnt-to-fresh mixtures. Correlations between magnitudes and directions of flamelet velocities were measured, and probability distributions were generated for magnitudes and for directions of flamelet velocities. The results showed significant differences in distributions of magnitudes and of directions for fresh-to-burnt and for burnt-to-fresh passage. They also showed differences between the lean and rich propane flames but not between the lean and rich methane-flame pairs, the characteristics of which tended to resemble those of the lean propane flames. Finally, flamelet motions for the V flame were noticeably different in many respects from those for the Bunsen flames, as explained by the different effects of thermal expansion on the mean flow in the two configurations, the fresh mixture being radially on the inner side of the Bunsen flames and on the outer side of the V flame. The exceptional characteristics of rich propane were explained by nonlinear enhancement of the average flamelet burning velocity through preferential oxygen diffusion to turbulence-induced flamelet bulges convex toward the fresh mixture. The differences in the probability distributions for the two passage directions were thought to arise from the laminar burning velocity measured in the burnt gas being not negligible in comparison with average flamelet velocities, thereby influencing distribution shapes for burnt-to-fresh passage.