The Effect of Incomplete Fuel-Air Mixing on Spark-Ignited Flame Kernel Growth

Results from an experimental study of the effect of incomplete fuel-air mixing on spark-ignited flame kernel growth in turbulent propane-air mixtures are presented. The experiments were conducted in a turbulent flow system that allows for independent variation of flow parameters, ignition system parameters, and the degree of fuel-air mixing. Measurements were made at 1 atm and 300 K conditions. Five cases were studied; a premixed and four incompletely mixed cases with 6%, 13%, 24% and 33% RMS (root-mean-square) fluctuations in the fuel/air equivalence ratio. The overall fuel/air equivalence ratio was unity in all cases. The flow characteristics were measured by LDV. The RMS fluctuation in the fuel/air equivalence ratio was characterized using NO2-based laser induced fluorescence. High speed laser shadowgraphy at 4,000 frames-per-second was used to record flame kernel growth following spark ignition, from which the equivalent flame kernel radius as a function of time was determined. The effect of incomplete fuel-air mixing was evaluated in terms of the flame kernel growth rate, “cyclic” variations in the flame kernel growth, and the rate of misfire.

The results show that fluctuations in local mixture strength due to incomplete fuel-air mixing cause the flame kernel surface to become wrinkled and distorted; and that the amount of wrinkling increases as the degree of incomplete fuel-air mixing increases. Incomplete fuel-air mixing was also found to result in a significant increase in “cyclic” variations in the flame kernel growth. The average flame kernel growth rates for the premixed and the incompletely mixed cases were found to be within the experimental uncertainty except for the 33%-RMS-fluctuation case where the growth rate is significantly lower. The premixed and 6%-RMS-fluctuation cases had a 0% misfire rate. The misfire rates were 1% and 2% for the 13%-RMS-fluctuation and 24%-RMS-fluctuation cases, respectively; however, it drastically increased to 23% in the 33%-RMS-fluctuation case.