EXTINCTION AND TEMPERATURE CHARACTERISTICS OF TURBULENT COUNTERFLOW DIFFUSION FLAMES WITH PARTIAL PREMIXING

A turbulent counterflow diffusion flame of natural gas stabilized between two opposed jets discharging from straight tubes of 25.4 mm diameter fitted with turbulence-promoting perforated plates has been examined in terms of its appearance, extinction limits, and mean and fluctuating temperatures as measured by numerically compensated fine-wire thermocouples. It was observed that the flame was flat, blue, and located around the symmetry plane, that its appearance did not change with initial premixing of the fuel stream with air, and that a premixed flame could also be stabilized provided the equivalence ratio was smaller than that of the rich flammability limit. The bulk velocity for extinction of the nonpremixed flame increased with tube separation and with initial premixing, but decreased with an increase in turbulent intensity. The extinction data collapsed to a single curve to within 20% if the total strain rate acting on the flame (bulk plus small-scale turbulent) was plotted as a function of the air volume fraction in the fuel stream, implying a critical total strain rate for extinction that depended only on the degree of partial premixing. Partial premixing increased the resistance of the flame to straining, from about 350 s-1 for pure fuel to about 600 s-1 for an air volume fraction of 0.8, consistent with experiments and predictions for laminar counterflow flames and with experimental data from piloted turbulent jet flames. The present results for the total strain rate at extinction provide a quantitative description of the effect of partial premixing on flame stability and can be used to predict the extinction of nonpremixed flames in other geometries. The maximum mean temperature of the flame did not change as extinction was approached and was about 1300 +/- 50 K for all flow conditions measured, while the rms fluctuations were about 450 K; this insensitivity is attributed to a low-frequency (as indicated by high-pass filtering the temperature time series) flame motion, which also resulted in broad temperature probability density functions.