Turbulent Flame Propagation Characteristics Of High Hydrogen Content Fuels

This final report describes the results of an effort to better understand turbulent flame propagation, especially at conditions relevant to gas turbines employing fuels with syngas or hydrogen mixtures. Turbulent flame speeds were measured for a variety of hydrogen/carbon monoxide (H2/CO) and hydrogen/methane (H2/CH4) fuel mixtures with air as the oxidizer. The measurements include global consumption speeds (ST, GC) acquired in a turbulent jet flame at pressures of 1-10 atm and local displacement speeds (ST, LD) acquired in a low-swirl burner at atmospheric pressure. The results verify the importance of fuel composition in determining turbulent flame speeds. For example, different fuel-air mixtures having the same unstretched laminar flame speed (SL,0) but different fuel compositions resulted in significantly different ST, GC for the same turbulence levels (u'). This demonstrates the weakness of turbulent flame speed correlations based simply on u'/SL,0. The results were analyzed using a steady-steady leading points concept to explain the sensitivity of turbulent burning rates to fuel (and oxidizer) composition. Leading point theories suggest that the premixed turbulent flame speed is controlled by the flame front characteristics at the flame brush leading edge, or, in other words, by the flamelets that advance farthest into the unburned mixture (the so-called leading points). For negative Markstein length mixtures, this is assumed to be close to the maximum stretched laminar flame speed (SL, max) for the given fuel-oxidizer mixture. For the ST, GC measurements, the data at a given pressure were well-correlated with an SL, max scaling. However the variation with pressure was not captured, which may be due to non-quasi-steady effects that are not included in the current model. For the ST, LD data, the leading points model again faithfully captured the variation of turbulent flame speed over a wide range of fuel-compositions and turbulence intensities. These results provide evidence that the leading points model can provide useful predictions of turbulent flame speed over a wide range of operating conditions and flow geometries.