Laboratory experiments were conducted at gas turbine and atmospheric conditions (0.101 < P 0 < 0.810 MPa, 298 < T 0 < 580K, 18 < U 0 < 60 m/s) to characterize the overall behaviors and emissions of the turbulent premixed flames produced by a low-swirl injector (LSI) for gas turbines. The objective was to investigate the effects of hydrogen on the combustion processes for the adaptation to gas turbines in an IGCC power plant. The experiments at high pressures and temperatures showed that the LSI can operate with 100% H 2 at up to φ = 0.5 and has a slightly higher flashback tolerance than an idealized high-swirl design. With increasing H 2 fuel concentration, the lifted LSI flame begins to shift closer to the exit and eventually attaches to the nozzle rim and assumes a different shape at 100% H 2 . The STP experiments show the same phenomena. The analysis of velocity data from PIV shows that the stabilization mechanism of the LSI remains unchanged up to 60% H 2 . The change in the flame position with increasing H 2 concentration is attributed to the increase in the turbulent flame speed. The NO x emissions show a log linear dependency on the adiabatic flame temperature and the concentrations are similar to those obtained previously in a LSI prototype developed for natural gas. These results show that the LSI exhibits the same overall behaviors at STP and at gas turbine conditions. Such insight will be useful for scaling the LSI to operate at IGCC conditions.

No abstract

No abstract

This paper describes the evaluation of an alternative combustion approach to achieve low emissions for a wide range of fuel types. This approach combines the potential advantages of a staged rich-burn, quick-mix, lean-burn (RQL) combustor with the revolutionary trapped vortex combustor (TVC) concept. Although RQL combustors have been proposed for low-Btu fuels, this paper considers the application of an RQL combustor for high-Btu natural gas applications. This paper will describe the RQL/TVC concept and experimental results conducted at 10 atm (1013 kPa or 147 psia) and an inlet-air temperature of 644 K (700°F). The results from a simple network reactor model using detailed kinetics are compared to the experimental observations. Neglecting mixing limitations, the simplified model suggests that NOx and CO performance below 10 parts per million could be achieved in an RQL approach. The CO levels predicted by the model are reasonably close to the experimental results over a wide range of operating conditions. The predicted NOx levels are reasonably close for some operating conditions; however, as the rich-stage equivalence ratio increases, the discrepancy between the experiment and the model increases. Mixing limitations are critical in any RQL combustor, and the mixing limitations for this RQL/TVC design are discussed.

No abstract