The main objective was to compare a flame stabiliser at constant pressure loss and identical isothermal aerodynamics with three modes of fuel injection: premixed, direct propane injection and direct kerosene injection. A Jet Mixing type of flame stabiliser was used at simulated gas turbine primary conditions. The influence of gaseous mixing effects was to deteriorate the combustion efficiency solely by increasing the CO emissions and to increase the NOx emissions. The flame stability was increased and low CO emissions were achieved at weaker mixtures. Liquid fuel atomisation effects resulted in a further deterioration in combustion efficiency due solely to un-burnt hydrocarbons. However, the NOx emissions were reduced indicating that local stoichiometric burning around single droplets does not occur.

Shear layer turbulent fuel and air mixing has been utilised in a simulated gas turbine primary zone combustor. Two methods of fuel injection and two values of the number of air injection holes have been investigated at a constant pressure loss of 4% at a reference Mach number of 0.047. The method of fuel injection and the number of air injection holes was found to influence the flame stability and NOx emissions. A large number of holes produced much higher NOx emissions which was not compensated for by the ability to operate at weaker equivalence ratios due to the greater flame stability. An optimum primary zone operating condition, for very low NOx and high combustion efficiencies involving a flame temperature of approximately 1600K was identified and there was a wide flame stability margin on this condition.

Shear layer turbulent fuel and air mixing has been utilised in a simulated gas turbine primary zone combustor. Two methods of fuel injection and two values of the number of air injection holes have been investigated at a constant pressure loss of 4% at a reference Mach number of 0.047. The method of fuel injection and the number of air injection holes was found to influence the flame stability and NO x emissions. A large number of holes produced much higher NO x emissions which was not compensated for by the ability to operate a weaker equivalence ratios due to the greater flame stability. An optimum primary zone operating condition, for very low NO x and high combustion efficiencies involving a flame temperature of approximately 1600K was identified and there was a wide flame stability margin on this condition.

Grid plate flame stabilizers for low NOx emissions have renewed interest in recent years due to their use in low NOx hydrogen gas turbine combustors. For non-premixed grid plate combustion, the difference in flame stabilizer design is in how the grid plate air flow is fueled. In the present work a simple four hole grid plate is investigated using CFD with three methods of fueling the air holes: radially inward fuel injection using 8 fuel nozzles per air hole (Grid Mix, GM 1 and Micromix); central fuel injection (FLOX method); and through a fuel annulus around each air hole (GM2). ANSYS FLUENT CFD predictions for GM2 are compared with axial gas composition traverses inside the combustor and with the mean combustor exit plane emissions. The three methods of fuel injection are also compared using isothermal CFD to determine which of the three methods offer the best mixing quality, which controls the relative NOx emissions. The predictions were for an equivalence ratio of 0.624 for the combustion stage and 0.5 for the isothermal study, using industrial propane. CFD modelling used RANS simulation with Realizable k-epsilon turbulence model, non-premixed combustion with the Steady Laminar Flamelet model. The temperature and mixing profiles obtained for GM2 were in reasonable agreement with the experiments and the other two fuel injection methods were then compared with GM2.