CORRELATION OF THE EFFECTS OF FUEL TYPE ON GAS TURBINE COMBUSTOR EFFICIENCY

PAUL A LEONARD, Purdue University

Abstract

The efficiency of conventional diffusion flame gas turbine combustors is analyzed using semi-empirical correlating techniques. Times characteristic of mixing and of carbon monoxide and unburned hydrocarbon quenching are assigned to model regions of the combustor where inefficiency is controlled by mixing. These times allow inefficiency to be scaled with combustor geometric parameters and operating conditions when fuel droplet evaporation effects are small and primary zone performance is not limiting. The effects of slow fuel droplet evaporation, observed as increased emissions and inefficiency as droplet size or fuel boiling range increase, and most pronounced at low power conditions, are not correlated by the mixing-controlled model. Detailed measurements are made in the exit plane of a disc-stabilized model combustor with propane, JP-5 and Diesel Fuel Marine. These data reveal that the mechanism by which fuel properties and atomization manifest their influence is primarily through redistribution of fuel into regions of the combustor where it is susceptible to quenching. This is confirmed in the examination of the structure of a JP-5 flame. It is also observed that when the primary zone becomes too fuel rich its performance becomes important to combustor emissions. Fuel effects observed in four conventional turbine combustors are correlated by an extension of the basic mixing-controlled inefficiency model. A perturbation term estimated as the ratio of the fuel droplet lifetime to the available mixing time is introduced into the basic model. The droplet lifetime is estimated as the ratio of the square of the Sauter mean diameter of the fuel spray and the evaporation coefficient, and thus is related to nozzle performance and to fuel physical properties of viscosity, boiling point, specific gravity, and surface tension. Either of two perturbation models is found to provide good correlation of the available data. The preferred correlation better reflects changes of the sensitivity of emissions to evaporation processes as combustor operating condition is varied. Additional exercise of the models suggests they may be used to study the departure of primary zone performance from conditions when it does not control combustor inefficiency. From this analysis it may be inferred there are trade-offs in primary zone design between low inefficiency at low power and low soot emissions at high power.

Degree

Ph.D.

Subject Area

Mechanical engineering|Energy

COinS