The objective was to study the effect of equivalence ratio of secondary stage combustible mixture injected into the cross flow stream of vitiated air in a two staged combustion system on the characteristics of the secondary stage combustion zone. The primary cylindrical combustor equipped with low swirl air blast nozzle operating with kerosene generates vitiated air. A methane injector was flush mounted to the inner surface of the secondary combustor. It was used to inject the premixed methane-air mixtures perpendicular into the crossflow of vitiated air. An optical, double shell, secondary combustor with three optical windows on its outer shell was used to image the secondary stage flames. The inner shell was a quadratic fused quartz tube which acts as a thermal barrier and the outer thick quartz windows mounted in the quadratic stainless steel chamber withholds the pressure. Chemiluminescence imaging technique equipped with ICCD camera was used to image the OH* emissions of the secondary stage flame.
The vitiated air was generated at 2 bar and 1700 K. The velocity of the vitiated air in the secondary combustor was 57 m/s. A premixed methane air mixture was injected into the cross flow stream of vitiated air. The momentum flux ratio between the jet and the vitiated air was maintained at 1.4. The equivalence ratio of the premixed methane-air mixture was varied from 0.5 to 1.0. As the equivalence ratio of the secondary stage combustible mixture moves towards stoichiometric condition, the secondary stage combustion zone becomes compact and also the distance between the burner and the combustion zone decreases.
The turbulent flame stabilized in the secondary combustor exhibited large scale structures and other unsteady phenomena that require time-resolved computational methods. Large eddy simulations (LES) are well suited to the calculation of such complex flows. The flame was embedded in a strong turbulent flow where auto-ignition and quenching are important, which poses a significant challenge for the reaction modeling. The presumed JPDF turbulent reaction model, which has been proven to be a reliable model for these challenging conditions, was successfully coupled with the LES simulation. The qualitative agreement between the results of simulations and measurements was quite satisfactory.