Comparison of Flame Describing Functions Measured in Single and Multiple Injector Configurations

Abstract

Recent investigations of combustion instabilities in annular systems indicate that considerable insights may be gained by using information gathered in singlesector experiments. Such experiments are, for example, employed to measure flame describing functions (FDFs), which represent the flame response to incident perturbations. These data may be used in combination with loworder models to interpret instabilities in multiple injector annular systems. It is known, however, that the structure and dynamical behavior of an isolated flame do not necessarily coincide with those of a flame placed in an annular environment with neighboring side flames. It is then worth analyzing effects that may be induced by the difference in lateral boundary conditions and specifically examining the extent to which the FDF data from singlesector experiments portrays the dynamical response of the flame in the annular environment. These issues are investigated with a new setup named TICCASpray, which comprises a linear arrangement of three injectors. The central flame is surrounded by two identical side flames in a rectangular geometry with key dimensions, sidewall separation, and spacing between injectors identical to those of the annular system MICCASpray. The describing function of the central flame is determined with techniques recently developed in singlesector experiments (SICCASpray). The FDFs obtained in the two configurations are compared for two swirler types having different swirl numbers and pressure drops. The effect of the swirl direction of the neighboring injectors is also explored by operating with co and counterswirl combinations. Differences between FDFs determined in the two test facilities, sometimes modest and in other cases less negligible, are found to depend on the flames' spatial extension and interactions. The general inference is that the FDFs measured in a singleinjector combustor are better suited if the flamewall interaction is weak, and provided that the area is equivalent to that of a single sector of an annular combustor. Nonetheless, using a multiinjector system would be more appropriate for a more precise FDF determination.