Effect of Different Fuels on Combustion Instabilities in an Annular Combustor

Abstract

Combustion instability in annular combustors of jet engines is a recurring issue. In the present study, the characteristics of instabilities for different fuels are investigated by combining the instability maps obtained in a laboratory-scale annular combustor equipped with multiple swirling spray injectors (MICCA-Spray) and flame describing functions (FDFs) from a single sector configuration (SICCA-Spray). Two types of liquid fuels are injected as hollow cone sprays: heptane, which is fairly volatile, and dodecane, which is less volatile. Experiments are also conducted with gaseous propane, premixed with air, which serves as a reference. An instability map is systematically drawn by varying the global equivalence ratio and thermal power. The data indicate that the amplitude and frequency of instabilities depend, for the same operating point, on the fuel injection conditions and fuel type. Overall trends show that premixed propane is unstable in a broad operating domain. Injection of liquid fuels induce changes in flame time lag that modify the unstable regions. For heptane, the instability map is closer to the propane reference map, whereas dodecane exhibits wider stable regions. An attempt is made to understand these features by examining the FDF, which gives the ratio of relative fluctuations in heat release rate to the relative fluctuations in velocity. The FDFs measured in a single sector configuration give access to gain and phase information that can be used to determine unstable bands and calculate an instability index guiding the interpretation of the differences in instabilities of the three fuels.

Swirler effects on combustion instabilities analyzed with measured FDFs, injector impedances and damping rates

“…Mixing between air and propane is achieved far upstream in the air line. In this experiment the spray atomizer is inoperative and the flame has a color and shape similar to the spray flame (shown in [53]). A small variation in (I(CH * )/I(OH * ))/(I(CH * )/I(OH * )) can be observed in the premixed case as well, at certain instants of the cycle.…”

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

“…The inner and outer diameters of the annulus are 300 and 400 mm, respectively. Strong azimuthal instabilities were first observed with a 200 mm long inner tube and an outer tube 600 mm long [10] or 700 mm long [29], but in recent studies [28], strong and persistent azimuthal instabilities were recorded with lateral tubes of equal length (400 mm). The instabilities of MICCA-Spray have until now been interpreted using FDFs measured in the single-injector facility, SICCA-Spray.…”

“…1 (a) since it forms the basis of the geometrical configurations for SICCA-Spray and TICCA-Spray. This system is presented in further detail in [28]. It is equipped with sixteen swirl-spray injectors and is now mainly fed with liquid heptane or dodecane.…”

“…However, this study was carried out at a single frequency, and the flames were only subjected to transverse acoustic modulations. Recent investigations carried out at the EM2C laboratory [28] indicate that FDFs measured in a single-injector cylindrical combustor (SICCA-Spray) having the same area as one segment of an annular combustion chamber (MICCA-Spray) can be used to analyze and approximately predict the instabilities observed in the annular combustor. However, it is also noted that the flame dynamics in the single-injector arrangement does not completely match that prevailing in the annular chamber.…”

Do flame describing functions suitably represent combustion dynamics under self-sustained oscillations?

“…With respect to the transfer function, the FDF allowed to render some of the complex behavior of practical systems and more specifically those linked to finite amplitude oscillations [46]. This has been a considerable advance because the describing function allowed to retrieve nonlinear dynamical features [47][48][49][50][51][52][53][54][55][56][57][58][59][60]. Models using the FDF yield amplitudedependent results that allow direct comparisons with experimental data since most instability experiments are carried out when the oscillations are established and have a finite value.…”