Comparison of Two Methods for Flame Transfer Function Measurements: MOOG Valve or Siren and the Impact of Results on the Instability Analysis in a High Pressure Combustor

Abstract

Thermoacoustic instabilities may occur in every gas turbine combustor and could be hazardous to the flame stability and the structural integrity. It is important to be able to predict how hazardous the instabilities are: at what frequencies will they occur and will they develop into high amplitude limit cycle oscillations? The former question can be answered with the help of the Flame Transfer Function (FTF). The FTF establishes the coupling between burner passage aerodynamics and combustion dynamics and can be used as an input to an acoustic model to predict the eigenfrequencies and their growth rate. In the present research two methods to measure the FTF are used with different signal excitation instruments: a MOOG Valve and a Siren. Both the methods are based on data from pressure transducers only. The FTF is measured here by determining the combustor pressure response of the flame to fluctuations in the fuel mass flow at the burner exit. A siren unit has been developed and mounted at the upstream end of the fuel supply line of a pressurized combustor and is designed to have a harmonic excitation. The experimental method to measure the FTF by means of factorization in known or measurable sub-functions is briefly explained. Subsequently the Siren method is demonstrated by means of extracting the FTF at elevated pressure and as a function of thermal power. The results are compared with the results obtained in previous work of a MOOG valve excitation unit. The experimental investigation of the FTF is carried out in a high pressure combustor rig named DESIRE which is able to perform thermoacoustic measurements up to 500 kW thermal power at 5 bar absolute pressure. The results are compared and discussed. Subsequently a 1-D acoustic network model is presented which predicts the onset of the limit cycle pressure oscillations in the DESIRE combustor, using the FTF as an input. Thermo viscous damping effects and measured reflection coefficients are also included into the network model to improve the model predictions. Finally, the measured and predicted dynamic behavior of the combustor are compared. The results indicate that the network modeling approach is a promising design tool as it gives good agreement between measured and predicted dynamic behavior of the combustor and instability analysis. Well-defined boundary conditions and thermo viscous damping effects are important for the accuracy of the acoustic network models.