Separation of Hydrodynamic, Entropy, and Combustion Noise in a Gas Turbine Combustor

Abstract

This paper deals with noise sources which are centra! to the problem of core engine noise in turbopropulsion systems. The sources dealt with are entropy noise and direct combustion noise, as well as a nonpropagating psuedosound which is hydrodynaniic noise. It is shown analytically and experimentally that a transition can occur from a combustion noise-dominant situation to an entropy noise-dominant case if the contraction of a terminating nozzle to the combustor is high enough. In the conibustor tested, entropy noise is the dominant source for propagationai noise if the combustor is choked at the exit. It is also speculated that there might be another unexplored noise source interior of the combustor. Analysis techniques include spectral, cross spectral, correlation, and ordinary and partial coherence analysis. Measurements include exterior and interior fluctuating and mean pressures and temperatures. T has been known for some time that there are at least two probable causes for core noise—entropy or indirect noise and direct combustion noise.! Direct combustion noise is caused by a fluid dilatation caused by a fluctuating heat release. Entropy noise is caused by hot (or cold) spots passing through the pressure gradients of. the turbine assembly. Both noise sources have the same fundamental cause—heat release fluctuations—but they are formed in a different manner. Entropy noise depends upon the heat release history following a fluid element through the combustor, whereas combustion noise depends on the instantaneous aggregate heat release rate fluctuation. The purpose of this program was to isolate the two suspected causes of core noise and determine their relative importance to the core noise problem. It is clear that core noise presents a noise floor in current turbopropulsion systems, but there is controversy concerning the strength of core noise relative to other sources.2'4 It is sufficient to remark here that core noise exists and is measurable. Entropy and combustion noise are not the only possibilities for core noise. Another possibility is vorticity- nozzle interaction nose,5 which is essentially a resistance of a nozzle to pass an axial velocity fluctuation. While sources other than entropy and combustion noise are not directly investigated here, the analysis techniques do reveal whether or not entropy and combustion noise are dominant over other sources. The analysis techniques presented essentially try to relate processes taking place inside a combustor to the noise radiated to the surroundings. One troublesome problem encountered in a prior program6 was the contamination of interior pressure fluctuation measurements by non- propagating psuedosound (hydrodynamic noise). Another purpose of this paper was to eliminate this contaminant as much as possible in order to concentrate on propagationai sound and its causes.