Signal-processing-based characterization of thermoacoustic coupling in a lean premix swirling gas turbine combustor

Abstract

This study investigates the strength of the thermoacoustic feedback coupling between the acoustic field and the thermal field in a combustor with the coherence function between dynamic pressure (DP) and heat release rate (HR) signals. The time domain signals of DP and HR - which are measured synchronously from a lean premix combustor experiencing combustion instability (CI) - were used. The coherence function between DP and HR signals in the time-frequency domain reflects the degree of linear interdependence and the mutual causality of signals from the two fields. The results of the coherence analysis revealed the following: (i) the coherences are far from unity due to random properties in both signals during stable combustion, thus implying decoupling of the two fields; (ii) they are close to unity around the acoustic eigenfrequencies of the combustor during unstable combustion, thus demonstrating the linear interdependence and mutual causality between the two signals frequency-wise; and (iii) the formation of nearly linear interdependence between the two signals at the first eigenfrequency precedes the onset of limit cycle oscillation. Based on these findings, we propose an equivalent harmonic oscillator model for a combustion process under severe instability and formulate the damping ratio expression accordingly.