Study of noise-induced dynamics in model gas turbine combustors

Abstract

This thesis presents a comprehensive analysis of the stochastic dynamics in confined combustion systems, particularly gas turbine combustors, where thermoacoustic instability poses a significant challenge. Thermoacoustic instability refers to self-induced, large-amplitude pressure oscillations resulting from a constructive feedback loop between acoustic waves and unsteady heat release within the combustion system. These oscillations can lead to severe mechanical and thermal stresses that can compromise the structural integrity of the combustor. Noise, an inherent feature in such combustion systems, complicates the phenomenon by affecting the amplitude and behavior of these oscillations. Recent studies have highlighted that noise can lead to complex dynamics in such systems, which vary with changes in operating conditions and combustor designs. Hence, it is crucial to investigate the effect of noise properties on thermoacoustic coupling. Given the challenges posed by thermoacoustic instability, the study emphasizes on the necessity of considering noise properties—correlation time (or color) and intensity—in developing effective prediction, suppression, and control strategies. The work investigates how noise properties interact with thermoacoustic coupling, with a particular focus on early warning prediction of thermoacoustic instability and helical instabilities associated with swirling flows in practical gas turbine configurations. The thesis is divided into two main parts. The first part examines how noise characteristics influence the reliability of various early warning indicators (EWIs) for predicting thermoacoustic instability in gas turbine combustors. A combination of experimental and numerical approaches is employed, including an electroacoustic Rijke tube simulator, stochastic Van der Pol oscillators, and a lean premixed flat flame combustion system. The study explores how different types of noise, both additive and multiplicative, affect EWIs based on signal amplitude distribution, frequency spectra, fractal, and complexity measures, as the system approaches bifurcation, considering noise color and intensity. These systems exhibit instability through both supercritical and subcritical Hopf bifurcations. The analysis is conducted in the subthreshold regime, where the stable focus remains the only possible asymptotic state. The findings reveal that variations in noise color can lead to non-monotonic trends in EWIs, reducing their reliability. Our results indicate that the coherence factor is a reliable indicator for the entire range of investigated noise color, while variance and decay rates of the autocorrelation function (ACF) are reliable when noise correlation times are either much smaller or larger than the system time scale. Kurtosis, permutation entropy and Jensen-Shannon complexity can be effectively employed in systems where noise exhibits minimal correlation time (resembling white noise). While the Hurst exponent proves a reliable indicator in systems where noise has correlation times much larger than the time scale of the system, multi-fractal spectrum width and skewness are deemed unsuitable as EWIs. These insights enhance the understanding of the effectiveness and limitations of various EWIs in predicting or monitoring impending instability. The second part of the thesis focuses on how inherent noise interacts with turbulent swirling flows in combustors. Swirling flows, used in gas turbine combustors for flame stabilization, are prone to precessing vortex core (PVC), a self-excited global hydrodynamic instability associated with vortex breakdown. This instability leads to large-scale coherent structures and significant flame fluctuations, potentially triggering thermoacoustic instability. It is, therefore, crucial to study the interaction between these coherent structures and inherent combustor noise. The aim is to understand how PVC responds to broadband noise excitation, aiding the development of strategies to mitigate thermoacoustic instability. To this end, a novel multiple swirl burner is developed, capable of operating with various hydrogen-enriched fuel blends using RANS simulations. A dual swirl burner configuration showed promise due to its enhanced mixing capabilities. The effects of acoustic excitation on the swirling flow field are examined using Schlieren image velocimetry (SIV), with Proper Orthogonal Decomposition (POD) and Spectral Proper Orthogonal Decomposition (SPOD) analyses employed to identify dominant coherent structures and their interactions with acoustic excitation. The study confirms the presence of single and double helical PVC, marginally stable modes excited by turbulent fluctuations. Acoustic excitation at frequencies lower than the PVC mode is found to suppress the PVC, while broadband forcing excited both single and double helical instabilities, increasing the likelihood of thermoacoustic instability. Summarizing, this thesis emphasizes the crucial role of noise characteristics in predicting and managing thermoacoustic instability in gas turbine combustors. The findings offer valuable insights for enhancing the design and calibration of monitoring systems, contributing to more reliable and effective control strategies in modern gas turbine systems.