Predicting the occurrence of thermoacoustic instabilities is of major interest in a variety of engineering applications such as aircraft propulsion, power generation, and industrial heating. Predictive methodologies based on a physical approach have been developed in the past decades, but have a moderate-to-high computational cost when exploring a large number of designs. In this study, the stability prediction capabilities and computational cost of four well-established classification algorithms—the K-Nearest Neighbors, Decision Tree (DT), Random Forest (RF), and Multilayer Perceptron (MLP) algorithms—are investigated. These algorithms are trained using an in-house physics-based low-order network model tool called OSCILOS. All four algorithms are able to predict which configurations are thermoacoustically unstable with a very high accuracy and a very low runtime. Furthermore, the frequency intervals containing unstable modes for a given configuration are also accurately predicted using multilabel classification. The RF algorithm correctly predicts the overall stability and finds all frequency intervals containing unstable modes for 99.6 and 98.3% of all configurations, respectively, which makes it the most accurate algorithm when a large number of training examples is available. For smaller training sets, the MLP algorithm becomes the most accurate algorithm. The DT algorithm is found to be slightly less accurate, but can be trained extremely quickly and runs about a million times faster than a traditional physics-based low-order network model tool. These findings could be used to devise a new generation of combustor optimization tools that would run much faster than existing codes while retaining a similar accuracy.

The combustion instability characteristics in a model dump combustor with an exhaust nozzle were experimentally investigated. The first objective was to understand the effects of operating conditions and geometric conditions on combustion instability. The second objective was to examine more generalised parameters that affect the onset of combustion instability. Three different premixed gases consisting of air and hydrocarbon fuels (C2H4, C2H6, C3H8) were burnt in the dump combustor at various inlet velocity, equivalence ratio and combustion chamber length. Dynamic pressure transducer and photomultiplier tube with a bandpass filter were used to measure pressure fluctuation and CH* chemiluminescence data. Peak frequencies and their maximum power spectral densities of pressure fluctuations at same equivalence ratios showed different trends for each fuel. However, the dynamic combustion characteristics of pressure fluctuations displayed consistent results under similar characteristics chemistry times regardless of the used hydrocarbon fuels. The results showed that characteristic chemistry time and characteristic convection time influenced combustion instabilities. It was found that the convective-acoustic combustion instability could be prevented by increasing the characteristic chemistry time and characteristic convection time.

In this study, the combustion instability and emission characteristics of flames of different H2/CH4 compositions were investigated in a partially premixed model gas turbine combustor. A mode shift in the frequency of instability occurred under varying experimental conditions from the first to the seventh mode of longitudinal frequency in the combustor, and a parametric study was conducted to determine the reasons for this shift by using the length of the combustor, a factor that determines the mode frequency of longitudinal instability, as the main parameter. Furthermore, heat load and fuel composition (H2 ratio) were considered as parameters to compare the phenomenon under different conditions. The GRI-3.0 CANTERA code, OH chemiluminescence and the Abel inversion process were applied to analyse the frequency mode shift. NOx emissions, which occurred through the thermal NOx mechanism, increased with increasing heat load and H2 ratio. The instability frequency shifted from the first to the seventh mode as the H2 ratio increased in the H2/CH4 mixture. However, 100% H2 as fuel did not cause combustion instability because it has a higher burning velocity and extinction stretch rate than CH4. Furthermore, the laminar flame speed influenced the frequency mode shift. These phenomena were confirmed by the flame shapes. The Abel inversion process was applied to obtain the cross section of the flames from averaged OH chemiluminescence images. Stable and unstable flames were identified from the radial profile of OH concentration. The combustor length was found to not influence frequency mode shift, whereas the H2 ratio significantly influenced it as well as the flame shape. The results of this experimental study can help in the reliable operation of gas turbine systems in SNG plants.

The main aims of this research are, at first, combustion instability study based on equivalence ratio oscillation, and, secondly, investigation various frequency modes of combustion instability, taking combustion chamber geometry into account. Considering the configuration of the simulated combustion chamber, excitation probability of the longitudinal modes is higher than that of transversal modes. The reason of this fact is that the resonance frequency values of the longitudinal modes are less than those of transversal modes. So, the most important frequency mode, during combustion instability, is the first longitudinal mode. In this paper thermo-acoustic instability model is utilized for pre-mixed gas turbines combustion chamber, founded on equivalence ratio oscillation. For this purpose Lieuwen method is developed in order to attain the phase difference between pressure and heat release oscillations. Results concluded from combustion instability simulation for the first longitudinal mode, considering its importance, are compared with experimental data and good agreement is observed.

The importance of nonreactive acoustic admittance at a burner exit plane on the flame-driving characteristic was discussed using the measured phase difference between the velocity and pressure oscillations, and the acoustic wave structure inside the duct. Also, the flame-driving characteristics were investigated by flame radiation measurements, and results were compared with those obtained previously from admittance studies. Results show that the comparisons are in good agreement. Also, the flame-driving characteristics are closed related to the acoustic wave structure at the burner exit plane, and could be qualitatively predicted by knowing the imaginary part of nonreactive burner acoustic admittance.

This research applied the method of direct acoustic admittance measurement to investigate effects of the flame character and the burner acoustic admittance on the flame-driving behavior. The acoustic admittance data were obtained by directly measuring the amplitudes of pressure and velocity oscillations, and the phase relations between the pressure and velocity oscillations. Results of this research reveal that the flame-driving characteristic is mainly dominated by the nonreactive burner admittance not the flame. Whether the flame is to drive or damp the acoustic field is detcrmined by the acoustic admittance of the burner system. The flame equivalence ratio plays a role mainly in determining the magnitude of the flame driving/damping.