Combustion dynamics of multi-element lean-premixed hydrogen-air flame ensemble

2021 ◽  
Vol 233 ◽  
pp. 111585
Author(s):  
Hyebin Kang ◽  
Kyu Tae Kim
2003 ◽  
Vol 19 (5) ◽  
pp. 782-794 ◽  
Author(s):  
Ying Huang ◽  
Hong-Gye Sung ◽  
Shih-Yang Hsieh ◽  
Vigor Yang

Author(s):  
Steven M. Cannon ◽  
Virgil Adumitroaie ◽  
Clifford E. Smith

A lean premixed fuel injector/combustor typical of industrial gas turbine combustors has been analyzed using 3D Large Eddy Simulation (LES) methods. The objective of the study was to evaluate the 3D LES modeling approach for predicting combustion dynamics and compare it with simpler unsteady Reynolds Averaged Navier Stokes (RANS) methods using 2D and 3D analyses. Large amplitude pressure oscillations were observed experimentally at the modeled operating conditions, and previous 2D axisymmetric unsteady RANS analysis has shown reasonable, but not perfect, engineering agreement with pressure measurements. Although the pressure amplitude was accurately predicted, the frequency was substantially in error. This study sought to see if 3D modeling would improve the agreement. 2D axisymmetric and full 3D calculations were performed with a state-of-the-art, unstructured-grid, parallel (domain decomposition) CFD code. For the unsteady RANS calculations, the RNG k-ε turbulence model was employed, while for the LES calculation the Smagorinsky subgrid turbulence model was employed. Surprisingly, the 2D unsteady RANS, 3D unsteady RANS, and 3D LES calculations gave nearly identical pressure oscillation predictions, and all calculations had the oscillation frequency around 280 Hertz. This work has shown that smaller turbulent structures captured with 3D LES have very little effect on capturing combustion instability driven primarily by a fuel time-lag.


Author(s):  
Poravee Orawannukul ◽  
Jong Guen Lee ◽  
Bryan D. Quay ◽  
Domenic A. Santavicca

The response of a swirl-stabilized flame to equivalence ratio fluctuations is experimentally investigated in a single-nozzle lean premixed combustor. Equivalence ratio fluctuations are produced using a siren device to modulate the flow rate of fuel to the injector, while the air flow rate is kept constant. The magnitude and phase of the equivalence ratio fluctuations are measured near the exit of the nozzle using an infrared absorption technique. The flame response is characterized by the fluctuation in the flame’s overall rate of heat release, which is determined from the total CH* chemiluminescence emission from the flame. The relationship between total CH* chemiluminescence intensity and the flame’s overall rate of heat release is determined from a separate calibration experiment which accounts for the nonlinear relationship between chemiluminescence intensity and equivalence ratio. Measurements of the normalized equivalence ratio fluctuation and the normalized rate of heat release fluctuation are made over a range of modulation frequencies from 200 Hz to 440 Hz, which corresponds to Strouhal numbers from 0.4 to 2.8. These measurements are used to determine the fuel-forced flame transfer function which expresses the relationship between the equivalence ratio and rate of heat release fluctuations in terms of a gain and phase as a function of frequency. In addition, phase-synchronized CH* chemiluminescence images are captured to study the dynamics of the flame response over the modulation period. These measurements are made over a range of operating conditions and the results are analyzed to identify and better understand the mechanisms whereby equivalence ratio fluctuations result in fluctuations in the flame’s overall rate of heat release. Such information is essential to guide the formulation and validation of analytical fuel-forced flame response models and hence to predict combustion dynamics in gas turbine combustors.


Author(s):  
B. T. Chorpening ◽  
D. L. Straub ◽  
E. D. Huckaby ◽  
K. J. Benson

The implementation of sophisticated combustion control schemes in modern gas turbines is motivated by the desire to maximize thermodynamic efficiency while meeting NOx emission restrictions. To achieve target NOx levels, modern turbine combustors must operate with a finely controlled fuel-air ratio near the fuel-lean flame extinction limit, where the combustor is most susceptible to instabilities. In turbine configurations with multiple combustors arranged around the annulus, differences in flow splits caused by manufacturing variations or engine wear can compromise engine performance. Optimal combustion control is also complicated by changes in environmental conditions, fuel quality, or fuel type. As a consequence, engines must be commissioned in the field with adequate stability margin such that manufacturing tolerances, normally expected component wear, fuel quality, and environmental conditions will not cause unstable combustion. A lack of robust combustion in-situ monitoring has limited the ability of modern turbines to achieve stable ultra-low emission performance over the entire load range. Of particular concern is the avoidance of lean blowout (LBO) and combustion dynamics. To minimize combustion temperature and NOx production, it is necessary to approach the LBO boundary. This paper describes continuing work on incipient lean blowout detection using flame ionization, investigating the impact of three different piloting and equivalence ratio reduction strategies applied in a pressurized, lean premixed combustor. This work builds upon previous research in the development of the Combustion Control and Diagnostic Sensor (CCADS). In previous papers, the detection of flashback, equivalence ratio, combustion dynamics, and lean blowout using CCADS has been investigated and described. Previous investigation of lean blowout, however, has been limited to a side pilot configuration. In this paper, lean blowout behavior for a side pilot and a centerbody tip pilot are compared. In addition, two different methods for decreasing equivalence ratio to approach LBO are investigated. These cases are found to have differing lean blowout behavior, and differing CCADS signatures. This paper also reports on the ion signal behavior due to combustion dynamics observed during the equivalence ratio sweeps, including passing through stability boundaries. Tests were performed at 5 atm using an industrial style, lean premixed combustor nozzle, equipped with CCADS electrodes, in a water-cooled, natural gas fueled, acoustically noisy combustor. Testing included sweeps of equivalence ratio from 0.65 to 0.45, crossing one or more stability boundaries. LBO was approached for configurations with a side pilot (on the inlet wall of the combustor, but set away from the premixer) and a centerbody tip pilot. The centerbody tip pilot and the side pilot both helped stabilize combustion, but combustion dynamics still occurred. Incipient LBO was apparent in all cases; however, the different flame structure encountered with each pilot configuration and fuel control strategy made the flame ionization signature differ for each case.


Author(s):  
B. Anson ◽  
I. Critchley ◽  
J. Schumacher ◽  
M. Scott

The present paper describes an active control system consisting of a fast-acting actuator valve, coupled with a control algorithm capable of adaptive phase and amplitude control for pressure oscillation suppression. Experiments were conducted using two separate combustion test rigs: a small, lean premixed, tubular combustor (75 kW) and a larger premixed annular DLE system (4 MW). Active control of pressure oscillations at frequencies of approximately 90 Hz and 300 Hz was demonstrated on the 75 kW rig. Up to 90% reduction in single-frequency dynamic pressure amplitude and 70% reduction in peak-mean was achieved. Simultaneous suppression of these two distinct modes was also demonstrated. The system was also demonstrated on a full-scale 4 MW combustion rig, with peak-mean dynamic pressure reductions between 33–45%. Instability reductions were achieved by pulsing fuel supplied to either a diffusion or a premixed flame. The response of the flame was notably better for the premixed case. For the premixed flame, combustor pressure pulses were realizable up to a frequency of about 400Hz, while the diffusion flame could only induce combustor pressure pulses up to approximately 50Hz. Adaptive frequency, phase and amplitude logic were developed, allowing automatic selection of the optimal values of these parameters to maximize suppression efficacy to suit the particular operating condition.


Sign in / Sign up

Export Citation Format

Share Document