Scale Adaptive Simulation of a thermoacoustic instability in a partially premixed lean swirl combustor

Abstract

The thermoacoustic oscillation of a turbulent, swirl-stabilized, partially premixed flame in the PRECCINSTA gas turbine model combustor is analyzed by means of a Scale Adaptive Simulation (SAS) method. In the critical regions of the combustor the SAS features a fine spatial resolution and thus corresponds to a Large Eddy Simulation (LES). Two combustion models are applied, a simple eddy dissipation model and a detailed finite rate chemistry (FRC) model. A previous LES for the same combustor by Franzelli et al. [Combust. Flame 159 (2012) 621-637] indicated that the acoustic impedance of the fuel supply plays a critical role. Therefore in the present work, the fuel channels and fuel plenum are included in the computational domain and thereby the fuel inlet impedance is inherently taken into account. The resulting fields of velocity, temperature and mixture fraction fit well to experimental data with a slightly better agreement for the detailed FRC model. For both combustion models the computed frequency of the thermoacoustic oscillation is close to the experimental value, whereas its amplitude is significantly overestimated by about 15 dB in comparison to measurements. The reason for this overestimation is analyzed using an additional measurement where acoustic damping due to vibrating side walls is suppressed. For the latter experiment both frequency and amplitude agree well with CFD results, which indicates that acoustic damping effects must be carefully taken into account for validation of CFD. The 3D time-resolved simulations further provide detailed insights into the interaction of flow and mixing in the swirler, which leads to a convective time-lag between oscillations of velocity and equivalence ratio in the flow of unburned gas that largely affects the heat release response of the flame. Taken together, the results show that SAS computations can accurately reproduce frequency and amplitude of thermoacoustic oscillations of turbulent partially premixed flames in a gas turbine combustor provided that proper modeling of fuel supply and acoustic boundary conditions is applied.

Interaction between self-excited oscillations and fuel–air mixing in a dual swirl combustor

“…1b. The mass flow rates in Table 1 at the fuel and air inlets, which are fully reflective, are specified [12]. Sensitivity of the far-field outlet boundary location from the combustor exhaust duct exit (see Fig.…”

“…Another possible reason is that the plenum which supplies methane to the fuel nozzles can also play as a resonator for the oscillations. A numerical study [12] showed improved results with the fuel plenum included for the PRECCINSTA burner and this effect on the dual swirl burner will be explored in a future work. The computed peak PSD agrees quite well with the measured values and the LES is also able to reproduce the subharmonic and higher frequencies as in Fig.…”

“…The compressible Large Eddy Simulation (LES) is becoming a promising tool to predict self-excited instabilities since the unsteady heat release and its coupling with the acoustics are intrinsically dealt with by the LES equations and sub-models without assuming a compact flame and acoustic linearity. Past LES studies of self-induced oscillations in laboratory [10][11][12] and practical [13,14] combustors provided insights into the physical mechanism of these oscillations [3].…”

Time-resolved study of transient soot formation in an aero-engine model combustor at elevated pressure

“…In order to support interpretation of the complex spatial structures seen in the OH-PLIF measurements, results from a recent detailed LES of the same flame by Grader et al [7] are employed. The LES uses the DLR finite-volume solver THETA that is optimized for combustion simulations on unstructured grids [15]. In addition to the momentum equations, transport equations for species mass fractions and specific enthalpy are solved fully coupled in order to ensure that internal feedback within the gas phase chemistry is inherently captured.…”

Model-Based Approach for Combustion Monitoring Using Real-Time Chemical Reactor Network

“…Da nonuniformity can also lead to the onset of flow instabilities resulting in LBO [9]. Experimental and numerical studies have investigated species distribution in different combustor geometries, injector designs and fuel compositions e.g., [10][11][12][13][14][15][16][17]. Lean flame blowout modeling for an aeroengine application using zonal modeling was reported in [18][19][20].…”