Experimental study of vortex-flame interaction in a gas turbine model combustor
Introduction
Modern gas turbine (GT) combustors are operated under lean premixed or partially premixed conditions in order to reduce emissions of NOx and CO. The highly turbulent flames in GT combustors are most often stabilized aerodynamically by swirl-induced vortex breakdown. This leads to recirculation of hot burned gases, which enhances ignition of unburned gas and thus helps to operate the flames under the desired lean conditions. The detailed mechanisms of the stabilization of turbulent swirl flames, however, are based on complex unsteady interactions of flow field and chemistry that are still not well enough understood today [1]. Further improvements of GT combustors with respect to emissions, fuel flexibility, and reliability therefore depend considerably on a better understanding of these interactions.
The flow field of turbulent swirl flames not only contains turbulent velocity fluctuations, but in many cases also features unsteady coherent vortex structures. The most common coherent vortex structure in swirl combustors is the so-called precessing vortex core (PVC) [2]. A literature survey provided in Appendix A shows that PVCs are frequently encountered in GT-typical swirl combustors. Consequently, not only turbulence-chemistry interaction [3], [4], [5] is important in swirl combustors, but also the interaction of coherent vortices with the flame. While flow pattern and frequency of PVCs have been frequently characterized, studies of their interaction with a flame are, however, limited to the pioneering works of Syred and coworkers [6], [7], [8] and a study of Li and Gutmark [9]. Although these works could show some effects of the PVC on mixing and flame shape, the detailed mechanisms of the interaction remain largely unclear.
Many previous works have investigated vortex-flame interaction at a rather fundamental level (see, e.g., Refs. [10], [11], [12], [13] and references therein). In these works, interaction of a single vortex or a vortex pair with a flame front is studied under well-defined conditions. The studies revealed that a vortex can interact with a flame in numerous ways. E.g., the vortex can influence the flame through transport and mixing, leading to changes of local gas composition and temperature. Furthermore, vortical structures may induce an enlarged flame surface or lead to strain-induced local extinction. The magnitude of each effect strongly depends on various parameters like, e.g., local gas composition and temperature, relative size and orientation of flame front and vortex, and time scales of chemistry and flow. Conversely, the flame affects the vortex due to the expansion induced by heat release and leads to changes of flow parameters such as density and viscosity.
In highly turbulent GT-typical swirl flames, interaction of a vortex with the flame is much more complex than under the well-defined conditions of the fundamental studies referenced above. This is mainly owing to the irregular turbulent velocity fluctuations that are superimposed on the coherent vortex flow structure. The turbulent flow can lead to effects like unsteady strain, mixing and a change of local flame surface orientation that can then compete and interact with the effects of the coherent vortex. Under such conditions, the detailed mechanisms of vortex-flame interaction are not well understood, and further detailed experimental studies are required.
The objective of this work is to elucidate the interaction of large-scale coherent vortices within turbulent swirl flames in a GT-typical combustor by means of modern laser-based methods. In particular, novel high-speed laser techniques are applied for simultaneous measurements of flow field and flame structure with high temporal and spatial resolution, which provide detailed insights into the mechanisms of the interaction. The study uses a GT model combustor, which at certain operating conditions exhibits thermoacoustic oscillations. These oscillations have been subject of numerous previous experimental studies, where several diagnostic techniques have been applied such as Raman spectroscopy, PLIF, and PIV at repetition rates between 5 Hz and 5 kHz. Recent reviews of these works are provided by Boxx et al. [14] and Steinberg et al. [15]. In another work, where the same combustor was operated at a fuel-lean condition, Stöhr et al. have investigated the dynamics of lean blowout [16].
The present study, by contrast, focuses completely on the role of coherent vortex structures. The combustor is therefore operated at conditions where no blowout and virtually no thermoacoustic oscillations occur, while coherent vortices such as a PVC are present. The operating conditions cover a range of thermal power from 10 kW to 35 kW. After a general characterization of the flames, the flow field with the coherent vortices is studied in detail using particle image velocimetry (PIV) and acoustic analysis. An important part of this study is the analysis of the flow field using proper orthogonal decomposition (POD). The usefulness of POD for the analysis of a PVC has been demonstrated in a recent publication [17]. In the second part, the interaction of vortex and flame is studied using simultaneous PIV and planar laser-induced fluorescence of OH with repetition rates up to 5 kHz.
Section snippets
Combustor
A sketch of the GT model combustor is shown in Fig. 1a. Dry air at ambient temperature enters the plenum (diameter 79 mm, height 65 mm) and then separately passes the inner (Fig. 1d) and outer (Fig. 1c) radial swirl generator. The two co-swirling flows enter the combustion chamber through a central circular nozzle (diameter 15 mm) and a surrounding annular nozzle (inner diameter 17 mm, outer diameter 25 mm contoured to an outer diameter of 40 mm). Non-swirling gaseous fuel is supplied through 72
Flame shape
Figure 2 shows averaged OH∗ chemiluminescence images of the flames with φ = 0.65 and thermal powers from 10 to 35 kW. The OH∗ signal (integrated along the line-of-sight) is used as a marker of the locations of heat release [20]. At P = 10 kW, the heat release zone lies between y = 5 mm and y = 45 mm in axial direction and within a radius of r < 35 mm from the central axis. When thermal power and thus flow velocities are increased, the upper end of the flame zone gradually moves up to y = 100 mm for P = 35 kW. For
Instantaneous flow fields
Figure 6 shows instantaneous velocity fields in the vertical section (z = 0 mm) for P = 10 kW and P = 35 kW. It can be seen that the instantaneous flow contains distinctive features that are not present in the average flow field. First, both instantaneous flow fields exhibit a zig–zag arrangement of vortices in the ISL marked by black arrows. This arrangement indicates the presence of a coherent 3D helical vortex in the ISL. Such a helical vortex is often found in swirl flames (cf. Appendix A) and is
General remarks
The aim of this section is to characterize the effects of the PVC on the flame and its role in the mechanism of flame stabilization. For all conditions, the flame is stabilized in the ISL, where also the PVC is located. This suggests that the PVC may strongly interact with the flame. In contrast, the vortex in the OSL and the exhaust tube vortex are largely separated from the flame zone, and therefore, presumably, have no direct effect on the flame. In the following, the effect of the PVC is
Summary and conclusions
The interaction of a helical precessing vortex core (PVC) with turbulent swirl flames in a gas turbine model combustor was studied experimentally. The combustor was operated with air and methane at atmospheric pressure and thermal powers from 10 to 35 kW. The flow field was characterized using particle image velocimetry (PIV), and the dominant unsteady vortex structures were determined by means of proper orthogonal decomposition. For all operating conditions, a PVC was detected in the shear
References (90)
Prog. Energy. Combust. Sci.
(2006)- et al.
Combust. Flame
(2006) - et al.
Exp. Therm. Fluid Sci.
(2010) - et al.
Proc. Combust. Inst.
(1975) - et al.
Prog. Energy Combust. Sci.
(2000) - et al.
Proc. Combust. Inst.
(2000) - et al.
Combust. Flame
(2011) - et al.
Combust. Flame
(2010) - et al.
Combust. Flame
(2010) - et al.
Proc. Combust. Inst.
(2011)
Combust. Flame
Combust. Flame
Combust. Flame
Prog. Energy Combust. Sci.
Combust. Flame
Exp. Therm. Fluid Sci.
Combust. Flame
Proc. Combust. Inst.
Proc. Combust. Inst.
Combust. Flame
Int. J. Heat Fluid Flow
Proc. Combust. Inst.
Proc. Combust. Inst.
Int. J. Heat and Fluid Flow
Combust. Flame
Combust. Flame
Proc. Combust. Inst.
Combust. Flame
Proc. Combust. Inst.
Combust. Flame
Exp. Fluids
Combust. Theor. Model.
Proc. Combust. Inst.
Combust. Sci. Technol.
Phys. Fluids
Exp. Fluids
Meas. Sci. Technol.
Annu. Rev. Fluid Mech.
J. Fluid Mech.
J. Fluid Mech.
Proc. ASME Turbo Expo
J. Fluids Eng.
J. Eng. Gas Turb. Power
J. Fluid Mech.
Cited by (250)
Tomographic particle image velocimetry measurement on three-dimensional swirling flow in dual-stage counter-rotating swirler
2024, Chinese Journal of AeronauticsFormation of a blue whirl controlled by tangential and radial airflows
2024, Combustion and FlameFlame dynamics and combustion instability induced by flow-flame interactions in a centrally-staged combustor
2023, Aerospace Science and TechnologyEffect of fuel-feeding excitation on the isothermal turbulent swirling flow structure in a lean gas turbine model combustor
2023, International Communications in Heat and Mass Transfer