Studies on Flow Dynamics and Spray Swirl Interactions in Gas Turbine Combustor

Abstract

Coupling of spray with the coherent structures of a highly turbulent flow has been a long-standing problem especially in the context of liquid fuel delivery systems in gas turbine combustors. The atomizer in a gas turbine combustor usually has one or more (radial/axial entry) air swirlers with a fuel nozzle being mounted centrally along the longitudinal axis of swirler. It is well known that swirling flows are highly three dimensional in nature and often induce multiple aerodynamically unstable modes whose frequencies are several orders of magnitude. The basic understanding of flow dynamics in gas turbine swirl cup is critical to achieving clean and efficient combustion in modern-day gas turbine combustors. In this work, we analyze the evolution of the hydrodynamic topology and consequent spray-flow interactions in a coaxial swirl injector assembly. The key results of the present work are discussed in four parts. In the first part, the global evolution and temporal dynamics of various vortex breakdown modes are discussed. Experiments are carried out for three sets of co annular flow Reynolds number 𝑅𝑒𝑎=4896,10545,17546. Furthermore, for each 𝑅𝑒𝑎 condition, swirl number ‘𝑆𝐺’ is varied independently from 0≤𝑆𝐺≤3. Three distinct forms of vortex breakdown namely, pre-vortex breakdown (PVB), central toroidal recirculation zone (CTRZ; axisymmetric toroidal bubble type breakdown) and sudden conical breakdown are explored in greater details. Energy ranked, and frequency resolved / ranked robust structure identification methods – POD, DMD respectively is implemented over instantaneous time resolved PIV data sets to extract the dynamics of coherent structures associated with each vortex breakdown modes. The dominant structures obtained from POD analysis suggest the dominance of KH instability (axial + azimuthal; accounts ~ 80 % of total TKE) for both PVB and CTRZ while the remaining energy is contributed by shedding modes. On the other hand, shedding modes contribute to the majority of the TKE in conical breakdown. The frequency signatures quantified from POD temporal modes and DMD analysis reveals the occurrence of multiple dominant frequencies in the range of ~ 10 – 400 Hz with conical breakdown. This phenomenon may be a manifestation of high energy contribution by shedding eddies in the shear layer. Contrarily, with PVB and CTRZ, the dominant frequencies are observed in the range of ~ 20 – 40 Hz only. In addition, the current work explores the hysteresis (path dependence) phenomena of conical breakdown as functions of Reynolds and Rossby numbers. It has been observed that the conical mode is not reversible and highly dependent on the initial conditions. In the second part, we have reported how the liquid sheet behaves in such swirling flows. The air flow rate across the swirler is progressively varied to probe the two-phase flow interaction dynamics across weak, transition and strong momentum coupling regimes. The liquid sheet breakup and gas – liquid phase interaction dynamics suggests strong one way coupling at higher MR values. The POD analysis implemented over the shadow images clearly delineates the superimposing of gas phase instabilities with liquid sheet. The breakup length scale and liquid sheet oscillations are meticulously analyzed in time domain to reveal the breakup dynamics of liquid sheet. Furthermore, the large-scale coherent structures of swirl flow exhibit different sheet breakup phenomena in spatial domain. For instance, flapping breakup is induced by counter rotating vortices in the flow field induced by vortex breakdown phenomenon. The breakup regime map is also constructed based to illustrate the various forms of breakup mechanism as a function of MR values. Finally, the ligament formation mechanism and its diameter, size of first-generation droplets are measured with phase Doppler interferometry (PDI). The measured sizes scale reasonably with KH waves. In the third part, the fundamental mechanisms of vortex-droplet interactions leading to flow distortion, droplet dispersion and breakup in a complex swirling gas flow field are discussed. In particular, how the location of droplet injection determines the degree of inhomogeneous dispersion and breakup modes have been elucidated in detail. The droplets are injected as monodispersed streams at various spatial locations like the vortex breakdown bubble and shear layers (inner and outer) exhibited by the swirling flow. Time-resolved particle image velocimetry (3500 frames/s) and high-speed shadowgraphy measurements are employed to delineate the two-phase interaction dynamics. These measurements have been used to evaluate the fluctuations in instantaneous circulation strength 𝛤′caused by the flow field eddies and resultant angular dispersion in the droplet trajectories 𝜃′. The droplet-flow interactions show two-way coupling at low momentum ratios (MR) and strong one way coupling at high momentum ratios. The gas phase flow field is globally altered at low air flow rates (low MR) due to the impact of droplets with the vortex core. The flow perturbation is found to be minimal and mainly local at high air flow rates (high MR). Spectral coherence analysis is carried out to understand the correlation between eddy circulation strength 𝛤′and droplet dispersion 𝜃′. Droplet dispersion shows strong coherence with the flow at certain frequency bands. Subsequently, proper orthogonal decomposition (POD) is implemented to elucidate the governing instability mechanism and frequency signatures associated with turbulent coherent structures. POD results suggest the dominance of KH instability mode (axial and azimuthal shear). The frequency range pertaining to high coherence between dispersion and circulation shows good agreement with KH instability quantified from POD analysis. The droplets injected at inner (ISL) and outer shear layer (OSL) show different interaction dynamics. For instance, droplet dispersion at OSL exhibits secondary frequency (shedding mode) coupling in addition to KH mode, whereas, ISL injection couples only at a single narrow frequency band (i.e. KH mode). Finally, we have analyzed the spray- flow field dynamics in the realistic injector configuration (i.e. high shear injector). High shear injector usually consists of a series of air swirlers (primary and secondary) with diverging flare at the exit and centrally mounted fuel nozzle. It should be noted that to precisely probe the characteristic features, experiments have been also conducted with independent primary and secondary swrilers. A parameter named dynamic pressure ratio (𝜉) is used to quantify the monemtum transfer pathways between primary and secondary swirler flow field across various test cases. The test cases which exhbit 𝜉<1 are identified as primary swirler dominant flow and for 𝜉≥1 are delineated as secondary swirler dominant flow. In other words, for 𝜉<1 momentum exchange will be take palce from primary to secondary swirler and vice versa for 𝜉≥1 condtion. . The results revealed that flow pertaining to the secondary swirler exhibits sharp narrowband frequency in the range of 0 – 60 Hz, whereas, the primary swirler flow exhibits wideband frequencies with distinct peaks at 200, 800 Hz. The POD analysis extended over combined primary and secondary swirler flows shows the persistence of wide band oscillations for the test cases pertaining to 𝜉<1 (i.e. Regime I). This is due to the dominance of primary swirler flow in the Regime I. On the contrary, the frequenncy signatures shift to sharp narrow band (0- 70 Hz) for the secondary swirler dominant cases (i.e. 𝜉≥1; Regime II). In addition, we have also reported the sensitivity of the high shear swirl cup with respect to the geometrical parameters. The geometric parameters like flow split ratio (γ) between primary and secondary swirler, geometric swirl number, area ratio (Δ), flow orientation (i.e. co and counter rotation), exit flare angle (ϴ) etc are considered. The length scale (𝑊𝐷𝑓⁄), which embodies the radial extend of the recirculation zone is used as criteria to distinguish the various test cases. It is found that, the magnitude of (𝑊𝐷𝑓⁄) is governed by near field swirl number (SN10) and Reynolds number for the cases where SNgeo, γ, Δ have been varied. Here, SN10 represents the experimentally measured swirl number at ~ 10mm from the exit of swirl cup. On the other hand, for the cases with variations being θ and flow orientation, (𝑊𝐷𝑓⁄) founds to be only a function of near field swirl number f (SN10). Next, the spatial distribution of the spray perceived from patternation studies shows a linear relationship with the magnitude of 𝑊𝐷𝑓⁄. It is interesting to note that, though the spatial spread of the spray scales with 𝑊𝐷𝑓⁄, however, the spatial uniformity and measured droplet remains insensitive to the test variables.