Atmospheric low swirl burner flow characterization with stereo PIV

The lean premixed prevaporized (LPP) burner concept is now used in most of the new generation gas turbines to reduce flame temperature and pollutants by operating near the lean blow-off limit. The common strategy to assure stable combustion is to resort to swirl stabilized flames in the burner. Nevertheless, the vortex breakdown phenomenon in reactive swirling flows is a very complex 3D mechanism, and its dynamics are not yet completely understood. Among the available measurement techniques to analyze such flows, stereo PIV (S-PIV) is now a reliable tool to quantify the instantaneous three velocity components in a plane (2D–3C). It is used in this paper to explore the reactive flow of a small scale, open to atmosphere, LPP burner (50 kW). The burner is designed to produce two distinct topologies (1) that of a conventional high swirl burner and (2) that of a low swirl burner. In addition, the burner produces a lifted flame that allows a good optical access to the whole recirculation zone in both topologies. The flow is studied over a wide range of swirl and Reynolds numbers at different equivalence ratios. Flow statistics are presented for 1,000 S-PIV snapshots at each configuration. In both reactive and cold nonreactive flow, stability diagrams define the domains of the low and high swirl topologies. Due to the relatively simple conception of the physical burner, this information can be easily used for the validation of CFD computations of the burner flow global structure. Near field pressure measurements reveal the presence of peaks in the power spectra, which suggests the presence of periodical coherent features for almost all configurations. Algorithms have been developed to identify and track large periodic traveling coherent structures from the statistically independent S-PIV realizations. Flow temporal evolution is reconstructed with a POD-based method, providing an additional tool for the understanding of flow topologies and numerical codes validation.