Velocity field response of a forced swirl stabilized premixed flame

Abstract

This study is motivated by the necessity to develop a low order prediction approach for unsteady heat release response characteristics in lean premixed gas turbine combustors. This in turn requires an accurate description of the coherent hydrodynamic oscillations induced in the combustor flow by acoustic forcing. Time resolved velocity and flame position fields are obtained using sPIV and OH-PLIF measurements on a single nozzle, swirl-stabilized, premixed, methane-air flame in a model "unwrapped" annular combustor rig. A natural acoustic oscillation in the rig at 115 Hz results in a coherent flow oscillation that is concentrated primarily within the shear layer between the annular jet flow and the central recirculation zone. A linear stability analysis performed about time averaged base flow fields shows that the flow does not have any self-excited hydrodynamic modes. We then compare predictions from a forced response analysis at a forcing frequency of 115 Hz, based on the linearized Navier-Stokes equations for this coherent response. Good qualitative agreement between linear forced response analysis predictions and experimental response results, is seen for the spatial variation of velocity oscillation amplitude fields, away from the burner centerline. Further, good quantitative agreement between predictions and the experimental response is seen for the phase speed of velocity oscillations along the shear layer between the annular jet and the central recirculation zone. This phase velocity is an important flow field characteristic that has a significant impact on the heat release response that results from these coherent velocity oscillations. Present methods for forced response analysis assume uniform forcing amplitude along the radial direction at the forcing location, as well as, open flows along the streamwise direction. Both these assumptions are not strictly true for the present burner which has a center body on its axis. This maybe the reason for somewhat poor qualitative and quantitative agreement between experiments and predictions at the centerline.