Book Description

Turbulent jet mixing noise is an important component of the overall noise generated by modern aircraft. The mixing noise is dominated by the noise from large scale coherent structures and is observed at shallow exit angles (30[tilde]50deg from axis) with narrow banded spectrum which peaks at St [identical to] fD/U = 0.2 [tilde] 0.3. In this work, we apply a range of theoretical methods to investigate the low frequency noise generation using jet flow data acquired from Large Eddy Simulation database (Bodony & Lele (2005)). Bi-orthogonal relations between regular and adjoint linear stability waves are formulated and applied to detect the amplitude coefficients of instability waves in nonlinear flows. We also apply solutions from Parabolized Stability Equations to the beamforming technique to determine the amplitude coefficients. To separate supersonic and subsonic components of flow variables from LES, we use wavenumber- frequency domain filter then we apply Proper Orthogonal Decomposition method to examine the overall pattern of filtered field. Four jet operating conditions are considered, cold Mj=0.5, 0.9, and 1.95 jets, and a heated Mj =0.97 jet at three frequencies St = 0.1, 0.3 and 0.5 and azimuthal mode numbers n = 0, 1 and 2. Adjoint based amplitude coefficient shows quite promising results for unsteady laminar flow computed using Direct Numerical Simulation. For turbulent flow, this method has limited success, whereas PSE with beamforming technique shows a close resemblance to the turbulent flow data from LES. The near-field norms of the supersonic component of pressure for different jets show a close correlation with the far-field noise intensity and are found to scale well with Lighthill's power law. For heated transonic and cold supersonic jets, the POD projection of the supersonic component resembles the projection of full flow field. However, for subsonic jets, POD projection of supersonic component is drastically different from the POD modes of the full flow field.