In the present study, subsonic turbulent jet noise is investigated employing reduced-order representations of the flow field and its noise source targeting 'least-order' approximations of the key processes. These representations utilize LES data for a compressible jet at Mach number 0.9 and Reynolds number 3600. The fluctuations of the velocity field and of the Lamb vector as noise source are investigated with three methods. Firstly, the streamwise development is characterised by a statistical analysis. Thus, the most active region of the flow field and the Lamb vector are observed at 11 and 8 jet diameters downstream, respectively. Secondly, an azimuthal mode decomposition is carried out. The first five azimuthal modes resolve most of the flow field and Lamb vector fluctuation. Thirdly, the dimension of the dynamics phase space is estimated by the proper orthogonal decomposition (POD). About 350 modes are necessary to resolve at least 50% of the fluctuation level of the hydrodynamics and even more modes are required for the noise source. As expected, the end of the potential core correlates with the location of a distinct peak in the noise source magnitude, thus indicating a highly active acoustical region. Intriguingly, the noise source efficiency per unit energy increases with higher azimuthal modes. The comparison of the compressible jet results with the incompressible LES at the same Reynolds number reveals a significant smaller energy concentration in the first azimuthal modes, i.e. the incompressible flow is dynamically more complex. The current results are part of an ongoing effort to predict far-field jet noise by reduced- order modelling of its hydrodynamic source.
Read full abstract