Identifying noisy and quiet modes in a jet

In the current jet noise study, an empirical modal decomposition is proposed which distills the noisy and quiet modes of the flow field. In particular, the POD of flows is generalised for an optimal resolution of the far-field noise as opposed to a least-order rep-resentation of the hydrodynamic fluctuation level. This decomposition technique, which we call ‘most observable decomposition (MOD)’, is based on a linear cause-effect relationship between the hydrodynamics (cause) and the far-field acoustics (observed effect). In the current study, this relationship is identified from a linear stochastic estimation between the flow field and the far-field pressure — taking into account the propagation time of sound. We employ MOD to turbulent jet noise at Ma = 0.9, Re = 3600 using CFD/CAA data from RWTH Aachen. While more than 350 POD modes are necessary to capture only 50 % of the flow fluctuation energy, a mere 24 MOD modes resolve 90 % of the far-field acoustics. Evidently, far-field noise acts as filter which ‘sees ’ only a low-dimensional subspace of the flow and ‘ignores ’ silent subspaces which contain a large amount of fluctuation energy. The MOD methodology yields ‘least-order ’ representations of any other observable as well — assuming a linear relationship between flow and observable. I