Farfield filtering and source imaging of subsonic jet noise

摘要

Jet noise is analysed using data-processing tools adapted to two particular structural traits of the far field: the strong polar dependence and the temporal intermittency. Proper Orthogonal Decomposition is used to probe the polar structure of the sound field, wavelet transform being used to interrogate the temporal signature. The far field is decomposed, using each of these approaches independently, into a component attributed to 'coherent structures', denoted CS, and a residuum, R. The criteria for the decomposition being different, spatial on one hand and temporal on the other, comparison of the resulting CS components is of considerable interest; both decompositions lead, for instance, to CS components that compare favourably with a wavepacket source Ansatz. Using the two techniques, an analysis methodology is established and applied to data from a Mach 0.9, isothermal jet; a series of metrics are thereby proposed by which to evaluate the data. The methodology and associated metrics are then used to explore the effect of varying Mach number on isothermal and heated jets. The following main results are obtained. Both the unfiltered low-angle sound spectrum and that of the CS component of the isothermal jets are found to scale best with Helmholtz number, indicating that the associated sound source is noncompact. In the heated jet, on the other hand, a Strouhal number scaling is observed, again for both the unfiltered low-angle spectrum and the CS spectrum, suggesting that the associated sources are in this case more compact. Where the intermittency of the farfield signature is concerned it is found that increasing the Mach number of isothermal jets has no discernible impact, whereas in the case of the heated jet this increase is accompanied by a decrease in the intermittency, indicating some kind of associated stabilisation of wavepacket source dynamics. Finally, the unfiltered data is used to perform source imaging, using a wavepacket Ansatz. This allows a more comprehensive eduction of the wavepacket parameters. The trends observed are consistent with known changes in the mean field and with linear stability theory. Finally, the directivity of the wavepackets obtained using the source imaging is compared with those educed from the data using the POD and wavelet filters. Good agreement between all three constitutes a strong evidence supporting the contention that such wavepackets underpin the said, polar and temporal, features of the farfield.