We present diagnostic experiments and reduced-order models aimed at understanding and mitigating supersonic jet noise from coherent wavepackets in the turbulent shear layer, generally accepted to be the source of peak aft-angle mixing noise. The work builds on a successful Caltech/UTRC modeling approach that predicts the evolution of the wavepackets as instability waves of the turbulent mean flow, as well as the noise radiated from their near field. The models are experimentally assessed for unforced and forced supersonic isothermal and heated Mach 1.5 jets from ideal expansion nozzles. A spinning valve device is used to inject air near the nozzle lip at frequencies up to a Strouhal number of about 0.25. Results indicate a 2-3 dB benefit near peak frequencies of the spectrum and a 2 dB OASPL reduction at a mass flow percentage of 3.2. For the same injection plenum pressure, steady blowing yields more noise benefit than the unsteady actuation schemes explored until now. However, this may be explained by the slight decrease in injection velocity incurred in going from steady to unsteady operation. The reduced-order models, based on parabolized stability equations (PSE), are found to be in good agreement, in terms of envelope and phase, with those educed from the experimental data of the unforced jet. For the actuation schemes we have considered to date, the model and experimental data support a tentative explanation for the observed noise reduction in terms of attenuation of the wavepacket amplitudes by the thickened shear layer. Wavepackets induced by the harmonic component of the actuation are linearly superposed on those produced by broadband turbulence, without significant interaction, such that they lead to the addition of tones to the far-field noise that are counterproductive as far as noise reduction is concerned.
展开▼