Time domain aeroacoustic source separation with statistical inversion

We present an application of statistical inversion for source separation of time domain microphone array signals observing an aeroacoustic source. Here, the acoustic emission of an individual vortex ring convecting past the edge a semi-infinite half-plane in an anechoic chamber is measured. Inherent to the production of individual vortex rings, a spherical shock wave is produced, polluting the acoustic measurement. The vortex/edge (V/E) interaction is observed by a circular 12-microphone array centered about the edge of the half-plane where the vortex interaction is significant. Synchronous high-speed Schlieren imaging captures the formation and convection of the individual vortex rings. Sources present in the experiment are non-stationary and impulsive, limiting the efficacy of frequency domain source separation methods, necessitating a time domain approach. A model approximating the microphone array signals is introduced, which is parameterized in terms of experimental setup and aeroacoustic source parameters. Statistical inversion of the approximating model enables the estimation of the underlying source waveform time series of the V/E interaction in addition to estimating its uncertainty. Finally, the estimated V/E source parameters are compared to prior results.