Comparison of inverse techniques for reproducing an extended, partially coherent sound field above a reflecting plane

The relative abilities of two advanced array processing techniques to reconstruct a partially correlated noise source similar to a high-speed jet are compared. M-SONAH is a modification of statistically optimized near-field acoustical holography that includes multiple types of wavefunctions [A. T. Wall et al., J. Acoust. Soc. Am., 137, 963–975 (2015)]. The hybrid method may be described as an advanced cross-beamforming algorithm with regularization [T. Padois et al., AIAA Paper 2013-2212]. A numerical, multiple wavepacket-based source distribution provides a reasonable model of the frequency-dependent directivity and coherence properties of a jet noise sound. An image source was included to account for a reflecting ground plane and the numerically modeled sound field was used an input to both the M-SONAH and hybrid methods. For M-SONAH, the image source was handled using an additional set of cylindrical wavefunctions and for the hybrid method, sources were assumed to be located both above and below the reflecting plane. While both methods provide accurate reconstructions, for this test case, the hybrid method performs better at higher frequencies using a realistic measurement aperture and density. [Work sponsored by the USAFRL SBIR program.]