Fast Fourier-based deconvolution for three-dimensional acoustic source identification with solid spherical arrays. (July 2018)

• Record Type: Journal Article
• Authors: Yang, Yang; Chu, Zhigang; Shen, Linbang; Ping, Guoli; Xu, Zhongming
• Is Part Of: Mechanical systems and signal processing. Volume 107(2018)
• Page Start: 183
• Page End: 201
• Publication Date: 2018-07
• Subjects: Fourier-based deconvolution -- Spherical harmonics beamforming -- Three-dimensional acoustic source identification -- Solid spherical arrays; Structural dynamics -- Periodicals; Vibration -- Periodicals; Constructions -- Dynamique -- Périodiques; Vibration -- Périodiques; Structural dynamics; Vibration; Periodicals; 621
• DOI: 10.1016/j.ymssp.2018.01.028
• Languages: English
• ISSNs: 0888-3270

Abstract

Highlights: Fourier-based deconvolution is realized for solid spherical array's 3D source identification. A novel deconvolution theory is built using the periodic boundary condition. Four deconvolution methods are formed to enhance resolution and remove sidelobe. Recovered source strength is proportional to the square of the focus distance. Recovered pressure contribution is scarcely affected by the focus distance. Abstract: Being capable of demystifying the acoustic source identification result fast, Fourier-based deconvolution has been studied and applied widely for the delay and sum (DAS) beamforming with two-dimensional (2D) planar arrays. It is, however so far, still blank in the context of spherical harmonics beamforming (SHB) with three-dimensional (3D) solid spherical arrays. This paper is motivated to settle this problem. Firstly, for the purpose of determining the effective identification region, the premise of deconvolution, a shift-invariant point spread function (PSF), is analyzed with simulations. To make the premise be satisfied approximately, the opening angle in elevation dimension of the surface of interest should be small, while no restriction is imposed to the azimuth dimension. Then, two kinds of deconvolution theories are built for SHB using the zero and the periodic boundary conditions respectively. Both simulations and experiments demonstrate that the periodic boundary condition is superior to the zero one, and fits the 3D acoustic source identification with solid spherical arrays better. Finally, four periodic boundary condition based deconvolution methods are formulated, and their performance is disclosed both with simulations and experimentally. All the four methods offer enhanced spatial resolution and reduced sidelobe contaminations over SHB. The recovered source strength approximates to the exact one multiplied with a coefficient that is the square of the focus distance divided by the distance from the source to the array center, while the recovered pressure contribution is scarcely affected by the focus distance, always approximating to the exact one.