Low-frequency noise source localization and identification method based on spherical arrays

Abstract

In view of the low spatial resolution of small-scale open spherical arrays in the low-frequency range and the instability of the frequency location corresponding to the “nodal” position, this paper proposes a noise source localization and identification method based on a virtual spherical array. Based on the inverse problem of sound field reconstruction theory, the expansion method of the virtual spherical array is examined, and a virtual spherical array with a larger aperture than the original spherical array is obtained. The original array and virtual array are combined to form a double-layer spherical array, and the two joint processing methods of “step selection” and “sound pressure-vibration velocity” are used in the spherical harmonic domain. Finally, the spherical harmonic function expansion beamforming algorithm is used to verify the noise source localization and identification effect of the above virtual spherical array and its joint processing with an actual spherical array. The simulation and experimental results show that the virtual large-aperture open spherical array can be effectively obtained by the proposed method and that the virtual double-layer spherical array obtained by the virtual array and its joint processing with the actual spherical array can effectively improve the low-frequency spatial resolution of the spherical array. Moreover, it solves the “nodal” problem of the original actual open spherical array and improves the accuracy of noise source localization and identification.