Fast prediction of broadband scattering from an underwater target using non-uniform sampling method based mode decomposition

Peizhen Zhang,Dongyang Fu,Peizhen Zhang,Xiukun Li,Jun Fan
DOI: https://doi.org/10.1109/coa.2016.7535635
2016-01-01
Abstract:Broadband signals are widely used in active sonar and synthetic aperture sonar imaging, correspondingly, which require calculating broadband echo in frequency domain. Finite element (FE) is an effective method to calculate the echo characteristics of an underwater target. However, the computing time is tremendous due to the large number of iterative computations on a 2D or 3D grid. Especially when the object is large, it is difficult or even impossible to use a scattering calculation for FEM in high frequency. A 2D-FE numerical model is built to calculate scattering acoustic field for the object immersed in water. Far field of scattering acoustic is decomposed into several modes. Numerical calculation shows that the target is divided into the unit grid comparable to the wavelength of incident wave at low frequency, and scattering intensity of each mode fluctuates relatively little. If the waves are at high frequencies, the scale of the target is much larger than the wavelength, and scattering intensity or characteristics of the target fluctuates rapidly with increasing frequency. A nonuniform sampling method based on mode decomposition is presented: that is, in wavenumber domain, interval of calculation frequencies of each mode decreases from low to high frequency. Conventional interpolation algorithm is available to obtain scattering intensity of frequency point by point. After interpolation, all results of different modes are summed to get broadband echo characteristics of the underwater object. Several typical targets, such as an elastic sphere and a shell-like elastic cylinder are used to verify the efficiency of this methodology. A substantial reduction in computation burden is achieved by avoiding point-by-point computation in wavenumber domain. This represents an advantage in the computational efficiency of the Finite Element method.
What problem does this paper attempt to address?