Comparison of numerical methods for noise source identification

Abstract

The identification of sound sources is difficult in the presence of many potential sources that interact. Nearfield Acoustical Holography (NAH), which refers to a process by which all aspects of the sound field in a three-dimensional volume can be reconstructed based on sound pressure measurements on a two-dimensional surface conceptually, allows the identification of complex noise sources that are otherwise difficult to characterize. Conventional NAH has been applied most commonly in simple geometries such as planar, cylindrical and spherical surfaces. Since most noise sources are not simple geometrically, it is not possible to project to the actual surface of the source in most instances, thus limiting the ability of conventional NAH to resolve the source structure accurately. Numerical acoustics procedures based on the Helmholtz integral equation have been developed that allow sound fields to be projected into arbitrarily-shaped interior or exterior domains from measured sound pressure field. Two commonly implemented procedures are the inverse Boundary Element Method (BEM) and the Helmholtz Equation Least-Squares (HELS) method. An Enhanced Surface Potential (ESP) based formulation was recently developed and promised to overcome some of the drawbacks of the BEM-based formulation. In this paper, these three methods are compared through a numerical example.