Identification of Noise Sources in an Aircraft Fuselage Using an Inverse Method Based on a Finite Element Model

Because of standing waves, noise source identification in enclosed sound fields is a nontrivial problem, and direct localization of sources is nearly impossible, if the damping ratio is low. In such a case it can be advantageous to apply an inverse method. One novel approach is presented in this paper. It is based on a combined measurement and calculation technique and can be applied to reconstruct the boundary values of both sound pressure and particle velocity. The experimental part is limited to sound pressure measurements in a subspace of the investigated interior. These data are associated to the nodal degrees of freedom of an acoustic finite element model. The latter is based on a frequency-domain formulation. If all sources are located on the boundary, the resulting set of algebraic equations can be split into two subsets. The first subset can be used to determine all unknown sound pressure data. The second one allows for calculating the particle velocity. In contrast to previous publications, the solution of the inverse problem is found by minimizing a cost function that corresponds to the acoustic energy of the enclosure. In addition the method of Lagrange multipliers has been used to consider constraints defined by subset one. This inverse approach has been applied to sound source localization in a fully equipped long-range aircraft fuselage using internal as well as external sources to excite cabin noise. Internal sources have been localized with success. Furthermore, it has been possible to compute reliable data, if the cabin noise has been excited by external sources.