A technique to determine the number of incoherent sources contributing to the response of a system

Abstract A technique, referred to here as Number of Incoherent Sources (NIS) analysis, is developed for determining the number of incoherent noise generating processes in a structural-acoustic system. The technique is based on finding the rank of matrices consisting of the auto- and cross-spectra of the signals measured by multiple source transducers located within the system. Ideally, the rank of the matrices is the same as the number of incoherent processes sensed by the transducers, and is computed by using a singular value decomposition. However, in addition to the physical sources that are of interest, the measurement system, data acquisition and signal processing also introduce additional artificial incoherent noise sources. Part of the paper is devoted to an examination of how these additional noise sources arise and how they may affect the identification of the number of incoherent sources in the physical system, together with recommendations on how to minimise their contributions. NIS analysis is also applied to a multiple-input/multiple-output system for determining the number of actual incoherent noise generating processes and the number of measurement noise processes associated with the inputs and the outputs which are incoherent with the inputs and the outputs. Verification studies are conducted which show good agreement with the theoretical development.