A Hybrid Method for Full Spectrum Noise and Vibration Prediction

Abstract

Predicting the response of a complex structural-acoustic system across a broad frequency range presents a number of challenges to an analyst. It is quite common to find that the uncertainty associated with the local dynamic properties of various subsystems of a system can vary greatly across the system. It is also common to find that the modal density and wavenumber content of the various subsystems can vary greatly across the system. Typically, this results in a mixture of strongly phase correlated (long wavelength) motion which spans many subsystems, superimposed with weakly phase correlated local motion that is confined to individual subsystems. This mismatch in the local statistical and dynamic properties of a system is often referred to as the mid-frequency problem. This paper provides a qualitative definition of the mid-frequency problem and suggests that a statistical description of the local dynamic properties of a system is an essential element of any mid-frequency prediction method. A hybrid approach to the mid-frequency problem is then described which employs a statistical description of the local modal properties of various subsystems in a system. The spatial statistics of the local modes are of particular interest and the way in which these statistics are encompassed in the hybrid analysis is discussed. Experimental investigations of the spatial statistics of a frame-panel structure are then presented and measurements of the acoustic power radiated by the structure are compared with numerical predictions.