Search Results

Surface pressure measurements using microphone arrays are still challenging, especially in an automotive context with cruising speeds around Mach 0.1. The separated turbulent boundary layer excitation and the side mirror wake flow generate both acoustic and aerodynamic components, which have wavenumbers that differ by a factor of approximately 10. This calls for high spatial resolution measurements to fully resolve the wavenumber-frequency spectrum. In a previous publication [1], the authors reported a micro-electro-mechanical (MEMS) surface microphone array that successfully used wavenumber analysis to quantify acoustic versus turbulence loading. It was shown that the measured surface pressure at each microphone could be strongly influenced by self-noise induced by the microphone “packaging”, which can be attenuated with a suitable windscreen.

Low frequency interior wind noise is typically dominated by underfloor flow noise. The source mechanisms are fluctuating surface pressure loading from both flow turbulence and acoustic field levels developed in the semi-reverberant cavity between floor and road. Previous studies have used computation fluid dynamics (CFD) to estimate the aero-acoustic loading applied to a vibro-acoustic model, which is then used to predict the transmitted interior wind noise. This paper reports a new perspective in two respects. First it uses novel surface pressure microphone arrays to directly measure the underfloor aero-acoustic loading in the wind tunnel. Second, it considers two different underfloor aerodynamic configurations - with and without lightweight aero cover panels, which are installed primarily to reduce aerodynamic drag.

A number of researchers [3],[4] have shown that Statistical Energy Analysis[1] provides a useful framework for test-based noise path analysis - as well purely analytical models. This paper describes the results of a comparative study of two such models - an analytic SEA model and a test-based SEA parameter model - for predicting the Noise Reduction performance of a light truck body. First, the different power balance matrix solutions and SEA subsystem definitions are reviewed. Second the results of the two independent models are compared. For some specific cases of structure-borne noise, the test-based method is shown to be more accurate than the analytic model. Third, means by which individual measured SEA parameters can be integrated into an analytic SEA model are proposed and evaluated. Results indicate that the scope for analytical model improvement, but also highlight the need for more rigorous “consistency” in the way test-based SEA parameters are measured.