Search Results

Automakers have reported that passenger perception of vehicle interior wind noise is strongly correlated to the non-Gaussian and non-stationary character of the exterior aero-acoustic wind loading. Researchers in other domains have shown that leptokurtic non-Gaussian loading (Kurtosis κ>3) can be synthesized by non-stationary modulation of otherwise Gaussian random loading. This paper introduces a transient statistical energy analysis (SEA) model for the aero-vibro acoustic transmission of non-stationary wind noise which uses the same approach - a modulation of otherwise Gaussian random fluctuating pressure loading, in each one third octave band. The authors have previously shown that the non-stationary character of random wind loading can be measured in a wind tunnel or on the road with a suitable surface pressure microphone array.

It has been shown that internal transmission of wind noise is dependent on the external aerodynamic and acoustic excitation around the automobile. Flow over the A-pillar and side-view mirror induces strongly convecting turbulence and associated acoustics which excite the side-glass. A useful tool to understand and quantify these physics is to perform temporal Fourier analysis (auto-spectra) and spatial Fourier analysis (cross-spectra and wave-number decomposition). This study demonstrates the uses of wave-number decomposition to quantify the mechanisms associated with turbulent convection and acoustical propagation. A CFD computation using the commercial codes STAR-CCM+ is performed for the flow over a generalized side-view mirror in a freestream of 38m/s. LES-enabled turbulence is solved in a fully compressible framework so as to capture all the local acoustical propagation well beyond 3kHz.