CFD eliminates unpleasant sunroof noise

Anders Tenstam, Epsilon HighTech Engineering AB, on assignment by SAAB Automobile AB, Sweden

Aeroacoustics is the field of science that deals with acoustic emissions from an aerodynamic source. Typically, it may be broadband sources from turbulent motion, or narrow-band noise generation from fluid instability acting on a solid surface. It may also cover generation mechanisms of the kind causing the sunroof buffeting noise.

The origin of buffeting noise is a shear-layer instability forming in the opening of a cavity subjected to grazing flow. In the shear-layer vortices are produced and they are convected downstream of the opening, eventually hitting the rear edge. When the vortex breaks, a pressure wave is produced which enters into the cavity. At a certain speed, the vortex shedding frequency in the shear layer will match an acoustic mode of the cavity. Often, as in many wind instruments (e.g. the flute), the resonance is in the form of a standing wave. For an automobile cavity, the resonance is in the form of a Helmholtz mode, a special case of a standing wave but with a distinctly lower frequency. This is the same sound generation mechanism as when blowing air over a bottle opening. The reason for the high amplitudes is partly the fact that the listeners (driver and passengers) are located within the resonant body itself!

Consequently, what the simulation must capture is the vortex generation and transport in the shear layer, the compressibility of the air "cushion" inside the compartment, and the subsequent resonance between the two governing time scales. Using STAR-CD, a simulation method was developed at SAAB Automobile, using well-documented benchmark geometry. Excellent agreement was found between simulations and measured results, both in estimated resonance frequency and sound pressure level. The method was applied to a real car geometry (the new SAAB 93 Sport Sedan) and, even there, an encouraging agreement was found between simulations and reality.

Simulations were performed for two different designs of sunroof opening. One where problems were known to occur in reality (artificial situation), and the other incorporating the modifications that engineers at SAAB had developed during the project phase of the new car. The important conclusion was that the difference in behavior between the two designs was also observed in the simulations. So for future projects, SAAB now has a robust methodology for using STAR-CD to predict buffeting problems before they are built into expensive prototypes.

Vortex location and periodic history downstream