Hi there,

could someone give me the keywords on checking sound generation with CFD.

If I have calculated a flow field, I guess that a periodic flow field (vortex shedding) could be a reason for sound generation. What about sound generated by turbulence ? Using a k-eps-modell does it make sense to evaluate k as a measure of turbulence intensity and so for sound generation ?

Thanks,

Victor

If you check the pressure as a function of time, this should be a good indicator of sound (noise). The amplitude of the changes of the pressure should be somehow proportional to the amplitude of the sound. Periodic changes of the pressure can even give you the frequency of the sound (i.e. you will know if it is a B flat or just a C...). You might also have different sources in the flow (just like an accord, G minor for example, is made of different notes), etc...

For sure some sound frequencies will be associated with the frequencies in the dynamic of the flow (vortices, etc...).

See for example the shedding frequencies of vortices in the wake of a circylar cylinder:

Sreenivasan, 1985, Frontiers in Fluid Mechanics (ed. Davis and Lumley), p. 41, Springer.

Patrick

Strouhal number, St = Frequency of vortices

Euler number, E = Pressure force/Inertia force

Reynolds number, Re = Inertia force/Viscous force

For simplicity, these three numbers can help us find relationship between pressure and frequency of the fluid flow.

Thanks Patrick,

I think I have understood the thing with the changing pressure due to vortex shedding. But what a about sound generation by pure turbulent motion. When I use a k-eps turbulence model I only solve for the mean quantities. Pressure fluctuations are avaraged out by definition. Can I take the level of k as an indicator for sound intensity.

Victor

I am not working with the k-e models so I have no clue what they do and don't.

Averaging in space will not affect global pressure changes but will affect local changes. And averaging in time will not give any hint of any sound of course, since we are looking for time changes in pressure.

Patrick

you can't really use you results from an ordinary cfd analysis to get detailed noise output. this is the area of comp. aero-acoustics (CAA). the technique for solving your problem are outlined in Morris et al, Journal of comp. physics vol 133 pp. 56 -74, 1997. basically in CAA the noise in a fluid flow is simulated by introducing noise sources in the domain and solving either the wave eqtn, linearized euler (most common), euler, or nonlinear disturbance equations (see the paper above) to predict the propagation of noise away from these sources. the sources may be modelled or come from the output of a RANS/LES calculation. CAA requires special spatial and temporal differencing techniques (Tam and webb, J. comp. Physics, vol 107, pp. 262-271 (1993). if you want more info you can get in contact with Dr Phillip Morris or Dr. Lyle Long at Penn State (in the aerospace dept.) and they can help you (if they feel like, but they're nice guys).