CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM Running, Solving & CFD (http://www.cfd-online.com/Forums/openfoam-solving/)
-   -   How to solve acoustics ? (http://www.cfd-online.com/Forums/openfoam-solving/63418-how-solve-acoustics.html)

wolle1982 April 8, 2009 03:12

How to solve acoustics ?
 
Hi everybody,

I want to simulate a case and get the acoustic emission of it. Since I'm not familiar with it, can someone give me a little crash-course?

- what solver
- what commands for acoustics
- any added functions before calculating


I've not tried yet, maybe the acoustics module is already implemented, so I only have to run a compressible calculation....??

thanks for help.

Wolfgang

ivan_cozza April 8, 2009 06:33

Hi Wolfgamg,
you're asking something like a bit of moon ;)!
At the moment there isn't neither any acoustic-specific solver in OpenFOAM, nor high order schemes in order to build up your own solver.
I tried some month ago to work on a Linearized Euler solver in OF, but the results where very bad, as the existing schemes are not suited for acoustics.
So, there are two ways:
1) "Easy" way: you can generate acoustic sources with a turbulent flow model (LES, DES, URANS, RANS + SNGR). Then you have to make them propagate with an external ad hoc code.
2) "Hard" way: start coding high order schemes for OF, then build up an acoustic propagation code (if you success, let me know!!!!)

Have a nice day,
Ivan

wolle1982 April 9, 2009 05:49

hi ivan,

thanks for the reply.

could you suggest a "external ad hoc code" or even send it to me? I want to get the acoustics of a 3D Wing tip in order to compare with other wing tips.

how exactly would your suggestion no. 1) work? would i have to recall this code for every timestep? does the acoustics-calculation affect the numerical solution? So, do I have to give the acoustic result back to the OF calculation?

In my case, I only want to compare acoustic results. So i don't need the absolute value of the emmited sound (dB). can I somehow interpret this out of the vorticity results??

any suggests would help.

greets,
wolfgang

ivan_cozza April 11, 2009 04:24

Quote:

Originally Posted by wolle1982 (Post 212442)
could you suggest a "external ad hoc code" or even send it to me? I want to get the acoustics of a 3D Wing tip in order to compare with other wing tips.

Wolfgang, actually with ad hoc code I mean home written one. The only acoustics commercial code I can remember is Actran, but we work with home made Fortran codes...

Quote:

Originally Posted by wolle1982 (Post 212442)
how exactly would your suggestion no. 1) work? would i have to recall this code for every timestep? does the acoustics-calculation affect the numerical solution? So, do I have to give the acoustic result back to the OF calculation?

My nr 1) suggestion imply an acoustic analogy approach, so previously you calculate the sound sources with a time dependent CFD model, later you use the whole time history as input for an acoustic analogy model (there are tons of this model,eg linearized euler eq., LPCE, integral methods, etc. etc...). After this secon run, you have your acoustic results!

Quote:

Originally Posted by wolle1982 (Post 212442)
In my case, I only want to compare acoustic results. So i don't need the absolute value of the emmited sound (dB). can I somehow interpret this out of the vorticity results??

You mean that you're only interested on the directivity of sound? Or that you just need to know if a configuration is noiser or quiter than another?
I the first case, you need the whole procedure, in the second maybe you can just look at an acoustic source field, and you can skip the acoustic propagation step...

Hope this could help you,
Ivan

wolle1982 April 14, 2009 05:54

hi ivan,

thx a lot for the moment. yes, i only need to know which configuration is noiser or quieter.
could you please describe a little more detailed how to add an "acoustic source field". or is the source for those acoustic programms just the vorticity or pressure values. I have not worked with fields yet.

greets

7islands April 14, 2009 09:31

Hi Wolfgang,
Perhaps you would anyhow want to start with reading up classic literature about acoustic analogy by Lighthill, Curle and Howe in order to get an idea about how acoustic sources should be calculated. I guess it may be too difficult even for Ivan to present you a definitive guidance because the detailed methodology quite depends on the target Mach numbers and other characteristics of your problem anyway.

However as you write, if you are only interested in acoustic sources but don't have to care that much about the propagation, it won't be so difficult to add those acoustic calculations into one of unsteady CFD solvers.

Takuya

wolle1982 May 15, 2009 12:48

as a first approximasation for acoustics, i would like to read out the k-values and the epsilon-values and calculate them for getting the acoustic power in each cell. after calculation it should be written in a seperate file, so i can visualize in paraFoam.

has some of you c++ guys any idea of how to do this as a c++-newbie like i am.???!

7islands May 15, 2009 23:53

Hi Wofgang,
Then how about the Calculator filter in ParaView?

Takuya

wolle1982 May 18, 2009 04:20

Hi 7islands,

thanks a lot, didn't know that function. Is exactly what I needed.

Greets


All times are GMT -4. The time now is 05:01.