[kharnabnew]

Hi all,
I have read Fluent documentation for aeroacoustics problems. It mentions three approaches among which the direct method is the most accurate but computationally expensive. However, there are no details on how to apply it contrary to the other methods.
Can someone explain what is the direct method and how to apply it?
Thank you in advance
NK

[LuckyTran]

The direct method is to simply run a transient simulation with small enough time-step and fine enough grid to resolve spatially and temporally the acoustic waves (pressure waves).

This method is extremely prohibitive in terms of computational cost and memory (since the acoustic waves are usually much faster than convective speed, and wavelength of acoustic waves are much smaller than hydrodynamic length scales). Numerical damping also usually kills such waves. For example it's well known that even a 2nd order backward euler scheme in time is too dissipative and fourth order scheme is necessary.

[kharnabnew]

Thank you LuckyTran,
I am interested in jets within small domains which should be ok for the direct method. How is the feedback loop accounted for in such cases? Does Fluent solve modified Navier-Stokes equations in the direct method?
Then, I guess it is a question of postprocessing. Usually, the most important results are contours of pressure waves to determine the sources, sound level through temporal profiles and characteristic frequencies. Could you, please advise on how to get these in the direct method?
Best regards
NK

[LuckyTran]

Unless I am wrong about the "direct method" otherwise:

You would just run a transient simulation and plot the instantaneous pressure. You might need to initialize it and perturb the initial pressure field to help generate the acoustic waves but that's the general idea.

[kharnabnew]

Thank you again for your answer,
I will appreciate if you could clarify your comment "You might need to initialize it and perturb the initial pressure field to help generate the acoustic waves"
Best wishes
NK

[LuckyTran]

To solve a problem you need:

the governing equations (the pde itself)
boundary conditions
initial conditions

You need the correct initial conditions for your problem to evolve in a meaningful way.

[aerosjc]

Originally Posted by LuckyTran

To solve a problem you need:

the governing equations (the pde itself)
boundary conditions
initial conditions

You need the correct initial conditions for your problem to evolve in a meaningful way.

Hi! May I ask for your help? Could you have a look at this question: Is the fluctuation pressure the same as the sound pressure? It's also using the direction method.

Thank you!

[aerosjc]

Originally Posted by kharnabnew

Thank you again for your answer,
I will appreciate if you could clarify your comment "You might need to initialize it and perturb the initial pressure field to help generate the acoustic waves"
Best wishes
NK

Hi, Have you already finished the case of simulation of acoustics by direct method May I ask for your help? Could you have a look at this question: Is the fluctuation pressure the same as the sound pressure? It's also using the direction method.

Thank you!