[oveis87]

Dear CFD users,

I am going to simulate a sound wave at the inlet of a muffler-shaped geometry.
The inlet of my model is pressure inle for a pulsating airflow.
The pulsation is as a white noise mimicking a speaker. The frequency range is 10 to 800Hz.

The flow is laminar.
My problem is that, when I extract the pressure vs time for a specified number of time step, the results of signals for transfer function estimate is very far from the experiment.

Does anybody know how i can simulate such a big range of frequency at the inlet of my model in fluent?

PS. I am doing this to find the resonant frequency of my model.

Thank you very much.

[LuckyTran]

What about your outlet? Is there any influence there? Are you using a non-reflecting outlet or are you using a reflecting outlet?

[oveis87]

Originally Posted by LuckyTran

What about your outlet? Is there any influence there? Are you using a non-reflecting outlet or are you using a reflecting outlet?

Dear LuckyTran,
Thank you for your response. There is nothing to influence the outlet. the outlet is zero gauge pressure as depicted in the attached picture.

The outlet should not be reflective.

Thanks,
Oveis

[LuckyTran]

My bad, I forgot the non-reflecting outlet is not yet implemented in Fluent.

With a constant pressure outlet, it is reflective! It is reflective because you fix the pressure to be constant 0 at the outlet, so any acoustic perturbations that arrive there must be reflected backwards to maintain this condition. All the sound you throw at it will be reflected backwards unlike in a real physical situation where the sound would radiate outwards. Does this affect your analysis of the transfer function?

[oveis87]

Originally Posted by LuckyTran

My bad, I forgot the non-reflecting outlet is not yet implemented in Fluent.

With a constant pressure outlet, it is reflective! It is reflective because you fix the pressure to be constant 0 at the outlet, so any acoustic perturbations that arrive there must be reflected backwards to maintain this condition. All the sound you throw at it will be reflected backwards unlike in a real physical situation where the sound would radiate outwards. Does this affect your analysis of the transfer function?

Yes, since it affects the static pressure at different points, this results in an inaccurate transfer function. Do you think that I have to write a UDF code for the outlet?

[oveis87]

Originally Posted by LuckyTran

My bad, I forgot the non-reflecting outlet is not yet implemented in Fluent.

With a constant pressure outlet, it is reflective! It is reflective because you fix the pressure to be constant 0 at the outlet, so any acoustic perturbations that arrive there must be reflected backwards to maintain this condition. All the sound you throw at it will be reflected backwards unlike in a real physical situation where the sound would radiate outwards. Does this affect your analysis of the transfer function?

I selected the non-reflecting boundary conditions for both inlet and outlet. However, since the solution is chosen for incompressible flow (ideal gas), I get divergence in continuity and energy equations!

What is your idea about this divergence?

[LuckyTran]

I am confused as to whether or not you are using a non-reflective outlet. You showed convincingly that you do not have this option in your screenshot.


I also don't get why you would try to study acoustic perturbations in incompressible flow.


You seem to be in over your head on this problem. Maybe you should practice on a simpler case like a simple duct.

[oveis87]

Originally Posted by LuckyTran

I am confused as to whether or not you are using a non-reflective outlet. You showed convincingly that you do not have this option in your screenshot.


I also don't get why you would try to study acoustic perturbations in incompressible flow.


You seem to be in over your head on this problem. Maybe you should practice on a simpler case like a simple duct.

When you recommended to use non-reflecting boundary, then I searched and found that I did not turn the compressible flow option on. Then I turned it on and found non-reflecting option at the boundary. But when used it, the solution was diverged!!

What do you think about this?

[LuckyTran]

Given that you were trying to do incompressible acoustics, I think you should just start over from basic steps. Run a compressible case without any pulsating pressure and make sure you can get converged results but using the non-reflecting options.


Only have you pass this gate do you apply the pulsating pressure.


Also make sure you don't collect any initial transient data but only after the sound from the inlet has completely filled/saturated the domain. I.e. you need to give the sound time to travel from the inlet to outlet and bounce back and forth a few times.

[oveis87]

Originally Posted by LuckyTran

Given that you were trying to do incompressible acoustics, I think you should just start over from basic steps. Run a compressible case without any pulsating pressure and make sure you can get converged results but using the non-reflecting options.


Only have you pass this gate do you apply the pulsating pressure.


Also make sure you don't collect any initial transient data but only after the sound from the inlet has completely filled/saturated the domain. I.e. you need to give the sound time to travel from the inlet to outlet and bounce back and forth a few times.

Thank you for your constructive recommendation. It was useful.

Regards,
Ov

[Haochey]

Hi,

I hope you worked this out. I have a similar issue that the solution is not converging when using non-reflective outlet. Do you have any suggestions?