[czx0415]

I have a question about the reflection of the default walls, can anyone please help me with it? Thanks in advance.

In the BCs of Fluent, the default walls are no-slip walls. Imagining an acoustic wave propagates in the water. When this acoustic wave propagates to the interface of the water and the wall, does the wall 100% reflects the wave energy & without any wave refraction through the wall?

By countrast, in the real world, for the acoustic wave propagation in the water, when this wave meets a wall, there is wave refraction through the wall so the wall "absorbs" some energy of the acoustic wave.

In Fluent, if the default wall does reflect 100% acoustic wave, can anyone please tell me how should I set in Fluent to make the default wall "absorbs" some energy of the acoustic wave (I mean with refraction), like in the real world discussed above?

Thank you.

[LuckyTran]

Pardon the semantics but it is important to your answer your query clearly.


If you are using Fluent (and you say you are) and you have a surface and have selected a "wall" as a boundary condition, then Fluent defaults to a no-slip wall. For further clarification, it means you do not have an "interface" which would generate a wall shadow-wall pair in Fluent. Now I have to assume you are doing some time-resolved (i.e. unsteady) simulation: whether that be unsteady laminar, unsteady RANS, LES, DES, DNS, whatever the case may be and that you use a corresponding viscous flow model (i.e. laminar, turbulence model, some type of Navier-Stokes). If all these apply, then you have the 100% acoustic reflection case. And this is the normal way of doing CFD. The wall reflection happens automatically from the kinematic blocking at a wall. In other words, any time you have a wall in "Fluent", it is acoustically reflecting unless you do something extraordinarily special.


So what are these special cases? You are doing greater simulation by coupling Fluent with an elastic solid modeler (e.g. two-way fsi by coupling Fluent with Mechanica). Or you are using an aeroacoustic module or vibro acoustics module. Or you have created your own very implementation of an elastic wall BC via a UDF. In all of these later cases where you do not directly brute force the combined aeroelastic problem, the absorption spectra is provided via user input and you cannot use "Fluent" to determine how much sound would have been absorbed (because you have to tell that amount to Fluent as an input).

[czx0415]

Originally Posted by LuckyTran

Pardon the semantics but it is important to your answer your query clearly.


If you are using Fluent (and you say you are) and you have a surface and have selected a "wall" as a boundary condition, then Fluent defaults to a no-slip wall. For further clarification, it means you do not have an "interface" which would generate a wall shadow-wall pair in Fluent. Now I have to assume you are doing some time-resolved (i.e. unsteady) simulation: whether that be unsteady laminar, unsteady RANS, LES, DES, DNS, whatever the case may be and that you use a corresponding viscous flow model (i.e. laminar, turbulence model, some type of Navier-Stokes). If all these apply, then you have the 100% acoustic reflection case. And this is the normal way of doing CFD. The wall reflection happens automatically from the kinematic blocking at a wall. In other words, any time you have a wall in "Fluent", it is acoustically reflecting unless you do something extraordinarily special.


So what are these special cases? You are doing greater simulation by coupling Fluent with an elastic solid modeler (e.g. two-way fsi by coupling Fluent with Mechanica). Or you are using an aeroacoustic module or vibro acoustics module. Or you have created your own very implementation of an elastic wall BC via a UDF. In all of these later cases where you do not directly brute force the combined aeroelastic problem, the absorption spectra is provided via user input and you cannot use "Fluent" to determine how much sound would have been absorbed (because you have to tell that amount to Fluent as an input).

Thank you very much for your reply, this is really helpful.

According to what you told me, I still have 2 questions, could you please help with them? Thanks.

Q1: Like you said, to make the wall absorbs sound, one way is to use FSI, which I know how to do with it. However, I am not sure what is 'aeroacoustic module' and 'vibro acoustics module'. Does these 2 modules which can be set in Fluent? Or they are two other packages of ANSYS, so that I need to link them to Fluent? (like FSI, I need to link Fluent package and Static Structural package together).

Q2: Also, you told me that I can 'create my own very implementation of an elastic wall BC via a UDF', to make the wall absorbs sound. In this way, could you please let me know which Macro should I use for this UDF? I think 'Define_GEOM' and 'DEFINE_GRID_MOTION' Macros can describe elastic walls, but I am not sure which Macro can tell the wall to absorb sound.

Once again, thank you very much for your help.

[LuckyTran]

I leave it up to you to discover what modules are available that are suitable for your case since you haven't described what type of wall absorption you want to simulate. Similarly for the UDF.

However, I will say that wall motion is definitely not enough. A moving wall is still a reflecting wall except that now your reflection coefficient becomes complex. That is, it has a non-constant phase relationship w.r.t. an incoming acoustic wave. If in your mind this is suitable, then your wall is actually non-refracting and non-absorbing. Moving hard walls are not elastic walls.

For broadband sound cases it may be possible to simulate a broadband transmission of sound through a wall by hijacking Fluent's non-reflecting boundary condition with net 0 velocity. There is a tunable time constant that you can use to adjust the amount how reflection/non-reflecting is the boundary. However, the nrbc has a broadband frequency response function and will not work for cases with distinct frequency responses (i.e. absorption caused by structural resonances).

[czx0415]

Originally Posted by LuckyTran

I leave it up to you to discover what modules are available that are suitable for your case since you haven't described what type of wall absorption you want to simulate. Similarly for the UDF.

However, I will say that wall motion is definitely not enough. A moving wall is still a reflecting wall except that now your reflection coefficient becomes complex. That is, it has a non-constant phase relationship w.r.t. an incoming acoustic wave. If in your mind this is suitable, then your wall is actually non-refracting and non-absorbing. Moving hard walls are not elastic walls.

For broadband sound cases it may be possible to simulate a broadband transmission of sound through a wall by hijacking Fluent's non-reflecting boundary condition with net 0 velocity. There is a tunable time constant that you can use to adjust the amount how reflection/non-reflecting is the boundary. However, the nrbc has a broadband frequency response function and will not work for cases with distinct frequency responses (i.e. absorption caused by structural resonances).

Thank you so much for your reply, which is really helpful!