Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Scattering by a half-plane
Posted 2025年7月18日 GMT+8 23:18 Acoustics & Vibrations, Modeling Workflow, Physics Interfaces 6 Replies
Please login with a confirmed email address before reporting spam
Hello,
I am simulating the scattering of acoustic waves by a half-plane. The domain consists of a rectangular box with the half-plane at the bottom. The box is surrounded by PMLs shown in purple in the image below. Periodic boundary conditions are used in the front and back faces and the background field is a plane wave in the x-direction. I am using the pressure acoustics, frequency domain module and I would like to add a background mean flow in the x-direction.
I have thought of two possibilities:
1) Use the linearized Euler Equations first just to add the mean flow, and then use the acoustics module, but I didn't manage to couple the two things.
2) Modify the wave equation to go from the Helmholtz equation:
to a convected Helmholtz equation
I tried using the dipole source as this will give the first order partial derivative, but the results I got don't seem to agree with the analytical model I have.
Is there any better way to include the effect of the background flow?
Thanks
Attachments:
Please login with a confirmed email address before reporting spam
Hi Jorge,
If you can assume an irrotational and inviscid background flow, then use the Linearized Potential Flow, Frequency Domain interface. You can specify the mean flow simply by a user-defined expression, or solve a CFD problem for the mean flow using the Compressible Potential Flow interface. You can find this example of the former: https://www.comsol.com/model/doppler-shift-1376
-Mark
Please login with a confirmed email address before reporting spam
Hi Mark,
Thanks a lot for your reply!
I understand your point of view, however for validation of my simulation I must radiate the pressure on the plate to the far-field and compute the directivity patterns at different non-dimensional frequencies. I am trying to use the acoustic pressure, frequency domain since the far-field computation is straightforward using the far-field module (K-H integral). For the flow I can assume only inviscid and irrotational flow, so I guess it should be either linearized compressible or linearized Euler.
Do you think it would be possible to include the mean flow effect by adding a source term or other modification in the acoustic pressure, freq domain module?
Thanks
Jorge
Please login with a confirmed email address before reporting spam
It sounds like you want to compute far-field directivity pattern for when there is a background flow. Pressure Acoustics assumes a quiescent fluid and also the Kirchoff Helmholtz equation used to evaluate the exterior field calculation in pressure acoustics is generally not valid if there is background flow condition.
There is currently no built in feature for evaluating exterior field when there is background flow. How far away from the source are you evaluating? Is it possible to make the domain larger to get some far-field response in the physical domain?
-Mark
Please login with a confirmed email address before reporting spam
Yes, indeed I have looked everywhere and couldn't find any way to introduce the background flow. From an acoustic perspective, the Linearized Euler, frequency domain should be equivalent to pressure acoustics? If that's the case then I could use a bigger domain to capture the far-field and have the background flow effect. I am interested in the directivity patterns sufficiently far from the source, I am thinking somewhere between 2 and 3 wavelengths. Do you think this would be enough?
Thanks for your kind help!
Please login with a confirmed email address before reporting spam
For the sake of computational efficiency, we recommend using Linearized Potential Flow, Frequency Domain interface. This solves the wave equation for an acoustic velocity potential in the presence of an irrotational and inviscid background flow. When there is no background flow, this is equivalent to Pressure Acoustics, Frequency Domain. Try solving first with a smaller domain and progressively increasing the size to compare directivity patterns as well as scale of the problem/computational resources used.
-Mark
Please login with a confirmed email address before reporting spam
Sounds good. I am just wondering what is the equivalent of a Background pressure field in Linearized Potential Flow, frequency domain?
Thanks
Jorge