Solving the problem of an organ pipe in air

[Organ_Mark]

I have been working on the problem of an organ pipe in air. Thanks to help form Saint Michael, I got a series of faces to define an organ pipe in a box of air using Salome. Most of the faces of the pipe are very high impedance to simulate a hard surfaces. One face of the pipe injects a pressure wave. The six faces of the surrounding box have impedance 343 which should let the energy escape like a plane wave (I think). There is one small segment of the pipe that has no face and connects the inside air with the outside air. I am using the Helmholtz solver.

The problem is, when I examine the solution, it looks like the walls of the pipe are not there. The source is working as it should and it propagates a mainly spherical wave to the surrounding box.

To summarize, there is air on both sides of a face which has zero thickness but finite impedance and it seems as it the face is not there.

[raback]

Hi Mark,

Imagine standard FEM where each node is associated with one pressure value. Do you have same nodes being used inside and outside of the organ pipe? As the pressure is the same they will have the same value.

You would have to have a mesh that has different values outside and inside. That can be done either by meshing or using more exotic discontinuous Galerkin approach where you enable jumps. The 1st is easier.

Is the pipe passive or do you want to solve a coupled system of vibration and wave propagation. The 1st is easier.

So this could be an easy meshing issue or at worst a bleeding edge stuff (coupled vibroacoustics with discontinous pressure field) where even a developer would spent days.

-Peter

[Organ_Mark]

Thank you very much Peter for the quick and learned reply. I think I don't know how to tell if I am using the same nodes inside and outside the pipe. I will tell you how I generated the geometry and I will show you my results. I suspect I am using the same nodes inside and outside. You can probably confirm the answer based on my results. I intended the pipe to not be real in the sense of a face that it only has impedance properties - not a pipe with mass and spring properties of its own.

I made the mesh by using Salome. I constructed the pipe by assembling a model that has the shape I wanted and then exploding the model. Then I selected the faces I wanted and used them as tools in the partition operation with the main object being the box of air. Then I get a composite that has one body (air connecting the inside and outside of the pipe) and a bunch of faces. I assign the faces to impedance of high or 343. One of the faces gets the source pressure assigned.

Please see the attached figures to get a better understanding of the results.

I think I need to know how to manipulate the mesh in the right way.

Thank you for help Peter.

Mark

[Organ_Mark]

I did a little experimenting and I may know what to do. Instead of a single face separating outside and inside the pipe, I can have 2 faces separating outside and inside. And the space between the faces can be very small, but there is no material between them to connect outside to inside. In this way you get interaction with the face from one side only which is what we want.

Is this the right idea? Is there a better way to do it?

[kevinarden]

Sometimes meshers make coincident nodes at faces. There is an option in Salome to merge coincident nodes to remove the problem.

[Organ_Mark]

Thanks for the help friends.

In the end I put in space for the walls of the organ pipe with each side of the wall having its own boundary condition. That seems to work fine although it did complicate the model a lot.

[raback]

Hi

Maybe you could share some nice pictures. I'm sure you got some?

-Peter