# Simulating open water acoustics in a tube

For the purpose of testing an underwater acoustic transmission it is possible to use a tube filled with water with a transducer at one end and a hydrophone at the other. The problem is that sound will bounce off the walls of the tube, making the results different than if the same test was done in open water.

Is there any material or method which can be used for constructing the tube or channel such that sound would not bounce off the sides of the tube and instead be absorbed, so that the resulting received signal would be similar to that received in an identical experiment in open water?

• You could make the walls absorbing, but that's the wrong strategy. The correct strategy is to use short pulses and only measure during the initial time before the first reflection can reach the hydrophone. A tube is the wrong geometry for that. – CuriousOne Feb 12 '16 at 0:20

## 1 Answer

I have dealt with some acoustics of waveguides but just in gas media. Honestly, I am not aware of any material of proposed qualities, but it's not any ultimate evidence.

Anyway, I gave some thought to this. Let me offer at least a focused discussion. Sadly, CuriousOne's comment is right. It is generally not a good idea, but let's investigate some limit cases.

• Standing waves: In an unbounded fluid field there would be no standing waves, so you should always be aware of any amplifications due to this phenomena, which is on the other hand typical fur tubes (generally, the first mode of duct medium oscillations is associated with the motion of the whole mass as one rigid body, the next present standard sets of nodes and antinodes based on boundary conditions). Beware of so called cut-off frequency above which not only modes along the tube, but the "cakes" and "circles" in transversal plane could be present (i.e. $\lambda$ enough small to fit in the tube transversally).

• Boundary layer: Beside the standing waves this is the second worst enemy. Mathematically it is still the chapter "boundary conditions". Vortex sheets located at the tube walls cause thermoviscous losses very well observable in impuls response measurements.

Solution (?)

I would try this one:

• To cancel the boundary layer effects, don't use a tube ($\approx$ 1D) but an aquarium or pool (3D). Do yourself a favor and use a rectangular one. :-)
• Calculate the aquarium eigenfrequencies (use the Helmholtz equation for simplicity).
• Measure the test signals (pulse (better) or sweep) and in obtained data processing cancel the peaks associated with eigenfrequencies.