[Bananenflanke]

Hi,

I am currently investigating sound waves in a pipe with the aim of determining the amplitude and phase of the sound wave. To determine the amplitude and phase, I measure the pressure at different positions along the pipe. As a model, one can imagine that the pressure in a flow is composed of a constant mean value, a value that fluctuates over time due to the sound waves, and possibly another fluctuating value due to pressure fluctuations that are transported with the flow (for example, eddies/turbulence). For the evaluation of the acoustics, only the acoustic pressure fluctuations are relevant and all other pressure fluctuations interfere with the evaluation. For an optimal acoustic evaluation of the pressure signals, a distinction must therefore be made between the acoustic pressure fluctuations and all other pressure fluctuations.

If it can be assumed that the acoustic pressure fluctuations dissipate slower than the other pressure fluctuations, the pressure could be measured far downstream. However, this methodology has the disadvantage that the calculation domain must be long. Other procedures can be found in the literature in which the non-acoustic pressure fluctuations are taken into account or filtered out during the evaluation, so that the calculation domain can be shorter. In general, these evaluation methods make use of the fact that the acoustic pressure fluctuations propagate at the speed of sound and the other pressure fluctuations propagate with the speed of the flow.

Problem definition:
In order to test the different evaluation approaches, I would like to superimpose sound waves with pressure fluctuations propagating with the flow velocity in a straight pipe. So far, I have generated sound waves in the calculation domain by specifying the velocity component at the inlet in the axial pipe direction in the form:



The real question is how I can generate the convective pressure fluctuations in a straight pipe?
It would be important that no further acoustics are created by the generation of the convective pressure fluctuations so that the evaluation results can be interpreted meaningfully. Furthermore, the frequency of the pressure fluctuations should be adjustable in order to be able to investigate the relevant frequency range (~10-1000 Hz). And it would be desirable if the amplitude of the convective pressure fluctuations could also be adjustable. Does anyone have an idea how I can realise this?

My first rough ideas, none of which I have been able to try out yet due to time constraints:
1) Use turbulence sources
2) Set the turbulence at the inlet
3) Provide an additional inlet in the pipe wall and let the fluid flow in radially at a certain frequency in an "impulsive" manner
4) Divide the inlet of the pipe into two surfaces and apply acoustic excitation to one surface and convective excitation to the other surface; the convective excitation takes place by specifying a radial velocity component and in a "pulse-like" manner

Since I have not yet dealt so intensively with turbulence modelling, I still have some unanswered questions about ideas 1) and 2). However, I could imagine, for example, that I can set the amplitude of the convective pressure fluctuation by specifying the turb. kinetic energy and realise a fluctuation by multiplying it with a sin function. With ideas 3) and 4), I can imagine that acoustics are also created in the process, whereby the sound waves initially propagate in a radial direction and thus possibly only have a small influence on the evaluation, but that would be one thing I would have to investigate.

Does anyone perhaps have further ideas that I can try out or comments on my previous ideas. Do you think they are useful or do you have any suggestions for improvement? I would be grateful for any kind of help!

Best regards and stay healthy!

[ghorrocks]

If it can be assumed that the acoustic pressure fluctuations dissipate slower than the other pressure fluctuations

What is the basis of that assumption? Why would one type of pressure wave dissipate slower than another type?

the other pressure fluctuations propagate with the speed of the flow.

Yes, but the pressure fluctuations from flow generated sources generates pressure waves which propagate just like acoustic waves. In fact, they often do emit sound waves - think of a flute or a rope humming in the wind....

I have generated sound waves in the calculation domain by specifying the velocity component

If you are interested in the pressure wave, then why don't you specify the pressure wave directly with a pressure boundary condition?

My first rough ideas, none of which I have been able to try out yet due to time constraints

Wow, why have you considered such a complex way of generating a pressure wave? Why not just use a pressure boundary and generate it directly?


While those are my comments to your post and what it raises directly, the whole concept of what you are trying to do does not seem right to me. If you are only interested in the propagation of acoustic waves in a system then you do not need CFD, you need an acoustic modelling package. The equations describing acoustic flow are very different than the Navier Stokes equations used by CFX and other CFD packages. The acoustic equations are linear and simple to solve, so you should have an answer in no time.

Trying to model acoustic wave propagation with any CFD package is doing it the really, REALLY hard way.

[Bananenflanke]

Hi ghorrocks,

Thank you very much for your answer! You are right that an acoustic modelling package would actually make the most sense for the investigations described. In the future, however, not only the sound propagation in a pipe but also the sound generation in much more complex components will be investigated and it has been decided that these later investigations will be carried out with a CFD package. The aim of the preliminary investigations is to gain as much experience as possible. For this reason, the investigations described above should also be carried out with a CFD package.

What is the basis of that assumption? Why would one type of pressure wave dissipate slower than another type?

Perhaps I expressed myself incorrectly - unfortunately, it's not my native language. I meant that a sound wave propagates further than a pressure fluctuation, which propagates at flow velocity, before it is completely dissipated. I didn't mean to state it as a fact either, but it was stated that way in a paper for a specific case of investigation.

Yes, but the pressure fluctuations from flow generated sources generates pressure waves which propagate just like acoustic waves. In fact, they often do emit sound waves - think of a flute or a rope humming in the wind....

Yes the pressure fluctuations propagate like sound waves, but they don't propagate at the speed of sound but at the speed of flow, right? And as long as these pressure fluctuations don't hit any obstacle (like the flute or the rope), no sound waves are generated. At least that is my understanding of the matter. According to this, pressure fluctuations that propagate with flow velocity in a straight pipe should not produce sound waves.

If you are interested in the pressure wave, then why don't you specify the pressure wave directly with a pressure boundary condition?

For plane waves, the amplitudes for pressure and speed can be converted into each other with the help of impedance (density times speed of sound), so it doesn't matter what you specify in the boundary condition. For me personally, however, it is easier to work with velocities.

Wow, why have you considered such a complex way of generating a pressure wave? Why not just use a pressure boundary and generate it directly?

For example, if I vary the pressure with a sin-function at the boundary condition, I create a sound wave but not a pressure fluctuation which propagates with flow velocity. But I want to have both at the same time.Therefore, this is unfortunately not the solution and the reason for the complex ideas.

So the goal is to create a sound wave and a pressure fluctuation which propagates with the flow velocity (like a wave). As much as possible should be known about the sound wave and the pressure fluctuation (frequency, amplitude) in order to be able to interpret the measurement data meaningfully. I can generate the sound wave as described above. But how can I generate the other pressure fluctuation?

Any suggestions are appreciated!

[ghorrocks]

I meant that a sound wave propagates further than a pressure fluctuation, which propagates at flow velocity, before it is completely dissipated. I didn't mean to state it as a fact either, but it was stated that way in a paper for a specific case of investigation.

This is incorrect. A pressure wave is a pressure wave, regardless of its source. A pressure wave generated by a flow feature propagates at the acoustic velocity, not the flow velocity. You might be getting confused with the velocity field of a flow feature, which is convected at flow velocity. But even this way of looking at it is problematic, as any wave is a change in pressure and velocity - to change the pressure the fluid has to move, and this movement is a velocity. So an acoustic wave is a velocity wave as well as a pressure wave.

if I vary the pressure with a sin-function at the boundary condition, I create a sound wave but not a pressure fluctuation which propagates with flow velocity.

As mentioned in my previous comment, pressure fluctuations propagate at the acoustic velocity regardless of source.

So the goal is to create a sound wave and a pressure fluctuation which propagates with the flow velocity (like a wave).

This is not possible, all pressure fluctuations propagate at the acoustic velocity.

[Bananenflanke]

Hi ghorrocks,

Thank you for your clarifying comments! It seems I got something mixed up.

You might be getting confused with the velocity field of a flow feature, which is convected at flow velocity.

In my previous comments I always meant the velocity field of a flow feature, which is convected at flow velocity. Since pressure and velocity are coupled, my understanding is that the velocity field should also result in a pressure field. Therefore, both the velocity field and the pressure field of the flow feature should move at flow velocity. If the velocity field of the flow feature is not stationary, the pressure field of the flow feature should also not be stationary. The values of the velocity field and the pressure field of the flow feature therefore "fluctuate" in time. These are the "fluctuations" that I want to create in the pipe in addition to a sound wave. In principle, this should be something like a vortex and a sound wave moving through the pipe.

That should be possible, or have I made some kind of mistake in my thinking?

[ghorrocks]

That clarifies things. Yes, you can impose a vortex on a flow field and the vortex will convect with the flow field. The vortex will have a velocity and pressure field associated with it. But it will also radiate pressure waves which radiate away at the acoustic velocity, and these pressure waves are actually coupled pressure/velocity waves as well. Isn't fluid mechanics wonderful stuff?

You will find that imposing something like a vortex at an inlet is not simple. You need to define the rotation motion of the vortex over time as it passes through the inlet plane. There will be some studies in the literature which have done this, I would recommend a literature review to see the various ways they have implemented it.

An idea which is probably simpler is to have the inlet boundary as a fixed velocity inlet, but shortly afterwards have a momentum source term which puts a rotational vortex in the flow shortly downstream from the inlet. This greatly simplifies it as you can simply apply a rotational momentum source for a period of time and turn it off - much simpler than applying it at an inlet.

[Bananenflanke]

Thanks for your quick reply and glad we could clear up the misunderstanding! Fluid Mechanics is really cool once you understand it

You are right that it would not be easy to create a vortex directly at the boundary condition. I really like your suggestion with the momentum source. I will try it out as soon as I can!

[cfd_de]

Originally Posted by ghorrocks

This is incorrect. A pressure wave is a pressure wave, regardless of its source. A pressure wave generated by a flow feature propagates at the acoustic velocity, not the flow velocity. You might be getting confused with the velocity field of a flow feature, which is convected at flow velocity. But even this way of looking at it is problematic, as any wave is a change in pressure and velocity - to change the pressure the fluid has to move, and this movement is a velocity. So an acoustic wave is a velocity wave as well as a pressure wave.

I am looking for the difference between the two types of Pressure waves. The pressure wave that is being generated because of a flow feature and the other pressure wave which is a sound wave.

If both move with the speed of sound in a medium then what's the difference between these two types of pressure waves ? How we can categorize and differentiate between these two types of pressure waves ?

In Pysics do we really say that this is a sound pressure wave and this is flow generated pressure wave ? do anything exists like this ? Sound waves are gain generated because of the flowing fluid, so i am confuse how we can make the difference ? Please help

[ghorrocks]

For an analogy, think about waves on the ocean. When a boat bobs up and down on the ocean, was it an ocean swell wave? Or a wind chop wave? Or the wake from a boat? Or a reflected wave from somewhere? Sometimes you can identify the source of a wave, but most of the time you cannot. There are just waves flying around everywhere and they combine to form the wobbly surface of the ocean.

It is the same with pressure waves in air. When a single source of wave has some clear identifying feature (maybe frequency, maybe direction) then you might be able to tell what the source of the wave was. But often this is not possible and there is just a myriad of waves heading all over the place and identifying which wave is which is not possible.

But there is no underlying physical difference between an acoustic wave and a flow generated wave. They are both just pressure waves.

[cfd_de]

Originally Posted by ghorrocks

For an analogy, think about waves on the ocean. When a boat bobs up and down on the ocean, was it an ocean swell wave? Or a wind chop wave? Or the wake from a boat? Or a reflected wave from somewhere? Sometimes you can identify the source of a wave, but most of the time you cannot. There are just waves flying around everywhere and they combine to form the wobbly surface of the ocean.

It is the same with pressure waves in air. When a single source of wave has some clear identifying feature (maybe frequency, maybe direction) then you might be able to tell what the source of the wave was. But often this is not possible and there is just a myriad of waves heading all over the place and identifying which wave is which is not possible.

But there is no underlying physical difference between an acoustic wave and a flow generated wave. They are both just pressure waves.

Yes you are right, it is really difficult to identify the source of the waves. We just look how they behave and their effects.

But once we say acoustic pressure wave and flow generated pressure wave there must be some difference regarding their behavior atleast ? isn't it ?

The pressure wave which we identify from simple CFD analysis by drawing contours on a plane is a flow generated pressure wave ? Right? This isn't an acoustic pressure wave? Right?

Also from a CFD analysis the pressure waves which get reflected from a boundary are also flow generated pressure waves ? Right?

[ghorrocks]

You seem to be still trying to label pressure waves as either flow generated or acoustic. As I said in my previous post, all pressure waves are just that - pressure waves. Sometimes you can trace the waves origin back to a source, but often you cannot. And whether you can trace it back depends on what you are simulating, what else is around, how complex it is and so on.