How is sound generated 1

[izax1]

Can sound be generated without a structure vibrating, and transmitting the vibration to air waves? In other words, if you have a noise problem, will eliminating structural vibration eliminate the noise? This is of course very academical, but I need to understand the sources of noise generation.

I have more details, but that will probalbly just confuse the basics.

Thanks for any help.

 

[MikeyP]

Yes. Turbulence caused by airflow around a rigid object can also generate sound waves. Jet noise is also caused by turbulence.

M

 

[izax1]

Thanks MikeyP

But I am still curious.

In real structures, there is no such thing as a rigid object. I would imagine that to some degree turbulent flow or a jet stream would excite the adjoining structure and cause the structure to vibrate. Will this have no influence at all on the noise generated??

 

[GregLocock]

Not much. It is probably worth pointing out that acoustics and vibration is an area where rough and ready approximations are often best practice. Due to masking and the evils of dB addition we often ignore minor sources of noise and just work on the biggest. Half the time that doesn't come back to bite us!

In this case you are talking about aero-excitation ( I think they call it) and the reason it isn't important for nois4e radiation from most engineering structures is that the impedance mismatch between the air and the structure is great, in both directions for the flow of the energy.

Also, of course, you can't radiate more energy than you put in.

Cheers

Greg Locock

 

[izax1]

OK. Thanks a lot for the acoustic training.

My real problem is a slide valve generating unwanted noise. The noise is generated and amplified at certain slide positions. We still need to do some measurements of pressures, frequencies and levels.

If I understand you guys correct, the noise can be either flow based or structure based. What can be done to find the real source. I take it that I first need to find the source (flow or structure) to eliminate the noise.

Is this done in a "scientific way", or is this based on experience and empirical formulas??

 

[GregLocock]

That's called a noise path analysis. It can be done in a scientific way, but needs a good understanding of the system, a lot of data, and good experimental technique. Empirical tests are often quicker and just as useful for simple systems.

In your case it sounds as though the source is in the fluids, the question is, is the sound transmitted to the oberver via the fluid or the structure?

It's hard to suggest empirical tests without knowing the system, but you can, for example, eliminate noise radiated from the body of the valve by cladding it in glass fibre and lead.

Similar gross changes to the system should eliminate/identify structure borne and fluid borne paths. Cheers

Greg Locock

 

[pjhype]

Flow-induced aeroacoustic noise is often caused by the laminar flow on the surfaces shedding off of the back surface and exciting the air causing noise (vortex shedding). What is usually done is to install something in the upstream flow to break up the laminar flow so that a turbulent flow exits off of the back surface. This will eliminate in-phase vortex shedding and eliminate the flow-induced noise.

Generally speaking, sound waves cannot pass through a "rigid" object. If sound is generated within an acoustic volume, it will excite the surrounding structure and the surrounding structure will radiate sound to the outside. Usually, when the frequency of the acoustic volume matches a structural panel frequency, you will get amplification of the sound to the outside. It is generally difficult to just stiffen the structure to eliminate the noise, as the noise still has energy and this energy will excite other parts of the structure.

For your problem, I am guessing that as you change the position of the slide, you are exposing the surrounding structure to more or less aero-acoustic excitation. If you want to eliminate sound amplification off of the body of the slide, you will need to perform a modal analysis to find out what frequency the body is vibrating at various positions of the slide. You can then stiffen the body (to move the resonance frequency out of the range of the aero-acoustic noise), or you can apply a viscoelastic damping treatment on the body (to dampen the structural vibration and attenuate the structure-radiated noise), or you can put a sound dampener on the body (to absorb the structure-radiated noise). Alternative (non-active) approaches would involve eliminating or attenuating the aero-acoustic noise altogether by changing the flow rate of the fluid in your system or placing a turbulent flow generator upstream of the slide to eliminate in-phase vortex shedding.

pj

 

[izax1]

Thanks a lot pj.

My experience is mainly with structural dynamics, and my first thought was then to shift the natural frequency of the structure to avoid resonance. I understand now I have to look deeper into te problem.

Thank you

 

[izax1]

I'm back again.

I have been talking to some other people about my sliding valve problem. Apparently, as in many engineering problems, the solutions are not quite obvious.

Maybe I did not point out clearly that the "noise" appear to be a single tone (one single frequency).

My simple mind tells me that if this noise is generated inside the valve (and probably by the flow), in order for someone outside the valve, the noise has to be transmitted through the valve housing, and the structure must needs to vibrate at that single frequency.

Assuming my above statement is true, why will not changing the natural frequency of the structure reduce the sound level??

I am aware that it may not be a simple answer to my question, but we intend proposing to the project now to go that route, i.e. measure the frequency of the tone and calculate the natural frequency of the structure to see if we have a match. Is this a waste of time????

Thanks for any additional input.

 

[GregLocock]

What frequency is the tone? What size is your structure? How thick? what material?

You may be lucky and changing the structure may cure it, on the other hand the most likely effect if the resonance is the problem is that you will just change the frequency of the problem, or if the structural resonance is not at the same frequency as the tone then you will scarcely change the radiated noise.

I repeat, clad the valve body in glass fibre and lead. If this does not change the SPL then modifications to the structure are irrelevant since some other path dominates.

You need to do some detective work before launching off into a model, there are too many possibilitis in this system as described. Cheers

Greg Locock

 

[pjhype]

bernt,

What if you cannot identify a structural mode that doesn't coincide with the tone? Or what if you do find a structural resonance that coincides with the tone, but the noise is being transmitted by some other mechanism? Do you tell your project team that you need to "go back to the drawing board?"

You should propose that you will identify the source of the noise in order to determine the best solution to reduce it (e.g. structural modification, flow modification, damping, acoustic barrier, etc). You should then develop a matrix of potential causes of the noise and design some experiments (or analyses) to prove or dis-prove these causes. That way, if your initial theory does not work out, you will have thought of other courses of action that you could pursue.

pj