PDA

View Full Version : Splitter silencer design



nh3simman
23-03-2007, 07:07 AM
Does anyone know the formula for insertion loss of a splitter type silencer?

I know that it must be related to:

surface area of acoustic material
cross sectional flow area
material absorption coefficient
The Woods practical noise guide gives a table of insertion loss for different lengths, air passage width. But this table also breaks the insertion loss into 8 frequency bands.

Is it safe to assume that the same equation would be applicable for all frequency bands?

nh3simman
25-03-2007, 04:53 PM
I find in the Buffalo Forge "Fan Engineering Handbook" this equation.

I = 12.6 (Ao/Ax) a^1.4

I is the insertion loss

a is the absorption coefficient of the material.
I assume that you need a different value for the 8 different frequency bands.

Ao is the surface area of the material.
Ax is the cross sectional flow area.

When I compare this equation to the Woods Practical Noise Guide, it doesnt work!!!:eek:


Someone, please make a suggestion...

nh3simman
26-03-2007, 10:32 AM
How hard can it be?

The silencer is just a piece of duct that is lined with acoustic material and has a few splitters.
http://www.coolit.co.za/silencer/Silencer.gif
If you buy one from the silencer manufacturers, you get exactly that.

Oh yes, you also get a guarantee of insertion loss. It is expensive because of the guarantee, not the technology.

By adding splitters, you reduce the effective airflow cross section and you increase the exposed surface of the acoustic media.

from the above equation then
I = 12.6 (Ao/Ax) a^1.4
adding splitters will increase I.

So, why not increase the number of splitters to the maximum.

By adding splitters, you increase the air pressure drop.

BUT this is the crux.

You also start to increase the self generated noise.

See how much noise you can make by blowing a small amount of air across a thin film of plastic. And this effect is not shown in the Buffalo Forge equation.

I'm thinking that the actual equation should be

I = 12.6 (Ao/Ax) a^1.4 - (self generated noise)


Does anyone know how you calcuate self generated noise?

Lc_shi
30-03-2007, 07:08 AM
good topic,where you got the formula?

nh3simman
30-03-2007, 11:15 AM
good topic,where you got the formula?

Hi LC

I found the insertion loss formula in the Buffalo Forge Handbook.

Only later after reading the text, I saw that the Woods Noise Control book has the formula in the text. No equation, go figure!

I have been searching on-line, in text books and asked the so called sound experts for years without getting a complete answer. Noise seems to be one of those black arts. I hate not knowing...

We had a talk by Mats Sandor from Systemair, Sweden. They have amazing sound facilities and are experts on sound. I asked him at the talk and the best he could supply was the Woods eqn.

But it doesn't work and is clearly not complete since it does not include the self generated noise.

If I am having so much dificults finding this information and the experts can't give a clear answer, then I have to question the "design" material used by the manufacturers.

What do you think?

nh3simman
03-04-2007, 04:53 PM
I agree with everything said so far.

Oh, hold on, I see I have been talking to mostly to myself. Thanks Lc at least you didn't leave me totally alone here.

My conclusions:


Nobody cares about splitter silencers.
Nobody knows about splitter silencers.
Everyone except me knows about splitter silencers and you have decided not to tell.
Noise is never a problem.

Feel free to make comments even if the thread is a years old. I have been searching for years anyway so what's a few more.

Josip
03-04-2007, 11:43 PM
Hi, nh3simman :)


I agree with everything said so far.

Oh, hold on, I see I have been talking to mostly to myself. Thanks Lc at least you didn't leave me totally alone here.

My conclusions:


Nobody cares about splitter silencers.
Nobody knows about splitter silencers.
Everyone except me knows about splitter silencers and you have decided not to tell.
Noise is never a problem.

Feel free to make comments even if the thread is a years old. I have been searching for years anyway so what's a few more.

It is not the truth that you were talking only to yourself. Many people just read posts without any replay.

For me number 1,2 and 4 :D sorry not my field but was nice to read your posts. Never too late to learn something new;)

Best regards, Josip:)

momo
04-04-2007, 03:54 AM
No audio/sound/noise control system analysis would assume same loss at all frequencies... at the simplest level a F with a half wavelength greater than the duct would get through...
It is a great area of formulas, numbers etc especially for some mortals that play around with recording studios etc. terms like bass traps, absorption, refraction, reflection.. creep in and you get into resonance and fluid dynamics deeply - great living to be made for maths minded, not me :o.
However it is more and more important since the European right to sleep quietly is progressively applied. Carnival and fiestas included (see press re near cancellation of Tenerife's this year:p)
The type you refere to seems to be used near ventilated compressor rooms; with external condenser fans: design and the larger and slower rpm the better.

nh3simman
04-04-2007, 04:06 AM
No audio/sound/noise control system analysis would assume same loss at all frequencies.

Hi momo, Quite right. I quoted an absorption coefficient (a) but as you imply, this value is usually given as a function of frequency band.
63 125 250 500 1000 2000 4000 8000 Hz

I have looked at some of the sound studio stuff to try get some clues about noise reduction but they have different criteria.

I once asked a sound expert if there was a formula that could be used to give the the insertion loss of a splitter silencer. He said NO. At our next meeting, I showed him the equation in the Buffalo Forge Handbook and he looked at me like I was crazy.

nh3simman
06-04-2007, 09:48 PM
I emailed this thread to an expert in the field of sound and he gave me some direct feedback.

Bad news I'm afraid. I am assured that there is no equation that can be used to calculate the insertion loss of a splitter type silencer.

A bit difficult to accept? You bet. Engineers design space shuttles, nuclear power station, we can predict the behavior of combustion in a cylinder, with a few simple relationships I can calculate the performance of a heat exchanger. But it is impossible to predict the noise insertion loss of air travelling through a lined passage!

The film heat transfer and pressure drop coefficients of a turbulent fluid flowing in a pipe are complex phenomena that relate to the boundary layer. Despite this complexity, we now have very reliable means of calculating these factors. These are experimental relationships that have stood the test of time. Surely, the people who work with sound do a similar thing?

I'm not convinced and will keep looking.

US Iceman
06-04-2007, 10:12 PM
You are not alone in your conundrum about silencers. :confused:

I spent a fair amount of time trying to figure this out many years ago and finally realized this is more black art than science.

I was researching this for muffling the noise from compressor discharge lines and oil separators. For low mole weight refrigerants this does not occur very often. For higher mole weight refrigerants this can be a very big problem for large screw compressors (and very big oil separators as they act like speakers).

The guys who can probably give you an answer are the same people who design high-performance mufflers for automotive engines.:cool:

The best information I could find related to finding the acoustic velocity and wavelength and trying to throw it out of phase to cancel the noise. In theory it sounds reasonable, but I never could generate any interest in developing this further than an exercise in frustration.

nh3simman
07-04-2007, 05:27 AM
The guys who can probably give you an answer are the same people who design high-performance mufflers for automotive engines.:cool:


Good point, I will do some searching in that field.

US Iceman
07-04-2007, 05:43 AM
The interesting thing about automotive mufflers is if they can "tune these" for different sounds they must have some way of predicting the sound bands.:cool:

Brian_UK
07-04-2007, 11:19 PM
As an aside to this discussion, which I admit I left alone years ago because I let the 'experts' get on with and give me the result to fit, I had to remove a sound attenuator on the inlet to an AHU the other day.

They had fitted it back to front, the shaped inlet was on the air outlet end with the flat pod ends facing the incoming airflow.

Makes you wonder sometimes.....

Also I admire the guys who do sound analysis.

Once had a recording studio that had a low level noise breakout which we couldn't track down.

We used a noise analysist along with a vibration analysist and between them they identified a faulty anti-vibration mount on the supply fan situated some three floors above the studio.

buddy3141
10-09-2007, 04:40 PM
You are having vibration problems with a pump, correct? You might take a look at possibly trying somethink like the pulsation dampener that is described at Hydril's website.
Good luck.

buddy3141
10-09-2007, 04:46 PM
A volume-choke-volume acoustic filter may be another option for reducing the pulsations, which cause piping vibration, that pumps inherently generate.