Intake Silencing Techniques 2

[Superfry]

I am working on a weight/noise reduction contract for a small military generator. I am trying to investigate methods of reducing intake noise and was wondering if anyone has any experience they can share on the subject. I've seen one type of intake silencer using a tubular silencing technique that uses specific length tubes protruding into the intake filter housing, but I have no idea as to how to size the tubing or plenum, etc. The engine is a yanmar L48 single cylinder diesel, 211cc.
Thanks in advance for any info.
-Tony

 

[GregLocock]

You should investigate the properties of quarter wave tuners, Helmholtz resonators, and expansion chambers. Given that it is a generator it'll run at virtually constant speed, so it should be fairly easy to optimise a tuned system. Quickest way to tune a quarter wave tuner is to build a sidebranch with a moveable piston. That is good fun as well. Low frequency Helmholtz resonators can be a bit tricky to get right.

Any decent acoustics book will have a chapter on this, Beranek or Harris spring to mind. Don't worry too much about exact calculations, they rarely work.

Don't tune it too exactly, the speed of sound will vary with the intake air temperature.





Cheers

Greg Locock

 

[Tmoose]

If you have room A 10X (2200 cc) expansion chamber close to the intake, with a small intake tube on the inlet will take a lot of sting out of the single cylinder pulse. The filters/absorbers/etc on the intake will have an easier time of it.

The intake stuff on snowmobiles is fun to look at and try to figure out.

 

[Rob45]

I tend to agree with Mr. Locock: a Helmholtz tuner will work well for a constant speed engine (and even many variable-speed engines use them), but I disagree that the calculations aren't any good: I have had great success with tuners designed just per the book.
Might depend on the book, of course.
Anyway, what most such books leave out is the actual sizing: in the Imperial system now, mind you - the chamber volume in a Helmholtz tuner should be sized so that the product of the tuned frequency (in cps or Hz, of course) and the chamber volume in cubic inches = 20,000 or more.
If you want to use meters and dynes and such, well, "the calculation of the factor is left to the student as an exercise," as they say.

The advantage of a narrow-band tuner such as this is that, for the silencer volume, weight, and such, the attenuation is much greater than for any comparable expansion volume or absorption-type silencer.

Foremost, however, you should keep in mind that one of the things you're doing with this silencer is attenuating the sound produced by the intact tract itself, as excited by the engine. That is, there is a "system" here, and the engine is merely exciting the system. So you make your job easier by seeing to that some acoustic resonant frequency of the intake "system" is not equal to any harmonic of the engine firing frequency.
For example, this Yanmar single running at e.g., 1800 rpm will have very strong pulses at 30 Hz, so you want to make sure that the intake does not have resonances at 30, 60, 90, or even 120 Hz. If it does, the intake noise would be much greater than if the tract resonance were a little different - let's say at 80 Hz.
If some unfortunate resonance like this does occur, you lengthen or shorten the intake as required to get away from it.
Have fun!
- Robert

 

[jds22]

Superfry,

What rpm is the generator working at?
This gives the dominant frequency of induction noise as it is the induction stroke that creates the pressure pulse.
As Rob45 says 1800rpm would have a strong peak at 30Hz,
but as it would be a 4 stroke cycle the dominant noise in this case would be 15 Hz or 0.5 order... 1 induction pulse every 4 strokes/2 revs.

Resonators are good but also an expansion chamber should be good at broard band attenuation.

What short of volume have you got to play with?
How much noise attenuation do you need?

Let me know and I could have a look at it.

John
Air Induction NVH engineer

 

[Superfry]

I know it's been a while, but this part of the project has just started up again.
John - I'd like to answer your questions and get your advise on a basic approach.
The engine will be running at either 2800rpm or 3600rpm (there are two different speed alternators it powers). Again it is 211cc 4 stroke single(12.9ci). I would like the intake to be around 3 liters. We would like as much noise reduction as possible from the intake with the main limiting factor being weight - perhaps around 1lb max. The generator appears to have two main noise contributors - the intake around 125-500Hz, and combustion knocking around 2kHz. A noise barrier can deal with the 2kHz problem, but the lower intake noise is much harder for a lightweight barrier to attenuate.
This is a low volume application (1000/yr), so I would like to keep the production costs down as well, which i imagine would limit the use of additional resonating chambers. Two approaches I have seen used and I'm favoring are just a simple tuned tube sticking into the intake volume, or the use of several small tubes for the intake (not sure what this is called or how to design them, but they seem to work well on other generators).
Also, I was wondering if anyone has any experience prototyping intakes like this. If so, what methods work well?
Are there any simple programs available for tuning intakes like this?
Thanks again,
Tony

 

[Rob45]

Tony:
I just got through tuning a very effective quarter-wave tuner for an air compressor intake system on a large truck Diesel engine. It provided right at 19 dB attenuation at the problem frequency.
These are also called sidebranch resonators, and are designed so that a closed-end tube connected to the offending intake system has a length equal to 1/4-th the problem frequency. Thus, a wave incident on this tube travels up, is reflected off the closed end, and arrives back at the source 180-degrees out of phase, providing a cancelling signal.
We were working with a 1" ID hose, and simply 'T'-d another 1" tube into it. Since our problem noise was at 224 Hz, we made the hose 1128fps/224 Hz/4 = 1.26' = 15" long. We plugged the end of course...
This reduced the noise level at the driver's ear at idle by 10 dB(A).

One consideration when using one of these tuners is that is needs to be located at a position in the intake where acoustic pressure is high, in order to be effective. You can calculate that position, you can measure pressures, or you can cut-and-try. All three methods work.

 

[distracted]

The problem with all reactive type elements is that they are always tuned to a specific frequency. The peak frequency of the engine varies with rpm. Also, on intakes you do not want to choke the flow as that kinda defeats the purpose.
There are a couple of possible solutions that can help this. If you use a helmholtz resonator you can put a small amount of foam or fibreglass in the bottom of the resonator cavity (or sometimes even at the mouth) this will broaden the attenution curve (but at some expense of the peak attenuation).
The other way that works is to use a sort of expansion chamber (more like the shape of a rabbit in a snake) where the duct expands and contracts. I would have to look it up to find the tuning & attenuation, but it is quite low pres. drop. Sharp on the upstream and more rounded on the downstream edge. Sometimes even a extended tube on the upstream side.

 

[Rob45]

Distracted:
You wrote "The problem with all reactive type elements is that they are always tuned to a specific frequency."

True enough, but you overlook the fact that most intake noise problems occur at a specific frequency.

First, the problem being discussed is with a military genset that will undoubtedly operate primarily at a fixed RPM - and there's your fixed frequency.
Then, just like exhaust systems, noise problems with intakes occur primarily at frequencies where the engine noise is driving some specific acoustic mode of the intake duct. That is, intake noise is not equally loud at all frequencies, but has peaks and valleys determined by the lengths of specific runs of the ducting.

That's part of what makes Helmholtz and quarter-wave tuners so commonly used in solving intake (and exhaust) noise problems. That and the high degree of attenuation achieved.

 

[steelcityDC5]

I've done a little bit of searching, and cant find anything. How exactly is intake resonance calculated? This is something I have been trying to find for quite some time now.


-Brian

 

[GregLocock]

Exactly? By using very expensive programs like Ricardo's WAVE.

There are approximate methods, for simple systems. I'd have thought an Engine Noise and Vibration book would cover that, but I don't know what mine is called, it is in storage.

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[steelcityDC5]

Well, it doesnt need to be too exact, but better than some of the online sites with an approximate runner length calculator.

I am currently working on an ITB and IM design project.


-Brian

 

[Rob45]

Steel:
If you know the temperature in the intake tract, you have the speed of sound; with that, and the intake length(s), you calculate the 1/4-wave frequency(s) for each length.
I.e., f= nc/4L where c=speed of sound, L = intake length, n = 1, 3, 5,...
Now you have a bunch of frequencies: Will engine firing excite any of these?
If so, you make a tuner for that frequency.
Either a Helmholtz or a side-branch, as I discussed months aog on this thread.

 

[porktyme]

Rob45,
What is the consequence of using a Helmholtz resonator that doesn't fit the sizing criteria (frequency*volume>20,000). I am designing a side branch necked Helmholtz resonator for an auto exhaust system and space contraints force me to use a resonator size ~8000 (120Hz*66in3). I am designing this system to reduce a fairly narrow band exhaust drone from ~110-130 Hz (specific calculations indicate 120-125Hz).

 

[magnograil]

The quarter wave side branches are called Quincke resonators. Very sharp tuning over a narrow band. Check out NACA report 1192 on muffler design (which, of course, also works on intakes). They made theoretical calculations and experimental measurements of over 70 designs.
A note, if two Quinke resonators are attached to the main branch across from each other they can cause a noise source from the shedding vortex much like blowing over the opening of a bottle.
Helmholtz resonators work at the lowest frequency for the size. Changing the Q will change the range of frequencies it is effective over. The broader the range the lower the suppression and vice versa.

 

[porktyme]

magnograil,
Excellent idea! I'll use a H-Q tube with an expansion chamber inline. That should save a lot of space. Do you have any idea where I can obtain the relevant equations for simple modeling of this setup (ball park estimates of tube size/length, I'll tweak it during field tests). The NACA report makes mention of the HQ tube but there's little detailed information. I already have a nice spreadsheet setup for necked side branch Helmholtz resonators, but the HQ tube equations have me a bit confused. Thanks.

 

[magnograil]

The Herschel-Quincke liners are used on turbo-fans. Download the NACA report from the Langley(?) site. it has the theoretical and experimental data. Just search on NACA reports.

 

[magnograil]

Sorry, Quincke tubes are not quarter-wave side branches. I just noticed the NACA 1192 report does not include Quincke tubes but does have quarter-wave side branches. Quincke tubes are two unequal length side branches. One of the NACA reports had the Quincke tube calculations. Unfortunately, the laptop at work was stolen over the weekend with all that data on it.
A paper on quarter-wave resonators was presented as a Journal of the Acoustical Society of America 109 (4) April 2001 by Radavich and Selamet with Novak (of FMCo).

 

[Rob45]

Porktyme:
First, if the Helmholtz tuner doesn't meet the minimum volume guideline, it's unlikely there'll be much (if any) attenuation. Going larger than the minimum OTOH doesn't provide any benefit either. Attenuation as a function of the volume x frequency product is a plot that has a sharp knee at approx. 18000 - 2000 "cubic inch-Hz."

Regarding a modified Herschel-Quincke tuner: see SAE paper 2005-01-2356 delivered at the just-completed Noise & Vibration Conference in Traverse City, Michigan. This paper has a good general expression for the modified H-Q tuner, which BTW is not quarter-wave tube, but a full-wave tube connected to a duct in two places, a half-wavelength apart. The modified H-Q tuner puts quarter-wave tubes opposite the H-Q tube connections in a sort of letter-'A' shape.

 

[jds22]

Some things to consider in the intake system.

1. Shell noise from the expansion chamber
2. Inline Helmholtz effect of expansion chamber and dirty side duct (see Mann & Hummel paper SAE 2001-01-1431)
3. Duct lengths - these will amplify noise at resonances
4. Expansion lengths - these will not attenuate at resonances
5. Cost of system ? tooling / piece price ?
6. Flow rate of system - estimated at 0.4 m^3/min (this will be quite a pulsed flow (single cylinder) though so will be different to usual Automotive applications
7. Pressure loss of system - expansion / contraction losses
8. Radiated noise / orifice noise level needed?
9. Filtration efficiency ISO 5011 for details

Other things have been touched on here but this is new ground. My worry would be the inline Helmholtz effect (this amplifies the noise as it in-line rather than side branch) and getting it below the primary harmonic 0.5 order with only 3 litres. In your case this is extremely hard as this is 25 and 30 Hz depending on the rpm.

In your favour is that this will be a very low flow engine so the diameter of the ducting can be extremely small which will imit the radiated noise and help improve the expasion ratio effect of the expansion chamber.

For Automotive NVH, there is an excellent book which will answer a number of these questions by Matthew Harrison, he was my supervisor at Southampton ISVR.

http://www.sae.org/servlets/productDetail?PROD_TYP=BOOK&PROD_CD=R-364

Compressor noise and high frequency induction noise from turbos etc is not really covered in depth.

Also another good book for internal flow systems by D S Miller ISBN 0-947711-77-5, this is my 'bible' for pressure loss calcs and good air flow management.