four pole noise analysis?

[johnab]

I have 4 pole noise analyis of air intakes system (db vs frequency)in company database. (someone did in past & has already left the company & nobody uses till today). As I was requested to use the software so trying to understand the previous model & have a question if anyone can help!

The Air intake system having an air cleaner, 3 resonator, 2 side branches, dirty & clean side tube (8 items). I am inserting dia & length for each component to find out the db vs frequency plot but when i saw the past model it has 16 length & 16 dia for the same model? How did that happen?
(I was thinking by myself that the duct coming into contact with atmosphere need corrective lentgh but couln't find out the solution)

Thnks,

 

[GregLocock]

How much experience do you have with analytical acoustics? Or experimental acoustics? That is rather a complex model, I very much doubt that a noob will get much meaningful data out of it. Sorry.

The end effect correction is 0.3D, off the top of my head.



Cheers

Greg Locock

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

 

[johnab]

Not a whole lot as novice to Auto NVH. Anyway, For the future model,I need to run some analytical part & then need to compare actual test in an anechoic environment.
I have few formula from past employee for effective length calculation. Anyway, can anyone explain/ refer book having simple example to quick start considering detail understanding will take time!

 

[GregLocock]

4 pole acoustics models for intake systems were largely developed by P.O.A.L. Davies at ISVR in the UK, so papers by him may be helpful.

The package he developed (I think) was called LAMPS. I never found it any practical use whatsoever, for exhausts.



Cheers

Greg Locock

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

 

[johnab]

I tried to look even by Munjal but didn't find besides mathmatical equations (definitely important from deriving formula but doesn't helping me to understand)

let me know if you know something specific that tells how to define the length, dia & other damping parameter that will help to plot the attenuation vs frequency plot?

(where frequency, F = ½ pie under root A^2 B / V * ALRho)

 

[bukac1]

Best source of info is: M.L Munjal,Acoustics of Ducts and Mufflers, John Willey & Sons, 19897.
However, the Four-pole method is an obsolete stuff. A method based on solution of the wave equation is much better.
I will be presenting a paper on this method at this year's 18th International Compressor Engineering Conference at Purdue University, July 17 - 20.
Good luck!

 

[johnab]

I understand that paper details would be confidential yet but can you little bit describe solution of wave equation method & comparison with four pole, if only possible

Thnks

 

[SomptingGuy]

There are several commercial codes out there that solve the wave equation in 1D for intake system analysis and most OEMs will have their own in-house varieties too.

Where I work we ditched the linear methods back in the late 1980's and all work is now routinely carried out using a proprietary 1D code.

There are many benefits. Most important for me are:

- Simultaneous performance prediction (noise vs performance tradeoffs), since you model the whole system: intake, engine, exhaust.

- Actual audio output. You get to hear what it'll sound like. Ideal for today's "sound quality" fixation.


BTW, what sort of company do you work for? An OEM, a supplier, an aftermarket manufacturer?

 

[johnab]

Thank you for reply, I work for Tier II supplier

 

[SomptingGuy]

Ok, cards on the table: I work for a supplier of afore mentioned 1D codes (as my profile hints). Professional ethics prevent me from saying which one, but I can assure you they are the way to go for intake system development. Forget the 4-pole stuff.

 

[johnab]

Somptingguy:

I'm not sure but r u talking @ GT-power (or similar)? as i heard using it for speed vs db plot

suggest me if any professional software available so i can think @ to purchase it?

 

[bukac1]

To explain application of direct solution of the wave equation in the design of mufflers requires solid knowledge of mathematics (partial differential equations, matrix algebra, solution of systems of equations in complex numbers,and programming.
I am sorry for you. You are, or your boss is, probably one of those mouse click happy engineers who believe that clicking a mouse will produce results. Besides clicking the mouse on some software you need to understand mathematics that is behind it.
Many years ago, when I started to study engineering, the dean of engineering told us: "Engineering is, it always was, and it always will be applied mathematics. If you don't like math, go somewhere else, but do not mess with engineering."

 

[SomptingGuy]

Direct solution of the wave equation is best left to mathematicians.

All the commercial (and in-house) 1D codes solve it using numerical methods - typically using finite volume or finite difference methods (although some older codes still use the method of characteristics approach). Their numerical accuracy is largely dictated by the quality of their models for duct flow, flow through orifices and area changes and flow through collectors/splitters.

Given that most commercial codes are adequately accurate when used sensibly, their commercial value can only be determined experimentally, by using them. An experienced user can typically cut the number of physical muffler prototypes down by an order of magnitude.

But as bukac1 points out, there are always those "who believe that clicking a mouse will produce results". This is true of all tools and I believe there's a well known saying connecting workers, blame and tools. The challenge for the tool providers is to provide an environment that guides users toward sensible modelling techniques and away from stupid ones.