Reducing acoustic energy loss from bellmouth open pipe

[edjza80]

Heya folks, i have done a chunk of research (not my field!), and am having trouble relating the infomation im reading back to my specific query... which is:

regarding... the delicate balance between reducing the pumping losses by applying a bellmouth to a duct termination, and at the same time, suffering the increased loss of radiated acoustic energy from within the pipe at this same bellmouth interface.

obviously the acoustic boundary for a standing wave is not as strong in a pipe with a bellmouth. How can i make this acoustic boundary better, without totally screwing up the flow character of the bellmouth?

ideas, or reading specifically focused at this problem would be appreciated.

cheers
ed

 

[edjza80]

err... this may have landed in the wrong subforum.... sorry!

 

[GregLocock]

Well you lucked in since I did a little work on that a long time ago.

Roughly speaking, I think you are worrying unnecessarily. The reflection coefficient of any feasible bell , given the dimensions involved, is very close to 1.

FWIW Pout/Ppipe for a tailpipe is around -30 dB, that is 95% plus of the sound pressure gets reflected.

keywords for your google search would include tube, termination, intensity in a duct, reflection coefficient, acoustic impedance.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[SomptingGuy]

One good paper that covers this area in huge detail is:

"Wave reflections from duct terminations"

A. Selamet and Z. L. Ji (Ohio State University)
R. A. Kach (Ford Motor Company, Dearborn, MI)

J. Acoust. Soc. Am. 109 (4) April 2001, pp 1304-1311

- Steve

 

[edjza80]

cheers steve, greg.

i read that paper quite some time ago for another reason, and didnt remember it covering acoustic radiation? me needs to go fid that one back out and have another look.

if the coeff is close to 1, id hate to think how loud these pipes would be if it were any worse!

 

[GregLocock]

Yup, it'd be like listening to the exhaust port. or worse.

couple o'hits

http://www.v-eight.com/tech_forum/viewtopic.php?f=9&t=62

Billzilla is solid, I don't know what he's on about this time.

and

http://www.profblairandassociates.com/pdfs/RET_Bellmouth_Sept.pdf

which is probably where you started.



Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[edjza80]

greg, thats my website (and my busted bellmouths) hehe

and i too have no idea what bill is talking about on that occasion.

im less concerned at the moment with the intake efficiency as thats pretty sorted, but by crikey it makes some noise. on a dyno you can hear the intake work through its harmonics and when it finally comes up to peak torque rpm and the planets all line up, jesus better stand back

just trying to figure a way to contain more of that noise in the pipe. if it had the additional effect of more energy in the intake well all good, but thats not the aim.

i saw one fellow who designs golfball sized spheres and mounts them in/above the throat of the bellmouth and claims that his solution, but it all sounds a bit washy.

 

[edjza80]

and for what its worth:

http://www.v-eight.com/multimedia/library/Wave reflections from duct terminations.pdf

from here:

http://www.v-eight.com/multimedia/library/

cheers
ed

 

[GregLocock]

Ah I thought you were trying to improve the pulse tuning of the intake, not make them quieter.

A square cut tube would be best. A reflector plate set back from the end of the tube would be neat, that'll give some cancellation.

Well, now I've got that paper, um, well which question doesn't it answer?

To make them quieter on production cars we use airboxes and helmholtz resonators and quarter wave pipes. They might increase your pumping losses by 0.5%.






Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[edjza80]

yeah, i last read that paper in 2006, so er, yeah, read it again last night

i guess the problem is we've got our intake reasonably optimised, and the bell design is a fixed factor we have to work with.

i guess an annular step in the bell could be worth experimenting with (use an insert in the existing bell to test it)

but whats not covered much are things like pulse plates, or that golfball idea, or other 'accessories' to work in conjunction with the pipe and its bell termination.

we run a carbon fibre airbox over the assembly so theres some possibility to dampen sound at that stage, but id like to think theres a way to keep the noise in the pipe. (and if that helps secondarily improve the intake efficiency, then all good)

 

[GregLocock]

It's a good paper.

Say you've got 6 cylinders and have the worst problem at 7200 rpm.

Then each stub is running at 3600 rpm, 60 Hz. So the wavelength is 5m. NOTHING you do on the scale of 10mm here or there will have the slightest effect on a wave as long as your car. A venturi would have to be half a metre long even for the wave to notice it.

In the paper a=0.025 (say) and k is 0.2, so H=ka=.005, so in all those graphs of reflection coefficient you are working on the extreme LHS, ie H=0. As you can see the reflection coefficient =1 in all cases at that end.

OEMs don't use things like golfballs because testing has always shown that increasing the frequency by combining the pulses, and then using quarter wave cancellation, and area transitions, is the most cost effective solution. Sometimes we use helmholtzs as well, but that's about it. The problem is that the wavelength is so long compared with the duct dimensions that interesting details only show up as blockages, not tuners.

The reflecting plane idea is a step in the right direction.

Your airbox should give you a 10 dB improvement - even without tuning. Make sure it is isolated from the engine, otherwise it is just a big loudspeaker. I'm not convinced that CF is the best material, designing airboxes is a bit of a suck it and see game. Softer plastics often works well, but you see all sorts of things used, an interesting question is how much effort it took to get them to work.

Thanks for posting that paper.




Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[SomptingGuy]

It is indeed a very good paper. I stumbled across it while searching our company library for something else, but was amazed by the range of termination types described and the detailed plots. Munjal describes end corrections for open and flanged terminations only and gives a single figure for each.


- Steve