Intake Design Paper

[Sam Zimmerman]

Hey guys,

I am giving a presentation for a combustion institute conference and have a rough draft of a paper for their proceedings. I am only allowed 6 pages so I can't fit everything but would appreciate feedback on the paper. I have never seen a post like this before here but what the hell. Let me know anything you disagree with.

Paper is here (http://www.webpages.uidaho.edu/~racing/05FFL-188.pdf)

[Mike Claffey]

Howdy, had a quick glance through,

just after the intro

"How his engine power output influenced by the location of resonances and anti-resonances across the speed range?"

Should this read "How is engine power..."?

Regards

Mike

UWA Motorsport
www.motorsport.uwa.edu.au (http://www.motorsport.uwa.edu.au)

[Chris Boyden]

Great job with the paper. The only thing more that I would like to see is stuff that you are still working on....like changes in cross sectional area from runner to plenum, restrictor to plenum and how that affects impedance and torque, etc.

When we did our flow testing, I found pretty interesting that the restrictor flowed 7 to 10 cfm more air @28" h2o by itself than bolted to the plenum w/ runners due to expansion and the cross sectional area change. However, this
is a far cry from the work you guys have done.

good luck with the presentation.

"How his engine power output influenced"
probably already caught this little mistake...but here it is anyway

[nathan s]

Sounds like money to me. I don't think you could really add much without making the paper hard to follow.

One thing I did notice is that you didn't mention what wave number you were using for your different runners. If I am not mistaken, your impedence is effected by how far your wave has gone through the pipe.

Otherwise, it looks good. Good luck with the presentation.

BTW...Is there any chance I could score a copy of notes from that conference? It sounds interesting.

[jonnycowboy]

Hi,
your paper looks great and also looks to be a new future reference for teams looking to tune their intakes!!
I'm currently designing the intake for the concordia car, also a cbr600f4i.
I used another formula:
RUnner length=(((720-CAMDURATION)*0.25*2550)/(RPM*ReflectiveValue))-(0.5*DIA)
where reflective value is either 1st (strongest, 2nd or 3rd.
I decided to go for 1st wave reflective which gives me a length of 28 inches, which is fine (we have the room). However I don't see how you could calculate the runner length from the impedance formulas you gave, perhaps if you could explain this a bit better here it might be more clear for us mathematical weaklings!
thanks!
-Jon

[Sam Zimmerman]

There is not a calculation per se for runner length in my paper, but an interpretation of the graphs produced. The example shown in the paper shows that where the two strongest resonances occurred were also where the two torque peaks occur. The resonances are the local minimums on the impedance vs. engine speed graphs. The downside to this method is that it involves interpretation rather than a direct calculation. The good part is that it takes the entire system into account and the entire engine speed range.

If I were designing an intake for a diesel generator that would only run at a single speed, other methods are far simpler. My argument is that for an engine that redlines at 12K or more, the simpler methods may get you in trouble by causing large anti-resonances at undesirable engine speeds (i.e. the torque curve shown in the paper).

Having said all that, your method jonny is a more precise variation of the Platner method shown in the paper, which worked great for Chrysler with their high performance intakes. I am just trying to find a solution that will offer a better torque curve prediction for high revving engines such as ours.

If I had time, this method could also be used to predict the phase angles in the pressure waves so one could precisely match the valve timing with the pressure waves in the intake and exhaust but that will have to be left for someone else's master's thesis.

[John Bucknell]

Good job Sam.

The multiple resonances is something which not many people catch on to, and you show directly that it is there and has many peaks. Hopefully everyone who needs a quick education on manifolding takes a look at this.

[B Lewis @ PE Engine Management]

Hi Sam,

The paper does a good job of describing the manifold as a system including the subject of anti-resonance (which very often is completely overlooked). We have quite a few SAE teams ask us for information on manifold design throughout the year. Would you mind if I reference your paper in the future?

[Sam Zimmerman]

I don't mind Brian, but I would rather you wait for the final version in about a month or so. http://fsae.com/groupee_common/emoticons/icon_wink.gif

[B Lewis @ PE Engine Management]

Thanks Sam, I will wait. Will you post a new link when you are finished?

[Sam Zimmerman]

Brian,

I lengthened the paper to include more on the derivation and the acoustical testing portions. It is in SAE format now. As always, I would still like to hear any feedback on the theory, test methods, conclusions, where anybody disagrees, etc.

Feel free to share this with anyone who is curious.

Paper is here. (http://www.webpages.uidaho.edu/~racing/05FFL-188.pdf)

[Chris Boyden]

Thanks for posting that. It's awesome.

[nathan s]

Sam you are my hero. Everything looks great.

[Engine Dude]

Sam or anyone,

In the Engelmen method,

What is the R value in the engine speed determination.

Also, when you made your plots of impedance vs. engine speed, when you calculate the impedance at each point of the intake system, what is the k and j in

Ze=Z/jtan(kL)

Cheers, p.s my maths is terrible

[Sam Zimmerman]

R is the compression ratio. The effective volume is V*(R+1)/(R-1). I had never noticed that I didn't define "R".

j is the square root of -1.

k is the wave number (w/c).

Sorry for the confusion. Notice that the impedance equation you quoted is only good to find the impedance at the entrance to a pipe (or cylinder) that is closed on the opposite end. Equation 27 is used for sections that are open on both ends. Start with your piston (or closed valve) and work your way all the way back to the inlet using a combination of equations 27, 28, and 29.

[Engine Dude]

Thanks Sam,

Ill give it a go.

Brilliant work!!

Regards

Sam
UniSA Racing
(South Australia)

[Engine Dude]

Sam,

When you say Z is a function of k which is a function of frequency or engine speed, does this mean that your wave number is (engine speed (rpm) / c) ?

Also when calculating your speed of sound in the pipes, how did you work out the isentropic bulf modulus?

In your paper you say beta = (p/�')*�o
How did you find the acoustic pressure and density for the system? Or is there some value you assumed?

Thanks again Sam

From Sam Parry.
UniSA

[LSU Dave]

Wow, that paper has some methodology in it that I have never seen or even considered...

On that note, I'm beginning the design of our 2006 intake, and I can't seem to find a good book on intake/plenum design. I know a few were recommended to me, but I have since forgotten.

This thread seems like the knowledgable place to go for that kind of information. So on that note, what does everyone recommend.


---------------------

Check the labels on figure 10... two open valve labels.

"Figure 12 shows a strong resonance from at 5000 RPM"

"The eleven inch runner
configuration does not show the expected anti-
resonance that the ten and eleven inch configurations produced"

The rest looks good... can't wait to model this for myself. Bravo.

[nathan s]

There are a lot of whitepapers on the SAE site. Plus there is Design and Simulation of Four Stroke Engines, Gordon P. Blair. This is a really good book, even though I only grasp a small portion of the knowlege in it http://fsae.com/groupee_common/emoticons/icon_frown.gif. Also, Winterbone has some good books on intake design.

[buddy]

Another interesting engine simulator... and some intake simulation papers.

http://www.lesoft.co.uk

buddy

[Engine Dude]

Sam,

When calculating the characteristic impedance, Z=po.c/S

Is c = w / k

Where k is the wave number and w is the speed of the engine in rpm?


In Winterbone it says wave number is 2*pi divided by wavelength.

then say k = (2*Pi*f) / a

If wavelength = a/f

Where f is the frequency.


How can you approximate the frequency of the plane waves travelling in the pipe sections?

Cheers
Sam
UniSA

[jd74914]

I know I'm dragging this thread back to life, but does anyone have a copy of Sam Zimmerman's paper that they could send to me? I would really like to read it.

Thanks.

James (jd74914 AT gmail.com)
UCONN Racing

[B Lewis @ PE Engine Management]

Hi James,

I think I have a copy. I will check and pass it along if I do.

[Hot Rod JayRad]

If you find that paper, could you send it to me as well? thanks much!

jniemiec at berkeley dot edu

[Chris Aho]

I'll take one too please http://fsae.com/groupee_common/emoticons/icon_smile.gif

caho at wpi dot edu

[Krister]

I would like a copy too.

Krister.Olsson at vok.lth.se


Krister

[jaydenanders]

could you send one of those my way at snowman_ja@hotmail.com

Thanks!

[Michael Palaszynski]

Zimmerman has contributed greatly to this forum. he's got some great insight!

could i get a copy as well!

thanks,

Mike Pal

[jpusb]

I'd like a copy too please.

juanp.fsaeusb@gmail.com

[Horace]

hlai02@students.poly.edu

por favor mi amigo

[samphlett]

I know this is an old thread, but I'd be interested in seeing that paper if anyone still has a copy. The original link is dead.

[Michael Palaszynski]

ooops i definitely forgot my email in my delirious state...

mikepal@umd.edu

[Krautsalat]

here you can download the paper ....

intake design paper (http://www.datenklo.net/dl-1biDx)

Krautsalat

[Michael Palaszynski]

thanks!

[Michael Hart]

Thanks!

[vmc]

If there is a floating copy of this I would love to get one as well...the above links server has been down...

Thanks

Never mind the server is back up and running...

[redzuki]

Very interesting work. Enjoyable read. I have noticed one point that has not been covered. Engelman H.W. in his 1973 ASME paper "Design of a Tuned Inlet Manifold" states that:

"this analog is valid only if the intake periods have little or no overlap"

Firstly: what is little?
Secondly: does he mean intake to intake overlap or inlet and exhaust overlap? I would presume the latter.

Kind regards,

alex

[HumbleJC]

Has anyone got this link to Sam's paper to work? I keep getting forwarded to some foregn page. If you have the paper could someone please email it to me?!
email is jordan.cachiaras@mnsu.edu Thanks

[G-Dan]

I would appreciate a copy too :

gabriel.pare-olivier.1 -at- ulaval.ca

[samphlett]

I'm surprised there isn't some (moderated) space on this site for members to upload things like this (engineering tips, pics, papers, etc).

[Dan G]

The paper is copyright of SAE, so unfortunately I can't post it up for public consumption. But your team can get it from SAE using this info:

http://www.sae.org/technical/papers/2005-01-3803

05FFL-188
Development and Validation of an Impedance Transfer Model for High Speed Engines
Sam Zimmerman, Dan Cordon, Michael Anderson, and Steven Beyerlein
Mechanical Engineering, University of Idaho
Copyright © 2005 SAE International

[Taylor Oelke]

could i please get a copy of this paper as well? oelke@usc.edu

[nickerss]

Originally posted by Taylor Oelke:
could i please get a copy of this paper as well? oelke@usc.edu

Did you even read the post directly above yours?? Just buy the damn thing. It's 11 bucks.

[Mikey Antonakakis]

We just bought it, and I read it, and I like it. I didn't absorb everything though, so I might have to read a few more times.

[Lucy]

Hi, I've been looking for the Engelman ASME paper "Design of a Tuned Intake Manifold" and I can't find it, not even in the ASME digital store, if someone please could tell me where to find it, I would appreciate it.

[Geoffct]

Your faculty adviser has a certain number of FREE papers from SAE each year, so you don't even need to spend the $11.

[EPMAl]

I've been trying to use Zimmerman's impedance calculation method to validate my intake design. I'm actually using CFD but since I'm using my design in a school project, my teacher wants me to calculate some characteristics about my intake, so I thought I would plot some graphics of the anti-resonance and resonance frequencies. I'm trying to validate my calculation on the geometry that is explained in the paper, but it doesn't seem to work. I'm using k=2*pi*RPM/(c*60). When calculating Z, I took rho as being constant but I not sure I can assume that though. I'm calculating the impedance starting from the cylinder (which is the only pipe closed at one end) and going all the way back to ambient air (runner than plenum than inlet pipe). I plotted the results using the log of the norm of my impedance. My results still doesn't correspond to what is shown in the paper. I was also wondering when you have a convergent or divergent section in your intake system (maybe like, let say, a venturi) can you use the mean pipe section area in the calculation or is there another way to approximate it? Can someone help?

[Sam Zimmerman]

Hey guys, I haven't been on this site forever. I'm glad to see some intake discussions. When I was around nothing would kill a thread faster than technical talk about intakes.

I receive requests every semester for help on my paper. After several attempts at trying to get students to poke holes in my paper and publish their own, I've resigned myself to the fact that all past FSAE students who contacted me were looking for quick turn-key programs that will work for them. I understand the schedule forces that.

Given that, I would suggest that you do some basic calculations, prototype an intake, use a speaker and mic as shown in my paper, and come to a solution rather quickly. I know that judges used to post here, I don't know if they still do but I would like to hear their 2 cents on what I am about to say.

Simple research, basic equations, and then planned prototype and design verification testing is a valid path towards a good prototype solution. If you want to really understand the acoustics, plan on a master's degree.

I'll finish with a question. Does anybody here know of a more recent SAE paper that discusses the impedance method in my paper?

[Sam Zimmerman]

Originally posted by Lucy:
Hi, I've been looking for the Engelman ASME paper "Design of a Tuned Intake Manifold" and I can't find it, not even in the ASME digital store, if someone please could tell me where to find it, I would appreciate it.

Lucy,

I got my copy through inner-library loan, I believe. You might also have to search under "Design of a Turned Intake Manifold" since Engelmann mis-spelled the title on his paper. http://fsae.com/groupee_common/emoticons/icon_smile.gif

[samphlett]

Sam,

Imagine you could cash in in your fame. A pint for every nugget?

[EPMAl]

Actually Sam, I was wondering if your methodology was taking into acount interference between cylinders or you calculated your system frequency on the assumption that the three other valves were closed during the whole cycle, thus not having pressure waves from the other cylinders interfering with the wave you're actually computing (from what I understand, you only used one cylinder as the wave source). I'm new to acoustics but to me, it seems we could have a more precise model if it was time-based which mean we could compute the interference effect between cylinders over an engine cycle (maybe using the acoustic pressure and velocity equations you presented in your work). After my post, I got your model running but still have some differences with your results.
http://i402.photobucket.com/albums/pp109/Motoal/acoustic_impedance.jpg
The curves for the runner-plenum interface for close and open valves look good, but they still are offset which means the only thing that could move my curves on the x-axis would be my equivalence between frequency and engine speed. I'm using w=2*pi*(2*N/60); because i figured that the frequency had to be twice the engine speed divided by 60 (twice because the wave travels twice to go from the valve and back to the valve and divided by 60 to transfer rpm to revolution per seconds). I can't really figure out what goes wrong here. Also, does someone has an opinion on using mean cross sectional area of convergent or divergent tubes to approximate the speed the waves travels throught them. It just seems logic to me. Don't worry Sam, for the time I spent working on this, I'm really not looking for a quick turn-key. Right now, it just became personal between me and acoustics, haha.

[EPMAl]

Ok, sorry, I just saw, when I posted it that I didn't translate the figure. Those are the two curves at the runner-plenum interface with open valve (in blue) and closed valve (in greenish blue).

[Sam Zimmerman]

Originally posted by EPMAl:
Actually Sam, I was wondering if your methodology was taking into acount interference between cylinders or you calculated your system frequency on the assumption that the three other valves were closed during the whole cycle, thus not having pressure waves from the other cylinders interfering with the wave you're actually computing (from what I understand, you only used one cylinder as the wave source). I'm new to acoustics but to me, it seems we could have a more precise model if it was time-based which mean we could compute the interference effect between cylinders over an engine cycle (maybe using the acoustic pressure and velocity equations you presented in your work).

Nice work. I did not take into account the interference, but it would be a great next step. The only thing that I would caution on is trying to get too exact with the acoustics model and not moving on to the prototyping and testing. If the changes you are making to the model don't shift the graph by more than a couple hundred RPM, you might end up finding that you are picking the fly shit out of the pepper. Give it a try though, I never did and I would love to know the results.


Also, does someone has an opinion on using mean cross sectional area of convergent or divergent tubes to approximate the speed the waves travels throught them. It just seems logic to me. Don't worry Sam, for the time I spent working on this, I'm really not looking for a quick turn-key. Right now, it just became personal between me and acoustics, haha.

You have definitely thought this through. For the changing cross-section tubes you will have a continually changing impedance. I simply modeled this with a do loop, essentially modeling the tube with the changing cross-section as 'n' tubes in series, each with a length of n/L. Be careful to only add the effective length addition to the ends and not to each step, if that makes sense.

I hope I was understanding your questions correctly.

Good luck, that will be one pint. I like porters.

[speedracer_007]

the webpage link doesnt have the paper any more if anyone having could forward it to me at my email id i would really appreciate it
speedracer_sid@yahoo.co.in
Thanks

[EPMAl]

You'll have to buy it from sae.org . It costs about 12-13 USD. You can than download it right from the website.

[Gaanja]

Hi sam,

the paper is a great.

I have a few questions

1) The cam lift during the intake suction stroke changes the area available for the air to fill the cylinder. The changing area actually affects the impedance of the entire system. Is this taken care of by assuming the piston position to be at half stroke and the cam lift to be at average lift position calculated over the cyle?

2) is there a separate way of analysing the impedances of an intake that has a continuous area change? say one shaped like a flower vase? or do we do a finite element approach by dividing them into separate tubes of very small thickness along the cross section and then sum them up?

3)Is it worth the time to look at the effect of the position of the butterfly valve on the system at a given rpm?

[williams13nr]

I know this is from over a year ago, so sorry to bring it back but I can not find an answer to my problem.

I am using the impedance modeling approach to our intake as many teams have done in the past, and came across the same problem that EMPAI had back in Feb of 09 above. I can not reproduce the exact graphs that Sam produced in his papers, using all the values he used. The plots look close for the engine, but the frequencies do not coincide, and are off by ~2-3000 RPM. I tried to model the simple resonator as he does first in the paper, and got the values to be right on. I am also using w=2*pi(2*N/60) for the frequency, which I believe is what Sam used (found on another thread about this paper). I am confused because there is no difference in the calculation for the runner with cylinder open and the resonator other than the frequency, yet one matches and the other does not.

Has anyone been able to successfully reproduce Sam's exact plots for the impedance given the different runner lengths? Any help would be greatly appreciated.