carbon fiber compressor housing? 1

[ramblerman]

first post.
I'm working on a hare-brained idea. Not that it wouldn't be easier to find the most suitable part already made by somebody else, but I like making things myself, and I've got the time.
I've been doing some work with Kevlar/vinyl ester resin to make underhood components for my project car. Now I have a hankering to supercharge it, and I don't want it to look like anybody else's. I'm working on a variable speed drive for a centrifical compressor (I started with the compressor side of a turbo unit) but now I want to get the compressor right. Most modern turbos work at pretty high rpm's to get their pressure, and using a large truck compressor seems like overkill, what with the heft of industrial size parts. Assuming I find the compressor wheel whose boost characteristics I like, why can't I make the housing out of carbon fiber? The heat issues of a turbo unit are all but non-existent, especially if I cool the bearing/gear housing.
any feedback?

 

[patprimmer]

Are you thinking of using C/F for the compressor housing or the gear box housing.

Compressor housing sounds fine to me, but I doubt if a resin matrix composite will have sufficient dimensional stability over the operating temperature range likely to be encountered in use.

Check out the co-efficient of thermal expansion of potential materials vs that of aluminium. Be aware that plastics might not have a linear rate of expansion vs temperature throughout the required temperature range, so try to get a size vs temperature graph.

Plastics generally expand more than metal by about a factor of 10. The C/F will reduce it somewhat, but still no wear near that of metals

Regards

eng-tips, by professional engineers for professional engineers
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[ramblerman]

I'm talking composite just for the compressor housing. The gear box will be aluminum with oil-cooled ball bearings. My thoughts are along the line of a designated sump/cooler and using ATF. With the work I've done underhood to minimize heat, the majority of heat will just be from the 'charger's bearings and gears, not from external (exhaust, engine block) sources. I should be able to keep that below 150 degrees without problem.
The change in dimension of the compressor housing due to heat should be minimal.
the gearbox will be set up to reach a peak pressure early (~170Kpa @ 2000 rpm), then maintain that to redline (5200 rpm) at 2/3 throttle or more, but provide little boost at less than 1/3 throttle, with the boost going from no boost to full boost from 1/3 throttle to 2/3 throttle. My concern in change of dimensionality also includes the expansion of the compressor housing due to pressure, as well as the fatigue issue of the composite constuction due to pressure. I can find info about c/f fatigue due to tension/compression/torsional stress, but is there any info about c/f under pressure?

 

[patprimmer]

I am sure there is, and I might have some in an archive "somewhere".

Be sure the data you use is for a similar composite, and not data extrapolated from fibres, not taking the resin matrix into account.

Carbon fibres have exceptional strength and modulus, even compared to steel. I think they have good fatigue, but VERY LOW elongation at break, so load sharing between fibres can be a problem and this must be considered in the design.

Regards

eng-tips, by professional engineers for professional engineers
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[globi5]

Pressure load in this particular case is comparable to a tension load. However I don't see how this could be an issue though. Compared to the pressure (and other forces) a truck tire or even a bicycle tire has to deal with (800kPa), the pressure inside the compressor housing is low.
Also note that c/f doesn't work well under compression loads (which shouldn't be an issue either).

 

[globi5]

Since pressure load is comparable to a tension load and the modulus of elasticity of c/f is several times higher than that of aluminum it should expand less than an aluminum housing under pressure (assuming same material thickness).

 

[patprimmer]

C/F is quite good under compression, it is Aramid (commonly known by the DuPont trade mark of Kevlar) that is sensitive to compression loads.

Take care not to presume fibre modulus is the same as the finished composite modulus. It is always less, the degree varies with fibre content and fibre orientation as well as resin matrix characteristics.

It is a poor C/F based matrix that is not substantially better re modulus when compared to aluminium, but it can happen with poor design.

Regards

eng-tips, by professional engineers for professional engineers
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[thundair]

If temperature is an issue polyamides can raise you from 275F to >500F

Also look at using unistrand prepreg for hoop strength for the housing. 50 modulus

Regards

 

[SBBlue]

" the gearbox will be set up to reach a peak pressure early (~170Kpa @ 2000 rpm), then maintain that to redline (5200 rpm) at 2/3 throttle or more, but provide little boost at less than 1/3 throttle, with the boost going from no boost to full boost from 1/3 throttle to 2/3 throttle."

=============================

You might want to give to give your design parameters a little bit of thought.

Assuming that we're talking about sea level, the pressure in your intake manifold at wide open throttle will roughtly be one atmosphere, or lets say 101 KPa (actually a little less, but work with me here. . . .)

At one-third throttle your intake manifold pressure will be about one-third atmosphere (34 KPa) and at two-thirds throttle about two-thirds atmosphere (67 KPa). (Now remember that these are only "ballpark" figures.)

Now I'm assuming that when you say you want your "boost" to be 170 KPa you mean that you want the pressure in the intake manifold to be 170 KPa above the ambient pressure (which is the generally accepted definition of "boost"). So we're talking about an absolute intake manifold pressure of 101 KPa (ambient) + 170 KPa ("boost"), or 271 KPa.

Since you're operating at two-thirds throttle, the compressor will have to increase the pressure from about 67 KPa to 271 KPa, or about a 4:1 ratio. There may be turobcharger compressors that can do that, but I think that the absolute best compression ratio is only about 3:1 at present. Realistically, you will need sequential turbos and/or superchargers to achieve that type of compression.

But let's say that you can do find a Handy-Dandy Ace Turbocharger Compressor, how much would you expect the temperature of the gas to increase as you do a 4:1 compression?

This can be determined using the cold air standard tables. If you have a 100% efficient compressor that doesn't exist now, you looking at a compressor output temperature of 343 deg F. If you have a state of the art cutting edge turbocharger compressor that is 80% efficient, your outlet temperature is 395 deg F. A Roots supercharger type compressor will maybe only be 50% efficient; in that case, you are looking at an outlet gas temperature of 472 deg F. An after cooler might be a good design consideration. I'm not sure how those temperatures will work with carbon-fiber construction.

And one final observation -- why design the turbocharger so it is providing boost on partial throttle? You are wasting energy by increasing exhaust back pressure when you do that.

 

[GregLocock]

Nice one.

Also bear in mind that aramid has a negative coefficient of thermal expansion, so it can be used to match the composite's expansion to that of the aluminium, with a bit of thought.

Cheers

Greg Locock

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

 

[globi5]

Isn't throttle opening angle and manifold pressure very rpm dependent, since with increasing airflow the throttle resistance goes up exponentially?

(At half throttle and 2000 rpm you might almost have atmospheric pressure in the manifold.)

 

[ramblerman]

To SBBlue;
No, absolute pressure of 170Kpa, boost of about 70Kpa (`10 psi boost). Sorry, I think in absolute, since my driving regularly goes from sealevel (south carolina) to 11k (Rocky Mountan Natl Park, Colorado). Yes, the boost parameters are open for question (remember, I'm just benchracing this right now)
My engine control is by MegaSquirt, a diy open source engine management system, and electronic boost control is available. However, at this stage boost control can only be defined by one parameter, either rpm or %throttle. So, to globi5, that's one of the questions! I'm still playing with the choice of draw-through or blow-through throttle plate. The compressor will be close enough to the manifold that I really think I have a choice. Throttle response will be faster if I use a blow-through, but since I am designing out turbo lag by mechanically driving it, throttle step-in might be sudden. I might lose throttle response but gain drivability with a draw-through throttle, as well as minimizing power loss at part-throttle. However, I have to explore the effect that will have on the compressor.
To patprimmer, thundair and GregLocock;
Thanks for the materials hints! I appreciate that. Since this is (as near as I can find) new ground for composites. I keep thinking that VorTech or Paxton would pick up on this, and since they (or anybody else, for that matter) haven't, I was beginning to think that it was unworkable.

 

[MikeHalloran]

Certainly, making a mechanically driven centrifugal supercharger is possible; B29s had 'em.

Making a mechanically driven centrifugal supercharger durable is another matter. My friend Wayne spent WW2 rebuilding the gearboxes that drove 'em. Actually, it was more of a salvage operation than a rebuild; the highly loaded planetary speed increasers, with selective fit gears (different tooth thickness to control lash), tended to fail catastrophically rather than progressively.

As for the carbon- fiber scroll, you face a learning experience there, too. Sure, the fibers will take a lot of heat, but the available matrices that are heat resistant are also brittle, and not real strong, and not easy to work with. Oh, and not transparent. Maybe you could slap a carbon fiber/ epoxy 'applique' over a standard compressor housing and call it a day.

My own personal rule of thumb suggests that if a project includes more than one element that pushes the limit of what's possible, the project will eat up even a government's budget, and never work worth a crap. I.e., you can't win a war on two fronts.

Your mileage may vary; I hope it does.









Mike Halloran
Pembroke Pines, FL, USA

 

[ramblerman]

MikeHalloren,
Are you Scottish? Murphy would have loved you! Yes, I know the probability of failure is high. That's what makes this fun. Beating Murphy is what it's all about. The local landfill has its share of my failures. But today's gears and bearings aren't WW2 quality. Anyway, I'll let you all know how it works.

 

[MikeHalloran]

Scottish? What a terrible thing to say to an Irishman!

I'm American. My ancestors came from Ireland.

They probably didn't really come from there, either. In Gaelic, the name means 'stranger from across the sea'.

They don't come stranger.



Mike Halloran
Pembroke Pines, FL, USA

 

[ramblerman]

LOL!
Actually, your comment about WW2 aircraft engines jogged part of my (admittedly chaotic) train of thought about this project. A friend was an airplane mechanic in WW2 and is also a real steam fanatic. We have had long talks about history's engineering failures. Yeah, centrifugal blowers for aircraft had distinct limitations, and I would submit that, beyond the then-current bearing and gear technology, to put a system like that under constant demand is to beg for failure. Which is why I like the idea for the everyday auto engine, with its low demand/high demand, low rpm/high rpm cycles IF one can come up with an adequately efficient and reliable cvt. When I realized that the engine control system I'm working the bugs out of has the capability to control a boost system I started working on this. So my wanting a c/f compressor housing is a slight rabbit trail.
On the other hand, part of my cvt work is keeping as much heat as possible away from it, so as I look at the compressor housing I scavenged from a turbo unit I realized that an important component in its design had to do with dealing with heat external to the compressor itself. Next thing you know, I'm wondering why I need this big heavy chunk of heat-transmitting aluminum when I'm already working with carbon fiber. So, your point is well taken, and I know that the primary problem is the transmission. But I think I have that solved.

 

[GregLocock]

globi5, you are right, that is why EECs use MAP as their primary control. Throttle angle is a lousy way of detecting load, but it is useful as a feed-forward parameter.

So, when we talk about 1/3 load typically we do not mean the throttle is open by 30 degrees, we mean it is generating 1/3 of the max torque available at that rpm.





Cheers

Greg Locock

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

 

[MikeHalloran]

Ramblerman, I think you need to re-read what SBBlue wrote. The gas flowing through the compressor doesn't absorb heat from the underhood environment; the heat is generated internally to the gas, just by compressing it.



Mike Halloran
Pembroke Pines, FL, USA

 

[tbuelna]

ramblerman,

minimizing the clearance gap between the compressor wheel and the scroll housing is essential for high efficiency. Most centrifugal compressor wheels are aluminum, so utilizing a scroll housing of aluminum (with a similar CTE) makes sense. Unless your scroll (diffuser) housing is exceptionally large, the weight savings between graphite/epoxy or cast aluminum will be negligible. The cost, however, will not be.

As for turbocharger turbine housings, they are usually heavy wall Ni-Resist iron castings. The heavy wall thickness is necessary to contain any fragmentation failures of the turbine wheel at the high rotational speeds the turbocharger spool experiences.