Would this flywheel concept work?

[Webbd050]

Im designing a flywheel based kinetic energy recovery system for a car. The flywheel needs to have a very high angular velocity to store enough energy. At high angular velocities centrifugal forces will act to pull the flywheel apart, away from the centre of rotation. If i could somehow, maybe by use of tightened bolts going from the rim to the centre axis, apply a force so that when the flywheel is stationary its effectively being crushed. In theory these forces would opposed the centrifugal forces when the flywheel rotates thus allowing the flywheel to rotate at higher speeds without the use of stronger materials.

 

[Baluncore]

You could achieve a better result by making your flywheel, or just it's rim, from the same material as the bolts.

 

[berkeman]

Im designing a flywheel based kinetic energy recovery system for a car. The flywheel needs to have a very high angular velocity to store enough energy. At high angular velocities centrifugal forces will act to pull the flywheel apart, away from the centre of rotation. If i could somehow, maybe by use of tightened bolts going from the rim to the centre axis, apply a force so that when the flywheel is stationary its effectively being crushed. In theory these forces would opposed the centrifugal forces when the flywheel rotates thus allowing the flywheel to rotate at higher speeds without the use of stronger materials.

Another consideration might be precession. If you only have the single horizontal flywheel turning at very high energies, and you crest a hill, you could get enough sideways torque to roll the car over, it would seem. You may need to consider 2 counter-rotating flywheels to eliminate the precession problem.

http://en.wikipedia.org/wiki/Precession

.

 

[Webbd050]

Hi, i have already designed it with 2 counter rotating flywheels, was just wondering if the concept of manufacturing them with an inward force would work, but thanks anyway.

 

[Baluncore]

Final optimisation should begin only once a concept has been shown to work. Radial bolts will not solve as many problems as they will cause. You do not need that complexity this early in the project.

Your challenge now, is not to be distracted by optimisation of flywheel design, but to find efficient ways to couple energy between the flywheel and the road. My guess is that you will use magnetic coupling with the flywheels as the rotors of a motor/generator. That will strongly influence the materials and structure of the flywheel.

Later you might consider gaining some advantage by shrinking a high tensile rim onto the flywheel. Another way would be to wind a carbon fibre onto the outer surface of the flywheel. With a fixed tension in the fibre of say 1 kg and 1000 turns you could pre-stress the flywheel with a hoop stress of 1000 kg. But first get your energy coupling technology resolved and show that a flywheel has better energy storage than an electric battery for the same mass.

 

[Webbd050]

What problems would radial bolts cause? How would you make the carbon fibres shrink around the flywheel? I only have a basic workshop to build the prototype in hence why i was thinking the bolt idea would be easier and cheaper to produce than a wrap. Thanks.

 

[AlephZero]

What problems would radial bolts cause?

To make your flywheel store and release energy, somehow you have to apply a torque to the rim of the wheel to change its RPM.

Think why bike wheels don't have radial spokes. Or if you can't think why, build a bike wheel with radial spokes and see what happens to the back wheel when you push the pedals, or apply the brakes. (Note: I accept no liability for personal injuries if you try that experiment!).

How would you make the carbon fibres shrink around the flywheel?

Just remember that everything you see on the Internet (even on PF) is a practical idea, even if it might have some theoretical basis

 

[Baluncore]

What problems would radial bolts cause?

Bolts would make it hard to build and balance. The gap between the bolt and material provides access for moisture to circulate and corrode. A threaded bolt provides many weak points from which a crack can spread. The threads would strip if not tapered like the fir tree roots of turbine blades. Bolts converge on the centre where the concentration of threads weakens the structure. At the periphery they are far apart and have less effect. Bolts would need to be tightened to a symmetrical torque, or the flywheel would change shape and balance at speed. &etc.

How would you make the carbon fibres shrink around the flywheel?

If an elastic thread is stretched as it is wound onto a cylinder then the stretch remains in the thread and acts as an elastic hoop. I often use the technique for fabrication and repairs. For low tech hoops I use a cheap polyester thread wound under tension, then soak with super glue. For higher tech I use Kevlar fibre and soak with liquid epoxy.

Gas cylinders and spheres are sometimes wire wound. Take a look down this page to wire winding.
http://www.codecogs.com/library/eng...spheres/thin-walled-cylinders-and-spheres.php

Built-up gun barrels are made by shrinking a series of steel tubes onto the barrel.
http://en.wikipedia.org/wiki/Built-up_gun

 

[Webbd050]

Thanks mate that's really useful :). Don't suppose you have a rough figure for the force you can get from the kevlar shrink wrap do you? Is it difficult to get hold of kevlar fibres? Thanks

 

[AlephZero]

But before you rush off to wrap kevlar round everything,, remember that gas cylinders and gun barrels don't rotate like your flywheel. If somebody thinks a bolted assembly would be hard to balance, something wrapped with a layer of squidgy fibers doesn't bear thinking about.

(And I see the carbon fiber has mow morphed into polyester thread and kevlar...)

 

[Baluncore]

Your challenge now, is not to be distracted by optimisation of flywheel design, but to find efficient ways to couple energy between the flywheel and the road.

Webbd050, I should not have to repeat myself.

 

[cjl]

But before you rush off to wrap kevlar round everything,, remember that gas cylinders and gun barrels don't rotate like your flywheel. If somebody thinks a bolted assembly would be hard to balance, something wrapped with a layer of squidgy fibers doesn't bear thinking about.

(And I see the carbon fiber has mow morphed into polyester thread and kevlar...)

Composite flywheels (primarily carbon fiber) have been used - it's a known technology. It isn't cheap, but it can definitely be done, and done well.

 

[AlephZero]

Composite flywheels (primarily carbon fiber) have been used - it's a known technology. It isn't cheap, but it can definitely be done, and done well.

Sure, but they are not something you can cobble together in your "basic" home workshop.
http://www.esm.psu.edu/labs/cmtc/flywheel.html (click the thumbnail images to enlarge them).

 

[256bits]

But before you rush off to wrap kevlar round everything,, remember that gas cylinders and gun barrels don't rotate like your flywheel. If somebody thinks a bolted assembly would be hard to balance, something wrapped with a layer of squidgy fibers doesn't bear thinking about.

(And I see the carbon fiber has mow morphed into polyester thread and kevlar...)

In additon, gas cylinders vessels are modeled as thin walled pressure vessels, not like a flywheel which is more apt to qualify as a thick walled pressure vessel if of a solid material, where plastic deformation would occur on a radius closer to the centre.

 

[xxChrisxx]

I don't want to make it look like I'm disagreeing with you because I am not. You make very valid points, but...

Composite flywheels (primarily carbon fiber) have been used - it's a known technology. It isn't cheap, but it can definitely be done, and done well.

By a man, in a shed, with 'basic' facilities?

Op you need to prove your concept. If it won't work at low speed, then it certainly won't work at high speed. Which means steel is good enough for now.

We know you can ultimately make a carbon composite flywheel, as its already been done. See, flybrid and similar concepts.

When you research this some more, getting a strong enough flywheel is the least of your worries. If you want a high speed flywheel to stay spinning for any lengh of time, its needs to be sealed in a vacuum. By extension this means a high speed shaft, and thus major sealing issues.

 

[Webbd050]

Thanks for all your replies. I am not necessarily going to be building the most high tech ultra fast flywheel i can think of its just that its for my uni project and a material analysis is a good way to get more marks haha. So ill probably come up with a few concepts one with all the high tech shizzle and then the cheap and cheerful steel one that ill probably end up building a prototype of. Yes I've thought about putting it in a vacuum chamber too, i was thinking just have it in a completely sealed chamber with the power being transferred through the chamber wall via a magnetic clutch/ gear but I'm not sure if id be able to do that. I know Ricardo have built a prototype that use that method. Can you build your own direct drive magnetic gear with just some of those neodymium magnets aligned properly? I know to change the ratio with a magnetic gear you also need a rotating cagey kind of thing with steel bars but would you need that for just a direct drive? Thanks

 

[Webbd050]

And sorry baluncore just getting ideas for down the road a bit :P i have thought about cvts aswell.

 

[cjl]

I don't want to make it look like I'm disagreeing with you because I am not. You make very valid points, but...



By a man, in a shed, with 'basic' facilities?

Op you need to prove your concept. If it won't work at low speed, then it certainly won't work at high speed. Which means steel is good enough for now.

Sure, though a man in a shed with basic facilities won't get one that is anywhere near its full potential (and realistically, for something you can make at home, I doubt you could make a composite flywheel that would match what you could easily do with steel and a lathe - it's fairly easy to make homemade composites, but it's difficult to get them anywhere near their theoretical strength).

We know you can ultimately make a carbon composite flywheel, as its already been done. See, flybrid and similar concepts.

When you research this some more, getting a strong enough flywheel is the least of your worries. If you want a high speed flywheel to stay spinning for any lengh of time, its needs to be sealed in a vacuum. By extension this means a high speed shaft, and thus major sealing issues.

This is also a major problem. Also many of the easy to work with resins for making composites aren't really made for vacuum use. That doesn't mean they can't be used, but they'd be less than ideal. Realistically, as I mentioned above, a steel flywheel is a better choice for any sort of homemade application. That still doesn't solve the drag/vacuum-sealing problem...

 

[Webbd050]

Can anyone think of something that i could buy off the shelf that i could use as the flywheel?

 

[Baluncore]

Can anyone think of something that i could buy off the shelf that i could use as the flywheel?

A truck wrecker... The flywheel from an old diesel engine would be a good start. It would be balanced as a unit.
It comes down to how you intend to couple energy between the flywheel and the outside world.

WARNING. A centrifuge or flywheel should be operated inside a containment vessel capable of retaining the fragments following a burst.

How much weight do you need?
You have bearing issues to resolve.
You have stability and resonance issues to overcome.
I presume you have seen; http://en.wikipedia.org/wiki/Flywheel_energy_storage

For safe speeds of cast iron wheels, see below.
The limitation on flywheel RPM comes down to material, construction and peripheral speed.

Cast-iron Wheels with Solid Rims: Cast-iron wheels having solid rims burst at a rim speed of 395 feet per second, corresponding to a centrifugal tension of about 15,600 pounds per square inch.

And on the same page;

Wheel having Tie-rods: Tests were made on a band wheel having joints inside the rim, midway between the arms, and in all respects like others of this design previously tested, except that tie-rods were used to connect the joints with the hub. This wheel burst at a speed of 225 feet per second, showing an increase of strength of from 30 to 40 per cent over similar wheels without the tie-rods.

So you can see that a solid wheel is stronger than a wheel having radial tie rods.
This answers the OP question.

Safe speeds will be about half the the actual bursting speed.

 

[nsaspook]

WARNING. A centrifuge or flywheel should be operated inside a containment vessel capable of retaining the fragments following a burst.

This is a no joke, here is a picture from a maglev tubopump that had a rotor hub crack at about 25,000 rpm. That thick skin is not just for vacuum containment.

 

[Webbd050]

My current design for the flywheel which gives the required performance increase of the vehicle has a 100mm outside diameter and 70mm inside diameter rim and spins at 60,000rpm. I've done FEA tests on it and according to that it should withstand the forces. The system is only intended for a 200kg 35bhp motorbike engined car.

With regard to bearings, i was thinking magnetic bearings on the central axle to support the weight if the flywheel with ordinary bearings to the sides to stabilise it, which i think is how the flybrid and ricardo systems do it but haven't looked into bearings in much detail.

There is very limited space available on the car so it needs to be small.

And yes I'm aware of the danger of bursting and the need for a suitable containment vessel. The university has an engine test cell where i can test it safely aswell.

 

[AlephZero]

My current design for the flywheel which gives the required performance increase of the vehicle has a 100mm outside diameter and 70mm inside diameter rim and spins at 60,000rpm.

Considering the rim speed would be about Mach 0.9, you would need to run it in a vacuum, otherwise you will waste a lot of energy through windage.

And your rim acceleration is about 200,000 G. That's more like an ultracentrifuge than a flywheel. Any design guidelines based on a materials handbook and cast iron are completely irrelevant IMO.

 

[Baluncore]

60,000 RPM = 1,000 revs / sec. 100 mm diameter = 12” circumference.
1,000 * 12” = 12,000”/sec = 1000 ft/sec.
So I must assume you will be using high tensile steel, not cast iron.
You could get a short length of HT 4” hydraulic rod and turn it to length in a lathe.

But how can you possibly couple 60,000 RPM to the outside world.
It would need to be an induction motor armature because magnets would fly apart.
In order to generate power, an induction motor armature needs to drive an existing AC supply or a resonant load.

 

[Webbd050]

That flywheel design was made out of titanium as my course director said he had a titanium billet somewhere i could use but obviously as its lighter than steel it has to spin faster to store the same amount of energy so it would probably be easier to use steel at a lower speed wouldn't it?

Could i not have gear ratio of say 3:1 between the flywheel and magnetic coupling inside of the actual casing so its 20,000rpm the magnets have to transfer? Would that be more manageable?

The ricardo system is 60,000 rpm and transfers it using magnets. http://www.ricardo.com/en-GB/News--...nd-generation-high-speed-flywheel-technology/

And the flybrid system is slightly over 60,000 rpm and just uses a seal around the driveshaft to maintain the vacuum and transfers it via a torodial cvt. http://www.flybridsystems.com/F1System.html

I have access to SAE technical reports at uni which explain them in much more detail and there doesn't seem to be anything exceedingly complex to the coupling system unless I am missing something.

 

[AlephZero]

I don't know what happened to that Ricardo project, but googling KInergy and Kinestore (and the lack of hits) is a good clue.

Most people don't have any intuitive idea about the dynamics of rotating systems with high centripetal accelerations, unless they build up some experience working with an existing engineering team, or survive the failures of their initial mistakes. To make an obvious point, at 200,000G, an "insignificant" local design feature or manufacturing defect with a mass of 1 gram has literally the same consequence as a 500-pound gorilla trying to rip your device apart...