Flywheel Energy Storage

[svetz]

Flywheel Energy Storage

Flywheels with magnetic bearings are 97% efficient, have an 85% round trip efficiency, are not adversely affected by temperature, have high C-Rates, zero degradation (do not degrade over time based on DoD or C-Rate), unlimited cycling, are made of inert/benign materials, the SoC can be precisely determined via rotational speed, are sealed devices that need no/little maintenance, and have long lifespans.

So, why don’t we see more of these? You can find them on Alibaba, but at three times the $/kWh as LiFePO4. So, why are they so expensive?

There’s the wobble issue: Left to its own, a spinning flywheel would slowly precess following the Earth's rotation. But that’s overcome by the magnetic bearings. There’s shock and vibration, the flywheel needs precision balancing; all these add up increasing the capital costs.

They’re also heavy… Amber Kinetics offers an 8kW/32 kWh flywheel that weighs in at 5 tons.

Back in 2013 the Velkess flywheel was supposed to be a drop-in replacement for 48V lead acid and cost around $1300/kWh. Their secret sauce to reducing costs was a flexible flywheel [video] that didn’t require precision and so would be cost effective.

So it seems like a great technology if someone can figure out how to reduce the manufacturing costs.

 

[John Frum]

A former colleague had worked at a station with a flywheel.
The wheel came off its mounts one day.
I forget his exact words but "holy hell" comes to mind.

 

[svetz]

I noticed Amber Kinetics buried their flywheels in the installation materials, as I saw it I was wondering how much earth it would take to absorb the instantaneous discharge of 32 kWh. Turns out it's about 28 kg of TNT.... less than I imagined. They have cute little hats too, makes me wonder if they have ballistic gel in them:

 

[RickP]

It‘s a shame we humans rarely prepare for an emergency until the emergency overtakes us. If the same effort was put into renewable, sustainable energy that goes into the next cell phone, television, or fashion trend, we’d already be there. The quote attributed to Ben Franklin “Justice will not be served until those unaffected are as outraged as those who are” certainly applies. As long as unlimited power just ”appears“ at the wall socket, people won’t value innovation in producing it in better ways.

 

[svetz]

The flexible flywheel is kind of interesting – the idea is the inner hub is flexible so it can dampen vibrations so the flywheel can be less perfect; leading to lower costs.

So that got me to thinking… Carbon fiber is five times stronger than steel, twice as stiff, and 20% the weight, and doesn’t fatigue like steel (also very brittle in comparison so don’t hit it), can be molded to vastly reduce production costs and prices have been falling as more uses have been found for it.

Imagine a hollow carbon fiber doughnut as the rotor. Light weight, easy to ship, strong. Once on site, you fill the doughnut with jelly (okay, maybe it’s just a ton of water or perhaps grease would do).

The mass of the rotor (i.e., the jelly) would be viscous, so any vibrational energy would be converted into heat (e.g., ultimate in flexible). Add baffles inside the interior of the rotor to prevent the jelly from rotating independently. The carbon fiber has little mass so would absorb/impart very little vibrational energy despite it’s stiffness. Also, manufacturing defects in the carbon fiber would have less impact since it has less mass.

So if you’re an ME looking for an idea for your thesis… well, let me know how it works out!

 

[svetz]

Turns out there are a lot of studies on partially filled cavities and rotation, mostly having to do with fuel in satellites using spin for stabilization. If there's a cavity (e.g., spent fuel), then there are unstable regions at certain spin rates as show to the right (from Computational Structural Mechanics& Fluid Dynamics Advances and Trends). It's also a function as to how many compartments (baffles) are in place. I saw a paper for a filled rotor, still trying to get my hands on it.

​

 

[Supervstech]

My father told me about when the USAF tried using flywheel batteries in mobile situations... he said it was kinda comical watching the grunts with one on their back try to turn corners ... the gyroscopic effect was “hilarious”...

 

[svetz]

Found a paper entitled "Vibration and Control of Viscous Liquid-Filled Rotor System Supported by Active Magnetic Bearings", in a language I can't read... but it seems clear from the diagram that there are some RPMs that have a harmonic (?) vibration. Not sure why. Currents in the baffled section from speed changes? So need a comparison with a solid "pefect" rotor.

​

 

[WoodsieLord]

I was hoping to see a thread devoted to flywheels. The energy density is clearly not as good as chemical batteries but I like them for the (let me use triple quotes here) " " " simplicity " " ". I know. Bearings, vacuum, and probably the same issues with lower speeds as super capacitors as batteries (slow rotating speeds would probably yield lower voltages, harder to harness/collect)
Yet, I feel like improvising a flywheel battery at home seems less difficult than making your own chemical batteries. Unless your name is Robert Murray Smith, that is...!
I've got to admit that I spend more time fantasizing my own flywheel than what I would like to acknowledge! Great thread
Regards!

edit: I can't resist. I've got to add that the thing I was thinking lately is the rotation of the earth vs the gyroscopic effect... what if the flywheel is shaped like a sphere with a structure allowing it to change it's axis freely ? Wouldn't that be beneficial ? Or the earth rotating thing does not bother that much as the bearings?

 

[svetz]

... what if the flywheel is shaped like a sphere with a structure allowing it to change it's axis freely ? Wouldn't that be beneficial ? Or the earth rotating thing does not bother that much as the bearings?...

I don't know that the flywheel being a sphere would make a difference or not; but you're right that freeing it on 3 axis (e.g., like a gyroscope mount) would resolve any pressure from precession (say that three times fast). Of course, how much of a force is that anyway?

 

[BiduleOhm]

Shape don't make a difference on the gyroscopic effect, but it's important for the energy storage part (you want the maximum mass on the outside to maximize Wh/kg)

The carbon fiber + liquid idea as another added advantage to being a lot more safe in case of failure

 

[DerekH]

The flexible flywheel is kind of interesting – the idea is the inner hub is flexible so it can dampen vibrations so the flywheel can be less perfect; leading to lower costs.

So that got me to thinking… Carbon fiber is five times stronger than steel, twice as stiff, and 20% the weight, and doesn’t fatigue like steel (also very brittle in comparison so don’t hit it), can be molded to vastly reduce production costs and prices have been falling as more uses have been found for it.

Imagine a hollow carbon fiber doughnut as the rotor. Light weight, easy to ship, strong. Once on site, you fill the doughnut with jelly (okay, maybe it’s just a ton of water or perhaps grease would do).

The mass of the rotor (i.e., the jelly) would be viscous, so any vibrational energy would be converted into heat (e.g., ultimate in flexible). Add baffles inside the interior of the rotor to prevent the jelly from rotating independently. The carbon fiber has little mass so would absorb/impart very little vibrational energy despite it’s stiffness. Also, manufacturing defects in the carbon fiber would have less impact since it has less mass.

So if you’re an ME looking for an idea for your thesis… well, let me know how it works out!

Carbon fiber has great tensile strength but is notoriously bad in shear and torsion. The peak power of that design would probably have to be painfully low to avoid over-torquing it and destroying the rotor.

 

[svetz]

Carbon fiber ... is notoriously bad in ... torsion. The peak power of that design would probably have to be painfully low...

The torsional strength is a really great point! Thanks for bringing that up.

Just to elaborate on your point about peak power, what I believe you're saying is that the torsional limit of fiber carbon would in effect set the equivalent of the maximum charge/discharge rate, because it's the limit as to how fast you can spin up or down the flywheel, i.e., change the momentum. Nothing to do with the top rotational speed.

I wonder if there's a geometric layout to the fibers that would improve that for a spinning flywheel? For example a clockwise spiral sandwiched with a counter-clockwise spiral layer?

 

[DerekH]

The torsional strength is a really great point! Thanks for bringing that up.

Just to elaborate on your point about peak power, what I believe you're saying is that the torsional limit of fiber carbon would in effect set the equivalent of the maximum charge/discharge rate, because it's the limit as to how fast you can spin up or down the flywheel, i.e., change the momentum. Nothing to do with the top rotational speed.

I wonder if there's a geometric layout to the fibers that would improve that for a spinning flywheel? For example a clockwise spiral sandwiched with a counter-clockwise spiral layer?

Yeah, exactly. Power input/output is directly related to acceleration/deceleration of the flywheel which is directly related to torque.

Max rotational speed defines your energy storage capacity and would also be limited by the shear characteristics of the carbon fiber which applies the centripetal force that keeps the jelly from becoming a sprinkler ?

Did a bit of casual reading on it just now out of interest and it seems like the fiber geometry relative to load and the resin material make a big difference, but too tired to do a deep dive on it atm. Very interesting topic though, like your thinking on it!

 

[svetz]

... Or the earth rotating thing does not bother that much as the bearings?...

I was curious too how much of a force that was... found this on Wikipedia:

Flywheel energy storage systems using mechanical bearings can lose 20% to 50% of their energy in two hours.[17] Much of the friction responsible for this energy loss results from the flywheel changing orientation due to the rotation of the earth .... This change in orientation is resisted by the gyroscopic forces exerted by the flywheel's angular momentum, thus exerting a force against the mechanical bearings. This force increases friction. This can be avoided by aligning the flywheel's axis of rotation parallel to that of the earth's axis of rotation.

I suspect minimized is better than avoided.

 

[svetz]

...Shape ... it's important for the energy storage part ...

So what is important for the energy storage part?

From the Engineer's Toolbox, the energy in a flywheel is Ef = 1/2 I ω^2, where ω is the angular velocity if radian/s and I is the moment of inertia.

For a rigid body, I = k m r^2. K is a geometry shape factor (see below).

What this tells us is that of the design elements mass is only a linear player. For example:

Increase by 10x    Affect on output power
           mass       10x
            rpm    3,944x                  [6.28 radians in one revolution, so (10x6.28)^2]
         radius      100x

The final factor is k, and the farther the mass is from the center the more power it can store. That is a bicycle tire shaped flywheel of the same speed /mass/radius holds nearly twice as much as a flywheel with a constant thickness.

Inertial constants of some common types of flywheels

• wheel loaded at rim like a bicycle tire - k =1
• flat solid disk of uniform thickness - k = 0.606
• flat disk with center hole - k = ~0.3
• solid sphere - k = 2/5
• thin rim - k = 0.5
• radial rod - k = 1/3
• circular brush - k = 1/3
• thin-walled hollow sphere - k = 2/3
• thin rectangular rod - k = 1/2

 

[BiduleOhm]

I suspect minimized is better than avoided.

Well if you align it correctly you can go to the point it's so small that it's negligible, but that means a weird angle relative to the floor if you're not at the equator or poles so not the easiest thing to do when we're talking very heavy things.

 

[svetz]

So, any math doing is worth being really crazy about. Take a deep breath and buckle up folks!

Prowseville Example
Imagine a ring around our hypothetical city of tomorrow whose job is to store energy for the city. Assumptions:

• Prowseville has the same energy needs as Los Angeles, about ~72,000 MWh/day
• Prowseville is roughly circular and has a diameter similar to LA, about 40 miles (64 km)
• the flywheel is a continuous train on a magnetic track encircling the city, so k=1.
• Power exchange occurs as the "cars" have magnets and the stations have coils, so the cars can be sped up or slowed down.
• The cars are cylindrical, made of graphite, filled with rock and measure 20' long by 2' diameter, they travel in a tunnel
• The tunnel has a vacuum
• need an even number of power-balanced counter rotating rings to reduce geological stress
• 85% round trip efficiency, 95% efficient

How many rings would it take to provide a day of power?

Math

• Need 72 GWh, at 95% efficient storage should have 76 GWh
• Each car has a volume of hπr^2, 20 x 3.14 x x 1^2 = 62 cuft
• We'll fill the cars with the excavated rock at 168 lb/cuft, so a mass of 10,416 lbs, or 4,735 kg.
• With a 40 mile diameter, the circumference is 2πr = 125 miles, 663,168 ft, 202,133 m or up to 33,158 cars
• The radius is 20 miles, or 32,186.9 m
• 33,158 cars for a rotor mass of 34,5377,894 lbs or 156,989,952 kg
• Let's say the speed is 40 mph, 64.4 kph, or 18 m/s
• I = k m r^2, 1 x 156,989,952 x 32186.9^2 = 1.6263195046e17 kg m2
• ω = 18m/s / 202,133 m/rev * 2π radians/rev = 2 radian/hr or 2/ 3600 r/s
• Ef = 1/2 I ω^2, 0.5 x 1.6263195046e17 x (2/3600)^2 = 25,431,134,170 Joules = 7 MWh storage at 40 mph.

Two rings (to counterbalance) would be 14 MWh... So, that's no where near enough power for a day.

This really illustrates the small effect mass has in comparison to increasing the rotational speed:

   speed                                     Car Diameter, MWh
 mph    m/s             w         2'        3'         4'         6'         8'
  20     9    0.000277638        1.8        4.0        7.0       15.9       28.2
  40    18    0.000555275        7.0       15.9       28.2       63.4      112.8
  60    27    0.000832913       15.9       35.7       63.4      142.7      253.7
 100    45    0.001388188       44.0       99.1      176.2      396.4      704.7
 150    67    0.002082282       99.1      223.0      396.4      891.9    1,585.6
 200    89    0.002776376      176.2      396.4      704.7    1,585.6    2,818.8
 300   134    0.004164565      396.4      891.9    1,585.6    3,567.6    6,342.4
 600   268    0.008329129    1,585.6    3,567.6    6,342.4   14,270.4   25,369.6
1000   447    0.013881882    4,404.4    9,910.0   17,617.8   39,640.0   70,471.2

From the excel table above, it would take two tracks of 6' diameter cars at 1000 mph to provide 1 day storage for Prowseville.

 

[BiduleOhm]

And at 1000 mph each 6' car has about 270 T (~ 600000 lbs) of centrifugal force you need to counteract...

 

[svetz]

I was actually surprised it was so high, my first "intuitive" guess was 40 mph, so really really far off. Could just be a conversion error too ;-)

Update: How fast is 1000 mph anyway?
A 1m radius wheel where the outside was moving 447 m/s (1000 mph) would be how many RPM?
Let's see... Circumference is 2πr = 6.28 m.... 447 m/s ÷ 6.28 m/rev = 71 rps, 71 rps x 60 s/m = 4270 rpm.
A two pole AC motor operating at 60 Hz will always run at approximately 3600 rpm, so that's not much faster. Still, easy to see why the rotors end up costing so much.