Flywheel generator used for frequency regulation 2

[rockman7892]

Can someone explain to me how a flywheel generator is used for frequency regulation in a generator?

I read an article recently that said that flywheels could be used on a generator to establish better frequency regulation. The article explained that using a flywheel the generator would have a quicker response to standard control methods used to maintain frequency. What are these these standard control methods? The article seemed to indicate that the control methods involved the ramping up and down of the generator.

The article also discussed the possibility of the flywheel application to being a better alternative to another method which involved reserve storage in the form of a battery. Can anyone explain how a battery is used for this frequency regulation. As an interesting fact it said with the advent of the smart grid all the car batteries from rechargable cars plugged into the grid at any instance could help with frequency regulation.

 

[KiwiMace]

The flywheel adds rotational inertia to the generator, making it harder for changes in load to effect the rpm and thus the frequency.

The battery method is to rectify the AC to DC, storage in batteries, and then conversion back to AC. This way the variations in frequency and voltage in the input are isolated from the output.

Both methods add storage, or capacity to the system. Storage acts to dampen the variations in the system.

 

[electricpete]

Clearly the flywheel helps the transient response of the generator.

As an example, let’s say generator is islanded and putting out 100MW in steady state.
Then there is a step change in load to 150MW. Assume the generator governor is slow to respond and examine the behavior for a short period of time before the generator governor responds. Consider the perturbation introduced by the step change P = 50MW. Use symbols W = stored rotational enrgy, J = rotational inertia, w = radian speed, t = time since stpe change.

W = 0.5*J * w^2
P = dW/dt = J * w * dw/dt
dw/dt = P / (2*J*w)
Since w changes by only a small percentage, the rate of change of radian speed w is practically inversely proportional to J. The higher the J, the slower the speed changes. Thus the dip in frequency during the time before the governor responds is minimized by increasing J. Clearly transient can be more complicated if we try to model generator response, or if generator is connected to a system with other sources.

I’m not 100% clear of definition of "frequency regulation". I would think it is change in frequency from steady state to steady state. In that case the flywheel has no effect.... it is only important during the transient (example above).

=====================================
(2B)+(2B)' ?

 

[waross]

I think that Pete is saying that with a big flywheel, the inertia of the flywheel puts energy into the system when hit with a block load and so the frequency drop will not be as much or as fast as if there was no flywheel.
But.
The frequency may still drop by the droop percentage before the governor fully opens the throttle, and the return to set frequency will be slower as the flywheel must also be accelerated back to set speed. Isochronous operation will complicate the response somewhat but the frequency will still drop with the application of a load and recovery may be a little slower.
All in all probably a good thing for short-time block loads such as motor starting.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[KiwiMace]

The car thing would not work unless the 'smart grid' means that the average electric car user has a two way meter that you would probably only bother with if you had solar cells or other generation capability. This plugged-in storage is useless unless the power can flow back out to the grid.

 

[wayne440]

The KWH meter will pass power either way, some even let you pay for it both ways too.

 

[KiwiMace]

Ok, didn't realise that. The charging unit would need an inverter as well. Personally i'd be very unhappy if my nearby station tripped and my nissan leaf or whatever drained out to prop the grid.

 

[bacon4life]

Yes, you are correct about the standard method of frequency controls, in which governers and the prime mover response are regulated by physical time constants on the order of seconds.

Was the article refering to adding additional mass to the generator? Or to adding a separate energy storage flywheel to the system such as

http://www.beaconpower.com/products/smart-energy-25.asp

The response of an idependentant flywheel is regulated by the electrical time constant of the power electronics controling the flywheel, on the order of cycles instead of seconds.

One way for car batteries to help with load regulation is just by having remote control, and leaving them plugged in for a little longer than they need to be. Assuming a very smart grid and charger, an 8 hour charging period and a 10 hour plugged in period, the grid could plan on charging your car in 9 hours. If the frequency drops, the charger would shut off for up to an hour. If the frequency rises, the charger would charge faster than usual. Either way at the end of 10 hours your car is charged, the grid just got to choose the exact timing of charging. Note that it might be a few more years before the technology might be in place to try this out on more than theoretical scale.

The car battery could also be a two way system, but there would be additional issues, such as who pays for losses, the two way converter, and the loss of battery life caused by cycling.

Mark

 

[cranky108]

Why?
If the generator were isolated a flywheel would extend time period of the frequency deviation, while reducing it's magnitude. A battery and inverter system can also reduce the magnitued, and reduce the period of frequency deviation. However both methods has there limits.

On the grid, there is a combinition of slow and fast responce machines, with a varity mass constants, which gives a much better responce to frequency deviations, while also reducing the magnitude of the deviations.
So a better answer would be flywheel, and reserve process outputs would cost the least per MW.
The battery thing works well on smaller systems.

What's the range of an electric car? Or projected range? Will there be a charging station when you reach that range? And what do you intend to do for 10 hours (with the kids screeming "are we there yet")? The electric car I just don't see it being more than a "town car", with people at some time needing a "road car".

I guess if we only use electric cars, then the interstate highways will become much less crouded.

 

[waross]

Thia question and the responses seem to be mixing grid frequency and diesel generator frequency issues.
If we are going to base a control system on frequency it may be well to compare the normally expected deviation limits of grid frequency with common frequency deviations of a diesel generator.
A diesel generator frequency may be as much as 3% below nominal frequency, or 1.8 Hz. Block loading may cause greater transient deviations. A flywheel will reduce the transient deviation but you pay for that with a greater recovery period.
The alarm limit for grid frequency may be at 59.95 Hz (0.05 Hz deviation) or 0.083%
Normal deviation may be half of that.
This gives rise to some cjallenges.
1> Accurately measuring the grid frequency with such small deviations.
2> Inferring the reasons for the frequency deviations. Even though the frequency may be 0.001 Hz low, the operators may have other reasons than loading to allow the frequency to drop.
The ratio of normal frequency deviation of generators versus the grid may be over 20:1 and one answer may not be valid for both cases.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[FreddyNurk]

Baconlife's question regarding what 'type' of flywheel is significant. I read the question as talking about products by Piller, Powercorp and others whereby a flywheel is connected to a generation system via a bidirectional inverter, whereas the other interpretation is to modify what is already fitted to a diesel engine.

As I understand it, the inverter driven flywheel is quite good at compensating for transients seen by the generator, and can also have applications in short cycle peak lopping systems.

The obvious comparison with the battery system is the capable rate of discharge of the battery versus the slowing down of the flywheel.

 

[rockman7892]

Thanks for all the replies guys.

The article I was referencing was in the latest IEEE Spectrum magazine and was highlighting flywheel generators by Beacon.

Is the flywheel itself built into the generator itself or does it somehow externally contribute?

I'm assuming with the battery reserve that the batteries are connected to the grid through an inverter? What size battery bank and specifics are we talking about in this case? (voltage/power capacity etc....) Is the battery bank always connected and contributing power to the grid or does it only turn on in cases of frequency dips?

I'm also curious how fast the flywheel generator is to react to frequency transients? Are we talking seconds where there is a noticable dip in frequency with problems or are we talking milliseconds where the frequency dip is almost undetectable to equipment on the grid? How much faster is this reaction than the typical frequency control methods discussed?

 

[mgtrp]

For your reference, BESS is a very large battery bank assisting with maintaining frequency while additional generators are brought online in the event of a major generation loss in Fairbanks, Alaska.

http://www.gvea.com/about/bess/purpose.php

 

[bacon4life]

Hopefully this is the article you read?

http://spectrum.ieee.org/energy/the-smarter-grid/flywheels-keep-the-grid-in-tune

The flywheel is the motor and generator, there is no prime mover such as steam or diesel providing energy. When the flywheel needs to speed up, the it acts as a motor, to extract energy, it acts as a generator. Here is an example of one:

http://www.activepower.com/solutions/ups-systems/flywheel-technology/

There are also designs that have a large diameter flywheel made out of exotic high strength materials, and a much smaller diameter metal rotor for the motor/generator.

Flywheel systems can be designed for response times from fractions of a cycle to the 4 second time for the flywheel system in the article. The very fast response would be appropriate to a UPS system that might provide only 60 seconds of backup while a diesel generator starts.

The frequency transients on a large grid are usually much slower. During a fault the frequency takes at least a few cycles before generators go out of synchronism and can often be 10's of cycles before having synchronism problems. Maybe other can comment on the even slower speed of "steady state" frequency response of the grid. Thus a flywheel designed specifically for grid frequency regulation would not need to react as fast a UPS flywheel.

The full scale response time of the flywheel in the article is 4 seconds, which is definitely faster than
steam plants, and somewhat faster than hydro plants.

 

[bacon4life]

Per this NERC whitepaper, the frequency response of the USA grid has been declining in recent years. Now it makes more sense to me why there is a market for flywheels.

http://www.nerc.com/docs/standards/sar/Frequency_Response_White_Paper.pdf

 

[mgtrp]

While that might be true, when you consider you need to pick up over 3000 MW of load to effect a 0.1 Hz deviation in frequency (although I believe this figure is down around ~2500 MW these days), you'd need an extremely large flywheel to have any noticeable effect!!