It's happening, the power grid is changing.

[DenzilPenberthy]

An interesting thing here I stumbled across yesterday:  The National Grid in the UK are inviting tenders for projects which can supply between 1 and 50 MW into the grid in sub 1s response times.  It is interesting to see this becoming a thing now as we reduce the amount of large steam driven rotating machinery for generation.

I wonder if this reduction in system inertia is also from the demand side? i.e. fewer directly connected large synchronous motors, instead most stuff nowadays is inverter driven so presumably wouldn't contribute to the general rotating mass of the system in the same way?

http://www2.nationalgrid.com/Enhanced-Frequency-Response.aspx

Copied and pasted:

Background
The changing generation mix in GB is reducing the contribution that synchronous generation makes
to the energy market, which in turn is reducing the level of system inertia, particularly on low demand
days. Lower system inertia affects the ability of the System Operator (SO) to manage the system
frequency within normal operating limits. This in turn will drive the procurement of larger volumes of
the existing frequency response products. NGET is therefore investigating alternative solutions, one
of which is the creation of an enhanced frequency response service. The aim of this service will be to
improve management of system frequency pre-fault, i.e. to maintain the system frequency closer to
50Hz under normal operation, however performance in post-fault conditions is also of interest.
NGET believes that the most appropriate way to develop this new service is to run a tender for four
year contracts for enhanced frequency response capability. This will give NGET experience in
integrating enhanced frequency response into the suite of system management options, as well as
providing a competitive route to market for new technologies. Whilst a number of the energy storage
solutions currently available have the ability to provide frequency response in these enhanced
timescales, NGET welcomes the participation of any technology types which can meet the required
parameters. Provision of additional services such as reactive power will also be taken into
consideration.
Ultimately, it is NGET’s aim to create a sustainable tendered market for the provision of enhanced
frequency response services from all technologies, work which is being progressed through the

SMART Frequency Control project1
. This would ensure that there is a route to market for both
existing and new providers, regardless of technology type, and also would realise the benefits of
competition in the procurement of these services.

[tggzzz]

Oh, that is interesting.

The sub-second response time means that hydro won't be sufficient; IIRC that takes 20s to reach full power.

Is there any indication how long this has supply power? If only 20s, then I suspect flywheels or compressed gas might work.

[DenzilPenberthy]

"Our initial analysis indicated that a battery with a 45 minute duration (minimum operational state of
charge to maximum operational state of charge) based on current frequency data and possibly one
that minimizes or at least meets the 95% availability. However, this is currently under review, and
we will be giving more guidance on the minimum duration required shortly. The maximum duration
that we would value from an EFR service is 30 minutes, any ability of the assets to deliver energy
beyond this timeframe has no additional value to NGET, however it is up to Applicants to size the
energy capacity of their assets (where necessary)."

So maximum 30 mins. They've stated battery here but elsewhere in the documentation they are very keen to stress they are completely agnostic about what technology is used.  In fact I read just now that they've received tenders so far from 68 projects, 60 of which have qualified to tender of 1.3GW in total, of which 880MW is battery based. It doesn't say what other ideas people have come up with.

 As you've stated, they already have frequency response arrangements in place with things like pumped hydro and diesels that can pick up in high 10's of seconds to low number minutes.

[DenzilPenberthy]

It is also interesting to think that once there is a decent install base of electric vehicles, this service could be provided by a vehicle fleet operator if the chargers were bi-directional and say 10% of the battery charge was made available to use.

[coppice]

So maximum 30 mins. They've stated battery here but elsewhere in the documentation they are very keen to stress they are completely agnostic about what technology is used.  In fact I read just now that they've received tenders so far from 68 projects, 60 of which have qualified to tender of 1.3GW in total, of which 880MW is battery based. It doesn't say what other ideas people have come up with.

Battery based might mean nothing more than having a minute or two of battery capacity while they get something else on line. UPSes are often described as battery based when they have a few minutes of battery, and a diesel generating set.

[DenzilPenberthy]

Battery based might mean nothing more than having a minute or two of battery capacity while they get something else on line.

Good point, I hadn't thought of that. 

Another interesting thing to note is that these projects are required to be symmetrical, i.e. act as a power sink in the event of an over frequency event

[Delta]

This is very interesting indeed!

To the shed!  Time to dig out some old car batteries, flywheels and hydraulic cylinders!  O0

[stmdude]

As you've stated, they already have frequency response arrangements in place with things like pumped hydro and diesels that can pick up in high 10's of seconds to low number minutes.

I've seen a diesel start within 1/50th of a second for emergency power on a cell-tower.  Explosive charges in the cylinder(s) will make them start in a real hurry.

Probably doesn't scale with the size of diesels needed for more than a couple of KW.

[station240]

I'd imagine battery and super capacitor based systems would do this easily enough.
Perhaps also flywheel based generators.

[Delta]

I don't think the technological task is particularly difficult.  For example
A series string of around 1000 12v lead acid batteries to give a DC bus voltage above the 15.5kV P-P voltage of the three phase 11kV MV net, being continuously float charged from the grid, whilst trickling a little bit through an inverter back to the grid.  That inverter could be cranked up to max chat in a few cycles!

Making it economically plausible (ie profitable) is the challenge.

I'm off to price up 12v LA batteries... 

[Someone]

I'd imagine battery and super capacitor based systems would do this easily enough.
Perhaps also flywheel based generators.

Microinverters coupled to super capacitors would be a very interesting way to deliver on this tender, while still utilising the inverters during normal operations. A controller that maintains a minimum charge in the capacitor and dumps excess power to the grid, then on demand can draw down on the capacitor storage. If consumers could access the variable market pricing there would be plenty of people doing this sort of thing already, but the current level of investment needed to access the market is a huge problem.

[johansen]

what i'm concerned about is the wet dream of using everyone's hybrid car battery as a free energy sink/source.

There's no way you are going to get paid fair market value for the inevitable KWH per dollar cost of a battery (lead acid is something like 30 cents per kwh, minimum, under ideal conditions and faithful maintenance, 20 percent depth of discharge, etc) lithium ion is higher; allegedly lithium iron phosphate has reached parity with lead acid.. that is if it can really can deliver 3000 charge cycles at 80 percent depth of discharge, and at the end of 3000 charge cycles, the amp hour capacity is still 80% of new. (anyone have any data on this?)


also why does anyone need a power grid that keeps the frequency within .1Hz?

[MrSlack]

Sticking the crazy Victorian lunatic hat on for a second, screw supercapacitors and battery banks. What I want is giant several hundred-ton flywheels all across the country. Think of them as massive capacitors except with AC in and AC out as you can have AC drive and standard generators. Sync them to spin slightly above a multiple of 50Hz and then when there is a grid demand switch them in at 50.5Hz and switch them back out when the energy spike is handled and the other plants have come online at 49.5Hz. Plus the linear decrease in frequency as energy is used means they can easily be used to sync phase with a control loop. They can initially have energy injected via solar, gas, steam, diesel, tread wheels (only joking) or from a quiet spot on the grid.

Now I'm going to find that elephant I was going to electrocute while wildly firing my blunderbusts into the air and observing my pocket watch through the fluff of my neckbeard.

[tggzzz]

What I want is giant several hundred-ton flywheels all across the country. Think of them as massive capacitors except with AC in and AC out as you can have AC drive and standard generators.

In the 70s/80s my father briefly surveyed energy storage for the CEGB/CERL. His focus was, IIRC, on medium-scale storage to see if there were alternatives to Dinorwig which was just coming online. Flywheels would not have been able to store "interesting" amounts of energy, but maybe the definition of "interesting" has changed - hence my question about how long any new scheme would have to store power.

One of the problems with flywheels is their failure mechanisms, particularly what happens when bearings disintegrate - something that is more likely to happen if you start extracting power within a second. If bearings disintegrate then the resulting damage will not be confined to the site, since the flywheel will carve a path across the countryside as its power is dissipated. Such behaviour is deprecated by safety inspectors!

ISTR seeing a picture the result when that happened to a turbine - there was a neat hole in the building's wall, just like in old cartoons!

[MrSlack]

Interesting stuff. The failure mode was what was interesting to me ... I could imagine a 500 tonne steel cylinder steamrolling a small village!

More sensibly I envisaged a horizontal disk dug deep into the ground. Although that may just result in an earthquake machine

[BillyD]

Interesting stuff. The failure mode was what was interesting to me ... I could imagine a 500 tonne steel cylinder steamrolling a small village!

More sensibly I envisaged a horizontal disk dug deep into the ground. Although that may just result in an earthquake machine

I wonder how much consideration would you have to give to aligning the wheel with the axis of rotation of the earth. A horizontally spinning disk will attempt to invert itself by 360 degrees once per day as the earth rotates, so there would be strain on the bearings to counteract this. Not to mention slowing down the earths rotation, oh the humanity....

[MrSlack]

I think I might actually be an evil James Bond villain in the making. One beeeeeelion dollars or I will stop the Earth spinning and move to my space station on the moon.

[MrSlack]

What about lots of smaller flywheels. Thousands of them?

[Kilrah]

Then the cost of multiplying the auxiliary equipment likely exceeds the savings from the smaller flywheels...

Envoyé de mon SM-G930F en utilisant Tapatalk

[DenzilPenberthy]

I was wondering last night about flywheels too.  I don't think even compressed gas or diesel engines with explosive charges in (I'll have to look that up, not heard of this before) would be any good. I can't see any way of spinning up an alternator of 1 to 50 MW size to synchronous speed in sub 500ms without destroying it.  Unless it was kept spinning all the time which would be very energy and maintenance inefficient.

I wonder if hydrogen fuel cells would be able to deliver power in that sort of response time? I've heard about some pilot projects generating H2 from excess wind energy and feeding it into the gas grid. Maybe store a bit of it and feed it into some shipping container sized fuel cells?

As I was thinking about flywheels I was picturing a landscape like the days of town gas here in the UK when 'gasometers' were a common sight in every town. Will we see large structures like this dotted about the place filled with spinning concrete blocks?
https://en.wikipedia.org/wiki/Gas_holder

Also as regards the need to maintain the system at precisely 50.000Hz - There are reasons for this but the real reason is that the frequency is driven by the balance of supply and demand. If there is excess power being fed in to the grid, everything speeds up and the system frequency increases. Conversely if there is more demand than supply, the frequency drops as the (traditionally synchronous) generators bog down and slow under the increased load.  If the frequency falls too far from 50Hz, the generators which are still supplying power will start to trip out as they are overloaded which then puts more load on the remaining ones and can cause a cascading failure of the whole grid.

[coppice]

I've seen a diesel start within 1/50th of a second for emergency power on a cell-tower.  Explosive charges in the cylinder(s) will make them start in a real hurry.

Starting piston engines with explosive charges is nothing new, but I've never heard of them starting, running up to speed, stabilising, and being ready to go on line in anything like 1/50th of a second. You can't even get a high output from a piston engine until it has warmed through.

[Delta]

I've worked at sea on Class 3 Dynamically Positioned vessels, where standby engines (4.3MW each) were kept pre-heated to 60C, with an external pre-lube pump running all the time.  With compressed air starters, these would be on line and on load in less than 10 seconds.  Not quite what's needed here, but impressive non the less.   (It didn't stop the idiot DPOs from wanting to run every available engine all the time with 15% load on the poor coking up blighters though)

[Someone]

I've seen a diesel start within 1/50th of a second for emergency power on a cell-tower.  Explosive charges in the cylinder(s) will make them start in a real hurry.

Starting piston engines with explosive charges is nothing new, but I've never heard of them starting, running up to speed, stabilising, and being ready to go on line in anything like 1/50th of a second. You can't even get a high output from a piston engine until it has warmed through.

Diesel backup has been around a long long time, and is well proven for response within seconds:
https://en.wikipedia.org/wiki/Diesel_rotary_uninterruptible_power_supply
In general a fast starting Diesel backup scales up to MW sizes and can be assisted in starting with compressed air. They form the essential safety systems of power stations.

[tggzzz]

In general a fast starting Diesel backup scales up to MW sizes and can be assisted in starting with compressed air. They form the essential safety systems of power stations.

In the UK the power source of last resort, i.e. "black start" if the entire network is down, is pumped hydro.

[Spikee]

have a giant elevated reservoir and pump water in it when having excess electricity supply. Quite cheap and gravity is pretty fast.