Battery storage leaves fossil fuels and regulators in state of inertia

Print Friendly, PDF & Email

The brain cells are working overtime at the headquarters of network owners, grid operators, generators, and regulators. Australia’s electricity grid is about to make the leap from analogue to digital, and everyone is scrambling to keep up.

Ever since the opening of the Tesla big battery next to the Hornsdale wind farm in South Australia last month, it is as though a new era has dawned for the management of Australia’s electricity supply.

This is not just about flexibility, smoothing out renewables, or responding to peak demand and speed of response – it is also about grid security and grid stability. And it is causing a massive re-think.

Here’s why:

There is now no doubt that wind and solar will provide the cheapest form of electricity supply, in Australia, and most other countries, now and into the future.

The first line of defense against this new reality for the defenders of incumbent interests is to ask what happens when the “wind don’t blow, and the sun don’t shine” – i.e. how to meet the peaks.

The most obvious retort is that the existing fossil fuel system has a phenomenal amount of in-built redundancy to deal with exactly these peaks – think of all the “peaking” gas generators used for just a few hours a year – and to deal with major outages, such as the near 20 trips of major coal and gas generators recorded so far this summer.

Studies by the likes of the CSIRO, the network owners, and the Finkel-commissioned review of battery storage, note that much less storage is needed than thought, if any at all, until the share of renewables heads towards 50 per cent.

The falling cost of storage, and the flurry of projects in small and large-scale batteries, and the numerous pumped hydro schemes being considered in nearly all mainland states, should answer the storage question.

The second line of defense for the fossil fuel industry is to point to the need for “synchronous” generation and system inertia. But just like the Maginot line deployed by then French in WWII, the incumbents are being outflanked by new technology and new ways of thinking.

This is not as minor as you might think. You will find references to “inertia” and “synchronous” generation littered throughout the documents produced by rule makers defending the status quo and the incumbents whose future depends on it.

It’s a fierce topic of debate – equivalent to the discussions that accompanied that shift from analogue to digital, from land lines to mobiles, from petrol cars to EVs. Just how much inertia is really needed? Not as much as you think, say some.

(Have a look at these past stories  – Gas turbines vs battery storage: S.A. in state of inertia over energy future; 

and Inertia in power system: We don’t actually need that much

A new study from Ireland, which like South Australia vies for global leadership in the amount of variable renewable energy supplies in its grid, shows that the conventional view of “inertia” may soon become redundant.

The study called “Batteries: Beyond the spin” (a reference to spinning turbines that traditionally provide that inertia), hails the dawn of a new era of digital inertia.

Principally, it says that 3,000MW of gas generators – the main providers of conventional inertia in Ireland – can be replaced by one-tenth of the capacity in batteries. It is time to go digital, it says, and change the rules accordingly.

Such reports are causing experts to have a rethink about grid security, and to recognise that while we may need “synchronous” generation to provide “real” inertia, possibly the only reason we need “real” inertia is to support the synchronous generation.

It is a nexus that begs to be broken, and because inverter-based technologies like wind and solar and battery storage operate in different ways – as different as digital to analogue – it’s got a lot of people thinking.

(Of course, in Australia, there will always likely be some spinning turbines, even if only from pumped hydro plants or even spinning thermal. But that’s not much comfort to coal and gas generators trying to figure out their future).

The study, by consultancy Everoze and the Queen’s University Belfast, with contributions from battery storage company AES and UK’s National Grid, promotes exactly this idea of “digital inertia”, and the lessons are as applicable to Australia, another island grid, as they are for Ireland.

The study finds that battery storage can arrest the fall in frequency and the rate of change. In short, even though they do not provide spinning mass, what the study calls a “digital inertia” response provides the same benefits – or greater – as inertia.

The study is well worth reading. It’s technical, but important, and provides a fascinating insight into how technology can change the game, and why regulators need to think differently so they can change the rules accordingly.

“It’s time to stop framing ancillary services around the incumbent technology, and create a genuinely level playing field,” the report says.

In a study on recent faults in the Irish grid, it found that 360MW of batteries could have provided the same amount of power, after 0.1 seconds, as the inertial response of 3000MW of synchronous generators.

“Whilst the system operators are right to explore a combination of options for managing RoCoF (rate of change of frequency), the QUB study demonstrates how batteries can fully replace the power and energy delivered by existing inertial response,” it says.

The experience in Australia with the Tesla big battery is already proving similar. What the Tesla battery has demonstrated is that while it will never beat fossil fuel generators on their immediate inertial response, it is substantially quicker to do whatever is required next.

While “ synthetic inertia” or “digital inertia” are not instantaneous, they are still very fast (and controllable), unlike synchronous units which are not controlled for about the first second of any event as they just contribute their inertia.

Australian experts say the issue in Australia is that the response times with governors (on thermal power plants) have been allowed to get longer, an issue that is causing concern for the market operator. Digital technology is considered to be precise and fast.

By unlocking the potential of digital inertia, the QUB study says it is possible to refine operational constraints, use existing gas plants more efficiently, and retire or mothball less efficient plants, and save significant costs.

But it needs a new approach to rule-making, a problem that many in the industry are facing in Australia.


  • john

    The sooner more large battery systems are put all over the grid and the sooner more PHES is put all over the grid the sooner the wholesale cost of power will plummet.
    Along with that solar and wind with huge amount of PHES to take up the slack when the wind does not blow or the sun does not shine the better.

    In fact with enough of the above there is no need for any other type of generation.
    Yes a fairly high cost of implementation but no underlying ongoing cost of usage this is the big point about this kind of energy utilization.

    • Peter F

      It is arguable that not even a lot of PHES is required for grid stability. Modern low wind turbines i.e those with large (therefore high inertia) rotors supplied with synthetic inertia controls can contribute about 5-6 times as much Fast Frequency Response energy to the grid as gas turbines of the same capacity. In addition Francis turbines in existing hydro can be run dry to provide more inertia. So a system with 4 GW of Francis turbines and 35 GW of wind will have far better short term frequency stability than a system with a mix of 35 GW of gas and steam plants

      • john

        Yes an i am hearing you but they need to be distributed all over the grid .

  • Cooma Doug

    When the coal fans start talking their doomsday stuff we should refer them to this artical.
    I think its getting close to the kind of thing that could be spoken in parliament. Add a few of the cost numbers we will have in a couple of years and the dumbest senator will listen.

    • A Blot

      Only trouble is when you start showing the coal lovers facts and figures that conflict with their beliefs they stick their fingers in their ears and respond with Blah blah…..

      • phred01

        The market will fix the coal lobby

        • John Norris

          Unless the coal lobby fixes the market…

          Edit: I think/believe/hope you’re right!

    • neroden

      “article”, FYI

  • Cooma Doug

    One thing that is very important is the predictability of renewables.
    Our ability to predict weather conditions accurately over long and short times. This will be another asset emerging in the market.
    I am only just starting to recognise it. We used to look at the numbers 20 years ago. If there was a 660mw coal gen in service we needed 660 MW extra in the synchronised capacity.
    All good but we never knew when or if we would need it. Then when it happens suddenly we need to make another security reset, again with no certainty.

    Wind and solar will be predictable to such an extent, the wind reduction and sun reduction periods will be bid onto the market with certainty. Rather than being a large sudden shock loss like coal and gas, it will be a market alligned load shift.
    Solar wind and storage working together, creating certainty and opportunities that will become obvious as they increase in market participation.
    I can see the transport industry having a synergy with solar and wind, increasing their value on the market significantly.

    • Peter F

      It is even trickier than that. Imagine you had two coal generators in service each capable of 660 MW of output splitting a 660 MW load between them so you had an effective 660 MW spinning reserve. If one generator tripped the system would still crash because the second coal generator could not take over the load quickly enough. Steam plants have an Inertia constant up to 9 and then few seconds with of excess steam in the steam chest so the system may struggle on for 10 seconds or so before under-frequency relays shut everything down.

      Our old system worked because the inertia of 20-35,000 MW of generators was connected at any one time but between them there was at least 660 MW (or whatever the largest unit was generating) of unused generation capacity that could ramp up relatively quickly. After about 30 seconds some hydro would start to spin up and perhaps save the day

  • Carl Raymond S

    A grid with no spinning inertia? Can’t be done.
    Why that would be like riding a train without a ticket, or reading a newspaper without paper, or driving a car without petrol. Preposterous!

    • Peter F

      or having a telephone without wires

      • Carl Raymond S

        Or a federal government without corruption. Oh wait, that actually hasn’t happened yet.

        • nakedChimp

          Just design a currency that doesn’t gravitate to the biggest private piles all the time.. something like a demurrage on liquid assets (cash, savings accounts, etc.) would be enough to do this..

    • technerdx6000

      Can’t tell if sarcasm or not

      • Carl Raymond S
        • Pixilico

          Once upon a time, newstands used to be full of newspapers and magazines…Cheers, mate!

          • Carl Raymond S

            And ticket windows had long queues. You had to allow 5 mins to buy one, or risk watching your train pull out. Today’s ticket window is tomorrow’s coal mine.

          • neroden

            To be fair, I do think we will always have paper tickets as a sort of emergency backup.

          • Carl Raymond S

            The emergency backup is ‘free till normal operations resume’. In 5 years of travel on Opal card I have not experienced that.

        • baseload renewables

          Here I was thinking it was about the added electrical load from the fridges by keeping an automatic sliding door open…but maybe it was warranted by the weather at the time.

          • Carl Raymond S

            It was hot. I’ve let go to getting agitated by every air conditioned shop with an open front door.
            I save it for people waiting in parked cars with a/c on, poisoning the passers by (and themselves).

  • Peter F

    Even that chart is kind to thermal power plants. In practice no known survivable grid event would lead to a rate of change of frequency (ROCF) of more than 1 Hz per second. A typical gas turbine has what is known as an inertia constant of about 3. i.e. the total spinning inertia is about 3 seconds worth of power at rated output so a 200 MW gas turbine has about 600 MWs.
    The energy in the rotor is Jw2 so in the first second if the frequency falls to 49 Hz ( an extremely dangerous condition) the energy given up will be 600 x(1- (49/50)^2) = 24 MWs the power will be 24 MW so best case it will be 12% of rated power for the first second.
    However if the ROCF relays are set at 0.3 Hz/s and frequency fall at that rate then the system will already be shedding load. If the frequency falls by 0.3 Hz/ second, still a fairly major event, then the inertia is worth 600 x(1- (49.7/50)^2) = 7.1 MWs or 3.5% of rated power. In the meantime the 20 MW battery could have supplied 16.5MWs

    • Carl Raymond S

      Thanks. Most enlightening.

  • Peter F

    There is a parallel in CNC machine tools. When I started designing them many years ago the idea was that the rotary inertia of the motor multiplied by the square of the reduction ratio between motor and load has to be at least 3 times that of the load.

    Now we build much more accurate and faster responding machines with load inertia 5-10 times that of the motor, (i.e. the required system inertia has fallen by 15-30 times) because the control systems are orders of magnitude faster and the amplifier currents like the batteries, can respond in 1,000th of the time they used to.

    Power systems are a lot more complicated but similar changes are possible

    • Island fisher

      Dont think there any politicians in Canberra who would be able comprehend this (LNP side anyway)

      • Pixilico

        Although thi$ they probably would.

    • Phil

      The rate (speed) of the change,The amount of the change,The duration of the change

      3 separate changes that need to be designed for
      And the technology to do that is far better today than ever to correct that

  • phred01

    I c a lot of stranded assets

    • Nick Kemp

      So do the banks which is one reason it is scaring coal and gas lovers, Even if the Libs want coal mines and power stations someone has to fund them

  • Pixilico

    Glad you came up with it, Giles. It makes the case for a renewables-only grid much more compelling than it already is:

  • neroden

    Even a single small hydro plant has enough spinning inertia to handle the <0.2 second time before the batteries kick in at full power. So does a solar thermal plant. So does a wind turbine.

    In the unlikely event that you literally have no hydroelectricity, no wind turbines, and no solar thermal on your entire grid — you're running a pure photovoltaic grid — you can just put in a loose spinning turbine (not a power plant, basically a sort of flywheel) and have spinning inertia.

    "Spinning inertia" is a red herring; we probably need none, and if we need any, the amount is tiny — a rounding error.

    • JonathanMaddox

      Most wind turbines are not synchronous but inverter-excited and are (at present) configured not to keep any potential generation in reserve. They will neither slow down nor increase power output in response to grid faults.

      The “loose spinning turbine” you mention is called a synchronous condenser. I have no idea why there aren’t more of them in places like Ireland and South Australia.

      • baseload renewables

        Is it possible to retrofit wind turbines with the hardware and software that can make this happen? Or is this possible on new builds only?

        • JonathanMaddox

          It’s a good question. I’m sure some turbines already have the ability to reduce output when instructed, and bring it back up to the maximum possible given the conditions on command, but the time lag would likely be on the order of tens of seconds, or whole minutes, rather than the mere microseconds required for “spinning reserve”.

  • JonathanMaddox

    I’m surprised there’s not one mention of synchronous condensers in the whole article. They are rotating generator units, unpowered but connected to the grid, spun up by grid energy and providing literal synchronous spinning inertia (and nothing else). They may have a heavy flywheel attached, or not. They may be purpose-built or they may be repurposed generators from otherwise decommissioned power stations. Germany has rather a lot of them, I honestly don’t know why Ireland and South Australia don’t.

    • Peter F

      I think the main problem with synchronous converters is that they are synchronous so the power they can supply is related to the variation in frequency (see above) so you can only use about 4-6% of the rotary energy so it is a big expensive unit for very little real energy.
      Asynchronous flywheels can contribute 90% of their energy but need a large expensive power converter. The flywheel itself can be much smaller because full load speed can be 2-3 times synchronous speed

    • Electric Boogaloo

      Aren’t synchronous condensers mainly employed for power factor correction on large industrial loads?

  • nakedChimp

    Not to mention that with a distributed grid with a lot of smaller suppliers all over the place, those large ‘sudden’ event’s (where one large supplier just vanishes and the remainder has to catch the fall) will be less likely or heavy.
    It’s a completely different playing field.

    And once the consumers of electricity start to install buffer batteries at their inputs to create their own island BEHIND the meter (read UPS, just bigger than what the term uses to describe) those kind of supply problems (0.1-0.5s drops in frequency and/or voltage till help kicks in) won’t matter anymore, as your ISLAND grid behind your UPS will never see those.

    • Peter F

      Many large loads are now behind inverters already, think DC power supplies or Variable speed drives on motors