The Hornsdale Power Reserve And What It Means For Grid Battery Storage

Renewable energy has long been touted as a major requirement in the fight to stave off the world’s growing climate emergency. Governments have been slow to act, but prices continue to come down and the case for renewables grows stronger by the day.

However, renewables have always struggled around the issue of availability. Solar power only works when the sun is shining, and wind generators only when the wind is blowing. The obvious solution is to create some kind of large, grid-connected battery to store excess energy in off-peak periods, and use it to prop up the grid when renewable outputs are low. These days, that’s actually a viable idea, as South Australia proved in 2017.

Building The World’s Largest Lithium-Ion Battery

The blackout was caused by storms taking down transmission lines, but the politics of the day laid the blame at renewable energy’s door. Credit: ABC News

On the 28th of September, 2016, the state of South Australia faced an unprecedented storm. Wind gusts exceeded 120 km/h, and pouring rains threatened to flood low-lying areas. Amid the chaos, major transmission lines and interconnectors were damaged, and the entire grid went down. 1.7 million people were suddenly without power.

Thankfully, in the hours following, civilization remained standing. Emergency crews were able to bring the grid back to life relatively quickly, with most parts of the state up and running within 24 hours. Despite this, the incident was quickly used as a political football, with the reigning pro-coal Federal Government placing the blame squarely on the state’s heavy use of renewable energy. This was provably false, but spawned a greater discussion about improving grid resilience and the use of grid storage for when solar and wind generation is impractical.

Despite the hype and bluster, or perhaps because of it, the battery was built in record time. Credit: The Lead SA

The polava drew the attention of Elon Musk, who threw down the gauntlet. A battery-based storage solution installed within 100 days of contract signature — or it’s free. Further billionaires weighed in, and the State Government set the project in motion. Construction actually began prior to the contract being signed on September 29th, 2017, with the battery finished 55 days later, technically well ahead of schedule. Some decried it as a PR stunt, but the Hornsdale Power Reserve now sports the world’s largest lithium-ion battery, connected right up to the grid.

Taking Advantage of a Fast and Ready Resource

The battery as installed features 129 MWh of storage, using Samsung 21700 lithium-ion cells, similar to those in the Tesla Model 3. Capable of a maximum discharge rate of 100MW, at full tilt the battery can operate for a little over an hour. However, in day to day operation, the battery is not used in such a way. Instead, the battery is contracted to operate in several different ways to suit the needs of the grid.

Three hours of output at 30 MW (or 90 MWh of capacity) is set aside for load shifting. This is where excess renewable energy is stored in the battery when not needed by the grid. This power is then later sold on the market during peak times when electricity prices are higher.

Next to this, 70MW of output is reserved for maintenance of grid stability. Power grids require careful management of their many connected generators, in order to make sure that the voltage and frequency of the grid remain in acceptable bounds. This is achieved through maintaining a balance between the supply and demand of electricity across the entire grid. Sudden changes in either supply or demand require quick reactions from attached generators in order to avoid major excursions that can risk taking the grid offline, or lead to rolling blackouts. Commonly, this happens when a major generator such as a coal plant has a fault, or when a soap opera ends in England and thousands of households all switch on the kettle at the same time. This is referred to as Frequency Control and Ancilliary Services (FCAS) in the local market, and is typically handled by gas-powered generators, which can respond on timescales of seconds to minutes.

A graph of the Loy Yang incident in December 2018, showing the Hornsdale Power Reserve’s quick response time to the frequency excursion. Credit: Renew Economy

The Hornsdale battery, however, is capable of much faster response times. On December 14th, 2017, the Loy Yang A coal generator tripped, causing the sudden loss of 560 MW of generation from the grid. Upon the main frequency dropping to 49.8 Hz, in mere milliseconds, the Hornsdale installation delivered 7.3 MW to the grid to prop up the frequency over a period of several minutes, while other infrastructure was brought online. This fast-discharge capability has allowed the battery to beat other generators to the punch. This has brought the battery’s owners, Nueon, significant profits from the provision of FCAS services to the grid, taking 55% of the market from existing operators. There have been complaints that the existing billing system is not actually fast enough to properly compensate the battery’s owners for its output, as the system was originally designed around conventional generation which is slower to respond.

Looking To The Future

The success of the project has proven that battery storage at a grid scale is now practical and, more importantly, profitable. The Hornsdale Power Reserve was built at a cost of just $50 million, and despite their initial mockery of the state project, the Federal Liberal government have confirmed funding for the expansion of the project. The battery will be upgraded to deliver up to 150 MW, with capacity expanding to 193.5 MWh. This will allow the battery more capacity to respond in the case of adverse events, as well as improve the profitability of load shifting operations. There will also be trials of “digital inertia” technology, where the battery’s inverters will be designed to replicate the mechanical inertia of large synchronous generators, further stabilizing the grid.

Other lithium-ion battery projects are cropping up worldwide, too. Plans for a New York plant are underway, aiming for a 316 MW output and a massive 2,528 MWh of capacity, allowing the battery to run for extended periods. California is also looking to host a pair of installations, with a 400 MW project in Long Beach and a 300 MW project in Moss Landing.

With the cost of lithium-ion batteries continuing to come down thanks to economies of scale and the electric car revolution, it’s likely that yet more projects will come to fruition. Where there is profit to be made in the load shifting of renewable energy, along with a need for highly responsive grid stabilization services, grid storage batteries make a strong business case.

130 thoughts on “The Hornsdale Power Reserve And What It Means For Grid Battery Storage

  1. Ah yes, an application where the lightness and watts per kilogram of Li-Ion are super necessary, hope they’re building the racking out of carbon fibre…

    Should have done this 50 years ago with Nickel-Iron and we’d still be using it. Maybe get twice the capacity per dollar doing it now still.

    1. Ah yes, 50 years ago, when the semiconductors and control systems capable of handling such power levels were as readily available as nowadays, at a price which was not prohibitively high with regards to the application and its profitability… Those good old days! /s

      1. we didnt need semiconductors and control systems to implement a battery stabilized grid. We didnt because we had abundant coal and petrochemical resources and lacked the long term vision to understand the results.
        And RW makes a very valid point about the longevity of properly maintained nife batteries. Lithium is awesome when you need light weight and high power but in a stationary application like this there are more affordable chemistries.

        1. Exactly my point, albeit somewhat implicit. True, there was no need back then for systems like these because of an invalid assumption of eternal overabundance of fossil fueled electrical power.

          But even if the need or will was there, technology just wasn’t at a point yet that would allow for an economically viable solution, regardless of the fact that nowadays Li-Ion is far from an (economically) optimal solution to the storage part of the equation.

          My point was, for now let’s at least be glad these kind of systems are finally being developed and implemented and not begrudge or belittle them on specifics. Let’s just view these kinds of setups as a Ford Model T compared to any modern car; we’ll get there!

        2. >we didnt need semiconductors and control systems to implement a battery stabilized grid.

          Actually, we did. Battery management is a whole different beast when you have tens of thousands of cells in the same battery. The scale of the thing back in say 1950 would have required a million dollar computer with IO up the wazoo to implement what today is done with a simple Arduino.

          1. Also, IGBTs to switch/invert currents from high voltage DC to HV-AC didn’t exist until 1979, and a practical devices weren’t available until 1982-84. It took another decade for the prices to drop low enough that they could actually be used in much anything.

            Before that, you’d have to combine your grid batteries with vacuum tubes or thyristor technology, severely limiting its practical efficiency and use.

    2. And unlike lead-acid batteries which can tolerate more electrical abuse and are almost 100% recyclable, what are they going to do with all the toxic waste after the Lithium Batteries have lost their capacity? Is disposal included in the price? And the fire risk?

      This still might be another Solydra or A123

      1. Lithium ion batteries are far too valuable to throw in a landfill when they reach end of life. They are already being recycled, and are more than 90% recyclable. The only thing holding the recycling industry back is a lack of cells that have reached end of life.
        Storage – and auto batteries are also easier to recycle as the packs are several kW as opposed to single cells in phones and other electronics.

          1. Almost nil. Li-Ions are difficult to recycle due to their explosive nature and the process costs more than just making new batteries.

            There’s two ways: throw them in an incinerator, in which case it doesn’t matter if the cells burst to flames, or freeze them up with liquid nitrogen prior to crushing and grinding to prevent it. The first method makes recovery and separation difficult, while the second is just expensive and uses a ton of energy.

        1. Nobody is recycling Lithium-Ion batteries simply because the cost of recycling makes recycled materials 5x the price of the same ores dug from the ground. It just doesn’t make any economical sense.
          If we added the cost of recycling to the batteries, then the cost of batteries would be back to $1000 per KWh, which would make them uncompetitive for any practical use.

          1. That is demonstrably false. Electric vehicle batteries are being recycled today, and one of the results is that recycled materials are actually cleaner than mined materials.
            The main reason recycling is so far quite limited is simply that modern electric vehicles have not been around very long. The batteries are still good, even in older vehicles, (except for the Nisson Leaf) and not ready for recycling.
            Look up Redwood Materials, started by the CTO from Tesla. I linked to a couple of articles in another post, but it is still awaiting moderation (probably because it contains links).

          2. >That is demonstrably false.

            If you take it literally. 2-3% is still as good as none.

            >recycled materials are actually cleaner than mined materials.

            But also more expensive, which is a dead end because demand for EVs depends strongly on the price of batteries. The market is still so small exactly because most people simply can’t afford it.

        2. >The only thing holding the recycling industry back is a lack of cells that have reached end of life.

          >by 2020, China alone will generate some 500,000 metric tons of used Li-ion batteries and that by 2030, the worldwide number will hit 2 million metric tons per year.

          >But very little recycling goes on today. In Australia, for example, only 2–3% of Li-ion batteries are collected and sent offshore for recycling
          >The recycling rates in the European Union and the US—less than 5%—aren’t much higher.

          There’s plenty of material to recycle, but no sense in recycling it because the cost is too great. Battery prices would have to go up for recycling to make sense, but they must go down to make EVs and grid storage sensible.

      2. “and the fire risk?”

        Give me a break. I bet you have a lithium ion battery in your pocket RIGHT NOW. Why aren’t you worried about the fire risk of that? Or maybe the risk isn’t actually anything like what you’re insinuating.

        Also, you just literally heralded *lead* as being less toxic…

        1. The manufacturer recommendation to firefighters for automotive lithium battery fire? Let them burn out on their own, and keep the surrounding flammables from burning. Bounce your water spray off the ground before it hits the vehicle to avoid electric shock. “There aren’t enough class D (metal fires) extinguishers in half the state to put out a Tesla battery fire.” (source: firefighter/fire extinguisher rep teaching lab personnel about proper use of fire extinguishers.)

        2. >Why aren’t you worried about the fire risk of that?

          What says I aren’t? It’s a calculated risk because we haven’t got anything better. People have literally died of li-ion explosions, such as the guy who got shot in the head by his own e-cigarette when the battery went.

      3. I don’t know, those Tesla batteries last a long time when not used to full cycles.

        Lead acid will be dead in under 500 sufficiently deep cycles.

        We really need a safe metal-air battery, but until then we probably are better off learning to recycle LTO cells, that last 20k cycles. Years of near constant deep cycling could be decades of fairly gentle use.

          1. LTO doesn’t calendar age in the same way as Li Ion, nor does it have the same embodied energy issues as lead acid AFAIK. It also doesn’t explode.

            It’s obviously not perfect, but once we can fully recycle them, and production is scaled to make it cheap, it might be the best we have for stationary stuff.

      4. Some useful reading on battery recycling:
        As mentioned in the last article, for now, Tesla are mostly recycling batteries from R&D, as there are not enough end of life batteries coming back from vehicles (or storage installations) yet. In other words, even the batteries that were installed in vehicles when Tesla first started are still working well in the field, and there is a healthy second hand market for batteries from wrecked vehicles for conversions.

    3. Nickel iron batteries last a long time but have problems too.
      <65% charge/discharge efficiency, moderately high self discharge, and they don't do steep charge or discharge cycles (but this can be solved by using a lot of them).
      Lithium batteries do seem to be expensive for this type of application. But if the price is right, they perform well. Shorter life than the nickel iron though.
      I still think the future for grid scale is flow batteries.

      1. If there was a dirt cheap tech with 30% efficiency and 10% a day self discharge it would still be worth using because between midnight and 6 AM every day there’s a problem getting rid of base load electricity that nobody wants to use. (Also wind plants could be unfeathered at that time). Also east or south (North in Aus) facing solar overproduces early in the day before solar gain related loads (i.e. air conditioning and refrigeration) peak, so sometimes it’s mere hours that you need to store power for (Which in some types of cells may mean you get better efficiency, the excess heat it made is still stored in it and will improve chemical efficiency of discharge.)

          1. Yeah, that should be great to pair with solar and A/C.

            I wonder how cheap/efficient you can get it for household if you just have the equivalent of a couple of big welding tanks to stash liquefied refrigerant in? I think most of the problem is the price of the damn stuff, and the environmental concern about losing a half ton of it with a leak vs a few pounds.

            Makes me wanna build an all in one window hanger, electric solar, with secondary ammonia evap system behind it (heating ammonia, like a flame evaporator, and cooling the solar panel) and a propane pressure system, with maybe a 5lb tank…. Don’t think the 4ft by 2ft solar panel you’d get on a single window would really be up to it though. 200W if you’re lucky… think you’d only run something equivalent to 3000btu… might do better on a direct solar to ammonia type.

          2. Though really it’s better to do it at building design stage. Take for example the roman hypocaust system, false floor on stacks and stacks of tiles, with a hot air plenum beneath it…. now in summer if you put a tall chimney on this, up into clear air so the slightest breeze makes it draw well… then at night you can pull cool air through the stacks of tile, giving you a ceramic cold reservoir, then use ventilation to draw all air through them in the day, cooling it off. Though you’d have to arrange it such that if you were using the roman method of heating, direct flue gases to underfloor, that you had everything well sealed enough to not give you carbon monoxide poisoning in winter. Modern practice would probably avoid this with a heat exchanger anyway.

            Theoretically one can emulate this in a crawl space or basement, but optimally requires a free standing house, and freedom to modify it.

      2. If you only need to offset supply and demand by 12 hours, and in reality the required time shift is less than that, then self discharge is the least important issue. NiFe is ideal for such massive scale UPS systems, because that is really all it is. The efficiencies and capacities can be improved too, using nanotech to change electrode structure. There are papers on that if you have a look in the usual places for them.

    1. yup. It was sufficient to maintain frequency stability and protect from portions of the grid faulting while slower responding primary generators adjusted. That is its job. It isn’t necessary to make up the full loss of supply instantly to prevent the grid faulting, but if other sources hadn’t responded quickly (4 seconds, according to the linked article, for Gladstone coal plant to respond sufficiently), things would have gone down.

      1. That’s somewhat debatable, as there was plenty of rolling reserves in the grid. The frequency could have dropped further, but this wouldn’t necessarily have tripped anything.

        Though there is a case if the grid has enough inverter-driven generators (wind/solar) because these don’t have the flywheel effect of larger generator units, so the grid becomes more fragile and requires tighter frequency adjustments to stop these units from “panicking” and dropping off-line.

  2. I don’t know why flow batteries aren’t talked about more for grid storage. They can scale arbitrarily large (just more / bigger electrolyte tanks).

    I also don’t know why nuclear power isn’t talked about more in the context of climate change. Yes, there are lessons to learn from TMI, Fukushima and Chernobyl, but those lessons aren’t “stop using nuclear power,” but rather, “stop using unsafe PWR and BWR designs.”

    There isn’t a silver bullet for climate change. It’s a silver shotgun, with efficiency improvements, grid storage and renewables, nuclear, carbon capture, battery improvements…

    1. Three Mile Island incident were a partial meltdown mainly due to unclear documentation and poor experience in how to design and run such a plant. After all, the facility were one of the first ones ever built. So one should not really hang it out for being “catastrophic” as media often portrays it. Not like other technologies still in their infancy have had problems of their own.

      The Chernobyl “disaster” were due to a test to see how quickly the reserve cooling system would kick into action. They did the test 3 times “successfully”, ie the test didn’t end badly and were well controlled. Though didn’t give the right results, as in the reserve cooling system didn’t do its job adequately.

      When the 4th test were scheduled, they stumbled onto a small issue, that postponed the test some 10 hours, so the next shift were now responsible for doing a test that they had never done before, had no training on how to do it. And then the supervisor made dubious decisions and didn’t even follow the documented test procedure…. That it ended as it did shouldn’t be a surprise to any observer.

      If this safety test were postponed to the next day, or to a day when the trained personal were at hand, then this accident most likely would have never happened.

      The Fukushima “accident”, is something that shouldn’t even be called an “accident”, but rather “an obvious case of unneeded cost cutting”, since they decided to build the plant on top of a hill that were some 30 meters high above the sea. Quite adequate protection from waves one might think.

      But the plans quickly changed, since lifting heavy generators, turbines and the reactor itself, among other construction equipment and supplies those 30 meters posed a logistical challenge. So they solution were simple, just blast the hill to dirt and lower it some 20 meters closer to the sea…. Then the facility also had problems with water leaking into its basement, where the cooling pumps lived.

      On the day the facility had its “accident”, it got hit by a 14 meter tall wave, that would have largely been quenched if the plant weren’t some 20 meters lower then originally planed…

      This wave also relatively easily made its way into the basement, cutting power to the cooling pumps, and the general facility as a whole. Despite that the plant had some design changes from the original plan to make it more resistant against flooding, since it after all were 20 meters lower then it should have been.

      Then the supervisors handling the accident didn’t want to use the largest source of cold water available, ie the pacific ocean… Since salt water could make the cooling pipes of the plant corrode.

      Now a quick word on corrosion, is that it takes time for it to happen… Pumping sea water through the pipes isn’t going to damage them, and they can be rinsed clean with fresh water when such arrives a couple of days later. (since yes a delivery were on the way at that time)

      If the pipes were to take damage from having sea water going through them for a couple of days, then one should start to ask the question of what low quality pipes they are using….

      So in the end, all of this could have been very easily avoided.

      Though, then there is the Ågesta power plant that had their turbine hall flooded with the reactor’s cooling water, cutting out both power and control systems for part of the facility, but this ended without any issues, this plant though later got shut down for good due to low oil prices…. The source of this accident were also due to a design oversight.

        1. Transhumanists are working on it. I don’t believe they’ll get there in any timescales relevant to climate change or nuclear power, or that they will be as error free as they think, or that anyone will want to be a cyborg…

          But they’re working on it!

      1. “cutting power to the coolant pumps”
        This made me wonder why this type of plant isn’t backed up by steam powered turbo pumps? Overly simplified but reactor looses coolant flow, starts making steam, steam pumps coolant (or backup coolant) until steam is reduced… anyone more familiar with the technology know why this isn’t done?

      2. more proof that humans are too stupid for nuclear power

        but of course humans are too stupid to realize that they are stupid.

        “but of course all of this could have been very easily avoided”. Sure, just ignore human nature, just like everyone else in the nuclear power industry. We know more about the behavior of protons than we do about the behavior of humans, isn’t that reassuring?

        1. It’s also human nature to learn from mistakes.

          Second point is, these accidents weren’t actually that bad. Objectively speaking, the three major accidents were completely overhyped by anyone with an axe to grind.

    2. flow batteries is a cool technology i think they are still struggling with energy density and cost of the electrolyte making a large scale operation impractical and too costly. if i remember correctly vanadium compounds seemed the most promising density wise but then again vanadium doesn’t grow on trees.

      1. Energy density isn’t a real concern at grid scale; “charging” a mountain lake with pumped water is a viable option. So is simply filling acres of train cars with dirt and pushing them up-hill.
        5/20/50 year ROI is the limiting factor.

      2. Vanadium is limiting the price, but the modern research on flow batteries is on organic chemicals such as quinones. They’re cheap and benign to the point that the problem is rather bio-fouling or bacteria eating your battery.

    3. Apart from the obvious problem of long time storage of nuclear waste, nuclear power plants are much too expensive and take too long to build.
      As for flow batteries, I agree and I think the technology is mature enough. There are flow batteries in use today. We don’t have to wait to see whether it’ll work.

    4. Nuclear has two big problems…

      1. it has a PR problem (thanks to Chernobyl, Three Mile Island, Fukushima… etc)
      2. it takes a long time to commission and decommission

      There are nuclear-style reactors that are vastly superior to the older designs that made nuclear famous for all the wrong reasons. The waste generated has even gotten to a point where storage isn’t the major headache that it was years ago.

      BUT…(1) you have to get these concepts past the non-technical folk who will not differentiate between the different reactor designs and will immediately reply “Not In My Back Yard”, citing all kinds of irrational reasons for opposing such projects. (Just ask the operators of the Lucas Heights reactor outside Sydney, which doesn’t even generate power.) and (2) you have limited time to get the reactors built and operational.

      I’d suggest the time for new reactors was about 10-15 years ago. Now, I think battery storage is a better option as it and renewables are quickly beginning to overtake what traditional sources are able to do.

      A small traditional station (just idling along, covering the base load), coupled with lots of battery storage and renewables (to take care of the peaks), would be a winning combination IMO. Coal power stations are great for frequency stability, but they literally take hours to wind up and down in power output. Especially the big ones. Renewables are great for being reactive to load fluctuations, but without that rotating mass they don’t have the inertia to keep frequencies stable, and they’re subject to the vagaries of the weather. (You don’t get power from PV at night for example.)

      Combining the two with some battery storage to me seems ideal.

      The traditional station basically would remain at a constant output for the most part. If we got lots of cloudy wind-free days, it could wind up a little more to cover the reduced renewables output. On good days, it could wind down a little and let the renewables take over. Load spikes would be smoothed out by battery storage.

      1. 3. human beings do not possess the necessary intelligence to handle nuclear materials.

        The proof is in the pudding: multiple accidents, multiple fatalities, radioactive material in the ocean, radioactive material in the Columbia river, nuke plants located on beaches with inadequate emergency exits, cooling towers that fall down because nobody bothered to check for corrosion, stupid plant workers who can’t be bothered to wipe their feet, etc. Truly a clown show for the ages.

          1. Bukit Merah was an incident with Mitsubishi operating a rare-earth mine/refinery, where they dumped the radioactive tailings on the local community. When this was discovered to produce leukemia and birth defects in the local population, they hollowed out an entire mountain and drove 50,000 truckloads of contaminated earth into it to contain the problem. This was all hush-hushed in the media except for some rare reports, because the REEs were needed to produce electronics, magnets for generators etc.

  3. Having had a casual interest in grid energy storage systems throughout the years.
    I do have to say that using lithium ion batteries is a “decent” solution. It “works”, but it is far far from ideal…. And as RW pointed out above, Lithium is more well suited for applications where power density and weight are of main concerns.

    The far more mundane lead acid battery is a more reliable solution for this large scale stationary application.

    Though, most lead acid batteries found in cars and such typically die due to the failure of one cell out of the normally 6 cells within a typical 12 volt battery, yes, even lead acid batteries wants their cells balanced, who knew….. Not to mention people discharging them far bellow what is long term reliable for such cells…

    Though, for grid stabilization, a synchronous condenser can actually be a far more cost effective solution. (This is in effect a large three phase AC motor/generator spinning a flywheel, sometimes in a vacuum to lower energy losses due to air resistance. Simple stuff indeed, but needs occasional maintenance.)

    For energy storage on the other hand, these synchronous condensers aren’t really “ideal”. Though, they are able to act very quickly on any changes in demand, then they also do some load balancing of the phases, and even a fair bit of power factor correction, all things that makes a grid more reliable.

    But in terms of energy storage solutions, there is one that is rather popular throughout the world, and that is pumped water, but this requires that one has some elevation difference at hand. Otherwise it is kinda useless. (Though, technically speaking, one could just dig a hole a few hundred meters into the ground and blast out a large cave, then just flood it when one needs power, and pump it out when one has an excess of energy on the grid. Downside here would be ground water seeping in…)

    Pumped water have the advantage of being relatively simple, and in areas where one already has a dam, then one just needs a pump for the parts of the year when melt-water/rain isn’t overly abundant. And large dams can also be built to handle rather quick changes in demand, making it a very well fitting compliment to both solar and wind.

    Another solution is compressed air energy storage. Currently there is some large scale setups of this in use, with charge/discharge losses of less than 20%, so very comparable to batteries, though, unlike batteries one doesn’t need fancy chemistry, only a large storage vessel, typically a cave, or a large “fiberglass”/composite tank.

    Some offshore wind farms are looking into using large balloons on the seafloor for storing compressed air, practically working as expansion vessels. Since this would be fairly cheap for the amount of energy stored. Then there is also the advantage that one can also sell the compressed air as a service to nearby industry. (Meaning that the industry doesn’t need to run their own, usually less efficient compressors)

    There is also flywheel based energy storage systems. But floating a couple of tons spinning at many thousands of RPM on top of a magnetic bearing within a vacuum might be a bit fancy.

    In the end, there is plenty of energy storage solutions, but lithium ion batteries isn’t really ideal for this application.

      1. It doesn’t have to be water. Gravity fed systems can be used with anything heavy. Mine shafts with a heavy weight, tall towers, a rail track up a mountain. These systems have the advantage that sand, concrete or scrap metal etc. won’t evaporate.
        Disused mine shafts have the advantage that the shaft is already there as well as a lot of the infrastructure, and environmentalists are less likely to get upset over changes to a landscape that has already been mined for decades. Even a useful number of prospective employees with experience in mine shafts are likely to be available.

        1. Using weights to store potential energy is rather inefficient though. Not from an energy standpoint.
          But rather from the standpoint of how much weights one will need.

          Even having 425000 tons (metric) suspended 500 meters high would only be about 577 Kw/h of stored energy. To paint a picture, this is one thousand, 20 foot intermodal containers worth of solid lead suspended half a kilometer in the air.

          If this were falling at 10 m/s, then it would only generate 41 MW of power for about 50 seconds. (Though, accelerating and de-accelerating is a thing….) About enough to power a smaller suburb.

          So practically speaking, this isn’t practical.

          Another energy storage method that doesn’t need water as well is compressed air.
          Here, in a near isothermal setup, compressing air to 200 bar, and using the same one thousand 20 foot intermodal containers worth of volume, then we can store 1102 Kw/h of energy.

          This is almost twice as much energy, but we should probably also consider that the weights need a huge shaft to descend into. (Adding the volume of the shaft should technically be done, but that just makes the weights look pathetic in comparison.)

          Storing compressed air is though more tricky, but composite materials surrounded by rock tends to survive high pressures without too much difficulty. And it isn’t like suspending literally thousands of tons of weights hundreds of meters into the air doesn’t pose its own set of challenges.

          Efficiency of the motors lifting the weights, or compressing the air, and the generators used to take the energy back out is though a different story. But there is commercial compressed air energy storage facilities in use currently with losses lower then 30%, and this is also improving as well.

          Not to mention, air is rather cheap compared to practically any weights… And it has phenomenally low shipping costs for some odd reason. Though, building vessels to contain the pressure is likely not even close to “free”…

          1. After recalculating some numbers. I were off by 3 orders of magnitude on how much energy those weighs would store…

            It isn’t 577 Kw/h, it is 577 Mw/h.

            Though, it still need a complicated mechanical setup, while compressed air only needs a storage vessel.

            And a lot of compressed air energy storage facilities use underground spaces like former salt mines, natural gass wells, and simple things like caves, or tunnels.

          2. Ok, I see you corrected your numbers in the next post, but for some reason I can’t reply to that one.
            Based on your revised numbers, I think mine shafts are perfectly feasible, but may be more suited to medium/local storage rather than huge installations. As said, the cost will be relatively low, because the mine shaft is already paid for and there would already be roads, power cables etc. from the time the mine was in operation.

          3. Stop saying “Kw/h”. It’s doubly wrong on both the unit and the prefix.

            kilo is small, Watts is capitals, and it’s kW-h or kWh, not “per hour”. Like foot-pounds, not foot per pounds.

        1. It’s not like shifting behavior isn’t a thing. Remember when telephone service was expensive, and some consumer phones automatically figured out who was the cheapest carrier, and routed through that? Then there’s the shifting DVRs brought about, although that wasn’t economically driven, but convenience. Electrical usage isn’t any different, and in some parts of the world with unreliable grids, time has to be taken into account.

        2. My power company already encourages this by charging me more for power during peak times, so I have consciously thought about it. The main form of “load leveling” I do is charging my plug-in hybrid at night when the rates get cheap.

          I guarantee you factories already think about this. Large commercial customers are surcharged for all kinds of things, including high peak current and low power factor.

          1. Yes, I also charge my Volt at night and usually run the washer and dryer at night, but I still have a peak in the evening for cooking, AC (in summer) etc.

            I somehow don’t think businesses that only have one or two shifts are going to be able to get their workers to come in at night, and they normally use most power when people are there.

            Behind the meter battery local storage (and solar) can make a lot of sense though, not only to buy at night and use in the day, but also to smooth out those expensive transients that occur in the morning when everyone comes in and starts turning stuff on.

          2. OHM [Owen Harald Morgan] wrote:

            “I somehow don’t think businesses that only have one or two shifts are going to be able to get their workers to come in at night,”

            People who work nights desire a “shift differential” i.e. more pay for working over night. The savings of the business using off-peak power would need to be comparable.

          3. >Behind the meter battery local storage (and solar) can make a lot of sense though

            It would make sense, but not financial sense with the current subsidy schemes. Net metering means your power sold to the grid is worth retail power prices (free electricity), so there’s no sense in storing it. The storage costs you more than using the grid as your virtual battery.

      1. There’s no unit of time measurement so I guess we won’t know. I wonder if that 560MW figure is per hour and the battery stepped in and provided 7.3MW total over some short timeframe. 560MW/h would be 9.3MW per minute. Then 7.3MW would be enough power for other backup generators to kick in within a minute of the 560MW one going down.

        1. Uuuuuuhhhmmm… The Watt is a unit of *power* defined as 1 joule (a unit of *energy* ) per second. (That is, time is built into the unit.) (Mega/giga/kilo/etc)Watts per hour is an almost useless unit. That would be saying Joules per second per hour. I think you may be confused with the Watt-hour, which is one Watt for one hour (a roundabout unit for *energy* equal to 3600 Joules).

          560MW and 7.3 MW are standalone amounts of power consumption. (And yes, that is a big mismatch! But I guess every bit helps.)

          1. I guess you’ve never paid for an energy bill before…. because watts per hour is exactaly how electricity is billed.

            It takes a given amount of energy to provide 1 watt for an hour. It also takes that same amount of power to provide 2 watts for half that time.

            In the case of batteries, it come down to how fast you drain it. If you drain it twice as quickly, you draw twice the power, aka twice the wattage.

            1 watt/hour means that the system uses 1 watt for an hour. Or 2 watts for 30min. Or 4 watts for 15min.

            Lastly, the logic of “per second” being in the unit twice being a problem is false. Take for example the units of acceleration: distance per second per second (often written as distance per second squared).

          2. While Doc Oct and Nathan are definitely confused with regards to units, I do think that the W/h performance of the power reserve really helped here. After all, it’s proven ability to ramp up power output at around 2TW/h allowed to to get up to 7.3MW in “milliseconds”, thus quickly stabilizing the grid. The gas generators which eventually supplied most of the power only worked at about 500 GW/h, a quarter of the rate.

          3. >The gas generators which eventually supplied most of the power only worked at about 500 GW/h, a quarter of the rate.

            This is why when you have lots of renewable power in the grid (high load/supply variations), the preferred method is to use large diesel generators running on natural gas (and a bit of oil for ignition). They’re faster to spool up and down than gas turbines. Wartsila-Sultzer is making bank selling huge 50 MW generator units, originally designed for cargo ships, to California because of their renewable energy mandates.

            But what can you do. Bureaucrats say jump, so the engineers jump.

        2. No, the numbers are correct. The HPR only injects a small amount of power but it is able to halt and start correcting the frequency transient created by the loss of the large coal units. Without that very fast frequency control our grid here in SA can cascade collapse as additional generators go offline to physically protect themselves due to the frequency deviations. HPR stops the deviations for a few seconds to minutes while other assets come online then it backs off. It doesn’t need a lot of power to hold the grid up.

    1. When a 560MW generator call it quits, that load goes (unbalanced) to the rest of the power plants, and even if they could supply that sudden load with no problems, the result is that the frequency of each one sags a little. And the grid NEEDS to be fully synchronized. Even a “small” difference of 0.2Hz between power plants can begin a domino effect of generators tripping out, on the worst scenario doing serious damage to the grid’s infrastructure (less likely if well mantained). Remember we’re talking megawatts here.

      If by inyecting mere 7.3MW this system manages to null the frequency sag, it is effectively saving the day because it’s preventing the entire grid from going into panic mode & shutting down. What I find impressive is how fast the systems calculate how much juice to give, the maths behind that should be bonkers.

      *stupid me clicked report instead of reply, sorry…

  4. It is a load leveler with 14 seconds of capacity if you match it against the state wide load, my UPS systems have 15 minutes. I am not impressed. I wish I could scam the energy markets with a battery bank the way they do as my overnight off peak power is much cheaper.

    1. Your UPS might power your computer for 15 minutes, but I very much doubt it would last as long powering your whole house, or your whole neighbourhood, or your whole city. If you make a UPS that can power a city for 15 minutes you’ll be very richly rewarded, I’m sure.

  5. Apart from political points, that is the real use of this battery – very short term power supply or arbitrage.

    Yet it was sold as ‘massive energy storage’.

    If you look at what is being produced right now – – you’ll see SA is making 1585MW at the moment – with gas=625MW, wind = 587MW, large solar = 139MW and small solar = 223MW.
    So the battery, flat out, can provide 6.3% of the network for an hour.

    Of course, that isn’t how it is used, it’s used as something to smooth out the other forms of supply (which it is good at doing) – but the public perception is that it can fill in (for example) solar at night. Nope. Well, maybe for 15 minutes.

    Don’t get me wrong, I think something like this is a good idea in the energy mix, and if SA – with it’s small population and rich in energy – can’t go renewable then nobody can..

    However, they are also relying on the other states providing them power when they need it (ie the network is the main battery) and if all the states did what SA has done – much harder due to some of them being many times the size – everyone would have a lot of trouble making it work. And we wouldn’t need a battery with 129MW hours, we would need something at least 10 times bigger…

    So without their gas generation, and the rest of the network (as their main battery), they would be stuffed…

    I like renewables. And a combination of them and gas is a reasonable way to go. But unless we get a lot better storage than we have now – and I don’t think there is anything on the horizon – it just isn’t going to work large scale without either nuclear (or coal) OR a complete change in how we use energy…

  6. “Lead acid will be dead in under 500 sufficiently deep cycles.”

    True, but no adequately-designed and maintained lead-acid bank will be discharged to that extent.

    I’m off-grid with 1320ah of lead-acid. Installed in 2009. That’s about 3,800 days, or charge/discharge cycles.

    Not quite the capacity when they were new, of course. but far from dead. A decent charge controller with an adequate bank of solar PV takes care of charge/discharge.

    And a backup generator. If you’re sufficiently motivated to take care of your battery, it’ll last a long time

    1. That’s actually a problem, because you have all those batteries sitting there for nothing, being only used for the top 10-20%. Lead batteries are difficult to scale up for grid use because you need so much lead that there’s not enough supply in the world if everybody starts doing it.

  7. Why do you expect a single ‘generator’ to supply the whole state at any time? No generator of any type does that. Poor argument.

    And “load leveler” is a small part of what the HPR is doing so well. Its huge successes are in FCAS (Frequency Control and Ancillary Services) – controlling voltage and frequency of the network when step transients introduce problems. There have now been a number of times that the HPR has halted and corrected frequency transients that would otherwise have caused cascade failures, and it has done it much faster than traditional systems can manage (or, more accurately, fail to manage), preventing quite a few blackouts in its first year of operation. What this fast, accurate frequency and voltage stabilisation has allowed is for the AEMO (the guys that run the grid over most of the Australian population) to dial back the amount of reserve gas generators spinning as stability backups. These gas generators are the ones that have forced our wholesale electricity prices in Sth Aust through the roof over the last decade and the HPR has single handedly smashed their cartel (to the tune of $40 million last year). The upcoming upgrade is expected to further halve the amount of gas redundancy required in the system saving even more millions of dollars.

    So when people point at the HPR and laugh because it can only supply a whole state for 14 seconds, we look on with pity at them and revel in the technology that has taken our grid from an unstable state to zero blackouts, with high levels of solar and wind renewables, in a single year. Laugh away.

  8. Hyundai already beat Tesla by building a bigger battery in Ulsan, South Korea.

    The fact of the matter is, Tesla’s Megabattery is just about 1,500 Model S batteries that Elon Musk dumped into a “project” because the Model S was selling poorly prior to the expected launch of the Model 3 and he wanted to shift some stock.

      1. Depends. The Powerwall is supposed to use a different type of battery – one that lasts longer in use – the one that was originally meant for Model 3 but they couldn’t make it work cheaply enough.

        1. Basically, the Model S and 3 currently use Nickel Cobalt Aluminum (NCA) Li-ion cells which have a great energy density and incredible power density (“Ludicurous acceleration!”) for a lithium-ion cell, and they’ve honed the process down with Panasonic to push the cost per kWh down – but the tradeoff is the tendency to burn and explode with a tendency of internal thermal runaway if the cells are heated above 160 C. They also don’t have a very great shelf-life with an expected lifespan of around 12 years if you baby them with thermal control and low currents.

          The Powerwall is supposed to use Nickel Manganese Cobalt (NMC) batteries which have greater cycle and shelf life, depending on the environment (temperature), but they cost more to make and the energy and power density aren’t so great – so no more “ludicurous mode”. Installing these cells in the cars would look like a downgrade because they’d have to use fewer cells to keep the price down, which would result in lower performance and range – partially compensated by making the car lighter and smaller – which would turn them identical to their competitors such as Nissan, VW, Seat etc.

          So there you see why Tesla is such a “trailblazer”. They make different compromises because they’re selling to a different crowd who don’t mind, or don’t understand/care what they’re buying into.

    1. It was also cheaper than Tesla’s $50 million at around €38 million for 50% more capacity. But that’s still expensive, because the useful lifespan is only little over 10 years. If the Tesla battery is cycled once every day, the throughput cost is about $106/MWh on top of the initial power price (eg. $50/MWh + $106/MWh) which is another reason why the battery cannot be used to provide everyday power – it can only compete with the most expensive generators.

      And if you want to talk about capacity, there’s a CAES plant in McIntosh, Alabama, which stores 2,860 MWh and can discharge over 26 hours, producing steady power equal to the Tesla battery.

      1. Man, a pressure vessel failure at a plant like that is a scary thought. 2.86 GWh is equivalent to about 2,300 tons of TNT. That’s 200 times the size of the fertilizer explosion at West, Texas. It’s nearly as big as the Halifax explosion, which killed 1,950 people. I hope it’s well away from any populated areas.

        1. One of the limiting factors for a CAES explosion is the rapid temperature drop when the gas starts to escape – if it comes out too fast, it loses pressure. That’s why the system is combined with a gas turbine – it actually uses the waste heat of the turbine to pre-heat the gas coming out of the cavern and keep the pressure high.

          Without a pre-heater the round-trip efficiency of the system would be poor. This is a problem for systems with low capacity – 2.86 GWh is still small beans for grid storage and for CAES – because the temperature of the system doesn’t have enough time to equalize. It works better when the charge/discharge rate is low – less than 10% of the capacity per day.

  9. How many years do the cells last in these systems? It seems like all of my phones / laptops / whatevers are down to about 50% capacity after a couple years.

    So I have to wonder how often these will need replacing, and how much it costs to replace them, and how much that affects the long term ROI.

      1. Almost all of Model S’s are 2014 or younger. The battery is so large that they experience very little cycling load, much less than a cellphone or a laptop would. In such use, even a consumer grade li-ion battery can last 8 years before it starts to show signs of wear.

        You should expect to see a ton of Model S cars getting scrapped around 2022.

      2. Hi temperatures are a lithium killer. What do you think happens when you continually put grid scale voltage through them and then discharge multi mwh.. ? Li-on are just not suitable for grid scale use. They are not safe, will need replacing after ever diminishing capacity, are toxic and an expensive pain to recycle when dead. Save them for where we really need them currently.

        1. “Hi temperatures are a lithium killer. What do you think happens when you continually put grid scale voltage through them and then discharge multi mwh.. ? Li-on are just not suitable for grid scale use. ”

          Don’t let the fact that it’s been running and performing brilliantly in the real world sway your opinion. 😆

    1. The initial 90 and a bit million spent on the system is well on track to be made back in less than two years. So ROI is a slam dunk. $40m saved in FCAS in the first year, plus Neon/Tesla making millions on playing the charge/discharge spot price market. System is expected to last ten years plus, so ROI is pretty darn solid.
      Even if it turns out worse than design by half an order of magnitude and is as bad as your phone, it’s still worth it.

  10. Why all this debate when Vanadium flow batteries are safe, durable, don’t degrade, can discharge to 0 with problem, have an electrolyte that retains its value and is then easy to recycle in to the standard steel industry route, aren’t explosive or flammable, can store 100s MWhs, can discharge over a long duration, can be maintained easily and safely, need no temperature management system etc.
    . For all the above reasons, and as Vanadium is back to a reasonable price level after last year’s spike, they are cheap over their full lifecycle. Why do we think that China are building giant versions…

    1. Vanadium batteries have a limit price around $150/kWh storage because of the supply of vanadium. That’s on par with the cheapest li-ion batteries on the market today, which is still too expensive in the long run.

      1. Lease agreements are being struck to lower capex costs based on the long term store of value intrinsic in the Vanadium contained. Primary Vanadium miners are also ramping up production hence the stable lower price. Eg Largo and Bushveld minerals

        1. The $150/kWh was a projection for the lowest production cost estimate. The actual price is currently much higher. Vanadium recycling (or lithium recycling for that matter, or even both together) doesn’t solve the question because the industry needs to expand rapidly to terawatt-hour level storage. The supply cannot physically keep up with demand, and the price goes up.

          Building megawatt-hour scale batteries is literally a million times too small to matter in the large scheme of things, because batteries are ultimately needed to provide seasonal scale energy storage of renewable electricity. The fossil fuel infrastructure currently holds 2-3 months worth of strategic reserves, and the battery electric infrastructure must ultimately meet that. For the US, that means approximately 900 Terawatt-hours of energy stored – somehow. Multiply by all the other countries in the world. It’s such a huge number that it won’t be met with all the lithium, vanadium, cobalt etc. you can possibly dig up in a hurry.

          This is why talking about batteries for grid storage and a solution to renewable energy woes is missing the point entirely. It’s a neat idea, but too little, too late.

          1. “The fossil fuel infrastructure currently holds 2-3 months worth of strategic reserves, and the battery electric infrastructure must ultimately meet that.”

            But, but…
            The Sun has 4 Billion years of strategic reserves!

          2. I don’t think it’s valid to compare like that. FF power generation holds reserves because it takes time to get the burnable product out of the ground then ship it to the power station. Wind and solar do not need reserves because the ‘product’ arrives on site all by itself with no mining or shipping. Unless the sun isn’t going to shine for 2-3 months and the wind isn’t going to blow for 2-3 months, then I don’t see how the same level of storage is required.
            And everybody also keeps forgetting the point that batteries such as the HPR are massively useful for system stability (i.e. the point of the original article here), they’re not there just to time shift power production-usage curves.

          3. >FF power generation holds reserves because it takes time to get the burnable product out of the ground then ship it to the power station.

            It is a fair comparison. Imagine if a tornado tears down a major power corridor connecting wind turbines and solar panels across state lines. Takes 2-3 months to re-build.

          4. > Unless the sun isn’t going to shine for 2-3 months and the wind isn’t going to blow for 2-3 months

            Uh… winter?

            And wind power too has 30-50% seasonal variability. Hydroelectric power has year-to-year variability, etc.

          5. For example:


            The combined output of wind and solar power in California varies from a low 1.9 TWh in January to a high 4.7 TWh in June 2017, or a factor of 2.5 difference over six months. Meanwhile, the CAISO electricity market remains relatively stable between 20 – 25 TWh per month sold.

            You need a LOT of batteries to even try to match the supply with the demand, and you have to do it all over the year, not just over days or weeks. Some 100 MWh “megabattery” is like a fart in the wind to the actual size of this problem.


            Prediction of the necessary storage capacity:
            >8333 TW h per season long term storage by 2050 worldwide.
            > The short term storage then corresponds to about 0.7% of the seasonal storage, and on average ∼8 kWh installed storage capacity per person on an earth with 7 × 109 inhabitants.

            So, we need just 83.3 million Tesla mega-batteries to satisfy the global demand for renewable energy storage by 2050. Easy peasy.

          7. >And everybody also keeps forgetting the point that batteries such as the HPR are massively useful for system stability

            Why is there suddenly more instabilities in the system? Oh right, more renewable power using the grid as a virtual battery over longer distances, making faults spread faster and wider. The fundamental problem is still the lack of adequate distributed energy storage capacity in the grid, and these tiny tiny batteries aren’t helping much.

  11. As someone who lives in the state (SA), it seems that the impact was somewhat understated.

    The CBD had power back after about six hours. Most suburbs had it back after around ten hours. Some regional areas (within 3hrs of CBD) didn’t have power back for four whole days.

    Extra flak was received because at least one of the major hospitals almost exhausted its generator fuel supply.

    With the suburbs seeing maybe 1-2 ten second outages a year, plus a 30min-60min outage roughly every three years, this was new and scary for many.

    Plenty of fuel for it to become a political football.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.