There is no argument about the vital role storage will play in a grid dominated by variable renewables, writes Jeremy Chunn, but its rise is sure to be a prolonged affair that is potholed with risk.
Renewables are the irritable toddlers we must learn to live with. We love them, yes, but their wild swings in energy – bouncing off the walls one minute, asleep over dinner the next – require a lot of patience. In terms of pushing up levels of wind and solar in the grid to replace coal and gas, the job of levelling out variable supply to match predictable demand goes to storage assets.
In its Integrated System Plan 2020, the Australian Energy Market Operator called for 6-18GW of new dispatchable resources – most of it storage – to set up the National Electricity Market for a transition to clean energy. “Market design needs to reward the increasing value of flexibility and dispatchability in complementing and firming variable generation,” the authors of the ISP wrote.
The federal government’s Technology Investment Roadmap, released in September, included as one of its “stretch goals” an ambition to bring the cost of electricity from storage to below $100/MWh. The government is pitching in $18 billion with the expectation of attracting $100 billion in private funds to prompt development in production of hydrogen, batteries, low-carbon steel and aluminium, carbon capture and storage and carbon sequestration.
The Morrison government accepted in the roadmap that “broad deployment of storage will facilitate more low-cost solar and wind electricity in the grid”.
Storage projects are falling into place around Australia. In October the NSW government ticked off planning approval for the 600MW Oven Mountain pumped hydro facility, cited to ease the way for new solar and wind generation in the Renewable Energy Zone being established in the state’s north. AGL Energy has a target of installing 1.2GW of new battery storage and demand response capacity by 2024, and let’s not forget the 1.7GW Battery of the Nation project in Tasmania, the 2GW in NSW planned for Snowy 2.0 and the many gigawatts in the project pipeline around the country, linked to wind, solar and hydrogen plants.
It appears the age of storage is upon us. Or is it?
Who will back pumped hydro?
Looking far ahead, BloombergNEF Asia-Pacific power market specialist Ali Asghar expects batteries and pumped hydro will do most of the storage work in tomorrow’s energy system.
Pumped hydro, he says, is a difficult technology for investors to grapple with. “There are a lot of uncertainties [around pumped hydro] and the economics look challenging,” Asghar says, listing long planning and construction timelines, long economic lifetimes, a mature technology with less room for cost reductions and highly site-specific cost structures. For these reasons he expects state and federal governments will step in to back pumped hydro projects, as they have at Snowy 2.0 and Oven Mountain in NSW.
“Other than Snowy Hydro, none has reached the point of a final investment decision, and there is a good reason for that,” he says. “Anything you hear for a pumped hydro project in terms of costs right now, it’s certainly not going to be the case once the feasibility studies are over. There is a greater probability of cost creeps.”
In an open letter to Prime Minister Scott Morrison and NSW Premier Gladys Berejiklian the $2 billion estimate for Snowy 2.0 has already been queried by a group of 37 industry experts, who say $10 billion will be closer to the mark.
The change in supply and demand over the next 10 years will be unprecedented, Asghar says, as renewables edge up and up in the mix and coal generators retire. “And on the demand side we’ve seen a record amount of rooftop solar generation, and that in itself changes the underlying demand profile in Australia,” Asghar says. “We’ll require a lot more flexible generation.”
As for coal closures, BNEF expects the 2GW Liddell plant in NSW to close in 2023, with the next closure in 2028-29 and further closures in the first half of the 2030s. Some closures will be almost predictable, as plants reach the end of their technical lives, shareholders put pressure on owners and degraded performance makes them uneconomical.
But here’s the crux of the problem. Owners of remaining plants may have made no commitments to back away and may even extend operations to milk revenues from gyrations in pricing as supply is stepped down. “Every time a coal plant closes that results in price spikes, which is great for generators,” Asghar says. “They can make a lot of money. Everybody is waiting for who blinks first.”
Owners of storage facilities built before the staggered exit of coal will also be able to partake in these jumps in revenue, of course. “Timing is everything,” Asghar says. “You don’t want to be building an asset in 2025 if you don’t need it until 2030.” The sweet spot for investment and construction, he says, is around 2028-30. On closer inspection everything appears to hinge on plans for the 1.3GW Vales Point coal station. “That’s the one that has no public commitment to retire,” he says.
Large amounts of storage are most likely needed around the late 2020s to early 2030s, he says, and until then developers will be watching the cost of lithium-ion technology fall – they hope. “There are a lot of people on the sidelines looking at this,” he says.
If the NEM really only needs a huge step up in storage a decade from now, no-one today can say with much certainty what will be the cheapest way to solve the problem. If anyone with private money wants to back pumped hydro in the meantime, they’d better start soon to have a facility up and running in time.
Many types of storage
Of course, it’s worth remembering that electricity can’t be stored. The solutions mentioned here involve chemical applications or rely on gravity, among other things. Every technology comes with a trade off in efficiency and many solutions only look good when they are connected to luxurious amounts of supply. The problem isn’t just to stabilise supply but to account for wayward output from variable generation sources. Even then it’s worth looking harder at the problem to understand its abstractions.
“I would encourage you to define storage as broadly as you can,” says Dr Brian Spak, who leads the CSIRO’s grids and renewable integration work and can identify deep reserves on the demand side.
Take thermal storage, for example, which includes the temperature of the air inside a building and the inertia associated with it. It also includes demand response for air-conditioning, where houses are pre-cooled during times of surplus solar or air-con units are set a degree or two warmer than some people’s chilly preferences. “That’s certainly a renewable integration avenue, and arguably a type or storage,” Spak says.
Australia already has the most successful demand response program in the world, he says, with the water heater programs in NSW and Queensland introduced to facilitate coal plants to run through the night. “There’s no reason why they can’t be adapted relatively simply to integrate solar during the day,” Spak says. “There’s an energy arbitrage value, an FCAS value and, if you put the project at the right part of the grid, you might be able to get some network value,” he says. “There are lots of things storage could do that the market doesn’t really enable.”
Renewables are producing in parts of the grid where there is no demand at the moment, Spak says. If you owned a solar farm with a battery on it, and you could forecast what the price was going to be at different times of the day, you could use the battery to shape output of your solar farm to where the grid needs it. “But there is no market signal for that now that incorporates marginal loss factors, because it’s levelized over the course of a year.”
The rise of electric vehicles will bring a useful variable load to the market, Spak says. “One way to help solve the problem of decarbonising the grid and having it run on renewables is to make the system bigger,” he says. “Don’t just try to do it with the existing electricity system; bring transport along for the ride. You’d get a lot of benefit from the flexibility of vehicles.”
Networks and retailers will likely adjust tariffs to incentivise energy use and limit PV exports during the middle of the day, as has happened in South Australia. “We want to make it in customers’ interests to run everything, but especially electric vehicle charging, in the middle of the day.”
Spak says the ability of having a legion of EV batteries connected to the grid to charge and discharge would simplify short-term storage problems.
The grid itself can also be classified as storage. For instance, if the east coast was linked by transmission to Western Australia, the evening peaks in the eastern states could be largely powered by solar. “A broader geographic coverage makes the problem easier to solve as well,” he says. “The challenge of renewable integration doesn’t just depend on how much renewables you have on the grid; it depends on the nature of the grid itself and how robust the transmission networks are.”
Developers of large-scale solar plants in poorly-connected parts of the grid where curtailment is a significant issue will agree that transmission needs to catch up if the market is to get the full benefit of our strong sun and wind resources. Building transmission is a headache everywhere in the world, he says, “but it’s an easy way to solve some of this problem.”
CSIRO modelling has shown the need for storage becomes evident as renewables account for about half energy supply, but the need for storage grows strongly around the 80% level. The solution relies on where the generation is, what type of generation it is, and where the storage is and what type it is.
The solution might be to detour around storage and build so much solar that the NEM can get through the darkest days of winter and in summer surplus solar is spilled. “There’s a lot of research that suggests renewable curtailment is a feature, not a bug, of high renewable systems,” he says. “You overbuild and sometimes you’re going to have more than you need and you don’t use it. We just have to get more comfortable with that approach.”
In that case, shaping and moving energy consumption will become more important than saving it and reducing consumption. “Building out the grid tends not to be a very cost-effective solution; we should be getting better utilisation out of the existing infrastructure, which includes water heaters, air-conditioners and pool pumps,” he says. “We just need to open the markets and create the opportunity for those devices to play a role in the energy future.”
A plan for dispatch
Who is the ultimate architect of the grid? There isn’t one. And the reason AEMO delivered its Integrated System Plans in 2018 and this year was to provide a vision for how the system is evolving and how a better version can be planned for. “The grid is a market,” says AEMO group manager forecasting, system design and engineering Nicola Falcon, “and a lot of the future development and generation is going to be driven by market incentives and private investors’ preferences to type of technology.”
The Integrated System Plan, used as an investor’s guide, sows the clues as to the services and needs the power system will require, and for regulators it offers guidance on the transmission needed to unlock the possibilities of this efficient future energy mix.
The ISP sets out that the grid will need between 6-19GW of “dispatchable resources”, most of it storage, but AEMO’s planners can’t foresee whether the market will provide the right incentives for it to be built in time to balance the staggered exit of a dozen or so coal-fired power stations.
AEMO’s report rounded storage into three categories: “shallow” storage for capacity, ramping and FCAS (including virtual power plants and two-hour large-scale batteries); “medium” storage for intra-day shifting (four-hour batteries, 6-12-hour pumped hydro and existing pumped hydro) and “deep” storage for variable renewable energy droughts and seasonal smoothing (24-hour pumped hydro and 48-hour pumped hydro).
The appetite for building assets in each storage class will in many respects rely on requirements for transmission to play a supporting role. An owner of a solar plant, for example, might decide to add storage on site to even out supply and boost revenues from a hybrid solution. “They’re not going to be waiting for transmission; they want to be able to do that just to make sure they can generate at the times that maximise value,” Falcon says.
Deeper storage solutions, however, will require coordination of new transmission to really set off the fireworks. “Consumers won’t be able to benefit from Snowy 2.0 unless the HumeLink [transmission] is developed to get generation to the load centres.”
In other cases it will make sense for storage to be built ahead of transmission, Falcon says, pointing to the Renewable Energy Zones announced by NSW and Queensland. “All of the solar and wind in those areas won’t be built simultaneously, so building some batteries ahead of the transmission can help smooth out some of that renewable generation,” she says.
Revenue can flood to owners of fast response batteries, as has been shown by Neoen’s success with the Hornsdale Power Reserve in South Australia, but there is a risk to owners in that opportunities will narrow as more storage assets enter the market. Is it that simple? “That argument holds if the system is not changing significantly,” Falcon says, “yet what we’re finding is that with large volumes of coal retiring, particularly beyond 2030, there is actually going to be potential for highly volatile prices, particularly if gas is required to fill the gap.”
Owners of storage assets may need deep reserves of patience and very sharp reflexes when required. If, however, an energy market that reflects the zero marginal cost of high levels of renewables penetration leads to wafer-thin arbitrage opportunities, new incentives may need to be provided so that battery owners can turn a profit for providing essential services.
At AEMO, Falcon sees EVs and VPPs as necessary energy sponges that can be activated during minimum demand periods, when PV systems are exporting like billy-o. “That’s going to dramatically improve the outlook for system security,” she says. Home batteries generally discharge stored solar during the evening peak and morning, she says, but a coordinated fleet in a VPP might achieve better value by being on call to earn money when required and top up if needed during overnight off-peak, as well as on sunlight during the day. “That could be a more efficient outcome for both the system and consumers,” she says. “It’s about being able to optimise their use in a much smarter way.”
Use it or lose it
Let’s face it, consumers aren’t that interested in electricity. For VPPs and EVs to really flex their muscles the aggregators that command them will have to explain their services to consumers and what it means to their household budgets. For the digital age of optimised delivery of energy to take hold in consumers’ minds, they will have to trust the provider who tells them they can “set and forget”. The industry watchdogs will have to be super vigilant, on everyone’s behalf.
Falcon says AEMO is keeping a close eye on hydrogen, to understand its role in the energy balancing act. It looks expensive in this market, but who’s to say costs won’t fall – as they are predicted to. “The potential long-term opportunities for hydrogen … could allow for electrolysers to be operating at times of high renewable generation – essentially as a solar soak,” she says.
For the past few years it’s felt as though a new form of energy storage is invented every day: underground compressed air, rail cars loaded with rocks, a nest of cranes that lifts and lowers concrete blocks, molten sand, molten salt, various forms of concentrated solar thermal and more. It’s up to investors to pick the right tech for the job, as long as they can do one of four things: fast-start short duration to respond quickly to supply and demand panics and deliver FCAS; two-four-hour storage to shift solar generation into the evening peak; 12-24-hour-plus storage to cover wind droughts, and; very deep inter-seasonal solutions, like hydro and hydrogen.
“We know we need a portfolio of different storage durations and it’s going to be up to the market and technology development to determine what’s the best technology for each of those different needs,” Falcon says. “It’s a case of needing a number of things to work in concert to ensure we get the best outcomes for consumers.”
There is little question that renewables coupled with storage and transmission can take much of the carbon intensity out of the electricity system – it’s beyond an ideological tit for tat. “The uncertainty is, what’s the pace of change going to be,” she says. “Clearly, to some degree at least, policy will have some role to play in how quickly we see that transition occurring.”
ARENA chief executive Darren Miller tells EcoGeneration the agency is assessing storage technologies on what they can do in the short term and long term. As lithium-ion batteries become cheaper, for example, will their role extend from balancing the system to providing more medium-duration storage of 6 to 8 hours and potentially more. “But we are very reluctant to back only one horse for something so important, so at ARENA we are looking at other forms of long-duration storage,” Millers says, listing solar-thermal, which can provide 12-14 hours of energy dispatch at night time, along with compressed air and hydrogen.
“At ARENA we are very focused on various technology options and we want to move them all forward rather than just attack one type of technology,” Miller says. “The risk of getting it wrong is that we have no options in 10 years’ time.”
Consumers’ role to play
Storage in the scale of hundreds of megawatts can serve a variety of purposes, says Professor Joe Dong, director of the UNSW Digital Grid Futures Institute, including providing ancillary services, grid stability and firming generation from nearby solar or wind farms. Smaller community-scale “grid batteries” will be well-placed at 11kV distribution feeders, to improve the feeders’ capacity to host more PV and distributed energy resources. Owners of residential PV systems with storage will use their batteries to store midday energy to use in the evening peak.
“Distributed energy storage when combined with demand side management can be a really nice resource to help reduce emissions and help households save electricity costs, depending on how evolved the control is,” Dong says.
For VPPs where assets are owned by a retailer-aggregator, residents will enjoy some bill savings but it’s likely most of the financial benefits will be enjoyed by the ones who are pulling the strings – the battery owner.
The slow entrance of electric vehicles onto Australian roads adds another layer of complexity to understanding how the grid will cope with the energy transition. Owners of EVs may be savvy and charge their vehicles during cheap off-peak periods or they might be oblivious to the grid’s woes and charge them during the evening peak, including super hot days when the network is stressed out.
Looking further into the future, however, Dong is anticipating the possibilities of linking retired EV batteries into powerful storage assets. They will be cheap to pick up, he says, and still have plenty of useful capacity – just not enough to power a car. “The investment return is mostly through the electric vehicle itself, so when the battery cannot be used by the EV you send it back for recycling. Before recycling, EV batteries are still 70% or 80% good, so you can use the remaining value.”
A menu of storage assets big and small will be called on to fulfil different roles, Dong says, with some providing basic energy balancing support or regulation support, others providing emergency support. “They can work together very nicely.”
There will be a boom in storage as there was in solar, Dong says. In hindsight it will be obvious and impressive. As it’s happening, however, it will include enormous risk, rapid successes and sudden failures. Hold on tight.