by Julian Spector
August 08, 2018

Energy storage has established its role on the grid, but only preliminarily.

Even the wildly successful Aliso Canyon batteries, which stepped in to serve local capacity after a massive Southern California gas leak, look tiny compared to a small gas peaker.

The bigger projects are coming, though, which means it’s not too early to consider where this ship is headed. We'll spend this week’s Storage Plus exploring the storage plant of the future. Once the growing pains recede and the cost curves come down, what will business as usual look like?

Bigger and longer-lasting, in a nutshell. Storage plants will look and function more like full-fledged power plants, especially in regions with aggressive renewables targets. But that simple directionality leads to several follow-on effects in terms of design and function.

Here we’ll examine how a typical storage facility might look in 2030, after the asset class has been normalized on the grid for more than a decade.

Plant design and function

The biggest battery in the U.S. right now is the 30-megawatt/120-megawatt-hour Escondido system that AES built for San Diego Gas & Electric's Aliso Canyon response.

That battery — actually a barracks of parallel containers, sheltering racks of batteries on a patch of asphalt outside San Diego — marked an early victory in the storage industry’s efforts to be taken seriously. The Southern California grid needed peak capacity in dense urban areas, and fast; the industry delivered a half dozen battery systems within half a year.

Still, 30 megawatts can only get you so far. That’s a small fraction of a typical gas peaker plant’s power capacity, and Escondido can only keep that up for four hours.

That’s not going to suffice in a world where intermittent renewables provide much of the grid’s electricity, producing steep rises and falls in net demand throughout the day.

That vision is not just theoretical: Numerous other states are ramping their wind and solar production, and California could soon pass a law requiring entirely zero-carbon electricity by 2045.

Such a grid will require energy storage to lift a heavier load in supplying flexible capacity, and that means bigger and longer-lasting battery plants.

We haven’t yet seen the upper bounds of what the industry can do.

Tesla’s record-holding South Australia plant (100 megawatts/129 megawatt-hours) wasn’t the product of diligent development planning so much as a breathless 100-day sprint. Back in 2014, AES won a contract for a 100-megawatt, 4-hour system for Southern California Edison that won’t come online until 2021.

New giants are already starting to emerge: PG&E’s Moss Landing project will replace several gas peakers with battery installations in the South Bay region. The largest will be a 300-megawatt/1,200-megawatt-hour battery.

Such scale makes certain demands on plant design. For instance, we will see a shift from the containerized architecture, seen at Escondido and elsewhere, to enclosed building designs for larger projects, said Mitalee Gupta, a storage supply chain analyst at GTM Research.

This stems from the mathematics of surface area: At a certain point, one can reduce construction expense by enclosing a large amount of battery cells in one large space, rather than many small metal boxes.

Metal containers will stay prevalent in smaller systems, she noted. They also provide greater flexibility, in case a customer wants to relocate a battery after its original deployment.

The economics of solar-paired storage will shift as well. The federal Investment Tax Credit, which rewards batteries that charge from co-located solar, will recede. For solar developers, there will still be value in adding on batteries as a way to firm up solar generation.

For storage developers, locational value may well drive projects toward load pockets where there simply isn’t space to add solar as well. As long as surplus renewable generation is on the grid, it doesn’t much matter where the batteries sit that charge from it. Location does matter for discharging at times of peak grid constraint, so that side of the equation should take precedence.

Technology: Newcomers or more of the same?

At present, lithium-ion technology appears capable of maintaining its dominance well into the 2030s.

That still allows for variation within the lithium-ion family of chemistries. Currently, nickel-manganese-cobalt leads the pack, thanks to its early prominence in the electric vehicle space.

Companies are already racing to minimize cobalt in their recipes, however, to reduce material cost and insulate themselves from reliance on that rare earth metal’s supply chain.

Lithium-iron-phosphate, NMC’s safer, more staid cousin, has also been gaining traction recently, as the market seeks an alternative to supply-constrained NMC cells. Gupta expects that trend to continue.

Outside of that, the prediction game grows perilous.

It’s entirely possible that by 2030, some of the long-duration alternatives to lithium-ion storage will be ready for wide-scale deployment. Several flow battery companies still stand, as does Ambri’s liquid metal battery and Fluidic’s zinc-air device.

If any of these can remain funded long enough to prove their technology, win some major utility contracts and ramp their manufacturing capacity, they could chip away at lithium’s market share, starting in cases where the longer duration proves especially valuable.

Such use cases have been few and far between, but should be eminently clear by 2030, the deadline for California to achieve 60 percent renewable electricity under the new bill.

In a category of its own, Form Energy hopes to solve uber-long duration storage, on the order of weeks to months. Such a breakthrough will be necessary to make use of sunny-month surpluses during the darker months of the year. That startup hopes to prove out two chemistries in the next five years, with deployment coming in the following five, if it works.

There may even be some new pumped hydro to speak of.

That most prevalent form of grid storage hasn’t grown in years due to its considerable environmental and siting challenges. But, the New York Times reported, the Los Angeles Department of Water and Power has been investigating a pumped-hydro retrofit on the Hoover Dam, which would allow it to absorb massive amounts of solar generation.

Recent experience suggests the only safe approach to lithium-ion alternatives is to treat them as impending bankruptcies until proven otherwise. Given the diversity of options, though, it’s hard to imagine none of them will break through, especially once there are mature market opportunities to pursue. A lot can change in a decade; imagine going back to 2008 and pitching PG&E on shutting down a modern gas peaker and replacing it with 300 megawatts of batteries.

“Lithium-ion will definitely dominate the market, but we will see other technologies gain market share,” Gupta predicted.

Where will these battery behemoths go?

So far, most utility-scale batteries have jumped into open spaces in existing substations. A utility-owned parking lot can upgrade to a storage facility with little structural improvement. But that low-impact approach won’t work for the massive storage plants to come.

Again, Moss Landing points the way: Dynegy subsidiary Vistra will install those 300 megawatts inside the turbine building of a Dynegy gas plant with two mothballed generator units.

If California follows through on its zero-carbon law, all remaining gas plants will have to retire. If you add up those facilities with already shuttered coal and gas plants, storage developers should have plenty of attractive sites to keep themselves busy.

This approach minimizes interconnection issues, because the sites are already designed to ship generation out through the transmission grid. Backfilling land already set aside for electrical production also avoids competing for land that could go to other uses (like plastering enough solar on the landscape to fulfill a legal mandate).

Trendsetter projects to watch

A few projects have been announced or built that will have an outsized impact on future design choices. Here are three to look out for to understand where the sector is headed.

First Solar's Battery Peaker: This project shows how to beat out all other options for peak capacity by taking cheap desert solar power and feeding it into a 50-megawatt battery, in a state with no policy drivers tilting the scales in favor of storage. This model makes sense for sunny areas with enough space; don’t expect solar popping up next to batteries in the L.A. Basin. Stay tuned to see if First Solar makes any money on its aggressive early entry.

Moss Landing: Can truly massive batteries replace gas plants, keep the grid up and running, and save ratepayers money at the same time? PG&E aims to find out, at the behest of California regulators. This will be the one that sets the model for retrofitting storage into decommissioned power plants. What’s degradation look like in a 300-megawatt peaker? When do the batteries need replacement? How does the plant fare in a prolonged heat wave?

Tesla's Hornsdale Plant: It’s incredible that this plant actually came together, but Elon Musk’s Twitter bet prompted his engineers to do something valuable: build a merchant battery that plays in a range of market products, balancing operational costs and revenues to determine the economic optimum. That intellectual achievement is already inspiring followups, and more will follow as market rules begin to pay for the full range of services storage can provide.