Three white metal cabinets, about eight feet tall, sit on a concrete pad at La Crema Winery in Sonoma County, California. It’s the kind of infrastructure you’d drive by a thousand times without noticing, like a hot water heater or an air conditioning unit. And compared to the gleaming metal maze of wine-making machinery next to it—air compressors, hoists, bins—the installation isn’t flashy. But the cabinets and their contents are part of an experiment that made headlines around the world this year. Behind the unmarked doors are enormous batteries.
The electric car company Tesla announced versions of these batteries intended for business and homeowners this spring with great fanfare. The La Crema set is a test system, paid for and managed by Tesla, that lets them see how the devices work in the real world. Planned for more than a year and finally started up in January, the batteries charge at night when energy is cheaper and, during the day, when energy is more expensive, discharge to meet the winery’s needs, explains Julien Gervreau, the senior sustainability manager at Jackson Family Wines, the company that owns La Crema.
As Gervreau walks around the cabinets, they emit a series of ominous, heavy clicks. He exchanges glances with Mike Mendenhall, La Crema’s production manager. “That might be discharging,” suggests Mendenhall. “You’ll hear those clicking sounds throughout the afternoon as the batteries go back into the grid.” The first time he heard it, he thought there was something wrong—Tesla assured them it’s normal—and it’s still so new that nobody at the winery feels like an expert yet. But the plain white box clicking to itself on the edge of a vineyard is one representative of a wholesale shift in the way we think about energy.
Our modern power system is a just-in-time affair. If the plants that generate electricity make more than the homes and businesses on the other end are using, that’s a disaster—it can cause serious damage to the devices plugged into the system. If they make less, that’s also a disaster—outages and blackouts can cost economies millions. So the power you need is constantly being made just as you use it. Over and over again, molecules in fuel oil are ripped apart—or water pushes a turbine or energy of a light particle is harvested—and electrons flow into the power grid at exactly the same moment as you are pulling them out.
There’s a certain amount of buffering inherent in the grid, which is a patched-together web of wires that crisscrosses the whole continent. The web is tacked down at various points by plants that generate power, and the infrastructure and the flow through it is managed by, among many smaller groups, ten large entities with names like the Electric Reliability Council of Texas and Southwest Power Pool—the one that manages the grid where NOVA is based is called Independent System Operator–New England. The grid can withstand small disconnects between supply and demand, but utility companies must have back-up power sources that can be up and running almost instantly.
It’s the just-in-time of our power system that makes adding in renewables like solar power tricky. Solar may be abundant during bright days, but on cloudy days there’s less of it. And at night it drops to zero, and the slack is picked up by natural gas or other plants that can run regardless of the conditions.
With enough energy to power the whole world for a year hitting the earth every hour in the form of sunlight, and few ways for us to hang onto it to use later, it’s water everywhere, and not a drop to drink. We’re presently standing in the desert during a rainstorm, our mouths open, and not a bucket to be seen.
It’s the just-in-time of our power system that makes adding in renewables like solar power tricky.
As more and more solar power has been fed into grids around the world, utilities are running into problems that didn’t exist just a few years ago. This winter, European grid operators were feverishly preparing to deal with a potential disaster: On March 20, 2015, the moon would pass between Earth and the Sun, preventing some of its light from shining on Europe. With 3% of Europe’s power coming from solar now—and with some countries, including Germany, getting as much as 7% annually, and 35% on sunny days, from it—it was an issue of continent-wide importance. Not only would the missing energy have to be replaced for more than an hour, but the eclipse would cause a drop in production much, much faster than nightfall does, so the switch would have to be executed with extreme grace. In the famous opening scene of The Raiders of the Lost Ark , Indiana Jones deftly exchanges a golden idol for a sack of dirt of exactly the same weight to avoid triggering a trap. The problem facing Europe was similar, except with as much as 30 gigawatts of power—the equivalent of what would be produced by 15 Hoover Dams—hanging in the balance.
Preparations at the level of the European grid began more than a year before the eclipse. The plan involved placing orders for additional electricity to be generated at hydroelectric and other plants to make up the difference. And in the end, that strategy—scheduling more electricity generation, carefully feeding it into the system, and removing it when the sun reappeared—worked for the European grid. They managed to maintain the same frequency of electrons coursing through the grid before, during, and after the eclipse.
But whether using stored energy—taking a small step away from the just-in-time system, which in everything is used just as its made—could be part of the solution was something Michael Koller wondered about. He is an engineer at EKZ, the local grid operator and utility company in Switzerland’s Zurich canton and owner of the largest battery installation in the country. It is an experimental set-up that they have been using and studying for three years, trying to see in what situations it would be helpful.
“There’s a variety of services that they can provide,” Koller says. “We think there’s a lot of potential, and we wanted to understand how it works so we can decide on a case-to-case basis when it is better to use storage or use conventional generation.”
EKZ’s batteries are available for hire to help stabilize disconnects between supply and demand. There was no need for them during the eclipse, in the end. But earlier this year, Koller is pleased to say, they were booked to step in when a nuclear power plant had to go offline. On a chart showing the power pouring through the grid, as the nuclear plant checks out at 4:45 am, you can see the batteries ramping up, helping keep the supply consistent.
It was a relatively small triumph. But in the future batteries could work in other, broader applications, Koller says, the most obvious and appealing of which is storing solar energy.
Last year the financial services company Barclays downgraded electric utility bonds, saying that thanks to the increasing cost-effectiveness of photovoltaics and the rise of storage, it was no longer such a safe investment to bank on the growth of traditional utilities. “In the 100+ year history of the electric utility industry, there has never before been a truly cost-competitive substitute available for grid power,” the report noted. “Based on our analysis, the cost of solar + storage for residential consumers of electricity is already competitive with the price of utility grid power in Hawaii. Of the other major markets, California could follow in 2017, New York and Arizona in 2018, and many other states soon after.”
Back at La Crema, glistening solar panels coat the facility roofs, and the money generated from selling that energy back to the utility company offsets up to 65% of the winery’s annual electric bill. At the moment, similar to what’s happening at EKZ, the batteries are being used to add a tiny bit of buffer to the grid. They perform a service called peak shaving, which means that when utility customers are using a lot of power from the grid, the winery starts drawing on the batteries for its needs instead. That lowers accordingly the amount of power that needs to be produced at that moment. Though the seasonal nature of the winery business would make it difficult—most of their electricity is used in just a few months in the fall, during the harvest—Gervreau is curious about using the batteries to power the entire winery directly from the sun, rather than indirectly. “If we can take a winery off the grid, that’s really the name of the game,” he says.
Chance are, you’re probably not planning to go off the grid anytime soon, despite Barclays’ observation about the growing affordability of that option. But there still might be a role for batteries like the ones at La Crema and EKZ in the management of your electricity because the just-in-time system has costs beyond handling eclipses or solar power.
Depending on where you live, the local utility may not generate the power you use itself. In Wellesley, Massachusetts, not far from the NOVA offices, the Wellesley Municipal Light Plant’s headquarters is a small brick building next door to the town fire department. Dick Joyce, the director, explains that for years now the Light Plant has been concerned with brokering deals with power plant companies to keep the town up and running rather than generating its own power. The plans made there for Wellesley’s power are a network of financial transactions extending as far as five years into the future, making arrangements for, say, 14% of the electricity needed in January 2017 to come from a hydroelectric plant in Canada, or 2% to come from wind, and so on.
A large fraction of what the town has to pay for this power is decided every year in the summer. “Either July or August is when the grid in New England has its peak demand,” Joyce explains. Everyone has their air conditioners going, and ISO–New England is pulling down an enormous amount of power to keep supply at the same level as demand. But, by design, using grid power during the peak is extremely costly. Whatever the town is using at the moment when the demand peaks for the year—probably about 4 pm or so on a late-summer day—they will pay a hefty penalty in proportion to it every month for the rest of the year. Reducing how much they have to rely on the grid during peak demand is something Joyce and his colleagues at the Light Plant are very interested in.
By design, using grid power during the peak is extremely costly.
A smart grid, one that automatically takes power-intensive devices offline during peak moments, can help with that. The growing efficiency of many devices helps as well. But storage could have a significant impact. If houses in Wellesley were attached to batteries, the Light Plant could plan for the moments when they know peak demand is coming—say, 4 pm on a July afternoon on the second or third day of a heat wave with the temperature creeping above 90 degrees. At 2 pm that day, they could flip a switch and start drawing from the batteries until 6 pm when things cool down and the demand falls off, thus limiting the amount they have to pull from the grid and the amount they have to pay all year in penalties. “You would only have to run them a few hours, maybe five days a year,” Joyce says. “But it would save so much money.”
Joyce has been following storage developments, including Tesla’s announcement this spring. “It’s interesting.” He pauses. “We’ll see.” To be useful to the Light Plant, each battery must have the capacity to take a whole house offline. He’s seen some that can only manage a sixth of what a typical Wellesley house needs, and he’s also not so sure how reliable they are. Tesla’s smallest battery for home use is advertised as being able to produce 5 kW, however, which might be enough capacity, so Joyce is interested to see how it fares with early adopters. “If that works, that would be fantastic.”
Should the batteries stand the test of time, the town would probably start offering a rebate to people who wanted to buy them or equivalent products, much as they have done for solar panels, and instruct them to charge them during the night. Perhaps the town itself might invest in an installation of some kind, somewhere down the line, when the technology is mature. The long-term effect would be that instead of a jagged curve with peaks and troughs, the town’s energy demand, and energy costs, would even out.
“That’s where everybody is going now,” Joyce observes. “You focus on the off-peak, smooth out your demand. Even if it was electricity at night that recharged those batteries, rather than solar, it doesn’t cost us a lot of money.” Looking at the long view, you could imagine a future with enough storage that the same amount of power is generated year-round in may places. Battery storage would provide a way to capture solar efficiently, but it would also make the rest of power generation a less-erratic, less-costly process.
The value of energy storage, beyond just batteries, is such that while it’s fairly new at the level of consumers and small utilities, bigger utilities have been exploring it for more than a century. They have built large-scale storage installations that have been used as back-up power in case of a plant going offline or demand changing suddenly. These installations make use of other technologies, like pumped-storage hydropower.
At one such plant at Northfield Mountain, in Western Massachusetts, visitors must drive a half-mile through solid rock to reach a subterranean chamber where four enormous turbines wait. At Northfield, as in other pumped-storage plants, the operators pump water uphill to a reservoir when electricity is cheap. Then, when costs go up and demand is high, they let it run downhill to spin the turbines, analogous to mill wheels, that convert that energy into electricity. The turbine chamber intercepts a 1,100-foot shaft running from the upper reservoir, which can hold 5.6 billion gallons of water, to the lower, which is 20 miles of the Connecticut River. Sometimes the energy stored there is held as back-up for the ISO–New England-managed grid, which is required to have a certain amount held in reserves in case another power plant goes offline. Sometimes the plant’s owner, GDF Suez, a French multinational utility company, sells the power directly.
Pumped-storage is an old idea—“A Ten-Mile Storage Battery,” crows a 1930 Popular Science headline about a plant in Connecticut, referring to the length of the upper reservoir—and it’s fairly efficient: It can return between 70 and 80% of the energy used to pump the water up. (For comparison, hydrogen fuel cells have about 40% efficiency. Tesla, in turn, claims its battery has 92% efficiency. Some owners of Tesla cars, which also bear the batteries, suggest that the average is closer to 85%—still very good.) And it’s widespread: In 2012, more than 99% of the world’s storage capacity was from pumped water installations.
But you can’t build a pumped-storage plant just anywhere. You must have two enormous reservoirs and a sudden decline in altitude between them, which often means a dam with its own slate of environmental side-effects. They are also not necessarily the most cost-effective way of dealing with the erratic changes of solar and other renewables, suggests Todd Strauss, senior director of energy policy planning at Pacific Gas and Electric (PG&E), one of the largest utility companies in the United States. It’s worth looking at other options, including batteries.
PG&E has 5.4 million electrical customers—including La Crema Winery—and in recent years, the number of customers with photovoltaic solar installations has climbed to 2–3%. That has changed the pattern of demand on the grid significantly. Now, some of the lowest net demand times are between noon and 3 pm—even though, paradoxically, that’s when a lot of power is being used—because of all the solar panels producing power. It means that PG&E needs to produce less power on sunny days than used to be normal. “There is a much higher uncertainty now,” Strauss says. “This is very different from ten years ago.”
To deal with this change, power plant operators can alter them so they produce less during low demand times. PG&E can also encourage people to use more power during hours when solar power is streaming in, perhaps shifting the best charging hours for electric cars from nighttime to noon. Batteries or some other kind of storage, such as flywheels, could be another way. In 2010, the California government passed a law requiring that utilities look into smaller-scale energy storage solutions, set to be operational by 2024, in part to encourage innovation by tech companies.
PG&E currently has two test battery installations, one near a solar farm and another out on a spur of the grid in the San Jose hills. Using them safely and effectively is not as simple as it might seem. For instance, batteries can be a significant fire hazard. “We’re still in the midst of ironing all those things out,” Strauss says.
In the end, PG&E is interested in storage solutions that combine safety, reliability, and cost-effectiveness, Strauss emphasizes. He says it’s most likely that there will be a suite of solutions rather than a specific technology to meet their goals.
Will batteries be a part of that suite? “We don’t know,” he says. Considering the case of photovoltaic technology, where the cost of an installation is now a quarter of what it was ten years ago, he suggests that if batteries can have a similar drop over the next decade, that would be a game-changer. (The economics of energy are usually summed up in terms of how cheaply a kilowatt-hour, or enough to power a 100-watt light bulb for ten hours, can be generated. Per-kilowatt-hour that they release, batteries, which must be charged with power that’s already been generated ahead of time, are still costly.)
An analysis released in January by Citigroup suggests just how precipitously cost must come down. If batteries are going to be used by utility companies, they need to fall below the per-kilowatt-hour cost of pumped-hydro storage to be attractive. However, the analysis suggests that a sufficient drop is likely to happen in the next seven to eight years. After that, Citigroup predicts the growth in demand will help drive the costs even lower.
Another factor to take into account is that the energy industry in general is remarkably conservative, says Koller, the EKZ engineer. “There is a very long time lag in the energy sector,” he says. “It took about 40–50 years from the discovery of crude oil until wide adoption. It’s the same for natural gas; it’s the same for electricity, actually.” In part, that’s because reliability is incredibly valuable—when an area does not have a reliable grid, businesses will flee. There are good incentives for sticking with what works.
Still, little by little, the energy world is changing. PG&E is experimenting with storage technologies that make batteries sound prosaic. Earlier this month, the California utility completed preliminary tests near Lodi investigating the feasibility of a facility that would store energy by pumping air into immense caverns using electricity from natural gas and let it flow out to spin turbines that generate electricity when demand is high, just as in a pumped-storage hydro plant. If built, it would be only the third large compressed air storage installation in the world, and the first new one to come online in decades.
Strauss says that PG&E will soon put out a call for companies interested in building the plant, which will let them assess whether the storage will be worth the financial investment. “That’s the big theme with all of the storage stuff,” he says, from his point of view. “At the end of the day, what does it cost to get it to work in the field?”
Out on the edge of the vineyard, back in Sonoma County, how it works in the field is what’s on Gervreau’s mind. Jackson Family Wines has 21 white metal boxes spread out across their wineries, and in the five months or so of operation, they seem to have worked well. He can watch the batteries’ behavior on a computer screen, with red and blue lines zigging and zagging across each other as the batteries charge and discharge in accordance with the wineries’ power needs. An algorithm controls the devices, and the plan is that it will grow more and more adept at anticipating what they should be doing at any given moment. When the batteries began to click earlier in the day, a look at the system’s monitor confirmed that they were indeed discharging. But it was in response to a sudden leap, maybe an air compressor coming on, and they stopped discharging soon after.
Gervreau looks up at the shining array of solar panels on the roof. He, like Strauss, is intimately familiar with the drop in the cost of photovoltaics and wonders whether batteries might see something similar. Jackson Family Wines did not pay for the batteries, nor do they manage their upkeep, so they did not have to weigh their costs against the benefits before installation. But he thinks that Tesla’s influence—their ability to build large numbers of batteries, create a demand for them, and sell them at the relatively inexpensive price of $3,000-$3,500—will catalyze a similar change in the economics for many people. “They are creating a market, and that’s a big first step,” he says.
As far as the dream of getting a winery off the grid, that would require, judging from a back-of-the-envelope calculation, at least double the battery capacity they currently have installed. In the meantime, he hopes to see what other tricks the batteries are capable of, in addition to peak shaving. In particular, it will be interesting to see whether they can keep the winery running during times when PG&E sends out alerts asking that certain companies stop pulling grid power in order to keep supply and demand balanced, in return for a fee. Sometimes they’re given as little as 30 minutes’ notice, and the day was forecast to be hot, with air conditioners coming on. Will storage help us move away from such a just-in-time system and towards a system where we make the most of our existing electricity infrastructure, as well as renewables?
Gervreau and Mendenhall keep checking their phones to see whether today is one of those days. “It’s a brave new world,” Gervreau says. “We’re all just figuring it out."