Second in a two-part series. Click here to read part one.
At 2 p.m. on a sunny afternoon in June of this year, solar power in California surged to a new milestone. Over the course of an hour, photovoltaic panels and other arrays generated more than 2,000 megawatts of electricity -- a record for the Golden State and for California ISO, its systems operator.
Seven hours later, as the sun set into the Pacific Ocean, that output fell to zero.
Intermittency, or the inconstant supply of wind and solar power, represents one of the most difficult roadblocks to the world's deployment of renewable energy. While these technologies can generate significant amounts of power under ideal meteorological conditions, their output tends to be as fickle as the weather.
At their current levels, today's renewables can easily be complemented by the baseload power plants -- coal, nuclear and natural gas -- already on the grid. In grid lingo, base load means they're meant to be the foundation, the major source of electricity. The problem is that this legacy infrastructure was not designed to mingle with renewables' intermittent supply levels. Scaling conventional thermal power plants up and down on an hourly or daily basis is economically inefficient, experts say.
The device that has traditionally assumed this chore is called a peaking plant, usually a diesel-powered backup generator. "The peaking plants in operation today" -- plants kept on standby, to be brought online only when demand peaks -- "are already inefficient because they only sell electricity for around 100 hours a year," explained Trieu Mai, a senior analyst and supervisor at the National Renewable Energy Laboratory (NREL). "It's already hard to find revenue for these plants, and with high levels of renewables, that would be increasingly challenging."
Keeping baseload power in the game
Proponents of traditional baseload power argue that its services will still be valuable in the future, whether producing electricity or performing other tasks in an integrated, low-carbon economy.
Nuclear power plants, for example, could back up renewables during periods of intermittency, said Paul Genoa, senior director for policy development at the Nuclear Energy Institute (NEI). Then when wind and solar power come back online, they could use their high operating capacity to perform other useful tasks that store their energy output rather than waste it or turn it off, such as hydrogen fuel production, he said.
"If we're really going to de-carbonize, we have to go beyond just the production of electricity," he said, noting that 40 percent of power in the United States is consumed as electricity.
Both Mai and Genoa say that markets will have to better reward capacity, or the assurance of standby power, if a renewable energy system is going to emerge.
Any serious challenge to the baseload model -- which has dominated throughout modern industrial history -- is decades out. But that hasn't stopped operators and engineers from exploring novel solutions.
Power when you want it
One solution may lie in power plants that "follow" load, specifically designed to ramp up generation when needed and reduce it when renewables resume their output.
Hydroelectric power plants were the original case studies in load following, able to increase or decrease power by gigawatts or tens of gigawatts each minute depending on the reserves of water available to them. In the United States, however, much of the country's hydroelectric potential has already been tapped.
And although some nuclear technologies do follow load -- as is the case in France, where the high penetration of nuclear energy necessitates some flexibility to meet changing demand -- they do not represent the most efficient use of nuclear's high capacity factor, Genoa said.
In California, which aims to get a third of its energy from renewables by 2020, systems operators are trying a different approach. The California ISO recently replaced two units of its El Segundo power plant with highly efficient, "flexible" gas-fired units capable of quickly scaling up and down to complement renewables.
Each of the new units, which were designed by Siemens Energy Corp. and are operated by NRG Energy Inc., can go from zero to 150 MW in the span of 10 minutes, and they have a combined maximum output of 550 MW -- enough to power 400,000 homes.
El Segundo is no prototype. It was designed to provide a return on investment and operates profitably despite its intermittency, said Bonnie Marini, a director of marketing for Siemens Energy.
And although there are only two of these "flexible" power plants -- both in California -- currently in operation in the United States, more are on the horizon, she said.
"We've seen that when it comes to renewable energy, California tends to be a first-mover," she said. "The next region that's seriously looking at flex plants is Texas, where there are significant amounts of wind energy. Really, you're looking at anywhere where large amounts of renewables are coming online."
In Europe, flex plant technology has been in operation since 2009, she added.
An antiquated notion?
Some analysts have even posited the idea that baseload power could be phased out on a grid "smart" enough to quickly redistribute the intermittent flows of renewable energy.
"Base load is a 20th-century concept. Heck, it's a 1950s concept," said Mark Cooper, a senior research fellow for economic analysis at the Vermont Law School Institute for Energy and the Environment. "We're going to have a system that uses intelligence, communication, micro control. We are living through a third industrial revolution" -- with digitization and communications technologies standing in for the steam engine -- "and not a bit of it has made its way into our electrical system. But it will."
An integrated system that balanced renewables over geographic areas -- with some load-following gas plants -- would be far more economical than keeping baseload power online, he said.
Utility operators have largely been skeptical of this viewpoint, arguing that such a system would require a vast overbuild of renewables and create unnecessary redundancies.
More likely, NREL's Mai said, is a balance between traditional and novel systems. An oft-cited study by NREL, of which Mai is co-author, found that the United States could potentially scale up to 80 percent renewables in its energy mix by 2035, given continuing improvements in technology.
At those levels, there would be far fewer baseload power plants in operation, he said. "During the summer, when demand is high, they'd be operating at baseload capacity," he said. "But during the spring, they'd be operating intermittently -- either on again, off again, or on a daily cycle."