by Julian Spector
August 31, 2018

Futuristic discussions about cleaning up the grid got a lot less theoretical this week when California lawmakers passed a 100 percent clean energy requirement for 2045.

The legislature approved the bill Wednesday, sending it to the desk of Governor Jerry Brown, an avowed climate hawk. He hasn't signed it yet, but I’m writing this column based on the assumption that the bill becomes law.

That leaves the overwhelming question of how California should actually accomplish this. And that’s a question the energy storage industry is best suited to answer.

It’s pretty clear now that states can ramp their renewable portfolio standards toward 50 percent without blowing up their grids, even though that was unthinkable for many just a few years ago. At the same time, wind and solar cost declines and deployment numbers continue to beat expectations.

Going from a high renewable penetration to eliminating fossil fuels, though, takes the state into wholly uncharted territory.

Nobody knows the answer here, but that’s OK, because the law allows 27 years and an uninhibited range of carbon-free technologies to figure it out.

Put another way, it’s theoretically possible to power the grid without fossil fuels, so the more pertinent question is how to do so in a cost-effective and minimally redundant manner. We could install enough solar power to meet demand on even the cloudiest days or during the darkest eclipses, but bankrupt clean energy pioneers don’t make for good role models.

To succeed, California must strike the right balance between cheap and plentiful renewables serving bulk generation and enough dispatchable resources to ensure power availability round the clock. In this installment of Storage Plus, I examine currently available data sets to sketch out a vision of how this mightily ambitious transition will unfold.

Shut down the gas

Renewables excel at providing cheap and abundant energy. The most challenging part of the transition will be clean capacity, ensuring the grid can deliver energy when and where it is needed at all times.

The most jarring change, then, will be California weaning itself off of natural gas, which has played the vital role of balancing the ups and downs of solar and wind. That fuel currently provides 34 percent of the state’s electricity, a reminder that gas still delivers bulk energy, not just peak power.

Some experts I spoke with believe the text of SB 100 could allow for some gas burning, provided clean electricity nets out dirty electricity over the course of the year. If 100 percent of the annual load comes from clean sources, then additional generation beyond 100 percent could be dirty. That strikes me as contrary to the intent of the law, but I’ll leave the final word on that to legal experts.

Assuming that the era of gas in California must end, figuring out an orderly retreat will be a crucial policy goal of the coming decade. We can peer into the future with the help of new research from the Union of Concerned Scientists, which modeled an “orderly and equitable” phase-out of the 89 gas peakers and combined cycle plants in California ISO territory.

“For California to realize the benefits of its clean energy transition and achieve its global warming emissions targets, it needs to reduce its dependence on natural gas electricity generation significantly,” UCS researchers Laura Wisland and Mark Specht write. “This transition should prioritize reducing natural-gas generation in communities most negatively affected by the pollution resulting from burning fossil fuels and by the social, economic and health burdens associated with global warming.”

Wisland and Specht model the economic retirement of gas plants, while still satisfying local capacity and planning reserve margin requirements. A plant retires if it becomes more cost-effective for the system to procure a new resource to meet grid reliability. In reality, numerous out-of-market forces can keep a plant alive longer than it should be, but the model represents what prudent and un-politicized grid investments would look like.

Local capacity requirements play a significant role in whether or not gas plants face economic retirement. That said, the analysis found that 24 percent of peaker capacity and 23 percent of combined cycle gas turbine (CCGT) capacity could retire without any risk to grid reliability — that’s 28 of 89 plants. Even more striking, the economic model would retire those plants this year, rather than waiting and continuing to pay the costs of operating them.

Consider that the fat in the grid steak, or the dead wood in the grid underbrush, or the lousy supermarket sugar cookies that one can easily cut out of a low-carb grid diet. California built out a number of plants in the ‘90s and early 2000s, before renewables and the shale revolution made energy much cheaper. The system has more than it needs, and that’s the no-brainer place to start retiring.

Regulators have already begun this process, first by pumping the brakes on redundant new gas plants, and by ordering PG&E to buy batteries instead of paying to keep open three Calpine gas plants. To fulfill that order, PG&E is procuring the world’s largest battery at Moss Landing, a 300-megawatt sign of things to come.

Whether that behemoth really saves ratepayers money compared to propping up uneconomic gas plants will say a lot about the prudence and timing of this transition.

Retirements accelerate in the UCS scenario where California must meet a cap of 30 million metric tons of CO2 equivalent — the more aggressive climate policy that the California Public Utilities Commission plans for — and hour batteries compete to provide local capacity requirements. Those conditions see 87 percent of peaker capacity and 30 percent of CCGT capacity retire by 2030.

The model replaces those capacity retirements with a portfolio that contains 12.4 gigawatts of solar, 4.3 gigawatts of wind, 5.9 gigawatts of 4-hour storage, 1.5 gigawatts of geothermal and 1.1 gigawatts of pumped hydro.

Image credit: Union of Concerned Scientists

Given that 300-megawatt battery plants are already underway, the prospect of getting 6 gigawatts online with more than a decade of lead time seems well within the realm of possibility. Crucially, this is calculated to cost less for the system than keeping the gas plants open.

In sum, according to the study:

  • California can retire roughly one-quarter of its gas capacity today without jeopardizing grid reliability. There’s really no reason not to do that (setting aside parochial concerns of local economic benefits from the plants).
  • If California pursues the more aggressive climate change mitigation policy of 30 MMT, 4-hour storage can pick up the slack and let almost all gas peakers retire by 2030, although 70 percent of combined cycle capacity will remain. That leaves 15 years to figure out a replacement for the few surviving peakers and the bulk of the CCGT fleet.
  • Other studies may come to different conclusions about how many plants can retire economically by what time. This should be a growing area of research as the state proceeds with the new objective.
  • If battery prices continue to beat expectations, they could provide an even more compelling alternative within the 2030 timeframe.

Renewables take over, but then what?

Crucially, the bill requires 100 percent carbon-free energy, not 100 percent renewable energy.

I analyzed the case for the more permissive target back when the bill was first proposed. At the time, my critique of 100 percent renewable energy earned me public accusations of being a “nuclear shill” from certain members of the solar industry intelligentsia. Senator Kevin de León ultimately chose the wiser option, and last I checked that has not irredeemably spoiled his reputation as a clean energy crusader.

The problem is, we don’t have many alternative energy sources that appear likely to compete with renewables for clean grid market share.

New large-scale dam building is off the table due to the environmental devastation it has wrought. Select pumped hydro storage facilities, in reservoirs that don’t disrupt river ecosystems, could move ahead with this bill as a driver.

Geothermal could play a bigger role on the grid. Carbon capture, or the intriguing emission-free gas technology being developed by Net Power, could come online in the next 27 years, but we aren’t there yet. Small modular nuclear reactors might achieve licensing in the 2020s, but at least some of the old-guard California environmentalists will still be around to oppose their carbon-free energy.

“In a decade, we could have a lot of zero-carbon on-demand options that can displace the remaining unscrubbed gas in the system and mitigate the costs of variability,” said Armond Cohen, executive director of the Clean Air Task Force. “The bill wisely keeps all options, and at the same time, also provides a clear incentive to bring forward innovation in new zero-carbon dispatchable technologies."

Until those tools develop, the current likely trajectory sees cheap renewables grow and grow until they trigger some difficult investment dynamics.

“The value of diversity will increase as saturation of the system from a single resource increases,” said Arne Olsen, a senior partner at Energy and Environmental Economics, who led a 2013 study on how California’s largest utilities could meet the 50 percent renewables target. “The more solar we add, the more valuable non-solar alternatives will be.”

Vast amounts of midday solar will make additional solar investments less attractive, and create more demand for anything that produces at other times of day, or can shift that solar generation to other times.

The super-high renewable grid has the hardest time in winter, when California’s solar productivity dips, Olsen said. Conversely, the summer sun brings large surpluses that have to go somewhere or be tossed away through curtailment.

Expanding the regional grid with a full day-ahead market would “help enormously” in dealing with surplus solar, Olsen noted. That would bring in the solar resources of the desert Southwest, which come online before the sun rises on the West Coast, as well as wind resources in the Rockies and hydropower in the Pacific Northwest.

“The more that you expand the horizon over which you optimize the grid, the more diversity you have in terms of resources [and] load,” he said.

Too much of a good thing

The swings from surplus to deficit will be extreme, said Wade Schauer, director of Americas research at Wood Mackenzie Power and Renewables. He ran a “back of the envelope” model of the CAISO system operating with 100 percent renewable generation in 2045, assuming no load growth.

That preliminary analysis showed that, from April through June, the average hourly generation surplus is 4,400 megawatts; there are 1,676 hours when generation exceeds load and 508 hours when generation is less than load.

From July through August, the average hourly deficit is 1,591 megawatts; there are 1,344 hours of generation less than load and 840 hours of generation greater than load.

The average hourly deficit is even greater from October through December, at 1,971 megawatts; there are 1,439 hours of generation less than load and 745 hours of generation greater than load.

This conundrum suggests three possible coping mechanisms within California.

The grid could vastly overbuild renewables so that it still produces enough during those deficit months. That would be a hard sell economically speaking.

The clean energy industry could make things easier by supplying storage technology that can take the surplus from April, May and June and spreading it out over the deficit months. Form Energy is trying to do that with super-cheap long-duration batteries, but there’s no guarantee that its approach will succeed, as the founders freely admit.

By passing this bill, California has created the best incentive yet for patient capital to invest in other super-long-duration technologies. This needs to happen to make an affordable 100 percent clean energy grid possible.

Looking beyond California, the state could export some of its surplus and import during the lean months, but Schauer is skeptical that a regional grid will be much help.

“The Pacific Northwest is already in surplus from April through June (hydro) and historically ships that to California,” he wrote in an email this week. “As Nevada, Utah, Arizona and New Mexico ramp up their renewables, they will also be surplus (at least in April and a little bit in May) and especially during the 1 p.m. peak solar output time. So, California won't be able to dispatch its surplus into those markets in 2045, and the duck curve will be across the Southwest, not just in California.”

The irony of California’s leadership on solar would be that its neighbors all enthusiastically follow suit, and everyone has surpluses at the same times. This is more of an economic problem: Curtailing is easy, but how do you continue investing in solar plants when they will be forced to throw away much of their production?

The deficit months pose a practical and economic problem, which is how to secure enough capacity without burning carbon.

Big X-factor: Taking control of demand

Everything else about modern society will change during this timeframe, which means help could come from the demand side of the equation. Ideally, changes in how we consume electricity would lessen the peaks and valleys of supply that the grid must serve, making the job easier.

Homes and businesses are already gaining exposure to time-differentiated rates, which shift consumption to times when the grid has an easier time supplying it. That should absolutely continue, perhaps into the realm of real-time pricing, which would be much more granular than today’s time-of-use rates.

Locating more generation and storage at the point of consumption further lessens the burden on the bulk power system. California’s building code will start requiring solar on most new homes in 2020, and the customer-sited storage industry just got a boost from the legislature’s decision to extend the Self-Generation Incentive Program five more years.

“In order to get to that 100 percent, if you want to create the most efficient system, you need to look at getting clean energy generation closer to where the energy is consumed,” said Alex McDonough, vice president for public policy at residential solar and storage company Sunrun. “You avoid the tremendous cost of the transmission and distribution system, and it brings customers into the solution, makes them part of the solution.”

The projected rise of electric vehicles offers a delightful opportunity to soak up surplus solar generation, assuming drivers are willing to adjust their charging schedules based on the dynamics of grid operation. That comes with a cost, though: What happens when that massive fleet of cars needs to fill up in the deficit months?

These demand-side resources are much further along than month-long storage, but it’s hard to know how they’ll develop in the next few decades. I’d count them as extremely promising tools for California’s transition, but it’s up to industry and utilities to deliver on that promise.

The last few thousand words should dissuade anyone from thinking this will be easy, be they nuclear shills or otherwise. That said, nobody has yet tried in earnest to make this happen; a legislative mandate can be a wonderful tool for clarifying one’s focus and dedication.

Really, that’s the point of moonshots, said Daniel Finn-Foley, an energy storage analyst at WoodMac Power and Renewables.

“The whole point of a moonshot is that you might not know how you'll do it when you set it, but you set it boldly anyway,” he said. “And people tend to look down on moonshots, saying, 'Oh, they're unrealistic, how will they be implemented, the technology isn't there, it's too expensive,' etc. But they also tend to forget: We landed on the moon. It was unrealistic and expensive and required new technology and implementation, but we did it. California can hit 100 percent too.”

The time for debate about possibility has passed. The law is the law, intimidating and invigorating as it is. Now California, after carving an improbable civilization out of arid desert, amid fires and earthquakes and floods and mudslides, simply because 40 million people would rather live there than anywhere else, must prove merely that it can exist without polluting itself just to turn the lights on.