It’s not often that we get to put the future grid to a real-world test and see how it performs.
That happened during California’s historic August heat wave and the electricity shortages it created. After years building up a nation-leading market for renewables, energy storage and distributed energy resources, California faced the heat wave with a grid system that looks more like what the future holds than what you see in most other states.
These new tools did not save the day. California’s electricity supply fell short of demand on Friday, Aug. 14 and Saturday, Aug. 15, prompting the managed retreat known as rotating outages. Supply nearly fell short again in the following days, but an all-hands-on-deck statewide response rustled up new capacity and shed some demand.
The proximate cause of the crisis was not “too many renewables” so much as “too little dispatchable capacity.” Even with an elaborate bureaucratic process to pay for sufficient capacity to handle events like these, the state found itself without enough post-sunset power. California banked on borrowing from its neighbors to meet its peaks, and that didn’t work because the heat wave hit neighboring states too.
The outages reveal how insignificant energy storage currently is relative to California’s electrical needs, and how crucial it will be to meeting the state’s goals while preventing a repeat of this embarrassment.
Nowhere near enough storage in place
The batteries on the CAISO grid worked great — there just weren’t enough of them.
At the start of the year, CAISO had just 136 megawatts of storage in operation. That figure jumped considerably with the arrival of LS Power’s Gateway battery, the largest in the world at 250 megawatts.
But that nation-leading fleet of batteries remains infinitesimal compared to California’s historical peak demand of around 50,000 megawatts. The supply shortfall CAISO expected heading into the afternoon of Aug. 17, for instance, was 1,400 megawatts. The existing battery fleet stood no chance of covering the gap.
Indeed, startup OhmConnect was able to muster 150 megawatts of flexible demand by signaling to residents to adjust their power usage during the peak hours. Flexible demand emerged as one of the winners from the week, a cost-effective and scalable solution that helped pull the state back from the precipice. But sometimes you need electrons, not the absence of demand for them. That’s where more batteries come in.
Reinforcements are on the way. LS Power worked overtime to activate more capacity at Gateway, adding 30 megawatts in the midst of the heat wave, while nabbing the title for world’s biggest battery. The massive Moss Landing project is under construction near Monterrey, as is the 100-megawatt battery near Oxnard that Strata Solar developed and recently flipped to Capital Dynamics.
An even bigger batch is coming next year, in response to the California Public Utilities Commission’s belated realization last fall that gas-plant retirements planned a decade ago would leave the state short on capacity. And Capital Dynamics upped the ante last month by committing to build 2,000 megawatts in the next few years with Tenaska.
The good news is that California won’t be caught as unprepared for a peak event as it was this time; the state will only have more battery capacity going forward. At the same time, the crisis builds the case for hanging on to old gas peaker plants that were supposed to shut down; it’s harder now to say the state doesn’t need those peakers when it clearly needed them this time.
How long is necessary?
The outstanding question in the peakers-vs.-batteries debate is how long batteries need to discharge to safely take over for gas plants that can run indefinitely. I recently covered research from Form Energy on how conventional batteries can replace around 10 percent of New York state’s peaker plants, but the addition of long-duration storage allows for replacing almost all of them.
The early evidence from California suggests this debate may be missing the point, at least for now.
On the first night of the crisis, CAISO’s battery operations log shows a clear discharge event lasting just two hours, from 6 p.m. to 8 p.m. The battery fleet delivered between 100 and 150 megawatts throughout that period; after the second hour, CAISO signaled for the batteries to charge, implying that the moment of supply stress had passed.
CAISO's log of battery activity on the day of the first blackouts shows a clear two-hour peak discharge window from 6 to 8 p.m. before the grid operator gave the batteries the signal to charge up again.
The next day, which also ended with rolling blackouts, saw batteries discharging up to 310 megawatts between 2:30 and 3 p.m., loading up in the afternoon and delivering sustained capacity between 6 and 8 p.m. The next few days saw more scattershot battery operations, but the clear two-hour peak discharge returned on Wednesday, Aug. 19.
The storage industry likes to talk about using batteries to save surplus solar power and move it into nighttime delivery. In the past, that heroic vision had almost nothing in common with how California’s batteries actually operated. Following CAISO’s economic dispatch signal, the battery operations chart typically looked like an EKG, with jagged spikes up and down all day, seemingly at random. The ups and downs reflected quick and valuable bursts of ancillary services. The absence of visible discharge during peak hours meant batteries weren’t helping with the peak.
The chart from the heat wave shows batteries finally living up to their peak-crushing potential. It also shows that you only needed to invest in a 2-hour duration battery to make the most of this market opportunity. If any developers had taken the initiative to build, say, a 12-hour duration flow battery (if such a thing existed in the market), they would have banked on the high prices between 6 and 8 p.m. and then wondered what to do with the extra 10 hours of untapped discharge potential.
What comes next?
The long-duration energy storage lobby (it exists now!) took the opportunity to define the crisis as evidence of the need for their product.
“As the state works to integrate even more intermittent renewable energy into the grid to reduce our reliance on fossil fuels and meet our climate goals, we need long-duration energy storage to ensure we have enough in-state, clean and reliable energy to keep the lights on,” said Julia Prochnik, executive director of the recently formed Long Duration Energy Storage Association of California.
That sets up a disconnect between the power market that exists today and the future state of a decarbonized grid.
While we still have gas plants on standby, even a peak grid crisis only calls for two hours of steady battery discharge. But natural gas started ramping up by 10 a.m. on Aug. 14, peaking at more than 25,600 megawatts just past 7 p.m. before gradually tapering off. Even at midnight, gas delivered more power than renewables did at any point in the day.
California's dependence on natural gas, visualized. For extra points, try to find where the battery cavalry shows up to save the day.
When those fossil-fueled plants shut down, California will need significantly more than two hours of capacity to replace them. The question is, how does a long-duration developer build a business case in the current world to pave the way to the future world? Regulators earlier this year identified a need for 1 gigawatt of long-duration storage by 2026, but the mechanism for procuring it is still a big old question mark.
There’s also ample work to be done on the customer side of the meter. I’d recommend reading Jeff St. John’s reporting on how distributed storage showed up to fight the peak, even though years of policy work still haven’t produced a comprehensive set of rules for their participation.
“What’s ironic is that the [CPUC] and CAISO have been putting up all these barriers to us participating, and then they called us and said, ‘Hey, can you do this for us for free?’” said Ted Ko, vice president of policy and regulatory affairs at Stem.
As the dust settles, California will have to use all available tools to prevent a repeat of the August shortfall. That may mean extending the life of some of the gas plants that were supposed to shut down. But the event could accelerate the use of large- and small-scale storage by making it clear that they’re needed now.
If you spend too long preparing for the future, it might just sneak up on you.