If you’ve been paying attention to energy issues in the media lately, you may have encountered a curious narrative that’s starting to gain traction among supporters of renewable energy. Namely, that the core problem with wind and solar photovoltaic (PV) generation, which is matching moment-to-moment demand with the intermittency wind and sunlight, isn’t really a problem at all but rather a smoke screen being thrown up by fossil fuel advocates to block further deployment of wind and solar capacity.
The argument goes something like this: Wind blows at night, sun shines during the day, offshore wind blows most of the time, and if you have enough of everything spread over a large enough area, deficits in one area are balanced by output elsewhere. Germany, with its large installed capacity of wind and solar PV and a robust initiative to transform its power sector, is often offered up as a proof of this position.
To anyone with hands-on experience in power generation, this argument surely seems nonsensical, and it should. No matter where the wind is blowing or the sun is shining, most forms of renewable generation are not dispatchable and cannot be matched with moment-to-moment demand without some form of backup generation. And unless output is precisely matched to demand, grid instability is the result.
What Intermittency Means for the Grid
Let’s get one thing clear at the outset: Renewable intermittency is a real issue and a real challenge, at least if you listen to the people who oversee the grid rather than policymakers and pundits.
I spoke to Steven Greenlee at the California Independent System Operator (CAISO) about this issue and asked what he thought about arguments that intermittency concerns are overblown.
“The situation is not a crisis,” he said. However, “our analysis has shown that renewable resource variability does pose some operational challenges.
California has an ambitious goal of 33% renewables by 2020 and has seen substantial additions of residential solar PV over the past few years. CAISO, which oversees grid reliability in California, has enthusiastically embraced renewable integration. But it has no illusions about what that means.
CAISO, he explained, “must match supply second to second to the net load to maintain system balance and reliability.” The problem is that output from renewables can drop dramatically, by hundreds of megawatts, in as short a period as 30 minutes.
CAISO has devoted substantial resources to advanced forecasting capabilities and increased incentives for flexible, fast-start generation resources. Why that’s critical is illustrated by a chart Greenlee provided to GAS POWER (Figure 1).
The problem, Greenlee explained, comes in the evening as solar goes offline. Conventional generation has to come online rapidly to replace it, but most such generation can’t move that fast from a cold start—it has to be “idling” at minimum load during the day. But that minimum load may not be needed, which creates the potential for overgeneration conditions. These units can’t just shut down during low-load periods, because they may be needed as soon as the sun starts to set.
And that’s just day-to-day swings. “An additional condition not illustrated in the graph is the variability within the hour, because of passing clouds or storms that interrupt the output of the wind and solar generation resources,” he said. “For these times the ISO also requires flexible resources to smooth out the renewables’ sudden intermittencies.”
What It Means for Generators
It is true, at least according to some studies, that enough installed wind and solar over a large enough area with enough robust, smart transmission capacity would be able to power an entire grid. But the devil is in the details (and the dollars).
A study published by researchers at the University of Delaware last year found that a mix of solar and on- and offshore wind could provide reliable power for a large grid more than 99% of the time. But making it work would require building an enormously redundant system, with a nameplate capacity of almost triple the load under some scenarios. Most of that excess capacity would be idled all but a few days of the year.
The example of Germany is perhaps not what renewable advocates ought to be using; in fact, it validates the problems uncovered in this study. The evidence is mounting that increased penetration from renewables at the expense of baseload generation is not producing the expected results.
An article published in Dissent magazine this summer lays out just what a basket case Germany’s Energiewende, or energy transition, has become:
Despite massive construction of new capacity, electricity output from renewables, especially from wind and solar, grew at a sluggish rate. Germany is indeed avoiding blackouts—by opening new coal- and gas-fired plants. Renewable electricity is proving so unreliable and chaotic that it is starting to undermine the stability of the European grid and provoke international incidents. The spiraling cost of the renewables surge has sparked a backlash, including government proposals to slash subsidies and deployment rates. Worst of all, the Energiewende made no progress at all in clearing the German grid of fossil fuels or abating greenhouse emissions—nor is it likely to for at least a decade longer.
The most recent numbers are not pretty. While Germany has added renewable capacity at a breakneck pace, total generation hasn’t grown as much with it. Germany added about 10 GW of wind and solar PV in 2012, pushing its total from 54 GW to 64 GW, an 18.5% increase. But actual output grew much more slowly, by only 8.5%. While solar PV generation grew impressively from 19.3 TWh in 2011 to 28 TWh in 2012, wind output actually declined from 48.9 TWh to 46 TWh. And if you’re doing the math here, that output represents some pathetically low capacity factors: 17% for wind and a mere 11% for solar.
In other words, in order to meet a hypothetical 63-GW load for an average day (based on total consumption of about 550 TWh last year) with just wind and solar at current capacity factors, the country would need a lot more than its 64 GW of installed capacity: 286 GW of solar and 185 GW of wind (assuming equal output from both under the “solar by day, wind by night” theory). That’s actually far more overcapacity than the 300% figure produced by the UD study. It amounts to about 750% of the load—a preposterous level of overbuilding.
Why It Matters
As bad as that may sound, it’s actually a rosier picture than reality would be, since of course wind and solar don’t produce equal outputs across an entire day or an entire week. And it’s there that we run into the very real intermittency problem.
Take a look at this graph from the Fraunhofer Institute (a German solar energy research group) showing the aggregated output from Germany’s wind and solar resources over every day in 2012 (Figure 2). You can see that despite impressive generation during the summer, there are numerous periods during the fall and winter when capacity factors fall into single digits and even low single digits.
Expand one of the worst weeks, in mid-November, and the problem is even more glaring (Figure 3).
To push this thought experiment to its conclusion, if we take one of several weeks in 2012 when Germany’s combined wind and solar capacity factors fell below 10%, the country would have needed at least 600 GW of installed wind and solar PV to meet all demand, a breathtaking 1,000% overcapacity (which, on the worst days, still wouldn’t have been enough).
Those arguing against the intermittency problem might point out that this is a worst-case scenario, but that’s really the point: What we have here is exactly one of those intermittent gaps that has to be filled by dispatchable generation. “Solar by day, wind by night” simply doesn’t work every day of the year.
One last figure from the Fraunhofer report will make the issue crystal clear, if it’s not already. This graph aggregates the total solar and wind generation over 2012 to show the net load the combined resources were able to meet (Figure 4).
Remember this is with an installed capacity that grew from 54 GW to 64 GW over the year. Yet there are significant white areas once you venture past the 15 GW diagonal. Only if you get down under around 6 GW to 7 GW does the white completely disappear.
Every one of those white spots is a day when conventional generation had to fill the gap. Worse—at the risk of belaboring the point—this data only drills down to single days. Break it down to the hour or the minute, and there are surely far more holes to fill. And that matters.
Because, lest we get too lost in abstract numbers and figures here, all this fluctuation has real-world costs—big ones. Der Spiegel reported in August that over the past three years, the number of short interruptions in the German grid has grown 29%, and service failures have risen 31%. Half of those failures caused production stoppages for industrial customers, with damages ranging from tens to hundreds of thousands of euros. For industrial consumers and those with large amounts of electronic equipment, even millisecond voltage fluctuations can be disastrous. (This has naturally been a boon to makers of backup batteries, who have seen sales in Germany grow strongly over the same period.)
Where We’re Going
If you’ve gotten this far, a bit of clarification may be in order. I suspect the narrative I refer to in the opening got started because of various people going around asserting, “Renewables are intermittent, so they’re a waste of money.” Other people, who know what they’re talking about, have been replying, “No, the intermittency is there, but it’s manageable, and focusing on it ignores other advantages of renewables.” But some other people, who don’t know what they’re talking about, appear to have been taking that rebuttal to mean “ ‘Intermittency’ is a myth, so we can safely dismiss anyone who brings it up.” It’s that latter group I’m concerned with.
Those folks need to understand that unless there is so much overcapacity that even generation with capacity factors in the sub-10% range can fill it, dispatchable generation has to be available to fill those worst-case weeks. And that means, if you have a huge fleet of wind and solar PV, which normally supplies the bulk of your power, dropping under 5% capacity factors on bad days, you need to have what amounts to an entire parallel generation fleet that can take over when the wind isn’t blowing and the sun isn’t shining. That’s not energy efficiency by any measure.
The bottom line is that a stable grid cannot exist without substantial dispatchable capacity of some sort. That much should be obvious from the CAISO chart in Figure 1.
Of course, dispatchable renewables do exist. Hydro, pumped storage, and geothermal can take the load—in areas with sufficient hydro and geothermal potential. Solar thermal with molten salt storage can operate round-the-clock in sunny areas. But the technology is still new and expensive, it’s not clear what grid-scale production would entail, and it’s of limited use in areas with poor solar potential (like Germany in the winter). And economic grid-scale storage (whether batteries, flywheels, or power-to-gas) is still a long way away.
What that means is that, for the near term at least, clean conventional generation from gas and nuclear will be indispensable to back up growing renewable output if the goal is to reduce carbon emissions. Germany, as I discuss elsewhere in this issue, is having increasing trouble managing this balance, as a decision to dump nuclear in an environment of high gas prices is causing a resurgence of coal and rising emissions—the exact opposite of what the Energiewende was supposed to accomplish. The U.S., by contrast, seems to be aimed a better direction, with a growing fleet of advanced gas plants coming online to balance expanding wind and solar output.
Here’s hoping a few head-in-the-sand renewable advocates don’t derail things.
—Thomas W. Overton, JD is POWER’s gas technology editor (@POWERmagazine, @thomas_overton).