clock menu more-arrow no yes mobile

Filed under:

Flattening the "duck curve" to get more renewable energy on the grid

"Hm? You want to do what to me?"
"Hm? You want to do what to me?"

Wind and solar power are pretty great, especially the part about how they don't create pollution that kills people and threatens the stability of advanced civilization.

But they do have their challenges.

In my previous post, I discussed one of those challenges: the "duck curve," i.e., the effect wind and solar energy have on daily demand for utility electricity.

Here's the key graph:

california’s duck curve. CAISO

Those lines show "net load" — which is total demand for electricity minus whatever renewable energy is on the grid — over a typical spring day in California.

As you can see, as more and more solar energy comes online during the sunny daytime hours, demand for utility electricity is pushed down further and further. That's the duck's belly.

But just as the sun is setting and solar energy is declining, people are getting home from work and turning on all their appliances. So net load rises very rapidly (the duck's neck) to an early-evening peak (the duck's head).

Grid operators do not like these big valleys and peaks, and they really don't like the steep, rapid ramping in between them. Keeping a power grid stable is hard enough already.

So the duck curve is a problem. The question is how can the duck be flattened? How can those peaks and valleys be smoothed out?

duck in flight
Flatten out, duck. Flatten out.

It's a fun puzzle, to be honest. You have these two variables, supply and demand, that have to be balanced at every moment. To some extent both can be controlled, moved around, mixed and matched. It's like a real-time Rubik's cube, only the solution keeps changing every hour.

Vis–à–vis duck flattening, I'll touch on two big, visionary ideas and then 10 smaller, more immediate, more practical measures that use existing technologies.

The main point to make is that we have a decent (if somewhat hazy) understanding of the long-term solutions to the duck curve, the kind of stuff we'll be dealing with in 2050 when wind and solar are getting toward 60, 70, 80 percent of grid energy.

But we have a very clear understanding of the sort of immediate steps we could take to accommodate more wind and solar than we have today. There's no reason the duck curve should delay the growth of renewables.

The two big, long-term solutions to duck curvature

The first big strategy to flatten the duck is interconnection. The more grids can be connected with one another to form larger grids over larger areas, the more spread out the potential for wind and solar will be and the more spread out load will be, both of which will serve to smooth out the peaks and valleys in the curve.

Theoretically, you can push the problem back for a long time with this strategy, hooking up more and more grids. Maybe someday we'll get to the much-discussed global grid.

Transmission today, transmission tomorrow, transmission forever.

But that's a long way off, and in practice, though the technology of high-voltage transmission is well-understood, the politics of it — how to get funding and rights of way, how to deal with NIMBYs — is not.

The second big strategy is energy storage. If you can store some of that wind and solar energy rather than automatically sending it to the grid, you make it "dispatchable," meaning you can time it. It becomes a movable piece of the puzzle.

A sufficient amount of energy storage would, almost by definition, flatten the duck and remove any limits on the integration of wind and solar. But at least at current prices, that would be prohibitively expensive.

Advocates of these two strategies are always arguing with one another about which is the ultimate long-term solution.

Researchers from NOAA recently published a study in Nature Climate Change showing that by a) siting renewables where they have the highest potential, and b) building high-voltage transmission lines to those areas, the US could cut greenhouse gas emissions 78 percent from 1990 levels in 15 years — with no new energy storage.

Wind power potential across the US.
Wind power potential across the US in 2012.

By linking the entire US into a single high-voltage grid, you can keep the duck flat enough to accommodate, to be technical about it, a metric shit-ton of new renewables. (As a bonus: According to NOAA, it can be done at comparable costs to a fossil-fueled grid.) Thus, NOAA researchers say, there's no need to wait for storage to scale up and get cheaper.

And that's true. But, as storage fans respond, getting high-voltage transmission lines built is extremely difficult and time-consuming. Meanwhile, storage is rapidly getting cheaper and seeping into the grid to fill the cracks.

It's kind of a pointless argument: They're both right. Both new transmission and lots of storage will eventually be necessary to get the grid to zero carbon. (Transmission will help wind more; storage will help solar PV more; see this study.)

But it's important to remember that we don't have to wait on new transmission lines or a scale-up of energy storage. There's a great deal that can be done immediately, with existing tools and techniques.

10 steps to flatten the duck quickly

flying duck (RAP)

To guide us through this somewhat nerdy territory, we have "Teaching the 'Duck' to Fly," by utility guru Jim Lazar of the always excellent Regulatory Assistance Project.

(The paper was originally released in 2014; since then, utilities have tried out many of the ideas, experts have offered feedback, and the new and improved second edition was just released.)

Lazar starts with a hypothetical California-like grid with a duck problem:

duck curve (RAP)

The blue line is total electricity demand. The green line shows how wind and solar duck it up.

Lazar then offers 10 practical ideas to start flattening the duck. I'll walk through them quickly. (Their relative contributions, and much supporting detail, can be found in the report.)

1) Target energy efficiency to the hours when load ramps up sharply

If load ramps up quickly during a certain time of day (mainly between 4 and 7 pm), then target the energy uses that cluster in those hours.

Lazar focuses on "residential lighting, air conditioning, and office building lighting controls." Technologies to substantially cut energy use in these applications is widely available and affordable, and would cut demand when it's most needed.

2) Acquire and deploy peak-oriented renewable resources

Policies that require utilities to purchase renewable energy generally do not say anything about the timing of when that energy is produced. But timing turns out to be important (a regular theme in the duck-flattening literature).

If renewable energy production needs to be pushed back a few hours later in the day, there are particular resources that can help. Most big hydropower facilities have some storage ("pondage") that allows them to alter the timing of their power production; they could be induced to push it later.

hoover dam
Saving some power for later.

Some wind sites produce energy more regularly in the evening hours; they could be favored in procurement. Solar panels could be turned to face west rather than south, which somewhat decreases their total production but, importantly, keeps them producing for up to two hours after south-facing panels stop.

And concentrated solar plants (CSP), which use the sun to heat fluid that turns a turbine, can store some of that heated fluid and delay some of their production.

In short, renewable energy can be pushed back into evening, when it will be more help with the peak (er, duck's head).

3) Manage water and wastewater pumping loads

This one sounds boring, but it turns out pumping water consumes 7 percent of total electricity in the US (and much more in California, where lots of water gets moved over long distances).

Current electricity rate structures encourage water and wastewater utilities to use small pumps that operate continuously, at a low level, throughout the day. Tweaks in the rate structure (I'll spare you the details) could instead encourage them to use large pumps that operate only in off-peak hours, shutting off during high ramp or peak times.

4) Control electric water heaters to reduce peak demand and increase load at strategic hours

Hot water is most intensively used in the morning and evening. But there's no reason water heaters should heat the water during those hours. Heated water can be stored a long time.

hot water heater
The hot water heater: one of clean energy's more humble champions.

Theoretically, the 45 million electric water heaters in the US could be connected to the grid and controlled; water heating could be shifted to those times when excess wind or solar energy needs to be absorbed.

Not only could load be shifted, but those same water heaters could effectively be treated as batteries that provide grid frequency and voltage control.

This would require simple load-control technology on the heaters and some new institutional arrangements with customers, but it's entirely doable. Lots of utilities are already doing it.

5) Convert commercial air conditioning to ice storage or chilled-water storage

Air conditioners are a huge part of peak load for most utilities. And a big part of that is commercial air conditioners, those giant units you see on malls and university buildings.

But just as water heaters don't need to heat water during the hours it's being used, commercial air conditioners don't need to make cold during the hours people use it. They can make it beforehand, when power is cheap, in the form of ice (or chilled water). Then when demand is peaking and power is expensive, they can use the stored ice to cool buildings (as a bonus: much more quietly).

In this case, ice is the battery that allows an air conditioner to shift its time of peak power consumption.

Ice-storage air conditioners are a widely available and affordable technology.

6) Rate design: Focus utility prices on the "ramping hours" to enable price-induced changes in load

This one's pretty straightforward. If you want people to shift their energy use from certain hours to certain other hours, reflect that in the price of energy. This is known as "time of use" pricing, whereby the price of electricity varies throughout the day.

Studies have shown that customers voluntarily shift their load in response to price signals. And eventually, much of that process can be automated by "smart" appliances and home energy control software.

7) Deploy electrical energy storage in targeted locations

Unlike thermal storage (hot water, ice) or mechanical storage (pumped water), electrical storage is still pretty expensive, though the cost is falling rapidly.

However, there are certain targeted places on the grid where compressed air, pumped hydro, or battery storage make sense. These need to be places where storage can produce multiple value streams:

  • Defer investments in new grid infrastructure
  • Time-shift power use, storing power when it's cheap and releasing it when it's expensive
  • Provide spinning reserves and frequency or voltage control to the grid
A pumped-hydro energy facility just outside Los Angeles.
A pumped-hydro energy facility just outside Los Angeles.

8) Implement aggressive demand response programs

I explained demand response in this post. It basically amounts to finding lots of little ways that customers can shift their demand, aggregating all those customers together, and treating the sum of their demand-shifting capacity as dispatchable power — "negawatts" instead of megawatts.

9) Use inter-regional power exchanges to take advantage of diversity in loads and resources

This is Lazar's nod to interconnection, only it's not about building new power lines; it's about using existing interconnections more strategically.

There are places in the US where separately managed grids already have links with one another. Those interconnections can be better used to trade power. Different regions have different times of production and different peak loads; they can help smooth out each other's peaks and valleys.

10) Retire inflexible generating plants with high off-peak must-run requirements

There are some big, older power plants (coal, nuclear, and natural gas) that are extremely inflexible. They cannot ramp up and down quickly. And it's very expensive to shut them completely off and restart them, so they must always be kept running at a minimum level. When renewable energy starts expanding too much, it bumps into these limits and is curtailed.

coal plant
A dinosaur in twilight.

What a modern grid needs above all is flexibility. Over time, these inflexible plants should be retired and replaced with flexible resources — newer natural gas, storage, and (some) renewables. The more flexible the overall grid is, the more wind and solar can be integrated.

So that's the 10 steps!

(It's worth pausing to acknowledge, as Lazar does, that every region and utility is different and each will need its own customized set of solutions.)

Supply and demand curves are not fated — they can be reshaped

Here's what all these measures put together do for Lazar's hypothetic grid:

duck curve, flattened
A flat duck.

That dotted line is the old net load curve — the duck. The thick green line is the new and improved net load curve, after all Lazar's strategies have been applied.

As you can see, the belly no longer hangs so low and the head no longer pokes up nearly so high. There are no more steep ramps. The duck has been flattened. It's flying!

And it's been accomplished using existing technologies and demonstrated policies. No breakthroughs or grand schemes required.

It's just a matter of breaking supply and demand down into their constituent parts, assessing which parts can be controlled or shifted (which is more and more of them), and rearranging them so that they balance throughout the day.

It's not rocket science. It's, uh, duck science. And it can enable any utility that's ready and willing to accommodate a lot more clean, renewable energy.

Further reading:

If you can't get enough of this ducking stuff, here are some ways to dive deeper:

These are all about the duck curve specifically. On the somewhat broader topic of integrating more wind and solar into the grid, a reading list would run several (kabillion) pages. Just start with this post and this post, which contain links to lots of other resources.