California’s Duck: More Quack than Bite

This is the long version of a post that appeared in the California Energy Markets Sept. 23 Issue, and also posted on TURN’s website here —

— — — — — — — — — — — — — — — — — — — — — — — — — — — — — -

If you’ve followed debates about California’s electricity system over the last several years, you’ve no doubt seen our friend the duck made out to be an evil creature ready to destroy California’s ambitious energy plans. The “duck curve,” as its known, represents a plot of net demand (electric demand minus renewable generation) on certain spring days. The low “belly” of the duck is caused by increasing solar generation and the “neck” results from reduced solar output as the sun goes down and demand for electricity increases into the late afternoon and evening. While it is often implied that this odd pattern represents potential catastrophe for grid management, I believe issues posed by the duck have been misunderstood. Problems found on “duck days” can be managed as California achieves 50% renewable electricity by 2030, and increasing renewable generation presents opportunities to create a cleaner, more flexible electricity system.

What Is the Duck Curve?

As displayed in the figure below, a duck that gets fatter over the years as additional solar generation comes online begins to form when we look at net load (electricity demand minus renewable generation) for certain days in the future.

Source: CAISO (2013), What the duck curve tells us about managing a green grid,

Notice the top of the graph identifies that we are looking at a particularly sunny day in spring when the glorious California sun beats down amid relatively cool temperatures (and thus low electric demand). This rather odd net load shape is not this extreme on most days of the year.

The duck poses two interrelated issues that present challenges for grid operation. The first is caused by the duck’s “belly” — as (primarily) solar generation increases in the afternoon, the supply of electricity may exceed demand. Since demand and supply of electricity must always be in nearly exact balance to maintain system reliability, this means excess power needs to be stored, exported, or curtailed, a term meaning a renewable source of energy is shut off. This sounds terrible because we’re wasting an essentially free (at the margin) renewable resource. The second issue is the “ramp,” or the steep rise of the “neck” of the duck between 4:00 and 8:00 p.m., which requires the rapid addition of other generation to quickly meet peak demand.

How Big is the Duck’s Bite?

I looked primarily to recent studies by the Union of Concerned Scientists (UCS) and General Electric for guidance on grappling with the duck, though many others are out there. Let’s take the second issue (ramping) first. The concern here is that as renewable generation (primarily solar) dwindles, gas plants will have to pick up the slack by adding energy to the system quicker than they are used to. Generators may be limited by their “ramp rate,” or amount of generation that can be added over a particular time period. However, a closer look shows that California is not only blessed with lots of sun, we also have a relative glut of quick-ramping generation assets in the form of natural gas plants. UCS found that new natural gas plants would not be needed through 2024, while GE states that “The duck curve does not necessarily require fast ramping capability (since the ramp is over a 2–3 hour period) but rather standard ramping rates, with expanded range over the multi-hour load ramp.”

Coming to this first issue presented by the duck’s “belly,” how big of a problem is curtailment? And is curtailment something that should be avoided at all costs?

The best way to measure the relative level of curtailment is to look at the percentage of renewable energy that must be curtailed over the course of a year due to oversupply conditions. At 50% renewable penetration, which California will reach by 2030, modeling conducted by UCS approximates this percentage to be around 5%, before reduction measures are employed. A study by the CAISO suggests curtailment levels of 3%-6% when renewable energy supplies 40% of electric demand in California.

Curtailing excess renewable generation does not necessarily create reliability problems or cause major disruptions. Fundamentally, it could increase the costs of renewables, depending on how the “cost” of the curtailed energy is covered. It may require overbuilding renewable generation capacity in order to meet the 50% target. While limiting curtailment levels is certainly a valid policy goal, some curtailment may be a cost-effective means of meeting environmental goals cost-effectively. That said, I think we can reduce curtailment to minimal levels as we approach 50% renewable electricity penetration.

Potential solutions for the ramping and curtailment issues are often intertwined, whereby measures to mitigate curtailment also decrease ramping needs. For instance, shifting electricity demand away from the peak and towards the duck’s “belly” decreases both curtailment and ramping requirements. Similarly, storing solar power during the day for production at night, and/or west-facing solar panels (which produce more electricity later in the day) provide comparable “two birds-one stone” effects. I won’t cover most of these solutions here, but a good start is a paper by James Lazar of the Regulatory Assistance Project. I discuss one solution below for reducing curtailment that requires shifting some responsibility for grid reliability to renewable generators, away from fossil fuel plants.

Renewables Can Contribute to System Reliability to Reduce Curtailment Levels

Looking back at the duck graphic above, you may have wondered why “overgeneration” can occur when net load does not go below 0. Let me explain. Below the “belly” of the duck sits a “stack” of generators, many (but not all) of which use fossil fuels like natural gas. The net load curve may dip below electricity supplied by these generators, which would result in oversupply of electricity. Of course fossil plants could also be “curtailed,” or turned off, but this is generally not practiced because some of these resources cannot come back online quickly enough due to operational constraints and system reliability concerns. Exacerbating the excess supply of electricity when there is low net demand are fossil plants’ minimum capacity, which mean plants can only be turned down to 30%-50% of their maximum operating capacity. Thus, even though these plants will decrease output to minimum levels as prices become low or negative, as would occur at the duck’s “belly,” they still provide a relatively large amount of electricity. System operators therefore tend to favor renewable curtailment to maintain system reliability.

This “minimum” level of output is further increased because fossil plants provide something called “downward operational reserves.” If needed, fossil plants ensure (and are paid for) a certain amount of capacity that, when called upon, can help decrease supply of electricity quickly if a large drop in load is experienced. This further inflates the “stack” of fossil generators that sit below the net load curve. UCS found, for a 50% renewable scenario, about 70% of renewable curtailment is due to the downward reserve requirement.

One particularly impactful solution that both UCS and GE discuss is to allow some portion of downward reserves to be taken over by renewable generators, rather than relying solely on fossil fuel plants. This would allow fossil generators to turn down output towards lower minimum levels (and/or potentially shut down). Downward reserves held by renewable generators involves more strategic curtailment, whereby a solar generator may be called to curtail in sub-hourly increments (rather than larger hourly blocks) to provide downward reserves. More strategic curtailment at finer levels of control results in less curtailment overall. If just a portion of downward reserves can be provided by renewables, UCS finds that curtailment can be reduced by 44% to around 2.7% of total renewable output.

These findings also demonstrate the relative importance of fossil plants’ minimum level requirements. The ability for natural gas plants to turn down to lower minimum production levels will be a primary lever of “flexibility” that must be explored as California increasingly shifts away from fossil fuels towards renewables. Therefore, we see that issues related to the duck curve derive from operational inflexibility of both renewable and fossil generators.

The duck and increasing renewable generation does not necessitate doomsday scenarios for California’s electric grid. Sensible measures to shift some responsibility for grid reliability to renewables can make a big impact along with other cost-effective solutions to increase net load during spring daytime hours and decrease the steep ramp. By the time California hits 40%-50% renewable output, we will see the duck fly right by.