• A 2021 law requires that the Utilities Commission and Duke Energy set a plan to cut CO2 emissions from electricity generation by 70% (from 2005 levels) by around 2030 and be “carbon neutral” by 2050
  • This law requires the commission to set the “least cost path” to meet those goals — without compromising the “adequacy and reliability of the grid”
  • Analysis shows that all four of Duke’s scenarios, heavy on wind and solar, would risk capacity shortfalls in the winter and summer and be higher-cost than a model scenario featuring far more reliable, zero-emissions nuclear generation

In 2019, Gov. Roy Cooper issued an arbitrary goal of reducing North Carolina’s carbon dioxide (CO2) emissions from energy generation by 70% from where they were in 2005. The target year for this goal was 2030, and it also called for “carbon neutrality” by 2050, the idea that the state emissions would be reduced to the level of CO2 that the state is able to remove or “offset.” The undergirding idea is the suspicion that CO2 is actually responsible for adverse weather events in recent years and computer-modeled predictions of worldwide cataclysm within 10 years.

While North Carolina’s CO2 emissions have been falling all century (see below), on October 2021, the General Assembly overwhelmingly passed House Bill 951 that essentially put the governor’s arbitrary goals into law. The new law featured some important language, however. Bill sponsor Sen. Paul Newton promised it would “ensure reliable energy generation with fiscal responsibility.” The only way it could achieve those promises, as I noted in a previous research brief, is “by strict adherence to the law’s text.”

The law directs the North Carolina Utilities Commission (UC) to work with utilities to develop a plan by the end of this year to achieve the goals. Among other things, the text requires that the actions taken must:

  • be “reasonable”
  • set forth the “least cost path consistent … to achieve compliance with the authorized carbon reduction goals”
  • “Comply with current law and practice with respect for least cost planning of generation” in achieving compliance
  • “maintain or improve upon the adequacy and reliability of the existing grid”

The law also allows the UC flexibility in the initial compliance date of 2030, allowing a delay of two years or more if the UC “authorizes construction of a nuclear facility or wind energy facility that would require additional time for completion.” With respect to cost and especially concerning “the adequacy and reliability of the existing grid,” new nuclear would be a much less expensive way of reducing CO2 emissions than most other generation.

In keeping with the law, Duke Energy Progress, LLC and Duke Energy Carolinas, LLC (jointly, “Duke”) filed its Carolinas Carbon Plan with the UC. The Duke plan includes four alternative portfolios for achieving the two phases of CO2 emissions.

With respect to the 70% reduction goal, as early as 2032, all four of Duke’s proposed scenarios would feature the following ranges by that time:

  • A loss of 4.9 megawatts (GW) to 6.2 GW of coal-fired electricity owing to plant retirements
  • 5.4 GW–7.7 GW of new solar capacity
  • 1.7 GW–2.2 GW of new battery storage
  • 0.6 GW–1.2 GW of new onshore wind energy capacity, and up to 1.6 GW of new offshore wind capacity
  • Up to 0.3 GW of new nuclear generation
  • Up to 1.7 GW of new pumped storage
  • 2.4 GW of new combined-cycle (CC) natural gas or hydrogen generation
  • 0.8 GW–1.1 GW of new combustion-turbine (CT) peaker natural gas or hydrogen generation

By 2050, of course, the changes would be more dramatic: a loss of 9.3 GW of coal and additions ranging from 18.1–19.9 GW solar, 1.7–1.8 GW onshore and 0.8–3.2 GW offshore wind, 5.9–7.4 GW of battery storage, 2.4 GW CC gas/hydrogen generation, 6.8–7.5 GW of CT peaker gas/hydrogen generation, 9.9–10.2 GW new nuclear, and 1.7 GW of pumped storage.

Even in the near term, Duke’s plans would retire a sizeable amount of reliable, quickly dispatched baseload generation and add a great deal of unreliable, intermittent generation. The former is the more emissions-intensive coal, while the latter are weather-dependent “renewable” resources of wind and solar. Nevertheless, even though the law makes ample space for new nuclear generation — which tops all generating sources in terms of reliability and is also a zero-emissions resource — Duke’s plan would allow for considerably less new nuclear than unreliable wind and solar and untested hydrogen.

“Wind and solar are making the grid more unreliable”

It therefore portends significant reliability problems. Regardless of the emissions reductions from wind and solar facilities, they are completely at the mercy of elements: when and if the sun shines and wind blows. For example, last September power customers in the wind-dependent U.K. were hammered by high prices when the wind unexpectedly stopped blowing. Five of the country’s eight nuclear power plants were offline. Energy experts warned that were it to happen in the winter, it would present “a real issue [for system stability.” The system operator requested the restart of a coal plant to bring more power to the grid — something it would not be able to do after 2024, when all coal plants are due to be shut down. As it was, consumers found themselves paying nearly seven times more for power in September 2021 than they were a year prior.

In July, the North American Electric Reliability Corporation (NERC) released its new 2022 State of Reliability report, highlighting fast emerging threats to grid reliability outside of peak demand. The Institute for Energy Research discussed several of NERC’s key findings regarding renewable resources and grid reliability, including (emphasis added):

According to NERC, no longer is the peak demand period the only clear risk period because risks can emerge when weather-dependent generation is impacted by abnormal atmospheric conditions or when extreme conditions disrupt fuel supplies. The impact of wide-area and long-duration extreme weather events, such as the cold weather in February 2021 in the South Central United States, and the heat in August 2020 in the West underscore the need to consider extreme scenarios in resource adequacy and energy sufficiency planning. Because there is less flexible generation that is fuel-assured, weatherized, and dispatchable occurring in many areas as the resource mix evolves toward renewable energy, the risk of energy shortfalls increases. Wind and solar are making the grid more unreliable as they gain share. Although margins in 2021 were assessed as adequate for traditional reliability criteria, NERC warned of potential seasonal shortages when accounting for more extreme conditions.

In its summer assessment, NERC warned of unexpected tripping of solar generation, which could become a major threat as more solar is interconnected to the grid. The inverter tripping challenge is one of the most risky issues to reliability as NERC expects 500 gigawatts of solar coming online in the next 10 years. The unexpected tripping has occurred during “normal grid disturbances” such as a lightning strike or a piece of equipment going offline. In 2016, California’s Blue Cut Fire tripped several transmission lines and caused almost 1,200 megawatts of solar energy capacity to go offline unexpectedly. Similar losses of solar generation also occurred between May and August in California and Texas last year. California and parts of the western interconnection are also at elevated risk, particularly in late summer as water resource levels decline and solar output begins to fall off earlier in the day.

Winter or summer, every Duke scenario risks capacity shortfalls

The John Locke Foundation’s Center for Food, Power, and Life (CFPL) analyzed the Duke scenarios and found that (emphasis added) “none of Duke’s plans are the least-cost means of providing reliable electricity to North Carolina residents.”

Each of the scenarios would greatly increase the amount of installed capacity on the grid well beyond the expected growth in population. While population is expected to increase by 32% by 2050, the amount of installed capacity in the Duke plans would increase by 84.3% (Portfolio 4) to 92.3% (Portfolio 1).

Overbuilding would add to the Duke plan’s cost to consumers, and it’s a tacit acknowledgment that much of this new capacity would indeed be unreliable. As stated in the CFPL analysis, “Without a doubt a significant reason for seeking such a massive buildout in installed capacity is the perceived need to overbuild renewable resources, which do not emit CO2 but which are inherently intermittent and unreliable, in order to overcome their intermittence” (emphasis added).

Alarmingly, the CFPL analysis found that by 2031-32 in the winter, when electricity demand is highest and power outages are deadliest, each scenario would no longer rely on dispatchable baseload generation to meet net load, but instead on the accredited capacity of wind and solar, storage, demand-side management, and load management resources. For example, here is the graph concerning Portfolio 1:

Duke Portfolio 1 Total Firm Capacity as a Percentage of Peak Demand

The analysis explained the risk: “as we learned in California and Texas, accredited capacities for wind and solar generation are not guaranteed, and an overreliance on these technologies could result in capacity shortfalls. Storage systems charged with intermittent renewables are not guaranteed to be available when power is needed most, which would potentially leave the grid short of capacity during peak demand periods.”

Furthermore, the CFPL analysis showed that, under hourly load estimates during a model week in August (the height of summer) by 2050, the Duke scenarios would experience “capacity shortfalls of 31 to 41 hours, which could be significant enough to trigger load-shedding” (i.e., rolling blackouts). Here is a representative graph concerning Portfolio 1:

Duke Portfolio 1 Hourly Load Shape, Capacity Shortfalls: August 24–30, 2050 (34 hours)

In no way could those outcomes signify “maintain[ing] or improv[ing] upon the adequacy and reliability of the existing grid” as required by law.

Nor, for that matter, do the Duke scenarios represent the best that can be done to comply with the law. The CFPL analysis presented a fifth model scenario outside of Duke’s four, which would be far more reliant on new nuclear generation. This model scenario achieved the law’s emissions reductions without compromising grid reliability. It also would be less expensive to consumers than any of the four Duke portfolios, showing that they also cannot satisfy the law’s requirement of providing the “least cost path” to compliance.