In 2015, Électricité de France put out a pamphlet titled Flexible nuclear generation to foster the development of renewable energy as one of its “50 Solutions for the Climate.” The pamphlet boasts of the flexibility of EDF’s nuclear fleet, showing a 1.3 GW nuclear power plant increasing and decreasing its output by 70% within 30 minutes.
The documentation centers on the Golfech plant, and shows these impressive feats of rapid ramping, with two cycles up and down within a 24 hour period. And while EDF noted that such cycling was prompted by daily changes in demand, it also stresses that such capabilities will be able to make nuclear power a good complement to the fluctuating output of wind and solar on the grid. However, in Northern Germany a situation has played out which challenges this line of reasoning. In February the Brokdorf nuclear power plant was taken offline after damage to its fuel rods was found. According to a local nuclear supervisory authority, the operation of the plant in “load-following” mode had contributed to unexpected oxidation of the rods. As of July, the plant was operating in “safe mode,” and politicians from Germany’s Green Party are calling on a Swiss reactor near the German border with similar problems to be shut down.
These two examples give very different accounts of whether or not nuclear power is suited to accompany high levels of wind and solar generation. The problem at Brokdorf comes as the U.S. nuclear industry is making a new push for subsidies for aging and uncompetitive plants, and as some U.S. energy journalists are promoting “flexible” nuclear as a solution to accompany high levels of renewable energy.
So can nuclear accompany high levels of renewable energy? This is a technical and economic question, and one that has ramifications for the future of the technology.
Wind, solar, and flexibility
Wind and solar are variable and intermittent. Their output varies not only seasonally, but from day to day and hour to hour. So does electricity demand, and the result is that a degree of flexibility has always been needed in power systems to adapt to changes in demand, including when people go to and from work, when factories shut on and off, and when weather leads large numbers of people to turn on and off air conditioning and electric heaters.
And while low-to-moderate amounts of solar typically reduce peak demand for electricity, much more flexibility is needed on power systems that have very high levels of wind and solar. In California, which currently gets more than 10% of its electricity from solar on an annual basis, other forms of generation need to supply an additional 10 GW of electricity over a three hour period to meet both the increase in evening demand and the fall-off in solar production when the sun goes down.
This means that come 5 p.m. gas-fired power plants across the state are increasing their output, which is called “ramping.” As the state puts more solar online to meet its 50% by 2030 renewable energy mandate, it will need even more ramping. Brendan Pierpont, an energy finance consultant at London’s Climate Policy Initiative, has worked on reports that model these details. He estimates that under a situation where 35% of California’s electricity comes from solar, during the most challenging day of the year, enough resources to meet roughly half of peak demand will need to ramp over a one hour period, and 80% during three hours.
California’s sole remaining nuclear power plant, the Diablo Canyon, does not ramp down during the middle of the day, despite frequent and recurring situations of negative prices, and even during days when a portion of the state’s solar output is curtailed on a system-wide basis. This is not unusual for the U.S. fleet, which typically runs 24/7 aside from times when individual reactors are shut down for maintenance and refueling.
Conditions are different in Europe. Nuclear power plants are regularly ramped up and down in France, to partially respond to the shift in electricity demand from day to night. Additionally, in other nations plants such as the Brokdorf facility are ramped to respond to fluctuations in wind and solar generation, although the vast majority of nuclear power plants are not.
Ramping, start-up and shut-down
The nuclear industry claims that all currently deployed boiling water reactors (BWR) and pressurized water reactors (PWR), which make up the entire nuclear fleet in the United States and the majority in Europe, can ramp quickly. The industry cites a ramping rate of 5% per minute between 50 – 100% of rated power, while EDF states that its reactors can ramp up to 80% up or down in 30 minutes, which may be due to changes in the control rods in French reactors.
Flexibility on a fleet-wide basis is not guaranteed to the same degree. While EDF estimates that it can ramp 15 GW with its nuclear fleet in 30 minutes, there is no significant operational record to go on for the U.S. fleet, and only a portion of the nuclear fleet is ramped in Germany. EDF noted in its comments to pv magazine that “France is the only country in the world to know the power of its nuclear fleet.”
However, IASS Potsdam Senior Fellow Craig Morris, who has written extensively about nuclear energy and renewables, has stated that “no nuclear fleet worldwide is ramping to any significant extent, so we actually have no idea whether ramping will work in practice.”
There is also contextual evidence which suggests that top-line statistics are not telling the whole story. In 2009, EDF stated that while its third generation PWR can ramp up at 5% of its maximum output per minute, that this is from 25% to 100% capacity and is limited to a maximum of two cycles per day and 100 cycles per year. The utility notes that higher levels of cycling are possible but that this is limited – 60% – 100% of capacity.
Plants may also not be allowed by regulation to ramp this quickly. Austria’s Ökologie Institut notes that while a study by the Institute for Energy Research found that Germany could ramp half of its total nuclear capacity in 15 minutes, that “the maximal technical capability for load following maneuvers neglects safety relevant restrictions, which are currently valid in the operating German nuclear power plants.”
And while ramping levels down to 20%, 25%, or 30% of rated power may be enough to accommodate moderate levels of renewable energy, for very high levels of renewable energy, having even an entire fleet of nuclear plants running even at 20 – 30% of rated output will mean oversupply of power at times when solar and wind are at full output, particularly if this happens during a time of low power demand.
To achieve maximum flexibility to accommodate high levels of wind and solar, is a question not only of how well remaining power plants can ramp, but how quickly they can go to 0% power, and return to full output. Newer combined cycle gas turbines (CCGT), designed for fast start up, can do this in around 30 minutes, and combustion turbines (CT) in less than 10. This is among the reasons that in its recent report on the integration of high levels of renewables, Flexibility: the path to low carbon, low cost electricity grids, CPI modeled grids with a mix of CCGT and CT gas plants, as well as lithium-ion batteries as the most suitable and inexpensive means of achieving the high levels of flexibility needed, beyond harnessing low cost demand-side flexibility and existing flexible generation.
Nuclear power plants cannot be quickly turned off and back on again, and for common plant designs the time to start back up and reach full load is one to two days. However, such plants can be run in hot standby mode, where the plant is temporarily disconnected from the grid, or system power modes, where they are not shut fully down but are producing only enough power to keep internal processes operating, and are not exporting power to the grid.
According to a 2010 article in the International Journal of Nuclear Power (ATW), German nuclear reactors are able to ramp from hot standby to full power in only one to two hours, and less than an hour in “house load.” This is quick enough where a fleet of such plants could theoretically be ramped in the evening to make up for waning solar output.
Wear and tear
Nuclear power plants share with CCGT and coal plants the characteristic that both generate power by boiling water to create steam to run a turbine, but the similarity ends there. Nuclear reactors are massively complex pieces of machinery, where powerful reactions are carefully controlled by plant operators, with multiple layers of safety mechanisms. Although there are other options for both designs, for both BWR and PWR, the mechanism for ramping typically involves the insertion of special control rods, along with a boron solution.
The nuclear industry admits that ramping results in additional wear on plant equipment, however there is disagreement between the nuclear industry and its critics regarding how much the control rods are affected by ramping nuclear power plants and the degree of the resulting effect on safe operation of these plants. EDF maintains that most of the effects are in the secondary circuits, or the non-nuclear part of the plant, such as pumps and valves, and describes the additional maintenance needed as being “marginal.”
“The pressure and temperature variations are much lower in the primary circuit (localized in the nuclear part), which avoids consequences on the materials,” EDF told pv magazine.
Meanwhile, a 2010 report by Austria’s Ökologie Institut describes a mechanism whereby frequent ramping deforms the plastic on control rods, with potential cracking if the power increase is too large. In the case of the Brokdorf plant, safety inspectors attributed accelerated oxidation of the plant’s rods to ramping.
But whatever the full extent of the impact on control rods, this is only one of the components subject to increased wear. Ökologie Institut’s NPP Output Flexibility states that a reasonable conclusion is that ramping wears out entire nuclear power plants faster. “One can indeed assume that because of frequent load-following cycles, thermal stresses, fatigue, and mechanical constraints, flexible [nuclear power plants] NPP are likely to age quicker than those operating at base load,” reads NPP Output Flexibility.
Regardless of who is closer to the truth as to the impact of ramping on the life of plants, it is clear that there are economic impacts to ramping nuclear plants. Some of these come from the increased wear. EDF, as cited by Ökologie Institut, is reported to have found that running plants at partial load increases unscheduled outages, at a cost of several million euros.
However, the economic consequences of ramping are not confined to wear on the plants. Compared to other forms of generation, particularly gas plants, nuclear power plants are characterized by high up-front costs and low fuel costs, with a mid-range of capital costs six times as high as that of CCGT (and an even greater differential for other gas turbines). Ramping down affects the profitability of projects, and lengthens the timelines for paying off initial costs. The economic impact of reduced output is more severe for nuclear power than for fossil fuel generation.
This is not a trivial matter. Other than “peaking” gas plants and diesel generation, new nuclear power plants are already the most expensive form of conventional generation per unit of electricity, with consultancy Lazard estimating a levelized cost of electricity of $97-136 per megawatt hour. In fact, the record of nuclear reactor construction start dates shows that the fall in global nuclear construction preceded the Chernobyl Disaster, and writers such as Craig Morris have concluded that it was cost, not safety concerns, which prompted the global move away from nuclear in the 1980s. If nuclear power plants are expected to produce less power and generate lower revenues, this further undermines an already precarious position.
Oil and water
Within certain ranges and limitations, nuclear power plants can ramp quickly. Whether this is sufficient for the needs of the energy transition is unclear, but it is possible that such plants could accommodate moderate levels of solar by ramping down in the morning and up in the evening, as part of a larger portfolio of different types of generation. For very high levels of PV, the picture is less clear, and there is no track record to go on.
There are strong reasons why nuclear power plant owners will work against plants being operated regularly in this manner, and the economics of building new nuclear power plants for load following are dismal. The nuclear industry has been forthcoming about the fact that there is an economic disadvantage to ramping nuclear generation. There are clearly technical issues as well, which at least further undermine the economics of ramping nuclear power plants, and may add additional safety concerns.
While developed nations should prioritize rapid decarbonization over short-term costs, there is not now and never will be an unlimited amount of money to pour into this problem. The nuclear industry knows this, and as such the attempt to cast nuclear power plants as a suitable accompaniment to high levels of wind and solar is ultimately a desperate act by an industry which is in severe crisis in both Europe and the United States.
Nuclear reactors may be able to ramp (within limitations), but ultimately nuclear is fighting for space on the grid with wind and solar. As such the building of new nuclear power plants, and in some cases the extension of licenses for old ones, can limit the transition to renewable energy.
EURACTIV, German nuclear damage shows atomic and renewable energy are unhappy bedfellows, 2017: https://www.euractiv.com/section/electricity/news/german-nuclear-damage-shows-atomic-and-renewable-power-are-unhappy-bedfellows/
Ökologie Institut, NPP Output Flexibility, 2010
Wärtsilä, Combustion engines vs. gas turbines, startup times https://www.wartsila.com/energy/learning-center/technical-comparisons/combustion-engine-vs-gas-turbine-startup-time
Climate Policy Initiative: Flexibility, the path to low-cost, low-carbon grids, 2017: https://climatepolicyinitiative.org/wp-content/uploads/2017/04/CPI-Flexibility-the-path-to-low-carbon-low-cost-grids-April-2017.pdf
ATW, Load cycling capability of German Nuclear Power Plants, 2010 http://www.ktg.org/kernenergie-wAssets/docs/fachzeitschrift-atw/2010/atw2010_09_waas_lastwechselfaehigkeiten_kkw_en.pdf
Lazard, Levelized cost of energy analysis 10.0, 2016: https://www.lazard.com/perspective/levelized-cost-of-energy-analysis-100/