NUC0014

Written evidence submitted by Mr Malcolm Grimston, Honorary Senior Research Fellow, Imperial Centre for Energy Policy and Technology, Imperial College

Abstract

This brief submission does not seek to address all of the questions before the Welsh Affairs Select Committee’s investigation into nuclear energy in Wales. There are others much better qualified than the present author to discuss the details of funding models, government support for new reactor concepts and the local economic benefits for the north Wales economy of nuclear nuclear capacity at Wylfa (and perhaps Trawsfynydd).

The comments below focus on electricity systems issue and the case for significant future nuclear investment, alongside some further renewable investment and, if feasible, technologies to allow for electricity ‘storage’. As such they address the main question What role can, or should, nuclear power play in achieving net zero and UK energy security? and, indirectly, What would the likely cost be to the taxpayer of the UK Government supporting the development of a new nuclear power station at Wylfa? Decades of evidence show that nuclear power is well suited to providing reliable, economic and environmentally acceptable supplies.

Key findings:

• Global interest in and public support for nuclear power has been growing, even before the present crisis in Ukraine, as concerns over the cost and security of fossil fuels and their environmental implications has increased.
• Almost all of the low-carbon capacity which will be delivering a zero-carbon economy by 2050 has yet to be built. Developing a mix of low-carbon sources, including nuclear, is the most cost-effective way of delivering on this aspiration.
• For a period such as July 11-21 2022 (a time of considerable stress on the UK grid associated with a heatwave which yielded the UK’s highest recorded temperatures to the point but a period that may become more typical as climate change develops), the effect of tripling nuclear output over current levels would be to reduce the use of gas for power (which ranged from 10GW to 18GW over the period) by a consistent 10GW throughout.
• By contrast, tripling wind and solar capacity, at a similar upfront cost (ignoring issues of transmission, balancing etc.), would result in a system prone to extreme fluctuations. At some times gas would still be generating 14GW; at others wind and solar output would be exceeding complete demand, resulting in negative real-time prices and presenting a much greater challenge in balancing supply and demand.
• Investment in nuclear power in Wales and the UK as a whole will bolster grid resilience at times of high demand and/or low availability of other low carbon sources, notably wind and solar.
• LCOE (Levelised Cost of Energy) is an increasingly misleading metric in systems with growing renewable penetration; when system costs are taken into account nuclear power remains a competitive source of low-carbon electricity.
• Nuclear power can be scaled rapidly once a manufacturing supply chain is established.
• Uranium is a very widespread resource globally and one with extremely high energy density, requiring very small land area, mineral requirements during construction and volumes of fuel relative to alternatives and creating very little waste. One consequence of this is that nuclear power is highly resistant to fuel price inflation, a vital consideration when seeking to ensure energy security at manageable cost at times of global energy shortages.

1. Introduction

It is often observed that energy policy, especially with respect to electricity, finds itself impaled on the horns of a trilemma. A successful energy policy will serve three goals, which at any time may be in tension with each other to a greater or lesser extent;

• preserving energy security;
• delivering economic supplies;
• protecting the environment.

In fact one can add to these a fourth: political and social acceptability. This is not entirely independent of the other three but has a separate existence from them, as the radically different energy policies pursued by countries with apparently similar energy systems, and indeed different political parties within any individual country, serves to demonstrate.

Recent crises in the global energy market, including but by no means restricted to the issue of Russian oil and gas supplies into Europe, duplicate previous events in illustrating that this is not a band of equals. In times of severe stress on the system politics comes first; insofar as they can be separated from the political imperative of the time, it seems clear that secure supplies comes next, followed by cost, with the environment often appearing as an optional extra, to be sacrificed easily. Japan’s response to Fukushima (involving an enormous increase in fossil fuel imports) and Germany’s recent decision to extend the lifetime of some of its coal-fired power stations are cases in point. (In July 2022 Japanese Prime Minister Fumio Kishida said he aimed to restart up to a further nine nuclear reactors to avoid an electricity shortage in the winter months in addition to the ten which have already restarted; there are even signs that Germany might delay closure of its final three reactors beyond the end of 2022.)

2. Nuclear politics

Nuclear power’s standing in the political sphere undergoes fluctuations that are perhaps a little more marked than those affecting other fuel sources. It is fair to say that at present nuclear power is in the ascendancy in many countries. Some 30 countries have announced their intention to build their first nuclear power plants in the near future – in four cases, viz. Belarus, Bangladesh, Egypt and Turkey, the first reactors are already under construction. South Korea and Belgium have reversed or delayed nuclear phase-out policies. Even in Germany public opinion towards nuclear power seems to be softening, especially with respect to retaining the last three operating nuclear plants which are due to close at the end of this year – a recent poll found that 70% of Germans wished for the plants to be kept running, including a majority of Green Party voters[1] – though it is still some considerable way from endorsing construction of new capacity.

In the UK as a whole, polling from the Nuclear Industry Association shows support for continuing to develop nuclear power alongside renewables and other low-carbon approaches growing, even before the current international energy crisis (Figure 1). Both the Government and the principal Opposition party (at national level) are supportive of new nuclear build.

Figure 1: proportion of people who agree that Britain needs a mix of energy sources to ensure a reliable supply of electricity, including nuclear and renewable sources

3. Energy security

There are three potential challenges for secure power supplies: a shortage of generating capacity; existing capacity being unavailable (e.g. shortage or interruption in fuel, including wind or sunlight for renewables; planned maintenance; breakdown; political decisions); and failure of the transmission and distribution system.

3.1              The need for new capacity

The UK has an imminent need for development of considerable volumes of new electricity capacity. By the end of this decade all existing coal capacity will be offline; four of the five operating nuclear power stations will have closed; many of the early CCGTs and windfarms will also have reached the end of their lives. Indeed, it is important to note that all of the low-carbon capacity that will (hopefully) bring us to carbon neutrality by 2050 has still to be built, with the possible exceptions of the Sizewell B nuclear power plant and some hydropower.

Quite how much new capacity is required will depend on the sources of that capacity. The National Grid Electricity System Operator (ESO) calculates ‘derating factors’ for various sources of electricity generation. Derating factors measure the ratio of the amount of power one can rely on coming from a particular power source against its theoretical, or ‘nameplate’, capacity (i.e. the maximum it could produce if conditions were perfectly favourable). So if a source ran with perfect reliability all the time its derating factor would be 1 or 100%; if one could never be sure when it was going to work and when not its derating factor would be 0. A source with 0% derating factor would need reliable backup equivalent to its entire nameplate capacity; a source with 50% derating factor would need reliable backup equivalent to half of its nameplate capacity and so on. NGC’s 2021 figures give derating factors of 91% for CCGT, 80% for nuclear, 78% for demand side reduction (DSR), 11% for offshore wind, 8% for onshore wind and 2% for solar power.[2] In other words, 100GW of solar power would need backup of between 98GW of reliable backup (almost certainly gas-fired in the UK context) to make a 100GW firm contribution to the grid, while 20GW of nuclear (which would provide approximately the same average output) would need just 4GW of backup. These factors have profound implications for the economics of low-carbon generation, as discussed below.

The following illustration refers to UK power demand and supply between July 10 and July 21 2022, in the period around the extraordinary heatwave that hit the country on July 18 and 19. Predictions of climate change suggest that such periods are likely to become more common in future years. Figure 2 shows actual supply and demand during this period; for simplicity, only the contributions from gas, nuclear, wind and solar are shown, alongside total demand.

Figure 2: UK power generation (GW), four-hourly intervals, July 10-21 2022[3]

During this period, not untypical of hot weather (when wind speeds tend to be low):

• the UK’s 27GW of wind capacity was generating at a maximum of 35% of rated capacity and at times was producing as little as 0.6GW and as much as 10.5GW;
• solar output fluctuated as expected during the day (and was of course zero overnight) – at the end of 2021 solar capacity stood at 14.6GW;
• nuclear produced a steady 5GW throughout from 6.5GW of capacity, some units being offline for statutory outage or operating at low power during investigations;
• the output from gas turbines in effect made up the difference but never dropped below 10GW, reaching as much as 17.5GW.

Let us consider two possible lower-carbon scenarios; one which would involve tripling the UK’s current nuclear generating capacity; the other tripling both solar and wind capacity. Assuming that existing capacity would need to be replaced, the new nuclear capacity which would have to be installed to triple present nuclear output would be of the order of 20GW, while that of wind and solar would be around 80GW and 45GW respectively. (These are inevitably extremely rough estimates and should be regarded as illustrative only.) The direct cost of installing this volume of capacity would be similar – perhaps between £100 and 150 billion in each scenario. However, nuclear investment would have the advantage of constant and predictable output; of reducing the need for gas-fired generating capacity for back-up requirements; and of not requiring a major new network of grid wires to carry the output from source to market. The former scenario (Figure 3), which would require building six new nuclear plants of the 3.2GW capacity of Hinkley Point C and Sizewell C, would allow gas demand to fall by 10GW throughout a period such as this, both reducing fuel use (and greenhouse emissions) and removing the need for a perhaps 16GW of gas-fired capacity (assuming an 80% derating factor).

Figure 3: UK power generation (GW), four-hourly intervals, July 2022 – tripled nuclear capacity

Figure 4 shows the effect of tripling both solar and wind capacity, assuming a commensurate increase in output. (Figures 4a-c separate out solar, wind and gas, for the sake of clarity of presentation; Figure 4d shows the effect on gas use of tripling nuclear output). Tripling solar capacity would reduce demand for gas during the daytime but would have no effect overnight. Trebling wind capacity would be less predictable. At some times it would have little effect on gas demand; at others trebling wind output would introduced an extra 20GW of power which would have a significantly deleterious effect on power prices, effectively rendering at least some of the wind output worthless (see Section 4.2).

Figure 4: UK power generation (GW), four-hourly intervals, July 2022 – tripled wind and solar

In this scenario, gas demand between July 10 and 21 2022 would have varied from 0 to 14.3GW; as the respective derating factors show, a major expansion of renewable sources of this nature would reduce the need for gas-fired capacity by only 9GW, though it would reduce fuel use considerably. In this scenario there are times when wind and solar together would be capable of generating more power than the system could absorb. This would be difficult to manage – either nuclear plants would need to be taken offline temporarily or renewable generators given considerable constraint payments to stop generating.

Figure 4a: UK power generation (GW), four-hourly intervals, July 2022 – tripled wind

Figure 4b: UK power generation (GW), four-hourly intervals, July 2022 – tripled solar

Figure 4c: UK power generation (GW), four-hourly intervals, July 2022 – effect of tripled wind and solar on gas use for power production

Figure 4d: UK power generation (GW), four-hourly intervals, July 2022 – effect on use of gas from tripled nuclear capacity

In terms of efficient use of resources, then, investing in new nuclear capacity would be considerably more effective in reducing dependence on imported gas and the emission of greenhouse gases, while reducing the need for gas-fired capacity and aiding grid stability, at least when output was varying as it did between July 10 and 21 2022.

(It should be noted that no 10-day period can be regarded as entirely ‘typical’ – there are times when for example the output from wind is consistently higher than it was between July 10 and 21 2022. Such times can offer even more extreme challenges in terms of balancing supply and demand and managing extended periods in which power prices were negative.)

3.2              Availability considerations

Uranium, the fuel for most nuclear power generation at present, is a plentiful and widespread mineral; the top five producers in 2021 were Kazakhstan, Namibia, Canada, Australia and Uzbekistan. Uranium, being an extremely concentrated energy source, is relatively easy to stockpile and the plants’ output is not materially affected by routine changes in the weather. In 2021 the average load factor (defined as the ratio of the energy that the available capacity actually produced during a specified period to the theoretical maximum energy that the capacity could have produced if it ran flat-out) of nuclear power stations globally was 77.5% – the figure was skewed by the Japanese fleet, the world’s third-largest, where many reactors have not been operating over much of the decade since Fukushima but are still technically ‘available’. The UK figure, at little more than 60%, reflected the age of the UK fleet and issues with the graphite in the earliest AGRs; in the US, which uses water-cooled reactors such as are proposed for the future UK nuclear programme, the figure was 92.7%.)[4] The load factors of offshore wind, onshore wind and solar in the UK in 2020 were 23%, 37% and 10% respectively.[5]

In the peak year for nuclear generation in the UK (1998), nuclear power accounted for some 27.4% of total UK electricity generation, at 99.5TWh. By 2021 output had more than halved to 45.9TWh, 14.7% of the UK total (Figure 5).

Figure 5: UK nuclear generation, TWh per year, 1965-2021[6]

Though this fall is precipitate and will continue as the AGR programme reaches the end of its life over the next six years, it should be noted that nuclear power can be installed very rapidly, if the desire is there to do so. Figure 6 shows the seven highest examples of development of low-carbon electricity generated per capita in various 10-year periods in different countries. As can be seen, the two most rapid deployments were of nuclear power, in France (1980-1989) and Sweden (1977-1986).

Figure 6: highest rates of development of low-carbon generation, MWh per capita, over 10-year periods

Nuclear power is sometimes painted as ‘inflexible’. In fact the experience of France over some decades shows that even ‘Generation 2’ GW+ scale nuclear power can cope with varying demand and deliver secure power supplies even at penetration levels over 70%. Modern large-scale reactor designs, such as the EPR being built at Hinkley Point (Somerset) and recently confirmed for Sizewell (Suffolk), and the US-based AP1000, are designed to load-follow and one might expect SMRs to be more flexible still. However, it is true that in a system with complementary capacity with lower capital costs and higher avoidable costs one would choose to run the nuclear stations whenever they were available, as relatively little money is saved by switching them off or reducing their output.

4. Economics

The economics of electricity supply are extremely complex, and become more so in systems with high levels of penetration of intermittent sources.

4.1              Traditional power supply systems

In ‘traditional’ power supply systems, consisting of a range of reliable, ‘dispatchable’ sources, the economics is relatively easy to calculate, at least in principle. To take a very simple system served purely by nuclear (20GW), coal (20GW) and gas (20GW), with 20GW ‘baseload’ (the lowest demand experienced, usually early in the morning in late July or early August for the UK) and a total of 60GW at ‘peak’ (usually early evening in late January). Assuming no carbon restraints, this system would run the nuclear stations (with high fixed costs but low ‘avoidable’ costs[7]) flat out; would being in coal stations (with higher avoidable costs and lower fixed costs) to cover the next 20GW of demand; and use gas-fired plants (CCGT or OCGT) at times of highest demand (Figure 7).

Figure 7: Illustrative cost profiles (% of total costs) of power sources, 10% rate of return

One consequence of this is that, because fuel prices represent a very low proportion of overall nuclear costs, nuclear power is relatively very resilient to fuel price inflation when compared to gas or coal-fired generation. (By the same token nuclear power benefits less at times of very low fossil fuel prices but this appears unlikely in at least the foreseeable future.)

The system operator – often the owner of the grid – would rank available capacity in terms of ‘merit’, high merit involving low avoidable costs, low merit involving high avoidable costs. On a half-hourly basis the system operator would ‘call up’ plant, starting with the highest merit, until enough capacity was committed to cover expected demand for that half-hour. There are various ways of managing this but one common one is that all generators are then paid at the contracted cost for the last plant to be called up. In practice, of course, the system would need to keep a ‘capacity margin’ to deal with the times when peak demand was higher than expected or when plant was unavailable for one reason or another. However, in such a system cost will be quite easy to calculate, in principle. The value of a unit of output does vary throughout the day and year but in a relatively predictable fashion so the average cost of a delivered unit (known as Levelised Cost of Energy or LCOE) – i.e. the total costs of running the plant throughout its lifetime (including construction and decommissioning) divided by the total number of units it will have generated – is a useful metric in determining the system costs and guiding future investment. Since the system operator can call up the amount of output required to meet predicted demand in real time, the value of a unit rarely if ever drops below zero and is set by the avoidable cost of the ‘marginal’ generator, i.e. the last plant to be called up to meet that predicted demand.

(Of course in reality things are more complex. For example, the rate of return demanded by the investor is crucial to the LCOE of sources, most notably nuclear. At a 5% demanded rate of return nuclear costs are reduced considerably when compared to 10%, while the effect of such a change on the LCOE of gas-fired capacity is much more modest since most of the costs is accounted for by the fuel, not the initial investment. A Regulated Asset Base approach to nuclear funding would be highly beneficial compared to the funding regime which has been employed at Hinkley Point C. Under RAB, the owner of the new plant can receive an income during the construction process rather than waiting for the plant to be completed and to start generating output. This significantly reduces the risk of schedule and cost overruns from the point of view of the owner and thereby reduces the demanded return on capital, though of course at the expense of moving some of the economic risk of the project from the investor to the consumer.)

4.2              The effect of increased renewable penetration

The picture becomes much more complex when significant amounts of intermittent capacity and output penetrate the system. Because electricity cannot be stored at grid scale (at least at present) – the 3GW of pumped storage at Dinorwig in north Wales is used more to balance supply and demand than to ‘store’ electricity as such – demand and supply must be maintained in real time within extremely narrow margins. When renewables are not generating very much – e.g. overnight for solar or in periods of low (or very high) wind speeds for wind – other plants, with higher running costs, have to be pressed into action and real time power prices rise. On the other hand, when renewables are generating at or near capacity, the real time power price is depressed. As renewable penetration grows this becomes a more prevalent issue (see Figure 3). Generating more electricity than is needed to meet demand can be as challenging as not generating enough – potentially it can cause surges in the system that can damage electronic equipment and in the worst case can melt the grid wires. As a result, even at quite moderate levels of penetration there are growing periods when electricity has a negative market value.

A striking example is provided by South Australia (Figure 8). Throughout September of last year, average power prices in the State were negative from 07:00 until 15:30 as renewables (mainly rooftop solar) swamped the market.[8]

Figure 8: average wholesale prices of electricity and sources, South Australia, September 2021

Similar periods of very low or negative wholesale prices have been observed in California, Germany and even the UK. Yet Australia had to suspend the entire operation of its National Electricity Market (NEM) in June 2022 as power prices reached unprecedented high levels, owing in part to a reduction in the availability of renewables in the previous months.

Such systems can also face challenges in terms of managing the daily transition from high solar output to zero and back. California’s power grid has now developed a characteristic ‘duck curve’ shape. Figure 9 illustrates the phenomenon. In 2012 the California Independent System Operator (CAISO) gave the effect its soubriquet as it is supposed to resemble the shape of a duck. The output of conventional power sources has to increase very rapidly around sunset to replace the considerable output of the State’s solar fleet at the very time that demand is also growing towards the evening peak; which creates strains on the substations and other infrastructure. The phenomenon has developed rapidly over the last decade. (In Hawaii the higher proportion of solar has led to an even more pronounced version known as the ‘Nessie’ curve.) While solar power may be of limited direct use in Wales, the phenomenon does illustrate the challenges of managing power sources whose output can vary very significantly owing to weather conditions.

Figure 9: hourly electrical load versus load without solar and wind – ‘duck curve’ – California, March 31, 2012-2020[9]

Obviously, even at quite significant levels of renewable penetration, there will still be considerable periods when demand is high and reasonable renewable output can find a market at a reasonable price. However, as penetration grows this phenomenon creates an awkward dilemma. Without major market intervention, renewable output tends to cannibalise itself. All else being equal, when renewables are generating strongly they don’t earn much money as they depress the market price; when they are not generating much the market price rises but they do not benefit greatly. Left to the free market renewables could not function; they would have to either expose their output to the real time market, or sign longer-term contracts and find themselves having to dip into the real time market to fulfil them when output was low. If the product is of very low or negative value then it is in a sense irrelevant how much it costs to produce it. In these circumstances LCOE becomes a highly misleading metric when coming to calculate the real costs of intermittent sources which fall to consumers.

4.3              Market distortion and hidden subsidies

As a result, governments sanction profound intervention in the marketplace. Renewable output is contracted on a ‘must-take’ basis backed up by guaranteed prices (or ‘strike prices’, via ‘contracts-for-difference’). In other words, power suppliers (the customers for wholesale power production) will have to take renewable output whenever it is available, in preference to other sources. This is in principle a two-way measure – when the market price is above the strike price agreed with the renewable generators then money flows from the generator to the consumer – and a similar approach has been taken with regard to the financing of the Hinkley Point C nuclear station currently under construction. However, the main effect is to displace the cost of the variability in renewable output away from the producers and onto the electricity consumers as a whole. The approach is backed up by ‘constraint payments’, whereby generating companies are paid to stop generating altogether. These payments have been growing enormously over the last decade in the UK (Figures 10a & 10b). (2021 was a relatively poor year for wind coverage, resulting in lower constraint payments; constraints in the first half of 2022 amounted to almost as much as the whole of 2021.)

Figure 10a: constraint payments to UK wind generators (£ million)[10]

Figure 10b: output constraints to UK wind generation (TWh)

4.4              Grid implications

A major new programme of nuclear power capacity would require considerable extension and strengthening of grid connections to these new plants. One advantage of new build occupying sites that have previously hosted nuclear (or other) power plants is that some such wire capacity, or at least its route, may already be in place. This being said, to construct say 3.2GW of new nuclear capacity at Wylfa would require a very substantial upgrade to the grid capacity that served the 1GW Magnox plant. However, the grid implications for installing sufficient renewable capacity to provide the same output as say 20GW of nuclear capacity would be much greater, first because the amount of nameplate capacity would be much larger and secondly because of the need to retain reliable back-up capacity for those periods when renewables were generating very little. In many countries, notably Australia and Germany, deployment of renewables has slowed dramatically as protests grow against the environmental consequences of power lines and other aspects of the technologies. Germany did not install a single offshore wind generator in 2021.[11]

In summary, the economics of power production is complex and LCOE is an increasingly inadequate metric to use in determining the overall costs of any particular policy. System costs, associated with the intermittency and low power density of renewables, represent a growing proportion of total costs as renewable penetration grows and these costs must be properly taken into account when designing a fit for purpose low-carbon system. In general, the system costs of nuclear power are very much lore than those of intermittent sources.

5. Environment

In environmental terms, nuclear power is very low carbon, on a par with if not below renewable sources. As it is a highly concentrated source of energy, the mineral requirements over the whole lifetime of a nuclear project are considerably less than those of any other low-carbon source of electricity (Figure 11). The volumes of potentially hazardous waste arising from the use of nuclear power are therefore very low when compared say to those arising from solar power; while there has been considerable progress in the long-term management of radioactive waste, notably in Scandinavia, as yet no similar programmes have been developed for the much larger waste streams from renewables.

Figure 11: minerals used in clean energy technologies, kg/MW, compared to other power generation sources[12]

It is striking that a considerable number of long-term environmental activists, such as George Monbiot, Mark Lynas, Michael Shellenberger and Chris Goodall, have in recent years changed their earlier stance and become firm advocates for major new nuclear power programmes.

6. Conclusions

Nuclear power has a long pedigree of providing secure, low-carbon electricity in over 30 countries globally, with many others recently having stated intentions to ‘join the nuclear club’. Wales, with its impressive nuclear pedigree, has the experience, infrastructure and opportunity to develop both a new large-scale plant (at Wylfa) and a Small Modular Reactor (at Trawsfynydd). At a time when the tripartite challenge of providing secure power supplies at reasonable cost and lower climate change consequences has never been clearer, the UK should place itself centrally at the forefront of the global revival of interest in installing large quantities of new nuclear power capacity.

August 2022