The 2020 UK Energy white paper and replacing the UK's nuclear fleet

by John

16th December 2020

Version 0.2


The British government has just published its energy white paper Powering our Net Zero Future. In his introduction to the paper Secretary of State for Business Energy and Industrial Strategy Alok Sharma pledges to building 40GW of offshore wind by 2030: "enough to power every home in the UK", and talks about shifting from gas to electricity for home heating. On nuclear energy the paper says merely that it aims "to bring at least one large-scale nuclear project to the point of Final Investment Decision by the end of this Parliament, subject to clear value for money and all relevant approvals" and that "retiring capacity will need to be replaced to keep pace with existing levels of demand". However its "illustrative 2050 mixes" envisage a doubling to tripling of nuclear capacity compared to the current fleet:

Fig 3.4 from BEIS White Paper

The graphic shows figures in "TWh" - actually TeraWatt-hours per year. This is a convenient measure for bean-counters who can translate it into prices and costs (a TWh is a million 1kWh "Units" of electricity for which we consumers pay about 15p each) but it doesn't relate easily to the figures we're interested in such as 40GW of offshore wind or 3.2GW power stations like Hinkley Point C. A Terawatt-hour per year is 1,000 Gigawatt-hours per year which is just a Gigawatt times the number of hours in a year: 8,760 — so 1 TWh/y = 0.876 GW. So here for convenience is the same graphic with a scale in GW:

Fig 3.4 from BEIS White Paper scaled in Gw

Note that these figures are average amounts of power actually generated. Real world power plants don't generate their full power 24*7. Nuclear power plants have to be shut down for refuelling, inspection and maintenance every year or two, whilst the output of wind and solar varies between their full, "nameplate", capacities, and zero, depending on the state of the sun / wind. The average output (known as "capacity factor") of nuclear plants can be over 90%, but for offshore wind the figure is around 40% and for onshore around 30%. Solar in the UK has a capacity factor of around 10%.

So to get the 50GW or so of renewable power illustrated would require between 125GW of installed, "nameplate", capacity if it were all offshore wind, and 500GW if it were all solar.

The BEIS model

Note that these are just illustrations of possible mixes; the BEIS modelled "almost 7,000 different electricity mixes in 2050, for two different levels of demand and flexibility, and 27 different technology cost combinations" producing "a dataset comprising of over 700,000 unique scenarios, allowing us to identify common features of a low emissions, low-cost electricity system". More details of their analysis are in their "Modelling 2050: Electricity System Analysis". This finds that:

  • There is no single optimal technology mix; many capacity mixes can meet different carbon emissions levels at low cost. This is true for all levels of demand modelled.
  • Electricity system costs are lowest when carbon intensity is between 5-25gCO2/kWh.
  • All low-cost solutions include significant levels of wind and solar. Wind and solar generation could more than quadruple by 2050.
  • System flexibility reduces system costs. It does this by reducing curtailment of wind and solar and flattening demand for electricity, and therefore the overall capacity required. Our modelled options include batteries, demand side response and interconnectors.
  • All low-cost solutions also require other forms of low-carbon generation to provide resilience during extended periods of low wind and solar irradiation. Our modelled options to provide this are nuclear, gas generation with Carbon Capture, Usage and Storage (CCUS), and short-term dispatchable generation from unabated gas and/or low-carbon hydrogen.
  • Moderate levels of low-carbon hydrogen could replace unabated gas fired generation and reduce the requirement for other low-carbon generation. The extent of the impact is dependent on the quantity and cost of hydrogen available for generating electricity. We have only modelled the impact of low-carbon hydrogen-fired generation, but technologies that can offer longer-term storage than current technologies (i.e. batteries) could have similar impacts.

Their model assumes "plausible 2050 capacity ranges for those low-carbon technologies that are deployable at scale":

min max
Gas CCUS 2 GW 30 GW
Offshore Wind 40 GW 120 GW
Onshore Wind 15 GW 60 GW
Solar 15 GW 120 GW
Nuclear 5 GW 40 GW


The most noticeable feature of the "illustrative 2050 mixes" graphic is the huge proportion of renewables (shown in green) - around 50GW. As noted in the table above these comprise onshore and offshore wind, and solar: the Department is still evaluating the role of Biomass and does not include it in their current models. Nor are wave, tidal or other renewable sources considered: the models explicitly consider only technologies currently capable of being deployed at large scale.

daily and weekly graph of nuclear, coal, CCGT and wind for 6 Dec 2020

With such a high proportion of intermittent renewables it would be interesting to know how the BEIS model would cope with conditions such as those experienced in the UK in first week of December 2020 when, despite temperatures not being particularly cold, the lack of wind (and of course solar, at night) led to coal-fired power being brought online to make up a shortfall in supply.

(Graphics from GridWatch)

Nuclear and CCS

A key finding from the BEIS models is that for the most ambitious degrees of decarbonisation - a carbon intensity of electricity of 5gCO2/KW or less - the lowest cost solutions require:

nuclear gas CCUS total nuclear + gas CCUS
with hydrogen 15 - 30 GW 15 - 30 GW 35 GW
without hydrogen 20 - 40 GW 15 - 30 GW 50 GW

Although in the BEIS models nuclear and gas with CCUS are more or less interchangeable there are good reasons to prefer nuclear:

  1. Energy security: despite recent discoveries the UK's gas reserves are generally in decline, making the country increasingly dependent on foreign gas supplies.
  2. Imported liquefied natural gas (LNG) has a higher carbon footprint than non-liquefied due to the energy required to liquefy gas.
  3. Russia, one of the world's biggest suppliers of natural gas, sources significant amounts of gas from Arctic waters, with destructive environmental consequences for these formerly pristine environments.
  4. Burning gas, even with CCS, may still result in air pollution from nitrogen oxides (NOx), depending on the technology used.
  5. Nuclear energy may be able to offer negative emissions through coupling with Direct Air Capture plants using low-grade heat mostly wasted by the nuclear plant.

Let us therefore look at the higher end of the ranges for nuclear capacity - 30 GW if hydrogen is available, 40 GW without hydrogen. How might this be achieved?

Meeting the challenge: rebuilding the UK nuclear fleet

Our current nuclear fleet has a total capacity of 9GW. Most of our power stations are British-designed Advanced Gas-cooled Reactors (AGRs) which are coming to the end of their lives and will all be retired over the next decade, by 2030. We also have one Pressurised Water Reactor (PWR) at Sizewell B. This is licensed until 2035 but EDF, the plant owner, has expressed the intention of applying for a 20 year lifetime extension to 2055. (This is quite normal for PWRs which have, over the decades during which they have been in service across the world, proven to be reliable, safe, and long-lived.)

We are currently building only one new nuclear power station, the 3.2GW twin EPR at Hinkley Point C. As our AGRs retire our nuclear capacity will drop by half:

Current committment: Hinkley Point C

If Sizewell C - another twin EPR - is approved and comes online in the early 2030s our overall capacity will still be lower than at present:
Assuming retirement of Sizewell B in 2035

Fleet extended to Sizewell C

If Sizewell B is granted a lifetime extension to 2055

With lifetime extension for Sizewell B

Bradwell B, where a consortium comprising China General Nuclear and EDF want to build a CGN-designed HPR1000 reactor, could add another 2.3GW, taking us to just under 10GW, a modest increase on current capacity:

Fleet with HPC, SZC and Bradwell B

There are outline plans to build another single 1.6GW EPR at Moorside. This gives us a little more capacity:

Fleet with HPC, SZC, Bradwell B and one EPR at Moorside

What about the Rolls-Royce consortium's 400MW SMR, which should be available towards the end of the decade? The plans for Moorside include some of these. If we assume one is brought online in 2029 and a further 3, one every couple of years, that takes us to 13GW:

Fleet with HPC, SZC, Bradwell B and one EPR + 4 RR SMRs at Moorside

Beyond the outline

This is the extent of current, even outline, plans. It leaves us well short of even the minimum capacity we need to fulfil ambitious but economical mitigation plans in the BEIS model. What if we had more ambitious plans for our fleet? Suppose we were to build 2 EPR reactors instead of one at Moorside and keep on building EPRs, completing them at a rate of one new twin reactor station (like HPC and SZC) every five years. This takes us to 20GW; the minimum capacity we need without hydrogen, a little more than the minimum with it:

Fleet with continued build-out of EPRs

If, after Bradwell B, we build further HPR1000 stations, completing a twin reactor station every ten years, and also continue building Rolls Royce SMRs every 2 years, we get to over 31GW by 2050:

Fleet with continued build-out of EPRs + HP1000s + RR SMRs

Reality check

Is this too ambitious? Are the assumptions made here, about three new types of reactor to be built in the UK, where we haven't built a nuclear power station since the last millenium, realistic?


A major assumption is that we can build new twin-EPR power stations, pretty much a carbon copy of Hinkley Point C and the proposed Sizewell C, starting a new one every five years, and that each will take around 10 years to build. Given the horrendous delays that have plagued the first two EPRs to be started, at Olkiluoto in Finland and Flamanville in France, this might seem delusionally optimistic. However since those two projects started, two EPRs have been started and completed, now up and running, at Taishan in China, taking about 10 years to build. They were a joint venture between China General Nuclear and EDF, who are also in partnerhip in the HPC project. Both parties have an interest in getting the project completed and bringing in revenue as quickly as possible, and experience in the successful construction of the earlier project.


Another assumption concerns the proposed Bradwell B twin-HPR1000 power station, and others using these reactors. The HPR1000 (also known as "Hualong One") is based on mature designs which have been developed and used for several decades in China, which themselves are based on a proven French design derived from a design by Westinghouse. Unlike the EPR it doesn't have radical and possibly problematic new features. China General Nuclear is keen to build one in the UK in order to gain approval from a respected nuclear regulatory authority in a Western country in order to gain world-wide sales. As such it has an interest in carrying out a swift, trouble-free build. CGN would doubtless be happy to have several more built in the UK. I have assumed the same time to build as EPRs, but actually the first reactor of this type, at Fuqing in China, has just been completed in 5 years.

Rolls Royce Consortium SMR

There are no examples of this reactor design that have yet been built. However it is a Pressurised Water Reactor and Rolls Royce have been building PWRs for submarines and aircraft carriers for decades so they have the technical capabilities. The contribution of these reactors to the overall plan is relatively minor.

Where would we put them all?

The UK has nuclear power stations at Dungeness, Hartlepool, Heysham, Hunterston, and Torness which are to be decommissioned and for which new power stations are not currently planned, plus sites such as Wylfa for which there were plans for new power stations. It seems likely that sites for as many new power stations as we could realistically build in the next 30 years should not be a hindrance to a reasonably ambitious new nuclear building programme.

Has anything like this been attempted before?

Yes. Over a period of 20 years France built around 60 reactors giving about 60 GW of nuclear power, starting from practically nothing. They did this in response to the threat of having to pay significantly more for their energy after the oil price shocks of the early 1970s. They now earn several billion Euros per year in electricity exports.

French nuclear build-out

Over a similar period Sweden also built a nuclear fleet which, whilst smaller in size than the French, added capacity at an even faster rate per capita of their population.

Far from being overly ambitious, the nuclear building scheme outlined above is comfortably modest compared to what can be achieved by governments with a clear vision and plan for what they want and need to achieve.


Please leave comments via the contact page