Applied Energy 364 (2024) 123128

Contents lists available at ScienceDirect

Applied Energy
journal homepage: www.elsevier.com/locate/apenergy

A nuclear future? Small Modular Reactors in a carbon tax-driven transition to

clean energy
Wanni Xie a , John Atherton a,b , Jiaru Bai a , Feroz Farazi a , Sebastian Mosbach a,b ,
Jethro Akroyd a,b , Markus Kraft a,b,c,d ,∗
a
Department of Chemical Engineering and Biotechnology, University of Cambridge, West Cambridge Site, Philippa Fawcett Drive, Cambridge CB3 0AS, United
Kingdom
b CARES, Cambridge Centre for Advanced Research and Education in Singapore, 1 Create Way, CREATE Tower, #05-05, Singapore, 138602, Singapore
c School of Chemical and Biomedical Engineering, Nanyang Technological University, 62 Nanyang Drive, Singapore, 637459, Singapore
d
The Alan Turing Institute, London, NW1 2DB, United Kingdom

GRAPHICAL ABSTRACT

ARTICLE INFO ABSTRACT

The study investigated the effect of a CO2 tax to encourage the adoption of Small Modular Reactors (SMRs)
as an alternative to fossil fuels for power generation in the UK. The trade-offs of different SMR placement
policy options with respect to the competing objectives of minimising transmission losses and population risk

∗ Corresponding author at: Department of Chemical Engineering and Biotechnology, University of Cambridge, West Cambridge Site, Philippa Fawcett Drive,
Cambridge CB3 0AS, United Kingdom.
E-mail address: [email protected] (M. Kraft).

https://doi.org/10.1016/j.apenergy.2024.123128
Received 31 August 2023; Received in revised form 15 February 2024; Accepted 27 March 2024
Available online 9 April 2024
0306-2619/© 2024 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
W. Xie et al. Applied Energy 364 (2024) 123128

Dataset link: https://github.com/cambridge-ca were investigated to understand the impact on the optimal placement and usage of SMRs at different carbon
res/TheWorldAvatar, 10.17863/CAM.99921 tax levels. Different assumptions about renewable power availability were explored. The study identified the
most cost-effective number of SMRs per site and optimised the power flow for cost efficiency. Regardless of
Keywords:
Small Modular Reactors (SMRs)
renewable power availability, a carbon tax in the range of £45–60/t was found to incentivise the full adoption
Carbon tax of SMRs with a levelised cost of electricity of £60/MWh versus £0–20/t at £40/MWh. The SMR placement
Placement policy influenced the performance and cost of the energy system, as well as whether a region acted as a net importer
Clean energy transition or exporter of energy. The most cost-effective solutions were achieved by balancing transmission loss and
Fossil fuel replacement population risk.
Electrical power generation

in their simulations, with a moderate tax (approx. €30/t) leading to

Nomenclature a significant cut in emissions by incentivising the replacement of coal
AMR Advanced Modular Reactor with natural gas.
ASEAN Association of South East Asian Nations The UK has outlined plans for the adoption of low-carbon energy
technology. The recent ‘Energy white paper’ [11] and ‘Ten-point plan
BMRS Balancing Mechanism Reporting Service
for a green industrial revolution’ [12] both identify offshore wind and
CAPEX Capital expenditure nuclear power as crucial components of decarbonising power genera-
CCS Carbon Capture and Storage tion. Point 1 of the ‘Ten-point plan’ targets 40 GW of offshore wind
CPF Carbon Price Floor capacity by 2030. This represents a substantial increase compared to
ETS Emissions Trading System current capacity (11 GW onshore, 12.7 GW offshore) [13]. Point 3
LCOE Levelised Cost of Electricity promotes the development of nuclear energy projects, including new
NSGA2 Non-dominated Sorting Genetic Algorithm large-scale plants, the deployment of small modular reactors (SMRs)
II and the development of advanced modular reactors (AMRs). This stands
OPF Optimal Power Flow in contrast to a decrease in the annual share of electricity provided by
nuclear power from 23% to 14.9% between 2000 to 2021 due to the de-
SMR Small Modular Reactor
commissioning of ageing plants [14]. Currently, only six conventional
WH SH High wind and high solar power availabil-
nuclear plants with a combined capacity of 6.8 GW remain in operation
ity
in the UK [13]. By 2035, most of the remaining plants will have reached
W H SL High wind and low solar power availability the end of their operational life, with only one new 3.2 GW European
W L SH Low wind and high solar power availability Pressurised Reactor plant currently under construction at Hinkley Point
W L SL Low wind and low solar power availability C. If no others are constructed, the nuclear capacity of the UK will be
W M SM Mid-level wind and mid-level solar power reduced to approximately one-third of its current level by 2050 [15,16].
availability However, in 2022, the UK published a target to introduce up to 24 GW
of new nuclear capacity by 2050, aiming to contribute approximately
25% of electricity generation. This represents a 3.5-fold increase in
nuclear capacity compared to 2022 [4,14,17] and is consistent with the
vision of the ‘Ten-point plan’, indicating a strong focus on developing
1. Introduction nuclear capacity over the coming decades.
The UK performed a techno-economic assessment of small modular
Climate change caused by greenhouse gas emissions is a global reactors (SMRs) in 2016 [18]. The analysis estimated the advantages
problem. Electric power generation accounts for approximately 16% to the UK economy and society, including £20 billion of undiscounted
(53.7 MtCO2 e in 2022) of greenhouse gas emissions in the UK [1]. gross value added if 70% of supply chain components were manufac-
The corresponding energy mix consisted of fossil fuels (41.5% natural tured in the UK between 2017 and 2040. Cárdenas et al. [19] examined
the impact of nuclear power as a baseload component of a carbon-free
gas, 1.6% coal), conventional nuclear power (16.7%), wind (28.8%),
electricity system in the UK. They showed that conventional nuclear
solar (4.6%), hydroelectricity (1.2%) and biomass (5.6%) (2022) [2,3].
plants would be expensive relative to renewable energy, but that SMRs
The UK must replace fossil-fired power generation with low-carbon
showed promise in achieving cost-effective power generation by sup-
alternatives if it is to cut emissions [1]. It must simultaneously accom-
plying approximately 80% of demand. Conversely, Asuega et al. [20]
modate a projected 50% increase in power demand by 2035, driven
concluded that while SMRs do not exhibit significant economic benefits
by support for electric vehicles, heat pumps, and hydrogen [4]. It aims
compared to large reactors in the UK, their advantage lies in the
to achieve 95% low-carbon power generation by 2030, and to reduce
factory approach to production, which mitigates construction delays
the emission intensity of power generation to less than 5 gCO2 /kWh by
and cost overruns. Steigerwald et al. [21] investigated the uncertainties
2035, compared to 182 gCO2 /kWh in 2022 [4,5]. associated with estimating production costs for SMRs. They showed
Different forms of carbon tax have been introduced to incentivise that SMRs may not be cost-effective, with no positive net present value
the development and deployment of low-carbon technology [6–8]. The observed over their lifespan, and estimated the mean Levelised Cost of
UK introduced a Carbon Price Floor (CPF) to support the EU Emissions Electricity (LCOE) for the 470 MW SMR prototype developed by Rolls-
Trading System (ETS) in 2013 [8]. The CPF taxes fossil fuels used for Royce as approximately 222 USD2020 /MWh. This is considerably more
power generation. It consists of two components: an ETS allowance than the £40–60/MWh estimate published by World Nuclear News
price and a Carbon Support Price that supplements the ETS allowance [22], the low end of which is similar to estimates of the future LCOE
price to reach a target carbon price. The impact of carbon pricing and between 2030 and 2040 for wind (which is more abundant than solar)
subsidies for renewable energy have been widely investigated. Mo and power in the UK [23].
Zhu [9], for example, studied how to design a CPF to promote invest- The potential of SMRs outside the UK has also been studied. Nian
ment in carbon capture and storage (CCS). They estimated that a price et al. [24] examined the economic competition between nuclear (con-
of €20/t would stimulate investment, while €30/t would be required ventional and SMRs), wind and solar power in the context of replacing
to sufficiently promote carbon abatement. More recently, Gugler et al. fossil-fuelled power generation, providing insights into net-zero path-
[10] compared the outcome of a CPF with subsidies for wind and solar ways for the Association of South East Asian Nations (ASEAN) coun-
power in the UK and Germany. Carbon pricing outperformed subsidies tries. They proposed a life cycle analysis framework to assess the LCOE

2
W. Xie et al. Applied Energy 364 (2024) 123128

and total emissions, highlighting that the driving factor for nuclear employs the Non-dominated Sorting Genetic Algorithm (NSGA2) [34]
power is to reduce carbon emissions by replacing fossil fuels. A team to address two objectives. Objective 1 aims to minimise the capital
from the University of Regina have extensively investigated where to expenditure (CAPEX) and risk costs associated with the placement of
site SMRs in Saskatchewan, Canada. Gao et al. [25] proposed a factorial the SMRs, while objective 2 aims to minimise the demand-weighted
optimisation-based SMR siting (FOSS) method to choose where to site distance between the SMRs and centres of demand. The distance metric
SMRs within a general electricity-system framework. They considered serves as a proxy for minimising transmission losses and ensuring
the competitive relationship between CCS and SMR and showed that balanced grid operations. The demand is represented as a load on
replacing decommissioned coal-fired power plants with SMRs would each bus of the transmission grid model. The transmission distance
contribute to at least 65.6% reduction in emissions by 2045, compared and risk cost metrics chosen as indicators for SMR site selection are
to a 2018 baseline. Gao et al. [26] extended the analysis to encompass based on the criteria outlined in the studies [35–37]. The candidate
all of Canada. The approach was modified by Zhang et al. [27] to locations for the placement of SMRs were the sites of existing fossil-
integrate climate, economic and social factors to reflect the long-term fuel (68), former fossil-fuel (8) and former nuclear (6) power stations.
effects of climate change to support site selection. Liu et al. [28] further The algorithm excludes consideration of the residual value or economic
extended the approach to identify patterns of SMRs and wind farm feasibility for reusing existing site infrastructure. The reason lies in (i)
siting in Saskatchewan. The site selection algorithm was observed to the ageing infrastructure at the chosen sites and (ii) that each SMR
give priority to power stations with large capacities and independent module incorporates all essential infrastructure. The bespoke reuse
transmission grids. infrastructure would run counter to the design intent of the SMRs,
While these studies raise questions about the economic performance which is to provide an economy of scale by producing multiple units
of SMRs, the assessments vary depending on the system specifications of the same design. The optimisation was subject to the constraints
and criteria chosen by researchers including the indicators used and that no more than four SMRs could be placed on a given site and that
the scope of analysis. Clearly, these issues contribute to the debate. SMRs could not be placed on two sites that are in close proximity. The
Notwithstanding this, the ‘Ten-point plan’ [12] specifies a clear target combined capacity of four SMRs is approximately 1.9 GW, similar to the
for the introduction of up to 24 GW of nuclear capacity in the UK, with capacity of the sites considered for replacement by SMRs. The choice
SMRs identified as a key part of the plan. Rolls-Royce is developing to frame the placement problem in this way ensures that the algorithm
a 470 MW ‘UK SMR’, with an estimated capital cost (CAPEX) of £1.8 can only place SMRs in locations that have existing connections to the
billion per SMR and an estimated LCOE of £40–60/MWh over a lifespan transmission grid, and that any newly placed SMRs are commensurate
of 60 years [22]. In March 2022, the UK government begin a three-step with the capacity of the connection. At each iteration, NSGA2 identifies
generic design assessment of the Rolls-Royce SMR. The first step ran a feasible solution expressed in terms of the location of each SMR, and
from April 2022 to March 2023 and agreed on the scope and schedule the corresponding CAPEX and risk cost.
for technical engagements. The second step started in April 2023 and The second step samples the feasible solutions by taking the
will perform a 16-month assessment of the fundamental acceptability weighted sum of the two normalised objectives to give a single weighted
of Rolls-Royce SMR design [29]. So far, Rolls-Royce has identified four objective. The choice of weight allows the exploration of the effect
potential sites for deployment of the SMRs, Trawsfynydd, Sellafield, of different placement policy options. The annualised total cost is
Wylfa, and Oldbury [30]. calculated for each sampled solution
A number of questions regarding the potential deployment of SMRs ∑
𝐿
𝐶OPEX + 𝐶Emissions
in the UK remain unaddressed. How many would be needed? Where 𝐶CAPEX + 𝐶Risk +
would they be placed? Are the sites suggested above sufficient? Which 𝑙=1 (1 + 𝑑)𝑙
𝐶T = , (1)
fossil fuel plants should be replaced and in what order? 1 − (1 + 𝑑)−𝐿
The purpose of this paper is to investigate what level of carbon tax 𝑑
would be required to incentivise the widespread adoption of SMRs and where 𝐶T represents the annualised total cost as a net present value
to estimate how many SMRs would be required to replace fossil fuels divided by an annuity factor. 𝐶CAPEX and 𝐶Risk denote the CAPEX and
for electrical power generation in the UK. We consider explicit carbon associated risk cost. 𝐶Risk is proportional to the population located
taxation, which directly assigns a monetary value to CO2 emissions. within a given radius of each site selected to host SMRs, where the
The study demonstrates a systematic algorithm for the optimal siting radius is proportional to the capacity of the SMRs on the site. 𝐶OPEX
of SMRs that takes into account the geospatial configuration of the and 𝐶Emissions indicate the annual operating expenditure (OPEX) and
existing transmission grid, and seeks to balance the trade-off between emission cost. The emission cost is proportional to the carbon tax,
siting SMRs away from centres of population but close to areas of which is provided as an input parameter. 𝐿 and 𝑑 are model param-
demand to minimise risk while maximising efficiency. It shows how eters corresponding to the lifespan of the SMRs and discount rate,
different siting criteria impact cost, transmission losses and the energy respectively.
independence of different regions, providing a framework to support The OPEX and emission cost for each sampled solution are deter-
future decisions. mined by an Optimal Power Flow (OPF) analysis using the 29-bus
model. The OPF calculates the optimal output from each generator,
2. Methodology the voltage magnitude at each bus and the transmission loss across
each branch in the network. This provides information to support the
Fig. 1 shows the optimisation algorithm used to assess the cost and evaluation of each sampled solution in terms of the contribution of
placement of the SMRs. For the purpose of this analysis, we consider each generator to the power mix, the stability of the system and the
SMRs with a capacity of 470 MWe at an estimated cost of £1.8 billion efficiency of power transmission. The emissions from each generator
per unit, with scenarios for an LCOE of both £40/MWh and £60/MWh, were estimated using emissions intensities from Staffell [38].
based on the nominal specifications of the Rolls-Royce design [22,31]. The second step repeats the OPF analysis for five scenarios for the
The analysis makes use of a 29-bus model of the UK high-voltage power availability of wind (W) and solar (S) power: WH SH , WH SL , WM SM ,
transmission system [32]. The model consists of 29 buses, 99 branches WL SH , and WL SL . The availabilities are categorised as high (H), medium
and 1129 generators. The results presented in this study were obtained (M), and low (L), and are determined by the maximum (H) and min-
using the 29-bus model, and cross-checked using a 10-bus model [33] imum (L) weekly average outputs, and average (M) annual outputs
to verify that they were insensitive to the choice of model. reported for wind and solar power by the Balancing Mechanism Report-
The algorithm consists of two steps. The first step selects sites for ing Service (BMRS) [2] via the National Grid: Live website [3] for 2022
the placement of SMRs. The number of SMRs is provided as an input. It (wind: 17 GW maximum weekly output, 2.89 GW minimum weekly

3
W. Xie et al. Applied Energy 364 (2024) 123128

Fig. 1. Algorithm used to calculate the cost and placement of SMRs.

output, 8.9 GW average annual output; solar: 2.9 GW maximum weekly scenarios with SMRs that have an LCOE of £40/MWh and £60/MWh.
output, 0.25 GW minimum weekly output, 1.4 GW average output). The carbon tax penalises emissions, triggering the adoption of SMRs as
Each scenario is used to define an operational rate as a proportion the level of tax is increased.
of total capacity (wind: 23.6 GW, solar: 4.74 GW) that is applied Fig. 2(a) shows a transition window (indicated by the blue shading)
uniformly to all wind and solar generators respectively. for SMR adoption for a carbon tax in the range £0–20/t for an LCOE of
The OPF analysis allowed the output of the SMRs to vary from 0 to £40/MWh and £45–60/t for £60/MWh. SMRs are adopted gradually
90% (the normal operating point for nuclear power generation [39]) within the transition window, where the precise point of adoption
of capacity. While it must be understood that is not feasible for real depends on the specifics of each site. Beyond the transition window,
SMR operations, it is a deliberate feature of the analysis and is used full adoption of SMRs is cost-effective. The final number of SMRs is
to provide insights about the SMR capacity that is required in each sensitive to the assumptions about the availability of renewables in
location. The analysis allowed the output from generators fuelled by each scenario as should be expected, but insensitive to the cost of the
coal, oil, and natural gas to vary from 0%–90% (their normal operating SMRs. This is because the number of SMRs needed to achieve zero
point [39]) of capacity. This provides insight into the required output emissions is determined by the required generation capacity once all
from each fossil-fuelled generator as the carbon tax increases and SMRs clean energy sources have been fully used, making it sensitive to the
are introduced. Low outputs (i.e, close to 0%) indicate that a generator availability of wind and solar. Conversely, the transition window is
is no longer required (under that scenario), and that it could potentially insensitive to the assumptions about the availability of renewables, but
be considered for decommissioning. The output of the wind and solar sensitive to the cost of the SMRs. The LCOE of the SMRs influences
generators was allowed to vary up to the limit imposed by the scenario. the economic feasibility of replacing fossil-fired generation with SMRs,
However, the OPF analysis always returned solutions in which wind defining the transition window without impacting the final SMR count.
and solar operated at the prevailing upper bound. Setting a distinct The location of the transition window is therefore sensitive to the LCOE,
lower bound broadens the feasible solution space, crucial for algorithm and if the LCOE exceeds the £60/MWh value published by Rolls-Royce,
convergence by allowing output fluctuations during intermediate it- the level of carbon tax required to incentivise the adoption of SMRs will
erations. The output of conventional nuclear, hydro, and bioenergy exceed £60/t.
generators was constrained in the ranges 0%–70%, 0%–25% and 35%– The maximum cost-optimal number of SMRs varies between scenar-
50% of capacity respectively. Again, the OPF analysis always returned ios, ranging from 12–13 SMRs (5.6–6.1 GW), depending on the LCOE in
solutions at the upper bounds of these ranges, where the upper bounds the most renewable-abundant scenario (WH SH ), to 49 SMRs (23 GW)
were selected to match the average output observed in 2022 [2]. This in the least renewable-abundant scenario (WL SL ). This highlights the
treatment was necessary to facilitate the numerical convergence of the significant impact of the assumptions about the availability of renew-
OPF solver. able energy on the required capacity of SMRs. It is noteworthy that the
Full details of the model specification, model parameters, data minimum cost-optimal number of SMRs is greater than zero at a carbon
sources and sites considered by the analysis are given in the Supplemen- tax of £0/t for the combination of £40/MWh LCOE and low availability
tal Information. The analysis was implemented as part of The World of renewables. Under these conditions, SMRs have lower operating
Avatar (TWA) project, which uses a knowledge graph to provide a costs than oil, coal and natural gas generators. Some scenarios show the
principled approach to integrating data from different sources [40]. The introduction of an additional SMR beyond the transition window. This
TWA is designed such that it would be straightforward to repeat the is due to the displacement of small residual amounts of fossil fuel from
analysis elsewhere in the world. the power generation mix. The capacities involved are small relative to
the capacity of an SMR, hence a high level of carbon tax is required
3. Results before this becomes cost-effective.
Fig. 2(b) shows how the CO2 emissions reduce as the number of
3.1. Cost-optimal number of SMRs, overall cost and emissions SMRs increases. The emissions for each scenario reduce to almost
zero across the transition window. There are some residual emissions
Fig. 2 shows how the cost-optimal number of SMRs, and the corre- beyond the transition window due to power generation from biomass
sponding emissions and cost vary as a function of the carbon tax for and waste incineration, which the OPF analysis allows to persist as

4
W. Xie et al. Applied Energy 364 (2024) 123128

Fig. 2. Cost-optimal number of SMRs, annualised cost and carbon emissions versus carbon tax for SMRs (weight = 0.5). Left: £40/MWh. Right: £60/MWh. Solid lines show
scenarios corresponding to the operation of an electrical system with SMRs. Dashed lines show data for scenarios corresponding to the operation of the current electrical system
without SMRs.

per the 2022 energy mix. The dashed lines show how the cost-optimal The model calculates emissions of 49.5 MtCO2 from the current
operation of the current energy system (without any SMRs and subject energy system in the WM SM scenario at a carbon tax of £30/t. This
to the same assumptions about the availability of renewables) would serves as a useful point of comparison because it corresponds to an
respond to the carbon tax. The model predicts a transition window energy mix that closely matches the average power mix in 2022 [3],
(indicated by the grey shading) where natural gas displaces coal and so allows an assessment of the model against the real energy system.
oil. Whether the SMR transition window occurs before or after this is The 49.5 MtCO2 estimated by the model is consistent with an esti-
sensitive to the cost of the SMRs. The switch to natural gas reduces mate of 49.7 MtCO2 calculated using the average power mix with the
emissions, but significant residual emissions remain after the transition same emissions intensities as the model, and consistent with officially
in the absence of SMRs. reported emissions of 53.7 MtCO2 [1]. Full details of the emissions

5
W. Xie et al. Applied Energy 364 (2024) 123128

pricing mechanisms to encourage the transition towards low-carbon

energy generation.
Fig. 3 shows the capacity of the current electrical power system
in Great Britain (2022), overlaid by the average generation (2022)
and capacity corresponding to different numbers of SMRs. The differ-
ence between average generation and overall capacity is indicative of
the required reserve capacity. The SMRs correspond to a significant
proportion of capacity and at the upper end, have a similar capacity
to current coal, oil and natural gas generators. 52 SMRs would be
sufficient to meet the current target of 24 GW nuclear capacity by
2050 [17], while 36 SMRs would be sufficient to provide 16.9 GW,
which is the difference between the target and current conventional
nuclear capacity.

3.2. Trade-offs in SMR placement

Fig. 4 shows that the competing objectives to place SMRs to min-

imise both site-demand distance and investment and risk cost form
a Pareto front. The panel on the left shows the Pareto front formed
by feasible solutions with 33 SMRs. The arrow labelled (A) shows a
Fig. 3. Capacity of the electrical power system in Great Britain (2022) overlaid by region of decrease in site-demand distance for relatively little increase
the capacity corresponding to different numbers of SMRs. The dashed lines show the
in cost, while the arrow labelled (B) shows a region of increase in cost
average demand in Great Britain (2022) [3] and the target for nuclear capacity in Great
Britain by 2050 [17].
for little reduction in site-demand distance. The competing objectives
are combined into a single objective. Low weights prioritise placement
near centres of demand as a proxy to minimise transmission losses.
High weights prioritise placement away from centres of population to
intensities and energy mix for this comparison are reported in the
minimise risk. The panel on the right shows the Pareto front for feasible
Supplemental Information.
designs with different numbers of SMRs. The figure is annotated with
Fig. 2(c) shows that the cost increases as the number of SMRs
lines representing the UK target of 24 GW nuclear capacity by 2050,
increases, with corresponding differences between scenarios. As should and the 16.9 GW difference between the target and current conven-
be expected, the total cost is sensitive to the assumptions made with tional nuclear capacity. The shaded area spans weights in the range
respect to the availability of renewables and the cost of the SMRs. 0.25–0.75. Designs in this region fall on the low-cost side of the apex of
The cost levels off beyond the transition window in the scenarios that the Pareto front and balance both objectives. The designs corresponding
introduce SMRs, corresponding to full adoption of SMRs and almost to the points labelled (A), (B) and (C) are considered in more detail
zero emissions. In contrast, the cost of scenarios that do not introduce below.
SMRs (dashed lines) increases linearly with carbon tax because of the Fig. 5 shows the placement of SMRs for designs with different
increasing penalty placed on emissions. After the onset of the transition weights to control the placement policy for the WM SM scenario and
window and under optimal power flow, a power system incorporating an LCOE of £60/MWh. Moving left to right, the figure shows how the
SMRs will be more cost-effective than a system based on the current designs change as the number of SMRs increases. The designs in the
generation mix, regardless of the level of carbon tax. Beyond the right-most column correspond to the points (A), (B) and (C) in Fig. 4.
transition window, the level of carbon tax simply affects the level of Fig. 5(a) shows designs where SMRs are placed near centres of
incentive (i.e, the difference in cost) to adopt SMRs. These findings demand to minimise transmission losses. The SMRs are mostly placed
highlight the potential economic benefits of adopting SMRs relative to close to the East Midlands with some on the west coast of Wales,
the status quo, and emphasise the importance of considering carbon replacing the use of coal and eventually natural gas as the carbon tax

Fig. 4. Pareto front formed by competing objectives to minimise site-demand distance and investment and risk cost (WM SM and LCOE of £60/MWh). Left: 33 SMRs. Right: 10–60
SMRs.

6
W. Xie et al. Applied Energy 364 (2024) 123128

Fig. 5. Impact of placement policy on SMR locations (WM SM and LCOE of £60/MWh).

is increased. Fig. 5(b) shows designs where SMRs are placed to balance At extreme weights, the designs become problematic. As the weight
proximity to centres of demand versus proximity to population. Fewer approaches 1, the placement of SMRs that would optimally operate
SMRs are now placed in the East Midlands, with more on the coast at low proportions of capacity becomes more prevalent, indicating
that although the designs are feasible in terms of the SMR placement
including the south and west coasts of England. Fig. 5(c) shows designs
algorithm (Step 1 in Fig. 1), they are not acceptable in other respects.
where SMRs are placed away from centres of population to minimise As the weight approaches 0, the designs become increasingly expensive
risk. SMRs are now also placed on the west coast of Scotland, causing as they move to more extreme points on the Pareto front, illustrated by
an increase in the capacity in Scotland from 4.96 to 6.86 GW. the arrow labelled (B) on Fig. 4.

7
W. Xie et al. Applied Energy 364 (2024) 123128

Fig. 6. Impact of placement policy on transmission loss and net demand (WM SM and LCOE of £60/MWh).

3.3. Impact of placement policy on energy independence

Fig. 6 shows the impact of the placement policy on transmission

losses and net energy demand for the WM SM scenario with 33 SMRs
and an LCOE of £60/MWh. The panel on the left shows that low
weights reduce transmission losses. The cost is reduced at intermediate
weights, at the expense of increased transmission losses in Scotland
and between Scotland and England. The reduction in cost arises from
the reduction in risk cost as SMRs are placed away from centres of
population. The panel on the right shows that high weights result
in higher transmission losses and higher costs. A comparison of the
three panels shows that the placement policy affects which regions are
net importers or exporters of energy, with Scotland exporting more
energy, and areas of England switching between being net importers
and exporters as the weight increases. Results from the other scenarios
are provided in the Supplemental Information for comparison.

4. Conclusions

This work shows a carbon tax transition window above which the
adoption of SMRs becomes cost-effective in place of power generation
from oil, coal and gas. The value of the carbon tax in the transition
window was insensitive to the assumptions made about the availability
of renewable wind and solar power, but sensitive to the cost of the
SMRs. A carbon tax in the range £45–60/t was observed to be required
for SMRs with an LCOE of £60/MWh, versus £0–20/t for an LCOE
of £40/MWh. Beyond the transition window, the SMR scenarios were Fig. 7. Frequently selected sites for SMR placement overlaid by population density.
always cheaper than scenarios based on the current generation mix.
Fig. 7 shows the sites and population density data [41] that were
considered by the placement algorithm, together with the sites that
They include Sellafield, Trawsfynydd and Wylfa, which were recom-
were most commonly selected for the introduction of SMRs. The com-
monly selected sites fall into two broad groups: sites in the East mended by Rolls-Royce [30], but not Oldbury. These observations were
Midlands that are close to centres of demand, and sites on the south and insensitive to the assumptions made about the availability of renewable
west coasts of Great Britain that are away from centres of population. wind and solar power and the cost of the SMRs.

8
W. Xie et al. Applied Energy 364 (2024) 123128

The optimal placement of the SMRs required a trade-off between References

minimising load-demand distance as a proxy for minimising trans-
mission loss, and minimising population risk. The placement policy [1] O’Sullivan C. 2022 UK greenhouse gas emissions, provisional figures.
Technical report, Department for Energy Security and Net Zero; 2023, URL
influenced the performance and cost of the resulting energy system,
https://assets.publishing.service.gov.uk/government/uploads/system/uploads/
as well as whether a region acted as a net importer or exporter of attachment_data/file/1147372/2022_Provisional_emissions_statistics_report.pdf.
energy. The most cost-effective solutions were achieved by balancing [Accessed February 2023].
load-demand distance and population risk. [2] Elexon. Balancing Mechanism Reporting Service. 2023, URL https://www.
bmreports.com/bmrs. [Accessed February 2023].
CRediT authorship contribution statement [3] Morley KR. National Grid: Live. 2023, URL https://grid.iamkate.com/. [Accessed
February 2023].
[4] Committee CC. Progress in reducing emissions 2022 Report to Parlia-
Wanni Xie: Writing – review & editing, Writing – original draft, Vi- ment. 2022, URL https://www.theccc.org.uk/publication/2022-progress-report-
sualization, Validation, Software, Methodology, Investigation, Funding to-parliament/. [Accessed February 2023].
acquisition, Formal analysis, Data curation, Conceptualization. John [5] National Grid ESO. Carbon Intensity API. 2023, URL https://carbonintensity.org.
Atherton: Writing – review & editing, Methodology. Jiaru Bai: Writ- uk/. [Accessed June 2023].
[6] European Court of Auditors. Energy taxation, carbon pricing and energy subsi-
ing – review & editing, Software. Feroz Farazi: Writing – review
dies. 2022, URL https://www.eca.europa.eu/Lists/ECADocuments/RW22_01/RW_
& editing, Validation, Supervision, Software, Data curation. Sebas- Energy_taxation_EN.pdf. [Accessed February 2023].
tian Mosbach: Writing – review & editing, Supervision, Software, [7] Bray S. Carbon taxes in Europe. 2022, URL https://taxfoundation.org/carbon-
Resources, Project administration, Methodology, Investigation, Fund- taxes-in-europe-2022/. [Accessed June 2023].
ing acquisition, Formal analysis, Conceptualization. Jethro Akroyd: [8] Hirst D. Carbon Price Floor (CPF) and the price support mechanism.
2018, URL https://commonslibrary.parliament.uk/research-briefings/sn05927.
Writing – review & editing, Writing – original draft, Visualization,
[Accessed June 2023].
Validation, Supervision, Software, Resources, Project administration, [9] Mo J-L, Zhu L. Using floor price mechanisms to promote carbon cap-
Methodology, Investigation, Funding acquisition, Formal analysis, Data ture and storage (CCS) investment and CO2 abatement. Energy Environ
curation, Conceptualization. Markus Kraft: Writing – review & editing, 2014;25(3–4):687–707. http://dx.doi.org/10.1260/0958-305x.25.3-4.687.
Supervision, Resources, Project administration, Methodology, Funding [10] Gugler K, Haxhimusa A, Liebensteiner M. Effectiveness of climate policies: Car-
bon pricing vs. subsidizing renewables. J Environ Econ Manag 2021;106:102405.
acquisition, Conceptualization.
http://dx.doi.org/10.1016/j.jeem.2020.102405.
[11] Secretary of State for Business, Energy and Industrial Strategy. Energy
Declaration of competing interest white paper: Powering our net zero future. 2020, URL https://www.gov.uk/
government/publications/energy-white-paper-powering-our-net-zero-future.
The authors declare that they have no known competing finan- [Accessed February 2023].
cial interests or personal relationships that could have appeared to [12] Department for Energy Security and Net Zero. The ten point plan for a green
industrial revolution. 2020, URL https://www.gov.uk/government/publications/
influence the work reported in this paper.
the-ten-point-plan-for-a-green-industrial-revolution. [Accessed April 2023].
[13] Department for Business, Energy & Industrial Strategy. Power stations
Data availability in the United Kingdom, May 2022 (DUKES 5.11). 2022, URL https:
//www.gov.uk/government/statistics/electricity-chapter-5-digest-of-united-
The data sources, such as DUKES, applied in this paper are thor- kingdom-energy-statistics-dukes. [Accessed February 2023].
oughly cited and can be readily accessed through the hyperlinks pro- [14] Matthew C, Walker A. Nuclear energy in the UK (2022). 2022, URL https://post.
parliament.uk/research-briefings/post-pn-0687. [Accessed February 2023].
vided in the references section. The codes developed for this work [15] Department for Energy Security and Net Zero and Department for Busi-
are available under an open-source license at https://github.com/ ness, Energy & Industrial Strategy. Nuclear energy in the UK (2019). 2019,
cambridge-cares/TheWorldAvatar. Archival copies of the datasets and URL https://www.gov.uk/government/publications/energy-trends-march-2019-
instructions to reproduce the work are available in the University of special-feature-article-nuclear-electricity-in-the-uk. [Accessed February 2023].
Cambridge data repository (doi: 10.17863/CAM.99921). [16] Haves E. Nuclear power in the UK (2021). 2021, URL https://lordslibrary.
parliament.uk/nuclear-power-in-the-uk/. [Accessed February 2023].
Declaration of Generative AI and AI-assisted technologies in the [17] UKGovernment. British energy security strategy. 2022, URL https://www.gov.
uk/government/publications/british-energy-security-strategy/british-energy-
writing process
security-strategy#nuclear. [Accessed April 2023].
[18] Department for Energy Security and Net Zero and Department for Business, En-
The authors used ChatGPT version GPT-3.5 during the preparation ergy & Industrial Strategy. Small modular reactors: techno-economic assessment.
of this work to enhance the readability and language of the manuscript. Technical report, 2016, URL https://www.gov.uk/government/publications/
After using this tool, the authors reviewed and edited the content as small-modular-reactors-techno-economic-assessment. [Accessed April 2023].
[19] Cárdenas B, Ibanez R, Rouse J, Swinfen-Styles L, Garvey S. The effect of a nuclear
needed and take full responsibility for the content of the publication.
baseload in a zero-carbon electricity system: An analysis for the UK. Renew
Energy 2023;205:256–72. http://dx.doi.org/10.1016/j.renene.2023.01.028.
Acknowledgements [20] Asuega A, Limb BJ, Quinn JC. Techno-economic analysis of advanced small
modular nuclear reactors. Appl Energy 2023;334:120669. http://dx.doi.org/10.
This research was supported by the National Research Foundation, 1016/j.apenergy.2023.120669.
Prime Minister’s Office, Singapore under its Campus for Research Ex- [21] Steigerwald B, Weibezahn J, Slowik M, von Hirschhausen C. Uncertainties
in estimating production costs of future nuclear technologies: A model-based
cellence and Technological Enterprise (CREATE) programme. Part of analysis of small modular reactors. Energy 2023;128204. http://dx.doi.org/10.
this work was also supported by Towards Turing 2.0 under the EPSRC, 1016/j.energy.2023.128204.
United Kingdom Grant EP/W037211/1. The authors would further like [22] World Nuclear News. Rolls-Royce on track for 2030 delivery of UK
to thank and acknowledge the financial support provided by the Cam- SMR. 2021, URL https://world-nuclear-news.org/Articles/Rolls-Royce-on-track-
bridge Trust, United Kingdom. Markus Kraft gratefully acknowledges for-2030-delivery-of-UK-SMR. [Accessed February 2023].
[23] Department for Energy Security and Net Zero and Department for Business,
the support of the Alexander von Humboldt Foundation, Germany. For Energy & Industrial Strategy. BEIS electricity generation costs (2020). Technical
the purpose of open access, the author has applied a Creative Commons report, Department for Energy Security and Net Zero and Department for Busi-
Attribution (CC BY) licence to any Author Accepted Manuscript version ness, Energy & Industrial Strategy; 2020, URL https://www.gov.uk/government/
arising. publications/beis-electricity-generation-costs-2020. [Accessed February 2023].
[24] Nian V, Mignacca B, Locatelli G. Policies toward net-zero: Benchmarking the
economic competitiveness of nuclear against wind and solar energy. Appl Energy
Appendix A. Supplementary data
2022;320:119275. http://dx.doi.org/10.1016/j.apenergy.2022.119275.
[25] Gao S, Huang G, Zhang X, Chen J, Han D. SMR siting for the electricity
Supplementary material related to this article can be found online system management. J Clean Prod 2021;297:126621. http://dx.doi.org/10.1016/
at https://doi.org/10.1016/j.apenergy.2024.123128. j.jclepro.2021.126621.

9
W. Xie et al. Applied Energy 364 (2024) 123128

[26] Gao S, Huang G, Zhang X, Han D. Small modular reactors enable the transition [34] Blank J, Deb K. Pymoo: Multi-objective optimization in Python. IEEE Access
to a low-carbon power system across Canada. Renew Sustain Energy Rev 2020;8:89497–509. http://dx.doi.org/10.1109/ACCESS.2020.2990567.
2022;169:112905. http://dx.doi.org/10.1016/j.rser.2022.112905. [35] Grimston M, Nuttall WJ, Vaughan G. The siting of UK nuclear reactors. J Radiol
[27] Zhang X, Huang G, Zhou X, Liu L, Fan Y. A multicriteria small modular reactor Prot 2014;34(2):R1–24. http://dx.doi.org/10.1088/0952-4746/34/2/r1.
site selection model under long-term variations of climatic conditions – a case [36] Zhang XY, Huang GH, Liu LR, Chen JP, Luo B, Fu YP, et al. Perspective
study for the province of saskatchewan, Canada. J Clean Prod 2021;290:125651. on site selection of small modular reactors. J Environ Inf Lett 2020. http:
http://dx.doi.org/10.1016/j.jclepro.2020.125651. //dx.doi.org/10.3808/jeil.202000026.
[28] Liu Y, Huang G, Chen J, Zhang X, Zheng X, Zhai M. Development of an [37] Department for Business, Energy & Industrial Strategy. National policy statement
optimization-aided small modular reactor siting model – a case study of for new nuclear above 1GW post 2025: siting criteria and process. 2018, URL
saskatchewan, Canada. Appl Energy 2022;305:117867. http://dx.doi.org/10. https://www.gov.uk/government/consultations/national-policy-statement-for-
1016/j.apenergy.2021.117867. new-nuclear-above-1gw-post-2025-siting-criteria-and-process. [Accessed June
[29] Business and the environment. GDA Step 1 of the Rolls-Royce SMR: statement of 2023].
findings. 2023, URL https://www.gov.uk/government/publications/gda-step-1- [38] Staffell I. Measuring the progress and impacts of decarbonising British electricity.
of-the-rolls-royce-smr-statement-of-findings/gda-step-1-of-the-rolls-royce-smr- Energy Policy 2017;102:463–75. http://dx.doi.org/10.1016/j.enpol.2016.12.037.
statement-of-findings. [Accessed February 2023]. [39] Department for Energy Security and Net Zero and Department for Busi-
[30] Environment Agency, Office for Nuclear Regulation, and Natural Re- ness, Energy & Industrial Strategy. BEIS electricity generation cost re-
sources Wales. Small modular reactor design completes first step of assess- port (November 2016). Technical report, Department for Energy Security
ment. 2023, URL https://www.gov.uk/government/news/small-modular-reactor- and Net Zero and Department for Business, Energy & Industrial Strat-
design-completes-first-step-of-assessment. [Accessed February 2023]. egy; 2016, URL https://www.gov.uk/government/publications/beis-electricity-
[31] World Nuclear News. UK confirms funding for Rolls-Royce SMR. 2019, URL generation-costs-november-2016. [Accessed February 2023].
https://www.world-nuclear-news.org/Articles/UK-confirms-funding-for-Rolls- [40] Akroyd J, Mosbach S, Bhave A, Kraft M. Universal digital twin – a dynamic
Royce-SMR. [Accessed February 2023]. knowledge graph. Data-Centric Eng 2021;2:e14. http://dx.doi.org/10.1017/dce.
[32] Belivanis M. Power systems test case archive. 2013, URL https://www.maths.ed. 2021.10.
ac.uk/optenergy/NetworkData/reducedGB. [Accessed February 2023]. [41] United Nations Office for the Coordination of Humanitarian Affairs (OCHA).
[33] Atherton J, Xie W, Aditya LK, Zhou X, Karmakar G, Akroyd J, et al. How United Kingdom: High resolution population density maps + demographic
does a carbon tax affect Britain’s power generation composition? Appl Energy estimates. 2020, URL https://www.maths.ed.ac.uk/optenergy/NetworkData/
2021;298:117117. http://dx.doi.org/10.1016/j.apenergy.2021.117117. reducedGB/. [Accessed February 2023].

10