THE IMPACT OF MODULARISATION STRATEGIES ON SMALL MODULAR REACTOR COST

Clara A. Lloyd, Anthony R. M. Roulstone, Campbell Middleton

University of Cambridge Engineering Department, Trumpington Street, Cambridge, CB2 1PZ, [email protected]

Small Modular Reactors (SMRs) based on established A recent paper by Ganda et al. [2] analyses LR
light-water technology have gained a lot of attention from construction cost data from a number of US sources,
the nuclear industry; however, the potential that SMRs including historical data and construction cost estimates.
have to reduce the cost of nuclear construction has been One of the sources Ganda discusses is a summary of US
under-studied. Modularisation is a cost reducing historical construction cost data that is presented in the
mechanism where a SMR power plant is subdivided into 1987 Energy Economic Data Base Programme (EEDB)
smaller units, or modules. These modules can be produced Phase IX Update Report [3]. Ganda finds that it is
offsite in a controlled environment, potentially offering important to consider both direct and indirect costs in cost
cost reductions that offset their apparently higher capital analyses because the contribution of indirect costs to the
costs. total construction cost is non-trivial. This work by Ganda
This paper will investigate the effects modularisation is central to establishing baseline LR construction costs.
and standardisation might have on SMR capital costs. Carelli et al. [4] develop a SMR cost estimation model
Modularisation and standardisation not only reduce direct based on specific cost-power scaling, where the cost of a
and indirect costs, respectively, but also enable activation single SMR is determined using scaling exponents and is
of other cost-reducing mechanisms, such as shifting then reduced using a series of multiplicative cost-savings
construction work from site to a factory, transferring related to SMR unit co-siting, replication and
learning between tasks, and achieving economies of standardisation (achieving learning), financial aspects
multiples. It will show that constructing a SMR using the (smaller units are easier to finance as they can be built in
same methods as current large reactors is not stages), and various design-specific solutions. Carelli
economically feasible and will demonstrate how estimates that, when SMRs take full advantage of all these
modularisation reduces SMR capital costs. cost-reducing strategies, the specific capital cost of a SMR
The primary constraints on module size are imposed will be 1.05 times that of a reference LR. This paper also
by weight and height transport limitations, linking reactor uses the cost-power scaling approach previously used by
size to ease of modularisation. This leads to an analysis of Carelli in [4] but considers a serial SMR build strategy,
which SMR components and structures should be targeted focussing on the cost reduction factors that will be common
for modularisation in order to achieve optimal cost to all SMRs of a single chosen design (that is,
benefits. modularisation and standardisation) and ignoring potential
site-specific cost reductions.
I. INTRODUCTION Work by Abdulla et al. [5] presents a qualitative study
in which the authors interview 16 nuclear industry experts
Historically, large nuclear reactors (LRs) have (vendors, regulators, engineers) to obtain an estimate of the
experienced severe budget and schedule overruns. The cost of building a SMR plant. Abdulla finds that the
industry has welcomed SMRs based on proven light water estimates vary by a factor of 2.5, and a new 225 MWe SMR
reactor technology for their potential to reduce both the plant could cost between $3,200/kWe and $7,100/kWe (in
very high construction costs and long build schedules of 2013 US dollars) [5]. While this paper develops specific
traditional LRs. There are a number of strategies that SMR costs based on the EEDB LR cost data and Carelli’s
SMRs might employ to bring about these cost and schedule power scaling rules, the results from the parametric model
reductions, including modularisation, standardisation, and presented in this paper are generally consistent with
production learning. The process by which a SMR is built, Abdulla’s findings. Moreover, the cost-reduction methods
however, must be radically different from the current LR identified by Abdulla’s study, such as factory fabrication
construction process, otherwise SMRs will fall victim to of units and a reduced SMR construction schedule, are
the issues that have plagued LR projects, particularly since central to the cost reductions in the work presented here.
the capital costs for nuclear power plants contribute about
70% of the total Levelised Cost of Electricity [1].
I.A. Modularisation Principles plant (NPP) and identifies how best to design each
component as a module. A summary of the feasible
Modularisation is the process by which a large, modules is given in Stone & Webster Table 4-1 (pp 145-
complicated product is broken down into smaller building 146). The key information used in this paper is:
blocks, or modules, according to a set of limiting •   Type of each module (including precast concrete
constraints. SMRs have a unique opportunity to leverage structures, structural steel, liner modules, etc.);
the benefits of modularisation as a build technique because •   Location where the modules are to be used (reactor
of their smaller physical size. The SMR modules can be containment, turbine hall, etc.);
constructed away from the nuclear construction site; this •   Quantity of modules that are required, together with
could be either in a shop, factory, or module assembly the module weight and dimensions (length, width, and
building, making parallel construction activities possible height).
and greatly improving productivity. In the case of SMRs, Stone & Webster also identify that transportation
transportation logistics impose dimensional and weight logistics limit the modularisation scheme. It is clear that
constraints on the modules. feasible transport logistics will be of greater importance for
One additional design philosophy that is essential for SMRs that are intended to be produced in volume;
successful modularisation is standardisation. producing greater numbers of modules increases the value
Standardisation targets the indirect costs associated with of easy and straightforward transportation logistics.
building a nuclear power plant by removing much of the Although the age of the Stone & Webster report is a
repeated upfront design work and by leveraging higher drawback, it provides a useful baseline for the work
learning to reduce both schedule and, by extension, cost. presented here, as it describes a modularisation scheme for
This is achieved through simplification of the construction, a NPP that is, to our knowledge, the only one of its kind
testing, and commissioning procedures. This paper published.
assumes a standard reactor system design and that
components and modules are similarly standard wherever II. PROBLEM SETUP
that is possible.
While modularisation directly offers one-off This paper seeks to determine the benefits, in terms of
productivity benefits, it also acts as an enabler for cost, of modularising SMR nuclear power plants of
continuous learning. Implementing standardised, modular different sizes. It uses transport constraints on the
construction into a nuclear build programme will help proposed modules to define what module dimensions are
facilitate a streamlined supply chain and off-site module feasible. SMR construction cost is the primary focus of this
build will potentially increase production learning rates. work, and is captured using the concept of overnight capital
Rosner and Goldberg [6] present a review of the scope for cost. Although operating and maintenance costs for SMRs
making SMRs competitive across larger programmes of are important to through-life economics, capital costs are
standard reactors; this is accomplished through production nonetheless predominant and modularisation will primarily
learning and economies of volume. A parametric study target these construction costs.
conducted by Chen et al. [7] investigates the effect of
production learning on specific SMR components. These The following terms are important to the argument:
studies highlight the importance of modularisation and •   Modularisation refers to dividing NPP structures,
standardisation for maximising production learning equipment, and/or components into modules, or
benefits, as well as designing the factory and supply chain ‘building blocks’, manufactured in a purpose-built
for optimising the conditions for production learning. factory, shop, or on-site module assembly building.
Analysis and discussion of these further benefits is outside Off-site modules are transported to the nuclear site for
the scope of this paper but has been discussed in other work installation and assembly. Modularisation introduces
at the University of Cambridge [8]. a set of one-off productivity improvements that serve
Designing a specific modularisation scheme for a to reduce direct costs [10, 11].
nuclear power plant is a challenge. The shipbuilding •   Standardisation refers to adopting a single SMR
industry, where modular design and build is the norm, design and, within that design, using as many common
provides some interesting information on the general components, modules, and equipment items as
principles and practice of modular construction, but data possible. Standardisation reduces indirect costs.
on specific modularisation schemes are not publicly •   Degree of Modularisation (DoM) is defined by the
available. Stone & Webster [9] developed a authors as the fraction of direct site costs, associated
modularisation scheme for a 790 MWe LR on behalf of the with the construction of a specific component, that are
US DoE. The Stone & Webster report provides a highly moved to a factory. In theory, DoM values can range
detailed analysis of modularisation and is useful for between 0, for a stick-built plant, and 1, for a plant in
establishing a modularisation scheme for reactors of which every structure is modularised.
various sizes. It assesses each part of the nuclear power
•   Effective Modularisation is again defined by the This sets a practical maximum DoM for a given
authors as the net fraction of direct site costs reactor.
transferred from the construction site to the production
12345  6 E
facility, for the whole plant. Effective Modularisation 𝑊𝑒𝑖𝑔ℎ𝑡  , =   𝑊𝑒𝑖𝑔ℎ𝑡  @&B (2)
12345  C&D
is useful because it gives an indication of the extent of
modularity for the whole plant.
D. Alternative modularisation schemes are considered by
further division on the Stone & Webster module
III. SMR MODULARISATION MODEL
dimensions. The transport feasibility and maximum
DoM are re-calculated according to the established
An overview of the SMR modularisation model used
criteria.
in this paper is given here; further details of the data used,
and the accompanying sources, are provided in the
III.A. Baseline Cost Breakdown
corresponding subsections.
This paper uses a parametric approach to estimating
A. A SMR baseline cost breakdown is determined using
SMR capital costs by applying power scaling laws to the
the detailed reference LR breakdown from EEDB [3]
available LR data, similar to the method used by Carelli
and applying specific cost-power scaling laws. The
[4], as described earlier. This paper is, however, concerned
specific cost-power scaling relationship used is shown
with the structure of the construction costs, both direct and
in Equation (1), where Specific Cost is the capital cost
indirect, and the effects modularisation and standardisation
of construction, in 2017 $/kWe, Power is the rated
have on reducing these costs. The only available data that
power output of the reactor, in MWe, and α is the
exist at the necessary level of detail are the data analysed
scaling exponent, from [4]. The subscript i refers to
by Ganda [2] and published by the US Energy EEDB Phase
the particular reactor under consideration and the
IX Update Report [3]. The EEDB data are from actual LR
subscript EEDB refers to the reference data used from
builds in the US, taken from annual reports published
[3]. Using the scaled SMR cost breakdown, a variable
between 1977 and 1987, although costs reported here have
Degree of Modularisation (DoM) is set up to allow a
been inflated to 2017 United States dollars. The data are
fraction of site costs to be moved to a factory.
structured by work type and cost type (labour, equipment,
:;< or material) and forms the most comprehensive and
123456
𝑆𝑝𝑒𝑐𝑖𝑓𝑖𝑐  𝐶𝑜𝑠𝑡, = 𝑆𝑝𝑒𝑐𝑖𝑓𝑖𝑐  𝐶𝑜𝑠𝑡../0
123457789
(1) detailed data set available. This paper uses the median data
from 1987, which provides cost data for an average 1144
B. The cost reductions that can be attributed to MWe PWR reactor built in the United States. As Ganda
modularisation and standardisation are determined notes, historical US-based data will not necessarily be
according to a set of rules from [12] and are applied to representative of global experience today. Changes in
the scaled cost breakdown from Equation (1). These regulation, location, labour markets, and material costs
are essentially a set of multiplicative factors that have over time will impact how accurately the inflated 1987
a net cost-reducing effect; the magnitude of which is average costs represent the cost of reactors built today.
dependent on the DoM. Analysing these influences is beyond the scope of this
work, although Berthelemy and Rangel [13] provide a
C. The Stone & Webster modularisation scheme is used thorough discussion of these factors. This data set does,
to modularise a range of reactors. The number of however, provide a consistent baseline for comparative
modules in a reactor is held constant and the weight of studies of different construction strategies for LRs and
each module for a given reactor is scaled relative to SMRs.
the Stone & Webster reference design, according to To obtain a cost breakdown for a reference stick-built
Equation (2). The variable Power refers to the rated SMR plant, the available EEDB cost data was scaled using
power output of the nuclear reactor, in MWe, and n is specific power-cost scaling according to Equation (1).
the scaling exponent. The subscript S&W is for the This provides a detailed cost breakdown for a range of
reference Stone & Webster plant and i is for the new reactors outputs between 150 MWe and 1500 MWe by
reactor plant. From this, the linear dimensions applying scaling exponents to the different types of cost,
(length/width/height) of the module were then according to the exponential factors in Table I. It is worth
determined (square root for 2-D modules and cube mentioning that this model is applying scaling laws beyond
root for 3-D modules). Constraints imposed by the the limits of existing data; therefore, the accuracy of the
transportation logistics were next applied to each results cannot be calibrated or confirmed against real
module to determine which modules are feasible values.
(transportable) and which are not (non-transportable).
 TABLE I. Two-Digit Code of Account headings [3]and weighted average scaling exponent as a proxy for size
corresponding scaling exponents [4]. scaling (n = 0.64), as it is reasonable to expect that the
Two-Digit Code of Account Heading Exponent, α direct costs (which are comprised of material and labour
accounts) are a reflection of the size difference between a
21 - Structures and Improvements 0.59
SMR and a LR.
22 - Reactor/Boiler Plant Equipment 0.53 The transport logistics of the reactor modules is
23 - Turbine Plant Equipment 0.83 developed according to the UK-specific criteria described
24 - Electric Plant Equipment 0.49 in Table II. This work identifies three possible categories,
called Transport Envelopes, into which modules can fall
25 - Miscellaneous Plant Equipment 0.59 based on their weight and/or dimensions. Transport
26 - Main Condenser Heat Rejection System 1.06 Envelope 1 is a straightforward road transport scenario,
with specific transport limitations; the module is the size of
III.B. Modularisation and Standardisation an ISO container (or less). Transport Envelope 2 is the
routine ‘relaxation’ of the weight and length limits in
The relationship between cost savings and Envelope 1 but still allows for accessible and relatively
modularisation are based on guidelines published by the straightforward transport in most of the UK (and likewise
EMWG (2007) in Chapter 11: Estimating Factory- continental Europe); the idea here is to use purpose-built
Produced Modular Units [12]. Some rules for vehicles to make use of the additional freedom available.
standardisation-related cost savings have also been Transport Envelope 3 is for any modules that exceed the
developed to reflect reductions in the indirect costs that limits in Envelope 2 and essentially means the modules
arise as a result of modularisation-enabled standardisation. cannot be transported by road (at least in the UK) and
These have been developed based on EMWG guidelines means the component cannot practically be modularised.
and expert opinion, in conjunction with the EEDB
definition of what each category includes. TABLE II. Transportation envelopes for NPP modules
•   Time based indirect cost categories are affected by the based on UK road transport limitations [14].
length of the reactor build schedule. SMR costs are Transport Weight Length Width Height
reduced proportionally to the SMR schedule Envelope (MT) (m) (m) (m)
reduction, which is indirectly a function of
modularisation. 1 28.8 12.032 2.34 2.292
•   Modularity based indirect cost categories are directly 2 47 27.4 4.3 3.2
affected by the influence of modularisation on direct 3 >47 >27.4 >4.3 >3.2
labour hours. The same percentage reduction of total
direct costs, achieved through modularisation, is Once the quantity and size/weight of SMR modules is
applied to these indirect costs. calculated, the distribution of these modules into the three
•   Finally, standardisation of the SMR facility (that is, transport envelopes defined in Table II can be determined.
building the same SMR design across multiple sites The maximum Degree of Modularisation is taken to be the
and plants) will directly impact detailed design work, percentage of modules, by number, that fit in either
reducing these costs by 80% through removal of work Envelope 1 or 2. The remaining modules are considered
not repeated for a standardised design. ‘un-modularisable’ as the excessive weight or dimensions
make transport by road nearly impossible. It should be
III.C. Transportation Constraints on Modularisation mentioned that it is expected there will be a small number
of very large items that cannot be modularised according
The reference modularisation scheme from Stone & to the scheme set forth here (for example the reactor
Webster [9] is used for data on module type, quantity of pressure vessel, steam generators, turbine generators, and
modules needed, as well as module weight, width, length, polar crane). The transportation analysis presented here is
and height. The Stone & Webster report shows that developed for the high volume of structural and equipment
transportation is the largest constraint on modules and will transport and the constraints are not intended to apply to
therefore limit how modular a nuclear power plant can be. special one-off equipment transport needs.
The transport constraints for SMRs are applied to a scaled
version of the modularisation scheme for the 790 MWe III.D. Modularisation Schemes
reactor. The physical size of structures is related to the
power output of the reactor power plant. From the relative Fig. 1 shows the maximum percentage of modules, by
geometries of small and large power plants, module size is number, that can be transported by road for a range of
assumed to follow an exponential, instead of a linear, reactor powers. The module weight and dimensions are
scaling form. Stone & Webster module weights are scaled based on scaling the Stone & Webster module scheme
according to Equation (2). This model uses Carelli’s cost
using a scaling exponent of n = 0.64 in Equation (2). As
expected, the model shows that smaller reactors have a 1.00
greater number of modules that fit within the transport
limitations, making modularisation more feasible for 0.80
SMRs (particularly for the M&E, liner, and reinforcing

Degree of Modularisation
steel categories). Fig. 1 also indicates that structural
modules, if designed according to the Stone & Webster 0.60
scheme, are generally not transportable and therefore most
structural elements cannot be modularised. This translates
to 564 modules, of a total 1417 proposed by the Stone & 0.40
Webster modularisation scheme, which cannot be
transported for any size of reactor. Of these, 379 modules 0.20
are precast concrete (which cannot be transported because
weight is the limiting factor) and 185 modules are
structural steel (where width is the limiting factor). 0.00
In order to increase the number of modules that fit in 0 300 600 900 1200 1500
Transport Envelope 1 and/or 2, it will be necessary to Reactor Power (MWe)
change the modularisation scheme developed by Stone &
Reinforcing steel modules Precast concrete modules
Webster. There are two ways this can be accomplished. Structural steel modules Mechanical Modules
•   Module Subdivision: break the existing modules Liner Modules
down further, so they are easier to transport. The total
modularised weight in each category remains the Fig. 1. Maximum Degree of Modularisation attainable
same, but each module dimension (width, length, from the original Stone & Webster scheme.
height) – and therefore the quantity of modules –
changes by a certain Module Division Factor (see
Table III). With this strategy, modules become 1.00
smaller and therefore more likely to fit in the defined
transport envelopes. The drawback to this is that there
will be a greater number of modules to transport, the 0.80
Degree of Modularisation

more subdivided the original scheme becomes.

•   Module Extension: modularise different structures
and components to those considered by Stone & 0.60
Webster. This is particularly relevant for SMRs,
which may, because of their smaller size, have 0.40
increased potential for modularisation in various
structures or components that Stone & Webster did not
consider feasible for the 790 MWe reference reactor. 0.20

Fig. 2 shows the maximum Degree of Modularisation,

which is equivalent to the percentage of modules by 0.00
number that can be transported, for a range of reactor 0 300 600 900 1200 1500
powers. The Stone & Webster modules are scaled using an Reactor Power (MWe)
exponent of n = 0.64 in Equation (2), and a Module
Reinforcing steel modules Precast concrete modules
Division Factor of 2.0 is applied. Comparison of Fig. 1 and Structural steel modules Mechanical Modules
Fig. 2 shows how further module subdivision increases the Liner Modules
percentage of transportable modules.
Fig. 2. Maximum Degree of Modularisation attainable
TABLE III. Relationship between module subdivision when a Module Division Factor of 2.0 is applied to the
and the number of modules required for an NPP. Stone & Webster scheme.
Module Division Factor Total number of modules
(Stone & Webster scheme) 1417 IV. SMR MODULARISATION SCHEMES
1.50 3373
Two modularisation schemes will be considered. In
2.00 6324
the first, the original Stone & Webster scheme is used and
3.00 15705 modules are not subdivided beyond the Stone & Webster
proposal, as shown in Fig. 1. In the second, the module output. The cost benefits of modularisation and further
dimensions proposed by Stone & Webster (length, width, module subdivision are summarised for a 300 MWe SMR
height) are divided by a factor of two, as shown in Fig. 4. and 1000 MWe LR in Table IV and illustrate the significant
In both cases, the maximum degree of modularisation for impact modularisation has on cost.
any system or category is determined by the fraction of
modules that can be transported in Envelope 1 and/or 2 (as
in Figs. 1 and 2). In both these schemes, the extent to
9000
which modularisation is applied to the SMR plant can be
varied by varying the DoM between 0.0 and the maximum
transportable percentage of modules. In all cases 8000

Specific Capital Cosst (2017 $/kWe)

modularisation is applied to the full plant equally (M&E
systems, civil structures, and equipment). 7000
Figs. 3 and 4 show the specific capital cost of
construction, in 2017 $/kWe, relative to reactor power, in 6000
MWe. Modularisation extends to the full reactor plant,
including structural elements as well as M&E systems.
5000
The model in Fig. 3 uses the original Stone & Webster
modularisation scheme; in Fig. 4 the original Stone &
Webster module dimensions are halved. For comparative 4000
purposes, each chart shows a reactor cost curve
corresponding to a stick-built reactor with no 3000
standardisation. This must be differentiated from a reactor 0 300 600 900 1200 1500
that has no modularisation (DoM = 0.0) but is built as part Reactor Power (MWe)
of a standardised programme and therefore has reduced No standardisation DoM = 0.0
development, licensing, and design costs. DoM = 0.2 DoM = 0.4
As expected, subdividing modules further, as shown DoM = 0.6 DoM = 0.8
by comparing Figs. 3 and 4, is a beneficial modularisation DoM = 1.0
strategy to implement, as it increases the maximum DoM Fig. 3. Specific capital cost of construction relative to
that can be achieved. This means, of course, that a greater reactor power for the original Stone & Webster scheme.
fraction of site costs can be shifted to a factory, resulting in
greater one-off productivity benefits but also increasing the
cost share that can achieve continuous, learning-related
9000
benefits. In both cases, there is a point at which
modularisation reaches a maximum benefit and further
modularisation has little effect. For the case in Fig. 3; this 8000
Specific Capital Cost (2017 $/kWe)

happens at DoM = 0.4 and in Fig. 4 this happens at DoM =

0.8 (or DoM = 0.6 for power outputs exceeding 750 MWe). 7000
Figs. 3 and 4 show how modularising reactors
becomes more difficult as the power output increases. For 6000
a given modularisation scheme, the module size increases
with reactor power, thus decreasing the maximum
5000
achievable DoM. For modularisation of the structural
elementsjh to be fully effective, however, Fig. 4 shows that
the modules need to be further subdivided so a greater 4000
number of modules are transportable. The cost
implications are significant. The minimum cost of a 300 3000
MWe SMR, when both structures and M&E are made in 0 300 600 900 1200 1500
modules, again to a maximum DoM of 60% is $7,040/kWe. Reactor Power (MWe)
This drops to $5,720/kWe when the modules dimensions No standardisation DoM = 0.0
are halved, and the plant is modularised to a maximum DoM = 0.2 DoM = 0.4
DoM of 80%. Given the transport constraints and Module DoM = 0.6 DoM = 0.8
Division Factors above, it is apparent that a DoM of 60% DoM = 1.0
will provide worthwhile one-off productivity benefits for Fig. 4. Specific capital cost of construction relative to
all reactor sizes; however, the corresponding Effective reactor power when a Module Division Factor of 2.0 is
Modularisation (that is, the net fraction of site costs that are applied to the Stone & Webster scheme.
shifted to a factory) will depend on the reactor power
TABLE IV. Comparison of modularisation schemes for a modularisation is not considered worth the small additional
300 MWe SMR and a 1000 MWe LR. The value in cost benefit.
brackets is the corresponding Effective Modularisation This work so far has relied on the notion of a variable
(Meff) the net fraction of site costs moved to a factory. Degree of Modularisation, which is as yet an abstract
SMR LR concept and is tied to cost fractions only. The next stage
300 MWe 1000 MWe of this work will be to identify what components and
No standardisation specific modular strategies should be adopted in order to
$10,100/kWe $6,300/kWe achieve the necessary overall Effective Modularisation.
No modularisation
Standardised This work also points out the importance of
No modularisation $8,570/kWe $5,370/kWe transportation in setting bounds on the modularisation
strategy. The next step is to consider the possibility of
Standardised $7,040/kWe $4,700/kWe developing a modularisation scheme that is extended
Max. modularisation (Meff = 0.35) (Meff = 0.18)
beyond that considered by Stone & Webster. SMRs may
Standardised have additional scope for modularisation because of their
Max. modularisation $5,720/kWe $4,100/kWe reduced size. Here too it will be important to perform case
Module subdivision (Meff = 0.66) (Meff = 0.46) studies identifying specific structural and M&E modular
(half) solutions that are deemed feasible for a SMR power plant.
Future work should also investigate the economic costs
It is important to point out the importance of introduced by different production learning rates and
production learning in relation to modularisation. scenarios, as well as transport options and the potential
Modularisation and standardisation both offer one-off implications module transport has on reactor lead-time and
productivity improvements that reduce SMR cost; project scheduling.
however, a range of continuous benefits can also provide
significant cost savings over the course of a SMR ACKNOWLEDGEMENTS
programme. Modularisation enables factory build,
meaning higher factory learning rates could be achieved, This research is being performed using funding
further reducing construction costs. Indeed, since the received from Arup Group Ltd. and the UK Engineering
achievable DoM values are higher for smaller reactors, the and Physical Sciences Research Council (EPSRC).
cost share that can achieve learning benefits is larger and
may also lead to greater cost reduction. In this context,
modularisation may also offer strategic benefits, where the BIBLIOGRAPHY
fact that modularisation enables increased learning is what
drives the modularisation decision. Analysis of this effect
is outside the scope of this paper but learning rates and 1. DECC, "Electricity Generation Costs 2013,"
supply chain considerations have been discussed in [8]. London, July 2013.
2. F. Ganda, J. Hansen, T. K. Kim, T. A. Taiwo and R.
V. FUTURE WORK AND CONCLUSIONS Wigeland, "Reactor Capital Costs Breakdown and
Statistical Analysis of Historical U.S. Construction
This paper develops a simple economic model to Costs," in ICAPP, San Fransisco, CA, 2016.
investigate the strategies for cost reduction of small nuclear 3. EEDB IX, "Phase IX Update (1987) Report for the
power plants by adopting modular construction. It also Energy Economic Data Base Program," United
highlights the importance of considering transportation- Engineers & Constructors Inc. and Oak Ridge
related constraints when developing these strategies. National Laboratory, Philadelphia, 1988.
Ultimately, the findings presented here are intended to help
4. M. D. Carelli, P. Garrone, G. Locatelli, M. Mancini,
guide SMR designers and vendors as to what decisions
C. Mycoff, P. Trucco and M. E. Ricotti, "Economic
they make regarding modularisation of the SMR power
features of integral, modular, small-to-medium size
plant, as well as helping identify what the expected
reactors," Progress in Nuclear Energy, no. 52, pp.
economic and logistic implications might be. This paper
403-414, 2010.
suggests that fully modularising the SMR plant is
necessary to achieve the maximum cost reduction; 5. A. Abdulla, I. L. Azevedo and M. G. Morgan,
however, this comes with the additional need to further "Expert assessments of the cost of light water small
subdivide the modules so that transport constraints can be modular reactors," Proceedings of the National
met. Given these conditions, the necessary DoM for a Academy of Sciences, vol. 110, no. 24, pp. 9686-
SMR that will provide significant construction cost 9691, 11 June 2013.
reduction is 60%; the effort required for further 6. R. Rosner and S. Goldberg, "Analysis of GW-Scale
Overnight Capital Costs," EPIC, Chicago, 2011.
7. X. Chen, A. Kotlyarevsky, A. Kumiega, J. Terry, B.
Wu, S. Goldberg and E. A. Hoffman, "Small
Modular Nuclear Reactors: Parametric modelling of
integrated reactor vessel manufacturing within a
factory environment. Volume I and II.," Department
of Energy, Office of Nuclear Energy, 2013.
8. R. Lyons, "The Effect of the Production System
Configuration on Small Modular Reactor Economics
(First Year PhD Report)," Cambridge University
Engineering Department, 2016.
9. Stone & Webster Engineering Corporation, "Plant
Systems/Components Modularisation Study," US
Energy Research and Development Administration,
Boston, 1977.
10. R. Rosner and S. Goldberg, "SMRs - Key to Future
Nuclear Power Generation in the US," EPIC,
Chicago, Il, 2011.
11. G. Maronati, B. Petrovic, J. W. Banner, C. C. White,
M. H. Kelley and J. J. Van Wyk, "TCIC Evaluation
of SMR Modular Construction Designs," Atlanta,
GA, 2016.
12. EMWG, "Chapter 11: Estimating Factory-Produced
Modular Units," in Cost Estimating Guidelines for
Generation IV Nuclear Energy Systems, Paris,
OECD, 2007, pp. 109-137.
13. M. Berthelemy and L. E. Rangel, "Nuclear reactors'
construction costs: The role of lead-time,
standardization and technological progress," Energy
Policy, vol. 82, pp. 118-130, 2015.
14. R. P. McDonnell, "Transport Strategy for Small
Modular Reactors," MPhil Thesis, University of
Cambridge, 2017.