ENTITIES AND PUBLIC POLICIES
THE EPR SECTOR
Thematic public report
Summary
July 2020
2
Summary of the Thematic Public Report of the Cour des comptes
g
AVERTISSEMENT
This summary report is intended to facilitate the reading and use
of the Audit Office report.
The Cour des comptes [Audit Office] is only accountable for the
report.
The responses from the administrative bodies and ministries
concerned appear in the follow-up to the report.
3
Summary of the Thematic Public Report of the Cour des comptes
Summary
1
An EPR reactor project drawn up under unfavourable
conditions
5
2
The building of the Flamanville EPR: an operational failure
with multiple causes
7
3
Serious consequences for the entire sector
11
4
An international strategy that cannot be continued under
the same conditions
13
5
The construction of a series of EPR2 in France: a
technological, economic and energy policy choice
15
Recommendations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
5
Summary of the Thematic Public Report of the Cour des comptes
An EPR reactor project drawn up
under unfavourable conditions
In 1989 the French group Framatome
and the German group Siemens started
to design a new nuclear reactor called, in
1992, the “European Pressurized Water
Reactor” (EPR) . This project drew on the
lessons learned from several decades of
electronuclear production and also met
increased security requirements after
the accidents at Three Mile Island, in the
United States, in 1979, and Chernobyl,
in Ukraine, in 1986 . The French and
German governments supported this
Franco-German reactor project from
its inception, and electrical engineers
from both countries joined it in 1992,
although they did not share the same
goals . German engineers intended to
evolve the “Konvoi” reactor used on
the other side of the Rhine, while EDF
wanted to evolve the N4 series, the last
reactor model then under construction .
After the decision by the Germans
to phase out nuclear energy in 1998,
France became the sole backer of this
project, whose acronym assumed the
meaning « Evolutionary Pressurized
Reactor » (EPR) . However, the broad
design
options
jointly
defined
by
engineers from both countries, although
they led to difficulties, were not called
into question .
View of the vessel and the EPR reactor building
(including corium recovery tank located under the vessel)
Source: IRSN
Double
containment chamber
Steam
generator
Cooling water
tank inside
the chamber
Discharge
channel
Corium
recovery tank
Core of the reactor
containing the fuel
Diagram of the corium recovery tank
6
Summary of the Thematic Public Report of the Cour des comptes
From 2001, the newly created Areva
group
developed
a
“turnkey”
EPR
sales strategy, thus opposing that of
EDF, which intended to remain the
frontrunner in “new nuclear power” in
France and abroad . The rivalry between
these two public groups, not arbitrated
by the public authorities of the time,
turned
into
dangerous
overbidding
within the French nuclear sector .
It was under these conditions that in
2003 Areva signed a contract to sell an
EPR to the Finnish electrical engineering
company TVO and that EDF launched, in
2004, the construction of the first EPR in
France, in Flamanville . This race between
the two French companies led to the
hasty launch of construction sites for
these first two EPRs, based on incorrect
technical references and insufficient
detailed studies . This lack of preparation
also led to an underestimation of
the difficulties in building the EPRs .
The nuclear industry has shown too
much self-confidence, inspired by the
successful building and operation of a
fleet of 58 reactors .
An EPR reactor project drawn up
under unfavourable conditions
7
Summary of the Thematic Public Report of the Cour des comptes
On the date of publication of the
report, the 3 .3 times increase in the
construction cost and at least 3 .5 times
increase in the commissioning time
of the Flamanville EPR compared
to the initial forecasts represented
a considerable deviation, even for a
“pilot” reactor .
This development is the result, first
of all, of an initial unrealistic estimate
of the duration and the cost of
building the Flamanville 3 EPR . While
the
average
reactor
construction
time worldwide was 121 months
between 1996 and 2000, the initial
construction time scheduled for the
Flamanville EPR was 54 months -
six months longer than the initial
time planned for the construction
of the Finnish Olkiluoto 3 reactor .
This
flagrant
underestimation
of
the construction time created huge
pressure to try to keep to the very
tight delivery times . The duration of
construction of the Flamanville EPR
is today estimated at 187 months,
before taking into account the impact
of the covid-19 epidemic, thereby
risking a further extension of this
duration .
Construction time of pilot reactors
(in months)
Source: Cour des comptes based on EDF data
The building of the Flamanville
EPR: an operational failure
with multiple causes
0
50
100
150
200
250
Fessenheim 1
Tricastin 1
Saint-Laurent
B1
Paluel 1
Chooz B1
Flamanville 3
(initial time)
Flamanville 3
(provisional
time)
8
Summary of the Thematic Public Report of the Cour des comptes
The
construction
engineering
time
was estimated at 5 million hours’ work;
22 million will be needed .
Almost
4,500 modifications have been made
since construction started, regularly
leading to suspension of the project, in
order to give engineers the time to deal
with the difficulties encountered .
This deviation is the result, furthermore,
of a lack of organisation of EDF’s
monitoring of the project and a lack
of
monitoring
by
the
supervisory
authorities . The Board of Directors did
not hold regular meetings to discuss
this strategic project, did not respond
to the alert messages from the audit
committee,
and
was
satisfied
with
the information it was provided with,
without taking any corrective measures .
The company did not have the structure
to execute a project of this scale: the
concept
of
«architectural
designer»
concealed a confusion between the
respective functions of the client and
of the project manager . Until 2015, the
project was not managed by a bona
fide
project team . The
contractual
relationship aggravated the weakness
of the technical management of the
project since the contracts did not
include, on their signing, either the
vagaries - which were foreseeable given
the «pilot» status of the reactor - nor the
incentives that would have encouraged
acknowledgment of the incomplete
nature of the design . Eleven out of the
12 main Flamanville EPR contracts
were thus subject to cost increases of
between 100% and 700% . The company
was late in organising itself to financially
coordinate this project: it was only after
2015 that it followed up the expenses
and then assessed the construction cost
on completion, now estimated in 2015
at an €12 .4 billion expressed in euros .
Furthermore,
the
administrative
authorities
in
question
have
not
performed their roles properly . While
the initial estimates of the construction
time and of the cost of Flamanville 3 EPR
were clearly underestimated, they did
not conduct either an assessment of the
project’s
socio-economic
profitability
or a specific analysis of the impact of
the successive problems encountered
during its execution . Nor did they
alert ministers to the importance of
the vagaries of the Olkiluoto 3 and
Flamanville 3 construction yards and
their consequences . The Government
was forced to embark on a costly
restructuring of the nuclear sector,
without the alarm signals having been
sent in due time .
The loss of technical skills and quality
culture in the nuclear industry has
today been admitted and highlighted
to
explain
the
EPRs’
construction
problems . But actors were not aware
of this at the start of the 2000s and this
diagnosis was only made late, in the face
of difficulties, and this notwithstanding
the gap of fifteen years between the
launches of the projects Civaux 2 (most
recently built French reactor in service)
and Flamanville 3 . In December 2019
EDF announced the implementation
of an action plan aiming to re-establish
the
necessary technical
skills
level
and quality culture . This examination
was therefore not carried out before
the decision was taken to launch the
construction of a new type of reactor .
EDF
decided
to
design
certain
components
of
the
reactor
in
a
so-called
failure
exclusion
process,
which
involves
strengthening
the
technical requirements in the design,
manufacture and in-service monitoring
of this equipment in order to make it
The building of the Flamanville EPR:
an operational failure with multiple causes
9
Summary of the Thematic Public Report of the Cour des comptes
extremely unlikely to fail . The former
Areva NP and its subcontractors were
not able to execute a number of parts
and welds adhering to these high
requirements . EDF did not inform the
nuclear safety authority of the existence
of a deviation from the failure exclusion
standard for crossing welds until 2017,
while it had known about them since
October 2013 . The delayed sending of
this key safety information to the safety
authority reflects a lack of smooth
interaction
between
the
sector’s
stakeholders and their safety authority .
Piping components covered by the failure exclusion
Source: IRSN.
Note: the double wall that figures on the diagram is that of
the reactor building
The financial impacts of these technical
and organisational inadequacies are
high . The repair of the crossing welds
alone leads to additional building costs of
around €1 .5 billion 2015 . The time spent
by EDF, between 2015 and 2019, trying
to convince the nuclear safety authority
that the discrepancies between the
technical requirements and what had
been achieved could be considered
acceptable, led to the suspension of the
project, which added to its cost .
The building of the Flamanville EPR:
an operational failure with multiple causes
Steam generator outlets
fixed point
11
Summary of the Thematic Public Report of the Cour des comptes
Risks
weigh
in
on
the
financial
situation
of
certain
companies
which
have
recently
benefited
from
a
recapitalisation
by
the
public authorities . The state raised
€4 .5 billion to provide capital to Areva
SA (€2 billion) and Orano (€2 .5 billion)
following
Areva’s
restructuring .
EDF received a capital injection of
€3 billion which enabled it to take
control of the reactor business of the
former Areva NP, now Framotome .
Due to the magnitude of EDF’s current
or
potential
claims
against
Areva
SA, the risk of financial failure of this
100% state-owned company cannot
be completely ruled out . The risks of
litigation brought by EDF against its
subsidiary Framatome are also likely to
weaken this company . The state must
therefore closely follow the results of
the current or future litigation processes
between these companies, of which it
is the main shareholder The strategy
of the state shareholder in this sector
needs to be confirmed .
Relationship between the industrial stakeholders in the sector involved
in the litigation process
Source: Cour des comptes, April 2020
*Companies outside of the scope of the relationship.
Note: The capital distribution of the different companies is as follows:
- For Framatome: the capital is 75% held by EDF, 19.5 % by Mitsubishi Heavy Industries
(MHI), and 5 % by Assystem
- For Areva SA: the capital is 100 % held by the state
- For EDF: the capital is 83.6 % held by the state, 12.9 % by institutional shareholders, 2 %
by individual shareholders and 1.3 % by employee shareholders (the remainder in treasury
shares).
Serious consequences
for the entire sector
Former
Areva group
dissolved
in 2017
A
reva SA
responsible
for completing
the end
of OL3's work
in Finland
and settling
the disputes
of its subsidiary
Areva NP
NewAreva NP
Subcontractors
Framatone
EDF subsidiary
since 2018
Areva TA
which became
TechnicAtome*
Orano*
new company
(formerly
Areva NC
)
Now
Principal
Service provider
relationship
EDF
Client
Ongoing litigation
12
Summary of the Thematic Public Report of the Cour des comptes
Serious consequences for the entire sector
The consequences of these deviations
obviously
impact
the
costs
and
profitability of the Flamanville EPR .
Its construction cost is estimated by
EDF at €12 .4 billion2015, to which
additional costs will be added which
could reach nearly €6 .7 billion2015 on
the commissioning of the reactor, still
being scheduled for mid 2023, including
around €4 .2 billion in financial costs .
Under these conditions, it is regrettable
that
neither
EDF
nor
the
public
authorities concerned calculated the
profitability forecast of the Flamanville
3 EPR, apparently considering it normal
that it should be absorbed into the
average profitability of all electronuclear
reactors . In the absence of data produced
by the company, the Cour has estimated,
based on hypotheses set out in the
report, that the cost of the electricity
produced by the Flamanville EPR could
be between €110 and €120 / MWh .
Evolution of the building cost of Flamanville 3 between 2006 and 2019
(billions of €2015)
Source: Cour des comptes based on EDF data
0
2006 2007 2008 2009 2010 20112012 2013 2014 2015 2016 2017 2018 2019
2
4
6
8
10
12
14
13
Summary of the Thematic Public Report of the Cour des comptes
The Olkiluoto project, in Finland, was
launched in 2005 by Areva, with its
commissioning scheduled for 2009 .
It experienced setbacks, delays and
additional costs which contributed to
the disappearance of the former Areva
group . The provisional acceptance
of the project was, in March 2020,
scheduled for the end of May 2021, but
uncertainty lingers and is, according
to a recent announcement by the
electrical engineering company TVO,
the plant’s purchaser, added to by the
probable impact of the covid-19 health
crisis on the works still necessary .
The construction cost would amount
to €8 .2 billion (in current euro) for
the part insured by Areva, without
taking into account that of the turbine
(€684 billion),
according
to
the
information provided by Areva SA, i .e .
nearly 4 times the amount planned in
the initial contract (€2 .28 billion for
the consortium formed by Areva and
Siemens) .
The EDF’s investments in the United
Kingdom
and
the
Hinkley
Point
project are high-risk operations . EDF
has disbursed €15 .7 billion for the
acquisition of British Energy and will
have to disburse €16 to 17 billion for
the construction of the two Hinkley
Point C reactors (HPC), if there are no
further deviations from costs . Risks
remain after the announcement of a
£3bn increase in construction costs
and of longer reactor construction
times in 2019 . The profitability of
the Hinkley Point project has been
revised downward several times since
the launch of the project . Financing is
provided by EDF, in the amount of its
participation in the project company’s
capital, which has had a huge impact
on its financial situation .
The Taishan 1 and Taishan 2 reactors
were successfully commissioned in
China in 2018 and 2019, but five years
late with respect to the schedule laid
out in the order, and with an additional
cost of 60% compared to the estimated
budget . Questions remain regarding
the setting of the purchase price of the
electricity produced by these reactors
and the project’s profitability for EDF .
The other EPR export projects are
full of uncertainties . Despite EDF’s
marketing efforts and those made by
the French government to offer India
very favourable financial conditions,
the negotiations, which began a long
time ago, are making little progress .
As for the EPR reactor construction
projects in Sizewell in the United
Kingdom, their execution is subject to
guaranteeing their financing, EDF no
longer having sufficient capacity to do
so .
An international strategy
that cannot be continued
under the same conditions
15
Summary of the Thematic Public Report of the Cour des comptes
Drawing lessons from the difficulties
encountered
in
building
the
EPR
reactors, but comforted by the fact
that those at Taishan are working well,
which
validates
this
technological
choice, EDF offers the ASN and the
administrative authorities a new EPR
model, called «EPR2» presented as
simpler and cheaper to build .
Three-dimensional representation of the EPR 2 nuclear island
Source: EDF
The construction of a series
of EPR2 in France:
a technological, economic
and energy policy choice
16
Summary of the Thematic Public Report of the Cour des comptes
The construction of a series of EPR2
in France: a technological, economic
and energy policy choice
By making this choice, EDF moved
away from the process of optimising
the EPR technology based on reaping
the benefits from the lessons learned .
The Olkiluoto 3 and Flamanville 3
projects have shown that prioritising
innovation
over
accumulated
experience entails risks and that the
cost of this innovation should not
be underestimated . However, it is
impossible to determine with any
reasonable degree of certainty that
the savings in building future EPR2s
compared to the cost of building
Flamanville-type EPRs will materialise .
However, the only hypothesis currently
being studied by the public authorities
in terms of new nuclear energy is that
of building six EPR2 type reactors, in
pairs .
The financial stake is major: the cost
of building three pairs of EPR2 is
estimated at €46 billion2018 . The
decision as to whether or not to build
future EPRs will have consequences
through to the 22nd century . It must
therefore be taken on the basis of
a feedback process that is not only
carried out internally at EDF, but
involves all of the players involved in
the building of all the EPR reactors
built or in the process of being built,
so that all stakeholders learn the
same lessons from the way the EPR
construction projects have unfolded .
In this scenario, electronuclear financing
methods must be implemented . EDF
cannot finance the construction of new
reactors by itself and can no longer
engage itself without guarantees on
the revenue generated by them . No
new project can be launched without
a form of public guarantee, whatever
the
mechanism
adopted .
However,
the burden thus transferred to the
consumer or to the taxpayer would only
be justified if the electricity produced by
the new nuclear power reactors proved
sufficiently competitive with respect to
other modes of electricity production,
renewable in particular, or if other
considerations justified keeping nuclear
power in the electricity mix .
This is why a complete analysis of the
electricity mix by 2050, presenting
challenges and solutions regarding
the assurance of supply, adaptation of
electricity transport and distribution
networks, management of radioactive
waste, dismantling of plants currently
in operation, and of course the cost
of operation of the electricity system,
which
must
be
conducted
before
any
decision
is
made
concerning
the development of a new fleet of
electronuclear reactors .
This decision having been put off
by the Government to beyond the
commissioning of the Flamanville
3 reactor, i .e . mid 2023 at the earliest,
it is possible to draw up, by then, both
the complete lessons learned about
the building of the EPRs and the
long-term planning of the electricity
mix recommended by the Cour .
17
Summary of the Thematic Public Report of the Cour des comptes
Recommendations
Recommendations on the execution
of large-scale projects
1.
Reconsider
the
concept
of
architectural designer by separating
the functions of client and project
manager
(EDF, 2020)
2.
Incorporate in contracts provisions
sharing the construction risk between
the
client
and
the
contractors
and others, engaging them in the
management of the works execution
schedule
(EDF, 2020)
3.
Carry out a six-monthly review of
the strategic projects and the risks
associated with them, at meetings of
the EDF board of directors
(EDF, MTES,
MEF, 2020)
4.
Make sure that the managers of
large-scale projects have authority
over the resources, especially the
engineering resources, required for
their execution (
EDF, 2020)
5.
List
in
a
shared
reference
document the terms of application
of the failure exclusion in order
clarify the industrial consequences
of the specifications in question
(EDF,
Framatome, immédiat)
Recommendations
concerning
the
preparation
of
forthcoming
decisions
6.
Calculate
the
provisional
profitability of the Flamanville 3
reactor and of the EPR2 and follow it
up
(EDF, 2020).
7.
Define,
before
engaging
in
international projects, their risk level
and expected profitability and their
financing conditions, and keep to
them
(APE, DG Trésor, EDF, 2020).
8.
Conduct a complete lesson learned
exercise on all the EPRs built or being
built in France and abroad, with all
the stakeholders concerned, before
launching
any
new
electronuclear
reactor project
(EDF, MTES, MEF,
2020).
9.
Extend
the
planning
of
the
electricity mix through to 2050,
before
taking
the
decision
to
launch
any
new
electronuclear
reactor project
(EDF, RTE, MTES,
MEF, 2020).