Construction Arbitrations in the Nuclear Sector

The nuclear sector has given rise to some of the most high-profile construction arbitration in recent years. The nature of nuclear power plant (NPP) projects means that many of these arbitrations take place in the full gaze of the nuclear industry press and environmental campaigning media. These arbitrations are often long running, and involve large teams of lawyers[2] and technical nuclear experts, in addition to the ‘usual’ delay and quantum experts. The sums at stake (frequently in the multiple billions of dollars) and the technical complexities of the issues in dispute can test even the most experienced construction counsel and arbitrators.

International arbitration in the nuclear sector

Four examples provide a good starting point to understand the nature of construction arbitration in the nuclear sector.

An International Chamber of Commerce (ICC) arbitration concerning the delayed completion and increased cost of the Olkiluoto 3 NPP in Finland ran from 2008 until 2018, when the parties reached a widely publicised global settlement. The NPP itself was originally planned to be operational in 2009; as of writing, regular electricity generation is anticipated to start in September 2020.[3] The press reporting of this arbitration was at times more akin to that of an investment treaty arbitration than a construction dispute; be it increases in sums claimed and counterclaimed,[4] failed attempts at settlement, hearing dates, the issuing of partial awards,[5] final settlement,[6] etc., the minutia of the arbitration is readily available for public consumption. In parallel, the Finnish nuclear regulator STUK has been producing regular reports on the project, setting out its detailed findings on a broad range of issues potentially relevant to the delay and financial claims before the tribunal.[7]

In contrast, a London Court of International Arbitration (LCIA) tribunal was reported as taking just three years to decide a dispute between a Hyundai-Samsung joint venture and the South Korean state-owned entity, Korea Electric Power Corporation, concerning variations, changes and disruption to the Barakah nuclear plant in the UAE.[8]

Another widely reported ICC arbitration concerned the San Onofre Nuclear Generating Station (SONGS) in California, USA. SONGS contains three reactor units, only two of which were operational at the relevant time (Units 2 and 3), Unit 1 having been permanently shut down in 1992. Mitsubishi Heavy Industries (MHI) designed and fabricated a number of replacement steam generators for Units 2 and 3 as part of a 10-year upgrade programme. In January 2012, after only 11 months of operation, one of the Unit 3 replacement steam generators experienced a radioactive coolant leak. Unit 3 was taken out of service in order to prevent further leakage and allow investigations as to the cause of the failure. Unit 2 (which was offline at the time due to planned maintenance) was also inspected. Investigations revealed that the replacement steam generators in both units were failing. In June 2013, the owner of SONGS, Southern California Edison (SCE), announced that it would permanently shut down the units, and begin decommissioning SONGS. In September 2013, the US Nuclear Regulatory Commission (NRC) issued its findings; NRC identified that the computer codes used by MHI to design the replacement steam generators were flawed, inaccurately predicting thermal hydraulic conditions in the steam generators, leading to tube vibration and wear, and ultimately to stem generator tube leak. One month later, SCE filed a Request for Arbitration against MHI seeking damages of not less than US$4 billion, subsequently increased to US$7.57 billion. In March 2017, the tribunal issued its award. MHI did not dispute the allegation that replacement steam generators in Unit 3 experienced unacceptable wear, but relied on a limitation of liability in the Purchase Order. The tribunal (by a majority) upheld the limitation provision and ordered MHI to pay the owners only US$125 million.[9] In addition, the owners were ordered to pay US$58 million to MHI in costs.

A third ICC arbitration in the sector determined liability as between the Bulgarian state energy firm NEK and the Russian NPP supplier, Atomstroyexport, following the cancellation of the Belene NPP project. The supplier sought more than €1 billion in compensation. In an award that was over 700 pages, the tribunal awarded the supplier €550 million in compensation, with the parties to discuss whether the NPP would be built or sold on to a third party.[10] One year later, Atomstroyexport supplied the reactor vessels for the abandoned project to NEK. Two years later, and Bulgaria is expected to lift the moratorium on construction of the Belene NPP, with Atomstroyexport one of the four potential suppliers who have registered interest to complete the project.[11]

Not only were these arbitrations played out in the public view, but in each case the arbitration was only one part of the story. Public scrutiny, criticism and court action from environmental groups, together with investigations and public reporting by the nuclear regulators, were running in parallel. Probity and cost audits, assessing the extent to which the financial consequences of these disputes can be passed through to consumers often follow. Further, the limited pool of owners and suppliers[12] (both Tier 1 and lower tiers) active in this sector means that in many cases, whatever the outcome of the arbitration, the participants will have little or no option but to work together over the coming decades. This is typical of arbitration in the nuclear sector, requiring counsel (and arbitrators) – to a far greater extent than in many construction arbitrations – to pay attention to the wider picture.

Nuclear safety – the overriding requirement

In conventional power projects, owners and suppliers juggle the competing constraints of time, cost and quality, with priorities driven by the commercial objectives as reflected in the underlying construction or supply contract.

For a nuclear project, nuclear safety – overseen by the local nuclear regulator – is superimposed over these constraints. Quality is non-negotiable, limiting the parties’ ability to compromise on time and cost. Every aspect of the project must comply with the relevant nuclear safety requirements, irrespective of how long or how costly it is to achieve compliance.

This overriding requirement generates three risk streams, one or more of which will be in issue in any construction dispute in the nuclear sector:

  • the licensing risk;
  • the risk of constructing in a nuclear environment; and
  • nuclear liability.

The licensing risk

Any nuclear project carries with it a substantial licensing risk:

  • Will the nuclear facility be licensed at all?
  • If so, how long will it take for the nuclear facility to be licensed?
  • What conditions will be attached to the licence?

Most nuclear regimes provide a two-track process: a high level licensing procedure, and day-to-day detailed regulation. Understanding these two separate processes, and knowing what are the ‘hold points’ – points beyond which the project cannot progress without approval from the regulator – is key when presenting or determining disputes in the nuclear sector.

High-level licensing

The high-level licensing process is often very political and public. While the terminology varies from regime to regime, most regimes include the following four gateways:

  • an initial licence to build a NPP;[13]
  • a construction licence (typically the ‘hold point’ for the pouring of First Concrete[14] or safety classified manufacture, for example, forging of the reactor pressure vessel or other components within the primary circuit);
  • the licence to permit the transportation to and loading of nuclear fuel, at the NPP; and
  • the operating licence required to commence operation of the NPP.

The first gateway will often have been passed before the supplier is committed to the project. The remaining gateways will typically be key milestones in the project schedule, often linked to the payment of liquidated damages for delay. As such, they will likely feature prominently in any construction arbitration where a supplier, key component vendors or subcontractors, are claiming additional time or money, or the owner is seeking to deduct liquidated damages. Nuclear experts (often retired regulators from other states, who have experience of the particular or equivalent nuclear regimes)[15] will usually be needed to explain the licensing process to the tribunal, and provide opinions as to the cause of delayed achievement of each gateway, or the foreseeability of any pre-conditions imposed by the regulator.

Detailed regulation

The less public, but no less important, licensing process relates to the oversight, supervision of and intervention into the day-to-day workings of the full supply chain by the nuclear regulator. While the most onerous regulations apply to contractors and suppliers whose work carries a high safety classification, the reality is that the nuclear regulator will take an interest (and potentially intervene) in all the tiers of the supply chain, supervising each phase of the project, from basic through to detailed design, procurement, manufacturing, factory testing, construction/erection and installation through to commissioning.

Common claim events arising from the implementation of the detailed regulatory regime include:

  • delayed approvals or extended and repeated approval cycles;
  • additional or revised requirements imposed during the approval process. For example, the regulator may require additional engineering studies to be carried out to justify the proposed design prior to approval;
  • additional testing required to demonstrate the quality of components, materials and as-built works;
  • the imposition of different ways of working during manufacture, construction or commissioning, which may cause delay, reduced productivity and/or additional cost; and
  • the imposition of changes to the NPP design as a pre-condition to approval of the works.

In a highly regulated environment, it is unrealistic for the parties to anticipate no changes or unforeseen regulator requirements. In addition, the nuclear regulator – charged with achieving the overriding objective of nuclear safety – will typically supervise and intervene on nuclear works to a much greater degree than an owner’s representative supervising conventional projects. Further, the regulator does not bear the cost of these interventions; the regulator’s only concern is (and should be) nuclear safety. However, changes, interventions and additional or different requirements can have substantial impacts on the project schedule and cost, giving rise to claims in arbitration as to who must bear the financial consequences between the owner and supplier or through the supply chain.

The relatively limited recent activity in NPP construction in many parts of the world means that it can be difficult for owners and the supply chain to predict what the regulators will require from them. On the one hand, while the nuclear regulations may not have changed substantially from the last NPP construction, the individuals who interpret and implement the regulations will likely be different and have their own views as to what compliance requires, which may, or may not, coincide with the views of their predecessors. Alternatively, the regulations may have been updated but not implemented since the last NPP construction in the jurisdiction, leaving the owner and supply chain with little guidance as to what the new regulations actually require of them. Parties will often seek to mitigate this risk through the use of existing NPP as ‘reference plants’ against which regulator-imposed change can be measured. However, unless the existing NPP is a ‘twin’ the comparison may not be straightforward. Indeed, where the new NPP involves significant developments from the earlier NPP, is a ‘first of a kind’, or is being constructed in a different state with a different regulator, or for a different owner with their own set of ‘owner preferences’, the utility of a reference plant as a benchmark can be of limited benefit, and may generate considerable technical and legal argument during an arbitration.

In practice, it is often this daily oversight and supervision of the project that gives rise to the majority of claims during the construction phase. Expert evidence from experienced nuclear experts as to the foreseeability and consequences of the regulator’s conduct will form an essential part of the supplier’s case, together with the usual evidence from delay, disruption and quantum experts as to the time and financial impacts.

The risk of constructing in a nuclear environment

Constructing in a nuclear environment poses additional risks to contractors beyond those found on any conventional major project. It is essential that counsel, experts and tribunals fully appreciate these risks, in order to be able to quantify the full impact of claim events on the project schedule and the supplier’s costs. In particular, the limited ability for the supply chain to work ‘at risk’, and the limits on the supplier’s ability to use normal project management techniques to mitigate potential delays and cost overruns, mean that when a nuclear project encounters difficulties, these can result in substantial delay and additional cost compared to equivalent difficulties on a conventional project. These constraints are addressed in turn.

Limited ability to work at risk

Many construction contracts will permit a contractor working on a conventional project to work ‘at risk’. The contractor can choose, for example:

  • to place long lead item procurement contracts before the component vendor (or even the contractor) has been approved in order to guarantee a manufacturing slot and delivery date, or fix a favourable price;
  • to start construction before engineering drawings have been issued ‘for construction’ in order to gain time; or
  • to undertake a major concrete pour before concrete test results are available, in order to take advantage of good weather.

The contractor does so knowing that if the vendor, drawings or test results are not approved, the owner will inevitably require opening up and investigation of the works at the contractor’s expense to check that the physical works conform with the contractual specifications. Where these investigations show that the work is non-compliant, the contractor may be instructed to remove and rework at the contractor’s time and cost. However, the contractor may well take the view that it has confidence in its own internal quality controls and the quality of its supply chain, engineers and supervisors, such that there is minimum risk in practice or that the risk is worth taking given the liquidated damages for delay that may be due if the contractor does not keep to the project schedule.

In the nuclear sector, many nuclear regulations prevent or severely restrict the ability of the supplier to work ‘at risk’, prohibiting suppliers (at Tier 1 and the lower tiers) from placing contracts or progressing work without prior formal approval. This is a particular issue where the works concern safety classified features of the NPP. With the supply chain unable to work ‘at risk’ (or constrained as to when this is possible), it can be very difficult (if not impossible) for delays or additional costs to be mitigated. As a result, any delay in approval by the owner or the regulator can have a very significant impact on the project schedule (and create substantial additional cost). These ‘hold’ points must be understood by the delay experts engaged by the parties, in order that they can properly assess the full impact of even small delays on the overall project.

Limited ability to manage internal processes

A further significant difference between conventional and nuclear projects is the extent to which the applicable nuclear regulations limit both the owner and the supplier’s ability to manage their internal processes.

In a conventional major project there will, of course, be project procedures in place that govern how the contractor manages the project. However, typically, these procedures are provided by the contractor, and can be amended or deviated from as the contractor sees fit. The situation is quite different when constructing in a nuclear environment, with the owner and supplier’s internal processes often requiring approval from the regulator, and limited to no ability to amend or deviate from the processes without prior approval. Where the regulator requires the supplier to work in a non-usual way, or is slow (or refuses) to approve proposed changes, this can create inefficiencies in the supplier’s workforce, or limit the supplier’s project team from improving productivity through changed work practices.

For example, in a conventional project the contractor can typically manage late design developments on the construction site, through a site-based engineering team who deal with buildability issues as and when they arise on the shop drawings. In a nuclear project, where limited work is possible without formal approved construction drawings, the supplier may be prohibited from adopting this approach. A small construction issue can therefore cause large areas of site works to stop while the design is reworked and approvals processed, rather than the supplier implementing a suitable workaround on-site.

Another example concerns welding; a significant part of any NPP project. On a conventional project, the contractor is free to reallocate welders around the site as and when work is available. In contrast, under many nuclear regimes, the approval of welders is highly regulated, with different welds requiring different certifications and approvals. Accordingly, the supplier is unable to reallocate welders freely, and must match specific welders to specific tasks. This is an important issue when considering loss of productivity claims, increasingly one of the most significant financial claims put forward in construction arbitration. Welding productivity is closely monitored on both conventional and nuclear projects, with experts normally able to assess with some accuracy the actual and anticipated productivity of welding teams, and quantify the loss of productivity. On nuclear projects, the greater variability of weld types, and the inflexibility of the workforce, makes the analysis much more complex than for conventional projects.

The intensive supervision of the works by the nuclear regulator – and in particular, the supervision of safety classified works – will also limit the supplier’s ability to revise construction, erection and commissioning sequencing in order to progress the works. For the same reason, the owner’s freedom on conventional projects to agree substitutes and compromises in order to keep the project on schedule (or to mitigate delays) is restricted, with almost any change requiring review and approval by the regulator before it can be implemented.

As for working at risk, the experts and counsel must understand each of these components of constructing in a nuclear environment, and be able to explain to the tribunal how the regulatory environment impacts on the owner and supply chain, and who bears the risk.

Nuclear liability[16]

Civil liability for nuclear damage

The third risk stream for participants in nuclear projects relates to their potential liability for nuclear damage.[17] This liability can arise only after nuclear material has been brought on-site: during or after fuel loading, commissioning, operation, maintenance or upgrade projects.

The Chernobyl disaster provided the clearest reminder that nuclear damage does not respect international borders; notwithstanding the significant emphasis on nuclear safety at every stage of a nuclear project, a nuclear incident[18] has greater potential to cause damage across a widespread geographical area than almost any other form of energy production or industrial process. Liability for nuclear damage suffered within the same jurisdiction as the civil nuclear installation is governed by domestic legislation. In addition, many states have signed up to a series of international conventions addressing civil liability for nuclear damage suffered in a state other than the state where the nuclear incident occurred.

The international nuclear liability regimes

The four most significant international conventions are:

  • the Paris Convention on Third Party Liability in the Field of Nuclear Energy 1960 (the Paris Convention) produced under the auspices of the Nuclear Energy Agency within the Organisation for Economic Co-operation and Development;
  • the Brussels Supplementary Convention 1963 to the Paris Convention (the Brussels Supplementary Convention);
  • the Vienna Convention on Civil Liability for Nuclear Damage 1963 (the Vienna Convention) produced under the auspices of the United Nations International Atomic Energy Agency; and
  • the Joint Protocol Relating to the Application of the Vienna Convention and the Paris Convention 1988 (the Joint Protocol).[19]

The Paris Convention and the Brussels Supplementary Convention are in force in most Western European countries;[20] the Vienna Convention has a broader geographical reach, in force in 40 states including some signatories to the Paris Convention.[21] While the detail of the conventions differ, the conventions are founded on the same basic principles:

  • liability for nuclear damage is channelled exclusively to the operator of the nuclear installation;
  • the operator’s liability is absolute irrespective of fault;
  • the operator’s financial liability per incident is subject to certain limits;
  • compensation claims must be brought within the state courts of the place where the nuclear incident occurred (normally the location of the nuclear facility) within a set period of time;
  • the operator must maintain insurance or other financial security commensurate to the liability. If this security proves insufficient to meet the specified liability, the state where the nuclear facility is located must make up the difference; and
  • the state court of the contracting state where the nuclear incident has exclusive jurisdiction over claims.

The Joint Protocol was introduced in 1988, following the Chernobyl disaster, which had exposed the difficulties in having two similar (but not identical) conventions. The Joint Protocol (in force in 28 states) seeks to extend the application of the Paris and Vienna Conventions as between the signatories, while avoiding duplication or inconsistencies.[22]

A fourth convention – albeit one that is currently in force in 10 states only[23] – is the Convention on Supplementary Compensation for Nuclear Damage. This convention supplements the Paris and Vienna conventions, bringing into the international liability regime notable absences (in particular, Canada, the United States and Japan) and establishing provisions for enhanced cover to be provided by the contracting states.

Nuclear liability and construction arbitrations

Given the channelling of liability to the operator, the supply chain for nuclear facilities located within a contracting state should not face direct claims in relation to nuclear damage. However, nuclear damage claims may still feature in construction arbitrations if the supply contract between the operator and the supplier allows the operator to pass liability for nuclear damage down to the supplier, or does not expressly exclude it.[24]

Practical issues to consider when arbitrating in the nuclear sector

Arbitrating in the nuclear sector gives rise to a number of practical issues that must be addressed in order to ensure an efficient construction arbitration.

The scale of the project and potential issues in dispute

The first issue is the need to appreciate the scale of NPP projects.

Even a project that proceeds from start to finish without any unforeseen difficulties will likely take more than 10 years from conception to operation.[25] Where a dispute arises in the early years of the project, there is every possibility that the arbitration may last for 10 years or more as it tracks the progress of the project itself. This has practical issues when staffing the arbitration, in particular, when choosing fact witnesses, experts and arbitrators.

The volume of documentation generated through the licensing, design, procurement, construction and commissioning of a NPP is unequalled in the construction industry. Hundreds of thousands of flow diagrams and isometrics, civil, mechanical, electrical and I&C drawings will be created, in addition to technical design justifications, work methodologies, procurement documentation, QC records, etc., to say nothing of correspondence, meeting minutes and reports. The volume of documentation – and server capacity required to store and access this quantity of technical documentation – will be impossible with any standard litigation software, and will likely require bespoke document management systems to be developed and maintained. Filtering the documentation into manageable quantities of relevant material for the arbitration team to work with, and the tribunal to review and base their decisions on, is a project in itself.

Traditional methods for analysing the impact of events on time and cost may not be appropriate

The complexity of a project to develop a NPP cannot be overstated. Every stage of the process will likely involve thousands of engineers, manufacturing and construction personnel in an iterative process that cannot fully be reflected in a traditional project schedule. Project schedules for NPP projects can easily include over 100,000 separate activities in each phase; the logic links between these activities cannot possibly reflect every iteration that will occur during the life of the project. While the schedules provide an essential tool to the project managers, they must be used with care by the delay experts. An expert who adopts a very schedule-based approach to delay analysis will quickly find themselves struggling to provide a meaningful opinion.

Another difficulty is caused by the fact that the scheduling software used on most construction projects (and by most delay experts) models the project on a linear basis. This is broadly representative of how a simple project responds to change and delay. However, a major project – and in particular, a project as complex as a NPP project – does not respond to change and delay in a linear fashion; it responds dynamically. The same is true of productivity and cost. The supplier or contract will require additional expert opinion on the actual consequences of change and delay, which will often far exceed the consequences assessed by reference to traditional linear analysis. While not often seen in conventional construction arbitration, reference to experts in systems dynamics may be required. A developing science in litigation or arbitration, systems dynamics has been widely used by management within the energy and other sectors for many years to predict outcomes when making significant investment decisions, etc. It is possible to create a systems dynamics model for the purposes of arbitration supported by appropriate independent expert evidence. However, in the author’s experience, many companies engaged in complex mega projects will have their own preferred modelling system used on a day to day basis to model risk and business decisions as part of its normal business management. Where such a model already exists, this model can be used by the relevant factual witness evidence (for example, a senior risk manager or similar) from the organisation, who can explain the build-up and historical reliability of the model. In the author’s view, a ‘tried and tested’ model that was in use long before a dispute arose and that can be explained by a well-informed fact witness, will likely be better received by a tribunal than a model created (or tailored) by an expert after the event and solely for the purposes of arbitration.[26]

Sampling and extrapolation

It is not uncommon for contractors on conventional projects to attempt to quantify multiple claim events by reference to sampling and extrapolation.

The approach most frequently presented in construction arbitrations uses non-statistical sampling. Most often due to constraints of time, resources or money, the contractor’s arbitration team will review in detail a limited sample of claim events that the contractor says caused a loss. This sample will be presented supported by factual witnesses or contemporary records, with technical experts opining on the technical aspects, and the quantum experts opining as to the resulting loss. The contractor’s fact witnesses or the experts will opine that the sample was representative, and the contractor will ask the tribunal to infer the overall loss from the sample. While this approach can result in a large claim number, it is frequently easy to undermine the analysis by demonstrating that the sample is unrepresentative.

For a nuclear project, the supplier may have literally hundreds of thousands of claim events. It is clearly impossible to present each of these claim events – and no tribunal would thank the arbitration counsel who put forward its client’s case on this basis. Engaging an expert statistician to develop and supervise a statistically robust sampling and extrapolation process can allow the financial impacts of a large number of claim events to be presented in a comparatively cost-efficient and digestible format, suitable for the tribunal to determine within a reasonable period of time.[27]

Export regulations

A further practical issue that arises on construction arbitrations in the nuclear sector is the need to comply with domestic and international regulations concerning the transfer of nuclear technology.

Nuclear technology often lies at the heart of a construction arbitration, with claims for additional time and money founded on alleged changes to existing technology or the time and cost incurred in developing new or enhanced technologies. Determination of these issues requires detailed information regarding the technologies to be transferred to the independent experts, counsel and the tribunal. Where that technology involves a transfer across an international border, the transfer may require advance permission from a regulator or export authority.

Counsel must be aware of these regulations, and ensure compliance across their internal and expert teams. The tribunal and arbitral institution (if applicable) must also be made aware of applicable regulations and terms of any permissions granted. For example, many international law firms and expert witness providers staff matters across offices, or hold data on servers in multiple jurisdictions to manage the risk of data loss in a given jurisdiction. However, the synching of data between servers in multiple jurisdictions may constitute the transfer of technology requiring advance permission from the relevant regulators. The use of cloud storage may also constitute a transfer of technology, as may a member of the legal or expert team (or the tribunal) travelling internationally with data about the technology on their laptop. Similarly, the filing of expert reports and submissions, or taking bundles to a hearing, may constitute the exporting of restricted nuclear technology.

Security issues

Every nuclear facility will incorporate security measures protecting the facility from both physical attack and cyberattack. The development, approval and incorporation of these measures into the facility is a matter of extreme sensitivity. Information regarding these measures is strictly controlled and often classified, with only a limited number of individuals within the supplier and the owner authorised to know what the measures are, let alone the status of the measures and the extent to which these measures may be a cause of (or impacted by) delay, disruption and additional cost.

Where a security measure gives rise to a claim, this will create additional practical difficulties in arbitration. Counsel, experts and arbitrators will likely require security approval from the relevant security agencies in order to be able to receive information regarding the measures. All information regarding the security measures will be subject to strict safeguarding requirements, and a failure to comply with the prescribed procedures may result in criminal sanctions against the individuals. These restrictions will permeate through the entire arbitration; for example, thought must be given when appointing the tribunal and staffing the counsel and expert teams to whether certain individuals may be prohibited from having access to information due to their nationality. Practicalities, such as whether a transcript can be taken of the part of the hearing where the measures are discussed, and if so, by whom, and on what terms, need to be considered. Where the arbitration is supervised by an institution that would normally receive a copy of the award (and in particular, for ICC arbitrations, where awards are scrutinised by the ICC Secretariat and ICC Court), discussions will need to take place at an early stage to agree with the relevant authorities how the arbitration can proceed.

‘Lessons learned’

Transparency is a fundamental part of the nuclear sector, with the nuclear sector investigating and publishing ‘lessons learned’ regarding every aspect of nuclear projects that did not proceed as planned. It is therefore entirely possible, and indeed, likely, that the nuclear regulator will not be alone in publishing findings as to the cause of problems on the project. Counsel on a construction arbitration in the nuclear sector may well find that senior personnel from its own client are publishing papers setting out ‘lessons learned’ on the very issues that are before the tribunal.


This brings this discussion back full circle to the start of the chapter. The nature of nuclear power is such that any nuclear project will be developed – and disputes determined – in the full gaze of the industry and, on occasion, the world’s press. Nuclear safety takes priority over everything, including arbitration strategy, however much money is at stake. The key skills for counsel working in this area are (1) to know what to expect; (2) be prepared to distil an unprecedented volume of material into manageable parts that can be understood and determined by the tribunal; and (3) present the public findings regarding the cause of your client’s claim events as a virtue, not a weakness. And remember, once you have presented or determined a construction arbitration in the nuclear sector, every non-nuclear construction arbitration you are involved with thereafter (however complicated) will seem straightforward!


[1] Jane Davies Evans is a barrister practising as counsel and arbitrator at 3 Verulam Buildings.

[3] See (accessed 14 August 2019).

[4] See for example, and (accessed 21 August 2018).

[9] See (accessed 21 August 2018). The redacted version of the dissenting opinion now published records (at paragraph 192) that the dissenting arbitrator (Johnathan D. Schiller) would have awarded the owners more than US$1 billion in damages. See$FILE/I1210013-Final%20Redacted%20Corrected%20Dissent%20(Public%20Version).pdf (accessed 21 August 2018).

[12] In this chapter, the term ‘supplier’ is used to denote the Tier 1 supplier who contracts directly with the owner unless otherwise stated.

[13] In some states this will involve a public referendum.

[14] ‘First Concrete’ is the term used in the nuclear construction sector in relation to the first pour of safety classified concrete, typically the base slab under the reactor.

[15] There is significant collaboration between many state regulators, together with the international bodies who oversee nuclear activities on an international basis, meaning that many regulators will have substantial personal experience of the regulatory regime of other states.

[16] A full description of nuclear liability is outside the scope of this book, but counsel appearing in construction arbitrations in this sector should be familiar with the basics of the international nuclear liability regime. See, for example, ‘Liability for Nuclear Damage’ (updated August 2018), World Nuclear Association, (accessed 21 August 2018).

[17] ‘Nuclear damage’ is defined in the Vienna Convention on Civil Liability for Nuclear Damage (the Vienna Convention) as:

(i) loss of life, any personal injury or any loss of, or damage to, property which arises out of or results from the radioactive properties or a combination of radioactive properties with toxic, explosive or other hazardous properties of nuclear fuel or radioactive products or waste in, or of nuclear material coming from, originating in, or sent to, a nuclear installation; (ii) any other loss or damage so arising or resulting if and to the extent that the law of the competent court so provides; and (iii) if the law of the Installation State so provides, loss of life, any personal injury or any loss of, or damage to, property which arises out of or results from other ionizing radiation emitted by any other source of radiation inside a nuclear installation.

[18] ‘Nuclear incident’ is defined in the Vienna Convention as ‘any occurrence or series of occurrences having the same origin which causes nuclear damage’.

[19] All three conventions have been supplemented by subsequent protocols.

[23] The convention is in force in Argentina, Canada, Ghana, India, Japan, Montenegro, Morocco, Romania, the United Arab Emirates and the United States. See (accessed 21 August 2018).

[24] Alternatively, the supplier may seek to pass down liability to lower tiered suppliers in arbitrations under subcontracts.

[25] To take an extreme example, the ITER project – an international 35-year collaboration to design and construct the first magnetic nuclear fusion plant currently under construction in France – commenced site works in 2007. Construction of the Tokamak Building commenced in 2014 with a seven-year build phase, First Plasma is planned for 2025 and Deuterium-Tritium Operation currently planned for 2035. See (accessed 21 August 2018).

[26] Non-linear analysis may also be appropriate for construction arbitrations outside the nuclear sector where the project involves multiple interactions that are not properly reflected in the traditional analyses. For example, systems dynamics is a tool relatively common within the oil majors. Until such time as non-linear analysis is widely accepted by construction arbitrators, the author recommends that this evidence is presented in conjunction with traditional delay and quantum analysis.

[27] In the author’s view, statistical sampling is also suitable for use on many construction arbitrations concerning conventional projects as a mechanism for saving time and cost.

Unlock unlimited access to all Global Arbitration Review content