ASN-IRSN merger: four principles to understand when discussing this project

Since the Government announced the ASN-IRSN merger project, an understandable emotion has affected part of the nuclear industry to the point of suggesting certain misconceptions about the governance of safety in France and its philosophy. Without taking sides and aiming to move towards a calm debate, Sfen proposes to take another look at some of the key elements at the basis of this organization.

Following the last Nuclear Policy Committee (CPN), the Government announced on February 8, in a press release, its plan to combine the technical expertise of the Institute for Radiation Protection and Nuclear Safety (IRSN) with that of the Nuclear Safety Authority (ASN). The executive specifies that it will be “vigilant in taking into account synergies, with the “Commissariat à l’énergie atomique et aux énergies alternatives” (CEA) and the “Délégué à la sûreté nucléaire et à la radioprotection pour les activités et installations intéressant la Défense” (DSND).” The Minister of Energy Transition, Agnès Pannier-Runacher, has asked the President of the ASN, the Director General of the IRSN, and the Administrator General of the CEA, to propose, by the end of February, “the first measures and a working method to implement these orientations, before a more detailed roadmap given the 2024 Finance Act.”

Emotions are particularly high among IRSN staff. The Institute has built up a real identity over the past twenty years, and its staff of around 1,800 people rightly feel that they have succeeded in their recognized public interest missions. We must also listen to the reflections on the risks to be avoided in the implementation of future reorganizations.

On February 14, during a special hearing of the Parliamentary Office for Scientific and Technical Choices (OPECST) or through the press, many personalities, parliamentarians, and stakeholders reacted both on the form (the sudden nature of the announcement, the absence of prior debate or impact study) and the substance. The opinions against this development of the two organizations mention the risk of regression in terms of safety control or a loss of confidence in its governance.

However, specific arguments deserve to be deciphered in that they may reflect certain misunderstandings about safety governance, expertise, or research. As a learned society, Sfen does not have an opinion on the reorganizations to be implemented, which are the responsibility of the operational conduct of the State’s missions. But sharing elements of analysis is part of its raison d’être. The nuclear safety control system’s solidity, efficiency, and credibility are essential to use this energy. Its fundamentals must be understood and shared by all.

1- The ASN controls nuclear safety, and the IRSN is not a counterweight

During the recent debates, certain positions could lead one to believe that the IRSN is the ultimate guarantor of safety and a counterweight to the ASN. Thus, if the ASN is “pressured,” for example, for the security of supply purposes, the IRSN could be a kind of “whistleblower,” one might think.

This vision seems utterly contrary to the spirit of safety governance advocated internationally and organized by the 2006 law on transparency and security in nuclear matters (TSN). The ASN has the status of an independent administrative authority under the control of Parliament. On behalf of the State, it is responsible for monitoring safety and radiation protection to protect people and the environment. Its President, Bernard Doroszczuk, reminded us in January[1] that its mission, as defined by law, does not take into account issues other than those of safety: “the assessment of the situation is carried out in a collegial manner and may lead us to take a decision relating to safety, […] each time proportionate to the safety issues at stake.

The ASN is recognized nationally and internationally for its rigor and impartiality. We may recall its decision of June 2020 concerning reworking the welds on the Flamanville EPR, which resulted in an additional delay and a significant additional cost for EDF. In addition, the ASN is subject to international “Integrated Regulatory Review Service” (IRRS) audits by its peers.

IRSN, on the other hand, is an EPIC (public establishment of an industrial or commercial nature). It is placed under the joint supervision of several ministries. It provides technical support to the ASN[2]: “the ASN relies on external technical expertise, in particular, that of the Institute for Radiation Protection and Nuclear Safety (IRSN).” Imagining that the IRSN would be a counterweight to the ASN is not only a misunderstanding but could also be the source of severe dysfunctions in safety governance.

2- The duality of expertise and decision: an organizational method rather than an absolute principle

The quality and effectiveness of control depend on the proper articulation of expertise and authority. It seems legitimate to reexamine the balance of the system at regular intervals and to examine the adjustments to be made. This is undoubtedly all the more justified since the organizational chains have become longer and more complex.

Recently, some people have spoken of a risk of “weakening nuclear expertise” because experts would no longer be required to concentrate on studying the risks, regardless of the decision to be taken.

However, it is incorrect to think that there is an irreducible duality between expertise on the one hand and authority on the other, the former living in a space free of contingencies, that of pure expertise, the latter having to assume the constraints of reality. This vision masks the indispensable dialogue and cooperation between the two: analysis is carried out within a given framework, it is never detached from the decision to be made, nor is it free of time. Thus, from the outset in France, control has been characterized by permanent dialogue (if necessary confrontation) between the safety authority and the operator, between technical support and the operator, and between the authority and technical support. Although there is an indispensable time for expertise and a time for decision-making, they live in a variety of spaces.

Let us remember that the current French organizational model is not universal. Other countries with highly respected safety governance have different organizations. The President of the ASN stated last January that “the situation is different abroad, where three types of configuration can be distinguished. First, safety authorities may have integrated expertise, as in the United States with the Nuclear Regulatory Commission (NRC) or Japan. The second model is based on separate entities, but the expert is placed under the authority’s supervision, for example, in Belgium (which is not the case in France). Finally, some countries have established a separation between the authority and several experts, of whom the authority is the sponsor. The authority chooses them according to their skills and may also call on experts from abroad, particularly for counter-expertise. This is the case in Great Britain and South Korea. On the other hand, whatever the model, expertise is a profession requiring appropriate organization, processes, and skills management.

Finally, based on technical dialogue and intellectual confrontation, the decantation of the proper safety decision is also based on eminently collective processes. This is the case within IRSN, within the eight permanent groups of experts placed at the ASN and coming from various scientific and technical backgrounds, and within the ASN itself, beginning with the college of commissioners.

3- R&D on nuclear safety is the result of close collaboration

The control of nuclear energy is based on a very solid R&D foundation. The primary role of R&D is to deepen knowledge of physics, calibrate numerical models, develop technical solutions, verify the characteristics of materials and evaluate their aging, etc. It also plays a fundamental role in training and developing new technologies. It also plays an essential role in training and skills development throughout the sector. This is why the major nuclear countries have, since the advent of nuclear physics, developed large public research organizations and laboratories (in France, the CEA). These must-have powerful technical means: research reactors, “hot” laboratories, critical models, etc.

Research advances through the work of a community. IRSN is a major player in this field through its teams and laboratories and, more often than not, through international cooperation. It regularly inspires and coordinates research in essential fields such as fuel behavior in accident situations, the phenomenology of severe accidents, etc. Much of IRSN’s research is conducted in CEA facilities or foreign laboratories. Numerous cooperation programs exist with the CEA, EDF, Orano, and Framatome. R&D is building a joint base of knowledge and calculation tools necessary for all technicians and experts. From the point of view of control, the essential thing is to ensure that the required research is carried out, to have access to the results, and to be able to dialogue with the experts. Moreover, the ASN must, by law, express its opinion on research topics to be studied in depth (publication of opinions) and discuss them with the various research actors.

Regarding skills, the link between expertise and R&D is undoubtedly the most crucial at IRSN in the industry. This was the main factor determining when IRSN was created, including R&D in its scope. Perhaps the situation should be examined by considering the mobility and career paths between expertise and R&D, within and outside IRSN. Maybe the most important thing is not so much that the R&D resources are internal as they can direct and stimulate programs, have access to results, cooperate with researchers, participate in specific experiments, and promote mixed careers.

More generally, it is important to ensure that expertise has access to the field: technical support must be rooted both in physics (thanks to R&D) and in the realities of installations and their operation (this is undoubtedly a way forward). These questions, which are being asked today about the IRSN, are also being asked about all engineering, especially nuclear engineering.

4- It is legitimate to ask questions about the organization of the future program

Regarding the nuclear power plants, while the fourth safety reviews are still in progress, it will be necessary to prepare for the transition beyond 50 years while examining the possibility of operating beyond 60 years. Today, there is a real consensus that the workload for inspection and expertise will increase considerably in the next few years, as will be the case for the industry. At the same time, France will launch a program to build six new EPR2 reactors, work on a small modular Nuward reactor, and support several start-ups that will develop disruptive technologies as part of the France 2030 program.

Since the Folz report in 2019, industrialists, led by EDF and Gifen, have undertaken a real overhaul of how they work to regain industrial excellence. In this context, the effectiveness of control can also be examined.

A mission conducted in 2015 by (among others) the General Inspectorate of Finance (IGF) had already suggested several avenues: prioritization of interventions according to safety issues, development of second-level control in certain areas, and greater clarity in the resources allocated to safety. Criticism has been leveled, in particular, at the time of appraisal. The Institute cannot be held solely responsible: some of these delays may come from the ASN, industry, or the layout and interfaces of the overall system. Rather than blaming one party or another, it is legitimate for everyone to work together to find ways to, in the words of the government press release, “streamline the ASN’s technical review and decision-making processes to meet the growing volume of activity associated with the revival of the nuclear industry. ■

[1] Hearing at the National Assembly on January 26, 2023 (Commission on the loss of sovereignty in energy matters)

[2] ASN 2022 report