Advanced Safety Assessment Methodologies: Extended PSA
Get Adobe Flash player

Context & Need for harmonisation of PSA

European context at the start of the project (2012)

In all European countries the Probabilistic Safety Assessment (PSA) methodology is used to confirm and enhance the safety of Nuclear Power plants (NPP) in complement to the deterministic approach. The role and importance of PSA is defined and emphasized in many national and international safety standards.

The European regulators involved in the Western European Nuclear Regulators’ Association (WENRA) have defined specific reference levels for PSA. Two of these references levels (RL) for NPP in operation are provided hereafter:

  • WENRA RL O 1.1 (Scope and content of PSA)

For each plant design, a specific PSA shall be developed for level 1 and level 2 including all modes of operation and all relevant initiating events including internal fire and flooding. Severe weather conditions and seismic events shall be addressed [1].

  • WENRA RL O 3.3 (Use of PSA)

PSA shall be used to assess the overall risk from the plant, to demonstrate that a balanced design has been achieved, and to provide confidence that there are no « cliff-edge effects »

Level 1 and level 2 PSA have been developed for internal initial events for most reactors in Europe. Internal fire, internal flooding and seismic PSA are also often developed. The external events PSAs, other than seismic, are more rarely fully developed.

In the recent period, it was recognized in most countries that a severe accident management program has to be implemented on NPPs and that robust level 2 PSAs for decision-making were needed.

The discrepancies between the status of existing L2 PSAs, within different European countries and organisations, identified during the period 2006-2008 within SARNET (European Severe Accident Network of Excellence), has conducted EC to support the ASAMPSA2 project in  FP7 (2008-2012) ( This project has identified some best-practices for the development and application of L2 PSA. This project is a significant effort of harmonisation of PSA methodology in the Euratom framework programme.

The 2011 accident at Fukushima has shown that extreme external events, with magnitude largely beyond the NPP design, can strike a NPP and make it impossible to control the plant. In the history of nuclear industry, some high amplitude external events [2] above the plant design conditions have already occurred in many countries but without off-site consequences. These events have then been investigated to reinforce the NPPs and the safety rules.

It is vital for the protection of populations and environment to be sure that all operated NPPs are adequately protected against such situations. If adverse conditions lead to a core cooling failure and to the core melt, releases of fission products have to be largely mitigated with the use of the severe accident management (SAM) provisions.

The PSA methodology is a structured approach able, in theory, to check that the NPP protections are sufficient. But what are the practices ? Are the existing PSAs methodologies practicable to identify the previously mentioned events and their consequences ? How being assured that existing and future PSA developments will contribute to appropriate decision-making on NPP safety enhancement and, at the end, protection of the public and environment?

[1] This means that these two hazards shall be included in the PSA, except if a justification is provided for not including them

[2] See for example (IAEA-TECDOC-1341, 2003)


Need for PSA harmonisation activities in Europe after  Fukushima accident

a) What should be harmonised for PSA after the Fukushima accident?

It is recognised today that the Fukushima Daïchi site protections against a realistically estimated tsunami were not sufficient. This fact can be considered as a failure in the definition of the deterministic design basis conditions of Fukushima Daïchi NPPs, but the PSA approach (if correctly implemented in complement of the deterministic design), should have led to the identification of this weakness and to a decision of site reinforcement.

Everyone having worked on risk assessment knows that the decision-making process based on risk induced by rare events is difficult and associated to a societal acceptance of risks. It is also recognized that many disasters associated to human construction were predicted but:

  • either have not been considered as frequent enough to justify reinforcement,
  • or the decision-making “process” (rarely a single stakeholder) has delayed the decision to implement the needed reinforcements.

The PSA methodology is, in theory, able to combine all components of risks (frequencies, consequences) but needs to be credible. Its relevance depends on the quality of PSAs content which covers an extremely large scope:

  • definition, characterisation and frequency of initiating events (internal events, external and internal hazards) and their combinations, including identification of “risk” sources and plant operating modes to consider
  • modeling of the accident sequences and of the NPP response (human and equipment) with, for instance a fault trees – event tree approach
  • assessment of accident consequences for each accident sequences,
  • presentation and summary of the results and their interpretation as input for the decision-making process.

Each step needs to be appropriately performed to obtain a final relevant risk assessment. For European countries, it seems that harmonisation of practices or technical exchanges will be particularly fruitful for all the four steps mentioned above but with a high focus on external hazards or in general high impacts events. For example:

  • what should be the “human risk assessment” (HRA) model in the case of a major earthquake or flooding ?
  • how to consider the containment efficiency after an earthquake in the assessment of accident consequences?

b) Link with the stress test effort conducted in countries and at European level

The Fukushima accident has lead EC and National Safety Authorities to request a public review, “stress tests” [] of all European NPPs, with the objective to assess the robustness of NPPs and to identify some possibilities of reinforcements where needed.

This review, organised by ENSREG, based on a deterministic approach, examined European NPPs resilience against events like earthquake or flooding, and the response in case of partial or total loss of the ultimate heat sink and/or loss of electrical power supply.

The review concluded that the level of robustness of the concerned plants is sufficient but for many plants, safety reinforcements have been defined or accelerated to face the possibility of beyond design events. The reinforcements include:

  • protective measures (against flooding, earthquake),
  • additional equipment (mobile equipment, hardened stationary equipment) able to control the NPP in case of beyond design event,
  • protective structures (reinforced local crisis centers, secondary control room, protective building for mobile equipment …),
  • severe accident management provisions, in particular for hydrogen management and containment venting,
  • new organisational arrangements (procedures for multi-units accidents, external interventions teams able to secure a damaged site …).

Action plans to implement these measures are now defined in all European countries.

In parallel, there is an interest to confirm with extended PSA [1], the high level of robustness of NPPs obtained after the implementation of the safety reinforcements described above. Building a meaningful risk assessment model based on both NPP and its environment is a difficult task which can be resource and time consuming, even if some guidance already exists on many topics.

Experience on the different parts of an extended PSA exists in the different European countries. The ASAMPSA_E project will provide an opportunity to examine what types of methodology have already been implemented and how efficient these are (optimisation of resources, potential for identification of NPP weakness…). This project should help the acceleration of the development of extended PSA in the different countries.

The position of the ASAMPSA_E project in the current context is represented on the graph below.

Position of the ASAMPSA_E project

Position of the ASAMPSA_E project


[1] See definition at the beginning of the text.


c) Severe accident (L2 PSA) – Possible additional PSA activities in Europe after the ASAMPSA 2 project

21 European organisations (plant operators, plant designers, Technical Safety Organisations (TSO), Safety Authorities and Research Centers) have collaborated within the EURATOM FP7 ASAMPSA2 project to the development of some European best-practice guidelines for L2 PSA developments and applications in relation with knowledge gained from research activities. These guidelines contribute to some harmonisation of practices within European Countries.

These guidelines can be applied by the European and non-European organisations. They can help in developing L2 PSA with high quality and some optimisation of the costs: many parts of the analysis are no longer topics of fundamental research activities but can be considered as developments often based on relatively well established simulations tools.

The ASAMPSA2 guidelines were discussed in March 2011 (7-9 March 2011) during a final workshop (Helsinki) with the End-Users and scientific (in particular from SARNET) community which recognised the quality of these documents and suggested some possible improvements. In addition, this workshop was an opportunity to identify some possible follow-up actions of interest. They are briefly summarised hereafter. The ASAMPSA_E project will address only some of these topics (2, 4 and 6).

Topic 1 – PSA2 methodology and tools

A need has been expressed for probabilistic tools able to incorporate deterministic models for the simulation of accident progression in an event tree and to calculate radioactive release for each accident sequence. Such tools should include Monte-Carlo simulations and offer possibilities for the dynamic reliability methods and importance analysis.

Despite previous harmonisation efforts, interface between L1 and L2 PSA remains a topic where some improvement of methodology and tools can justify further efforts.

This topic is not included in ASAMPSA_E.

Topic 2 – Extension of L2 PSA guidance for shutdown reactor states and external events

An extension of the existing ASAMPSA2 guidelines for shutdown states has been recognized as a topic of interest but the major point for guidance improvement was the extension to external hazards. In particular the significant contribution of external hazards to the global risk was discussed during the final ASAMPSA2 workshop.

This topic is included in ASAMPSA_E.

Topic 3 – Summary document on L2 PSA requirements (quality indicators)

The End-Users have expressed interest for a summary document from the ASAMPSA2 guidelines that would provide main requirements (quality indicators) to be associated to L2 PSA development and application.

This topic is not included in ASAMPSA_E.

Topic 4 – Risk metrics

The elaboration of L2 PSA risk metrics that could be applied commonly for all European plants and that could be understood easily by “non L2 PSA experts” is seen as an additional possibility for harmonisation at European level.

This topic is included in ASAMPSA_E.

Topic 5 – Research / Database on experimental data available for L2PSA needs

When developing a L2 PSA, it is required to justify many assumptions related to the phenomena on the basis of available experimental results (or use of qualified simulation tools). A need has been expressed for some specific documents explaining the existing relation between experimental results and possible assumptions in L2 PSA. This may be extremely useful and cost effective to maintain all existing knowledge, and to make easier the transfer of knowledge towards younger generation in a pragmatic and structured way.

SARNET recent findings may be a starting point for such an activity with establishing a database of available experiences for different phenomena (collected in one document, or considered as a complement in the ASAMPSA2 guideline) and formalization of structured expert judgment. Interest of a collaborative action would be to compare experience and conclusions of experts from different organisations.

Nevertheless, it is recognised that the objective cannot be the definition of generic data (like split fractions for event trees) that could be applicable to all nuclear power plants, whatever their design.

This topic is not included in ASAMPSA_E.

Topic 6 – Plant Safety: link between severe accident management options and L2 PSA results

Different severe accident management strategies have been implemented on some NPPs, even if the initial plant design is similar. International comparison between the different approaches should help experts understanding the origin of differences and to contribute to the reinforcement of the local options.

L2 PSA can be an efficient tool to support such a comparison.

Sharing experience, at European level, based on L2 PSA outcomes, can be an efficient way to discuss the efficiency of the different possible severe accident management strategies.

This topic is included in ASAMPSA_E.

Topic 7 – Training and review services by ASAMPSA2 experts

Some specific services can be organized and provided by the experts involved in the ASAMPSA2 project for reviewing existing L2 PSA or for training.

This topic is not included in ASAMPSA_E.

Events calendar

12-14 September 2016

Reviewing and discussing the ASAMPSA_E guidance reports for the development and use of extended PSA for NPPs

Location: Vienna University
More information and access reports >>

23-27 November 2015

Event: ASAMPSA_E technical meetings

Location: IRSN, Fontenay-aux-Roses


  • - Earthquake, flooding, extreme weather, lightning, biological infestation, man-made hazards, aircraft crash: event modelling and implementation in PSAs

  • - Risk metrics, PSA and DiD, screening methodologies

  • - Use of level 2 PSA for severe accident management optimisation

  • - Recent results of research for level 2 PSA.

 More events 

28 Partners

28 organisations in 18 European countries are partners in the ASAMPSA_E project.

3 additional associated partners are members of the External Experts Advisory Board (EEAB).

The project is open to collaboration with other organisations that have a broad experience in the field.

Take a look at all partners here.