ASAMPSA2 Project

Within the European community responsible for fission reactor safety (plant operators, plant designers, Technical Safety Organizations (TSO), Safety Authorities) a need to develop best practice guidelines for the level 2 PSA methodology has been repeatedly expressed, with the aim of both fulfilling the requirements of safety authorities in an efficient way, and also promoting harmonization of practices in European countries in order to use results from level 2 PSAs with a greater confidence.

Existing guidelines, like those developed by the IAEA, propose a general stepwise procedural methodology, mainly based on US NUREG 1150 and high level requirements (for example on assessment of uncertainties). While it is clear that such a framework is necessary, comparisons of existing level 2 PSA, performed and discussed in SARNET L2 PSA work packages, have shown that the detailed criteria and methodologies of current level 2 PSAs strongly differ from each other in some respects. Currently in Europe integration of probabilistic findings or insights into the overall safety assessment of Nuclear Power Plants (NPPs) is quite differently understood and implemented.

Within this general context, the project objectives are to highlight common best practices, develop the appropriate scope and criteria for different level 2 PSA applications, and promote optimal use of the available resources. Such a common assessment framework will support a harmonized view on nuclear safety, and help formalize the role of probabilistic safety assessment.

A common assessment framework requires that some underlying issues are clearly understood and well developed. Some important issues are:

The main characteristic of this coordination action is to bring together the different stakeholders (plant operators, plant designers, TSO, Safety Authorities, PSA developers), irrespective of their role in the safety demonstration and analysis: this should promote a common definition of the different types of L2 PSA and help develop common views.

The aim of the coordination action is to build a consensus on the L2 PSA scope and on detailed methods deemed to be acceptable, according to different potential applications. In any methodology, and especially one developed from a wide range of contributing perspectives, there will be a range of outcomes that are considered acceptable. To represent this range, the project will consider a ‘limited-scope' methodology and a ‘full-scope' one, based on what is currently technically achievable in the performance of a level 2 PSA. In this respect it should be noted that what is technically achievable may not be cost effective, but for the purpose of this project is taken to represent the upper bound of what may be considered ‘reasonable'.

a limited description of the main reactor systems associated with standard data on the reactor materials and on the severe accident phenomenology and human actions reliability will lead to a simplified L2 PSA that would include some indication of the main accident sequences that contribute to the risk of atmospheric releases due to a severe accident. For example, ‘limited-scope' methods could apply to a L2 PSA performed with a limited number of top events in the event-tree and mainly dedicated to identification of accident sequences contributing to the Large Early Release Frequency (LERF). Such a L2 PSA can, nevertheless, include very detailed and complex supporting studies for the quantification of these top events. Engineering judgment may also help in the quantification of the top events of limited scope L2 PSA but the justification of this engineering judgement is considered as a key issue.

sophisticated methods that consider the full range of reactor initial states and possible accidents together with detailed physical phenomena modeling and a uncertainty analysis: as a consequence these L2 PSAs allow not only identification of the most sensible sequences with their occurrence probability and the related fission product release to the environment, but also the uncertainty range of the results, the weak points in the reactor system and operation, and also to identification of the accident phenomena which need further assessment to improve the result's relevance.  In such a wide ranging L2 PSA, quantification of sequences leading to large early release is clearly not the only objective.

In reality, most current level 2 PSAs are at an intermediate level between these two approaches, but this representation is foreseen as a pragmatic way to organize the coordination action, allowing discussion both on simple and elaborated methodologies. It should be assumed that the need for application of an advanced method is established from the results obtained by an earlier simplified study in regard to specific requirements of the national safety authorities.

Evidently, the second type of approach is time consuming and supposes a qualified dedicated team. Some applications do not warrant this level of detail and, additionally, some small stakeholders (especially utilities) cannot afford this level of commitment. Clearly, the scope should be appropriate to the application and life stage under consideration and the detailed methods should represent an acceptable balance between best practice and available resources. Level 2 PSAs results obtained using differing approaches or for differing scopes should not be directly compared.

The project will focus on the definition of the objectives and the scope of both methodologies by making the best use of the End-User needs for level 2 PSA (Plant operators, Plant designers and Safety Authorities), existing know-how and current practices.

The main objectives of the ASAMPSA2 project are :

For the ASAMPSA2 project the L2 PSA ‘End-Users' community is larger than ‘End-Users' partners of the ASAMPSA2 Consortium. This very important point is clarified in §3.1.