Faire avancer la sûreté nucléaire

La Recherchev2


Severe Accident Research Network (SARNET). Level 2 PSA work package: comparison of partners methods for uncertainties assessment



Email :

Mot de passe :

B. Chaumont, M. Haesendonck, S. Vidal, J. Eyink, H. Loeffler, G. Radu, V. Kopustinskas, A. Ming, S. Güntay, V. Gustavsson, I. Ivanov, J. Dienstbier, A. Bareith, E. Hollo, G. Lajtha,
OECD/NEA/CSNI Workshop on evaluation of uncertainties in relation to severe accidents and level 2 PSA, Aix-en-Provence, 7-9 novembre 2005,
Rapport DSR 95

Type de document > *Rapport/contribution à GT (papier ou CD-Rom), *Congrès/colloque

Mots clés > sûreté, EPS-2, incertitude, SARNET (projet)

Unité de recherche > IRSN/DSR/SAGR

Auteurs > CHAUMONT Bernard

Date de publication > 29/11/2005


The PSA2 work package (PSA2 WP) is a part of the Joined Programme Activity of the European Severe Accident Network (SARNET) related to level 2 PSA methodologies. The general objectives of this work package is to provide a comparison of the different methodologies used or under development for level 2 PSA application by the partners involved in the work package and to promote their harmonization. The PSA2 WP is organized into three main topics: methodologies in general, methodologies for uncertainties assessment, and dynamic reliability methods. The different tasks initially defined for these three topics are shortly described and the partners involved identified.
Attention is then paid on the methodologies used so far by the different partners to assess the uncertainties in their level 2 PSA. A review of partners approaches to assess – as far as possible - the different sources of possible uncertainties is done for the different following topics:
- uncertainties propagated from the level 1 PSA,
- uncertainties (in sense of approximation) due to the binning of the level 1 sequences in Plant Damage,
- uncertainties related to the structure of the Accident Progression Event Tree,
- uncertainties related to the probabilities of stochastic events (system failure or recovery, human actions, some physical phenomena such as ignition of hydrogen combustion or triggering of steam explosion),
- uncertainties elated to the modelling of the different physical phenomena,
- uncertainties related to the cut-off frequency used in the probabilistic quantification of the Accident Progression Event Tree;
- uncertainties related to the binning of level 2 sequences in Release Categories (variables not considered, values of eventual continuous variables).
First conclusions of the comparison are given in terms of improvement needs and then of perspectives of the work for the following period of work.