Service interruption on Monday 11 July from 12:30 to 13:00: all the sites of the CCSD (HAL, EpiSciences, SciencesConf, AureHAL) will be inaccessible (network hardware connection).
Skip to Main content Skip to Navigation
Journal articles

De l’humain dans le bayésien

Abstract : Risk analysis of industrial systems has to include not only technical but also human and organisational factors. EDF (Electricité de France) and INERIS are presently developing a tool for "transverse" risk analysis that considers these three dimensions. We have chosen a Bayesian network as the basic tool for this development. A conceptual framework is proposed, based on the System-Action- Management principles developed by Paté-Cornell and Murphy. Consisting of three layers, it includes the technical level through a ”bow-tie” developed within the European ARAMIS project, human interventions - considered as major contributions to safety barriers - and organisational factors, which are studied at a global level. Simultaneously, a classification of the main sources of uncertainty is suggested, which is adapted to our approach. Specific recommendations are made, related to the global modelling approach of the industrial system and to the practical use of Bayesian networks. The final goal is to assess safety barriers in their environment, by providing probabilistic estimations that are able to help decision makers. We aim to apply this approach to a concrete industrial problematic, including interviews with key players in the plant in question.
Document type :
Journal articles
Complete list of metadata

Cited literature [14 references]  Display  Hide  Download
Contributor : Gestionnaire Civs Connect in order to contact the contributor
Submitted on : Thursday, September 6, 2018 - 11:19:52 AM
Last modification on : Friday, September 7, 2018 - 1:11:29 AM
Long-term archiving on: : Friday, December 7, 2018 - 1:48:11 PM


Publication funded by an institution


  • HAL Id : ineris-01869090, version 1



Régis Farret. De l’humain dans le bayésien. Rapport Scientifique INERIS, 2007, 2006-2007, pp.58-59. ⟨ineris-01869090⟩



Record views


Files downloads