®®®® SIIA Público

Título del libro: American Nuclear Society - International Topical Meeting On Probabilistic Safety Assessment And Analysis, Psa 2008
Título del capítulo: A method for estimation of cutset truncation error in psa based on cutset growth trend

Autores UNAM:
PAMELA FRAN NELSON EDELSTEIN;
Autores externos:

Idioma:

Año de publicación:
2008
Palabras clave:

Fault tree analysis; Truncation error


Resumen:

This work proposes a method for estimating core melt frequency beyond cutset truncation. The basic concept behind the proposal is to look at the number of cutsets as the primary fault tree result, regarding total probability as a derived quantity. All the information needed to estimate total probability is already contained in the distribution of the number of cutsets as a function of truncation probability. Top event probability is the integral of the cutset probability weighted by the number of cutsets, and this integral can be obtained analytically for the proposed growth model. The method is applied to several fault trees of varying complexity and origin, including full-scale PSA models. Once the number of cutsets is large enough to average out risk outliers (up to a thousand), the results for all examples reproduce actual Boolean reduction within few percent for all truncation values. This unexpected agreement between fault tree results and a basic two-parameter distribution may indicate that cutset growth is dominated by simple combinatorial principles, allowing to draw conclusions from limited amounts of cutset information. The results obtained provide valuable insights for PSA specific convergence, since the method is capable of estimating what cutoff is needed to attain the 5% ASME PSA standard criteria, along with the number of cutsets and top event probability at that level. Given that the proposed model can also predict the behavior of a subset of cutsets from the PSA, several applications are also suggested to reduce potential non-conservative results.


Entidades citadas de la UNAM: