Severe accident (SA) codes and their core degradation models have to deal with strongly nonlinear and discontinuous phenomena. In the application of uncertainty quantification to SA simulations, the combination of such phenomena may lead to a strong increase in the uncertainty propagated through the simulation, as well as to the chaotic behavior of the output variables. In this framework, the application of the limit surface search method of the RAVEN tool is proposed for a case where cliff-edge effects of SA phenomena determine a bifurcation of an output figure of merit. The algorithm is based on a predictive method making use of a support vector machine model, and it is applied with the aim of separating those input values that lead to different phenomenologies among the uncertainty calculations. The case study is in regard to the uncertainty analysis of the ASTEC code simulation of the QUENCH6 experimental test conducted in the framework of the International Atomic Energy Agency Coordinated Research Project I31033.

Adaptive Sampling Limit Surface Search Application to Identify Cliff-Edge Effects in Severe Accident Uncertainty Analysis

Bersano, Andrea;Mascari, Fulvio
2025-01-01

Abstract

Severe accident (SA) codes and their core degradation models have to deal with strongly nonlinear and discontinuous phenomena. In the application of uncertainty quantification to SA simulations, the combination of such phenomena may lead to a strong increase in the uncertainty propagated through the simulation, as well as to the chaotic behavior of the output variables. In this framework, the application of the limit surface search method of the RAVEN tool is proposed for a case where cliff-edge effects of SA phenomena determine a bifurcation of an output figure of merit. The algorithm is based on a predictive method making use of a support vector machine model, and it is applied with the aim of separating those input values that lead to different phenomenologies among the uncertainty calculations. The case study is in regard to the uncertainty analysis of the ASTEC code simulation of the QUENCH6 experimental test conducted in the framework of the International Atomic Energy Agency Coordinated Research Project I31033.
2025
adaptative sampling
ASTEC
Machine learning
severe accident
uncertainty quantification
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.12079/80090
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 0
social impact