Integrated Deterministic and Probabilistic Safety Analysis (IDPSA) of dynamic systems is challenged by the need of implementing efficient methods for accidental scenarios generation (that are to be increased with respect to conventional PSA, due to the necessary consideration of failure events timing and sequencing along the scenarios) and for their post-processing for retrieving safety relevant information regarding the system behavior (that, in the context of IDPSA consists in the classification of the generated scenarios as safe, failed, Near Misses (NMs) and Prime Implicants (PIs)). The large amount of generated scenarios makes the computational cost for scenario post-processing enormous and the retrieved information difficult to interpret. To address this issue, in this paper we propose the use of an ensemble of Semi-Supervised Self Organizing Maps (SSSOM) whose outcomes are combined by a locally weighted aggregation: we resort to the Local Fusion (LF) principle for accounting the classification reliability of the different SSSOM classifiers, for the type of scenario to be classified. The strategy is applied for the post-processing of the accidental scenarios of a dynamic U-Tube Steam Generator (UTSG).
Local fusion of an ensemble of semi-supervised self organizing maps for post-processing accidental scenarios
Di Maio, Francesco;Zio, Enrico
2017-01-01
Abstract
Integrated Deterministic and Probabilistic Safety Analysis (IDPSA) of dynamic systems is challenged by the need of implementing efficient methods for accidental scenarios generation (that are to be increased with respect to conventional PSA, due to the necessary consideration of failure events timing and sequencing along the scenarios) and for their post-processing for retrieving safety relevant information regarding the system behavior (that, in the context of IDPSA consists in the classification of the generated scenarios as safe, failed, Near Misses (NMs) and Prime Implicants (PIs)). The large amount of generated scenarios makes the computational cost for scenario post-processing enormous and the retrieved information difficult to interpret. To address this issue, in this paper we propose the use of an ensemble of Semi-Supervised Self Organizing Maps (SSSOM) whose outcomes are combined by a locally weighted aggregation: we resort to the Local Fusion (LF) principle for accounting the classification reliability of the different SSSOM classifiers, for the type of scenario to be classified. The strategy is applied for the post-processing of the accidental scenarios of a dynamic U-Tube Steam Generator (UTSG).File | Dimensione | Formato | |
---|---|---|---|
2017_03_13_Local Fusion of an ensemble of SSSOM for post processing acci....pdf
accesso aperto
Dimensione
806.06 kB
Formato
Adobe PDF
|
806.06 kB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.