Scenario Grouping and Classification Methodology for Postprocessing of Data Generated by Integrated Deterministic-Probabilistic Safety Analysis

Abstract

Integrated Deterministic-Probabilistic Safety Assessment (IDPSA) combines deterministic model of a nuclear power plant with a method for exploration of the uncertainty space. Huge amount of data is generated in the process of such exploration. It is very difficult to “manually” process and extract from such data information that can be used by a decision maker for risk-informed characterization, understanding, and eventually decision making on improvement of the system safety and performance. Such understanding requires an approach for interpretation, grouping of similar scenario evolutions, and classification of the principal characteristics of the events that contribute to the risk. In this work, we develop an approach for classification and characterization of failure domains. The method is based on scenario grouping, clustering, and application of decision trees for characterization of the influence of timing and order of events. We demonstrate how the proposed approach is used to classify scenarios that are amenable to treatment with Boolean logic in classical Probabilistic Safety Assessment (PSA) from those where timing and order of events determine process evolution and eventually violation of safety criteria. The efficiency of the approach has been verified with application to the SARNET benchmark exercise on the effectiveness of hydrogen management in the containment.

Metamodeling and On-Line Clustering for Loss-of-Flow Accident Precursors Identification in a Superconducting Magnet Cryogenic Cooling Circuit

“…Based on Spectral Clustering (SC), we classify the = + transients in clusters [28,44,74,75]. Then, we post-process each c-th cluster to extract its main features, in terms of prototypical time evolutions towards failure and of the corresponding component failures (i.e., the accident precursors) [45,46,76] whose generic element is the membership degree of the i-th transient of the database respect to the c-th cluster: the i-th transient is said to belong to the cluster with the membership exceeding a certain limit (in this paper, = 0.7).…”

“…In the second step, we employ a Spectral Clustering (SC) algorithm based on the Fuzzy C-Means (FCM) [44] to group similar scenarios together, which allows typifying the main patterns of the system evolution towards failure configurations (e.g., a LOFA). By so doing, we can reveal the "prototypes" of component failure modes (i.e., the precursors) that most likely drive the system to abnormal conditions [45,46]. Thirdly, we exploit the information collected within an On-line Supervised Spectral Clustering (OSSC) to timely associate new Energies 2021, 14, 5552 3 of 37 developing scenarios (measured during plant functioning) to the proper clusters and to identify the respective LOFA precursors [47].…”

“…• POD-based Kriging metamodels are employed for the first time to simulate-at a reduced computational cost-a large number of time-varying (possibly accidental) behavior of a machine for nuclear fusion; • Classical techniques available for the post-processing of the clusters produced by the FCM-based SC method [44][45][46] are originally tailored to identify both the LOFA occurrence times and the components failure modes; •…”

Integrated deterministic and probabilistic safety assessment of a superconducting magnet cryogenic cooling circuit for nuclear fusion applications

“…The Integrated Deterministic and Probabilistic Safety Assessment (IDPSA) framework (Kirschenbaum et al, 2009) allows such an analysis by combining (deterministic) phenomenological models of the system dynamics with (probabilistic) failure models . IDPSA techniques have been successfully applied to analyze nuclear fission systems: see, e.g., (Galushin and Kudinov, 2015) for an application to a hypothetical LOCA transient in typical French 900 MWe PWR; (Jankovsky et al, 2018) for the dynamic analysis of a sodium cooled fast reactor; and (Grishchenko et al, 2019) for the uncertainty quantification of a steam explosion scenario in a Nordic type BWR. In this work, IDPSA is employed, to analyze the response to abnormal transient conditions of the cooling system of a SC magnet for nuclear fusion applications, namely a single ITER Central Solenoid Module (CSM) in a reference (cold) test facility (Spitzer et al, 2015).…”

Postprocessing of Accidental Scenarios by Semi-Supervised Self-Organizing Maps

“…A first step could be distinguishing failed scenarios from safe scenarios, for example, by a fuzzy--means (FCM) classifier [6], a Mean-Shift Methodology (MSM) [7], or a decision tree [8]. Methods have been proposed for the identification also of PIs and Near Misses.…”