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

In the International Thermonuclear Experimental Reactor, plasma is magnetically confined with Superconductive Magnets (SMs) that must be maintained at cryogenic temperature of 4.5 K by one or more Superconducting Magnet Cryogenic Cooling Circuits (SMCCC). To guarantee cooling, Loss-Of-Flow Accidents (LOFAs) in SMCCC are to avoided. this work, we develop a three-step methodology for prompt detection LOFA precursors (i.e., those combinations component failures causing LOFA). First, randomly generate accident scenarios Monte Carlo sampling typical components and simulate corresponding transient system response deterministic thermal-hydraulic code. phase, also employ quick-running Proper Orthogonal Decomposition (POD)-based Kriging metamodels, adaptively trained reproduce output long-running code, decrease computational time. Second, group generated Spectral Clustering (SC) employing Fuzzy C-Means (FCM), order identify main patterns evolution towards abnormal states (e.g., Third, an On-line Supervised (OSSC) technique associate time-varying parameters measured during plant functioning prototypical groups obtained, which may highlight related (in terms failures). We apply proposed simplified model cooling circuit single module ITER Central Solenoid (CSM). The framework developed promptly detects 95% events around 80% precursors.