The complex phenomena occurring in a nuclear reactor necessitate accurate modelling and simulation to determine the radiological consequences of design basis accidents. The R2CA European project aims to improve simulation codes, overcome semi-empirical approaches, and re-assess the radiological consequences of accidental scenarios in light water reactors. The fuel performance code FRAPCON has been coupled with SCIANTIX, a grain-scale 0D code for the physics-based description of inert gases in oxide fuel. A limitation emerged during the coupling concerned the evaluation of the hydrostatic stress, an important physical quantity that serves as input for the SCIANTIX calculations. The hydrostatic stress is not computed by FRAPCON due to the adopted mechanical rigid pellet modelling. Then, to overcome this limitation, a surrogate model based on an artificial neural network is proposed in this work. The improved coupled system is evaluated against the Risø AN3 experiment. In the second part of the work, the focus was on the mechanistic modelling of fuel behaviour under oxidative conditions, which is crucial for accurate fuel modelling during loss-of-coolant accidents. A physics-based model describing the oxidation of UO2 has been implemented into SCIANTIX, which is then assessed against separate-effect experiments. Furthermore, a preliminary assessment is conducted against the high-temperature annealing experiment VERCORS 5.
I complessi fenomeni che si verificano in un reattore nucleare richiedono modellizzazioni e simulazioni accurate per determinare le conseguenze radiologiche degli incidenti di riferimento postulati. Il progetto europeo R2CA mira a migliorare i codici di simulazione, superando gli approcci semi-empirici e rivalutando le conseguenze radiologiche degli scenari incidentali nei reattori ad acqua leggera. Il codice di prestazione del combustibile FRAPCON è stato accoppiato con SCIANTIX, un codice 0D su scala di grano per la descrizione fisica dei gas inerti nel combustibile. Durante l’accoppiamento è emerso un limite riguardante la valutazione della tensione idrostatica, una quantità fisica importante che serve come input per i calcoli di SCIANTIX. La tensione idrostatica non viene calcolata da FRAPCON a causa dell’adozione di un modello meccanico di barretta rigida. Per superare questa limitazione, in questo lavoro viene proposto un modello surrogato basato su una rete neurale artificiale. Il nuovo sistema accoppiato viene valutato rispetto all’esperimento Risø AN3. Nella seconda parte del lavoro, l’attenzione si è concentrata sulla modellazione meccanicistica del comportamento del combustibile in condizioni ossidanti, essenziale per una modellazione accurata del combustibile durante un incidente di perdita del fluido refrigerante. Un modello che descrive l’ossidazione dell’UO2 è stato implementato in SCIANTIX. Questo è stato valutato utilizzando esperimenti a effetti separati. Inoltre, viene condotta una valutazione preliminare utilizzando l’esperimento di ricottura ad alta temperatura VERCORS 5.
Fission gas release during design basis accidents : surrogate and physics-based models for conventional and meso-scale nuclear fuel codes
PETROSILLO, GIACOMO
2022/2023
Abstract
The complex phenomena occurring in a nuclear reactor necessitate accurate modelling and simulation to determine the radiological consequences of design basis accidents. The R2CA European project aims to improve simulation codes, overcome semi-empirical approaches, and re-assess the radiological consequences of accidental scenarios in light water reactors. The fuel performance code FRAPCON has been coupled with SCIANTIX, a grain-scale 0D code for the physics-based description of inert gases in oxide fuel. A limitation emerged during the coupling concerned the evaluation of the hydrostatic stress, an important physical quantity that serves as input for the SCIANTIX calculations. The hydrostatic stress is not computed by FRAPCON due to the adopted mechanical rigid pellet modelling. Then, to overcome this limitation, a surrogate model based on an artificial neural network is proposed in this work. The improved coupled system is evaluated against the Risø AN3 experiment. In the second part of the work, the focus was on the mechanistic modelling of fuel behaviour under oxidative conditions, which is crucial for accurate fuel modelling during loss-of-coolant accidents. A physics-based model describing the oxidation of UO2 has been implemented into SCIANTIX, which is then assessed against separate-effect experiments. Furthermore, a preliminary assessment is conducted against the high-temperature annealing experiment VERCORS 5.File | Dimensione | Formato | |
---|---|---|---|
2023_07_Petrosillo_1.pdf
accessibile in internet per tutti
Descrizione: Thesis
Dimensione
1.92 MB
Formato
Adobe PDF
|
1.92 MB | Adobe PDF | Visualizza/Apri |
2023_07_Petrosillo_2.pdf
accessibile in internet per tutti
Descrizione: Executive Summary of the thesis
Dimensione
678.27 kB
Formato
Adobe PDF
|
678.27 kB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/211357