Les événements récents au Japon sur les centrales nucléaires de Fukushima ont montré que des accidents conduisant à la fusion du cœur peuvent survenir bine plus souvent qu’on ne l’avait supposé et que leur impact sur l’environnement et la vie publique est considérable. Pour les réacteurs actuels, un des moyens principaux pour stopper la progression de l’accident est de tenter de refroidir le plus rapidement possible les matériaux en utiliser une injection d’eau de secours. Suivant l’instant de déclenchement de l'injection d'eau dans un cœur dégradé (appelée renoyage) les zones du cœur présentent des degrés de dégradation variables. Ceci conduit à des écoulements 3D double phase dans la cuve à cause des hétérogénéités de porosité et de forme des matériaux à refroidir. La modélisation de ces écoulements est primordiale pour les études de sûreté. A l’IRSN, une partie de ces études se fait grâce au code ICARE-CATHARE. Ce code de calcul est utilisé en Europe par des entreprises nucléaires et sert à calculer l’évolution d’un accident dans un réacteur, en se concentrant sur l’état du cœur et du circuit primaire. L’objectif de cette thèse a été de développer un modèle de renoyage 3D (implanté dans ICARE-CATHARE) capable de traiter les configurations du cœur dégradé lors d'un accident grave. Le modèle proposé est caractérisé par un traitement du déséquilibre thermique entre les phases solide, liquide et gazeuse. Il inclut aussi deux équations de quantité de mouvement (une pour chacune des phases fluides). Une des améliorations faites au cours de cette thèse a été de bien distinguer les lois de transfert de chaleur pour différents régimes d’ébullition. On a ainsi proposé un modèle combinant les situations d’ébullition nucléée et d’ébullition en film. Les calculs permettent de mettre en évidence certaines caractéristiques multidimensionnelles de l’écoulement lors du renoyage, en particulier lorsqu’un fort gradient de pression est engendré dans le milieu poreux par l’écoulement de vapeur. En parallèle, l’IRSN a lancé un programme expérimental (essais PRELUDE et PEARL) dont l’objectif est de permettre la validation du modèle sur un dispositif 2D représentatif du renoyage de particules à haute température. L’analyse des résultats expérimentaux a permis de vérifier certains choix faits pour les lois physiques du modèle macroscopique. Néanmoins, la validation reste très globale puisqu’on ne dispose pas de mesures locales. La validation quantitative sur les données expérimentales a montré que le modèle fournit des résultats satisfaisants. Le modèle est capable de prédire la vitesse de progression du renoyage dans le cœur, la production du vapeur (instantanée et cumulée) et le pic de pression pour différents diamètres de particules et différents débits d’injection testés. / The TMI-2 accident and recently Fukushima accident demonstrated that the nuclear safety philosophy has to cover accident sequences involving massive core melt in order to develop reliable mitigation strategies for both, existing and advanced reactors. Although severe accidents are low likelihood and might be caused only by multiple failures, accident management is implemented for controlling their course and mitigating their consequences. In case of severe accident, the fuel rods may be severely damaged and oxidized. Finally, they collapse and form a debris bed on core support plate. Removal of decay heat from a damaged core is a challenging issue because of the difficulty for water to penetrate inside a porous medium. The reflooding (injection of water into core) may be applied only if the availability of safety injection is recovered during accident. If the injection becomes available only in the late phase of accident, water will enter a core configuration that will differ from original rodbundle geometry and will resemble to the severe damaged core observed in TMI-2. The higher temperatures and smaller hydraulic diameters in a porous medium make the coolability more difficult than for intact fuel rods under typical loss of coolant accident conditions. The modeling of this kind of hydraulic and heat transfer is a one of key objectives of this. At IRSN, part of the studies is realized using an European thermo-hydraulic computer code for severe accident analysis ICARE-CATHARE. The objective of this thesis is to develop a 3D reflood model (implemented into ICARE-CATHARE) that is able to treat different configurations of degraded core in a case of severe accident. The proposed model is characterized by treating of non-equilibrium thermal between the solid, liquid and gas phase. It includes also two momentum balance equations. The model is based on a previouslydeveloped model but is improved in order to take into account intense boiling regimes (in particular nucleate boiling). Moreover, the criteria characterizing the transition between different flow regimes were completed. Currently, the French IRSN sets up two experimental facilities, PEARL and PRELUDE. The aim is to predict the consequences of the reflooding of a severely damaged reactor core where a large part of the core has collapsed and formed a debris bed e.g. particles with characteristic length-scale: 1 to 5mm. This means the prediction of debris coolability, front propagation and steam production during the quenching after the water injection. A series of experiments performed in 2010-2012 at the PRELUDE facility has provided a large amount of new data that are summarized. On the basis of those experimental results, the thermal hydraulic features of the quench front have been analyzed and the intensity of heat transfer regimes is estimated. A three-equation model for the twophase flow in a heat-generating porous medium was validated. The quantitative validation of model with experimental results was realized and showed that the model provides satisfactory results. The model is able to predict the quench front velocity in the core, steam production (instantaneous and cumulated) as well as the pressure increase during reflood for different particle diameters and different injection liquid flows.

During the case of a hypothetical severe accident in a light water reactor, core damage may occur and molten fuel may interact with water resulting in explosive interactions. A Fuel-Coolant Interactions (FCI) consists of many complex phenomena whose characteristics determine the energetics of the interactions. The fuel melt initially undergoes fragmentation after contact with the coolant which subsequently increases the melt surface area exposed to coolant and causes rapid heat transfer. A substantial amount of research has been done to understand the phenomenology of FCI, still there are gaps to be filled in terms of the uncertainties in describing the processes such as breakup/fragmentation of melt and droplets. The objective of the present work is to substantiate the understanding in the premixing phase of the FCI process by studying the deformation/pre-fragmentation of melt droplets and also the mechanism of melt jet breakup. The focus of the work is to study the effect of various influential parameters during the premixing phase that determine the intensity of the energetics in terms of steam explosion. The study is based on numerical analysis starting from smaller scale and going to the large scale FCI. Efforts are also taken to evaluate the uncertainties in estimating the steam explosion loads on the reactor scale. The fragmented core is expected to form a porous debris bed. A part of the present work also deals with experimental investigations on the coolability of prototypical debris bed. Initially, the phenomenology of FCI and debris bed coolability is introduced. A review of the state of the art based on previous experimental and theoretical developments is also presented. The study starts with numerical investigation of molten droplet hydrodynamics in a water pool, carried out using the Volume Of Fluid (VOF) method in the CFD code ANSYS FLUENT. This fundamental study is related to single droplets in a preconditioning phase, i.e. deformation/pre-fragmentation prior to steam explosion. The droplet deformation is studied extensively also including the effect of the pressure pulse on its deformation behavior. The effect of material physical properties such as density, surface tension and viscosity are investigated. The work is then extended to 3D analysis as a part of high fidelity simulations, in order to overcome the possible limitations of 2D simulations. The investigation on FCI processes is then continued to the analysis on melt jet fragmentation in a water pool, since this is the crucial phenomenon which creates the melt-coolant pre-mixture, an initial condition for steam explosion. The calculations are carried out assuming non-boiling conditions and the properties of Wood’s metal. The jet fragmentation and breakup pattern are carefully observed at various Weber numbers. Moreover, the effect of physical and material properties such as diameter, velocity, density, surface tension and viscosity on jet breakup length, are investigated. After the fundamental studies, the work was extended to reactor scale FCI energetics. It is mainly oriented on the evaluation of uncertainties in estimating the explosion impact loads on the surrounding structures. The uncertainties include the influential parameters in the FCI process and also the code uncertainties in calculations. The FCI code MC3D is used for the simulations and the PIE (propagation of input errors) method is used for the uncertainty analysis. The last part of the work is about experimental investigations of debris coolability carried out using the POMECO-HT facility at KTH. The focus is on the determination of the effect of the bed’s prototypical characteristics on its coolability, in terms of inhomogeneity with heap like (triangular shape) bed and the radial stratified bed, and also the effect of its multi-dimensionality. For this purpose, four particle beds were constructed: two homogeneous, one with radial stratification and one with triangular shape, respectively. The effectiveness of coolability-enhanced measures such as bottom injection of water and a downcomer (used for natural circulation driven coolability, NCDC) was also investigated. The final chapter includes the summary of the whole work. / Under ett svårt haveri i en kärnkraftsreaktor kan en härdsmälta bildas och smältan växelverka på ett explosivt sätt med kylvattnet. En sådan FCI (Fuel-Coolant-Interaction) inbegriper flera fysikaliska processer vilkas förlopp bestämmer hur stor den frigjorda energin blir. Vid kontakt med vattnet fragmenteras först härdsmältan vilket i sin tur leder till att en större yta exponeras för kylvattnet och att värmeöverföringen från smältan snabbt ökar. Mycket forskning har ägnats åt att förstå vad som sker under en FCI men det finns fortfarande luckor att fylla vad beträffar t ex osäkerheter i beskrivningen av fragmentering av såväl smälta som enskilda droppar av smält material. Syftet med detta arbete är främst att underbygga en bättre förståelse av den inledande delen av en FCI genom att studera dels hur enskilda droppar av smält material deformeras och splittras och dels hur en stråle av smält material fragmenteras. Vi studerar särskilt vilka parametrar som mest påverkar den energi som frigörs vid ångexplosionen. Problemet studeras med numerisk analys med början i liten skala och sedan i full skala. Vi söker också uppskatta de laster som explosionen utsätter reaktorns komponenter för. En annan viktig fråga gäller kylbarheten hos den slaggansamling som bildas under reaktorhärden efter en FCI. Slagghögen förväntas ha en porös struktur och en del av avhandlingen redogör för experimentella försök som genomförts för att utvärdera kylbarheten i olika prototypiska slaggformationer. I avhandlingens inledning beskrivs de fysikaliska processerna under en FCI och kylningen av en slaggansamling. Det aktuella kunskapsläget på dessa områden presenteras också utgående från tidigare experimentella och teoretiska studier. Studierna i avhandlingen inleds med numerisk analys av hydrodynamiken för en enskild droppe smälta i en vattentank där VOF-metoden i CFD-programmet ANSYS FLUENT används. Denna grundläggande studie rör en enskild droppe under förstadiet till fragmentering och ångexplosion då droppen deformeras alltmer. Deformationen studeras ingående också med hänsyn tagen till inverkan av en tryckpuls. Inverkan av olika egenskaper hos materialet, som densitet, ytspänning och viskositet studeras också. Arbetet utvidgas sedan till en beskrivning i 3D för att undvika de begränsningar som finns i en 2D-simulering. Studierna av FCI utvidgas sedan till en analys av fragmentering av en stråle smälta i vatten. Detta är en kritisk del av förloppet då smälta och vatten blandas för att ge utgångstillståndet för ångexplosionen. Beräkningarna genomförs under antagande att kokning inte sker och med materialegenskaper som för Wood´s metall. Mönstret för fragmentering och uppsplittring studeras ingående för olika Weber-tal. Dessutom studeras effekten på strålens uppsplittringslängd av parametrar som diameter och hastighet för strålen samt densitet, ytspänning och viskositet hos materialet. Efter dessa grundläggande studier utvidgas arbetet till FCI-energier i reaktorskala. Här ligger tonvikten på utvärdering av osäkerheter i bestämningen av den inverkan explosionen har på omgivande konstruktioner och komponenter. Osäkerheterna inkluderar eventuell bristande noggrannhet hos såväl de viktiga parametrarna i FCI-processen som i själva beräkningarna. Den sista delen av arbetet handlar om experimentella undersökningar av slaggformationens kylbarhet som genomförts i uppställningen POMECO-HT vid avdelningen för kärnkraftsäkerhet på KTH. Vi vill bestämma effekten av formationens prototypiska egenskaper på kylbarheten. För detta ändamål konstruerades fyra olika formationer: två homogena, en med radiell variation i partikelstorlek och en med triangulär variation. Vi undersökte också hur förbättrad kylning kan uppnås genom att tillföra kylvatten underifrån respektive via ett fallrör (kylning genom naturlig cirkulation). I det avslutande kapitlet ges en sammanfattning av hela arbetet. / <p>QC 20150507</p>

Severe accidents arising from the fusion of a nuclear reactor core must be anticipated to enhance the efficiency of their mitigation. Such accidents have occurred at TMI-2 (1979) and Fukushima (2011). Following a loss of coolant accident, core heating and oxidation of the fuel cladding followed by reflooding (injection of water) may lead to the collapse of fuel rods and formation of porous debris bed in the core. Steam produced upon reflooding may activate the exothermic oxidation of Zircaloy leading to partial melting of materials. Such evolution generates zones with reduced porosity limiting coolant penetration and/or impermeable blocked zones. In this situation, the efficiency of injecting water into the core to stop the progress of degradation and prevent the reactor core melting may be significantly reduced. In this scope, IRSN launched PEARL program to investigate the thermal hydraulics of reflooding of hot debris beds surrounded by a more permeable zone simulating the presence of intact or less damaged zones in the core. The PEARL experiments were modeled and simulated using ICARE/CATHARE code to assess the evolution of a bottom reflooding of a superheated debris bed surrounded by a bypass of larger permeability. The thermal hydraulics of the quenching process has been analyzed and the effect of each of the initial conditions on the reflooding behavior was assessed. The effect of pressure was investigated and related to the entrainment of injected water at quench front level into the bypass. An analytical model was then developed to investigate thoroughly the reflooding of a superheated heterogeneous porous medium, composed of two layers of contrasting permeability and porosity, and to describe the water entrainment in the bypass. This model computes the main variables characterizing the reflooding process such as quench front velocity, water-to-steam conversion ratio, and the flow rate of water entrained in the bypass. It provides good qualitative and quantitative results for the two-phase flow redistribution as compared to experimental results. This model has several advantages. It is written in a rather general form including the Forchheimer correction terms and non-zero cross-terms in the generalized Darcy-Forchheimer momentum equation. Variations of proposed momentum equations including changes in correlations andinterfacial friction laws can be tested easily and efficiently. Comparison of the calculations against experimental results indicated that it is necessary to include an interfacial friction law to obtain good predictions. This model allows performing fast evaluations of the efficiency of cooling bycomputing the fraction of the injected flow rate that participates in cooling. Upscaling to the reactor scale is straightforward and calculations were performed to assess the impact of geometric parameters of the debris bed (particle size, porosity, dimensions) as well as thermal hydraulic conditions (temperature, pressure, injection flow rate) on the reflooding process. Thus the model is very useful to estimate the total quenching time and the maximum temperature that could be reached by the hot debris bed at large scales. This allows assessing the probability of a successful quenching of a hot debris bed formed during a hypothetical accidental scenario.

Ce travail de thèse porte sur l’étude des pertes de pression pour des écoulements monophasiques et diphasiques inertiels au travers de milieux poreux. Son objectif est d’aider à la compréhension et à la modélisation des transferts de quantité de mouvement à l’intérieur de lits de particules, en lien avec la problématique de la gestion d’un accident grave dans un réacteur nucléaire. En effet, lors d’un tel accident, la dégradation du coeur du réacteur peut amener celui-ci à s’effondrer pour former un lit de débris, que l’on peut assimiler à un milieu poreux à haute température et dégageant de la chaleur. Ce travail de thèse s’inscrit dans un projet de recherche en sûreté nucléaire visant à prédire la refroidissabilité d’un lit de débris par injection d’eau, ou « renoyage ». Une étude expérimentale des pertes de pression pour des écoulements monodimensionnels monophasiques et diphasiques à froid est proposée dans des situations représentatives du cas réacteur, en termes de granulométrie, de formes de particules et de vitesses d’écoulement. Les expériences réalisées apportent un complément important aux données existantes, en permettant notamment d’explorer les domaines d’écoulements diphasiques avec nombres de Reynolds liquides non nuls, tout en mesurant le taux de vide, ce qui est essentiel pour une modélisation. Des modèles prédictifs pour les pertes de pression à l’intérieur d’écoulements monophasiques et diphasiques au travers de lits de particules sont établis à partir des structures d’équations obtenues par une prise de moyenne volumique des équations de conservation locales. L’observation des écoulements monophasiques montrent que des lois de type Darcy-Forchheimer avec une correction quadratique en vitesse de filtration sont à même de prédire les pertes de pression avec une précision supérieure à 10%. Une étude numérique a montré que ce résultat est applicable pour un lit désordonné de particules peu rugueuses. L’étude des écoulements diphasiques montre qu’une structure d’équations de type Darcy-Forchheimer généralisée, incluant des termes supplémentaires pour prendre en compte les effets inertiels et les frottements interfaciaux, permet de reproduire le comportement des pertes de pression dans cette situation. Un nouveau modèle est proposé, et comparé aux données expérimentales et aux modèles utilisés dans les codes de simulation des accidents graves. / This work deals with single and two-phase flow pressure losses in porous media. The aim is to improve understanding and modeling of momentum transfer inside particle beds, in relation with nuclear safety issues concerning the reflooding of debris beds during severe nuclear accidents. Indeed, the degradation of the core during such accidents can lead to the collapse of the fuel assemblies, and to the formation of a debris bed, which can be described as a hot porous medium. This thesis is included in a nuclear safety research project on coolability of debris beds during reflooding sequences. An experimental study of single and two-phase cold-flow pressure losses in particle beds is proposed. The geometrical characteristics of the debris and the hydrodynamic conditions are representative of the real case, in terms of granulometry, particle shapes, and flow velocities. The new data constitute an important contribution. In particular, they contain pressure losses and void fraction measurements in two-phase air-water flows with non-zero liquid Reynolds numbers, which did not exist before. Predictive models for pressure losses in single and two-phase flow through particle beds have been established from experimental data. Their structures are based on macroscopic equations obtained from the volume averaging of local conservation equations. Single-phase flow pressure losses can be described by a Darcy-Forchheimer law with a quadratic correction, in terms of filtration velocity, with a better-than-10 % precision. Numerical study of single-phase flows through porous media shows that this correlation is valid for disordered smooth particle beds. Twophase flow pressure losses are described using a generalized Darcy-Forchheimer structure, involving inertial and cross flow terms. A new model is proposed and compared to the experimental data and to the usual models used in severe accident simulation codes.

In Nordic design of boiling water reactors, a deep water pool under the reactor vessel is employed for the core melt fragmentation and the long term cooling of decay heated corium debris in case of a severe accident. To assess the effectiveness of such accident management strategy the Risk-Oriented Accident Analysis Methodology has been proposed. The present work contributes to the further development of the methodology and is focused on the issue of ex-vessel debris coolability. The height and shape of the porous debris bed are among the most important factors that determine if the debris can be cooled by natural circulation of water. The bed geometry is formed in the process of melt release, fragmentation, sedimentation and packing of the debris in the pool. Bed shape is affected by the coolant flow that induces movement of particles in the pool and after settling on top of the bed. The later one is called debris bed self-leveling phenomenon. In this study, the self-leveling was investigated experimentally and analytically. Experiments were carried out in order to collect data necessary for the development of a numerical model with an empirical closure. The self-leveling model was coupled to a model for prediction of the debris bed dryout. Such coupled code allows to calculate the time necessary to have a coolable configuration of the bed. The influence of input parameters was assessed through sensitivity analysis in order to screen out the less influential parameters. Results of the risk analysis are reported as complementary cumulative distribution functions of the conditional containment failure probability (CCFP). Sensitivity analyses identified: effective particle diameter and debris bed porosity as the parameters that provide the largest contribution to the CCFP uncertainty. It is found that the effect of the initial maximum height of the bed on the CCFP is reduced by the self-leveling. / Kokvattenreaktorer av nordisk typ har en djup vattenbassäng under reaktorkärlet som kan utnyttjas för att kyla härdsmältan och de fragmenterade härdresterna vid ett svårt reaktorhaveri. För att bedöma effektiviteten av en sådan haverihantering har man föreslagit användande av en riskorienterad metodik för haverianalysen (ROAAM, från engelska ”Risk-Oriented Accident Analysis Methodology”). Föreliggande projekt fokuserar på kylbarhet hos härdresterna utanför reaktortanken och bidrar till den pågående vidareutvecklingen av ROAAM till ROAAM+. Höjden på och formen för den porösa ansamlingen av härdrester (här också kallad partikelbädd) är bland de viktigaste faktorerna som avgör om resteffekten kan kylas bort med hjälp av naturlig cirkulation av vattnet i bassängen. Ansamlingens geometriska form skapas under hela processen från utsläpp av  härdsmältan via fragmentering och sedimentering i bassängens botten. Formen kan sedan förändras med tiden genom att partiklar rör sig och omfördelas i kylflödet. Detta fenomen kallas en självnivellerande process. I detta arbete studeras denna självnivellerande process experimentellt och analytiskt. Experimenten utfördes i en särskild experimentuppställning utformad för att att samla in data och parametrar som behövs för att simulera fenomenet och utveckla en beräkningsmodell som sluts empiriskt. Denna modell kopplades sedan till en modell för beräkning av dryout i partikelbädden. Genom denna koppling av de två beräkningsprogrammen är det är möjligt att beräkna tiden för partikelbädden att nå en kylbar konfiguration. Inverkan av variationer i modellens indata studeras med hjälp av känslighetsanalys. Härigenom identifierades de minst inflytelserika parametrarna såsom effektiv drifttid, partikeldensitet, experimentell ovisshet i de empiriska samband som används för att sluta modellen, samt omlokaliseringstid efter det att reaktorn snabbstoppats (SCRAM).  Dessa parametrar avfördes sedan från den fortsatta känslighetsanalysen. Ett artificiellt neuralt nätverk tränades för att användas i stället för den kopplade koden och möjliggöra den beräkningseffektivitet som krävs för att studera hur osäkerheter i indata förs vidare i riskanalysen. Resultaten är presenterade i form av komplementära, kumulativa fördelningsfunktioner för den betingade sannolikheten för brott på reaktorinneslutningen (CCFP, från engelska ”conditional containment failure probability”). Det visas att CCFP kan variera inom ett brett område beroende på de valda kombinationerna av frekvensfunktioner för ingångsparametrarna. Resultaten visar att effektiv partikeldiameter och hög porositet är de två parametrar som ger de största bidragen till osäkerheten i CCFP. Vi har också funnit att fenomenet självnivellering har en gynnsam inverkan på CCFP och leder till lägre utsläppsrisk. Det vore värdefullt att förfina de modeller som beskriver bildandet av den initiala partikelbädden. Detta är särskilt viktigt i de scenarier där det finns kort tid för självnivellering innan partikelbädden börjar smälta igen, dvs när man har relativt hög initial temperatur i partikelbädden och/eller hög specifik värmeeffekt. / <p>QC 20170315</p> / APRI

Les accidents graves résultant de la fusion d’un coeur de réacteur nucléaire doivent être anticipés pour améliorer l’efficacité de leur mitigation. De tels accidents sont survenus à TMI-2 (1979) et à Fukushima (2011). Suite à un accident de perte de refroidissement, l’échauffement du coeur et l’oxydation de la gaine de combustible suivie d’un renoyage (injection d’eau) peuvent entraîner l’effondrement des barres de combustible et la formation d’un lit de débris dans le coeur. La vapeur produite lors du renoyage peut activer l’oxydation exothermique du Zircaloy, entraînant la fusion partielle des matériaux. Cette évolution engendre des zones à porosité réduite limitant la pénétration de l’eau et/ou des zones imperméables. Dans cette situation, l’efficacité de l’injection d’eau dans le coeur pour arrêter la progression de la dégradation et empêcher la fusion du coeur du réacteur peut être considérablement réduite. Dans ce cadre, l’IRSN a lancé le programme PEARL visant à étudier la thermohydraulique du renoyage des lits de débris chauds entourés d’une zone plus perméable simulant la présence de zones intactes ou moins endommagées dans le coeur. Dans cette thèse, les expériences PEARL ont été modélisées et simulées avec ICARE/CATHARE pour évaluer l’évolution d’un renoyage d’un lit de débris surchauffé entouré d’un bypass de perméabilité plus grande. La thermohydraulique du processus a été analysée et l’effet de différents paramètres (géométrie, conditions aux limites) sur le comportement de renoyage a été évalué. Sous certaines conditions, l’entraînement de l’eau dans le bypass a été identifié et évalué. Un modèle analytique a été mis au point ensuite pour étudier de façon approfondie le renoyage d’un milieu poreux hétérogène surchauffé composé de deux lits de débris de perméabilité et de porosité différentes et pour décrire l’entraînement de l’eau dans le bypass. Ce modèle calcule les principales variables caractérisant le processus de renoyage, telles que la vitesse du front de trempe, le taux de conversion eau-vapeur et le débit d’eau entraîné dans le bypass.Il fournit de bons résultats qualitatifs et quantitatifs concernant la redistribution du débit d’eau par rapport aux résultats expérimentaux. Ce modèle a plusieurs avantages. Il est écrit sous une forme plutôt générale incluant les termes de correction de Forchheimer et les termes croisés non nuls dans l’équation de Darcy-Forchheimer généralisée. Les différentes options des équations de quantité de mouvement proposées, y compris les changements dans les corrélations et les lois de frottement interfacial, peuvent être testées facilement. La comparaison des calculs avec les résultats expérimentaux indique qu’il est nécessaire d’inclure une loi de frottement interfacial pour obtenir de bonnes prédictions. L’extrapolation à l’échelle du réacteur est simple et des calculs ont été effectués pour évaluer l’impact des paramètres géométriques du lit de débris (granulométrie, porosité, dimensions) ainsi que les conditions thermiques et hydrauliques (température, pression, débit d’injection). Ainsi, le modèle est très utile pour estimer le temps de trempe total et latempérature maximale qui pourraient être atteinte dans le lit de débris à grande échelle. Cela permet d’évaluer la probabilité de réussite du renoyage d’un lit de débris chauds formé lors d’un scénario accidentel hypothétique. / Severe accidents arising from the fusion of a nuclear reactor core must be anticipated to enhance the efficiency of their mitigation. Such accidents have occurred at TMI-2 (1979) and Fukushima (2011). Following a loss of coolant accident, core heating and oxidation of the fuel cladding followed by reflooding (injection of water) may lead to the collapse of fuel rods and formation of porous debris bed in the core. Steam produced upon reflooding may activate the exothermic oxidation of Zircaloy leading to partial melting of materials. Such evolution generates zones with reduced porosity limiting coolant penetration and/or impermeable blocked zones. In this situation, the efficiency of injecting water into the core to stop the progress of degradation and prevent the reactor core melting may be significantly reduced. In this scope, IRSN launched PEARL program to investigate the thermal hydraulics of reflooding of hot debris beds surrounded by a more permeable zone simulating the presence of intact or less damaged zones in the core. The PEARL experiments were modeled and simulated using ICARE/CATHARE code to assess the evolution of a bottom reflooding of a superheated debris bed surrounded by a bypass of larger permeability. The thermal hydraulics of the quenching process has been analyzed and the effect of each of the initial conditions on the reflooding behavior was assessed. The effect of pressure was investigated and related to the entrainment of injected water at quench front level into the bypass. An analytical model was then developed to investigate thoroughly the reflooding of a superheated heterogeneous porous medium, composed of two layers of contrasting permeability and porosity, and to describe the water entrainment in the bypass. This model computes the main variables characterizing the reflooding process such as quench front velocity, water-to-steam conversion ratio, and the flow rate of water entrained in the bypass. It provides good qualitative and quantitative results for the two-phase flow redistribution as compared to experimental results. This model has several advantages. It is written in a rather general form including the Forchheimer correction terms and non-zero cross-terms in the generalized Darcy-Forchheimer momentum equation. Variations of proposed momentum equations including changes in correlations andinterfacial friction laws can be tested easily and efficiently. Comparison of the calculations against experimental results indicated that it is necessary to include an interfacial friction law to obtain good predictions. This model allows performing fast evaluations of the efficiency of cooling bycomputing the fraction of the injected flow rate that participates in cooling. Upscaling to the reactor scale is straightforward and calculations were performed to assess the impact of geometric parameters of the debris bed (particle size, porosity, dimensions) as well as thermal hydraulic conditions (temperature, pressure, injection flow rate) on the reflooding process. Thus the model is very useful to estimate the total quenching time and the maximum temperature that could be reached by the hot debris bed at large scales. This allows assessing the probability of a successful quenching of a hot debris bed formed during a hypothetical accidental scenario.