Melt Relocation Research Articles

Physical model features and validation status of three-dimensional simulation model for BWR in-vessel core degradation

After the severe accident at Fukushima Daiichi Units 1–3 in 2011, Regulatory Standard and Research Department, Secretariat of Nuclear Regulation Authority (S/NRA/R) initiated the development of a Boiling Water Reactor (BWR) in-vessel detailed core degradation model (called “Multifunction Model”) for decreasing computational uncertainties of severe accidents as a fundamental research activity. The objective of this study is to present the validation status of the Model against in-vessel core degradation phases, focusing on the description of important physical models and validation scheme. The Model focuses on in-vessel core degradation from cladding temperature rise to molten material flow to a lower part of reactor vessel, since there are a multitude of large uncertainties in simulating in-vessel core degradation. To simulate in-vessel core degradation more realistically, physical models for all the dominant phenomena, which are (a) multi-phase, multi-component and multi-velocity field, (b) three-dimensional configuration of internal components in a Reactor Pressure Vessel (RPV), (c) multiple melting points and chemical interactions, (d) candling phenomenon, (e) melt blockage on spacers, and (f) molten corium interaction with water pool in an RPV lower head, have been developed at S/NRA/R. The Model of in-vessel core degradation phenomena was assessed against the DF-4 experiment at Sandia National Laboratories (SNL) and the QUNECH-06 experiment at Karlsruhe Institute of Technology (KIT) in terms of estimating the cladding temperature escalation and hydrogen generation, and the CORA-18 experiment at KIT in terms of estimating axial- and lateral-directional motion of molten corium. Additionally, the Model of metallic melt relocation around a BWR core support plate, and molten corium breakup in the lower head was validated by the XR-2 at SNL, and the FARO L-19 and the KROTOS K-37 at Joint Research Centre (JRC) facilities. The Model demonstrated a reasonable capability to simulate the main features of in-vessel core degradation phases from a core zone to a lower head zone. In the next stage, sensitivity analyses of the Model will be carried out for future practical application from 2016 to 2017.

RELAP/SCDAPSIM/MOD3.5 analysis of KIT's QUENCH-14 experiment

The QUENCH-14 experiment was performed within the ACM series (Advanced Cladding Materials) performed by “Karlsruhe Institute of Technology” (KIT), Germany, to investigate the performance of M5® cladding material. During the experiment the peak temperatures exceeded 2000 K (the maximum temperature was estimated at 2249 K); therefore, a local melting of the cladding occurred. The experiment was terminated by reduction in the electrical power followed by water injection from the bottom of the test bundle. There was no breakaway oxidation or melt relocation. The test conditions used in the QUENCH-14 were comparable to the QUENCH-6 experiment that used Zircaloy-4. Simulations presented in the article were performed with MATPRO Zircaloy-4 properties and both QUENCH-6 and QUENCH-14 experimental conditions.

Development of numerical simulation method for melt relocation behavior in nuclear reactors: validation and applicability for actual core structures

Development of numerical simulation method for melt relocation behavior in nuclear reactors: validation and applicability for actual core structures

An integrity of reactor vessel head and ICI nozzle under in-vessel vapor explosion loads

An integrity assessment of lower head and ICI nozzle of the reactor vessel of Sinkori (SKR) #3/4 units under in-vessel vapor explosion loads has been performed. The core melt relocation parameters were chosen within the ranges of physically realizable boundaries. The premixing and explosion calculations were performed using TRACER-II code. Using the calculated explosion pressures imposed on the lower head inner wall and ICI nozzles, strain and stress were analyzed using ANSYS code. Then, the calculated strain results and the established failure criteria were used in determining the failure probability of the lower head and the ICI nozzles. Strain analyses showed that the failure of the lower head and the ICI nozzle due to the vapor explosion was not possible under the present framework of assessment.

Modeling corium jet breakup in water pool and application to ex-vessel fuel–coolant interaction analyses

In light water reactor core melt accidents, the molten fuel can be brought into contact with coolant water in the course of the melt relocation in-vessel and ex-vessel as well as in an accident mitigation action of water addition. For the last several decades, the potential risk of energetic molten fuel coolant interactions (FCIs, steam explosions) has drawn substantial attention in the safety analysis of reactor severe accidents. In this paper, an improved melt jet breakup model is presented and analyses of an energetic fuel–coolant interaction in a PWR cavity (1) partially filled (4m deep) and (2) completely filled (7m deep) with water are presented. The TRACER-II code was used in the analyses. For jet breakup model, the full dispersion equation of Kelvin–Helmholtz instability for the melt jet–vapor film–water was solved numerically and the solutions were correlated for use in the TRACER-II code. The new jet breakup model was benchmarked using FARO L28 test data. In reactor calculations the mixing calculations showed that the average melt drop size was much smaller in 4m deep pool with 3m free-fall than in 7m deep pool. The explosion calculations showed that the peak pressure at the center of mixture was ∼90MPa in 4m deep pool, ∼25MPa in 7m deep pool. It also showed that the maximum impulse at the cavity wall was found at the lower wall in both cases and it was 50kPas in 4m deep pool and 150kPas in 7m deep pool.

Numerical analysis of grid plate melting after a severe accident in a Fast-Breeder Reactor (FBR)

Fast breeder reactors (FBRs) are provided with redundant and diverse plant protection systems with a very low failure probability (< 10 − 6/reactor year), making core disruptive accident (CDA), a beyond design basis event (BDBE). Nevertheless, safety analysis is carried out even for such events with a view to mitigate their consequences by providing engineered safeguards like the in-vessel core catcher. During a CDA, a significant fraction of the hot molten fuel moves downwards and gets relocated to the lower plate of grid plate. The ability of this plate to resist or delay relocation of core melt further has been investigated by developing appropriate mathematical models and translating them into a computer code HEATRAN-1. The core melt is a time dependent volumetric heat source because of the radioactive decay of the fission products which it contains. The code solves the nonlinear heat conduction equation including phase change. The analysis reveals that if the bottom of grid plate is considered to be adiabatic, melt-through of grid plate (i.e., melting of the entire thickness of the plate) occurs between 800 s and 1000 s depending upon the initial conditions. Knowledge of this time estimate is essential for defining the initial thermal load on the core catcher plate. If heat transfer from the bottom of grid plate to the underlying sodium is taken into account, then melt-through does not take place, but the temperature of grid plate is high enough to cause creep failure.

LIVE-L4 and LIVE-L5L Experiments on Melt Pool and Crust Behavior in Lower Head of Reactor Pressure Vessel

The LIVE-L4 and LIVE-L5L experiments investigated the thermal-hydraulic behavior of the corium pool in the reactor pressure vessel lower head with the three-dimensional test vessel LIVE. The simulant material is a noneutectic binary mixture of 20% NaNO3-80% KNO3. Transient and steady-state parameters such as melt temperature and heat flux distribution through the vessel wall as well as crust formation characteristics were obtained. The two tests demonstrated that transient events like repeated melt relocation and change of decay power density facilitate crust deformation and change of crust thickness. Massive crust formation in a noneutectic melt pool leads to a change of melt pool composition and a decrease of melt-crust interface temperature. The melt temperature and heat flux at the same pool height and same power density can be roughly compared independent of heating history and initial melt pouring pattern. The dimensionless melt temperature as well as the dimensionless heat flux through the wall during the steady state are independent of power density if the pools have the same height. But, they are dependent on the pool height. For a low pool, the gradients with height of both melt temperature and heat flux through the vessel are larger than those for a high pool.

Determination of the hydrogen source term during the reflooding of an overheated core: Calculation results of the integral reflood test QUENCH-03 with PWR-type bundle

During a severe accident, one of the main accident management procedure consists of injecting water in the reactor core by means of various safety injection devices. Nevertheless, the success of a core reflood is not guaranteed because of possible negative effects: temperature escalation, enhanced hydrogen production, enhanced release of fission products, core degradation due to thermal shock, shattering, debris and melt formation. The QUENCH-03 experiment was carried out to investigate the behavior on reflooding at high temperature of LWR fuel rods with little oxidation. Posttest calculations with the ASTEC-CATHARE V2 code were made for code assessment and validation of the new reflooding model. This thermal–hydraulic model is used to detect the quench front position and to calculate the heat transfer between fuel and fluid in the transition boiling region. Comparisons between the calculational and experimental results are presented. Emphasis has been placed on clad temperature, hydrogen production and melt relocation. The effects of core state damage (initial temperature at reflooding onset) and the reflood mass flow rate on the hydrogen source term were investigated using the QUENCH-03 test as a base case. Calculations were made by varying both parameters in the input data deck. The results demonstrate (and confirm) the existence of a minimum flow rate for a successful reflood.

Experimental and calculation results of the integral reflood test QUENCH-14 with M5® cladding tubes

The QUENCH-14 experiment investigated the effect of M5® cladding material on bundle oxidation and core reflood, in comparison with tests QUENCH-06 (ISP-45) that used standard Zircaloy-4 and QUENCH-12 that used VVER E110-claddings. The PWR bundle configuration of QUENCH-14 with a single unheated rod, 20 heated rods, and four corner rods was otherwise identical to QUENCH-06. The test was conducted in principle with the same protocol as QUENCH-06, so that the effects of the change of cladding material could be observed more easily. Pre-test calculations were performed by the Paul Scherrer Institut (Switzerland) using the SCDAPSIM, SCDAP/RELAP5 and MELCOR codes. Follow-on post-test analyses were performed using SCDAP/RELAP5 and MELCOR as part of an ongoing programme of model validation and code assessment. Alternative oxidation correlations were used to examine the possible influence of the M5® cladding material on hydrogen generation, in comparison with Zircaloy-4. The experiment started with a pre-oxidation phase in steam, lasting ∼3000 s at ∼1500 K peak bundle temperature. After a further temperature increase to maximum bundle temperature of 2073 K the bundle was flooded with 2 g/s/rod water from the bottom. The peak temperature of ∼2300 K was measured on the bundle shroud, shortly after quench initiation. The electrical power was reduced to average value of 2 W/cm during the reflood phase to simulate effective decay heat level. Complete bundle cooling was reached in 300 s after reflood initiation. The development of the oxide layer growth during the test was essentially defined by measurements performed on the three Zircaloy-4 corner rods withdrawn successively from the bundle. The withdrawal of Zircaloy-4 and E110 corner rods after the test allowed a comparison of the different alloys in one test. One heated rod with M5 cladding was withdrawn after the test for a detailed analysis of oxidation degree and measurement of absorbed hydrogen. Post-test examinations showed neither breakaway cladding oxidation nor noticeable melt relocation between rods. Different from the QUENCH-14 (M5®) findings, the QUENCH-12 test with the E110 claddings performed under similar conditions had resulted in intensive breakaway effect at cladding and shroud surfaces during the pre-oxidation phase and local melt relocation on reflood initiation. The hydrogen production in QUENCH-14 up to reflood was similar to QUENCH-06 and QUENCH-12 bundle tests. During reflood 6 g hydrogen were released which is similar to QUENCH-06 (4 g) but much less than during quench phase of QUENCH-12 (24 g). The reason for the different behaviour of the Zr1%Nb cladding alloys is the different oxide scale properties of the materials.

Synopsis and outcome of the QUENCH experimental program

Abstract The paper gives an overview of the main outcome of the QUENCH program launched in 1997 at the Karlsruhe Institute of Technology (KIT), formerly Karlsruhe Research Center (FZK). The research program comprises bundle experiments as well as complementary separate-effects tests. The focus of the experiments performed from 1997 to 2009 was on scenarios of severe accidents whereas that of the future test program will be on large-break loss-of-coolant accidents (LOCA) in the frame of design-basis accidents, and debris coolability, in the frame of severe accidents. The major objective of the program is to deliver experimental and analytical data to support the development and validation of quench and quench-related models as used in code systems that model severe accident progression in light water reactors. So far, 15 integral bundle QUENCH experiments with 21–31 electrically heated fuel rod simulators of 2.5 m length have been conducted. The following parameters and their influence on bundle degradation and reflood have been investigated: degree of pre-oxidation, temperature at initiation of reflood, flooding rate, influence of neutron absorber materials (B 4 C, AgInCd), air ingress, and influence of the type of cladding alloy. In six tests, reflooding of the bundle led to a temporary temperature excursion driven by runaway oxidation of zirconium alloy components and resulting in release of a significant amount of hydrogen, typically two orders of magnitude greater than in those tests with “successful” quenching in which cool-down was rapidly achieved. Considerable formation, relocation, and oxidation of melt were observed in all tests with escalation. The temperature boundary between rapid cool-down and temperature escalation was typically in the range of 2100–2200 K in the “normal” quench tests, i.e. in tests without absorber and/or steam starvation. Tests with absorber and/or steam starvation were found to lead to temperature escalations at lower temperatures. All phenomena occurring in the bundle tests have been investigated additionally in parametric and more systematic separate-effects tests. Oxidation kinetics of various cladding alloys, including advanced ones, have been determined over a wide temperature range (873–1773 K) in different atmospheres (steam, oxygen, air, and their mixtures). Hydrogen absorption by different zirconium alloys was investigated in detail, recently also using neutron radiography as non-destructive method for determination of hydrogen distribution in claddings. Furthermore, degradation mechanisms of absorber rods including B 4 C and AgInCd as well as the oxidation of the resulting low-temperature melts have been studied. Steam starvation was found to cause deterioration of the protective oxide scale by thinning and chemical reduction. The most recent topic of the QUENCH program has been investigation of the behavior of advanced cladding materials (ACM) in comparison with the classical Zircaloy-4. Although separate-effects tests have shown some differences in oxidation kinetics, the influence of the various cladding alloys on the integral bundle behavior during oxidation and reflooding was only limited.

AgInCd control rod failure in the QUENCH-13 bundle test

The QUENCH off-pile experiments performed at the Karlsruhe Research Center are to investigate the high-temperature behavior of Light Water Reactor (LWR) core materials under transient conditions and in particular the hydrogen source term resulting from the water injection into an uncovered LWR core. The typical LWR-type QUENCH test bundle, which is electrically heated, consists of 21 fuel rod simulators with a total length of approximately 2.5 m. The Zircaloy-4 rod claddings and the grid spacers are identical to those used in Pressurized Water Reactors (PWR) whereas the fuel is represented by ZrO 2 pellets. In the QUENCH-13 experiment the single unheated fuel rod simulator in the center of the test bundle was replaced by a PWR-type control rod. The QUENCH-13 experiment consisting of pre-oxidation, transient, and quench water injection at the bottom of the test section investigated the effect of an AgInCd/stainless steel/Zircaloy-4 control rod assembly on early-phase bundle degradation and on reflood behavior. Furthermore, in the frame of the EU 6th Framework Network of Excellence SARNET, release and transport of aerosols of a failed absorber rod were to be studied in QUENCH-13, which was accomplished with help of aerosol measurements performed by PSI–Switzerland and AEKI–Hungary. Control rod failure was initiated by eutectic interaction of steel cladding and Zircaloy-4 guide tube and was indicated at about 1415 K by axial peak absorber and bundle temperature responses and additionally by the on-line aerosol monitoring system. Significant releases of aerosols and melt relocation from the control rod were observed at an axial peak bundle temperature of 1650 K. At a maximum bundle temperature of 1820 K reflood from the bottom was initiated with cold water at a flooding rate of 52 g/s. There was no noticeable temperature escalation during quenching. This corresponds to the small amount of about 1 g in hydrogen production during the quench phase (compared to 42 g of H 2 during the pre-reflood phases). Posttest examinations of bundle structures revealed the presence of only little relocated AgInCd melt in the form of rivulets, mainly in the coolant channels surrounding the control rod simulator and at axial elevations between the third (0.55 m) and first spacer grids (−0.1 m). Results of QUENCH-13 on the onset of absorber rod failure are in agreement with CORA results of nine experiments each containing one or more AgInCd/stainless steel/Zircaloy-4 control rod assemblies. Bundle degradation triggered by early melt formation was, however, more pronounced in the CORA experiments with maximum bundle temperatures of ∼2300 K (compared to ∼1800 K in QUENCH-13). Consequently, QUENCH-13 allowed studying the initiation of absorber rod failure by eutectic reactions of SS-Zr, and later on of AgInCd-Zr, as well as the redistribution of the absorber material within the test bundle. Furthermore, input data for modeling of aerosol release during severe accidents are considered as benefits of the experiment.

Modifications in SCDAP code for early phase degradation in a CANDU fuel channel

The existence of horizontal fuel channels surrounded by moderator in a CANada Deuterium Uranium (CANDU) reactor type constitutes the major feature which prevents the direct application to this reactor design of the severe accident physical models developed for PWR/BWRs. During a large loss of coolant accident (LLOCA) with a coincidence of a loss of emergency core cooling (LOECC), particular deformation phenomena take place inside a CANDU fuel channel. Another peculiarity appears at melt relocation in horizontal geometry. Brief discussion of accident phenomena is included. SCDAP/RELAP5 is a widespread and detailed computer code for severe accident analysis that can be adapted to benchmark the CANDU dedicated tools, MAAP4-CANDU and ISAAC. The paper summarizes the applications to this reactor type previously performed by the authors with the standard (not modified) SCDAP/RELAP5 (RELAP/SCDAPSIM/MOD3.4). To obtain a code version able to simulate a CANDU channel during the early phase of a severe accident, two new models are developed inside the SCDAP routines structure. These models concern the metallic melt relocation inside the horizontal fuel channel and fuel bundle slumping. Major features of the original LIQSOL (Liquefaction-flow-Solidification) model in SCDAP are described, together with the assumptions based on experimental evidence that led to the development of the modified model. Also, two sub-models that form the basis of fuel bundle slumping treatment are presented: (a) in-built resizing of hydrodynamic sub-channels inside the fuel channel, and (b) heat transfer at direct contact, fuel to fuel and fuel to shroud. The effects of the new models on the overall performance of the code are tested in several different cases and against the results of the standard RELAP/SCDAPSIM/MOD3.4(bi7). The plans to validate the new models are discussed, as well as some general adaptation directions to different phases of a severe accident in CANDUs, foreseen for SCDAP.

Model for melt blockage (slug) relocation and physico-chemical interactions during core degradation under severe accident conditions

The model describing massive melt blockage (slug) relocation and physico-chemical interactions with steam and surrounding fuel rods of a bundle is developed on the base of the observations in the CORA tests. Mass exchange owing to slug oxidation and fuel rods dissolution is described by the previously developed 2D model for the molten pool oxidation. Heat fluxes in oxidising melt along with the oxidation heat effect at the melt relocation front are counterbalanced by the heat losses in the surrounding media and the fusion heat effect of the Zr claddings attacked by the melt. As a result, the slug relocation velocity is calculated from the heat flux matches at the melt propagation front (Stefan problem). A numerical module simulating the slug behaviour is developed by tight coupling of the heat and mass exchange modules. The new model demonstrates a reasonable capability to simulate the main features of the massive slug behaviour observed in the CORA-W1 test.

Thermal expansion coefficient of steels used in LWR vessels

Because of the impact that melt relocation and vessel failure have on subsequent progression and associated consequences of a light water reactor (LWR) accident, it is important to accurately predict the heat-up and relocation of materials within the reactor vessel and heat transfer to and from the reactor vessel. Accurate predictions of such heat transfer phenomena require high temperature thermal properties. However, a review of vessel and structural steel material properties in severe accident analysis codes reveals that the required high temperature material properties are extrapolated with little, if any, data above 700°C. To reduce uncertainties in predictions relying upon this extrapolated high temperature data, new thermal expansion data were obtained using pushrod dilatometry techniques for two steels used in LWR vessels: SA 533 Grade B, Class 1 (SA533B1) low alloy steel, which is used to fabricate most US LWR reactor vessels; and Type 304 stainless steel (SS304), which is used in LWR vessel piping, penetration tubes, and internal structures. This paper summarizes the new data and compares it to existing, lower temperature data in the literature.

High temperature thermal properties for metals used in LWR vessels

Because of the impact that melt relocation and vessel failure has on subsequent progression and associated consequences of a light water reactor (LWR) accident, it is important to accurately predict the heatup and relocation of materials within the reactor vessel and heat transfer to and from the reactor vessel. Accurate predictions of such heat transfer phenomena require high temperature thermal properties. However, a review of vessel and structural steel material properties in severe accident analysis codes reveals that the required high temperature material properties are extrapolated with little, if any, data above 700 °C. To reduce uncertainties in predictions relying upon this extrapolated high temperature data, INL obtained data using laser-flash thermal diffusivity techniques for two metals used in LWR vessels: SA 533 Grade B, Class 1 (SA533B1) low alloy steel, which is used to fabricate most US LWR reactor vessels; and Type 304 Stainless Steel SS304, which is used in LWR vessel piping, penetration tubes, and internal structures. This paper summarizes the new data, compares it to existing data in the literature, and provides recommended correlations for thermal properties based on this data.

Evaluation of Containment Failure Probability by Ex-Vessel Steam Explosion in Japanese LWR Plants

The containment failure probability due to ex-vessel steam explosions was evaluated for Japanese BWR and PWR model plants. A stratified Monte Carlo technique (Latin Hypercube Sampling (LHS)) was applied for the evaluation of steam explosion loads, in which a steam explosion simulation code JASMINE was used as a physics model. The evaluation was made for three scenarios: a steam explosion in the pedestal area or in the suppression pool of a BWR model plant with a Mark-II containment, and in the reactor cavity of a PWR model plant. The scenario connecting the generation of steam explosion loads and the containment failure was assumed to be displacement of the reactor vessel and pipings, and failure at the penetration in the containment boundary. We evaluated the conditional containment failure probability (CCFP) based on the preconditions of failure of molten core retention within the reactor vessel, relocation of the core melt into the water pool without significant interference, and a strong triggering at the time of maximum premixed mass. The obtained mean and median values of the CCPF were 6.4x 10−2 (mean) and 3.9x 10−2 (median) for the BWR suppression pool case, 2.2x10−3 (mean) and 2.8x10−10 (median) for the BWR pedestal case, and 6.8X10−2 (mean) and 1.4x10−2 (median) for the PWR cavity case. The evaluation of CCFPs on the basis of core damage needs consideration of probabilities for the above-mentioned preconditions. Thus, the CCFPs per core damage should be lower than the values given above. The specific values of the probability were most dependent on the assumed range of melt flow rate and fragility curve that involved conservatism and uncertainty due to simplified scenarios and limited information.

Advanced treatment of zircaloy cladding high-temperature oxidation in severe accident code calculations: Part III. Verification against representative transient tests

The treatment of zirconium oxidation kinetics in severe accident (SA) codes has been the subject of many discussions and controversies in recent years. The main problem was the existence of several correlations which could lead to large differences in the calculated results. It appeared clearly that there was a need to converge towards a common understanding of the physical processes that must be modeled (oxygen diffusion, blanketing effect, etc.) and an agreed database among code developers and users. It would help reducing an important source of uncertainties in SA calculations. The kinetic correlation database, obtained as a result of examination of complementary experimental data in Parts I and II, is applied here to analyze a few high-temperature separate-effects tests and bundle experiments where Zry oxidation reaction played a dominant role. The ICARE/CATHARE computer code developed by IRSN is used to check the validity of the high-temperature correlations derived in Parts I and II. The physical modeling provided by the code includes detailed account of specific features of chemical interactions between fuel rod cladding and steam. In particular, high reaction rates at T > 2000 K are moderated by two effects, examined in Part I: steam blanketing during thin oxide layers growth and transition to oxidation of α-Zr(O) phase after total consumption of primary β-Zr in cladding metallic part. When applied to separate-effects tests, the evaluated parabolic correlations have shown their applicability to different types of temperature transients taking into account Zry oxidation specifics in rod geometry. The bundle integral experiments QUENCH-06 and PHEBUS B9+ did not lead to extremely large temperature excursions. Calculated temperatures, hydrogen production and oxide thickness, as well as parameters of melt relocation were found to agree well with experimentally measured values. As a result of this study, we believe that the new best-fitted correlations, obtained in agreement with available experimental data, can be used in further studies and can improve predictive power of the codes. The continuation of the current work will be the application of ICARE/CATHARE code with best-fitted Zry oxidation correlations to NPP accident scenarios.

Analysis of the degradation of a PWR control rod during a loss of coolant accident, using the ICARE2 code

Conditions leading to AIC control rod damage during a loss of coolant accident in a PWR geometry, even in absence of violation of the LOCA licensing criteria, are investigated using several versions of the ICARE2 code (IPSN). Before being applied to the reactor case, the code and the modelling procedure are validated against the out-of-pile severe fuel damage experiment CORA-5. Three particular initial configurations are considered for the subsequent control rod damage analysis: nominal control rod and guide tube geometry, zircaloy guide tube bowing with concurrent cladding thickness reduction and finally control rod cladding perforation. For each of these cases the thermal, mechanical and chemical behaviour is presented. Phenomena such as ballooning and cladding failure of fuel rods, guide tube failure, melt relocation and final fluid channel cross-section modification are described. Finally, the conclusions of numerous sensitivity studies are discussed and some suggestions are given for possible improvements of the ICARE2 code.

In-vessel melt-water interaction caused by the failure of the crust above the core support plate under molten pool

In this work, the consequences of postulated severe accidents in the European Pressurized Water Reactor which lead to molten pool formation inside the core and melt relocation into the lower head containing residual water are investigated. The path of relocation investigated is controlled by core support plate failure. Different failure cross sections are investigated. The results show that in none of the cases was any danger qf failure of the upper head of the reactor vessel leading to any risk for the containment integrity.

Modeling of chemical interactions of fuel rod materials at high temperatures II. Investigation of downward relocation of molten materials

In Part II of the modeling of chemical interactions of fuel rod materials at high temperatures, qualitative results on the nature of Zr-rich melt oxidation and interactions with fuel rods allow further interpretation of the post-test examinations of structures (debris) formed in the CORA tests under more complicated conditions, namely during downward relocation of the melt. In this situation, the molten mass extensively oxidizes and simultaneously dissolves UO 2 pellets and ZrO 2 scales of the cladding. The analysis of these simultaneous physico-chemical processes on the basis of the kinetic oxidation/dissolution model developed in Part I of the paper, allows a new interpretation and explanation of the CORA tests results concerning relocation dynamics of the major part of the melt (slow relocation of melt in the form of massive slug rather than quick relocations of droplets and rivulets), formation of local blockages (debris) in the interrod space and accumulation of the melt in the core region in the form of molten pool.