The research activities in the light water reactor (LWR) severe accidents domain at Karlsruhe Institute of Technology (KIT) are concentrated on the inand ex-vessel core melt behavior. The overall objective is to investigate the core melt scenarios from the beginning of core degradation to melt formation and relocation in the vessel, possible melt dispersion to the reactor cavity and to the containment, corium concrete interaction and corium coolability in the reactor cavity, and hydrogen behaviour in reactor systems. The results of the experiments contribute to a better understanding of the core melt sequences and thus improve safety of existing and, in the long-term, of future reactors by severe accident mitigation measures and by safety installations where required. This overview paper describes the experimental facilities used at KIT for severe accident research and gives an overview of the main directions and objectives of the R&D work.

In the course of a severe accident in a light water reactor, the interactions of corium with the concrete structures of the reactor cavity (Molten Corium Concrete Interactions or MCCI) may have a significant impact on the long-term integrity of the containment. The 2D behaviour of the melt pool contained in the reactor cavity under dry or top flooding conditions is considered as one of the key phenomena.The "scaling" issue is usually resolved by in a first step identifying the impact of physical mechanisms on the process and in a second step evaluating these mechanisms at scaled conditions regarding time and length. The conditions for the MCCI change with time due to the evolution of the melt's state defined by e.g., its composition, temperature and solid fraction, and due to the change of cavity contour and the decreasing decay heat.Here, simplified models are investigated with the objective to infer from laboratory-scale experiments how basic and important parameters like the temperature of the melt and the erosion depth evolve with time if transposed to reactor scale. Due to the simplifications in the models under consideration, the MCCI is analysed assuming "ideal" boundary conditions as e.g., an evolution of a cavity contour with time while retaining its geometrical shape (sphere, cylinder, etc.). Based on these idealised assumptions, generic trends for physical parameters like melt temperature, heat flux at the pool boundary surface, concrete fraction in the melt, viscosity, etc. can be deduced.Simple scaling methods are introduced and checked for consistency by comparison calculations with the MCCI MEDICIS module of the ASTEC integral code. Finally they are applied to a scaling problem under ideal and simplified initial and boundary conditions and the resulting generic trends of the physical parameters are evaluated at reactor scale. Such methods are very useful to better understand the MCCI phenomenology although more detailed MCCI codes are indispensable to simulate more complex accident sequences or to take into account complex boundary conditions. (C) 2014 Elsevier Ltd. All rights reserved.

Within the framework of large-scale MOCKA (KIT, Germany) experiments, a series of experiments have been performed to study the interaction of a simulant oxide (Al2O3, ZrO2, CaO) and metal melt (Fe) in a stratified configuration. To allow for a longer-term interaction, additional heating was provided by alternating additions of thermite and Zr metal to the melt. Since the heat generated by the thermite reaction and the exothermal oxidation reaction of Zr is mainly deposited in the oxide phase, prototypic heating of both melt phases is achieved. This allows the investigation of concrete erosion by metal melt as well as by the oxide which was not possible in all former experiments. Current tests in the MOCKA (KIT, Germany) program are focused on assessing the influence of concrete reinforcement (rebars) on the cavity erosion behaviour using a simulant oxide-iron melt in a stratified configuration. The experiments are performed in siliceous concrete crucibles with an inner diameter of 25 cm containing 12 wt.% reinforcement. In these experiments, the overall downward erosion by the metal melt was of the same order as the sideward one. In addition, the lateral erosion in the overlaid oxide melt region was about the same as in the metal melt region. The former experiments (BETA, COMET-L) and MOCKA tests on siliceous concrete without reinforcement have produced results with pronounced downward erosion by the metal phase. This pronounced downward erosion of the siliceous concrete without rebars seems to be inherent for melts containing a significant fraction of iron.

Recently performed experimental programmes at the French VULCANO and the German MOCKA and SICOPS facilities aimed at the further elucidation of various phenomena of molten core-concrete interaction (MCCI). Questions on these phenomena arose during the scientific discussion of MCCI in the last years. The large-scale MOCKA (KIT, Karlsruhe) experiments study the interaction of a simulant oxide (Al2O3, ZrO2, CaO) and metal melt (Fe) with concrete. To allow for a long-term interaction, internal heating was provided by alternating additions of alumino-thermite and Zr metal to the upper oxide layer of the stratified melt. Since the heat generated by the thermite reaction and the exothermal oxidation reaction of Zr is mainly deposited in the oxide phase, prototypic heating of both melt phases is achieved. Recent tests in the MOCKA (KIT, Germany) program are focused on assessing the influence of a typical 6 wt.% reinforcement in the concrete on the erosion behaviour. The experiments were performed in siliceous concrete crucibles with an inner diameter of 25 cm and a height of 1.3 m. In these experiments, the overall downward erosion by the metal melt was of the same order as the sideward one. In addition, the lateral erosion in the overlaid oxide melt region was about the same as in the metal melt region. Experiments with prototypic UO2-containing melts have been conducted in parallel in the VULCANO (CEA, Cadarache) and SICOPS (AREVA, Erlangen) facilities. In VULCANO a plasma arc furnace melts the oxide corium while three 1-L steel induction furnaces melt the steel. All these melts are poured in a concrete cavity where the decay heat is applied to the oxide phase via 40-kHz induction coils. VULCANO VBS tests with oxide and metal have shown an increased ablation in front of the metallic masses. More unexpected results concern a very high steel oxidation (especially in VBS-U1 with limestone-rich concrete in which almost all the steel was oxidised) and the fact that post test examinations did not show the expected horizontal steel layer. Induction heating was also used for decay heat simulation in the laboratory-scale SICOPS tests, using a "cold crucible" technique and a higher frequency of about 1.3 MHz. Tests were performed to study the influence of an additional metal phase on the 1D erosion behaviour. For these mixed melt tests, concrete specimen of 10 cm diameter, made of siliceous concrete with grain aggregates, were used. Oxide melts were either simulant or prototypic melts. Metallic melt was form by addition of steel (steel 37.1) pellets after oxidic melt has been produced. It was found that the impact of the metal phase depends on heating power and gas release rate. The analysis suggests that the mechanism of concrete attack by metal melt can be understood as a thermal destruction mechanism.

To complement the experimental data on melt pool behaviour in the vessel lower head Karlsruhe Institute of Technology (KIT) performs large-scale tests through the LIVE program. The objective of the LIVE-L6 experiment was to investigate the thermal hydraulic behavior of stratified melt pool in the reactor vessel lower head. In the test two melt layers were simulated: the lower with heat generation and the upper unheated layer. Both layers were composed of a non-eutectic melt (KNO_3-NaNO_3) as a simulant fluid. The layers were separated by a 2 mm thick copper plate which was located at 333 mm vessel height. In the homogeneously heated lower part of the melt pool several power plateaus were realized, starting from 18 kW, stepwise reduction to 5 kW and then increase of the power again to 18 kW. At each power plateau steady-state conditions were reached. The outer surface of the vessel was cooled with water. Besides the transient behavior, for which the LIVE-L6 test provides qualified data on temperature evolution in the molten pool and heat transfer from the lower melt layer to the upper part of the melt, the experiment addresses other important phenomena, such as the local distribution of heat flux, and the influence of solidification on the thermal-hydraulics of the pool, i.e. the possible existence of a mushy region and its impact on the heat transfer. In the post-test analysis crust thickness profile along the vessel wall, the crust composition and the morphology were determined. The LIVE-L6 experimental results are being used for the assessment of correlations and development and validation of mechanistic models for the description of molten pool behavior. The paper summarizes the objectives of the LIVE program and the main results obtained in the LIVE-L6 experiment.

The development of a corium pool in the lower head and its behavior is still a critical issue and is of great importance to assess the severe accident progression consequences to ensure the nuclear plant safety. Therefore, experimental efforts are a vital element of the assessment process, providing hard data and insights of the complicated multi-component, highly turbulent corium pool dynamics. It is essential to consider the whole evolution of the accident, including e.g. formation and growth of the in-core melt pool, characteristics of corium arrival in the lower head, and molten pool behavior after the debris re-melting. These phenomena have a strong impact on a potential termination of a severe accident. The general objective of the LIVE program at the Karlsruhe Institute of Technology (KIT) is to study these phenomena experimentally in large-scale 3D geometry and in supporting separate-effects tests, with emphasis on the transient behavior. The LIVE-L4 experiment was performed using a non-eutectic melt (KNO3-NaNO3) as a simulant fluid. Besides the transient behavior, for which the LIVE-L4 test provides qualified data on temperature evolution in the molten pool and crust growth rates, the experiment addresses other important phenomena, such as the local distribution of heat flux, and the influence of solidification on the thermal-hydraulics of the pool, i.e. the possible existence of a mushy region and its impact on the heat transfer. In the post-test analysis crust thickness profile along the vessel wall, the crust composition and the morphology were determined. The results of this experiment also allow a comparison with findings obtained earlier in other experimental programs. The LIVE-L4 experimental results are being used for the assessment of correlations and development and validation of mechanistic models for the description of molten pool behavior. These calculations are complemented by analyses with the CFD code CONV (thermal hydraulics of heterogeneous, viscous and heat-generating melts) which was developed at IBRAE. The CONV code was applied to simulate the LIVE-L4 test: a) assuming homogeneous heat generation in the liquid and b) accounting for wire heaters used to simulate the heat generation in the melt. Though the results of calculations demonstrate satisfactory agreement with the experimental measurements, deficiencies in the code prediction have been identified regarding e.g. the prediction of the crust thickness. The paper summarizes the objectives of the LIVE program, the main results obtained in the LIVE-L4 experiment and the results of the post-test calculations performed with the CONV code.

LIVE-L4 and LIVE-L5L experiments investigated thermal hydraulic behavior of corium pool in the RPV lower head with a 3D test vessel LIVE. The simulant material is 80%-20% KNO{sub 3}- NaNO{sub 3}. Transient and steady-state parameters such as melt temperature, heat flux distribution through the vessel wall as well as crust formation characteristics were obtained. The two tests demonstrated that transient events like melt relocation and change of decay power facilitate crust deformation and change of crust thickness. The dimensionless melt temperature and heat flux through wall during the steady state can be well described independent of power density. However the dimensionless melt temperature and heat flux are dependent on the pool height. A low pool has stronger focusing effect at the upper surface than in a high pool. (author)

The COMET-L3 experiment considers the long-term situation of corium/concrete interaction in an anticipated core melt accident of a light water reactor after the metal melt is layered beneath the oxide melt. The experimental focus is on the cavity formation in the basemat and the risk of a long-term basemat penetration by the metallic part of the melt. The experiment investigates the two-dimensional concrete erosion in a cylindrical crucible of 60cm in diameter fabricated from siliceous concrete in the first phase of the test, and the influence of surface flooding in the second phase. The initial mass of the melt was 425kg steel and 211kg oxide. Decay heating in the two-component metal and oxide melt is simulated by sustained induction heating of the metal phase that is overlaid by the oxide melt.

Benchmarking work was recently performed for the issue of molten corium concrete interaction (MCCI). A synthesis is given here. It concerns first the 2D CCI-2 test with a homogeneous pool and a limestone concrete, which was used for a blind benchmark. Secondly, the COMET-L2 and COMET-L3 2D experiments in a stratified configuration were used as a post-test (L2) and a blind-test (L3) benchmark. More details are given here for the recent benchmark considering a matrix of four reactor cases, with both a homogeneous and a stratified configuration, and with both a limestone and a siliceous concrete. A short overview is given on the different models used in the codes, and the consistency between the benchmark actions on experiments and reactor situations is discussed. Finally, the major uncertainties concerning MCCI are also pointed out.

Core melt solidification phenomena during external reactor vessel cooling is investigated in LIVE tests with different external cooling conditions and melt pouring positions. A non-eutectic simulant melt (80-20 mole% KNO(3)-NaNO(3)) is used in the LIVE tests. It is found out that when the vessel is cooled with water at the beginning of the melt pouring, the cooling is more effective than in the case of delayed water cooling condition, in which the vessel is first cooled with air and then flooded by water. The initial water cooling leads to a faster growth of crust layer, lower crust thermal conductivity and thinner crust layer than those under the delayed water cooling condition. The initial water cooling leads also to higher heat flux through the vessel wall during the steady state and shorter crust growth period in comparison with the delayed water cooling condition. The solidification of the melt is probably under supercooling condition. The pouring position near the vessel wall results in considerable asymmetric heat flux distribution at one latitude. The heat flux at the position of melt pouring is higher than the one at other locations.

Behavior of the corium pool in the lower head is still a critical issue in understanding of PWR core meltdown accidents. One of the key parameter for assessing the vessel mechanical strength is the resulting heat flux at the pool-vessel interface. A number of studies [1]–[3] have already been performed to pursue the understanding of a severe accident with core melting, its course, major critical phases and timing and the influence of these processes on the accident progression. Uncertainties in modeling these phenomena and in the application to reactor scale will undoubtedly persist. These include e.g. formation and growth of the in-core melt pool, relocation of molten material after the failure of the surrounding crust, characteristics of corium arrival in residual water in the lower head, corium stratifications in the lower head after the debris re-melting [4]. These phenomena have a strong impact on a potential termination of a severe accident. The main objective of the LIVE program [5] at FZK is to study the core melt phenomena both experimentally in large-scale 3D geometry and in supporting separate-effects tests, and analytically using CFD codes in order to provide a reasonable estimate of the remaining uncertainty band under the aspect of safety assessment. Within the LIVE experimental program several tests have been performed with water and with non-eutectic melts (mixture of KNO3 and NaNO3) as simulant fluids. The results of these experiments, performed in nearly adiabatic and in isothermal conditions, allow a direct comparison with findings obtained earlier in other experimental programs (SIMECO, ACOPO, BALI, etc.) and will be used for the assessment of the correlations derived for the molten pool behavior. The information obtained from the LIVE experiments includes heat flux distribution through the reactor pressure vessel wall in transient and steady state conditions, crust growth velocity and dependence of the crust formation on the heat flux distribution through the vessel wall. Supporting post-test analysis contributes to characterization of solidification processes of binary non-eutectic melts. Complimentary to other international programs with real corium melts, the results of the LIVE activities provide data for a better understanding of in-core corium pool behavior. The experimental results are being used for development of mechanistic models to describe the in-core molten pool behavior and their implementation in the severe accident codes like ASTEC. The paper summarizes the objectives of the LIVE program and presents the main results obtained in the LIVE experiments up to now.