Various NC systems are relevant within the nuclear technology. As already mentioned in the present report, various computational tools have been applied to the prediction of NC features of those systems. The complexity level of the concerned tools largely varies depending upon the simulated system and the objectives of the simulation. Simple analytical models solving approximate equations resembling the fundamental principles of the physics making reference to systems assumed in steady state conditions up to sophisticate codes including numerical solutions of partial differential equations in the three-dimensional space and in transient conditions can be distinguished.

In the following discussion, attention is focused on the application of thermal-hydraulic system codes. Under this category codes like APROS, ATHLET, CATHARE, RELAP5 and TRAC are included, all based upon the solution of a main system of six partial differential equations. Two main fields, one per each of the two phases liquid and steam are considered and coupling is available with the solution of the conduction heat transfer equations within solids interfaced with the fluid phases. A one-dimensional solution for the characteristics of the fluid is achieved in the direction of the fluid motion in time dependent conditions. It should be emphasized that more sophisticate models are also available including three­dimensional solutions and multi-field approaches in two and multiphase fluids. However the present qualification level of those sophisticated computational tools is questionable as well as their actual need in the design or in the safety applications.

The mentioned codes have been applied to the simulation of data measured in experimental facilities of different dimensions and complexity as well to the prediction of the NC performance of existing nuclear power plants and of advanced reactor concepts. The discussion is limited hereafter to the main findings from the applications of the system codes to the simulation of NC in Integral Test Facilities (ITF) that are simulators of PWR, BWR and WWER-440. Mention is made of the scaling problem and of the uncertainty expected in the predictions of NC in NPP.

Many experiments were performed in order to investigate the thermal-hydraulic response of the system in conditions similar to CAREM operational states. The influence of different parameters like vapor dome volume, hydraulic resistance and dome nitrogen pressure was studied. Perturbations in the thermal power, heat removal and pressure relief were applied [2].

As an example Figures 6 to 8 show the power, pressure and mass flow rate evolution during a heating power increase perturbation. As it can be seen a 5% power increase during 150 seconds produce pressure and mass flow rates increases of 2% and 3%. It was also observed that during this transient the temperature changes were very small.

The dynamic responses at low pressure and temperatures, and with control feedback loops were also studied.

It was observed that around the operating point self-pressurized natural circulation was very stable, even with important deviation on the relevant parameters.

255

235

4.1. Experimental results

The heat flux distribution for experimental runs corresponding to the nominal system pressure, Pn, of 2, 3, and 4 bars, and including pure vapor and different mixture of air and vapor, are presented in Fig. 2, 3, and 4 respectively [4,6].

An increase in system pressure increases local heat flux and this can be attributed to the increase in wall subcooling degree that enhances the thermal driving force for heat transfer. Moreover, higher system pressure associated with the higher inlet temperature leads to a greater number of molecular collisions helping in the diffusive transport of energy. However, in our experimental investigation, the dependency of the wall subcooling degree, either measured (Tc-Tw) or predicted from Gibbs-Dalton Law (Ts-Tw), on system pressure is such that the wall subcooling degree remains nearly the same for the same inlet air mass fraction and for the different system pressure. This implies that the vapor mass flow rate may dominate over system pressure, concerning the effect on local heat flux, for cases with air/vapor mixture (Fig. 5). The situation is rather different in pure vapor runs, that is increase in system pressure has a strong effect on enhancement of predicted, and even measured, wall subcooling degree and hence on increase of local heat flux (Fig. 6).

Experiments ran as expected. The main objective of the experiments was to assure that natural circulation begins when core is heated despite of initially non-existent density differences. On the other hand the experiments were planned to prove that natural circulation is efficient enough to remove the heat from core even though vessel level decreases continuously andflow changes into two-phase flow. Experiments 1-4 were clearly for varying the position of the leak and radial distribution of the core power. Test number 5 was known to be difficult to perform already during the planning of the experiments. The purpose of this experiment was to investigate a situation, when the valves both in depressurizing and in level control system would not open. In this case the coolant flow path is only through the leakage. The objective was to find out a certain power level, which can generate a steam flow large enough to prevent the coolant flow back to the reactor vessel through a horizontal pipeline. The power control system in PACTEL is designed to operate up to 1 MW power level, so the power below 50 kW was expected to give difficulties. Table III describes the power variation used in the experiment.

TABLE III. TIME VARIATION OF POWER IN EXPERIMENT 5

The facility for SCRAM-system experiments consists of a scram tank, a blowdown tank, electric heaters, piping, valves and measurement instrumentation [9]. The main interest focused on the scram tank. Electric heaters located below the water level generate and maintain the steam volume and a layer of saturated water in the tank. After the scram signal the energy of the steam volume is used to move the control rods into the core. In the experiments orifices simulated the response of the control rods. High design pressure leading to a high steam temperature and high temperature differences in the facility during the system operation challenges the integrity of the whole scram system.

5. CONCLUSIONS

Lappeenranta University of Technology in co-operation with VTT Energy has carried out several experiments to investigate advanced light water reactor safety systems. The experiments studied passive safety injection system with PACTEL facility, long-term cooling of the WWER-640 reactor concept also with PACTEL and new SCRAM system of the SWR — 1000 reactor concept with separate test facility. Experiments have been successful and given useful information of different safety systems proposed for innovative reactor designs. Experiments have also proved that PACTEL integral test facility is flexible to use and to modify for various possible applications.

The following test series related to natural convection are under consideration:

— boiling in large water pools

— WWER steam generator behaviour with out/with non-condensables

— two phase flows in typical reactor geometries (e. g. hot leg, downcomer)

— operational mode and effectiveness of passive safety systems.

A first feasibility study has been performed regarding a reactor with internal circulation flows. The Argentinean CAREM PWR Project has been used as a model. In Fig. 16 a comparison between the CAREM Reactor and its modelling in TOPFLOW is shown.

5. CONCLUSIONS

With the NOKO facility high quality data have been produced. They were used to study phenomena and used as a basis for comparison with code calculations.

Based on the experience gained with the NOKO facility the TOPFLOW facility will be improved and its test spectrum extended. Especially the development of two-phase instrumentation will be a major goal.

In some designs, natural circulation in the primary circuit is being used as the core heat removal mechanism under normal operation conditions including full power operation. Generally, due to low driving forces of natural circulation this mechanism is being utilized for relatively small to medium size reactors. Some examples of the reactor concepts based on natural circulation of the primary coolant over the full range of the normal power operation are given below. Most of these concepts make use of passive systems also to ensure or to back up some safety functions.

HSBWR (Japan): The HSBWR [2] is a BWR design of up to of 600 MW(e). The HSBWR adopts natural circulation of primary coolant and passive safety systems to improve the economy, maintainability, and reliability by system simplification. In this reactor concept pumped re-circulation systems are eliminated in order to reduce the number of components driven by external force. A riser of 9 m height is installed above the core in order to increase the driving force for natural circulation. The 3.7 m length of the fuel assemblies, with an active length of 3.1 m, is determined by the objective of avoiding seismic resonance between the fuel bundles and the reactor building, independent of the conditions of the ground on which it is constructed. Elimination of the steam separators results in a reduced flow resistance in the primary circuit, thereby providing an increased natural circulation flow rate.

The power density of the reactor core is typically lower in a natural circulation reactor than in a forced circulation reactor, but the lower power density allows a longer continuous operation. The short heated length of the fuel and the low power density provide good thermal-hydraulic characteristics. With the 8 x 8 type fuel assembly selected, the power density is 34.2 kW/L; the number of fuel assemblies is 708, and the equivalent core diameter is 4.65 m. The uranium enrichment of the refuelling batch, at equilibrium, is 3.6%, and the average fuel burn-up is 39 GWd/t for operation cycles of 23 months.

SBWR (USA): The simplified boiling water reactor (SBWR) [2] is a 600 MW(e) design. It is based on natural circulation, with a nominal core power output of 2000 MW(th), and incorporates a number of innovative features to achieve plant simplifications. Among others, it includes introduction of passive safety systems, instead of or as supplement to the traditional active safety systems. The core configuration consists of 732 bundles — 648 interior bundles and 84 peripheral bundles. The core power density is 41.5 kW/L.

Reactivity control is maintained by movement of control rods and by burnable poisons in the fuel. The reactor pressure vessel (RPV) is a vertical, cylindrical pressure vessel, with a removable top head, and head flanges, seals and bolting, and with venturi-shaped flow restrictors in the steam outlet nozzles. The RPV is 6 m in diameter, with a wall thickness of about 158 mm with cladding, and 24.5 m tall from the inside of the bottom head to the inside of the top head. The RPV height permits natural circulation driving forces to produce sufficient core coolant flow by increasing the internal flow path length.

The large RPV volume provides a large amount of water above the core, which translates directly into a much longer period of time before core uncovery can occur as a result of loss of feed water flow or a loss-of-coolant-accident (LOCA). This gives an extended period of time during which automatic systems or plant operators can re-establish reactor inventory control using any other systems capable of injecting water into the reactor. The large RPV volume also reduces the reactor pressurization rate that develops when the reactor is suddenly isolated from the normal heat sink, which eventually leads to actuation of the safety-relief valves.

AHWR (India): The 235 MW(e) Indian Advanced Heavy Water Reactor (AHWR) [3] is a vertical, pressure tube type, boiling light water cooled, heavy water moderated reactor. The reactor core is designed for use of thorium-based fuel, and for achieving a slightly negative void coefficient of reactivity. The AHWR incorporates several passive safety systems. These include core heat removal through natural circulation, direct injection of emergency core coolant system (ECCS) water into the core, passive systems for containment cooling and isolation, gravity driven water pool (GDWP) to facilitate core decay heat removal and containment cooling for three days without invoking any active systems or operator action.

During normal reactor operation, the full reactor power is removed by natural circulation. The necessary flow rate is achieved by locating the steam drums at a suitable height above the centre of the core, taking the advantage of reactor building height. By eliminating nuclear grade primary circulating pumps, their prime movers, associated valves, instrumentation, power supply and control system, the plant is made simpler, less expensive, and easier to maintain as compared to options involving forced circulation in the primary coolant circuit. The above factors also lead to considerable enhancement of system safety and reliability since pump-failure related transients have beeneliminated by design.

CAREM (Argentina): CAREM [22,23] is an innovative 100 MW(e) PWR reactor design with an integrated self-pressurized primary system through which the coolant circulation is achieved by natural circulation. The CAREM design incorporates several passive safety systems. The entire primary system including the core, steam generators, primary coolant and steam dome are contained inside a single pressure vessel. The strong negative temperature coefficient of reactivity enhances the self-controlling features.. The reactor is practically self — controlled and need for control rod movement is minimized. In order to keep a strong negative temperature coefficient of reactivity during the whole operational cycle, it is necessary not to utilize soluble boron for burnup compensation. Reactivity compensation for burnup is obtained with burnable poisons, i. e. gadolinium oxide dispersed in the uranium di-oxide fuel.

Water enters the core from the lower plenum. After being heated the coolant exits the core and flows up through the riser to the upper dome. In the upper part, water leaves the riser through lateral windows to the external region. Then it flows down through modular steam generators, decreasing its enthalpy by giving up heat to the water in the steam generator. Finally, the coolant exits the steam generators and flows down through the down-comer to the lower plenum, closing the circuit.

CAREM uses once-through straight tube steam generators. Twelve steam generators are arranged in an annular array inside the pressure vessel above the core. The primary side coolant flows through the inside of the tubes, and the secondary side water flows across the outside of the tubes. A shell and two tube plates form the barrier between primary and secondary circuits.

AST-500 (Russia): The 500 MW(th) reactor design is intended to generate low temperature heat for district heating and hot water supply to cities. AST-500 is a pressurized water reactor with integral layout of the primary components and natural circulation of the primary side coolant. Features of the AST-500 reactor include natural circulation of the coolant under reduced working parameters and specific features of the integral reactor, such as a built-in steam-gas pressurizer, in-reactor heat exchangers for emergency heat removal, and an external guard vessel.

V-500 SKDI (Russia): V-500 SKDI (500 MW(e)) [23] is a light water integral reactor design with natural circulation of the coolant in a vessel with a diameter less than 5 m. The core and the steam generators are contained within the steel pressure vessel. The core has 121 shroudless fuel assemblies having 18 control rod clusters. Thirty six fuel assemblies have burnable poison rods. The hot coolant moves from the core through the riser and upper shroud windows into the steam generators located in the downcomer. The coolant flows due to the difference in coolant densities in the downcomer and riser. The pressurizer is connected, by two pipelines, to the reactor pressure vessel and the water clean up system.

NHR-200 (China): NHR-200 [23] is a design for providing heat for district heating, industrial processes and seawater desalination. The reactor power is 200 MW(th). The reactor core is located at the bottom of the reactor pressure vessel (RPV). The system pressure is maintained by N2 and steam. The reactor vessel is cylindrical. The RPV is 4.8 m in diameter, 14 m in height, and 197 tons in weight. The guard vessel consists of a cylindrical portion with a diameter of 5m and an upper cone portion with maximum 7 m in diameter. The guard vessel is 15.1 m in height and 233 tons in weight.

The core is cooled by natural circulation in the range from full power operation to residual heat removal. There is a long riser on the core outlet to enhance the natural circulation capacity. The height of the riser is about 6 m. Even in case of interruption of natural circulation in the primary circuit due to a LOCA the residual heat of the core can be transmitted by steam condensed at the uncovered tube surface of the primary heat exchanger.

In many passive safety systems natural circulation is an essential part of the system. The main aim is to use density differences in a pool or in a closed loop to transfer thermal energy from a source to a heat sink without using other energies than gravitational energy. This process can take place with or without phase changes.

Passive systems with phase changes can develop considerable driving forces. The pressure differences often are in the range from 1 kPa-100 kPa. Examples are the passive containment cooling system, developed by General Electric, or the Emergency Condenser, developed by Siemens.

Passive systems without phase changes have much smaller driving forces with pressure differences in the range from 1 Pa to 1000 Pa. An example is the building condenser, developed by Siemens. The natural circulation with single-phase flow is valid both on the primary and on the secondary side, as long as the temperature differences are small. But small driving pressures are not equivalent with ineffective”. For example: heating up water from 5°C to 10°C gives a density difference of 0.3% kg/m3 (or 0.03%).

If the pool height is 5 m, then the maximum driving pressure is about 15 Pa. This corresponds to a vertical flow velocity of 0.17 m/s. Suggesting that the rising plume has a diameter of 2 m, the mass flow is about 500 kg/s and the thermal flux is about 10.5 MW.

In other passive systems natural circulation with pressure differences in the range <<1 Pa can determine whether a component can reach the right operating conditions or not. Assuming that radiolytic gas is accumulated from the original concentration 10"5 by a factor 100 in a “bull eye” of a conventional RPV level measurement, then the density difference is about 0.03 kg/m3. In a “bull eye” with a gas space of 5 cm height the driving pressure results to 0.015 Pa. Nevertheless in a “bull eye” that was properly formed this small pressure difference was sufficient to initiate a counter-current flow (a special form of natural circulation) of the enriched gas mixture against the fresh steam inside the connecting pipe. So in this “bull eye” no accumulation of radiolytic gas above the ignition concentration was possible. In another “bull eye”, which was not properly designed, the opposite was true.

So the effectiveness of systems with natural circulation is not primarily a question of high driving forces but of the proper design. Normally a simple calculation is sufficient for the designer to calculate the driving forces in a balance with flow resistances. It is not necessary to know each detail of the flow fields like degree of turbulence, exact velocity profiles in different regions of the circuit or the exact form and flow velocity of a raising plume in a pool. Rough calculations normally are sufficient to decide whether a system using natural circulation is properly designed or not.

Fundamental research seems to be necessary in all cases where components normally are tested with pure steam instead with nuclear steam, which always has a small content of radiolytic gas. To avoid handling pure hydrogen and pure oxygen in a laboratory it could be possible to generate “radiolytic gas” by an electorolytical apparatus continuously and in small quantities. Such experiments can be very instructive about the accumulation of the radiolytic gas, about the natural circulation effects in components with stagnating steam atmosphere and about the possibilities to get a passive transport of the enriched steam back to the RPV or to another line with intensive steam flow.

The opposite of natural circulation is stratification. In most cases where stratification takes place, it is neglected both by the designer and by the computer code. Experiments in the PANDA facility showed that stratification could become the essential effect in containment’s atmosphere. So a deeper knowledge about the counterparts “natural circulation” and “stratification” seem to be unavoidable to do the right calculations with mixed or with separated media in a containment or in a pool.

Y. A. Kurakov

Minatom RF, Moscow, Russian Federation

Y. G. Dragunov, A. K. Podshibiakin, N. S. Fil,

S. A. Logvinov, Y. K. Sitnik

OKB Gidropress, Podolsk, Russian Federation

V. M. Berkovich, G. S. Taranov

Atomenergoproject, Moscow, Russian Federation

Abstract. Elaboration and introduction of NPP designs with improved technical and economic parameters are defined as an important element of the National Program of nuclear power development approved by the Russian Federation Government in 1998. This Program considers the designs of WWER-1000/V-392 and WWER — 640/V-407 power units as the priority projects of the new generation NPPs with increased safety. A number of passive systems based on natural circulation phenomena are used in V-392 and V-407 designs to prevent or mitigate severe accidents. Design basis, configuration and effect of some naturally driven systems of V-392 design sited at Novovoronezh are mainly reflected in the present paper. One of the most important mean for severe accident prevention in V-392 design is so called SPOT — passive heat removal system designed to remove core decay heat in case of station blackout (including failure of all diesel generators). This system extracts the steam from the steam generator, condenses it and returns water to steam generator by natural circulation. The SPOT heat exchangers are cooled by atmospheric air coming by natural circulation through a special direct action control gates which operate passively as well. Extensive experimental investigation of the different aspects of this system operation has been carried out to validate its functioning under real plant conditions. In particular, full-scale section of air heat exchanger-condenser has been tested with natural circulation steam, condensate and air paths modeled. The environment air temperature and steam pressure condensing were varied in the wide range, and the relevant experimental results are being discussed in this paper. The effect of wind velocity and direction to the containment is also checked by the experiments.

1. INTRODUCTION

The Program of Russian Federation nuclear power development for 1998-2005 years and for the period till 2010 (approved by Russian Federation Government Resolution No. 815 dated July 21, 1998) has defined that the elaboration and implementation of new generation NPP designs with enhanced safety is the necessary factor of nuclear power extension in Russia. The new generation NPP projects shall meet up-to-date national and international requirements and envisage: (1) the probability of limiting release and serious core damage at beyond-design accidents less than 10 and 10 per reactor-year, respectively; (2) reduction of urgent evacuation area to 300-500 meters and emergency planning area to protect the population in case of beyond-design accidents to 700-3000 meters.

The safety of new NPP is provided by consistently implementing the defense-in-depth principle, based on the application of a system of barriers in the way of ionizing radiation and radioactive substance release into the environment, and also by realizing the engineering and organizational activities to protect these barriers. The National Program of nuclear power development considers the design of 1000 MW power unit with WWER-1000/V-392 reactor to be the priority project of new generation plants. This unit is so designed that radiation effect on the population and the environment is considerably below the allowable values established by the up-to-date regulatory documentation. The permit of Russian Federal Nuclear and Radiation Safety Authority (Gosatomnadzor) was granted to construct the power units with V — 407 reactor plant on the Sosnovy Bor and Kola sites and two units with V-392 reactor plant on the Novovoronezh site.

Extensive application of passive safety means, using natural physical processes, along with the traditional active systems is a specific feature of both these designs. The IAEA Conference on “The Safety of Nuclear Power: Strategies for the Future” [1] has noted that the use of passive safety features is a desirable method of achieving simplification and increasing the reliability of the performance of essential safety functions, and should be used wherever appropriate. However, the application of passive means is connected with some problems, which have to be solved by each plant designer. The passive systems have their own advantages and drawbacks in comparison with the active systems both in the area of plant safety and economics. Therefore a reasonable balance of active systems and new passive means is adopted in V-392 design to improve safety and public acceptability of nuclear energy.

One important problem related to the implementation of the passive means is that, in the most cases, sufficient operating experience of the passive systems/components under real plant conditions does not exist. Besides, the existing computer codes for transient and accident analysis are not sufficiently validated for the conditions and phenomena which are relevant to the passive system functioning (low pressure, low driving pressure and temperature heads, increased effect of non-condensable, boron transport at low velocities, and the like). As a result, the time — and money-consuming research and development works may be needed individually for each reactor concept to validate the operability of the passive safety means proposed in the design. Therefore, the extensive experimental investigations and tests have been already performed and are being planned to substantiate the design of the safety features proposed for new units with WWER-1000/V-392 and WWER-640/V-407 reactor plants.

P. K. Vijayan, A. K. Nayak, M. H. Bade,

N. Kumar, D. Saha, R. K. Sinha

Bhabha Atomic Research Centre,

Mumbai, India

Abstract. Scaling methods for both single-phase and two-phase natural circulation systems have been presented. For single-phase systems, simulation of the steady state flow can be achieved by preserving just one nondimensional parameter. For uniform diameter two-phase systems also, it is possible to simulate the steady state behaviour with just one non-dimensional parameter. Simulation of the stability behaviour requires geometric similarity in addition to the similarity of the physical parameters appearing in the governing equations. The scaling laws proposed have been tested with experimental data in case of single-phase natural circulation.

1. INTRODUCTION

Natural circulation is being increasingly employed in many innovative designs of nuclear reactor cooling systems. The basic advantage of natural circulation systems is the enhanced safety due to its passive nature. One of the basic requirements, which arise prior to the incorporation of such systems in nuclear reactors, is the assessment of their performance. Generally, in the nuclear field the assessment is carried out by validated computer codes. Code validation is usually done with data obtained from scaled test facilities. The topic of the present paper is the scaling laws used for constructing such scaled facilities for simulating single and two-phase natural circulation. Such a scaled facility in nuclear parlance is also known as an integral test facility (ITF).

Scaling laws make possible the comparison of the performance of different natural circulation systems and to extrapolate the data from small scale to prototype systems. Scaling laws for nuclear reactor systems are arrived at using the governing conservation equations. Pioneering work in the field of scaling laws for nuclear reactor systems have been carried out by Nahavandi et al. (1979), Zuber (1980) and Heisler (1982). The scaling law proposed by Zuber is also known as the power-to-volume scaling philosophy and is widely used for the construction of scaled test facilities simulating nuclear reactor systems. Power to volume scaling can be applied to both forced and natural circulation systems. However, the power-to — volume scaling philosophy has certain inherent distortions (especially in downsized components), which can suppress certain natural circulation specific phenomena like the instability (Nayak et al. (1998)). Hence it is necessary to examine the scaling laws which are specific to natural circulation. In the present paper, the reported scaling laws for single — and two-phase natural circulation loops are reviewed with respect to their use for predicting the steady state and stability behaviour. Then scaling procedures for single and two-phase systems are presented and tested against the available data on the steady state and stability performance of natural circulation loops. This exercise has shown that the steady state behaviour of single-phase loops can be simulated by a single dimensionless parameter. This is also true for uniform diameter two-phase loops. The simulation of the stability behaviour requires simulation of at least three dimensionless parameters in addition to geometric scaling. For the stability behaviour, scaling of the characteristic equation (obtained by the linear stability analysis method) appears to be the appropriate method for achieving similarity.

2. REVIEW OF SCALING LAWS

Several scaling laws have already been proposed for the design of scaled loops simulating natural circulation phenomenon. Such scaling laws are available for both single — and two — phase natural circulation loops.

G. Grotzbach, L. N. Carteciano, B. Dorr

Forschungszentrum Karlsruhe GmbH, Institut fur Reaktorsicherheit,

Germany

Abstract. An optional sump cooling concept for the European pressurized water reactor EPR was investigated at the Research Center Karlsruhe. This concept foresees to utilize single phase natural convection in water to remove the decay heat from the core melt. The natural convection was investigated by the SUCOS-2D and -3D scaled experiments. A numerical investigation and interpretation of these experiments was performed by means of the computer code FLUTAN. In this paper, the numerical investigation of SUCOS-3D is summarized. Fol­lowing the results of the former 2d experiments and the numerical analysis of both experiments, an unexpected temperature distribution is found in this 3d experiment. Basing on the experimental data it had to be postulated that one of the horizontal coolers was slightly tilted against the main flow direction. Additional numerical in­vestigations show that a slope of only one percent would explain the experimental flow field. Conclusions are also drawn on the limits of scalability and transferability of the experimental results to a reactor sump. A de­tailed transformation will only be possible by applying well validated CFD-codes and experienced code users. As the flow in the reactor sump will be turbulent and this flow is strongly three-dimensional and time — dependent, only the method of Large Eddy Simulation is considered of being an adequate tool for reliable trans­formation of the gained experience to analyses for the reactor sump at 1:1 scales.

1. INTRODUCTION

The final safety barrier after a core melt down accident is the core catcher in the reactor sump. An optional cooling concept for the European Pressurized water Reactor EPR utilizes passive safety features to remove the decay heat from the sump. After the accident, a dry distribution and stabilization of the core melt in the sump region of the reactor (see Fig. 1) is foreseen. Then cooling of the core melt begins with the water from the in-containment refueling water storage tank. Water cooled heat exchangers and condensers are present in the reactor sump re­gion in order to remove the decay heat from inside the containment. The decay heat is trans­ferred from the core melt to the sump water by evaporation, natural convection, and conduc­tion. In the first days the convection of the sump water is in two-phase conditions; about ten days after the accident, single-phase natural circulation conditions are reached.

The single phase natural convection is experimentally and numerically investigated in the Re­search Center Karlsruhe in the program SUCOS (SUmp COoling Small) (Knebel et al. 1995). The aim of this program is to obtain quantitative results to be transferred to the prototypic condition in order to make a statement on the feasibility of the single phase sump cooling. Two scaled test facilities (1:20) are applied in the program: SUCOS-2D (Fig. 2), which represents a two dimensional plane slab (580 x 275 x 235 mm) of a simplified reactor sump geometry, and SUCOS-3D (Fig. 3), which is a three dimensional scaled geometry (1298 x 580 x= 275mm) of the sump. Water was heated by a heated copper plate at the bottom of the pool simulating the core melt and cooled by horizontal and vertical heat exchangers in areas where they are pro­tected against vapor explosion consequences.

A numerical investigation of this sump cooling concept and the related model experiments is performed by using the FLUTAN code (Willerding at al. 1995). This thermal — and fluid — dynamical computer code is developed in the Research Center Karlsruhe for the numerical analysis of the passive heat decay removal in new reactor systems. It was already extensively validated and applied to analyses of model experiments for the decay heat removal in the fast breeder reactor SNR-300 (Weinberg et al. 1996). It is used here to investigate and interpret numerically the single-phase natural convection in the experiments SUCOS-2D and 3D. The aim of this numerical investigation is to confirm the feasibility of the sump cooling concept and to analyze in more detail the experiments.

The first step of this investigation consisted of the numerical simulation and interpretation of an SUCOS-2D experiment (Carteciano et al. 1999). The most important result was that SUCOS — 2D cannot be well reproduced with a two-dimensional calculation whereas three-dimensional calculations reproduce the experiment quite well. The simulations showed that significant
three-dimensional experiment specific phenomena are present in the experiments, but these 3d phenomena are of much less relevance in the reactor sump. Furthermore, the analysis of the calculation of SUCOS-2D gave information on the requirements for the modeling of geometry and boundary conditions for simulations of an experiment of SUCOS-3D which is the second and final step of the numerical investigation of the single-phase sump cooling concept. This step of analyzing the SUCOS-3D experiment in detail is documented in (Carteciano et al. 2000). Here, the most important results of this study are discussed to give at the end an out­look on the current status of methods to transfer all these results of the model experiments by CFD tools to EPR scales.