CATHARE is a system code developed by CEA, IPSN, EDF and FRAMATOME for PWR safety analysis. It can model light water reactors or test facilities using several available modules. Two-phase flows are described using a two-fluid six-equation model and the presence of non-condensable gases can be taken into account by one to four additive transport equations. The code allows a three-dimensional modelling of the pressure vessel. Successive sets of closure laws or "revisions " are developed in an iterative methodology of improvement. The Revision 6 of the closure laws is implemented in the Version V1.5. It includes models,
for the reflooding, the film condensation in presence of non-condensable gases, the interfacial friction in the core, the flashing.

CATHARE has a modular structure. Several modules can be assembled to represent the primary and secondary circuits of any PWR or of any analytical test or system test facility. They are 0-D, 1-D, 3-D modules available. All modules can be connected to walls, or heat exchangers with a 1-D conduction calculation. Many sub-modules are available to calculate the neutronics, the fuel thermomechanics, pump characteristics, accumulators, sources, sinks, etc.

All modules use the 2-fluid model to describe steam-water flows and four non-condensable gases may be transported. The thermal and mechanical non-equilibrium are described. All kinds of two-phase flow patterns are modelled: co-current and counter-current flows are modelled with prediction of the counter-current flow limitation. Heat transfer with wall structures and with fuel rods are calculated taking into account all heat transfer processes ( natural and forced convection with liquid, with gas, sub-cooled and saturated nucleate boiling, critical heat flux, film boiling, film condensation). The interfacial heat and mass transfers describe not only the vaporization due to superheated steam and the direct condensation due to sub-cooled liquid, but also the steam condensation or liquid flashing due to meta-stable sub­cooled steam or superheated liquid.

The range of parameters is rather large: pressure from 0.1 to 16 MPa, liquid temperature from 20°C to 1800°C, fluid velocities up to supersonic conditions, duct hydraulic diameters from 0.01 to 0.75 m.

An important experimental program was carried out as a support for the development and validation of the code.

Passive pressure pulse transducers (PPPT) are small heat exchangers with the primary and secondary sides enclosed in a small housing (see FIG. 4). In this design, the primary side is outside the heat exchanger tubes and the secondary side inside. The primary side is connected to the RPV without isolating elements, and thus forms part of the RPV. The pressure on the secondary side during standby condition is more or less atmospheric. The PPPTs were tested at the emergency condenser test facility at Julich in five different design variants.

The PPPTs function similarly to the emergency condensers: as long as the RPV water level is higher than the reference level, there is stratification on the primary and secondary sides and negligible heat loss. When the RPV water level drops below the reference level, the water level in the PPPTs drops accordingly until steam comes in contact with the heat exchanger tubes. In principle there is natural circulation on both the primary and secondary sides. The water inside the tubes is heated up by the steam and the condensed steam returns to the RPV. When the water inside the tubes evaporates, the pressure on the secondary side increases and would reach primary-side pressure after a longer period of time. A switching device is actuated once the secondary side reaches a pressure of 6 bar (gage) in order to obtain a shorter actuation time period.

Pressure differentials of several Pa on the primary side lead to pressure increases of several bar on the secondary side. This is a good example showing that passive devices are not necessarily connected by small driving or acting forces.

On the secondary side, forced or natural circulation should be prevented in order to obtain short actuation times. Otherwise the main secondary-side water inventory must be heated until a level of 6 bar (gage) is reached. In the last of the five design variants tested, nearly stratified conditions were achieved on the secondary side and the actuation time was only about 5 s.

1.1. Experiment Research Plan

Experimental studies about passive safety characteristics for AC600/1000 ar divided the following three steps:

(1) Mechanism demonstration tests include emergency residual heat removal test through natural circulation flow of SG secondary side, core make-up tank performance test and wind tunnel test for passive containment cooling. Above three mentioned tests finished by the end of 1996.

(2) Part function demonstration tests include emergency residual heat removal test through natural circulation flow of SG secondary side and atmosphere loop and thermal hydraulic transient behavior test research during small break LOCA for core make-up tank, pressurizer, accumulator and reactor pressure vessel. Above two mentioned tests will be finished by end of this year. Meanwhile, NPIC will cooperate with Karlsruhe Technology Center of Germany to do air natural circulation flow test for AC600/1000 passive containment cooling system.

(3) Comprehensive function demonstration tests. NPIC plans to construct the following two large test facilities: a. Integral thermal hydraulic test facility. It is a tandem system of primary coolant cycle, secondary side cycle of SG and air flow cycle, which can be used to research station black out accident, small break LOCA and computer code development. b. Integral containment cooling test facility. It is used to simulate and research comprehensive characteristics of passive containment cooling system for AC600/1000.

The last series of the EC funded project included additional five SBLOCA experiments. VTT Energy and LTKK ran the experiments in August and September, 1997. Based on the experiences from the previous series, the following new experiments were performed in the third and final series of the EC funded programme:

GDE-41 3,5mm cold leg break close to DC, CMT position; increased driving force for CMT flow

GDE-42 3,5 cold leg break close to DC, additional IL flow orifice

GDE-43 1 mm cold leg break close to DC, long recirculation phase; disappearance of driving force for injection

GDE-44 3,5 mm cold leg break close to DC, cold CMT; PBL heating

GDE-45 3,5 mm cold leg break close to DC, PBL connected to pressurizer (Korean design of PSIS)

The objective of the GDE-43 experiment was to investigate the PSIS behaviour in a situation, when the break size is small and, consequently, CMT recirculation phase is long. If the CMT recirculation phase is long, the whole PSIS may become full of hot water before the CMT begins to inject water, and the driving force for injection disappears. This may have effects on the beginning of safety injection from the CMT. In the GDE-43 experiment, the break was located in the Loop 2 cold leg close to the downcomer, in the similar manner as in the GDE-24 and GDE-34 experiments. The break size was 1.0 mm in diameter, which is clearly smaller than in the two other simulation cases.

In the GDE-45 experiment the PBL connected the top of the pressurizer to the top of the CMT. The CMT sparger design effectively reduced condensation in the tank and no severe condensation occurred in the experiment. The CMT injection flow stagnated once, however, when the water flowing from the pressurizer disturbed the hot liquid layer in the CMT, steam condensed in the tank and CMT pressure dropped slightly. The CMT injection flow rate was oscillating. This was an outcome of the specific loop geometry of PACTEL. PACTEL has loop seals in both hot and cold legs of the primary circuit. The water accumulation to the pressurizer and the PSIS injection line resulted in an earlier core heat-up than in the similar experiments with PBL connection to the cold leg.

In Fig. 1 the configuration in the NOKO facility is shown; for more details about the facility and the effectiveness of the emergency condensers see [1], [2] and [3]; three tubes from the emergency condenser were used. At distances of about 1000 mm thermocouples were installed along the height of the vessel using existing flanges, see Fig. 2.

At two positions, 800 mm and 2000 mm above the return line, see Fig. 1, special grids equipped with up to 12 thermocouples were installed. The position of the thermocouples is
shown in Fig. 3; in Fig. 4 a photo of this grid is shown. The objective was the evaluation if cold or hot water plumes would develop inside the vessel.

The flooding of the reactor vessel to a level above the upper line of the emergency condensers is mandatory.

Before the start of the test the fluid in the pressure vessel was heated up to a — uniform — temperature of about 100°C while the condenser tank was at a temperature of about 30°C.

T 2.01 (2) T 2.02 (1) T 2.03(1)

T 2.04 (1) T 2.05(1)

T 2.06 (2)

T 2.07 (1) T 2.08(1)

T 2.09(1) T 2.10(1) T 2.11 (1)

T 2.12 (2)

FIG. I. The configuration in the NOKO-test FIG. 2. Pressure vessel with instrumentation, facility.

When the integrity of the RPV is lost by core melt-through the corium has to be collected in a core catcher inside the reactor cavity where it must be cooled for the long term to prevent basemat melt-through. A typical example of this concept is the EPR design [20] where the core catcher is based on a large spreading area with provisions to avoid corium/concrete reaction and to achieve long term cooling of the molten core materials. One of the main features of this spreading concept is that, after the failure of the RPV, no immediate discharge of the corium onto the spreading area should occur. One reason for this delay is to collect the corium. The reactor cavity has to be designed to withstand the thermal and mechanical loads from the RPV failure. Therefore it has to be dry to prevent a steam explosion. Thus it should contain sufficient sacrificial material to decrease the temperature of the corium and ensure its spreadability at lower temperatures. The sacrificial material shall oxidize the remaining Zr without release of non-condensable gases (or at least with very limited amount) and bind non­volatile fission products. This delayed corium discharge will be achieved in a passive way by melt-through of a gate between the reactor cavity and the spreading area. After spreading, the corium is cooled from below by special cooling elements (similar to that used as plate condensers described above) embedded in concrete. This basemat cooling is achieved first by passive and later on by active measures. During the grace period (within 12 h after beginning of the severe accident) the spread corium is flooded from the top due to the melting of special plug-in fuses in the connection lines to the IRWST. The water evaporates at the corium surface and the condensate flows back to the IRWST. By spraying from the dome, fission products in the containment atmosphere are partly washed out.

An optional cooling concept for the corium in a spreading area utilises a spreading compartment separated by a concrete structure in a lower (spreading area) and an upper cavity [21]. Both cavities are connected by vertical channels (a riser and a downcomer). This system is flooded with the water level approximately in the middle between top and bottom of the upper cavity. Condensers at the top of the upper cavity an heat exchangers submerged in the water at the bottom of the upper cavity and in the downcomer drive a single phase NC flow for corium cooling. The cooling water flow is driven by pumps.

(a) System codes have reached a highly developed modelling status and a large acceptance. They can reproduce accurately enough most existing safety related experiments, so far as the dominant physical mechanisms are known and understood;

(b) System thermal-hydraulic codes have been successfully applied for normal operation and accident conditions in the licensing of existing natural circulation BWRs. However, there is a consensus that current system thermal-hydraulic codes typically being used for safety analyses are not sufficiently validated to address all of the relevant conditions and phenomena of natural circulation based innovative designs (low pressure, low driving heads, increased effect of non-condensable gases, effect of buoyancy at low velocities, etc);

(c) Different codes are used for design and to study the different phenomena. The capabilities of the different codes to represent each case should be evaluated. More robustness is needed in the codes with regard to effectively addressing natural circulation phenomena;

(d) Consideration should be given to improvements in the code validation matrix with regard to natural circulation phenomena;

(e) Limitations exist whenever natural circulation phenomena are predominantly of a higher dimensional nature. 1D codes can be improved in this case by introducing special, component-related models that do not change the 1D structure (like reported for the core makeup tank modelling in APROS) or by introducing more-dimensional components (e. g. 2D downcomer). For development of more optimised designs, enhanced confidence in safety analyses, etc. There is a need for more accuracy of the codes and more detailed descriptions of new cooling concepts. This leads to a tendency to use, at least for local detailed analyses, 3D and time-dependent CFD codes. CFD codes are used even for designing experiments and their instrumentation arrangement. They can also be reliably used to study in a more qualitative manner the relevance of certain phenomena in flow and heat transfer problems, so far as the governing physics is included in the equations or models. Work is needed to improve CFD codes by appropriate modelling and related experiments, e. g. for the phenomena of non­condensable gases and instabilities as related to natural circulation;

(f) For more quantitative use in natural convection, especially with time dependent flow, improvements of the turbulence models are needed. Sometimes, LES is a necessary alternative. With two-phase flows, a quantitative use of CFD is currently limited to mainly homogeneous flows; for other flow regimes, the modelling of interfacial phenomena needs improvements. The turbulence modelling in two-phase flows can only be accepted as a first step. New correlations and models for two-phase flow should be developed, especially, in the turbulent flow regime. R&D is necessary to improve the models and to extend them for all flow regimes;

(g) Interfacial phenomena are still open to investigation. The interfacial heat transfer, interfacial mass transfer, and interfacial shear stress should be modelled and incorporated into the thermo-hydraulic codes;

(h) Theoretical and semi-theoretical studies could be considered international co-operative activities. This co-operation will provide better understanding of nuclear power technology and probably will reduce public reaction against nuclear technology.

F. D’Auria

University of Pisa, Italy M. Frogheri

University of Genoa, Italy

Abstract

The present paper deals with the Natural Circulation (NC) phenomenon in Pressurized Water nuclear Reactors (PWR). In the first part, data gathered from relevant experiments in PWR simulators are considered. These allowed the establishment of a flow map that has been used for evaluating the NC performance of various reactor concepts. In the second part, a theoretical study has been completed to assess the power removal capability by NC from the core of a PWR having the current geometric configuration. Taking as reference a PWR equipped with U-tubes steam generators, two-phase conditions occur in the core at power levels less than 20% nominal power. Therefore, for core power larger than this value the reactor cannot be classified any more as a PWR. The study shows that from a thermohydraulic point of view, the core can operate at power levels close to the current nominal value without experiencing thermal crisis. Limited consideration has been given to the neutronic design of the core.

1. INTRODUCTION

Natural Circulation (NC) is an important mechanism in several industrial systems and the knowledge of its behaviour is of interest to nuclear reactor design, operation and safety. In the nuclear technology, this is especially true for new reactor concepts that largely exploit the gravity forces for the heat removal capability. Natural circulation in a PWR occurs due to the presence of the heat source (core) and the heat sink constituted by the steam generators. In a gravity environment, with core located at a lower elevation than steam generators, those driving forces generate a flowrate suitable for removing nuclear fission decay power. At present, the NC core power removal capability is only exploited for accident situations, basically to demonstrate the inherent safety features of the plants.

The evaluation of the NC Performance (NCP) in experimental facilities simulating the integral system behaviour of a PWR has been the object of previous activities, e. g. Refs. [1], [2] and [3]. The NC scenarios occurring at different values of the primary system mass inventories were considered. Data have been gathered and analyzed coming from the PWR simulators Semiscale, Spes, Lobi, Bethsy, Pkl and Lstf. The thermohydraulic design of all the considered facilities has been achieved by adopting the same set of scaling laws that can be synthesized as follows: power-to-volume-scaling, full-height, time-preserving. Reference is made to both single phase and two-phase natural circulation. In order to evaluate the NCP of the mentioned facilities, significant information comes from the analysis of the trend of the core inlet mass flowrate and the primary loop mass inventory. The flowrate and the residual masses have been normalised taking into account of the volume of each facility and of the power level (typically n times 1% of the nominal core power, where n ranges between 1 and 5) utilized in the selected experiment. Four main flow patterns were characterized depending upon the value of the mass inventory of the primary loop (see also Ref. [4]):

a) single phase NC with no void in the primary system excluding the pressuriser and the upper head;

b) stable co-current two-phase NC with mass flowrate increasing when decreasing primary system fluid inventory;

c) unstable two-phase NC and occurrence of siphon condensation (Ref. [5]);

d) stable reflux condensation with liquid flowing countercurrent to steam in the hot legs: flowrate is sufficient to remove core power till loop mass inventory achieves values as low as 30-40% of the nominal values.

A Natural Circulation Flow Map (NCFM) has been obtained from the envelope of experimental data: this constituted a significant result of the research. The NCFM has been used for evaluating the NCP of the Pactel and the RD-14M facilities, simulators of the WWER-440 and of the CANDU reactors, respectively. The application of thermalhydraulic system codes allowed the extension of the use of the map to the study of NCP in operating reactors and in reactors under design. Westinghouse PWR (equipped with U-Tubes Steam Generators, UTSG), Babcock & Wilcox PWR (equipped with Once Through Steam Generators, OTSG), WWER-1000 (equipped with Horizontal Tubes Steam Generators, HTSG), EPR (designed by a Siemens-Framatome Consortium, equipped with UTSG), AP — 600 (passive type reactor designed by Westinghouse and equipped with UTSG) and EP-1000 (a passive type reactor under design by Westinghouse equipped with UTSG) have been considered. Significant results are discussed in the first part of the paper, also providing judgements in relation to the NCP of the considered systems.

The second part of the paper deals with the use of the NCFM for the analysis of the NCP of an existing UTSG PWR outside the design and the operating limits: in those conditions the system may not be classified any more as a PWR. In order to analyze the new system performance, a qualified thermalhydraulic system computer code and a qualified nodalisation (as far as possible) are adopted, e. g. Refs. [6] and [7]. The NCFM is used as a reference to confirm the qualification level of the numerical tools and to give an idea of the distance between the considered thermohydraulic configuration and the values of the relevant NC parameters that are consistent with the current design limits. Therefore, the main purposes of the paper can be synthesized as:

• to characterize the NC as a system phenomenon in PWR;

• to give an overview of the NCFM based upon experimental data and of its application;

• to show the possibility for PWR in the current geometrical layout to operate in NC at 100% core power.

The achieved results can be useful in the framework of the design of advanced Nuclear Power Plants.

The scaling laws presented in section 3 are tested against experimental data for both steady state and stability behaviour.

5.1. Steady state behaviour of single-phase natural circulation

The steady state scaling laws were tested with data obtained from simple low pressure and high pressure loops. The data included both in-house experimental data and those compiled from literature.

5.1.1. In-house data

For testing of the scaling laws, experiments were carried out in three uniform diameter rectangular loops with horizontal heater and horizontal cooler (Vijayan et al. (1992)). These loops (see Fig. 3) had the same length and height but different internal diameters. These experiments helped to establish the adequacy of Grm/NG as the scaling parameter for the steady state flow as the data from the three loops could be expressed in the form of Eq. (14) with the same value of C and r. Subsequently, experiments were conducted to study the effect of the orientation of the heater and the cooler as different types of nuclear reactors have different orientations of the heat source (core) and the heat sink (steam generator). For example, PHWRs have horizontal core and vertical steam generators, PWRs have vertical

core and vertical steam generators and WWERs have vertical core and horizontal steam generators. In view of this, the heater and cooler orientations studied included the horizontal heater horizontal cooler (HHHC), horizontal heater vertical cooler (HHVC), Vertical heater horizontal cooler (VHHC) and vertical heater vertical cooler (VHVC). Further details of the experimental loop (see Fig. 4) and data generated can be obtained from Bade (2000) and Vijayan et al. (2000). The steady state data collected for different orientations of the heater and cooler are plotted without and with consideration of local pressure losses in Figs. 5a and b respectively. The experimental data is observed to be very close to the theoretical correlations for all orientations of the heater and cooler confirming the validity of the correlations (15) and (16). Considering the local pressure losses improves the agreement with the theoretical correlations (see Fig. 5b).

17 80 17

FIG. 3. (left) Uniform diameter loops with different external diameters & identical lengths. FIG. 4. (right) Experimental loop to study the effect of orientation.

Subsequent to this, experiments were carried out in the high pressure nonuniform diameter figure-of-eight loop FISBE (Facility for the Integral System Behaviour Experiments, see Fig. 6), which simulates the Narora Atomic Power Station. Steady state data from FISBE are compared with the present correlation in Fig. 7, which shows good agreement. It may be noted that for these tests the loop was an all pipe loop with a tubular heater section, a single U-tube steam generator and pumps replaced by pipe sections. Since the SG used only one U — tube, more than 95% of the loop hydraulic resistance was due to this. Although the loop had several elbows, bends, Tees and other fittings, these were mainly concentrated in the large diameter pipe sections. Hence the contribution of the local pressure losses to the total hydraulic resistance became negligible giving good agreement with the turbulent flow correlation.

W. J.M. De Kruijf, T. H.J. J. Van Der Hagen, R. Zboray, A. Manera

Interfaculty Reactor Institute, Delft University of Technology

R. F. Mudde

Kramers Laboratorium voor Fysische Technologic, Delft University of Technology Netherlands

Abstract. At the Delft University of Technology two thermohydraulic test facilities are being used to study the characteristics of Boiling Water Reactors (BWRs) with natural circulation core cooling. The focus of the research is on the stability characteristics of the system. DESIRE is a test facility with freon-12 as scaling fluid in which one fuel bundle of a natural-circulation BWR is simulated. The neutronic feedback can be simulated artificially. DESIRE is used to study the stability of the system at nominal and beyond nominal conditions. CIRCUS is a full-height facility with water, consisting of four parallel fuel channels and four parallel bypass channels with a common riser or with parallel riser sections. It is used to study the start-up characteristics of a natural-circulation BWR at low pressures and low power. In this paper a description of both facilities is given and the research items are presented.

1. INTRODUCTION

Natural circulation is a key item in the design of innovative natural-circulation-cooled Boiling Water Reactors (BWRs). Instead of using recirculation pumps to provide the cooling flow for the core, the core flow is driven by the density differences between the two-phase mixture in the core and the essentially single-phase flow in the downcomer. The natural-circulation core flow is enhanced by using a riser section on top of the core. Because the core flow cannot be controlled by means of a pump, the recirculation core flow is an internal variable of the system. Natural circulation has been used in the early stages of reactor development. Both the Experimental Boiling Water Reactor (EBWR) [1] and the Vallecitos Boiling Water Reactor [2] could be used with natural circulation. Later, the Humboldt Bay atomic unit and the Dodewaard plant have been operated as commercial BWR/1 plants with natural circulation. More recently the interest in natural circulation as a possibility for the core cooling has been renewed. This can be seen in the design of the Simplified Boiling Water Reactor, based on which new designs such as the ESBWR have been proposed [3]. The trend in these designs with respect to the reactor core is towards larger cores and higher power, combined with larger risers to enhance the natural-circulation core flow.

Because the core flow responds to changes in power the stability of a natural-circulation BWR is somewhat different from the stability of a forced-circulation BWR. Therefore, the stability of a natural-circulation BWR requires special attention. It has been shown that two different instability types exist for such a reactor, denoted by type-I and type-II [4]. Type-I oscillations are typical for natural-circulation BWRs and are driven by the gravitational pressure drop over the core and riser. Type-II oscillations are driven by the interplay between single-phase and two-phase friction in the core. This division in different types is not sharp. The transition from one type to the other occurs gradually. Although the character of both types of oscillations is different one could describe both of them as density-wave oscillations.

The type-I oscillations may occur during the start-up phase of the reactor, because this unstable region broadens as the pressure decreases and because it is associated with low — power operating conditions. The flashing of water in the riser at low pressures induces this type of oscillation. The neutronic feedback is not important for this type of instability. The core region is essentially single-phase and thus the power oscillations will be small. Moreover, at such a low level power oscillations cannot cause any damage to the fuel. However, large flow oscillations should be avoided in view of their possible effect on structural materials. Different types of oscillations could be possible in view of the different possible configurations of parallel fuel bundles with common or parallel riser sections.

Type-II oscillations may occur during high-power/low-flow conditions in both natural — circulation or forced-circulation BWRs. Because the density in the core region fluctuates, the nuclear feedback is essential in the analysis of these types of oscillations. A division is made between core-wide, regional, and local oscillations. In core-wide oscillations the total power in the core will vary and all fuel bundles oscillate in-phase. This mode is favoured neutronically because of the subcriticality of the regional modes. In regional oscillations the total power will remain nearly constant and the power distribution will vary periodically. This mode is favoured thermohydraulically because the total flow will be nearly constant. A third type of oscillation is a local thermohydraulically unstable fuel bundle which induces small changes in the power and the power distribution.

The pressure dynamics and the feedwater (inlet subcooling) dynamics of the system should also be taken into account; this might especially be important for type-I oscillations in natural — circulation BWRS for which water flashes into large volumes of steam giving rise to large flow oscillations.

At the Delft University of Technology two thermohydraulic test facilities are being used to study the instability types in natural-circulation BWRs: DESIRE for type-II oscillations and CIRCUS for type-I oscillations. A description of DESIRE is given in Section 2, a description of CIRCUS is given in Section 3. Both facilities can be used to produce valuable experimental data needed for further model development in the system codes applied for nuclear power plant analyses[5],